article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
we focus on the nonperturbative dynamics of the strong interaction that can be learned thanks to the clean separation of the electroweak and qcd parts in the hadronic decays of the @xmath2 lepton .
advances in the study of the two meson processes have been presented in this workshop @xcite,@xcite .
+ the decay amplitude for the considered decays may be written as : @xmath3 where our lack of knowledge of the precise hadronization mechanism is encoded in the hadronic vector , @xmath4 : @xmath5 symmetries help us to decompose @xmath4 depending on the number of final - state pseudoscalar ( @xmath6 ) mesons , @xmath7 . for three mesons in the final state ,
this reads : @xmath8 and @xmath9 @xmath10 , @xmath11 , correspond to the axial - vector current ( @xmath12 ) while @xmath13 drives the vector current ( @xmath14 ) .
the form factors @xmath15 and @xmath16 have a transverse structure in the total hadron momenta , @xmath17 , and drive a @xmath18 transition .
the scalar form factor , @xmath19 , vanishes with the mass of the goldstone bosons ( chiral limit ) and , accordingly , gives a tiny contribution .
this is as far as we can go without model assumptions , that is , we are still not able to derive the @xmath10 from qcd .
our claim is that the most adequate way out is the use of a phenomenologically motivated theory that , in the energy region spanned by tau decays , resembles qcd as much as possible .
+ in any case , the approximate symmetries of qcd are useful .
they rule what is the theory to be used in its very low - energy domain and guide the construction of higher - energy theories .
weinberg s theorem @xcite leads to conclude that a description in terms of the relevant degrees of freedom in the range of energies spanned by tau decays that incorporates the ( approximate ) symmetries of qcd therein will be valid .
some phenomenological wisdom would tell that this procedure will be most adequate , indeed .
well below the @xmath20(770 ) mass , @xmath21pt @xcite , @xcite is the effective theory of qcd .
being @xmath22 gev , one needs to include the resonances as explicit degrees of freedom . according to vector meson dominance @xcite , those of spin
one would prevail .
this is a result derived within resonance chiral theory ( r@xmath21 t ) @xcite worked in the large-@xmath0 limit of qcd .
r@xmath21 t is built upon the approximate chiral symmetry of low - energy qcd for the lightest pseudoscalar mesons and unitary symmetry for the resonances .
the perturbative expansion of r@xmath21 t is guided by its large-@xmath0 limit @xcite . at lo in 1/@xmath0 ,
meson dynamics is described by tree level diagrams obtained from an effective local lagrangian including the interactions among an infinite number of stable resonances .
we include finite resonance widths ( a nlo effect in this expansion ) within our framework @xcite .
we are also departing from the @xmath23 limit because we consider just one multiplet of resonances per set of quantum numbers and not the infinite tower predicted that can not be included in a model independent way .
for convenience , we choose to represent the spin - one mesons in the antisymmetric tensor formalism @xcite .
+ the relevant part of the r@xmath21 t lagrangian is @xcite , @xcite , @xcite , @xcite : @xmath24 where all couplings are real , being @xmath25 the pion decay constant in the chiral limit .
the notation is that of ref .
@xmath6 stands for the lightest pseudoscalar mesons and @xmath26 and @xmath27 for the ( axial)-vector mesons .
furthermore , all couplings in the second line are defined to be dimensionless .
the subindex of the operators stands for the kind of vertex described , i.e. , @xmath28 gives a coupling between one vector and three pseudoscalars .
for the explicit form of the operators in the last line , see @xcite , @xcite , @xcite .
+ symmetries of low - energy qcd are not enough for r@xmath21 t to describe it up to the perturbative region .
it is mandatory to match order parameters of spontaneous chiral symmetry breaking with partonic qcd related quantities .
the matching of @xmath29 and @xmath30-point green functions in the ope of qcd and in r@xmath21 t has been shown to be a fruitful procedure @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
additionally , we will demand to the vector and axial - vector form factors a brodsky - lepage - like behaviour @xcite . + while symmetry fully determines the structure of the operators , it is the qcd - ruled short - distance behaviour who restricts certain combinations of couplings rendering r@xmath21 t predictive provided we fix these few remaining parameters restoring to phenomenology .
the set of couplings determined by this procedure depends on the particular process under study .
in particular some couplings that were fitted phenomenologically in @xmath31 and @xmath32 are predicted now .
there are 24 unknown couplings in @xmath33 ( [ full_lagrangian ] ) that may appear in the calculation of three meson decays of the @xmath2 .
they are reduced when computing the feynman diagrams involved .
there only appear @xmath34 , @xmath35 , three combinations of the @xmath36 , four of the @xmath37 , two of the @xmath38 and four of the @xmath39 .
the number of free parameters has been reduced from 24 to 15 .
+ we require the form factors of the @xmath40 and @xmath41 currents into @xmath42 modes vanish at infinite transfer of momentum . as a result , we obtain constraints @xcite , @xcite among all axial - vector current couplings but @xmath43 , that are also the most general ones satisfying the demanded asymptotic behaviour in @xmath44 , studied along these lines in @xcite . proceeding analogously with the vector current form factor results in five additional restrictions @xcite . from the 24 initially free couplings in eq .
( [ full_lagrangian ] ) , only five remain free : @xmath45 , @xmath46 , @xmath47 , @xmath48 and @xmath49 . after fitting @xmath50 -using some of the relations in ref .
@xcite- only @xmath45 and @xmath48 remain unknown .
we obtained them from babar data on @xmath1 .
relating the measured isovector component of @xmath1 to the total isovector cross - section , as done in ref .
@xcite , is not trivial at all @xcite .
the issue is discussed in detail in our article @xcite .
we conclude that the measured isovector component in @xmath51 does not provide the total isovector component for the ( partially ) inclusive decay @xmath1 .
once some assumptions are used , one can employ @xmath52 to relate @xmath1 to @xmath53 .
we fitted our expressions to babar data @xcite obtaining @xmath54 .
our results for the decay widths of the considered channels are consistent with the pdg values @xcite .
to confront our predictions for the spectra with the forthcoming experimental data is a necessary task .
+ we have obtained several sets of allowed values for those quantities that were still free after the high - energy conditions were imposed @xcite .
we get two predictions : * the ratio of the vector current to all contributions is , for all charge modes : @xmath55 * the spectra of the two independent @xmath56 decay modes of the @xmath57 .
we show in figure [ k+k - pi- ] that one for @xmath58 .
the shape of the spectral function is similar for @xmath59 . in both cases
, the axial - vector current dominates the low - energy region , while the vector one peaks at higher values of @xmath60 .
we conclude by emphasizing that the phenomenological approach that we follow includes every relevant ( and known ) piece of qcd : we preserve the correct chiral limit at low energies that gives the right normalization to our form factors , we employ large-@xmath0 qcd arguments to use our theory in terms of mesons and it fulfills the high - energy conditions of the fundamental theory at the mesonic level . while we remain predictive by including only the lightest multiplet of axial - vector and vector resonances , our lagrangian can be enlarged in a systematic way to account for the contribution of higher resonance states , whose parameters may be fitted to experiment . +
this framework can easily be extended to study exclusive hadronic channels in @xmath62 collisions , as we plan to do .
once the spectra for @xmath63 are presented , our work may shed some light on the issue of the importance of isospin violating corrections when relating tau and @xmath62 hadronic cross sections .
+ our study can be very useful for the b - factories babar and belle , that are already analyzing huge samples with good quality data , the tau - charm factory bes - iii and a future super-@xmath64 factory ; specially , if our expressions are implemented in the currently used libraries for both types of processes , namely tauola @xcite and phokhara @xcite .
9 z. h. guo , these proceedings .
j. portols , these proceedings .
s. weinberg , .
j. gasser and h. leutwyler , , .
j. j. sakurai , currents and mesons ( university of chicago press , chicago , ill . , 1969 ) .
j. h. khn and f. wagner , .
a. pich , in proceedings of the _ tau charm factory workshop _ ed .
l. v. beers , slac ( 1989 ) . g. ecker , j. gasser , a. pich and e. de rafael , .
g. t hooft , , * 75 * , 461 ( 1974 ) .
e. witten , .
d. gmez dumm , a. pich and j. portols , .
g. ecker , j. gasser , h. leutwyler , a. pich and e. de rafael , .
v. cirigliano , g. ecker , m. eidemller , a. pich and j. portols , .
p. d. ruiz - femena , a. pich and j. portols , .
d. gmez - dumm , p. roig , a. pich , j. portols , to appear .
g. amors , s. noguera and j. portols , .
m. knecht and a. nyffeler , .
v. cirigliano , g. ecker , m. eidemller , r. kaiser , a. pich and j. portols , .
g. p. lepage and s. j. brodsky , .
p. roig , aip conf .
* 964 * ( 2007 ) 40 [ arxiv:0709.3734 [ hep - ph ] ] .
d. gmez dumm , a. pich and j. portols , .
r. barate _
[ the aleph collaboration ] .
roig , arxiv:0810.1255 [ hep - ph ] .
b. aubert _ et al . _
[ babar collaboration ] , phys .
d * 77 * ( 2008 ) 092002 .
yao et al .
, journal of physics g * 33 * , 1 ( 2006 ) .
s. jadach , z. was , r. decker and j. h. khn , comput .
commun .
* 76 * ( 1993 ) 361 .
g. rodrigo , h. czyz , j. h. khn and m. szopa , .
|
we present our study of the hadronization structure of both vector and axial - vector currents leading to decays of the tau into two kaons and a pion .
the cornerstones of our framework are the large-@xmath0 limit of qcd , the chiral structure exhibited at low energies and the proper asymptotic behaviour , ruled by qcd , that is demanded to the associated form factors .
the couplings of the theory are mostly constrained by this procedure and upon the analysis of babar data on @xmath1 we are able to predict the hadronic spectra .
|
dwarf spheroidals ( dsphs ) are the faintest galaxies known . they are gas poor and have pressure supported stellar components ( mateo 1998 ) . among them some stopped forming stars about 10 gyr ago and other have extended star formation histories ( hernandez et al .
they are typically clustered around the largest galaxy in a group . both mass loss from supernovae winds ( dekel & silk 1986 ) and environmental mechanisms like tidal and ram pressure stipping ( einasto et al . 1974 )
have been invoked to explain their properties .
suppression of gas accretion and/or photoevaporation during the reionization epoch likely played a role as well ( bullock et al . 2000 ) . in this paper
we describe the results of n - body / sph simulations of disky dwarf galaxies subject to the combined action of the main environmental mechanisms , tides , ram pressure , and both a cosmic and local ultraviolet radiation field .
the model galaxies and the choice of the orbits are consistent with the predictions of @xmath2cdm models ( mayer et al .
2002 , 2005,hereafter ma05 ) .
these simulations allow for the first time to explore the overall effect of the environment .
knowing the present and past mass of dsphs is crucial in order to compute the effects of both environmental and internal mechanisms that might affect their evolution .
moreover , the mass of dsphs has important implications on the missing satellites problem ( moore et al .
1999 ) , namely the fact that cdm predicts an order of magnitude more satellites than observed around the milky way or m31 .
white ( 2000 ) has suggested that this problem could be alleviated if the peak circular velocities of the halos of dsphs , and thus their masses , are higher than those obtained from their central velocity dispersion under the assumption of a flat ( isothermal ) rotation curve .
this is expected if dsphs sit at the center of an extended halo with an non - isothermal ( e.g. nfw ) profile , thus having circular velocity profiles that keep rising well beyond the apparent tidal radii of the galaxies .
stoehr et al .
( 2002 ) fitted the kinematics of dsphs inside subhaloes of a cosmological @xmath2cdm simulation using jeans modeling and king profiles for the light distribution of dsphs and found that these galaxies live in halos having a peak circular velocity @xmath3 km / s .
the halo density profiles employed in their analysis are much flatter than an nfw profile at the scale of the core radius of dsphs . 2truecm km / s ( thick lines)and @xmath4 km / s ( thin lines ) .
the solid lines represent the initial conditions while the short - dashed show the curves after 3 orbits .
the thick long - dashed line shows the curve for the same dwarf model with @xmath5 km / s on a different orbit .
pericenter distances for the various orbits are indicated in the plot ( see ma05).,title="fig:",width=336,height=336 ] kazantzidis et al .
( 2004 ) have studied the tidal disruption of individual cdm satellites with as many as @xmath6 particles and of subhaloes in a hi - res @xmath2cdm simulation .
they repeated the analysis of stoehr et al .
( 2002 ) and found that the observed velocity dispersions of draco and fornax can only be reproduced in subhaloes with @xmath7 km / s . in their simulations halos
maintain cuspy profiles down to the force resolution , which is smaller than the core radius of dsphs ( @xmath8 pc ) . as a result the rotation curves rise more steeply than those of stoehr et al .
( 2002 ) , explaining the discrepancy .
the masses of the tidally truncated subhaloes of kazantzidis et al .
( 2004 ) are in the range @xmath9 , in agreement with more recent analysis based on newer data on stellar velocities ( e.g. wilkinson et al .
( 2001 ) ) .
these masses are still quite high , yielding @xmath10 for draco . however peak circular velocities are at most a factor of 2 higher than those used in moore et al .
( 1999 ) , with the consequence that little changes for the missing satellite problem .
the initial @xmath11 of dsphs was larger than 30 km / s since mass loss produced by tidal shocks can lower @xmath11 ( ghigna et al .
kravtsov et al .
( 2004 ) ) , using hi - res cosmological simulations without baryons , find that @xmath11 can drop by a factor of 2 in 10 gyr or so .
mayer et al . (
2002 ) and ma05 always find a smaller decrease in @xmath11 in their large set of sph simulations that follow the evolution of satellites with baryonic disks embedded in cdm haloes ( figure 1 ) .
the different results might be due to the fact that some of the satellites in kravtsov et al .
( 2004 ) have orbits with pericenters smaller than those considered by ma05 .
however , the inclusion of baryons also makes the satellites more robust to the tidal disruption of the inner region , which is what matters for the evolution of @xmath11 . whether or not the baryonic disk becomes bar unstable
also is important , since bar - driven inflows can deepen the central potential of the dwarf ( kazantzidis , mayer et al .
, in preparation ) .
high - resolution cosmological simulations with hydrodynamics will eventually be able to quantify the overall significance of these effects . in any case , the results of ma05 suggest that the initial @xmath11 of dsphs was in the range @xmath12 km / s .
this has important implications .
first , it means that photoevaporation did not play a major role during the early evolution of satellites at high @xmath13 since it is only effective for @xmath14 km / s ( susa & umemura 2004 ; shaviv & dekel 2003 ) .
this is consistent with the lack of a clear signature of the reionization epoch in the star formation histories of dsphs ( grebel & gallagher 2004 ) .
second , blow - out of the gas of dsphs due to supernovae winds can not have occurred since it requires @xmath15 km / s ( e.g. maclow & ferrara 1999 ) .
2truecm as a function of the number of orbits for the satellites in the cosmological hydrodynamical simulation of governato , mayer et al .
( 2004).@xmath16 is measured within the half mass radius of the dwarfs .
the mean number of orbits for each of the three groups of dwarfs appears on the horizontal axis.,title="fig:",width=307,height=307 ]
aside from the different gas content , three other facts have to be considered when comparing dsphs and gas - rich , similarly faint dwarf galaxies known as dwarf irregulars ( dirrs ) .
first , except for the faintest among them all dirrs in the local group and nearby groups exhibit substantial rotation while dsphs do not .
second , stellar profiles in the two classes of galaxies are similarly close to exponential .
third , a morphology density relation exists such that dsphs are clustered around the primary galaxies while dirrs are found at much larger distances from them .
interestingly , karachentsev ( these proceedings ) finds that the fraction of dsphs over the total number of dwarfs in several nearby groups ( including the local group ) decreases drastically at distances larger than about @xmath17 kpc from the primary galaxies , these being comparable to the virial radius of haloes hosting bright spiral galaxies in @xmath2cdm models .
this clearly suggests that only dsphs are bound satellites of the primaries and thus the environment must be playing a crucial role in differentiating dirrs from dsphs .
of course there are outliers like the local group dsphs cetus and tucana , located at more than 500 kpc from , respectively , m31 and the mw , but this is seen also in cosmological simulations , where a few satellites on very plunging orbits can have apocenters exceeding the virial radius of the primary ( ghigna et al .
1998 ) .
mayer et al .
( 2001a , b ) and ma05 have shown that repeated tidal shocks at pericenters of their orbits within the halo of a massive spirals can transform disky dwarfs into objects resembling dsphs .
the timescale of the transformation is a few orbital times ( several gyr ) .
the mechanism behind the transformation has to do with non - axisymmetric instabilities of stellar disks .
first , tidal shocks induce strong bar instabilities in otherwise stable , light disks resembling those of present - day dirrs .
second , the bar buckles due to the amplification of vertical bending modes and turns into a spheroidal component .
finally , tidal heating / tidal mass loss thicken / remove the disk outside the buckled bar .
we stress that stellar mass loss can be minimal when the disk sits deep in the potential well of a massive halo ( see the gr8 run in mayer et al .
2001b ) and yet the transformation occurs .
the remnants have nearly exponential profiles ; the brightest among them , those coming from progenitors having relatively massive disks ( @xmath18 ) , develop a central steepening of the profile due to the particularly strong bar , and overall their profile resembles that of the bright dwarf elliptical satellites of m31 , like ngc205 .
most importantly , the bar sheds angular momentum outwards and as a result the systems end up with low rotation , with a typical final @xmath19 .
the transformation of course needs the dwarf to be on a bound orbit ; for the mean apocenter / pericenter ratio found in cosmological simulations , @xmath20 , mayer et al .
( 2001b ) determine that an apocenter distance comparable to the distance of leo i ( 250 kpc ) , the farthest dsph satellite of the mw , is a limiting case for producing a dsph from a disky dwarf in less than 10 gyr ( tucana and cetus should thus be on radial orbits ) . in brief
this `` tidal stirring '' accounts for most of the similarities and differences between dirrs and dsphs , including the existence of the morphology density relation , by postulating progenitors of dsphs with light , low surface brightness disks embedded in massive halos like dirrs .
this does not mean that such progenitors were identical to present - day dirrs .
since they formed at high @xmath13 they likely had assembly histories quite different from present - day dirrs , and thus their stellar populations and metallicities were probably different .
the missing piece now is how to place this model in the context of hierarchical structure formation .
unfortunately , to date cosmological simulations with hydrodynamics have not allowed a robust analysis of the structural evolution of satellites due to their limited resolution ( typically the force resolution is 500 - 1 kpc , so the tiniest dsphs like draco are not resolved at all ) .
searching for evidence of tidal stirring in these simulations is thus quite hard , but it can be done for a few well resolved dwarfs .
we selected the brightest satellites ( @xmath21 ) of the large spiral galaxy in the cosmological hydro simulation of governato , mayer et al .
( 2004 , hereafter gm04 ) as well as similarly bright dwarfs outside the virial radius of the same galaxy .
figure 2 shows their @xmath16 within the effective radius versus the number of orbits performed within the main system .
clearly @xmath16 correlates well with the number of orbits as expected within the tidal stirring scenario .
we also found that the shape of the stellar components of the dwarfs goes from more disky to more spheroidal with increasing number of orbits .
the absolute values of the @xmath16 have to be taken with caution because numerical two - body heating ( e.g. mayer 2004 ) is certainly an issue for these objects ( halos have only a few thousand particles at this scale ) .
however the trend is evident .
2.2truecm , small symbols are for those with a disk gas mass fraction of @xmath22 ( see section 4 ) .
red is used for adiabatic runs , blue for runs with radiative cooling , black for runs with radiative cooling+cosmic uv background and yellow for runs with cooling , uv and star formation .
the cross and star symbols are used for two runs with radiative cooling and a local uv radiation field yielding a maximum heating rate , respectively , 5 times and 10 times higher than that associated with the cosmic background.,title="fig:",width=3,height=3 ]
mayer et al . ( 2001b ) showed that tidal stripping alone could not produce the low gas fractions found in dsphs starting from a gas - rich disky dwarf .
gas consumption by star formation does not change the results significantly for mean star formation rates consistent with those inferred for dsphs ( hernandez et al .
ma05 study the combined effect of ram pressure and tidal stripping .
they construct two - component models for the milky way halo in which a dark matter halo consistent with the results of @xmath2cdm simulations has an embedded diffuse gaseous component with a temperature of @xmath23 k and a density of about @xmath24 atoms/@xmath25 at 30 kpc from the center , consistent with the values inferred from ovi absorption measurements and the existence of the magellanic stream ( sembach et al .
dwarf galaxies are placed on eccentric orbits with pericenters of 30 or 50 kpc . in these simulations
gas - rich disky dwarfs ( @xmath26 ) embedded in massive dark haloes typically lose 90% of their gas content .
in fact ram pressure increases by a factor up to 10 the amount of stripped gas mass compared to the case in which only tides are included .
ram pressure strongly depends on the depth of the potential well of the dwarfs . while for dwarfs with
@xmath27 km / s most of the gas content is easily removed , for more massive dwarfs the end result depends a lot on the orbit and on the temperature evolution of their gas ( see figure 3 ) .
the pericenter distance sets the strength of the ram pressure force mostly through the dependence on the orbital speed of the galaxy at pericenter .
the temperature evolution determines whether the gas component stays extended or becomes concentrated in the central , deeper part of the potential well as a result of tidally induced bar - driven inflows .
compression from the outer medium heats the gas .
the gas is rapidly heated to @xmath28 k , where the cooling function peaks ( the initial temperature is @xmath29 k ) at which point it cools radiatively to @xmath30 k in a fraction of a dynamical time ; this cold gas easily sinks towards the center due to the torque exerted by the bar and can not be removed by ram pressure . instead , if compressional heating is not radiated away , as when the gas evolves adiabatically , the increased pressure opposes the bar - driven inflow , keeping more gas at larger radii where the it is easily stripped .
for stripping to be very effective for dwarfs with initial @xmath31 km / s it is sufficient that the temperature of the gas is kept above @xmath30 k ( the virial temperature of these halos is @xmath32 k ) .
this requires some heating source to counteract radiative cooling .
ma05 find that the ( uniform ) cosmic uv background at @xmath33 ( haardt & madau 1996 ) can achieve this .
they find that if the progenitors of draco or ursa minor fell into the milky way at @xmath1 then ram pressure combined with tides was able to remove their entire gas content in a couple of orbits .
this translates into a timescale of about 2 - 3 gyr for a reasonable orbital time ( namely one consistent with the current orbital distance ) . in this scenario
the observed truncation of the star formation in draco and ursa minor more than 10 gyr ago occurred as a consequence of the infall of these galaxies into the milky way halo .
draco would have formed most of its stars before infall since most of the gas becomes ionized while approaching pericenter for the first time .
however , ma05 also find that the exact gas fraction of gas that becomes ionized and then stripped is quite sensitive on the initial gas density ( figure 3 ) . starting from initial conditions having an exponential disk comprising about @xmath34 of the virial mass(comparable to disk mass fractions of dirrs ) , a disk with @xmath35 gas fraction becomes entirely ionized while one with @xmath36 gas fraction remains neutral in the center and is only marginally stripped .
the mass of neutral gas could produce a stellar mass comparable to the present - day luminosity of draco assuming star formation goes on for another few gyr .
this would yield a star formation history more extended than that of draco , possibly closer to that of carina and fornax . on the other end a progenitor with a mostly gaseous disk that undergoes complete
stripping is an attractive case since it would naturally produce the large @xmath37 of draco without the need of invoking an extremely low disk mass fraction in the initial conditions ( ma05 ) .
one obvious solution is to make the case for a higher heating and photoionization rate ; no dramatic increase of the rates would be needed since the neutral gas has a temperature only 20% lower than @xmath30 k in the gas rich models .
the local uv flux coming from the primary galaxy during a phase of intense star formation could have been stronger than the average cosmic background .
mashchenko et al .
( 2004 ) calculate that the fuv flux of m31 derived from its @xmath38 luminosity is higher than the present - day value of the cosmic uv background out to 10 - 20 kpc from its center .
current starbursts have fuv luminosities @xmath39 higher than the mw and m31 , in the range @xmath40 erg / s ( leitherer et al .
according to cosmological simulations at @xmath41 the milky way was hosting a major central starburst as the bulge was being assembled from a big merging event ( gm04 ) .
a starburst is also predicted by models of the star formation history in the galactic bulge , which yield peak star formation rates of up to @xmath42/yr ( elmegreen 1999 ) , comparable to the most spectacular among present - day starbursts .
lower quantities of dust expected given the lower metallicity of the gas at high redshift imply an escape fraction of ultraviolet photons higher than that in present - day starbursts .
but let us be conservative and just assume that the fuv luminosity of the bulge was comparable to that of a major present - day starburst ( @xmath43 erg / s ) .
under this hypothesis one obtains that at 30 kpc from the bulge the heating rate was 10 times higher than that associated with the metagalactic uv background at @xmath41 .
we have run new simulations including such a local , time - dependent uv background ( the intensity of the flux is modulated by the orbital distance of the dwarf ) and found that ram pressure stripping at the first pericenter passage can be greatly enhanced for a gas - rich dwarf . even the most dark matter dominated among our models , with @xmath4440 km / s , loses more than 90% of its gas along the first orbit and the remaining gas is completely ionized .
the ionized gas is then stripped along the second orbit ( figure 3 ) .
this simulation reproduces naturally both the early truncation of the star formation in draco and its present - day high @xmath37 ratio ( @xmath45 ) .
a coherent picture is emerging .
dwarf spheroidals live in tidally truncated halos with moderate masses , @xmath9 , yet this implies @xmath10 in some of them ( e.g. draco and ursa minor ) .
however , since simulations show that most of the dark matter halo was stripped as the dwarfs entered the primary halos , their original dark matter content was higher , resulting in an initial @xmath46 km / s .
in fact tidal stripping can remove mass even quite close to the center , reducing their @xmath11 by @xmath47% ( see figure 1 ) .
the tidal interaction with the massive halos of the primaries is also responsible for the morphological transformation of small disks into spheroidal systems .
gas stripping instead is the result of the combination of tidal and ram pressure stripping . however a heating source is required in order to keep the gas extended and allow efficient stripping for dwarfs with @xmath46 km / s .
the more attractive scenario is one in which the heating was provided by the cosmic uv background at high redshift aided by an even stronger local fuv field produced by the primary galaxies during an epoch of intense star formation . according to galaxy formation models
this epoch should coincide with the time of bulge formation , at @xmath48 .
this would explain the sudden truncation of star formation that occurred in draco and ursa minor about 10 gyr ago .
dwarfs that entered the milky way halo later would have kept some gas for at least one or two orbits and thus undergo extended star formation .
our simulations show that the remaining gas falls towards the center as a result of bar driven inflows at pericenter passages .
the central gas density there increases and a burst of star formation occurs , as shown in mayer et al . ( 2001b ) .
this explains the bursty star formation histories of carina and fornax .
ram pressure and tidal stripping are reduced significantly at more recent epochs for two reasons ; the first is that in @xmath2cdm models orbits of satellites have on average larger pericenters and longer orbital times at lower redshifts ( see the cosmological simulations of ghigna et al .
( 1998 ) for example ) , the second is that both the local and cosmic uv background fade by orders of magnitude at @xmath49 . in summary
we predict that dsphs with extended or truncated star formation histories come from similar progenitors that fell into the primary halos at different epochs .
one variable whose evolution with time is not well known is the density of the hot gaseous corona of the milky way which sets in part the strength of ram pressure . however
new simulations of galaxy formation currently in progress suggest that the density of such halo was higher at higher redshift , which would reinforce our scenario .
i thank stelios kazantzidis , chiara mastropietro , ben moore and fabio governato for the many stimulating discussions on dwarf galaxies and cristiano porciani for his help on issues regarding the effects of the ionizing radiation .
governato , f. , mayer , l. , wadsley , j. , gardner , j.p . ,
willman , b. , hayashi , e. , quinn , t. , stadel , j. , & lake , g. , 2004 , apj , 607 , 688 grebel , e.k .
& gallagher , j.s . , 2004 , apj , 610 , l89 haardt , f. , & madau , p. , 1996 ,
apj , 461 , 20 hernandez , x. , gilmore , g. , & valls - gabaud , d. 2000 , mnras , 317 , 831 kazantzidis , s. , mayer , l. , mastropietro , c. , diemand , j. , stadel , j. , & moore , b. , 2004 , apj , 608 , 663 kravtsov , a.v . ,
gnedin , o.y . , & klypin , a.a .
2004 , apj , 609 , 482 leitherer , c. , i - hui , l. , calzetti , d. , & heckman , t.m . , 2002 , apj , 303 mac low , m. , & ferrara , a. , 1999 , apj , 513 , 142 mashchenko , s. , carignan , c. , & bouchard a. , 2004 , mnras , 352 , 168 mateo , m. , 1998 , ara&a , 36 , 435 mayer , l. , governato , f. , colpi , m. , moore , b. , quinn , t. , wadsley , j. , stadel , j. , lake g. 2001a , apj , 547 , l123 mayer , l. , governato , f. , colpi , m. , moore , b. , quinn , t. , wadsley , j. , stadel , j. , lake g. 2001b , 559 , 754 mayer , l. , moore , b. , quinn , t. , governato , f. , & stadel , j. 2002 , mnras , 336 , 119 mayer , l. , mastropietro , c. , wadsley , j. , moore , b. , & stadel , j. , submitted to mnras , astro - ph/0504277 moore , b. , ghigna , s. , governato , f. , lake , g. , quinn , t. , stadel , j. , & tozzi , p. , 1999 ,
apj , 524 , l19 sembach , k. r. et al .
2003 , apjs , 146 , 165 stoehr , f. , white , s.d.m .
, tormen , g. , & springel , v. , 2002 , mnras , 335 , l84 susa , h. , & umemura , m. , 2004 apj , 600 , 1 wilkinson , m. , kleyna , j.t . , evans , w.n . , gilmore , g.f . ,
irwin , m.j . , & grebel , e.k .
, 2004 , apj , 611 , l21
|
we discuss the role of environmental mechanisms in the evolution of dwarf galaxy satellites using high - resolution n - body+sph simulations that include simultaneously tidal forces , ram pressure and heating from ionizing radiation fields .
tidally induced bar - buckling instabilities can transform a rotating disky dwarf into a pressure supported spheroidal .
efficient gas removal requires instead a combination of tidal mass loss and ram pressure stripping in a diffuse gaseous corona around the primary system .
the efficiency of ram pressure depends strongly on how extended the gas remains during the evolution .
bar driven inflows that tend to drive the gas to the bottom of the potential well can be opposed by the heating from external radiation fields .
we show that even fairly massive dwarfs ( @xmath0 30 km / s ) would be stripped of their gas over a few gyr if they enter the milky way halo at @xmath1 thanks to the effect of the cosmic uv background .
gas mass loss can be much faster , occurring in less than 1 gyr , if dwarf satellites have their first close approach with the primary at the epoch of bulge formation .
indeed at that time the primary galaxy should have a fuv luminosity comparable to that of major present - day starbursts , resulting in a local uv field even more intense than the cosmic background .
|
quarks and leptons are believed to be the fundamental particles of matter yet their nature is still far from being understood .
one of the exciting puzzles is a formula involving the masses of the three leptons , discovered by yoshio koide @xcite : @xmath0 see table 1 for the corresponding numerical values ( from @xcite ) .
[ tbl:1 ] .lepton masses .
[ cols= " < , < , < , < , < " , ] * 4 . coda . * whether this intriguing connection with geometry will contribute to an understanding of the masses of leptons remains an interesting question and requires further investigation .
the analogy described above may turn out to be merely superficial , but given the current state of understanding about the matter , any interesting structural parallels are worthy of our consideration in the effort of reconstructing deeper patterns .
i am grateful to philip feinsilver for his encouraging interest and priceless comments .
j. kocik , a matrix theorem on circle configuration ( _ arxiv:0706.0372v2)_. y. koide , fermion - boson two body model of quarks and leptons and cabibbo mixing , _ lett .
nuovo cimento _ * 34 * ( 1982 ) 201 .
y. koide , _ mod .
_ a5 ( 1990 ) 2319
. y. koide , challenge to the mystery of the charged lepton mass formula ( 2005 ) , hep - ph/0506247 . _ _
|
a remarkable formal similarity between koide s lepton mass formula and a generalized descartes circle formula is reported . +
* keywords : * lepton mass , koide formula , descartes circle theorem , geometry .
|
micro - electro - mechanical systems ( mems ) technology has been developed rapidly in recent years.@xcite~@xcite the characteristic length scale of mems is typically the order of microns , and the ratio of the mean free path to the characteristic dimension ( i.e. knudsen number @xmath2 ) can not be negligible .
the dynamics associated with microchannels can thereby exhibit rarefied phenomena and compressibility effects .
the former is the emergence of a slip velocity at the wall boundary in microchannel flows .
the latter is the significant nonlinear pressure drop of gas flowing in a long microchannel . because conducting experiments in micrometer - size is a big challenge , numerical simulations of mems become very important tools of investigation .
but the methods commonly used in simulation of mems , such as molecular dynamics ( md ) , direct simulation monte
carlo ( dsmc)@xcite and direct numerical simulation of boltzmann equation , usually requires a tremendous amount of computer time and memory .
recently , for its intrinsic kinetic nature , the lattice boltzmann method ( lbe ) has been become an attractive method for simulation of microscalar flows where both microscopic and macroscopic behaviors are coupled.@xcite form its birth nearly 20 years ago ( 1988)@xcite , the lattice boltzmann method ( lbm)@xcite~@xcite has met with significant success for the numerical simulation of a large variety of fluid flows , and has emerged as an alternative numerical technique for simulating fluid flows .
the lbe method usually solves the bhatnagar - gross - krook ( bgk ) model equation , a simplified model boltzmann equation , on a discrete lattice , which is usually named the lattice - bgk ( lbgk ) model.@xcite since theoretical connections between the lbgk model and the boltzmann equation@xcite and the preliminary link between lbgk model and the burnett - type equations@xcite were established , the lbe model can be valid for rarefied gas flow provided the mach number is small .
moreover , since the standard bgk equation is able to simulate highly non - equilibrium gas flows , the lbgk model should also be applicable to rarefied microflows theoretically.@xcite~@xcite one of the most important issues of lbgk model for simulating microchannel flows is the introduction of @xmath2 number into lb models .
the popular treatment is the construction of the relationship between @xmath2 and relaxation time @xmath0 of lb models .
_ used the explicit lbe formulation and related the nondimensional relaxation time @xmath3 to knudsen number @xmath2 as : @xmath4 for a microchannel of height @xmath5 and gas density @xmath6 .
the factor 0.5 attributes to the explicit treatment of the collision term and @xmath7 is chosen to best match the simulated mass flow rate with experiments.@xcite later , lim _ _
et al.__@xcite proposed a different relation between @xmath2 and @xmath3 for the explicit lbe formulation without the correction factor of 0.5 as in nie _
et al._. their @xmath2 for a long microchannel was defined as @xmath8 , where @xmath9 and @xmath10 are the local pressure and pressure at the outlet of the microchannel , respectively .
_ gave different definitions of @xmath3 with different constant factors among various lattice models.@xcite recently , lee _ et al .
_ proposed a definition of @xmath0 about fully implicit lbe.@xcite another one is the implementation of boundary condition .
_ used the half bounce - back rule for the slip effect at the surface .
et al . _ used a specular reflection model to generate slip effect .
et al . _ adopted the maxwellian scattering kernel to address the gas molecule and surface interactions with an accommodation coefficient @xmath7 .
proposed a wall equilibrium condition according to the assumption of rough surface on the characteristic length of gas molecules .
for the former issue , the previous research is still related to the framework of navier stokes equations .
this is because the expression of the kinetic viscosity coefficient derived from the lb equation recovering the navier stokes equations , is still adopted in the makeup of the relationship of @xmath2 and @xmath0 .
furthermore , when @xmath2 approaches zero , the boltzmann equation can be reduced to the navier stokes equation .
but all the existing lbe for microscale flow can not be reduced to the navier stokes equation because of using the relationship between @xmath2 and @xmath0 .
the previous researchers always focus on recovering the navier
stokes equations precisely from the lbgk model and ignoring the deviation of the lbgk model and boltzmann equation .
this deviation may not do any effect in simulation of macroscale flows .
however , the character scale of microflows is much smaller . in such a case , keeping the collision consistency of the lbgk model and boltzmann equation should be paid special attention .
this collision consistency requires the collision frequency of lbgk model is equal to that of bgk model equation . with the further study on the relaxation factor @xmath0 , we found that these two collision frequency are equivalent when @xmath1 and the simulation results are the most approximate .
this principle is called collision consistency of lbgk model . on the other hand ,
the boundary treatment is based on the maxwell slip model straightforwardly .
there is not any adjustable parameter in our model and is similar to the newmann boundary condition in some degree .
the implement is not related to the navier
stokes equations any more , so it can be applied in simulating microflows not only in slip regime but also in transition regime .
to begin with , the bgk equation is : @xmath11 here @xmath12 and @xmath13 are the molecular distribution function and equilibrium distribution function , respectively .
@xmath14\ ] ] where @xmath15 and @xmath16 are molecular velocity and macroscopical velocity respectively ; @xmath17 is the collision time ; @xmath18 represents the collision frequency : @xmath19 .
where @xmath20 is molecular mean velocity ; @xmath21 is molecular mean free path .
lbgk model equation is : @xmath22 here @xmath23 is particle distribution function : @xmath24\ ] ] @xmath25 is lattice discrete velocity and @xmath26 is weighing factor , @xmath17 is the collision time . lbgk equation ( [ lbgk ] ) can be further discretized in space and time .
the completely discretized form of eq .
( [ lbgk ] ) , with the time step @xmath27 and space step @xmath28 , is : @xmath29\end{aligned}\ ] ] where @xmath30 is a point in the discretized physical space , and it is worth notice that @xmath0 commonly considered as a nondimensional relaxation time , @xmath31 , represents the collision consistency of lbgk model by the analysis of the physical signification of @xmath17 and @xmath27 .
the above discrete lbgk equation is usually solved in two steps : collision step and streaming step .
so the physical process in lbgk model can be regarded as a large number of particles take place one collision , through the distance @xmath32 and the time @xmath27 .
therefore , @xmath32 and @xmath27 can be regarded as the mean free path and the collision time in the lbgk model , respectively , and @xmath33 represents the collision frequency in the lbgk model .
consequently , @xmath31 means the ratio of the collision frequency of the lbgk model to that of the bgk model . from the gas dynamics , the collision frequency @xmath17 is in direct proportion to the density , and in dependence on temperature , but independence of molecular velocity . in order to furthest approach to bgk model
, the collision frequency of the lbgk model should be equal to that of the bgk model , i.e. @xmath34 . especially in the microflows ( @xmath35 ) , the collision consistency must be much more ensured for the collision frequency of molecule in the microflows is far lower than that in the macroscopical flows ( @xmath36 ) .
it must be emphasized that lbgk model is between the bgk equation and navier stokes equations .
we formerly attached importance to recover the lbgk to navier
stokes equations in the simulation with lbgk model , but now we should pay sufficient attention to the consistency of the lbgk model with the bgk equation , especially in simulation of the microflows .
in the lbgk model , @xmath27 is defined as : @xmath37 and in the bgk model , @xmath17 is defined as : @xmath38 using eq .
( [ dt ] ) and eq .
( [ tau0 ] ) , the relaxation factor @xmath0 can be determined as : @xmath39 where @xmath40 is the number of the lattice in the characteristic length @xmath5 , @xmath41 .
regarding @xmath20 approximately equal to @xmath42@xcite , eq .
( [ tau ] ) can be simplified : @xmath43 according the collision consistency , the relation of @xmath2 to the number of the lattice is : @xmath44 from eq .
( [ n ] ) , the number of lattice should be in inverse proportion to the kn , rather than be determined randomly . under the collision consistency , eq .
( [ fievol ] ) can be simplified as : @xmath45 the distribution function @xmath23 evolved in the lattice actually is the equilibrium distribution function @xmath46 .
hence , if the velocity on the boundary can be determined , the distribution function of the boundary can also be determined . to the slip flow regime ,
the macroscopical slip - velocity on the boundary can be obtained by : @xmath47 where @xmath48 is the boundary slip - velocity ; @xmath49 is accommodation factor . using second order difference formula , eq .
( [ us ] ) . can be discretized as : @xmath50 where @xmath51 and @xmath52 are the velocity of the first and the second point adjacent the wall , respectively .
in this paper , we study microscalar gas poiseuille flows in a constant external pressure gradient in order to validate the collision consistency .
the compressibility effect becomes negligible and only the rarefaction effect is accounted for .
we study three cases with different @xmath2 number value ( @xmath2=0.02 , 0.05 , 0.1 ) , and the number of grids adopted in the simulations is chosen according to the collision consistency .
results will be compared with the analytical results and the results of more fine grids and relaxation factor chosen according to the eq .
( [ taus ] ) . ) .,scaledwidth=75.0% ] in case 1 , when @xmath2=0.1 , and the pressure gradient @xmath53 .
according to the collision consistency , relaxation factor @xmath1 , and the grid numbers are @xmath54 . from the calculating result ( see fig . [ 01 ] ) , we can see the results agree closely with the analytical results when @xmath1 , and the relative error is between @xmath55 . when @xmath56 and @xmath57 , the fine grids are increased , but the simulative relative error is very big because of departing away from the collision consistency . )
.,scaledwidth=75.0% ] in figure [ 005 ] , when @xmath58 and the pressure gradient @xmath59 , relaxation factor @xmath1 according to the collision consistency , we can see the simulating results agree closely with the analytical results , and the relative error is between @xmath60 .
we increased the grids by 1.5 and 2 times and got the relaxation factor according to eq .
( [ taus ] ) , but the relative error was still very large . ) .,scaledwidth=75.0% ] figure [ 002 ] shows the comparison of calculating results obtained in different grids and relaxation factors when @xmath61 and the pressure gradient is 2.0 . as can be seen from the graph , the results accord closely with the analytical results , but in the other two cases , the accuracy of the simulating results decreased instead of increasing although the grids increased 1.5 and 2 times , about four times less than that of when @xmath1 .
all the calculating errors of different grids and relaxation factors in these three cases are shown in table 1 . for micro poiseuille flow ,
the best calculating results are obtained since we adopted the relaxation factors in according to the collision consistency .
so the simulating precision is 3 to 4 times more than that of the other results . from table 1 , we can see , the errors are almost the same under the same relaxation factors .
this also illustrates the influence of the boundary error on the whole flow simulating error .
we now investigate the simulation for higher @xmath2 number in this subsection .
the ratio of the microchannel length @xmath62 to height @xmath5 is chosen as 100 , and a 2000@xmath6320 regular grid is applied , which has been verified by grid - dependence .
we focus on the streamwise velocity profile @xmath64 at the exit normalized by the centerline maximum velocity and the pressure deviation @xmath65 from the linear distribution @xmath66 .
note that we choose the pressure deviation @xmath67 rather than the pressure distribution itself in order to show the difference more clearly .
= 0.194).,scaledwidth=75.0% ] along the microchannel ( @xmath2=0.194).,scaledwidth=75.0% ] fig .
[ u194 ] shows the velocity profile at the outlet of the microchannel with @xmath2=0.194 .
compared with the reliable dsmc result , nie _
_ over - predicted the slip velocity obviously and the velocity distribution is more higher in nearly the whole outlet cross - section . in our simulation , both the slip velocity and exit velocity distribution are in consistent with the results of dsmc method .
moreover , pressure deviation @xmath67 is shown in fig . [ dp194 ] .
the _ qualitative _ difference with dsmc method has been founded in nie _
_ s numerical result .
our present result is very closely to the analytical solution by arkilic _
et al._. and the deviation from the dsmc result appears only in the middle part along the microchannels .
from our study , the physical meaning of the relaxation factor @xmath0 in lbgk model is : the ratio of the collision frequency in lattice system and in the bgk model .
lbgk model situates between bgk model and macro navier - stokes equations so the collision consistency needs analyzing when lbgk model is dispersed from the bgk collision equation .
this is very essential to adopting lbgk model to simulate the microflow because the continuum assumption may be break down in the microflow .
therefore we should assure the same collision frequency in lbgk model and bgk model , that is collision consistency . under collision consistency , it is simple to deal with the boundary and determine the relation factor . adopting the collision consistency principle @xmath1 , the results of the simulating poiseuille flow accord closely with the analytical results . without it ,
the precision of simulating results is three to four times less even if the greater grids are adopted .
|
the collision consistency between the bgk collision model equation and lattice - bgk ( lbgk ) model is proposed by researching the physical significance of the relaxation factor @xmath0 in lbgk model . for microscalar flow
in which the continuum hypothesis is not still satisfied , the collision consistency @xmath1 should be ensured when using the lbgk model for simulating microflows .
the results of simulating microchannel poiseuille flow with constant pressure gradient under collision consistency by using lbgk model are well consistent with the analytical solutions , and the accuracy of these results is three or four orders of magnitude higher than those that do nt satisfy the collision consistency .
|
there is increasingly strong evidence that gamma - ray bursts ( grbs ) are associated with star - forming galaxies [ 1,2,3,4 ] and occur near or in the star - forming regions of these galaxies [ 2,3,4,5,6 ] .
these associations provide indirect evidence that at least the long grbs detected by bepposax are a result of the collapse of massive stars .
the discovery of what appear to be supernova components in the afterglows of grbs 970228 [ 7,8 ] and 980326 [ 9 ] provides tantalizing direct evidence that at least some grbs are related to the deaths of massive stars , as predicted by the widely - discussed collapsar model of grbs [ 10,11,12,13,14 ] . if grbs are indeed related to the collapse of massive stars , one expects the grb rate to be approximately proportional to the star - formation rate ( sfr ) .
+ + we have calculated the limiting redshifts detectable by batse and hete-2 , and by _ swift _ , for the sixteen grbs with well - established redshifts and published peak photon number fluxes . in doing so
, we have used the peak photon number fluxes given in table 1 of [ 15 ] , taken a detection threshold of 0.2 ph s@xmath0 for batse and hete-2 and 0.04 ph s@xmath0 for _ swift _ , and set @xmath1 km s@xmath0 mpc@xmath0 , @xmath2 , and @xmath3 ( other cosmologies give similar results ) .
figure 1 displays the results .
this figure shows that batse and hete-2 would be able to detect half of these grbs out to a redshift @xmath4 and 20% of them out to a redshift @xmath5 .
_ swift _ would be able to detect half of them out to redshifts @xmath6 , and 20% of them out to a redshift @xmath7 , although it is unlikely that grbs occur at such extreme redshifts .
consequently , if grbs occur at very high ( @xmath8 redshifts ( vhrs ) , batse has probably already detected grbs at these redshifts , and hete-2 and _ swift _ should detect them as well . the soft x - ray , optical and infrared afterglows of grbs are also detectable out to vhrs .
the effects of distance and redshift tend to reduce the spectral flux in grb afterglows in a given frequency band , but time dilation tends to increase it at a fixed time of observation after the grb , since afterglow intensities tend to decrease with time .
these effects combine to produce little or no decrease in the spectral energy flux @xmath9 of grb afterglows in a given frequency band and at a fixed time of observation after the grb with increasing redshift : @xmath10 where @xmath11 is the intrinsic spectral luminosity of the grb afterglow , which we assume applies even at early times , and @xmath12 is the comoving distance to the burst .
many afterglows fade like @xmath13 , which implies that @xmath14 in the simplest afterglow model , where @xmath15 [ 16 ] .
in addition , @xmath12 increases very slowly with redshift at redshifts greater than a few . consequently , there is little or no decrease in the spectral flux of grb afterglows with increasing redshift beyond @xmath16 .
+ + in fact , in the simplest afterglow model where @xmath17 , if the afterglow declines more rapidly than @xmath18 , the spectral flux actually _ increases _ as one moves the burst to higher redshifts !
an example of this is the afterglow of grb 000131 .
its peak flux @xmath19 was in the top 5% of all batse bursts and the break energy @xmath20 in its spectrum was 164 kev , yet it occurred at a redshift @xmath21 .
we have calculated the best - fit spectral flux distribution of the afterglow of grb 000131 from [ 17 ] , as observed three days after the burst , transformed to various redshifts .
the transformation involves ( 1 ) dimming the afterglow , ( 2 ) redshifting its spectrum , ( 3 ) time dilating its light curve , and ( 4 ) extinguishing the spectrum using a model of the ly@xmath22 forest ( for details , see [ 15 ] ) . finally , we have convolved the transformed spectra with a top hat smearing function of width @xmath23 .
this models these spectra as they would be sampled photometrically , as opposed to spectroscopically ; i.e. , this transforms the model spectra into model spectral flux distributions .
figure 2 shows the resulting spectral flux distribution .
the spectral flux distribution of the afterglow is cut off by the ly@xmath22 forest at progressively lower frequencies as one moves out in redshift .
thus high redshift afterglows are characterized by an optical `` dropout '' [ 4 ] , and vhr afterglows by a near infrared `` dropout . ''
we conclude that , if grbs occur at very high redshifts , both they and their afterglows can be easily detected .
theoretical calculations show that the birth rate of pop iii stars produces a peak in the sfr in the universe at redshifts @xmath24 , while the birth rate of pop ii stars produces a much larger and broader peak at redshifts @xmath25 [ 18,19,20 ] .
therefore one expects grbs to occur out to at least @xmath26 and possibly @xmath27 , redshifts that are far larger than those expected for the most distant quasars .
figure 3 places grbs in a cosmological context . at recombination , which occurs at redshift @xmath28 ,
the universe becomes transparent .
the cosmic background radiation originates at this redshift . shortly afterwards
, the temperature of the cosmic background radiation falls below 3000 k and the universe enters the `` dark ages '' during which there is no visible light in the universe . `` first light , '' which occurs at @xmath29 , corresponds to the epoch when the first stars form . ultraviolet radiation from these first stars and/or from the first active galactic nuclei re -
ionizes the universe . afterward , the universe is transparent in the ultraviolet .
+ + -0.2truein qsos are currently the most powerful probes of the high redshift universe .
grbs have several advantages relative to qsos as probes of cosmology .
first , grbs are expected to occur out to @xmath29 , whereas qsos occur out to only @xmath30 .
second , very high redshift grb afterglows can be 100 - 1000 times brighter at early times than are high redshift qsos .
this makes possible very sensitive high dispersion spectroscopy of the metal absorption lines and the lyman @xmath22 forest in the spectrum of the afterglows .
third , no `` proximity effect '' on intergalactic distances scales is expected for grbs and their afterglows , in contrast to qsos .
thus grbs may be relatively `` clean '' probes of the intergalactic medium , the lyman @xmath22 forest , and damped lyman @xmath22 clouds , even in the vicinity of the grbs .
the important cosmological questions that observations of grbs and their afterglows may be able to address include the following : @xmath31 information about the epoch of `` first light '' and the earliest generations of stars from merely the detection of grbs at very high redshifts ; @xmath31 information about the growth of metallicity in the universe in the star - forming entities in which the bursts occur , in damped lyman @xmath22 clouds , and in the lyman @xmath22 forest from observations of the metal absorption line systems in the spectra of their afterglows ; @xmath31 information about the large - scale structure of the universe at vhrs from the clustering of the lyman @xmath22 forest lines and the metal absorption - line systems in the spectra of their afterglows ; and @xmath31 information about the epoch of re - ionization from the depth of the lyman @xmath22 break in the spectra of their afterglows .
below we consider the first of these questions : the epoch of `` first light '' and the earliest generations of stars .
observational estimates [ 21,22,23,24 ] indicate that the sfr in the universe was about 15 times larger at a redshift @xmath32 than it is today .
the data at higher redshifts from the hubble deep field ( hdf ) in the north suggests a peak in the sfr at @xmath33 [ 24 ] , but the actual situation is highly uncertain . in figure 4
, we have plotted the sfr versus redshift from a phenomenological fit [ 25 ] to the sfr derived from submillimeter , infrared , and uv data at redshifts @xmath34 , and from a numerical simulation by [ 18 ] at redshifts @xmath35 .
the simulations done by [ 18 ] indicate that the sfr increases with increasing redshift until @xmath26 , at which point it levels off .
the smaller peak in the sfr at @xmath36 corresponds to the formation of population iii stars , brought on by cooling by molecular hydrogen .
since grbs are detectable at these vhrs and their redshifts may be measurable from the absorption - line systems and the ly@xmath22 break in the afterglows [ 4 ] , if the grb rate is proportional to the sfr , then grbs could provide unique information about the star - formation history of the vhr universe .
we have calculated the expected number @xmath37 of stars as a function of @xmath38 assuming ( 1 ) that the grb rate is proportional to the sfr , and ( 2 ) that the sfr is that given in figure 4 ( see [ 15 ] for details ) .
the left panel of figure 5 shows our results for @xmath39 for an assumed cosmology @xmath40 and @xmath41 ( other cosmologies give similar results ) .
the solid curve corresponds to the star - formation rate in figure 4 ; the dashed curve corresponds to the star - formation rate derived by [ 24 ] .
figure 5 shows that @xmath39 peaks sharply at @xmath42 and then drops off fairly rapidly at higher @xmath38 , with a tail that extends out to @xmath43 .
the rapid rise in @xmath39 out to @xmath42 is due to the rapidly increasing volume of space .
the rapid decline beyond @xmath42 is due almost completely to the `` edge '' in the spatial distribution produced by the cosmology .
in essence , the sharp peak in @xmath39 at @xmath42 reflects the fact that the sfr we have taken is fairly broad in @xmath38 , and consequently , the behavior of @xmath39 is dominated by the behavior of the co - moving volume @xmath44 ; i.e. , the shape of @xmath39 is due almost entirely to cosmology .
the right panel in figure 5 shows the cumulative distribution @xmath45 of the number of stars expected as a function of redshift @xmath38 .
the solid and dashed curves have the same meaning as in the upper panel .
figure 5 shows that for the particular sfr we have assumed , @xmath46 of all stars ( and therefore of all grbs ) have redshifts @xmath47 .
if the long grbs are indeed produced by the collapse of massive stars , one expects grbs to occur out to @xmath27 , redshifts that are far larger than those expected for the most distant qsos .
we have shown that both grbs and their afterglows are easily detected out to these vhrs .
grbs can therefore give us information about the star - formation history of the universe , including the earliest generations of stars .
the absorption - line systems and the ly@xmath22 forest visible in the spectra of grb afterglows can be used to trace the evolution of metallicity in the universe , and to probe the large - scale structure of the universe at vhrs .
finally , measurement of the ly@xmath22 break in the spectra of grb afterglows can be used to constrain , or possibly measure , the epoch at which re - ionization of the universe occurred .
castander , f. j. , & lamb , d. q. 1999 , apj , * 523 * , 593 fruchter , a. s. , et al . 1999 ,
apj , * 516 * , 683 kulkarni , s. r. , et al . 1998 , nature , * 395 * , 663 fruchter , a. s. 1999 , apj , * 516 * , 683 sahu , k. c. , et al .
1997 , nature , * 387 * , 476 kulkarni , s. r. , et al .
1999 , nature , * 398 * , 389 reichart , d. e. , 1999 , apj , * 521 * , l111 galama , t. j. , et al . 2000 ,
apj , * 536 * , 185 bloom , j. s. , et al .
1999 , nature , * 401 * , 453 woosley , s. e. 1993 , apj , * 405 * , 273 woosley , s. e. 1996 , in gamma - ray bursts , eds . c. a. meegan , r. d. preece , & t. m. koshut ( new york : aip ) , 520 paczyski , b. 1998 , apj , * 494 * , l45 macfadyen , a. i. , & woosley , s. e. 1999 , apj , * 524 * , 262 wheeler , j. c. , et al .
2000 , apj , * 537 * , 810 lamb , d. q. , & reichart , d. e. , 2000 , apj , * 536 * , 1 wijers , r. a. m. j. , rees , m. j. , & mszros , p. 1997 , mnras , * 288 * , l51 andersen , m. i. , et al .
2000 , a&a , * 364 * , l54 ostriker , j. p. , & gnedin , n. y. 1996 , apj , * 472 * , l63 gnedin , n. y. , & ostriker , j. p. 1997
, apj , * 486 * , 581 valageas , p. , & silk , j. 1999 , a&a , * 347 * , 1 gallego , j. 1995 , apj , * 455 * , l1 lilly , s. j. , et al .
1996 , apj , * 460 * , l1 connolly , a. j. 1997 , apj , * 486 * , l11 madau , p. , pozzetti , l. , & dickinson , m. 1998 , apj , * 498 * , 106 rowan - robinson , m. 1999 , ap&ss , * 266 * , 291
|
we show that , if the long grbs are produced by the collapse of massive stars , grbs and their afterglows may provide a powerful probe of cosmology and the early universe .
|
experimental studies of magnetic and electronic excitations in various band ferromagnetic systems continue to be of strong current interest , as evidenced by intensive neutron scattering studies of spin - wave excitations throughout the brillouin zone in ferromagnetic manganites highlighting magnon damping and anomalous zone - boundary softening,@xcite angle resolved photoemission spectroscopy ( arpes ) studies of iron to investigate many - body interaction between quasiparticles at the fermi level,@xcite and spin polarized electron energy loss spectroscopy ( speels ) studies of surface spin waves in ultrathin fe films showing strong spin - wave softening due to reduction of exchange interaction.@xcite driven by recent advances in the resolution of experimental probes , these studies provide valuable insight into details of the microscopic mechanism and characteristics of band ferromagnetism .
realistic multi - band calculations of spin - wave dispersion using an itinerant - electron model , in bulk bcc fe e.g. , have so far been carried out only in the random phase approximation ( rpa ) owing to the complexity of the band structure.@xcite recently , a tight - binding model involving 9 orbitals ( 4s , 4p , and 3d ) per fe atom has been used to calculate spin - wave dispersion in the rpa,@xcite and electron self - energy corrections in the ferromagnetic phase of iron were studied in light of recent arpes experiments on fe .
on the other hand , band ferromagnetism being an intrinsically strong - coupling phenomenon , spin - wave excitations in a single - band ferromagnet are strongly renormalized by correlation effects , as studied recently by incorporating self - energy and vertex corrections within a systematic inverse - degeneracy @xmath3 expansion scheme in which the spin - rotation symmetry of the hamiltonian and hence the goldstone mode@xcite are explicitly preserved order by order beyond the rpa.@xcite for the single - band hubbard model , the correlation - induced minority - spin spectral - weight transfer was shown to result in strong spin - wave energy renormalization ( quantum correction ) , and the interplay of lattice , band dispersion , and band filling effects was studied on the competition between the delocalization and exchange energy contributions to the spin stiffness , which fundamentally determines the stability of the ferromagnetic state in an itinerant ferromagnet.@xcite so how are quantum corrections generally affected by orbital multiplicity and hund s coupling ?
this question is of fundamental importance in view of the multi - band nature of transition - metal ferromagnets , but has not been addressed so far in the literature . in the special case
when the hund s coupling ( inter - orbital interaction ) is identical to the intra - orbital interaction , as obtained in the generalized @xmath2-orbital hubbard model,@xcite the quantum corrections are simply suppressed by the inverse - degeneracy factor @xmath0 . however , for arbitrary hund s coupling , the role of orbital degeneracy on quantum corrections to spin - wave excitations has not been investigated so far . in this paper
, we will extend the above goldstone - mode - preserving approach for the study of correlation - induced quantum corrections to a multi - band ferromagnet with arbitrary hund s coupling .
we will show that orbital multiplicity and hund s coupling strongly suppress the quantum corrections and spin - wave energy renormalization in a band ferromagnet .
we will further show the existence of an effective quantum parameter which , in analogy with @xmath1 for quantum spin systems and @xmath0 for the generalized @xmath2-orbital hubbard model , plays the role of @xmath4 in effectively determining the magnitude of quantum corrections in a multi - band ferromagnet . a variety of methods have been employed to investigate the role of orbital degeneracy and hund s coupling on the stability of metallic ferromagnetism , as briefly reviewed below .
the magnetic phase diagram was calculated by finite - temperature quantum monte carlo simulations within the dynamical mean - field theory ( dmft ) , and hund s coupling was shown to effectively stabilize ferromagnetism in a broad range of electron fillings even for a symmetric dos.@xcite it was pointed out that this stabilization is different from the mechanism based on asymmetric dos , which leads to ferromagnetism in the single - band hubbard model .
finite - temperature magnetism of iron and nickel was investigated using the lda + dmft approach which combines dmft with realistic electronic structure methods , and many body features of the one - electron spectra and the observed magnetic moments were described.@xcite the coulomb interaction energy values used were @xmath5 ev for fe ( ni ) , and @xmath6 ev for both fe and ni , obtained from constrained lda calculations . the role of lattice structure and hund s coupling
was investigated for various three - dimensional lattice structures within the dmft using an improved quantum monte carlo algorithm that preserves the spin - su(2 ) symmetry .
it was shown that the earlier ising - type dmft calculations@xcite overestimate the tendencies toward ferromagnetic ordering and the curie temperature .
both the lattice structure and orbital degeneracy were found to be essential for the ferromagnetism in the parameter region representing a transition metal.@xcite other numerical techniques such as exact diagonalization,@xcite density matrix renormalization group,@xcite slave boson,@xcite and the gutzwiller variational scheme@xcite have also been employed . the special case of two orbitals per site and quarter filling yields an insulating ferromagnetic state with staggered orbital ordering , as suggested by the equivalence to an `` antiferromagnetic '' state in the pseudo - spin space of the two orbitals , and first proposed as a mechanism for stabilization of ferromagnetsim by roth.@xcite however , the estimated curie temperature was found to be too low for transition - metal ferromagnets by a factor of 10 .
the insulating ferromagnetic state at quarter filling has also been investigated at strong coupling,@xcite and using the exact diagonalization method.@xcite
we consider a degenerate two - orbital hubbard model @xmath7 where @xmath8 and @xmath9 refer to the two degenerate orbitals at each lattice site @xmath10 and @xmath11 are the local spin operators for the two orbitals @xmath12 in terms of the fermionic operators @xmath13 and the pauli matrices @xmath14 .
the hopping terms @xmath15 for nearest neighbours and @xmath16 for next - nearest neighbours .
the above model includes an intra - orbital hubbard interaction @xmath17 , and an inter - orbital hund s coupling @xmath18 .
as our objective is to investigate the role of hund s coupling on quantum corrections in the orbitally - degenerate ferromagnetic state , we have not retained the inter - orbital density - density interaction term @xmath19 , also conventionally included in the orbital hubbard model .
this density interaction term is important in the context of orbital ordering in manganites,@xcite especially near quarter filling,@xcite as expected from its structural similarity with an antiparallel - spin interaction in the pseudo - spin @xmath20 space , which will favor antiferromagnetic orbital ordering . the orbital hubbard model above has been written in an explicitly spin - rotationally - symmetric form .
this continuous symmetry implies the existence of goldstone modes in the spontaneously - broken - symmetry state,@xcite which will play a central role in our study of correlation effects in the ferromagnetic state in which self - energy and vertex corrections are systematically incorporated such that the goldstone mode is explicitly preserved order by order .
we assume a ferromagnetic ground state with polarization in the @xmath21 direction , and examine transverse spin fluctuations representing both collective spin - wave and single - particle stoner excitations .
we consider the time - ordered transverse spin fluctuation propagator in this broken - symmetry state : @xmath22 | \psi_0 \rangle\ ] ] where the orbital indices @xmath23 , and the fermion spin - lowering and spin - raising operators @xmath24 .
the transverse spin fluctuation propagators can be expressed exactly in terms of the irreducible particle - hole propagators @xmath25 , as shown diagrammatically in fig .
the corresponding coupled equations are : @xmath26 where the interaction term @xmath27 for @xmath28 and @xmath29 for @xmath30 is shown as the wavy line in fig .
1 , and summation over repeated indices is implied . as external probes such as magnetic field or the neutron magnetic moment couple equally to electron magnetic moments in the two degenerate orbitals
, it is physically relevant to consider the sum @xmath31 of the intra - orbital and inter - orbital contributions .
indeed , it is particularly convenient to solve the coupled equations for the total transverse propagator , and we obtain : @xmath32 where @xmath33 represents the total irreducible particle - hole propagator , for which a systematic expansion will be discussed below .
in analogy with the inverse - degeneracy ( @xmath0 ) expansion for the generalized @xmath2-orbital hubbard model,@xcite we consider a systematic expansion : @xmath34 for the irreducible propagator @xmath35 in orders of fluctuations .
the zeroth - order first term @xmath36 is simply the bare particle - hole propagator , whereas the higher - order terms @xmath37 etc .
represent quantum corrections involving self - energy and vertex corrections , as discussed below . in the inverse - degeneracy expansion scheme,@xcite
the diagrams were systematized in terms of the expansion parameter @xmath0 , with the @xmath38-order term @xmath39 involving @xmath40 powers of @xmath0 .
thus the expansion parameter @xmath0 played , in analogy with @xmath1 for quantum spin systems , the role of @xmath4 . in the following
we will show that the dimensionless factor @xmath41 ( and a similar factor for the @xmath2 orbital case ) plays the role of the expansion parameter for arbitrary hund s coupling .
this allows for a continuous interpolation between the single - orbital limit with no hund s coupling and the generalized @xmath2-orbital limit with @xmath42 .
_ random phase approximation : _ retaining only the zeroth - order term @xmath36 yields the random phase approximation ( rpa ) , amounting to a `` classical '' description of noninteracting spin - fluctuation modes .
as the hopping term is diagonal in orbital indices , the zeroth - order term involves only the intra - orbital contribution : @xmath43 where the hartree - fock level band energies @xmath44 involve the exchange splitting @xmath45 between the two spin bands .
the superscripts @xmath46 refer to particle ( hole ) states above ( below ) the fermi energy @xmath47 . here
the magnetization @xmath48 is assumed to be identical for both orbitals @xmath12 in the orbitally degenerate ferromagnetic state .
for the saturated ferromagnet , the magnetization @xmath49 is equal to the particle density @xmath40 for each orbital . at the rpa level ,
the two ( intra- and inter - orbital ) components of the transverse spin propagator are easily obtained by solving the coupled equations ( 3 ) , and we obtain : @xmath50_{\rm rpa } % = \frac{\chi_0({\bf q},\omega)(1-u\chi_0({\bf q},\omega ) ) } { ( 1-u\chi_0({\bf q},\omega))^2 - ( j\chi_0({\bf q}\omega))^2 } = \frac{1}{2 } \left [ \frac{\chi_0({\bf q},\omega)}{1-u^+ \chi_0({\bf q},\omega ) } + \frac{\chi_0({\bf q},\omega)}{1-u^- \chi_0({\bf q},\omega ) } \right ] = [ \chi^{-+}_{\beta \beta}({\bf q},\omega)]_{\rm rpa } \\ & & [ \chi^{-+}_{\alpha \beta}({\bf q},\omega)]_{\rm rpa } % = \frac{\chi_0({\bf q},\omega ) j \chi_0({\bf q},\omega ) } { ( 1-u\chi_0({\bf q},\omega))^2 - ( j\chi_0({\bf q } \omega))^2 } = \frac{1}{2 } \left [ \frac{\chi_0({\bf q},\omega)}{1-u^+ \chi_0({\bf q},\omega ) } - \frac{\chi_0({\bf q},\omega)}{1-u^- \chi_0({\bf q},\omega ) } \right ] = [ \chi^{-+}_{\beta \alpha}({\bf q},\omega)]_{\rm rpa } \end{aligned}\ ] ] where the two effective interaction terms above are @xmath51 .
as seen , the propagators involve linear combinations of two modes , which can directly be identified as in - phase and out - of - phase combinations with respect to the orbital degrees of freedom .
these two modes represent gapless ( acoustic ) and gapped ( optical ) branches , as shown below .
we also introduce kernels for these two propagators which will be used later in the expressions for quantum corrections : @xmath52_{\rm rpa } = ( [ \chi^{-+}_{\alpha \alpha}]_{\rm rpa } - \chi_0)/\chi_0 ^ 2 = \frac{1}{2 } \left [ \frac{u^+}{1-u^+ \chi_0 } + \frac{u^-}{1-u^- \chi_0 } \right ] \ ] ] @xmath53_{\rm rpa } = [ \chi^{-+}_{\alpha \beta}]_{\rm rpa}/\chi_0 ^ 2 = \frac{1}{2 } \left [ \frac{u^+}{1-u^+ \chi_0 } - \frac{u^-}{1-u^- \chi_0 } \right ] \ ] ] the in - phase mode with effective interaction @xmath54 corresponds to the usual goldstone mode ( acoustic branch ) .
this is expected as the exchange splitting @xmath55 in the @xmath56 energy denominator involves the same effective interaction @xmath57 . on the other hand ,
the out - of - phase mode with effective interaction @xmath58 yields gapped excitations ( optical branch ) . a typical spectral function plot is shown in fig [ spctrl ] , with the inset showing dispersion of the acoustic and optical branches , and onset of stoner excitations .
_ quantum corrections : _ in the following we consider first - order quantum corrections to the irreducible particle - hole propagator ( 7 ) .
we consider the relatively simpler case of a saturated ferromagnet in which the minority - spin particle - hole processes are absent as the minority - spin band is pushed above the fermi energy due to coulomb repulsion . in this case
the effective antiparallel - spin interactions at lowest order reduce to the bare interactions @xmath17 and @xmath18 , and the effective parallel - spin interactions reduce to a single term involving the majority - spin particle - hole bubble . generally , these effective interactions involve a series of bubble diagrams , with even and odd number of bubbles , respectively .
diagrammatic contributions to the first - order quantum correction @xmath59 are shown in figure [ fig:2 ] .
the interaction lines are either @xmath17 or @xmath18 , depending on the orbitals of the connecting fermion lines .
the external ( @xmath12 on the right ) orbital degrees of freedom are also summed over to include both the intra- and inter - orbital contributions as in eq .
the physical meaning of the four diagrams has been discussed earlier.@xcite diagrams ( a ) and ( d ) represent corrections to the irreducible propagator due to self - energy corrections to the spin-@xmath60 particle arising from spin and charge fluctuations , respectively .
the shaded part in diagram ( a ) represents the propagator @xmath61 .
diagrams ( b ) and ( c ) represent vertex corrections , where the shaded part represents the kernel @xmath62 introduced in eqs .
( 12,13 ) . in diagram
( b ) , the opposite - spin _ particle - particle _ interaction suppresses the spin-@xmath60 particle spin-@xmath63 hole correlation , yielding a negative correction to @xmath25 .
diagram ( b ) therefore represents suppression of the magnetic response due to particle - particle correlations .
all four diagrams involve a spin - charge coupling , as indicated by the spin-@xmath63 particle - hole bubble , present explicitly in diagrams ( c ) and ( d ) and implicitly in ( a ) and ( b ) .
integrating out the fermion frequency - momentum degrees of freedom , the first - order quantum corrections to the irreducible particle - hole propagator are obtained as : @xmath64 @xmath65 @xmath66 ^ 2 \nonumber \\ & & \times \sum_{\bf k '' } \left ( \frac{1}{\epsilon^{\uparrow + } _ { \bf k''-q+q}-\epsilon^{\uparrow -}_{\bf k''}+\omega-\omega -i\eta } \right ) \end{aligned}\ ] ] @xmath67 we note that in the @xmath68 limit of decoupled orbitals , we recover the single - band hubbard model results.@xcite as collective spin - wave excitations are represented by poles in ( 5 ) , spin - rotation symmetry requires that @xmath69 for @xmath70 , corresponding to the goldstone mode .
since the zeroth - order term @xmath36 already yields exactly @xmath71 , the sum of the remaining terms must exactly vanish in order to preserve the goldstone mode .
for this cancellation to hold for arbitrary @xmath18 and @xmath17 , each higher - order term @xmath39 in the expansion ( 7 ) must individually vanish , implying that spin - rotation symmetry is preserved order - by - order .
this cancellation is demonstrated below for the first - order quantum correction obtained above . towards this end , we first note that the boson term ( quantity in braces ) in eq . ( 15 ) for @xmath72 can be expressed as : @xmath73 which is of identical form as the boson terms in eq .
( 14 ) and ( 16 ) for @xmath74 and @xmath75 . with @xmath76 for @xmath77
, we obtain from eqs .
( 14 - 17 ) : @xmath78 \ ; , \end{aligned}\ ] ] which yields identically vanishing contribution for each spin - fluctuation mode @xmath79 , thus preserving the goldstone mode .
we note that this mode - by - mode cancellation is quite independent of the spectral - weight distribution of the spin - fluctuation spectrum between collective spin - wave and particle - hole stoner excitations .
furthermore , the cancellation holds for all @xmath80 , indicating no spin - wave amplitude renormalization , as expected for the saturated ferromagnet in which there are no quantum corrections to magnetization .
evaluation of the @xmath81 integral in eqs .
( 14 - 17 ) has been discussed earlier.@xcite using a spectral representation it is convenient to carry out the @xmath81 integral numerically so as to include all three ( acoustic , optical , and stoner ) contributions from the magnon ( @xmath82 ) and the particle - hole terms . a typical spectral function plot of the rpa - level magnon propagator @xmath83_{\rm rpa}$ ] shows ( fig .
[ spctrl ] ) the low - energy ( acoustic ) , intermediate - energy ( optical ) , and high - energy ( stoner ) contributions , with the inset showing the dispersion of acoustic and optical branches and onset energy of the stoner branch .
we will now show that hund s coupling results in a strong suppression of quantum corrections , and this is the central message of this paper .
the suppression is approximately by an overall factor of @xmath41 for the two orbital case , and by a factor of @xmath84 for the @xmath2 orbital case .
the magnitude of the overall factor is asymptotically exact in the two limits @xmath68 and @xmath85 . in these two limits
the factor approaches 1 and 1/2 , respectively , for the two orbital case , whereas it approaches 1 and @xmath0 , respectively , for the @xmath2 orbital case .
indeed , for our degenerate two - orbital hubbard model ( 1 ) , this overall factor plays the same role as @xmath0 in the generalized @xmath2-orbital hubbard model,@xcite and generalizes this earlier result to the case of arbitrary hund s coupling . in the generalized @xmath2-orbital hubbard model ,
the form @xmath86 of the interaction term implies identical inter- and intra - orbital interactions ( @xmath42 ) . to see this overall suppression factor
, we consider eqs .
( 14 - 17 ) for the quantum corrections @xmath25 . in eq .
( 14 ) , assuming similar contributions from the acoustic and optical modes , which is a good approximation in the @xmath87 limit when the two modes become nearly degenerate , the boson term @xmath88 yields negligible contribution in view of the opposite contributions ( 11 ) of the acoustic and optical modes .
hence , only the contribution from the first boson term @xmath89 survives , leaving an overall factor @xmath90 on carrying out the @xmath91 integration . comparing with the corresponding factor @xmath92 for the single orbital case with identical interaction @xmath93 ,
yields an overall relative factor of @xmath41 . in view of eq . ( 18 ) , the quantum correction @xmath72 also involves the same boson term and hence yields the same overall factor .
similarly , the quantum corrections @xmath75 and @xmath94 together yield the same boson term and hence the same overall factor .
the magnitude of the overall factor is also exact in the @xmath85 limit , as shown below . the boson term in eq .
( 14 ) can be identically written , in view of eqs .
( 10,11 ) , as : @xmath95\ ] ] in terms of the acoustic and optical branches .
therefore , in the @xmath85 limit , the contribution of the optical mode vanishes , leaving an overall factor of 1/2 , as also resulting from the expression @xmath41 .
band ferromagnetism being a strong - coupling phenonomenon , quantum corrections to spin stiffness and magnon energy for the single - orbital hubbard model with @xmath96 have been shown to yield strong renormalizations ( reduction relative to the rpa values ) in two and three dimensions.@xcite the strong suppression of quantum corrections shown above highlights the critical role of hund s coupling in stabilizing ferromagnetism in realistic systems like transition metals such as fe , ni etc . with multiple @xmath97 orbitals .
we propose that in such systems , hund s coupling favours ferromagnetism by strongly suppressing the quantum corrections .
we now consider the net quantum correction @xmath25 for small @xmath98 in order to obtain the renormalized spin stiffness , which provides a quantitative measure of the stability of the ferromagnetic state with respect to long - wavelength fluctuations . in eqs .
( 14 - 17 ) , writing the antiparallel - spin particle - hole energy denominator as @xmath99\ ] ] and expanding in powers of the small band - energy difference @xmath100 we find that besides the zeroth - order cancellation for @xmath101 , the first - order terms in @xmath102 also exactly cancel .
this exact cancellation implies that there is no quantum correction to the delocalization contribution @xmath103 in the spin stiffness constant ; only the exchange contribution in the spin stiffness is renormalized by the surviving second - order terms in @xmath102 , and we obtain for the first - order quantum correction to stiffness : @xmath104 \ ; , \end{aligned}\ ] ] where @xmath105 is the lattice dimensionality .
here we have introduced effective interactions in the transverse channel : @xmath106 as @xmath107 is odd in momentum , the second and third terms in ( 23 ) give vanishingly small contributions due to partial cancellation .
the behaviour of quantum correction to spin stiffness with band filling is shown in fig . 4 for different values of the hund s coupling @xmath18 .
here we have kept @xmath108 fixed so that the exchange band splitting and the classical spin stiffness remain unchanged . the quantum correction to spin stiffness decreases sharply with @xmath18 and reduces to exactly half the magnitude
when @xmath109 , as shown in fig .
5 for two different band fillings .
inset shows a comparison of the normalized quantum correction calculated from ( 23 ) with the approximate form @xmath110 , which asymptotically approaches the calculated result in the two limits @xmath111 and @xmath112 , as discussed in section v. these results clearly show that even a small hund s coupling is rather efficient in strongly suppressing the quantum correction .
6 shows that the renormalized spin stiffness @xmath113 rapidly changes sign from negative to positive with increasing @xmath18 , highlighting the effective stabilization of ferromagnetism by hund s coupling , shown here for the simple cubic lattice with @xmath114 and fixed @xmath115 .
as discussed in section v , generalizing to the @xmath2-orbital case , the quantum correction to spin stiffness should be approximately suppressed by the factor @xmath84 . with increasing hund s coupling ,
this factor rapidly approaches @xmath0 , particularly for large @xmath2 ; the generalized @xmath2-orbital hubbard model therefore provides a good approximation for transition - metal ferromagnets with @xmath116 3d orbitals .
we have examined the role of this @xmath0 suppression of quantum corrections on the spin stiffness .
7 shows the renormalized spin stiffness @xmath117 for different number of orbitals @xmath2 , evaluated for a bcc lattice with @xmath118 , bandwidth @xmath119ev , coulomb interaction energy @xmath120ev , and the lattice parameter @xmath121 for fe . in a recent band - structure calculation,@xcite
the interaction energy considered is @xmath122ev ( so that the magnetic moment evaluated per fe atom is equal to @xmath123 ) and the bandwidth from the calculated dos plot is seen to be about 4ev .
our calculated values for the renormalized spin stiffness for @xmath124 are close to the measured value 280mev@xmath125 for fe .
the spin stiffness is seen to involve a quantum reduction of about @xmath126 near optimal filling .
turning now to the magnon dispersion over the entire brillouin zone , the renormalized magnon energy @xmath127 for finite @xmath128 is obtained from the pole condition @xmath129 in eq .
( 5 ) , where @xmath130 , and the four contributions to the first - order quantum correction @xmath131 are given in eqs .
( 14 - 17 ) .
while the bare particle - hole propagator @xmath132 remains real in the relevant @xmath80 range , the quantum correction @xmath131 is complex for any finite @xmath133 due to the coupling with charge fluctuations , resulting in finite zero - temperature magnon damping.@xcite both collective and stoner excitations are included in evaluating the @xmath81 integral , as discussed below ( 19 ) .
the effect of hund s coupling on the renormalized magnon dispersion is shown in figs . 8 and 9 , which provide comparisons of the bare ( @xmath134 ) and
renormalized ( @xmath127 ) magnon dispersions along symmetry directions for the simple cubic and square lattices , respectively .
again @xmath108 was kept fixed so that the bare magnon energy remains unchanged . while hund s coupling increases the magnon energy in the entire brillouin zone , the effect is particularly dramatic for long wavelength modes , which rapidly crossover from negative - energy to positive - energy modes .
even when the bare magnon dispersion exhibits nearly heisenberg form , with energies at x , m , r approximately in the ratio 1:2:3 as in fig .
8(b ) , the renormalized magnon dispersion shows strong anomalous softening at x relative to r. this indicates that magnon renormalization due to spin - charge coupling results in the `` generation '' of additional exchange couplings @xmath135 etc . within an equivalent localized - spin model with the same magnon dispersion .
the renormalized spin stiffness and magnon energies obtained above essentially determine the finite - temperature spin dynamics and therefore the curie temperature @xmath136 .
finite - temperature spin dynamics in a band ferromagnet and reduction of magnetization due to thermal excitation of spin waves has been recently discussed in terms of the spectral weight transfer across the fermi energy.@xcite due to the spin - flip scattering of electrons accompanying the thermal magnon excitations , a portion of the majority - spin spectral weight is transferred to the minority - spin band above the fermi energy , while an equal amount of minority - spin spectral weight is transferred to the majority - spin band below the fermi energy . from the self - energy correction in eq .
( 14 ) , and using the resolution of the transverse spin propagators ( eqs .
10,11 ) into acoustic and optical modes , the curie temperature is approximately given by : @xmath137 \sum_{\bf k } \left ( \frac{1}{\epsilon_{\bf k - q}^{\downarrow + } - \epsilon_{\bf k}^{\uparrow - } } \right)^2\ ] ] within a renormalized spin - fluctuation theory,@xcite where @xmath127 and @xmath138 refer to the magnon energies for the acoustic and optical modes .
now , finite hund s coupling results in i ) gapped optical modes with higher energies ( @xmath139 ) than the acoustic modes , ii ) reduced weightage of the optical modes due to the @xmath140 factor in above equation , and iii ) enhancement of the acoustic magnon - mode energies ( even at fixed @xmath108 ) .
all three factors suppress the rhs of above equation , directly resulting in enhanced curie temperature . while this enhancement of @xmath136 is in general agreement with dmft studies where only local excitations are incorporated , eq .
( 25 ) highlights the sensitivity to long - wavelength modes as well in our goldstone - mode - preserving approach .
negative - energy long - wavelength modes as in figs . 8 and 9 would result in vanishing @xmath136 even if bulk of the ( short - wavelength ) modes have positive energy .
long wavelength modes play a particularly important role in low - dimensional systems .
the divergence of the rhs of eq .
( 25 ) in one and two dimensions yields vanishing @xmath136 in accordance with the mermin - wagner theorem.@xcite
the role of orbital degeneracy and hund s coupling on quantum corrections to spin - wave excitations in a band ferromagnet was investigated .
a spin - rotationally - symmetric approach was employed in which self - energy and vertex corrections are incorporated systematically so that the goldstone mode is explicitly preserved order by order .
the present study of quantum corrections for arbitrary hund s coupling allows for a continuous interpolation between the orbitally independent case @xmath141 equivalent to the single - band hubbard model and the orbitally symmetric case of identical inter- and intra - orbital coulomb interactions ( @xmath109 ) equivalent to the generalized hubbard model with @xmath2 degenerate orbitals per site , for which the first - order quantum corrections are suppressed by the factor @xmath0 .
we find that even a relatively small hund s coupling is rather efficient in strongly suppressing the quantum corrections , especially for large @xmath2 , resulting in a strong enhancement of ferromagnetism .
this mechanism for the enhancement of ferromagnetism due to strong suppression of quantum corrections by hund s coupling implicitly involves an interplay of several band and lattice characteristics .
competition between the delocalization @xmath103 and exchange @xmath142 contributions to spin stiffness results in a subtle interplay of lattice , dimensionality , band dispersion , spectral distribution , coulomb interaction , and band filling effects , as investigated for several two and three - dimensional lattices.@xcite for a two - orbital hubbard model , the first - order quantum correction to spin stiffness was obtained diagrammatically and evaluated as a function of the hund s coupling strength @xmath18 .
we showed that the effective quantum correction factor decreases rapidly from 1 at @xmath143 to 1/2 at @xmath109 , and its behaviour with @xmath18 closely follows the approximate form @xmath144 obtained from our diagrammatic analysis . in the intermediate - coupling regime , with interaction strength @xmath17 comparable to the bandwidth , the renormalized spin stiffness was evaluated in the whole range of band fillings , and was found to rapidly crossover from negative to positive values with increasing hund s coupling , highlighting the strong role of orbital degeneracy and hund s coupling on the stability of the ferromagnetic state with respect to long wavelength fluctuations .
we also obtained , for both the square and simple cubic lattices , the renormalized spin - wave energy dispersion @xmath127 for momenta along symmetry directions in the brillouin zone . while hund s coupling results in an enhancement of the magnon energy in the entire brillouin zone
, the effect is particularly dramatic for long wavelength modes , which rapidly crossover from negative to positive energy .
generalizing to the @xmath2 orbital case , an effective quantum parameter @xmath145 was obtained which , in analogy with @xmath1 for quantum spin systems and @xmath0 for the generalized hubbard model , plays the role of @xmath4 for quantum corrections in a band ferromagnet . for large @xmath2 , this quantum parameter decreases rapidly with the hund s coupling @xmath18 and saturates to @xmath0 as @xmath85 .
this strong suppression of quantum corrections due to orbital degeneracy and hund s coupling is quite significant for the 3d transition - metal ferromagnets fe , co , ni , where @xmath124 . with @xmath146 , as considered in the recent rpa calculations for iron,@xcite we obtain a value of @xmath147 for the quantum parameter . while the smallness of this quantum correction parameter accounts for why rpa calculations of the spin stiffness for fe with realistic band structure yield values in close agreement with the measured value of 280 mev.@xmath125 , it also highlights the significant magnitude of the correlation - induced quantum corrections involved in the measured spin stiffness values for transition - metal ferromagnets .
|
an effective quantum parameter is obtained for the band ferromagnet in terms of orbital degeneracy and hund s coupling .
this quantum parameter determines , in analogy with @xmath0 for the generalized hubbard model and @xmath1 for quantum spin systems , the strength of quantum corrections to spin stiffness and spin - wave energies .
quantum corrections are obtained by incorporating correlation effects in the form of self - energy and vertex corrections within a spin - rotationally - symmetric approach in which the goldstone mode is explicitly preserved order by order .
it is shown that even a relatively small hund s coupling is rather efficient in strongly suppressing quantum corrections , especially for large @xmath2 , resulting in strongly enhanced stability of the ferromagnetic state .
this mechanism for the enhancement of ferromagnetism by hund s coupling implicitly involves a subtle interplay of lattice , dimensionality , band dispersion , spectral distribution , and band filling effects .
|
light - cone hamiltonian diagonalization methods offer a number of attractive advantages for solving nonperturbative problems in quantum field theory , such as a physical minkowski space description , boost invariance of the bound - state wave functions , no requirement for fermion doubling , and a consistent fock - state expansion well matched to physical problems . in the discretized light - cone quantization ( dlcq )
method , the light - cone hamiltonian @xmath0 of a quantum field theory is diagonalized on a discrete fock basis defined by assuming periodic boundary conditions in the light - cone coordinates @xcite .
the eigenvalues of @xmath0 give the mass spectrum of the theory , and the respective eigenfunctions projected on the free fock basis provide the frame - independent light - cone wave functions needed for phenomenology @xcite including the amplitudes needed to compute exclusive @xmath1 decays @xcite , deeply virtual compton scattering @xcite , and other hard exclusive processes @xcite .
the dlcq method has been successfully used to solve a large number of one - space and one - time theories @xcite , including supersymmetric gauge theories @xcite .
it also has found application in analyzing confinement mechanisms @xcite , string theory @xcite , and @xmath2-theory @xcite . the application of dlcq to physical , ( 3 + 1)-dimensional space - time quantum field theories is computationally challenging because of the rapid growth of the number of degrees of freedom as the size of the fock representation grows .
a promising alternative is the transverse lattice method @xcite which combines light - cone methods in the longitudinal light - cone direction with a space - time lattice for the transverse dimensions .
recently dalley @xcite and burkhart and seal @xcite have extended the transverse lattice method to estimate the shape of the valence light - cone wave function of a pion , a key input to much hadron phenomenology .
burkhart and seal have also given an explicit calculation of the isgur wise function for semileptonic @xmath1 decays @xcite .
another major difficulty in applying dlcq to quantum field theory in 3 + 1 dimensions is the implementation of a nonperturbative renormalization method .
most methods of regulating nonperturbative calculations in the light - cone representation , such as momentum cutoffs , do not allow a correct renormalization even of perturbative calculations .
the problem can be traced to the fact that any momentum cutoff violates lorentz invariance as well as gauge invariance @xcite . since
dimensional regularization is not available in dlcq , one needs to introduce new fields or degrees of freedom to render the ultraviolet behavior of the theory finite .
one intriguing possibility is to analyze ultraviolet - finite supersymmetric theories and then introduce breaking of the theory .
the heavy supersymmetric partners then regulate the ordinary sector of the theory in a manner analogous to pauli
villars ( pv ) regulation @xcite .
string theory also provides mechanisms for regulating quantum field theory at short distances which are equivalent to an infinite spectrum of pv particles @xcite .
the introduction of pv fields can thus regulate a theory covariantly , after which the discretized momentum grid of dlcq acts only as a numerical tool in the manner of performing a numerical integral . in our previous work @xcite
we have shown that a model field theory in 3 + 1 dimensions can be solved by combining dlcq with pv regulation of the ultraviolet regime . in our first application
@xcite , a model theory was constructed to have an exact analytic solution by which the dlcq results could be checked , for both accuracy and rapidity of convergence .
the model was regulated in the ultraviolet by a single pv boson , which was included in the dlcq fock basis in the same way as the `` physical '' particles of the theory .
we then extended this approach to a more realistic model which mimics many features of a full quantum field theory @xcite . unlike the analytic model which contained a static source , the light - cone energies of the particles in this model have the correct longitudinal and transverse momentum dependence .
an important question is whether the generalized pv method with a finite number of fields can regulate a field theory at all orders .
paston and franke @xcite have studied the relation between perturbation theory in the light - cone representation and standard feynman perturbation theory , and they have developed general rules for testing regularization procedures . for full yukawa theory , paston , franke and prokhvatilov @xcite
have shown that one pv boson and two pv fermions can regulate the theory in such a way as to allow a correct perturbative renormalization . in this paper
we shall apply generalized pv regularization and discrete light - cone quantization to the nonperturbative solution of ( 3 + 1)-dimensional yukawa theory in a single - fermion truncation .
we allow any number of bosons in the yukawa theory but only one fermion in the fock representation ; fermion pair terms and any other terms that involve anti - fermions are neglected . we shall thus consider a field - theoretic model where one particle , which we take to be a fermion of mass @xmath2 , acts as a dynamical source and sink for bosons of mass @xmath3 .
in addition , three heavy pv scalars , including two with negative norm , will be used to regulate the theory so that the chiral properties of the renormalized theory are maintained , at least to one loop in perturbation theory .
in particular , the mass of the renormalized fermion constituent vanishes as its bare mass vanishes .
a distinct advantage of our approach is that the counterterms are generated automatically by the pv particles and their negative - metric couplings .
we emphasize that the pv fields are added from the beginning , at the lagrangian level , to facilitate our nonperturbative calculations , rather than being invoked for subtractions at a diagrammatic level , which would limit the implementation to perturbation theory .
note that the pv fields enter singly , doubly or triply at all three - point vertices .
this contrasts with use of generalized pauli
villars spectral integrals over mass to regulate divergences .
the extra degrees of freedom of the pv scalars and their negative - metric couplings introduce new computational challenges .
however , the matrix eigenvalue problem can be solved for the lowest - mass state by the use of a new , indefinite - metric lanczos algorithm which we describe in an appendix .
we also calculate the light - cone wave function of each fock - sector component and the values for various physical quantities , such as average multiplicities and average momenta of constituents , bosonic and fermionic structure functions , and a form factor slope .
we also verify that with our choice of pv conditions , the dlcq calculations of the nonperturbative theory at weak coupling coincide with the covariant perturbation theory through one loop , although numerical resolution does start to become a problem even on a supercomputer . in our convention ,
we define light - cone coordinates @xcite by @xmath4 the time coordinate is taken to be @xmath5 .
the dot product of two four - vectors is @xmath6 thus the momentum component conjugate to @xmath7 is @xmath8 , and the light - cone energy is @xmath9 .
we use underscores to identify light - cone three - vectors , such as @xmath10 for additional details , see appendix a of ref .
@xcite or a review paper @xcite .
the following is an outline of the remainder of the paper . in sec .
[ sec : yukawa ] we discuss the regularization and renormalization of the yukawa hamiltonian .
our numerical methods and results are presented in sec .
[ sec : results ] .
section [ sec : conclusions ] contains our conclusions and plans for future work .
details of the numerical diagonalization method are given in the appendix .
we write the lagrangian for yukawa theory as @xmath11\psi -m\psibar\psi - g\phi\psibar\psi\,.\ ] ] the corresponding light - cone hamiltonian has been given by mccartor and robertson @xcite .
here we consider a single - fermion truncation and therefore neglect pair terms and any other terms that involve anti - fermions .
we also neglect longitudinal zero modes . to regulate the theory , we include three pv bosons .
the resulting hamiltonian is @xmath12 b_{\underline{n}+\underline{m},-s}^\dagger b_{\underline{n},s } a_{i\underline{m } } + \mbox{h.c.}\right ) \nonumber\\ & & + \frac{mg}{\sqrt{8\pi}l_\perp } \sum_{\underline{n}\underline{m}}\sum_{si}\frac{\xi_i}{\sqrt{m } } \left(\left[\frac{1}{n / k}+\frac{1}{(n+m)/k}\right ] b_{\underline{n}+\underline{m},s}^\dagger b_{\underline{n},s } a_{i\underline{m } } + \mbox{h.c . }
\right ) \\ & & + \frac{g^2}{8\pi l_\perp^2 } \sum_{\underline{n}\underline{m}\underline{m}'}\sum_{sij } \frac{\xi_i\xi_j}{\sqrt{mm ' } } \left[\left(b_{\underline{n}+\underline{m}+\underline{m}',s}^\dagger b_{\underline{n},s } a_{i\underline{m } ' } a_{j\underline{m } } \frac{1}{(n+m)/k}+ \mbox{h.c . }
\nonumber \\ & & \rule{0.25in}{0 mm } \left .
+ b_{\underline{n}+\underline{m}-\underline{m}',s}^\dagger b_{\underline{n},s } a_{i\underline{m}'}^\dagger a_{j\underline{m } } \left(\frac{1}{(n - m')/k}+\frac{1}{(n+m)/k}\right)\right]\ , , \nonumber\end{aligned}\ ] ] where @xmath2 is the fermion mass , @xmath13 is the physical boson mass , @xmath14 and @xmath15 are the mass and norm of the i - th pv boson , and @xmath16 .
the nonzero commutators are @xmath17 = ( -1)^i\delta_{ij}\delta_{\underline{m},\underline{m}'}\ , , \;\ ; \left\{b_{\underline{n},s},b_{\underline{n}',s'}^\dagger\right\ } = \delta_{\underline{n},\underline{n } ' } \delta_{s , s'}\,.\ ] ] the different boson couplings are denoted by @xmath18 , where @xmath19 corresponds to the physical boson ; the other @xmath20 are chosen to arrange cancellations discussed below .
fermion self - induced inertia terms cancel in a sum over bosons and therefore do not appear .
the bare parameters are the coupling @xmath21 and the mass shift @xmath22 .
the number of pv flavors is determined by the cancellations needed to regulate the theory and restore chiral invariance in the @xmath23 limit @xcite .
the discrete chiral symmetry @xmath24 , @xmath25 should itself guarantee a zero self - mass in this limit , if not for the fact that the symmetry is typically broken by a cutoff .
the one - loop self - energy of the fermion is @xcite @xmath26\ , , \nonumber\end{aligned}\ ] ] with @xmath27 a cutoff such that @xmath28 . in order that the self - energy be finite and proportional to @xmath29 , the relative coupling strengths @xmath20 must satisfy the constraints @xmath30 in addition
, the norm of the i - th pv field must be chosen as @xmath15 .
the third constraint in ( [ eq : constraints ] ) is peculiar to the one - loop calculation . for higher - order or nonperturbative calculations ,
it must be replaced by a more general renormalization condition .
however , to simplify the numerical work , we use the one - loop constraint and then check for failure of cancellation in the zero - mass limit .
as it stands , @xmath0 contains infrared divergences associated with the instantaneous fermion interaction , which is singular at the point where the instantaneous fermion has zero longitudinal momentum .
the divergences are partly cancelled by crossed - boson contributions . to cancel the remainder we need to add an effective interaction , modeled on the missing fermion z graph .
the effective interaction is constructed from the pair creation and annihilation terms in the yukawa light - cone energy operator @xcite @xmath31\xi_i b_{\underline{p},s}^\dagger d_{\underline{q}-\underline{p},s}^\dagger a_{i\underline{q } } + \mbox{h.c . } \\ & & + \frac{mg}{2l_\perp\sqrt{2l } } \sum_{\underline{p}\underline{q}si } \left[\frac{1}{p^+\sqrt{q^+}}-\frac{1}{(q^+-p^+)\sqrt{q^+}}\right]\xi_i b_{\underline{p},s}^\dagger d_{\underline{q}-\underline{p},-s}^\dagger a_{i\underline{q } } + \mbox{h.c . }
\nonumber\end{aligned}\ ] ] the denominator for the intermediate state is @xmath32 to guarantee the cancellation of the singularity in the numerical calculation , the instantaneous interaction is kept only if the corresponding crossed - boson graph is permitted by the numerical cutoffs .
the fock - state expansion for the single - fermion eigenstate of the hamiltonian is @xmath33 where @xmath34 is the number of physical bosons , @xmath35 the number of pv bosons of flavor @xmath36 , @xmath37 , and @xmath38 is the flavor of the j - th constituent boson .
it solves the eigenvalue problem @xmath39 .
the normalization is @xmath40 mass renormalization is carried out by rearranging the eigenvalue problem into one for @xmath22 at fixed @xmath29 @xmath41 \tilde{\phi}(\underline{q}_i ) \nonumber \\ & -&\int\prod_j dy'_j d^2q'_{\perp j}\sqrt{xx ' } { \cal k}(\underline{q}_i,\underline{q}'_j)\tilde{\phi}'(\underline{q}'_j)= \delta m^2\tilde{\phi}(\underline{q}_i)\,.\end{aligned}\ ] ] here @xmath42 is the fermion momentum fraction , @xmath43 is shorthand for the interaction kernel , and the @xmath44 are new wave functions . the coupling is fixed by setting a value for the expectation value @xmath45 for the boson field operator @xmath46 .
this quantity is useful because it can be computed fairly efficiently in a sum similar to a normalization sum @xmath47 with @xmath48 being the norm of the state with boson flavor partitioning @xmath49 .
the principal tools for the solution of the hamiltonian eigenvalue problem are dlcq @xcite and the lanczos diagonalization algorithm @xcite .
the first converts the problem to a matrix form , and the second quickly generates one or more eigenvectors .
the use of negatively - normed states makes the ordinary lanczos algorithm inapplicable ; however , an efficient generalization has been developed for this situation .
it is discussed in the appendix .
the constraint for the coupling renormalization is solved iteratively .
the discretization is based on the standard dlcq approach where longitudinal and transverse momenta are assigned discrete values @xmath50 and @xmath51 , with @xmath52 and @xmath53 chosen length scales .
momentum conservation requires the individual @xmath54 to sum to a fixed constant @xmath55 , where @xmath56 is the total longitudinal momentum .
the integer @xmath55 is called the harmonic resolution @xcite , because longitudinal momentum fractions @xmath57 are resolved to order 1/k .
the positivity of longitudinal momenta forces a natural cutoff such that @xmath58 . also
, the eigenvalue problem for @xmath29 is independent of @xmath52 ; the length scale cancels between @xmath59 and @xmath60 .
the boundary conditions in the longitudinal direction are chosen to be periodic for the boson fields but antiperiodic for the fermion field .
this means that the integers @xmath54 are even for bosons , and the corresponding fermion momentum index is odd . the harmonic resolution @xmath55 is then also odd for the single - fermion state considered here .
the transverse direction requires an explicit cutoff @xmath27 , which we impose on individual light - cone energies @xmath61 the total transverse momentum is taken to be zero .
the integers @xmath62 and @xmath63 are limited to a range @xmath64 $ ] , which , along with the cutoff , determines the transverse scale @xmath53 .
given this discretization , the eigenvalue problem is converted to a matrix problem by a trapezoidal approximation to any momentum integrals .
we have found useful modifications @xcite which include non - constant weighting factors near the integral boundaries .
these weights are adjusted to compensate for the dlcq grid being incommensurate with the boundaries . in the present calculation , these weights are kept only in the two - body sector where maximal symmetry can be maintained
. for higher fock sectors , sensitivity to cancellations is of greater concern than boundary effects , and the weights disrupt the cancellations . unlike an ordinary eigenvalue problem ,
the presence of negatively normed states allows unphysical states to be lower in mass than the physical one - fermion state .
criteria must be employed to select the correct state in a numerical calculation .
we used the following : a positive norm , a real eigenvalue , a minimum number of nodes ( preferably zero ) in the parallel - helicity boson - fermion wave function , and the largest bare - fermion probability between 0 and 1 .
each of these characteristics can be computed without constructing the full eigenvector , provided one saves a few components of each lanczos vector @xmath65 ( see the appendix ) . as a check on the calculation
, we took advantage of an exact solution that exists for the unphysical situation of equal - mass pv bosons @xcite . in a particular ( null ) basis
the matrix representation is purely triangular .
each wave function of the dressed fermion is then directly computable in a finite number of steps . for comparison
, we also solved the problem using brillouin
wigner perturbation theory , for which @xmath66 the integrals were computed numerically with the same discretization as the nonperturbative dlcq calculation .
the main effort in the perturbative expansion is then matrix multiplication , just as for the lanczos algorithm .
various quantities can be computed from the wave functions @xmath67 for the different fock sectors .
we compute the slope of the no - flip form factor of the fermion , structure functions for bosons and the fermion , average momenta , average multiplicities , and a quantity sensitive to boson correlations .
the form factor slope @xmath68 is given by @xcite @xmath69\ , .
\nonumber\end{aligned}\ ] ] the physical boson structure function for bare helicity @xmath70 is defined as @xmath71 the normalization is such that the integral yields the average multiplicity @xmath72 .
we separate two pieces , for parallel and antiparallel bare helicity @xmath73 and have @xmath74 the average boson momentum is treated analogously , with @xmath75 as a measure of the correlations in the multiple - boson fock sectors , we compute @xmath76 where @xmath77 and @xmath78 is the same as @xmath79 except that only states with two or more bosons are included .
calculations at low resolutions tend to have difficulty for stronger coupling .
this can be seen already at third order in perturbation theory , where loop integrals are poorly approximated and subtractions between loops with different boson flavors magnify the errors .
fully averaged quantities such as @xmath72 are less affected by this , but the structure function @xmath80 can be quite poorly represented .
an example is given in fig .
[ fig : badfb ] . clearly one cure is to work at higher resolution .
limitations on computer storage then require truncation in the number of constituents .
luckily , even at the strongest coupling that we consider , states with large numbers of constituents are unimportant , and yet the results differ significantly from lowest - order perturbation theory .
the relative importance of different numbers of constituents can be seen in table [ tab : fsprobmax4 ] , where we list probabilities for the various fock sector contributions to a calculation with moderate resolution .
truncation to a maximum of two bosons is seen to offer a very good approximation . for weaker couplings ,
lower resolution is sufficient .
.[tab : fsprobmax4 ] fock sector probabilities @xmath81 , where @xmath82 , @xmath83 is the number of physical bosons and @xmath35 the number of pauli villars bosons of type @xmath36 .
the helicities @xmath84 and @xmath70 refer to the physical and bare fermion , respectively .
the numerical and physical parameters are @xmath85 , @xmath86 , @xmath87 , @xmath88 , @xmath89 , @xmath90 , @xmath91 , and @xmath92 .
probabilities smaller than @xmath93 are not resolved with any accuracy .
[ cols="^,^,^,^,^,^,^,^ , > " , ] as a check on the logarithmic pv coupling constraint in eq .
( [ eq : constraints ] ) , we compute the bare mass shift @xmath22 when @xmath29 is much less than @xmath94 .
figure [ fig : zerom2 ] shows its behavior as a function of the longitudinal resolution for various bare couplings .
if the logarithmic constraint was sufficient for nonperturbative calculation , @xmath22 should go to zero as @xmath29 approaches zero and the resolution approaches the continuum limit .
this appears to work well only for weak coupling .
we have used the discretized light - cone quantization method to successfully solve for the mass and light - cone wave functions of a dressed fermionic state in a yukawa theory in @xmath95 space - time dimensions .
no _ a priori _ constraint on the number of bosonic constituents was necessary ; however , since the fermion constituent was treated as heavy , states containing fermion anti - fermion pairs were truncated .
we have found that the eigensolution at strong coupling displays features which significantly deviated from first - order perturbation theory . numerical resolution in this domain of strong coupling was not a limiting factor in this analysis . the regularization procedure which we have chosen , with three pv scalars , functioned well .
however , better convergence of the dlcq method at strong coupling could be obtained by constraining the pv couplings non - perturbatively , rather than using the one - loop perturbative constraints in ( [ eq : constraints ] ) .
a number of properties of the yukawa - theory eigensolution could be extracted from its light - cone fock - state wave function .
this illustrates the power of the dlcq method in making the fock - state wave functions of the eigenstates explicitly available .
it is also possible to use the derived wave functions to compute the pauli and dirac form factors of the dressed fermion state at general momentum transfers @xcite .
there are additional calculations which might be done within the context of the zero fermion pair approximation .
we could consider two - fermion states and study true bound states and scattering solutions .
we could also consider dressed spin-3/2 states and the analog of n@xmath96 transitions .
extension of these methods to pseudo - scalar yukawa theory would make the n@xmath96 connection stronger .
although the approach used in this paper to solve ( 3 + 1)-dimensional yukawa theory at strong coupling has been successful , future progress for solving other quantum field theories will require more efficient analytic methods and numerical algorithms .
an alternative ultraviolet regularization procedure , using only one pv scalar and one pv fermion @xcite may potentially provide a more efficient approach for solving yukawa theories .
not only is the constraint on the couplings trivial , but the light - cone hamiltonian is also much simpler .
the simplifications occur because the instantaneous fermion interactions ( the terms in ( [ eq : hlc ] ) of order @xmath97 ) cancel .
moreover , the dlcq matrix for the remaining three - point interactions is much more sparse , allowing calculations at higher resolutions .
our next step will be to test this alternative regularization .
one can consider two other possibilities for pv regularization @xcite of full yukawa theory .
one is to use two heavy fermions and one heavy scalar @xcite ; the other is to use one less heavy fermion but make the transverse momentum cutoff part of the regularization rather than just a numerical procedure . in each case
a @xmath98 term must be added .
we plan to explore both of these methods .
quantum electrodynamics and quantum chromodynamics in physical space - time , including the phenomenon of chiral symmetry breaking remain the central challenge to dlcq methods .
one attractive possible approach is to use broken supersymmetry as an effective ultraviolet regulator of the light - cone hamiltonians of gauge theories .
the pv method also has applicability to the renormalization of non - abelian hamiltonian gauge theories on the light - cone .
paston , franke and prokhvatilov @xcite have recently extended their analysis to the nonperturbative regulation of light - cone qcd , including the regularization of the infrared singularities introduced by using light - cone gauge .
they find that a combination of light - cone gauge , pv fields , higher derivative regulation , and carefully chosen momentum cutoffs can regulate the theory in such a way as to provide agreement with feynman calculations using the mandelstam
leibbrandt prescription @xcite for the spurious singularity in the gauge propagator .
the resulting dynamical operator is rather complex and several regulating fields are needed , but the regulation procedures appear suitable for numerical calculations .
we plan to test these methods in abelian theory , including the calculation of positronium bound states @xcite and the non - perturbative calculation of the anomalous magnetic moments of leptons at strong coupling @xcite .
this work was supported in part by the minnesota supercomputing institute through grants of computing time and by the u.s .
department of energy , contracts de - ac03 - 76sf00515 ( s.j.b . ) , de - fg02 - 98er41087 ( j.r.h . ) , and de - fg03 - 95er40908 ( g.m . ) .
the ordinary lanczos algorithm @xcite was designed for diagonalization of real symmetric or hermitian matrices .
a more general form , the biorthogonal lanczos algorithm @xcite , can be applied to non - symmetric cases but is quite cumbersome . in the case of a complex symmetric matrix
, the biorthogonal algorithm can be reduced to a form nearly as simple as the real symmetric case @xcite ; this approach was used in previous work @xcite where imaginary couplings made negative norms unnecessary .
the complex - symmetric approach is not easy to implement for yukawa theory because the hamiltonian is fully hermitian . instead
, negative norms are assigned , and the eigenvalue problem becomes one with indefinite metric . for this case
the biorthogonal algorithm can still be reduced to a simpler form .
let @xmath99 represent the metric signature , so that numerical dot products are written @xmath100 .
the hamiltonian matrix @xmath101 is constructed to be self - adjoint with respect to this metric , which means that @xcite @xmath102 .
the lanczos algorithm for the diagonalization of @xmath101 then takes the form @xmath103 where @xmath104 is taken as a normalized initial guess and @xmath105 .
this initial guess is generated with use of high - order brillouin
wigner perturbation theory .
to determine when to stop the lanczos iterations , the convergence of the eigenvalue and parts of the boson - fermion wave function are monitored . just as for the ordinary lanczos algorithm , the original matrix @xmath101 acquires the following tridiagonal matrix representation @xmath106 with respect to the basis formed by the vectors
@xmath65 : @xmath107 by construction , the elements of @xmath106 are real .
the new matrix is also self - adjoint , but with respect to an induced metric @xmath108 .
the eigenvalues of @xmath106 approximate some of the eigenvalues of @xmath101 , even after only a few iterations .
approximate eigenvectors of @xmath101 are constructed from the right eigenvectors @xmath109 of @xmath106 as @xmath110 .
brodsky , `` qcd aspects of exclusive b decays , '' hep - ph/0104153 .
m. beneke , g. buchalla , m. neubert , and c.t .
sachrajda , nucl .
b * 591 * , 313 ( 2000 ) , hep - ph/0006124 .
keum , h .-
li , and a. i. sanda , phys .
d * 63 * , 054008 ( 2001 ) , hep - ph/0004173 .
s.j . brodsky and d.s .
hwang , nucl .
b * 543 * , 239 ( 1999 ) , hep - ph/9806358 .
brodsky , m. diehl , and d.s .
hwang , nucl .
b * 596 * , 99 ( 2001 ) , hep - ph/0009254 .
m. diehl , t. feldmann , r. jakob , and p. kroll , nucl .
b * 596 * , 33 ( 2001 ) , hep - ph/0009255 ; * 605 * , 647 ( 2001 ) . for a review ,
see s.j .
brodsky , `` hadronic light - front wave functions and qcd phenomenology , '' hep - ph/0102051 . o. lunin and s. pinsky , phys .
d * 63 * , 045019 ( 2001 ) , hep - th/0005282 .
gross , a. hashimoto , and i.r .
klebanov , phys .
d * 57 * , 6420 ( 1998 ) , hep - th/9710240 . f. antonuccio , o. lunin , s. pinsky , and a. hashimoto , jhep * 07 * , 029 ( 1999 ) ; j.r .
hiller , o. lunin , s. pinsky , and u. trittmann , phys .
b * 482 * , 409 ( 2000 ) ; j.r .
hiller , s. pinsky , and u. trittmann , phys .
d * 63 * , 105017 ( 2001 ) .
paston , v.a .
franke , and e.v .
prokhvatilov , theor .
phys . * 120 * , 1164 ( 1999 ) [ teor . mat .
fiz . * 120 * , 417 ( 1999 ) ] , hep - th/0002062 .
paston , e.v .
prokhvatilov , and v.a .
franke , hep - th/0011224 .
s. mandelstam , nucl .
* b213 * , 149 ( 1983 ) ; g. leibbrandt , phys .
d * 29 * , 1699 ( 1984 ) .
|
we apply pauli villars regularization and discretized light - cone quantization to the nonperturbative solution of ( 3 + 1)-dimensional yukawa theory in a single - fermion truncation .
three heavy scalars , including two with negative norm , are used to regulate the theory .
the matrix eigenvalue problem is solved for the lowest - mass state with use of a new , indefinite - metric lanczos algorithm .
various observables are extracted from the wave functions , including average multiplicities and average momenta of constituents , structure functions , and a form factor slope .
# 1 _ # 1
|
heavy ion reactions at intermediate energy offer great promise to measure spectroscopic properties of nuclei far from stability , but one needs a tractable reaction theory to interpret the experiments . in this respective the availability of higher energy heavy ion beams is most welcome , because it becomes a reasonable theoretical approximation to neglect exchange of nucleons between the colliding nuclei .
one can therefore consider the interaction in each nucleus as that of an external ( complex ) potential field . within the framework of this basic approximation and with given potentials , we will here address the question of the accuracy of further simplified models of the reaction cross sections .
a number of theoretical models have been proposed and calculated @xcite-@xcite in which different approximations were made . in this work we focus on two of the models , the eikonal model @xcite-@xcite and the transfer - to - the - continuum ( tc ) model @xcite-@xcite .
there are several cross sections that are measured and calculated in the models .
the simplest measurement is the single - neutron removal cross section , in which only the projectile residue , namely the core with one less nucleon , is observed in the final state . besides the integrated removal cross section , denoted by @xmath0 ,
the differential momentum distribution @xmath1 is also measured .
a particularly useful cross section is @xmath2 , the removal cross section differential in longitudinal momentum .
if the final state neutron can also be measured , the corresponding coincident cross section @xmath3 is called the diffractive breakup cross section . the difference between the removal and diffractive breakup
is called the stripping cross section .
this paper is organized as follows . in sec .
ii we summarize the essential ingredients of the tc model and the eikonal model , and discuss the accuracy of the neglect of finite interaction times in the eikonal model . in sect iii we discuss the different treatment of the neutron - target s - matrix in the two models .
we will not discuss the accuracy of the hankel function approximation separately , but it of course plays a role in the comparison of cross sections that we make in section iv which contains also our conclusions .
all theoretical methods used so far rely on a basic approximation to describe the collision with only the three - body variables of nucleon coordinate , projectile coordinate , and target coordinate .
thus the dynamics is controlled by the three potentials describing nucleon - core , nucleon - target , and core - target interactions . in most cases the projectile - target relative motion is treated semiclassically by using a trajectory of the center of the projectile relative to the center of the target @xmath4 with constant velocity @xmath5 in the @xmath6 direction and impact parameter * b@xmath7 * in the @xmath8 plane . along this trajectory
the amplitude for a transition from a nucleon state @xmath9 to a state @xmath10 is given by @xmath11 where @xmath12 is the neutron - target interaction .
the state @xmath9 will be the bound state of the nucleon in the projectile , while the final state @xmath10 is a continuum state .
the detailed derivation of eq.([1 ] ) from a scattering amplitude containing the full time dependent propagator can be found in sec.ii of ref .
there it was shown to hold under the hypothesis that the breakup process is limited to peripheral projectile - target trajectories and that it is due mainly to the neutron interaction with the target potential .
the probabilities for different processes can be represented in terms of the amplitude as @xmath13 where @xmath14 can be momentum , energy or any other variable for which one measures a differential cross section .
the effects associated with the core - target interaction will be included by multiplying the above probability by the @xmath15-dependent probability for the core to be left in its ground state . thus the differential cross sections with respect to the longitudinal momentum
is @xmath16 where @xmath17 is the ( longitudinal ) recoil momentum of the neutron ( see eq .
( 2.3 ) of @xcite ) and @xmath18 is the spectroscopic factor for the initial single particle orbital . the cross section can be further divided into a stripping cross section @xmath19 and a diffractive breakup cross section @xmath20 depending on whether the removed neutron is detected in the final state or not .
we accordingly will consider these individual probability distributions , @xmath21 and use a similar notation for the cross sections .
we first summarize the transfer - to - the - continuum ( tc ) model .
this model treats the time - dependence of the reaction explicitly , thus conserving energy .
it uses the on - shell neutron - target scattering matrix , therefore making it in principle model - independent with respect to that interaction .
it also makes use of the asymptotic form of the neutron wave function in the projectile .
this is an asset in that the formulas have an analytic limit , but a disadvantage in that the results are only reliable at peripheral impact parameters .
the two breakup probabilities are given by the following expressions : @xmath22 @xmath23 where the factor @xmath24 is a transfer probability which depends on the details of the initial and final states , and on the energy of relative motion .
it is given by @xmath25 where @xmath26 is the angular momentum of the bound neutron in the initial state with respect to the core .
the variable @xmath27 has the interpretation as the angular momentum of the neutron with respect to the target .
also @xmath28 the arguments of the legendre polynomials @xmath29 and @xmath30 are @xmath31 and @xmath32 , @xmath33 and @xmath34 .
@xmath35 is related to the initial state binding energy by @xmath36 .
the variable @xmath37 has the interpretation as the modulus of the transverse component of the neutron momentum . in our notation @xmath38
is the energy of the neutron relative to the target in the final state . for diffraction
this is the same as the final laboratory energy of the neutron if the target recoil kinetic energy is neglected . in the case of stripping ,
if it goes through compound nucleus formation , @xmath38 is the excitation energy of the compound state above the neutron threshold in the residual nucleus . for inelastic scattering
it is the energy of the breakup neutron before it scatters from the target .
this is equivalent to the sum of the excitation energy of the target final state and the final neutron energy relative to the target .
if the target recoil kinetic energy is neglected the final kinetic energy @xmath39 of the ejectile is given by the energy conservation condition @xmath40 where @xmath41 is the initial incident energy of the projectile in the laboratory , q is the reaction @xmath42 value given by @xmath43 and @xmath44 is the initial neutron binding energy in the projectile . with this approximation eq .
( [ ec ] ) relates @xmath17 to the the projectile residue parallel momentum . finally , @xmath45 is the asymptotic normalization constant of the initial bound wave function @xmath46 it is obtained by fitting a realistic radial wave function to the hankel form @xmath47 outside the potential radius . in this way
the transfer to the continuum results are model dependent . on the other hand
the dimensionless quantity @xmath48 has been called reduced normalization in the contest of spectroscopy done with transfer reactions @xcite .
if one considers @xmath49 as the prefactor of the theoretical cross section obtained with the hankel function , the ratio between the experimental cross section and the theoretical cross section would determine its value .
eqs.([5]-[7 ] ) were derived assuming no overlap between neutron - core and neutron - target potential .
this assumption can be avoided if one makes an eikonal approximation to the basic expression for the amplitude , eq.([1 ] ) , as shown in appendix b. then one derives @xcite
@xmath50 |\tilde { \psi}_i ( { \bf b_n - b_c } , k_z)|^2 , \label{pg}\ ] ] where @xmath51 is the transverse coordinate of the neutron with respect to the target .
the neutron - target @xmath52-matrix is approximated by the eikonal form @xmath53 , related to the optical potential @xmath12 by @xmath54 .
finally , @xmath55 is the longitudinal fourier transform of the initial state wave function .
also the connection between eqs.([5]-[7 ] ) and eq.([pg ] ) is made by replacing the sum over partial waves in eqs.([5],[6 ] ) by the integral over impact parameters as in eq.([pg ] ) , and by evaluating the longitudinal fourier transform in eq.([pg ] ) using the asymptotic initial state wave function , eq.([in ] ) .
the one - dimensional fourier transform of the initial wave function can be calculated analytically in the case of an hankel function approximation eq.([in ] ) yielding @xmath56 where @xmath57 .
the total breakup probability is obtained from the integral of eq.([pg ] ) involving @xmath58 @xmath59 and @xmath60 are the kinematically allowed minimum and maximum neutron parallel momenta discussed in the following .
therefore although eq.([pg ] ) describes the neutron - target rescattering in the eikonal approximation , it still satisfies neutron energy and momentum conservation . on the other hand ,
a sudden approximation or frozen halo form of the eikonal method has been used in @xcite where energy conservation was neglected .
this approximation can be derived from eq.([pg ] ) when the integration limits in eq.([phsp ] ) can be extended to @xmath61 .
then eq.([phsp ] ) is just be the longitudinal density , and the formulae for the tc and eikonal model become identical .
in fact in this limit the removal cross section reduces to @xmath62 |s_{ct}(b_c)|^2|{\psi}_i ( { \bf b_n - b_c},z)|^2,\label{creik}\ ] ] which is consistent with the breakup cross section of @xcite . in eq.([creik ] ) @xmath63 is an eikonal s - matrix , as defined after eq.([pg ] ) . the steps necessary to obtain eq.([creik ] ) from eq.([pg ] ) and eqs.([4]-[6 ] ) can be justified in the high energy limit as follows .
if the neutron binding energy is not too large , the final energy or momentum distributions are strongly peaked at the incident energy per nucleon .
therefore it is possible to average the dependence of the neutron target optical model s - matrix of eqs.([5],[6 ] ) over the full range of neutron continuum energies and assume that the s - matrix can be approximated with the eikonal values @xmath63 obtained at @xmath64 . on the other hand , for the eikonal approximation to be good it is necessary that the parallel component of the neutron momentum @xmath17 be large with respect to the transverse component @xmath65 . in particular such condition
must be satisfied for the minimum values of both . from the definition of @xmath17 it is easy to see that its lowest possible value is @xmath66 in correspondence to @xmath67 . and that the minimum value of @xmath68
. therefore we get @xmath69 if @xmath70 . for a real halo with separation energy around
0.5mev the condition
@xmath69 is satisfied at all initial energies .
increasing the binding energy it is necessary to go to higher incident energies in order for the parallel momentum component to be larger than the transverse one . for a typical
binding of 10mev or more the conditions for the sudden eikonal approximation to be valid are satisfied from about 80a mev .
it is useful also to consider the values of neutron - target center of mass momentum @xmath71 where @xmath72 is the neutron - target reduced mass . here
too we see that for a typical diffuseness of a=0.5fm the semiclassical condition @xmath73 is satisfied starting from about 80mev .
finally if such condition is satisfied we can extend the lower limit in the @xmath17 integral to @xmath74 .
there is however also an upper limit to the final neutron or ejectile momentum value in the parallel direction , due to the energy and momentum conservation .
it was discussed in ref.@xcite where it was shown that @xmath75 m is the neutron mass .
@xmath76for a @xmath77 projectile . in terms of the maximum neutron final energy in the continuum
the above condition reads @xmath78 if this upper limit is also extended to @xmath79 then the @xmath17 integral gives a @xmath80-function and inserting eq.([pg ] ) and the damping factor @xmath81 in eq.([cross ] ) we finally get eq.([creik ] ) .
these very strict conditions necessary to extend the theoretical limits of the @xmath17 integral to infinity , can be somewhat revised in practical calculations . to see how accurate the sudden approximation is ,
we have calculated the integral @xmath82 under various conditions of angular momentum , neutron binding energy in the projectile , and projectile velocity . for values of the parameters of interest
there can be a rather large reduction for small values of the neutron transverse radius in the projectile , @xmath83 .
however , the approximation becomes increasingly accurate as the transverse radius is made larger . in fig .
1 we show the regions of energies in which the ratio @xmath84 is within 5% , 5 - 10% and 10% of being unity , for a @xmath85-wave orbital at a transverse radius corresponding to strong absorption .
the region of good agreement is somewhat larger for @xmath86- and @xmath87-waves .
thus it seems that the sudden approximation should be acceptable for weakly bound neutron orbitals ( @xmath88 ) and beam energies greater than 40a mev , provided the reaction samples the initial wave function at distances rather larger than the sum of the radii of the reaction partners .
it must be noted that here as in eq.([cross ] ) we have adopted the so called no - recoil approximation @xcite which consists in factorizing out from the matrix element the core - target s - matrix .
another important remark has to do with the extension of the lower limit in the @xmath17 integral to @xmath74 .
values of @xmath17 smaller than @xmath59 obtained when @xmath67 would correspond to final states in the projectile below the breakup threshold for the diffraction term . for the stripping term they would correspond to final bound states in the target .
taking @xmath89 is analogous to the completeness relation introduced in @xcite .
for this reason in eq.([creik ] ) the diffraction term has usually been corrected by subtracting terms like @xmath90 . in the transfer to the continuum model
this correction is not as important because the neutron final energy is always positive and the @xmath17 integration is performed in the kinematically allowed region .
estimates of absolute breakup cross sections published so far have been made either using eq.([cross ] ) with the breakup probability given by the transfer to the continuum theory eqs.([4]-[8 ] ) @xcite or with eq.([creik ] ) @xcite .
there are a number of assumptions contained in both models , the validity of which we are going to study in this paper .
experimentalists often use light targets such as @xmath91 or @xmath92 to study spectroscopy by projectile fragmentation reactions .
the definition of an optical potential for light targets is a very delicate issue which has been discussed in the literature for long time @xcite .
there are two main issues , i ) global parameterizations are based on large nuclei ; ii ) the imaginary optical potential changes drastically for light nuclei .
recently these problems have been faced in the contest of halo breakup studies @xcite where different choices have been made by different authors .
bertsch et al.@xcite used the varner et al .
parameterization @xcite .
such parameterization has been obtained taking into account relatively low energy proton and neutron cross sections ( 10 - 26mev ) on heavy targets , a=40 - 209 . in ref.@xcite
a neutron target s - matrix was constructed in the optical limit of glauber theory .
the corresponding optical potential does not need to be given explicitly , but the behavior of the transmission coefficient given in @xcite would suggest a strongly absorbing potential of volume type .
bonaccorso @xcite extended to high energies a parameterization of a phenomenological n-@xmath91 optical potential fitted to low energy data .
this potential seemed however to overestimate the free particle cross sections at high energies @xcite and therefore ref.@xcite also considered a microscopic optical potential calculated according to the method of jeukenne , lejeune and mahaux ( jlm ) @xcite .
the jlm potentials are more complicated to calculate and to use than a simple standard parameterization .
then for the purpose of the present paper we have attempted to fit the available experimental n-@xmath91 total cross sections with a new potential of the standard form .
the potential is given by a real woods - saxon well and it has both volume and a surface derivative woods - saxon forms for the imaginary part . in obtaining the parameters given in table i
we have been guided by some existing parameterizations @xcite obtained by fitting low energy data .
we have modified such parameterizations to get a smooth behavior of the free particle neutron-@xmath91 cross section in the energy range 10 - 180 mev .
the free - particle cross section have been obtained in an optical model calculation .
a good agreement of calculated free neutron angular distributions with the data of ref.@xcite has also been obtained .
this potential satisfies very well the subtracted dispersion relations as given for example in ref.@xcite at energies larger than 40mev . for lower energies
the agreement is less good at small radii . in fig .
2 , top left , we show the experimental cross sections @xcite together with the optical model cross sections , where the full curve is the total cross section , the dot - dashed curve is the elastic cross section while the dashed curve is the reaction cross section . at the bottom
left we show the same quantities calculated in the eikonal approximation .
the effect of the eikonal approximation is to reduce the calculated cross sections , in particular at the lower energies .
the reduction is more pronounced for the elastic cross section , probably because quantum mechanical reflection effects on the potential barrier can not be reproduced by the eikonal approximation . for comparison
we show at the top right and bottom right respectively the same quantities calculated with the varner parameterization .
varner potential does not reproduce the high energy data , also it gives a dominance of the elastic scattering up to about 80mev . here too the eikonal calculations underestimate the elastic scattering at low energies .
our potential has a rather strong surface term .
it gives also large volume integrals ( @xmath93 ) in accordance with the light nuclei systematics@xcite .
the microscopic origin of it can be understood since @xmath91 is weakly bound and it has rather high breakup probability .
also it has been known for long time @xcite to have an unusually large mean square radius and to be strongly deformed .
finally we notice that we get a very large elastic vs reaction cross section ratio at low energy .
the jlm potential gives the same behavior .
the little experimental free particle data available show similar trend and it would be interesting to see if breakup reactions around 20a mev which will soon be feasible at ganil , will parallel such a behavior . in fig .
3 we show the behavior of the term @xmath94 and of the transmission coefficient @xmath95 when calculated by the optical model ( solid and dashed line respectively ) and by the eikonal approximation ( dotted and dot - dashed line ) . the results are shown as a function of the angular momentum @xmath27 and in the case of the eikonal calculations the semiclassical relation @xmath96 was used to make the connection between the angular momentum and the neutron impact parameter . we give results at energies ranging from 20 to 100 mev . an interesting characteristic of these results is that the eikonal seems to concentrate the scattering at larger impact parameters than in the optical model case .
such effects have been seen by other authors @xcite in different situations that ours and sometimes it has also been shown @xcite that higher order eikonal corrections can improve the agreement with the quantum mechanical calculations .
it is possible that in our case the effect is amplified by the deep surface part of the imaginary potential .
it is well known that the eikonal approximation works well in the presence of strongly absorbing volume potentials and differences would also be less extreme for potentials that have more diffuse edges .
this is because the reflection effects at the barrier that modify the neutron trajectory are not taken into account in a semiclassical approach . from the differences seen here
, we would expect quite large differences between predictions of the two treatments of the s - matrix .
the quantification of these differences is the subject of section iv .
we start this section by showing some results for the total probability @xmath97 obtained from the integrand of eq.([cross ] ) after integrating over @xmath17 and using eq.([prs ] ) .
the @xmath98 values are given in table iii at each energy .
they were obtained , as explained in appendix a , from the s - matrix calculated by folding the neutron - target optical potential of table i with the projectile density , at all energies but 20a mev . at such a low energy the optical limit of the eikonal model can not be justified , but calculations by carstoiu @xcite in second order eikonal approximation , folding the jlm potentials agree with our @xmath98 at all energies , including 20a mev .
same values are also obtained from the optical potential of @xcite . for our numerical comparison of the various methods
, we have chosen to study the breakup from the s and p states in @xmath77 which were recently measured by navin et al.@xcite .
the initial bound state parameters are given in table ii . in fig .
4 we show the total probabilities for diffraction obtained in the tc calculation by the solid lines .
the dashed lines are for the absorption or stripping .
energies are as in fig .
the dotted lines are the calculations from the diffraction integrand of eq.([pg ] ) while the dot - dashed lines are for the stripping . because of the shift in the s - matrices of fig .
3 , we see here that the eikonal calculations have their maxima at slightly larger core - target distances than the optical model calculations . at the lowest energy the optical model gives almost equal diffraction and stripping probabilities .
the eikonal calculation gives a rather larger diffraction probability . in all other cases
the stripping is the dominant term .
this effect was first pointed out in normally bound nuclei @xcite and it has been confirmed by some experiments @xcite and other theoretical calculations @xcite .
the tc calculations discussed in this paper are performed using the asymptotic , hankel form of the initial state wave function , which can be the origin of some divergence if the projectile and target potential have an important overlap during the reaction .
this problem was discussed in great detail in ref.@xcite where it was shown to be more important for the stripping term since it is proportional to the neutron target phase shift , while the diffraction term , being proportional to the square of the phase shift converges more rapidly .
this situation reflects the fact that diffraction reactions involve a more peripheral part of the initial wave function than stripping reactions . here
we can easily see from eq.([pg ] ) that the integral over @xmath99 , the neutron target impact parameter is well behaved provided the neutron - target potential and the corresponding s - matrix fall off more rapidly than the tail of the initial wave function . a practical way to satisfy such a condition and to avoid any divergence in the sum over neutron partial waves in eq.([6 ] )
is to use a small value for the diffuseness of the imaginary part of the optical model ( cf .
sec.iv of @xcite ) . otherwise , as it is shown in appendix b it is possible to start again from eq.([1 ] ) and obtain a formula which contains the realistic single particle wave function and an optical model s - matrix , still keeping the information on the proper kinematical limits for the neutron parallel momentum .
the only hypothesis involved is that high neutron angular momenta are dominant . to show the amount of error that is introduced by the hankel approximation in the tc method
, we show in fig .
( 5 ) the integrand functions of the stripping and diffraction terms of eq.([pg ] ) , after integrating over @xmath17 , obtained from the realistic bound state wave function and the corresponding terms in the sum over partial waves of eqs.([5],[6 ] ) in the case of the incident energy of 78a mev .
the calculations were done at the fixed impact parameter @xmath100 between the projectile and target .
left figure is for the diffraction term while right figure is for the stripping term .
we can see that the calculation for stripping according to eq .
( [ 6 ] ) which is done with the hankel function and optical model s - matrix ( diamonds ) shows a slower decrease in the interior of the projectile ( small b@xmath101 ) than the calculation with the realistic wave function ( crosses ) done according to eq.([a10 ] ) of appendix b. therefore the comparison between these two results gives an indication of the effect of the hankel function approximation . on the other hand the eikonal calculations ( full curve ) , done with the full wave function have higher maxima which occur at a larger distance than the optical model calculation .
this effect is due to the behavior at large impact parameters of the eikonal s - matrix , as we have discussed before .
the comparison of the full curve with the crosses gives an indication of the effect of the eikonal approximation . in the end
the integrated probabilities obtained with the three possible approximations contained in eqs .
( [ 5 ] ) and ( [ 6 ] ) , eq.([a10 ] ) and eq.([a7 ] ) differ by not more than about 10% .
this leads to about 20% difference in the total cross sections because of the further integration over @xmath15 , the core - target impact parameter . from the peak values of the diffraction and stripping curves
it is easy to see that the stripping is almost three times larger than the diffraction .
the values of the total cross sections are given in table iii . because the eikonal probabilities are shifted towards larger distances we get total cross sections which are larger by about 20% than the tc calculations at all energies but 20a mev .
at this lowest energy two effects tend to compensate each other : the eikonal free particle cross section is largely underestimated while there is an overestimate of the phase space kinematically allowed .
then the eikonal and optical model calculations seem to give very close results .
considering all the effects discussed above we conclude than present models used to calculate the breakup cross section might tend to overestimate the true value , but the error should be of the order of 20% or less .
on the other hand presently available experimental data are inclusive with respect to the target which was unobserved . in the case of @xmath91 it
is possible that some part of the cross section comes from reactions in which the target itself underwent breakup .
these reactions could account for about 20% of the measured cross sections and they are certainly not taken into account by presently available theoretical models .
we suggest therefore that future experiments detect the final excitation state of the target in coincidence with breakup events ( as done for example in ref.@xcite for `` normal '' heavy ions ) and at the same time that theoretical models be improved through a more realistic treatment of the core - target interaction , as mentioned in sec.iv and in appendix a. we propose also an improved calculation of the breakup amplitude using the new formula eq.([a10 ] ) derived in appendix b that combines the better treatment of the wave function in the eikonal model with the better treatment of the target interaction in the transfer - to - continuum model .
we leave to a future study to apply this equation to the reaction data .
ejectile momentum distributions are not shown in this paper .
they have been discussed in great detail in @xcite where it was shown that there can be noticeable differences in the results of the tc and eikonal methods .
present data do not have enough statistics to distinguish clearly between the two models . *
acknowledgments * this research was supported by the u.s .
department of energy under grant de - fg03 - 00-er41132 .
we wish to thank david brink for several discussions and suggestions , florin carstoiu for communicating his results before publication and juan carlos pacheco for checking that our n - target optical potential satisfies the dispersion relations .
we discuss here the theory of the core - target s - matrix , needed to calculate @xmath102 in eq.([cross ] ) .
the best source of information comes from the ion - ion reaction cross section which is related to the core - target s - matrix by @xmath103 here we have assumed that the semiclassical replacement of the partial wave sum by an integral over @xmath104 is permitted . unlike the nucleon - nucleus scattering , this is always a good approximation , because of the smaller wave length involved , and one can safely calculate this s - matrix in the eikonal approximation .
the usual procedure is to define an optical potential for the core - target scattering , and calculate the s - matrix from it . in practice the behavior of the s - matrix
is determined by two parameters .
the first and most important is the strong absorption radius @xmath98 , defined as the distance of closest approach for a trajectory that is 50% absorbed from the elastic channel .
the reaction cross section practically speaking is determined by the strong absorption radius .
the next most important parameter is the thickness @xmath105 of the absorption region . in optical model fits , it is closely related to the asymptotic behavior of the imaginary potential @xmath106 . just as the reaction cross section
strongly constrains the strong absorption radius , measurements of the elastic angular distribution provide information about @xmath105 . in this work
we shall use a simple parameterized form @xcite for @xmath107 , namely @xmath108 it would be correct to use the elastic scattering s - matrix if the experimental cross sections were reported for projectile breakup leaving the target in its ground state .
in fact the target final state is difficult to measure and most experimental cross sections are inclusive with respect to the target final state
. thus the measured cross sections should be somewhat larger than one would calculate with an elastic core - target s - matrix .
a way to include absorption of the projectile core without regard to the target is to construct an optical potential convolution of a nucleon - ion potential with the target density . in the calculations of this work
we have determined strong absorption radii by this method , using the optical potential of sec .
we also take a diffuseness parameter @xmath109 .
in this section we give a more general formulation of the breakup problem which is less dependent on the reaction model assumptions of the text .
we start with eq .
( 1 ) , again assuming a straight - line trajectory . using a galilean transformation for the initial wave function as described in ref.@xcite , eq .
( 1 ) becomes the @xmath116 integral in eq.([a3 ] ) can be calculated by parts when the initial binding energy is small and the final neutron energy is close to the neutron incident energy and the neutron scattering angle is small such that @xmath117 . then this is unfortunately off shell but we will do the following high energy approximations similar to what we have done above in the eikonal case . in the large angular momentum , small scattering angle limit , define @xmath124 , we can substitute the partial wave sum with an integral and also use @xmath125 where @xmath126 is the transverse component of @xmath114
. then .2 in * fig.1 * ratio of phase space integrals with and without momentum cutoff , for a @xmath85-wave neutron wave function .
the effect of the cutoff is to include less than 90% , between 90% and 95% and more than 95% of the initial momentum distribution as marked on the figure . .2 in * fig.2 * top left : experimental cross sections for neutron scattering on @xmath91 together with the optical model cross sections using the potential of table i. the full curve is the total cross section , the dotdashed curve is the elastic cross section while the dashed curve is the reaction cross section .
bottom left : same quantities calculated in the eikonal approximation .
top and bottom right : the same quantities calculated with the varner potential .
.2 in * fig.3 * the behavior of the term @xmath94 and of the transmission coefficient @xmath95 calculated by the optical model ( solid and dashed line respectively ) and by the eikonal approximation ( dotted and dotdashed line ) .
results are shown as a function of the angular momentum @xmath27 . in the case of the eikonal calculations
the semiclassical relation @xmath96 was used to make the connection with the neutron impact parameter .
results are given at incident energies ranging from 20 to 100 mev .
.2 in * fig.4 * total breakup probabilities from the 1/2@xmath130 state in @xmath77 , as a function of the ion - ion impact parameter for diffraction obtained in the tc calculation by the solid line .
the dashed line is for the stripping .
energies are as in fig .
3 . dotted and dotdashed lines are diffraction and stripping probabilities from the eikonal calculation .
.2 in * fig.5 * the integrand function of the diffraction ( a ) , and stripping ( b ) term of eq.([pg ] ) after @xmath17 integration , full curve , obtained from the realistic bound state wave function and the corresponding terms , diamonds , in the sum over partial waves of eqs.([5],[6 ] ) in the case of the incident energy of 78a mev .
crosses are the results of a calculation of eq .
( [ pg ] ) in which the eikonal phase shifts have been substituted by the optical model phase shifts according to eq.([a10 ] ) .
all calculations done at fixed impact parameter @xmath100 between the projectile and target .
|
spectroscopic properties of nuclei are accessible with projectile fragmentation reactions , but approximations made in the reaction theory can limit the accuracy of the determinations .
we examine here two models that have rather different approximations for the nucleon wave function , the target interaction , and the treatment of the finite duration of the reaction .
the nucleon - target interaction is treated differently in the eikonal and the transfer - to - continuum model , but the differences are more significant for light targets .
we propose a new parameterization with that in mind .
we also propose a new formula to calculate the amplitude that combines the better treatment of the wave function in the eikonal model with the better treatment of the target interaction in the transfer - to - continuum model . *
pacs * number(s):25.70.hi , 21.10gv,25.60ge,25.70mn,27.20+n * key - words * breakup , absolute cross sections , eikonal approximation , optical potential , s - matrix , t - matrix .
|
many applications require realistic 3d human shapes .
for instance , 3d human models are used to design products that fit a target population .
typically , these shapes need to have certain characteristics or to be samples from a population .
although it is possible to digitize humans using 3d imaging technologies , it is impractical to find and scan suitable human subjects for each individual application . on the other hand
, there is a long history of using anthropometric measurements to describe the human shape .
these measurements are linear and curvilinear distances between anatomical landmarks or circumferences at predefined locations . in spite of providing limited shape information , they are readily available , and therefore widely used .
intuitively , the linear measurements should encode information about the body shape .
if accurate 3d human shape can be inferred from a simple set of measurements , traditional anthropometric data can be leveraged to create 3d datasets without resorting to expensive scanning equipment .
the early attempts at solving this problem deform a generic human model to fit the measurement data .
however , the models thus obtained are not necessarily human forms because generating a human shape from a sparse set of measurements is an under - constrained problem .
more sophisticated methods use knowledge of the human shape learned from a database of 3d scans . a statistical model can be built from a set of registered 3d scans and a relationship between the measurements and the shape space can be established . in doing so , detailed 3d shapes can be predicted from sparse and partial shape data .
apart from predicting 3d shapes from measurements @xcite , other partial shape information , such as marker positions @xcite and 2d images @xcite , has also been used to reconstruct 3d shapes .
existing methods that predict the 3d shapes using a statistical model limit the generated shapes to the space spanned by the shape model ( more specifically , within two standard deviations of the shape variability ) and local variations of human shapes that are outside this space can not be predicted accurately .
this means that these methods require a 3d database that accurately represents a target population ( for instance , human subjects of different ethnicities and different weight classes must be present in the database ) . to apply these techniques , one must carefully choose a target population and acquire a representative database .
in reality , the acquisition of a 3d anthropometry data that represents a target population well is complicated and expensive ( the acquisition of the civilian american and european surface anthropometry resource took 4 years and cost $ 6 million ) . in this paper
, we address the problem of estimating 3d human body and face shapes given a set of anthropometric measurements .
our main goal , in contrast to previous work , is to predict shapes that are inside the space of human shapes , but still account for local variations not captured by the training data .
this extrapolation allows us to create human shapes based on relatively small 3d databases .
as most of the previous approaches , our method does not take posture or expression changes into account .
that is , we assume that all of the training data is in a standard pose .
this scenario is commonly assumed since in a typical 3d anthropometry survey , the human subjects are asked to maintain a standard posture .
the rest of the paper is organized as follows .
section [ approach ] gives an overview of the approach , and section [ refinement ] outlines the details of a shape refinement step , which constitutes the main contribution of this paper .
we aim for design applications in engineering where accuracy is essential .
we show , through experiments , that straightforward application of existing techniques does not produce satisfactory results ( section [ experiments ] ) .
this section reviews existing work that aims to estimate 3d human shapes based on partial input information .
we categorize the related literature by the type of input data that is used to predict human shapes .
decarlo et al .
@xcite presented an early attempt at using anthropometric measurements to compute human face shapes .
the method generates a random set of measurements based on a set of rules that capture , for instance , typical facial proportions .
to create a 3d face shape , the method deforms a template model to fit these measurements using a free - form deformation technique called variational modeling .
variational modeling allows to give the measurements as constraints when deforming the template . to ensure that the estimated shape remains in the shape space of human shapes ,
much care is taken when creating the measurements that are used as constraints .
while this method generates human - like face shapes using a statistical model , the accuracy of the shapes created by this method is limited by the sparse set of measurements used to represent the training data .
that is , the results are overly smooth and do not contain realistic details .
it is not straight forward to extend this technique to allow for the generation of human body shapes from an arbitrary set of anthropometric measurements . by combining statistical learning with a mesh - based technique ,
our method allows the generation of realistic and detailed human body shapes from an arbitrary set of anthropometric measurements .
the following approaches estimate human body shapes from a discrete set of measurements on the body , all proceeding by learning a linear or near - linear mapping between the training set of human shapes and the parameter space of measurements and then using this mapping to predict a shape based on a new set of measurements .
while these approaches work well when the predicted shape is inside the shape space spanned by the training set , they do not allow for extrapolations from this shape space .
wang @xcite uses a parameterized database of human shapes consisting of _ feature patches _ to find a new human shape based on a given set of measurements .
feature patches are initialized to smooth patches that are subsequently refined to capture geometric details of the body shape .
while this refinement does not maintain the smoothness of the patches , the final results are visually smooth and do not contain realistic localized shape details .
the approach finds models in the database with measurements similar to the given set of measurements and computes the new human shape as a linear combination of these models .
wei et al .
@xcite use a similar approach , but the models are represented using a layer - based representation .
allen et al .
@xcite compute a new triangular mesh based on a given set of measurements . starting from a parameterized database of human bodies in similar poses , the approach performs principal component analysis ( pca ) of the data
this yields one pca weight for each training shape .
the training database is used to learn a linear mapping from the set of measurements measured on the training data to the pca space .
this mapping is called _
feature analysis_. feature analysis can be used to compute a new pca weight based on a new set of measurements , and the learned pca model allows to compute a new triangular mesh from this pca weight .
chu et al .
@xcite perform feature analysis on a database of human shapes consisting of feature patches to find a new model consisting of smooth patches .
seo and megnenat - thalmann @xcite represent the human body as a triangular mesh with an associated skeleton model . as with the approach of allen et al .
, this approach reduces the dimensionality of the data using pca .
this yields a set of pca weights .
the approach learns a mapping from the set of measurements measured on the training data to the pca space using an interpolating radial basis function ( rbf ) with gaussian kernel @xcite . as in the approach of allen et al .
, this mapping produces a shape based on a new set of measurements .
hasler et al .
@xcite apply the two previously reviewed approaches to a new representation of human body shapes that is posture invariant .
their method simultaneously models body pose and shape .
recently , baek and lee @xcite presented a technique that uses hierarchical clustering to build a statistical model of the training data .
the approach proceeds by clustering the training database and by performing a multi - cluster analysis of the training data .
to predict a body shape based on a set of input measurements , the approach finds the shape within the learned shape space that best describes the measurements using an optimization of shape parameters .
it is shown experimentally that accurate and visually pleasing body shapes are estimated when the input body sizes are inside the shape space spanned by the training data .
this approach is conceptually similar to the first optimization step of our algorithm .
hence , we expect that this approach does not allow to model shape variations that are outside of the shape space spanned by the training data .
anguelov et al .
@xcite aim to estimate a 3d human body shape based on a sparse set of marker positions .
this technique is useful when motion capture data is available .
anguelov et al.s scape model represents the human body as a triangular mesh with an associated skeleton model . as with the approach of allen et al .
, this approach reduces the dimensionality of the data using pca .
this yields a set of pca weights .
anguelov et al .
aim to compute a new triangular mesh based on a set of marker positions located on the body .
this is achieved by adjusting the pca weights to solve a non - linear optimization problem .
this method searches the solution in the learned pca space .
hence , it can not find local variations not present in the training database . finally , we review approaches that aim to estimate human body shapes based on a set of input images .
the following approaches proceed by learning a correlation between a training set of 3d face or body shapes and a set of derived 2d images and by using this learned correlation to predict a shape based on a new set of 2d images .
these approaches work well when the predicted shape is inside the shape space spanned by the training set .
however , they do not handle optimizations outside the learned shape space of 3d models .
blanz and vetter @xcite estimate a 3d face shape from a single input image in neutral expression .
they start by building a parameterized database of textured 3d faces and performing pca on the shape and texture data . given an input image , the learned pca space is searched to find the textured shape ( and parameters related to rendering the model ) that best explains the input image .
seo et al .
@xcite estimate a body shape from two images of a human in a fixed posture .
starting from a parameterized database of human meshes in similar poses , the approach performs pca of the 3d data .
given the two images , the learned pca space is searched to find a set of pca weights that corresponds to a 3d shape that matches the input images well .
chen and cipolla @xcite aim to estimate the human body shape in a fixed pose based on a given silhouette . starting from a parameterized database of human meshes in similar poses and a set of corresponding silhouettes , the approach performs pca of the 3d and 2d data separately .
the approach then computes a mapping from the pca space of the silhouette data to the pca space of the 3d data using a shared gaussian process latent variable model ( sgplvm ) @xcite .
given a new silhouette , the approach maps the silhouette into silhouette pca space and uses the sgplvm to map to the pca space of the 3d meshes .
ek et al .
@xcite use a similar approach to estimate the pose of a human body based on a given silhouette .
guan et al .
@xcite estimate both the shape and pose of a human body shape from a single photograph with a set of markers to be identified by the user .
the approach is based on the scape model .
when adjusting the pca weights , the shape is deformed to best match the image in terms of a shape - from - shading energy .
hasler et al .
@xcite estimate both the shape and pose of a human body from a photograph of a dressed person .
this approach requires a manual segmentation of the background and the human in the image .
this section outlines the proposed approach . as input to the method
, we are given a database of triangular manifold meshes @xmath0 of human bodies or faces with similar posture or expression and a set of measurements @xmath1 to be considered .
let @xmath2 in @xmath1 denote the measurements corresponding to @xmath3 .
furthermore , we are given a set of distances @xmath4 .
our aim is to estimate a shape @xmath5 that interpolates the distances @xmath4 . as previous methods
, the approach proceeds by learning a correlation between the shapes and the measurements .
when predicting a new shape , our approach finds an initial solution based on the learned correlation . unlike previous approaches , however , our approach refines this solution to fit the measurements using two steps of non - linear optimization .
first , we optimize the shape of the model with respect to the learned shape space .
that is , we aim to find the point in the learned shape space that best describes the measurements .
this gives a realistic human shape drawn from the distribution fitted to the training data .
hence , this shape can only contain local shape variations present in the training data . to account for other shape variations ,
we perform a second mesh - based optimization .
this optimization deforms the shape to fit the measurements as close as possible while satisfying a smoothness constraint without using prior knowledge of the body shape , and therefore predicts shapes with local variations not present in the training database .
when using a mesh - based deformation that optimizes the sought measurements , we can only expect to obtain a realistic human body shape if we start with a body shape that is close to the solution .
for this reason , we start from the shape in the learned shape space that best describes the sought measurements as opposed to starting directly from the point in shape space predicted by the learned correlation . the following sections describe the details of the approach .
anthropometric measurement represents a valuable source of human shape information , from which we can infer the 3d shapes for design applications . to learn a relationship between the anthropometric measurements and a set of 3d models , the measurements computed from the models should be equivalent to the measurements conducted by a human operator .
the measurements can be one of the three types : euclidean , geodesic , and circumference .
an euclidean or geodesic distance can be easily computed from two vertices on the model .
a circumference can be computed by intersecting ( part of ) the model with a plane , finding the ( possibly closed ) polygonal chain of the intersection that contains a specified vertex , and measuring the circumference of the convex hull of this chain .
this type of measurement is shown for the hip and waist circumferences in figure [ measurements_many_human ] . the reason we measure the length of the convex hull instead of the length of the chain itself is that many anthropometric measurements , such as the chest circumference , measure the length of a convex hull .
we specify the intersecting plane @xmath6 using the specified vertex @xmath7 and a normal direction @xmath8 .
furthermore , we specify a part of the model that is to be intersected with the plane @xmath6 .
this is necessary for some measurements , such as for the chest or hip circumference , where we wish to only intersect the torso ( and not the arms ) of the model with @xmath6 .
note that the intersection of @xmath6 with a subset of @xmath9 consists of a set of polygonal chains , where every vertex of the chain is either a vertex of @xmath9 or the intersection of an edge of @xmath9 with @xmath6 .
we now consider learning a correlation between the set of shapes @xmath3 and the corresponding measurements @xmath2 . to do this
, the database of the human shapes first needs to be parameterized , a process that computes point - to - point correspondences among the shapes .
this is in general a difficult problem @xcite . in practice ,
anthropometric markers identifying salient anatomic positions are often used to guide correspondence process .
such markers are provided in 3d anthropometric surveys , for example , the civilian american and european surface anthropometry resource ( caesar ) database @xcite . in this work ,
the known marker positions are used to deform a template shape to each subject of the database @xcite .
the euclidean and geodesic measurements are specified by their two endpoints and the type of distance to be considered and the circumference measurements are given by one plane ( specified by a vertex and a normal vector ) and the part of the body to be intersected with that plane .
since the database is parameterized , the vertices and body parts can be given in terms of their vertex and triangle numbers on the mesh .
at this point , a set of distances @xmath10 can be computed for @xmath0 .
we then learn a mapping between @xmath1 and the space of human body shapes using feature analysis . to this end
, we first perform pca on the meshes @xmath3 . in pca space ,
each shape @xmath3 is represented by a vector @xmath11 of pca weights .
pca yields a mean shape @xmath12 and a matrix @xmath13 that can be used to compute a new shape @xmath5 based on a new vector of pca weights @xmath14 as @xmath15 . recall that the aim is to create a new shape based on a new point @xmath4 in @xmath1 . to achieve this goal ,
feature analysis learns a linear mapping from @xmath2 to @xmath16 .
this yields a matrix @xmath17 that can be used to compute a new vector of pca weights @xmath14 based on a new point @xmath4 as @xmath18 , which generates a shape @xmath5 based on a new point @xmath4 as @xmath19 . to give all pca
coordinates the same weight , we normalize each entry of the pca weights by its corresponding pca eigenvalue before performing feature analysis .
the mapping between the measurements and the pca weights of the 3d shapes learned in the previous section allows us to find an initial shape @xmath20 given a new set of measurements @xmath4 .
however , straightforward application of feature analysis has the disadvantage that the shapes obtained for atypical @xmath4 are not contained in the space of human shapes ( e.g. see figure [ atypical_prediction ] ) .
since the subsequent refinement process depends on this initial shape , it is important to restrict it to be within the range of human shape . given the pca weights obtained by feature analysis @xmath18 , we restrict @xmath14 such that each dimension @xmath21 $ ] is at most @xmath22 times the standard deviation of the pca space along dimension @xmath23 .
this choice is based on the assumption that the shape space is modeled as independent gaussian distributions and most of the shapes are located within @xmath22 standard deviations of the mean for a suitable parameter @xmath22 .
hence , this step restricts the reconstructed shape to stay within the space of human shapes that was learned using pca . in our implementation
, we use the normalized weight vector to find an initial shape @xmath20 .
the shape generated from the pca weights can only serve as an initial estimation , because the shape space spanned by the pca is limited by the sample size and may not account for all of the human shape variation . to find a shape that respects the required measurements while staying in the learned shape space , we need to refine @xmath20 .
the following section presents our novel method for shape refinement in detail .
this section outlines a novel approach to refine the initial estimate @xmath20 to respect the required measurements while staying in the learned shape space .
we formulate the refinement problem as an energy minimization problem .
let @xmath24 denote the vertices of @xmath5 and let @xmath25 denote their position vectors .
furthermore , let @xmath26 be the vector of pca weights corresponding to @xmath5 , where @xmath27 is the pseudo - inverse of @xmath13 . for euclidean measurements , we are given the target length @xmath28 of the euclidean distance of the segment @xmath29 between two vertices @xmath30 and @xmath31 .
the shape @xmath5 that satisfies all given euclidean distance constraints minimizes @xmath32 where @xmath33 is the set of all euclidean distances that have a desired length , and @xmath30 and @xmath31 are the endpoints of @xmath29 . for geodesic measurements , we are given the target length @xmath34 of the geodesic path @xmath35 between two vertices @xmath30 and @xmath31 .
let @xmath36 denote the geodesic length of @xmath35 . in the following , we assume that the relative length of each edge @xmath37 of @xmath35 with respect to @xmath36 does not change during the deformation of the mesh . this is a reasonable assumption as we wish to preserve relative edge lengths during the deformation .
furthermore , this assumption translates into an easy optimization problem . with this assumption
, we can compute the target length @xmath38 of @xmath37 as @xmath39 , where @xmath40 is the current length of @xmath37 .
then , the shape @xmath5 that satisfies all given geodesic distance constraints minimizes @xmath41 where @xmath42 is the set of all geodesic paths that have a desired length , and @xmath43 and @xmath44 are the endpoints of @xmath37 . for circumference measurements , we are given the target length @xmath45 of the circumference @xmath46 of the convex hull of the polygonal chain passing though vertex @xmath30 and contained in the intersection between the mesh and the plane @xmath6 though @xmath30 with normal direction @xmath8 .
we compute the length @xmath47 of the circumference on the mesh as outlined in section [ sec_measurements ] .
let @xmath48 denote all of the points on the convex hull of the polygonal chain . with a similar reasoning as above , in the following ,
we assume that the relative length of each edge @xmath37 of the convex hull with respect to @xmath47 does not change during the deformation of the mesh .
hence , we can compute the target length @xmath38 of @xmath37 as @xmath49 , where @xmath40 is the current length of @xmath37 .
then , the shape @xmath5 that satisfies all given circumference constraints minimizes @xmath50 where @xmath51 is the set of all circumferences that have a desired length , and @xmath48 and @xmath52 are the endpoints of @xmath37 .
recall that @xmath48 and @xmath52 are not necessarily vertices of the mesh .
our aim is to deform @xmath20 , such that @xmath53 is minimized .
we solve this problem using two consecutive steps by minimizing @xmath54 with respect to @xmath14 followed by minimizing @xmath54 with respect to @xmath55 .
the first minimization finds the shape in the learned shape space that best describes the measurements .
this step finds a realistic human body shape whose measurements are close to the sought ones .
the second minimization then further deforms the shape using a mesh - based optimization without using prior knowledge on the shape .
this minimization allows the shape to deform locally in ways not present in the training data . during this minimization
, a smoothness term is used to keep a realistic body shape .
first , we discuss how to minimize @xmath54 with respect to @xmath14 .
this step finds the shape in the learned shape space that best describes the measurements .
we solve the optimization problem using a quasi - newton approach .
this approach has the advantage of having a near - quadratic convergence rate if the initial solution is close to the minimum .
since this approach requires analytic gradients , we next discuss the derivatives of the different energy terms .
the derivative of @xmath56 with respect to @xmath57 is @xmath58 where @xmath59 is the set of distances in @xmath60 with endpoint @xmath30 .
the derivative of @xmath56 with respect to @xmath14 is @xmath61 .
the derivatives of @xmath62 with respect to @xmath57 and @xmath14 are similar to @xmath63 and @xmath64 . to compute the derivative of @xmath65 with respect to @xmath57 , recall that each vertex @xmath48 of the convex hull of the polygonal chain is either a vertex of the mesh or the intersection of an edge of the mesh with @xmath6 .
hence , every @xmath48 can be expressed as a convex combination of at most two vertices of the mesh .
using this formulation allows us to compute the derivatives of @xmath65 with respect to @xmath57 and @xmath14 in a similar way to @xmath63 and @xmath64 .
note that the geodesic path between @xmath30 and @xmath31 and the circumference measurement specified by @xmath30 and @xmath8 may change when the mesh is deformed . to remedy this problem
, we update the geodesic paths and the circumference measurements and solve the resulting optimization problem @xmath66 times .
this finds the point @xmath14 in pca space corresponding to the shape that has measurements closest to the desired ones . as before , we normalize @xmath14 such that each dimension @xmath21 $ ] of @xmath14 is at most @xmath22 times the standard deviation of the pca space along dimension @xmath23 to ensure that the shape stays within the learned shape space . finally , @xmath14 is used to compute the shape @xmath67 as @xmath68 .
second , we minimize @xmath54 with respect to the vertex positions @xmath55 .
this term ensures that local variations not present in the training data are predicted correctly by locally deforming the shape to fit the required measurements .
however , simply minimizing @xmath54 may result in meshes that are not smooth .
hence , we encourage close - by parts of the mesh to deform similarly by considering the smoothness energy @xmath69 where @xmath70 is the translation vector by which @xmath57 is moved and @xmath71 is the one - ring neighborhood of @xmath30 .
the derivative of @xmath72 with respect to @xmath57 is @xmath73 note that the smoothness term helps ensure the predicted shape stays withing the shape space of human shapes .
the vertex positions are initialized to the vertex positions of @xmath67 .
we minimize the combined energy @xmath74 with respect to @xmath57 , where @xmath75 is the weight for the smoothing term .
as before , we update the geodesic paths and the circumference measurements and solve the resulting optimization problem @xmath66 times , yielding the final result @xmath5 .
the proposed method depends on the parameters @xmath76 , and on the termination criteria of the energy minimization . in our experiments ,
we stop the energy minimization if the relative change of the energy is smaller than @xmath77 times the machine accuracy , if the norm of the gradient is smaller than @xmath78 , or if 100 iterations have been performed .
the parameter @xmath22 determines the size of the acceptable shape space for reconstruction .
this parameter provides a way to trade off high measurement accuracy and likely human shapes .
if the measurements given to the method are known to be accurate measurements of a real human , @xmath22 can be set to a high value .
if the measurements given to the method are randomly generated , however , @xmath22 needs to be low to avoid non - human shapes .
hence , in our experiments with real data , we set @xmath79 and in our experiments with synthetic data , we set @xmath80 .
the parameter @xmath75 determines the relative weight of the smoothing term . in our experiments , we set @xmath81 .
the parameter @xmath66 determines how often the geodesic paths and circumferences are recomputed .
we need to set @xmath66 large enough to achieve the required accuracy , yet small enough to obtain an efficient approach . in our experiments
, we set @xmath82 unless specified otherwise .
we leave it for future work to find the optimal parameter settings automatically .
this is a challenging problem because the result depends on the parameter values , and because including the parameter values in the set of parameters to be optimized gives a highly nonlinear optimization problem .
this section demonstrates the effectiveness of our method using both synthetic experiments and experiments with real data .
we implemented the proposed algorithm using c++ .
the implementation uses dijkstra s
algorithm @xcite to compute geodesic distances , the quickhull algorithm @xcite to compute the convex hull to determine circumferences , and the limited - memory broyden - fletcher - goldfarb - shanno scheme @xcite to solve the optimization problems .
we choose these standard algorithms for their ease of use . to parameterize the database used for training , the efficient approach by xi et al .
@xcite is employed .
we compare our results to the results obtained using feature analysis and to the results obtained by using a sgplvm mapping from the space of the measurements to the pca space of the models .
both approaches are reviewed in section [ related ] .
we use the code by ek et al .
@xcite to compute the sgplvm .
the synthetic experiments aim to reconstruct human face and body shapes . for each synthetic experiment
, we learn a multivariate gaussian distribution @xmath83 from the available sets of measurements .
first , 200 sets of measurements are generated by randomly sampling 200 points from @xmath83 .
note that since these samples are drawn from the learned distribution , most of the measurements are similar to the ones in the training set .
we denote this set of measurements by @xmath84 .
second , 200 sets of measurements are generated by randomly sampling 200 points that are located on the ellipsoidal surface @xmath85 , where @xmath86 , @xmath87 is a constant , and @xmath88 is the vector that contains the diagonal elements of @xmath89 .
note that the larger the @xmath87 is , the farther the measurements are from the training set .
let @xmath90 denote this set of measurements . in our experiments , we set @xmath91 .
the experiments based on real data aim to reconstruct human body shapes from a set of measurements .
we consider two types of input data .
for the first type of input data , we digitally measure the distances in @xmath1 on a number of parameterized 3d human body scans and use these measurements as input to our algorithm .
we denote these measurements by @xmath92 .
note that in this case , the ground truth of the 3d body shape is known and can be used to evaluate the results . for the second type of input data
, we asked a number of volunteers to measure each other .
this results in a set of real - world measurements of real subjects .
we denote these measurements by @xmath93 . as in this case
, no ground truth 3d body shape is known , we validate the accuracy of the 3d reconstruction visually by comparing the 3d shape to silhouettes of photographs of the volunteers .
this experiment shows the accuracy we can expect when a person who is not an expert in anthropometry tries to use our method to build a digital clone of herself .
this section validates our approach using a set of synthetic experiments on 3d human face shapes . the training set consists of 50 faces from the caesar database @xcite .
after parameterization , each face contains 6957 triangles .
this experiment considers the seven measurements shown in figure [ measurements ] . here
, each dimension measures the geodesic distance between the pair of points shown in the same color .
note that several geodesic paths may overlap .
we generate new face shapes that aim to interpolate the measurements in @xmath84 and @xmath90 using our approach , feature analysis , and sgplvm .
table [ error_table ] shows the errors of the generated faces for each method .
we define an error for each dimension as the absolute value between the distance on the generated shape and the input measurement for this dimension , and use this to compute the error in mm as the average error over all dimensions .
note that feature analysis yields the lowest average errors and that sgplvm yields the highest average errors for all experiments .
@xmath94 + feature analysis & 0.67 & 1.63 & 6.76 + sgplvm & 5.35 & 5.28 & 10.88 + our approach & 0.68 & 2.00 & 7.27 + several of the shapes predicted using feature analysis are clearly outside of the shape space of human faces , as shown on the right of figure [ reconstructed_shapes ] .
this is undesirable because human perception is sensitive with respect to unrealistic variations in face shape .
in fact , most of the faces predicted using feature analysis in @xmath94 are clearly outside of the shape space of human faces .
it is desirable to have variability among the computed shapes .
both feature analysis and our approach predict significantly different shapes for all of the experiments .
sgplvm produces slightly different face shapes for different samples of @xmath84 and @xmath95 , but for all samples in @xmath94 , the same face shape ( visually similar to the mean shape ) is produced . in summary ,
our approach is the only one of the compared approaches that yields a variety of face shapes while always estimating shapes that are in the space of human faces .
furthermore , our approach yields lower errors than sgplvm .
this section validates our approach using a set of synthetic experiments and experiments based on real data .
the training set consists of 360 bodies of the caesar database .
after parameterization , each body contains 60000 triangles .
this experiment considers 34 measurements : 14 euclidean measurements , 4 circumference measurements ( hip , waist , chest , and head circumferences ) and 4 circumferences defined by four geodesic measurements each ( knee and arm circumferences ) .
figure [ measurements_many_human ] shows the measurements on one shape .
we generate new body shapes that aim to interpolate the new measurements in @xmath84 and @xmath90 using our approach , feature analysis , and sgplvm .
table [ error_table_body ] shows the errors of the generated bodies for each method .
the errors are computed in the same way as in the previous experiment .
note that our approach yields the lowest average errors for all experiments .
@xmath94 & @xmath92 + feature analysis & 3.29 & 14.26 & 36.86 & 3.09 + sgplvm & 4.91 & 12.19 & 22.88 & 4.55 + our approach & 1.74 & 6.39 & 15.49 & 1.10 + for most samples in @xmath96 and for several samples in @xmath97 , feature analysis predicts shapes that are outside of the shape space of human bodies .
for the samples in @xmath84 , feature analysis yields visually pleasing shapes .
sgplvm always yields body shapes that are in the shape space of human bodies .
however , for the samples in @xmath96 , the variation of the predicted body shapes is low .
some of the shapes predicted using @xmath96 with our approach contain small local artifacts due to the optimization with respect to the vertex coordinates .
note that setting @xmath75 to a higher value will reduce these artifacts at the cost of increased error . for @xmath97 and @xmath84
, our approach always predicts globally and locally realistic body shapes .
figure [ reconstructed_shapes_body ] shows four of the body shapes in @xmath84 predicted using our approach .
we can see that a large variety of shapes can be computed . for this experiment
, we consider 50 additional parameterized body shapes from the caesar database and use each of these models to digitally measure the 34 relevant measurements .
we call this set of measurements @xmath92 . for the body
shapes that aim to interpolate the measurements in @xmath92 , table [ table : error ] shows the errors of the euclidean measurements and circumferences of the generated bodies .
the error of the knee and arm circumferences is computed as the sum of the errors of the individual dimensions that contribute to the circumference .
we compute the average and maximum errors in mm for each dimension .
note that our approach yields the lowest average errors for all dimensions and the lowest maximum errors for most dimensions .
figure [ true_bodies ] shows the predictions for three different subjects .
each mesh is shown using a color coding corresponding to the signed distance from the parameterized body scan that was used to compute the measurements in @xmath92 .
note that sgplvm yields predictions with the highest errors and also predictions that are visually most dissimilar from the ground truth .
the predictions obtained using feature analysis and our approach are of similar quality .
note however that the predictions obtained using our approach are more accurate in localized areas of the body such as the waist area of the subject shown in the first row .
this is due to the deformation of the mesh that aims to obtain the correct waist circumference .
[ cols= " < , > , > , > , > , > , > " , ] finally , we demonstrate the ability of our approach to predict shapes that are not covered well by the training data . recall that this is relevant in practice because predicting human shapes based a small training sample avoids the complicated and expensive acquisition of large anthropometric databases .
we use as training data a set of 35 of the bodies of the caesar database .
all of the bodies in the training set are male bodies with typical body shape , see figure [ typical_male_training ] .
we choose five additional male bodies with body shape variations not covered by the training data and we use each of these models to digitally measure the 34 relevant measurements .
we then use these measurements for prediction . for this experiment , we set @xmath80 since the shapes are not represented well by the training data . furthermore , we set @xmath98 to increase the accuracy of the prediction .
figure [ atypical_prediction ] shows the results .
the first row shows the bodies of the caesar database used for prediction , the second row shows the result using our approach , the third row shows the result using sgplvm , and the last row shows the result using feature analysis .
note that our approach predicts body shapes that are similar to the ground truth for most shapes although a limited training set was used .
for the body in the last column , the result using our approach is not as large as the ground truth .
the reason is that the input body shape is far from the training data .
the results using sgplvm are all close to the mean shape and far from the ground truth .
the results using feature analysis are outside of the shape space of human body shapes .
this shows that our approach best handles the case where the predicted shapes have variations that are outside of the shape space spanned by the training data .
we conducted experiments based on human face and body shapes using both real and synthetic data .
we conducted an extensive evaluation by estimating 600 face shapes and over 650 body shapes . in summary
, we demonstrated experimentally that our method has the following properties . *
our method estimates realistic body or face shapes that are outside of the learned shape space . *
our method yields results that are more accurate than sgplvm , and that , unlike feature analysis , always stay within the shape space of human body or face shapes . * our method accurately predicts local shape variations not present in the training set ( which accounts for accurate predictions of the waist shapes in figure [ true_bodies ] ) . * our method yields visually pleasing results even when a small database is used for training .
this is important in practice because the acquisition of a large database for training is costly . * even for inaccurate input measurements acquired by non - experts ,
our method is able to find visually consistent results . *
our method is computationally efficient .
the running time of the optimization steps is linear in the number of vertices .
to predict one human body shape , our ( non optimized ) implementation takes about 35 seconds on a standard pc with a dual core cpu and 8 gb of ram .
we presented a novel approach to estimate human body and face shapes based on measurements .
we demonstrated that the approach can produce a large variety of likely shapes based on a relatively small training set and a small number of measurements .
we showed experimentally that , unlike feature analysis , the approach always produces shapes that are inside the space of human shapes .
furthermore , the approach yields higher accuracy than sgplvm .
that is , unlike previous approaches , our approach predicts shape variations not present in the training data while maintaining a realistic human body shape .
this allows to generate accurate body shape estimates based on simple modalities without the need to acquire a large - scale 3d database that represents the target population .
we thank the volunteers for participating in the experiment . furthermore , we thank pengcheng xi for helpful discussions and for providing us with the training data , neil lawrence for providing us with the sgplvm code , and the anonymous reviewers for helpful comments .
this work has partially been funded by the cluster of excellence _ multimodal computing and interaction _ within the excellence initiative of the german federal government .
|
the recent advances in 3-d imaging technologies give rise to databases of human shapes , from which statistical shape models can be built .
these statistical models represent prior knowledge of the human shape and enable us to solve shape reconstruction problems from partial information .
generating human shape from traditional anthropometric measurements is such a problem , since these 1-d measurements encode 3-d shape information .
combined with a statistical shape model , these easy - to - obtain measurements can be leveraged to create 3d human shapes . however , existing methods limit the creation of the shapes to the space spanned by the database and thus require a large amount of training data . in this paper , we introduce a technique that extrapolates the statistically inferred shape to fit the measurement data using non - linear optimization .
this method ensures that the generated shape is both human - like and satisfies the measurement conditions .
we demonstrate the effectiveness of the method and compare it to existing approaches through extensive experiments , using both synthetic data and real human measurements .
|
to investigate properties of equilibrium configurations of rotating neutron stars , so far , most neutron star models have been obtained by assuming neutron star matter to be a one - constituent perfect fluid ( for a review , see , e.g. stergioulas 2003 ) .
this treatment of equilibrium states of neutron stars seems to be quite reasonable as a first approximation for examining global properties of neutron stars such as the gravitational mass , the radius , or the maximum rotation frequency .
however , it has long been suggested that neutrons in the inner crust and neutrons and protons in the core of neutron stars are in superfluid states when the interior temperatures cool down below @xmath1 ( e.g. , shapiro & teukolsky 1983 ) . since the interior temperature of neutron stars
is believed to cool down quickly via the neutrino emission ( e.g. , baym & pethick 1979 ) , it is likely that each of many observable neutron stars , except newly born ones , has a core containing superfluids .
although superfluidity in the interior might be one of the important ingredients that affect the structures of neutron stars , the superfluidity has been neglected in most studies on equilibrium configurations of rotating neutron stars .
thus , it is necessary to examine how superfluidity affects fundamental properties of rotating neutron stars .
the superfluidity in neutron stars has been mainly argued in connection with the glitch phenomenon and the post - glitch relaxation observed in pulsars .
the glitch is a sudden decrease of an observed pulse period @xmath2 of a pulsar , whose largest magnitude is @xmath3 , that is followed by a period of continuous relaxation . in most explanations of glitch phenomena , a two - component model has been employed to describe rotational dynamics of neutron stars containing superfluid neutrons . in those treatments ,
one of the components is a mixture of the charged particles such as protons ( or ions ) and electrons , which are assumed to co - rotate because the charged components are strongly coupled each other due to short - range electromagnetic interactions ( easson 1979 , alpar et al .
1984 ) , and the other the superfluid neutrons , which are supposed to rotate separately from the charged components because the superfluid neutrons are inviscous and loosely couple to the charged components . in this paper
, we call the mixture of the charged components `` protons '' for brevity .
it has been shown that those two - component superfluid models have succeeded in explaining observed features of the glitches and the post - glitch relaxations qualitatively ( anderson & itoh 1975 ; alpar et al . 1984 ; sedrakian et al . 1995 , link & epstein 1996 ) .
those results mean that the existence of superfluid components inside neutron stars is supported not only from the theory of nuclear physics but also from observations of pulsars .
if they co - rotate , as mentioned in prix & rieutord ( 2002 ) , two constituents in a rotating superfluid star behave as one - constituent ordinary fluid , because the so - called entrainment effect , due to which the momentum of one superfluid constituent is dependent on the mass current of all superfluid constituents , does not operate and two constituents are in a chemical equilibrium state . on the other hand , specific characteristics of neutron stars due to the superfluidity are considered to be observed when two constituents rotate at different rotation rates . to study the effect of rotation velocity difference between two constituents on the global structures of rotating neutron stars with superfluidity , prix ( 1999 ) derived basic equations for the equilibrium configurations and obtained analytical solutions for slowly rotating superfluid stars within the framework of newtonian dynamics , generalizing the so - called chandrasekhar - milne expansion for ordinary fluid stars ( see , e.g. tassoul 1978 ) , in which rotational effects are treated as perturbations to non - rotating spherical stars . to describe dynamics of superfluids
, he made use of a variant of the so - called two - fluid formalism , which had been mainly developed by carter , langlois , and their co - workers ( carter 1989 ; carter & langlois 1998 ; langlois , sedrakian , & carter 1998 ; prix 2002 ) . in order to obtain analytical solutions , prix ( 1999 ) employed an equation of state whose functional form is @xmath4 , where @xmath5 and @xmath6 are the pressure and density , respectively .
more recently , prix , comer , & andersson ( 2002a ) extended the prix formalism so as to include the entrainment effect between two superfluids in order to investigate its effect on the stellar structures . in the same paper , prix et al .
( 2002a ) also argued the effect of the so - called symmetry energy , which can be included as a parameter in the equation of state , and found that the effect of symmetry energy is significant in determination of the density distributions of the neutrons and the protons .
as for relativistic rotating superfluid stars , andersson & comer ( 2001 ) calculated equilibrium configurations of slowly rotating superfluid neutron stars , extending the slow rotation approximation formalism devised by hartle ( 1967 ) ( see , also hartle & thorne 1968 ) for ordinary fluid stars to relativistic two - constituent superfluid stars .
very recently , prix , novak , & comer ( 2002b ) have obtained preliminary results for rapidly rotating superfluid stars , without the slow rotation approximation .
the purpose of this paper is to improve our understanding of properties of equilibrium configurations of rapidly rotating neutron stars with superfluids .
we calculate equilibrium configurations of rapidly rotating neutron stars with neutron and proton superfluids . in this paper
, we are concerned with rotating neutron stars of two superfluids whose angular velocities are different .
we however assume that the rotation velocity difference between two superfluids is very small in comparison with the angular velocity of the star
. this assumption could be reasonable for models of superfluid neutron stars because observations of pulsar glitches show that the rotation velocity of the protons may differ from that of the neutrons , but the amount of the rotation velocity difference is not very large .
we can therefore treat the effect of the rotation velocity difference between two superfluids on the structures of rapidly rotating neutron stars as a small perturbation to equilibrium configurations of rapidly rotating neutron stars in which two superfluids co - rotate .
for the dynamics of superfluids inside stars , we employ a newtonian version of the two - constituent formalism developed by prix ( 2002 ) . in order to obtain basic equations for determining the effect of the rotation velocity difference between two superfluids , equations of superfluid hydrostatic equilibrium and the poisson equation
are expanded in terms of the rotation velocity difference between two superfluids . to obtain the numerical solutions of equilibrium configurations , we make use of a variant of the so - called self consistent field method ( scf ) devised by ostriker & mark ( 1968 ) .
the equation of state for the superfluid we use is the analytical one employed by prix et al .
( 2002a ) , which is a natural extension of an @xmath7 polytropic equation of state for a barotropic ordinary fluid to the case of two superfluids . in
2 we present the basic equations employed in this paper for the rapidly rotating superfluid stars .
numerical results are given in 3 , and 4 is devoted for summary and discussion .
in this paper , a neutron star is assumed to be composed of superfluid neutrons and a mixture of superfluid protons and normal fluid electrons , because the interior temperatures of old neutron stars are much lower than the transition temperature to neutron and proton superfluids , which is considered to be @xmath8 k ( see , e.g. , epstein 1988 ) .
we further assume that this mixture of charged particles is perfectly in the state of charge neutrality because of strong coupling between the protons and the electrons by electromagnetic interactions .
the electrons therefore co - move with the protons , and their motions can be described with one fluid velocity @xmath9 . through this paper , we call the mixture of the the protons and the electrons `` protons '' for brevity as mentioned before . to describe the dynamics of the superfluids , we make use of the two - constituent formalism for newtonian superfluid dynamics , which has been recently developed by prix ( 2002 ) . we take account of no `` transfusion '' between the neutrons and the protons .
the particle numbers of the neutron and the proton must be therefore conserved separately .
the particle number conservation equations for the neutron and the proton are then given by @xmath10 where @xmath11 and @xmath12 are the number densities of the neutron and the proton , and @xmath13 and @xmath14 denote the fluid velocities of the neutron and the proton , respectively . here , @xmath15 means the covariant derivative in the three dimensional flat space .
the mass densities of two superfluids can be written as @xmath16 where @xmath17 is the mass of the neutron , and we assume the mass of the proton to be equal to that of the neutron . the fundamental quantity of the two - constituent formalism for the newtonian superfluid dynamics is the internal energy density @xmath18 , which is a function of @xmath19 , @xmath20 , and @xmath21 , where @xmath22 and @xmath23 .
note that we have neglected the entropy @xmath24 carried by the normal fluid of the electron for simplicity .
this treatment is justified for neutron stars whose internal temperatures are sufficiently low .
this function @xmath18 defines several basic thermodynamical quantities which describe the dynamics of two superfluids in terms of the total difference as follows @xmath25 where @xmath26 and @xmath27 denote the specific chemical potentials of the neutron and the proton , respectively .
the function @xmath28 represents the strength of the entrainment effect . by using a variant of the gibbs - duhem relation , the generalized pressure of two fluids
is defined by @xmath29 and we can obtain the relation , @xmath30 according to prix ( 2002 ) , the euler equations for two superfluids can be given by @xmath31 @xmath32 where @xmath33 and @xmath34 . here , @xmath35 means the lie derivative along the vector field @xmath36 .
note that we have considered no direct interaction force between two superfluids such as the mutual friction or the magnus - type force . in equations ( [ euler1 ] ) and ( [ euler2 ] ) ,
the function @xmath37 is the gravitational potential , which is determined with the total mass density @xmath6 , given by @xmath38 , through the poisson equation @xmath39 where @xmath40 is the gravitational constant .
considering the exterior differentiation of equations ( [ euler1 ] ) and ( [ euler2 ] ) , we can obtain the vorticity equations , given by @xmath41 where @xmath42 and @xmath43 are , respectively , the exterior derivatives of the one - forms @xmath44 and @xmath45 , which are defined by @xmath46 we consider stationary and axisymmetric equilibrium configurations .
thus , the time and azimuthal derivatives of physical quantities must vanish because a star is assumed to be in a stationary and axisymmetric state .
in this paper , we further assume that the neutrons and the protons , respectively , rotate with the angular velocities @xmath47 and @xmath48 around the axis of rotation , and that no meridional circulation is present in a superfluid star .
the fluid velocities for two fluids are then given by @xmath49 where @xmath50 is the rotational killing vector .
in this paper , we employ the spherical polar coordinates @xmath51 .
the components of @xmath50 can be written as @xmath52 in the spherical polar coordinates .
first of all , let us consider the vorticity conservation for the flow given by equation ( [ fluid_velocity ] ) .
substituting equation ( [ fluid_velocity ] ) into equation ( [ vorticity ] ) , we obtain @xmath53\wedge d\omega_n=0\,,\ ] ] @xmath54\wedge d\omega_p=0\,,\ ] ] where @xmath55 is the distance from the rotation axis , defined by @xmath56 .
thus , these vorticity conservation equations are automatically satisfied if the rotation laws for two superfluids are given as follows : @xmath57 where @xmath58 and @xmath59 are arbitrary functions of @xmath47 and @xmath48 , respectively .
in other words , the rotation laws of superfluid stars must be strongly restricted because the rotation velocities @xmath47 and @xmath48 may depend on the entrainment functions @xmath60 and @xmath61 , whose functional forms are in general determined with the matter distributions of stars .
when the effect of the entrainment between two superfluids does not operate , that is , the case of @xmath62 or of @xmath63 , the vorticity conservation equations become relatively simple as follows @xmath64 the functions @xmath47 and @xmath48 must not therefore depend on @xmath65 in this situation . in other words , we can choose the rotation velocities @xmath47 and @xmath48 freely as long as @xmath66 and @xmath67 are fulfilled .
note that this condition is the same as that of barotropic ordinary fluid stars . for simplicity , in this paper , we assume that both the neutrons and the protons are uniformly rotating , i.e. @xmath68 and @xmath69 , because vorticity conservations are automatically satisfied for uniformly rotating configurations . assuming the uniform rotations and integrating the euler equations ( [ euler1 ] ) and ( [ euler2 ] )
, we can obtain integrated equations for hydrostatic equilibrium states , given by @xmath70 where @xmath71 and @xmath72 are integral constants . here , we have used the relationships @xmath73 we are interested in configurations in which two superfluids are rotating rapidly with different angular velocities , @xmath74 .
we do not therefore make use of any slow - rotation approximation such as the chandrasekhar - milne expansion .
we however assume the difference between rotation velocities of two superfluids to be very small .
we can then treat the effect of the angular velocity difference @xmath75 on the stellar structure as a perturbation to rapidly rotating stars whose rotation rates of two superfluids are exactly the same . for superfluid neutron stars , although the angular velocity @xmath47 may differ from the angular velocity @xmath48 because the interaction between two superfluids may be very weak due to the superfluidity , it is believed that this difference between two angular velocities is small in comparison with the averaged angular velocity of the star . thus , this treatment could be appropriate for neutron star models with superfluidity .
we expand physical quantities appeared in equations ( [ poisson ] ) and ( [ beq1 ] ) up to the first order in @xmath76 as follows : @xmath77 @xmath78 @xmath79 @xmath80 where @xmath81 , @xmath82 , @xmath83 , and @xmath84 are given in terms of partial derivatives of the internal energy density @xmath18 , @xmath85 , and @xmath86 by @xmath87 @xmath88 here , quantities following `` @xmath89 '' denote perturbed ones .
note that since we consider the accuracy up to the first order in the rotation velocity difference , no entrainment effects appear in equations we solve . substituting those expanded quantities into equations ( [ poisson ] ) and ( [ beq1 ] )
, we obtain the zero - th order equations , given by @xmath90 @xmath91 @xmath92 equations ( [ eq0 - 1 ] ) and ( [ eq0 - 2 ] ) lead to a relationship between chemical potentials of the neutron and the proton @xmath93 in this paper , we take @xmath94 and @xmath95 to be @xmath96 , assuming that the two superfluids are in chemical equilibrium in the unperturbed state .
thus , our basic equations for obtaining unperturbed rapidly rotating stars can be written by @xmath97 on the other hand , the equations for the perturbed quantities are given by @xmath98 @xmath99 @xmath100 in order to solve the poisson equations ( [ eq0 - 3 ] ) and ( [ perturbation - eq3 ] ) , boundary conditions are required .
the appropriate boundary conditions at spatial infinity and at the center of the star are given by @xmath101 in order to obtain equilibrium configurations , the internal energy density @xmath18 must be specified . in this paper
, we employ the same analytical internal energy density as that used in prix et al .
( 2002a ) , which is given by @xmath102\ , , \label{def - ena}\end{aligned}\ ] ] where @xmath103 , @xmath0 , and @xmath104 are constants .
note that the entrainment that is included in the equation of state of prix et al .
( 2002a ) is neglected because the entrainment appears in second order in @xmath105 or @xmath106 .
this internal energy density is a natural generalization of an @xmath7 polytropic equation of state for a barotropic ordinary fluid into the two superfluids because it is written in the general quadratic form of @xmath19 and @xmath20 , given by @xmath107 where @xmath108 , @xmath109 , and @xmath110 are constants .
the chemical potentials for the internal energy density ( [ def - ena ] ) are then given by @xmath111\ , .
\label{def - mu}\end{aligned}\ ] ] on the other hand , @xmath19 and @xmath20 can be written in terms of @xmath26 and @xmath27 as @xmath112\,,\quad \rho_p={kx_p\over\sigma+1}(\sigma\tilde\mu_n+\tilde\mu_p)\ , , \end{aligned}\ ] ] which lead to @xmath113 let us consider the situation where the neutrons and the protons are in chemical equilibrium , which is given by the condition @xmath114 . from equation ( [ def - mu ] ) , we can obtain @xmath115 this equation means that @xmath116 inside the star and that @xmath103 represents the proton fraction when chemical equilibrium between the neutrons and the protons is achieved . on the other hand , the parameter @xmath0 of the internal energy density ( [ def - ena ] )
is interpreted as the so - called `` symmetry energy '' term ( prix et al .
2002a ; prakash , lattimer , & ainsworth 1988 ) . the appropriate range for @xmath0 might be @xmath117 and we consider only three values for @xmath0 , i.e. @xmath118 , @xmath119 , @xmath120 , in this paper ( prix et al .
2002a ) . for numerical calculations ,
it is convenient to introduce non - dimensional physical quantities as follows : @xmath121 where quantities with hat are non - dimensional ones , and @xmath122 and @xmath123 are defined by @xmath124 where @xmath125 is determined so as to be @xmath126 for the unperturbed star , and @xmath127 is the largest distance from the stellar surface to the center of the unperturbed star . here , @xmath122 and @xmath128 can be given freely because of the polytropic equation of state .
since , for unperturbed states , chemical equilibrium between two superfluids are assumed , the chemical potential @xmath129 can be written as @xmath130 by virtue of the analytical equation of state ( [ def - mu ] ) , where @xmath131 is the total mass density .
then the master equations for unperturbed stars are reduced to @xmath132 @xmath133 where @xmath134 and @xmath135 are the legendre polynomials . here , the gravitational potential @xmath136 has been written in the integral representation , in which a proper green function is employed to include physically appropriate boundary conditions both at @xmath137 and at spatial infinity ( [ boundary - condition ] ) . in this integral representation , equatorial
symmetry of the matter distribution has been assumed because we are concerned about stars having this equatorial symmetry .
note that these equations are exactly the same as those of polytropic ordinary fluid stars with polytrope index @xmath7 .
thus , solutions to equations ( [ unp1 ] ) and ( [ unp2 ] ) can be obtained by the so - called hachisu s self consistent field scheme ( hscf ) , once the axis ratio , @xmath138 , is given ( hachisu 1986 ) . here ,
@xmath138 is defined by @xmath139 where @xmath140 is the minimum distance from the stellar surface to the center of the star .
note that solutions @xmath131 , @xmath136 , and @xmath141 are universal functions in the sense that those are independent of the proton fraction @xmath103 . after getting a solution @xmath131 , and giving the proton fraction @xmath103
, we can obtain the mass densities and chemical potentials of two superfluids through the formulas @xmath142 the surface of the star @xmath143 for unperturbed stars is defined by @xmath144 where @xmath5 is the generalized pressure , which is given , for our internal energy density , by @xmath145\,\,.\end{aligned}\ ] ] pressure @xmath146 for the unperturbed state can be explicitly written as @xmath147 the non - dimensional stellar surface @xmath148 for an unperturbed star is then given by a solution of the algebraic equation @xmath149 . in this paper , we consider perturbed states whose densities of two superfluids @xmath19 and @xmath20 at the center of the star have the same values as those of unperturbed states . because of our equations of state ( [ def - mu ] ) , this requirement is equivalent to the assumptions of @xmath150 at the center of the star .
thus , the integral constants in equations ( [ perturbation - eq1 ] ) and ( [ perturbation - eq2 ] ) may be determined so as to be @xmath151 for the analytical internal energy density ( [ def - ena ] ) , perturbations of the total mass density @xmath152 is given in terms of perturbed chemical potentials @xmath153 and @xmath154 as @xmath155 from equations ( [ perturbation - eq1 ] ) and ( [ perturbation - eq2 ] ) , thus , we obtain @xmath156 where the relations of the integral constants ( [ perturbed_constants ] ) have been assumed . note that the perturbed total density is dependent on the proton fraction @xmath103 but not on the symmetry energy parameter @xmath0 . from equation ( [ peq-1 ] ) , it is found that perturbations @xmath152 , @xmath157 , and @xmath158 can be represented in terms of three functions independent of @xmath105 , @xmath106 , and @xmath103 as follows : @xmath159 where the three functions @xmath160 , @xmath161 , and @xmath162 are the solutions of the equations @xmath163 @xmath164 note that three quantities @xmath160 , @xmath161 , and @xmath162 are universal functions in the sense that those are independent of the parameters @xmath103 and @xmath0 , and that those are dependent only on the structure of the unperturbed star . with equations ( [ perturbation - eq1 ] ) and ( [ perturbation - eq2 ] ) ,
the chemical potentials of two superfluids are given by @xmath165 perturbations of the mass densities , on the other hand , can be written in terms of @xmath153 and @xmath154 as @xmath166\,,\quad \delta\hat\rho_p={\hat k x_p\over\sigma+1}(\sigma\delta\hat\mu_n+\delta\hat\mu_p)\,,\end{aligned}\ ] ] note that , for our internal energy density , the chemical potentials do not depend on the symmetry energy parameter @xmath0 , while the mass densities do .
the perturbations of the generalized pressure @xmath167 is given by @xmath168 if we write the stellar surface with accuracy up to the first order of the rotation velocity difference as @xmath169 the first order corrections for the stellar surface @xmath170 can be due to equation ( [ def - surface ] ) written in terms of @xmath153 and @xmath154 as @xmath171 where @xmath172 and @xmath172 are defined by @xmath173 because @xmath174 and @xmath175 are satisfied , two surfaces of @xmath176 and @xmath177 can be interpreted as the surfaces of the neutron and proton superfluids , respectively . although we can define the respective fluid surfaces of two superfluids as the zero - density surfaces ( prix 1999 ; prix et al .
2002a ) , we make use of the definition with the chemical potential for the surfaces because the definition with the chemical potentials is considered to be natural generalization of the surface definition for ordinary fluid stars , for which the surface is defined as a zero - pressure surface not as a zero - density surface .
note that as we can see from equation ( 35 ) , in general , `` zero - chemical potential surfaces '' do not coincide with `` zero - density surfaces '' .
in other words , the densities do not necessarily vanish on the zero - chemical potential surfaces because the equi - potential surfaces in general incline to the equi - density surfaces in the two - fluid model . in our numerical procedure of solving equations ( [ peq1 ] ) and ( [ peq2 ] ) , in order for the poisson equation ( [ peq2 ] ) to satisfy the boundary condition ( [ boundary - condition ] ) explicitly , equation ( [ peq2 ] )
is converted into the integral representation , given by @xmath178 in this paper , equations ( [ peq1 ] ) and ( [ p2 ] ) are numerically solved with a variant of the so - called self consistent field scheme ( ostriker & mark 1968 ) . to obtain solutions ,
we follow the following steps : i ) by assuming an initial guess for @xmath160 , compute the gravitational potential through the two - dimensional integration ( [ p2 ] ) .
ii ) determine the value of parameter @xmath162 from equation ( [ perturbed_constants ] ) .
iii ) by using the obtained @xmath161 and @xmath162 , solve equation ( [ peq1 ] ) for the perturbed density @xmath160 .
iv ) compare the obtained perturbed density with the one used in obtaining the perturbed potential . if the relative changes of @xmath160 are less than @xmath179 at all grid points , then the obtained perturbed density distribution is considered to be a converged solution .
if the condition for the relative changes is not satisfied , go back to step ii ) and there the obtained perturbed density is treated as a new initial guess for @xmath160 for the next iteration cycle . in actual computations , we make use of equidistantly spaced discrete meshes in the radial direction ( @xmath180 ) . in order to calculate integrations in @xmath181 direction ,
we employ a classical trapezoidal rule . as for the angular variable @xmath182
, we take the angular grid points at @xmath183 , where @xmath184 s are zeros of the legendre polynomial of order @xmath185 , i.e. @xmath186 , and there are @xmath187 grid points for @xmath188 .
we employ the gaussian quadrature ( see , e.g. , abramowitz & stegun 1964 ) to evaluate integrations in the @xmath182 direction . in the present investigation
, we take the number of mesh points to be @xmath189 @xmath190 .
the legendre polynomials in equations ( [ unp2 ] ) and ( [ p2 ] ) are added up to @xmath191 .
in this paper , we assume the proton fraction to be @xmath192 because a typical proton fraction in the cores of old neutron stars is expected to be around @xmath192 . for the parameter of the symmetry energy , @xmath0 , since we have little information about the range of @xmath0 , we investigate three cases , i.e. @xmath193 , in order to examine the effect of the symmetry energy on the stellar structures .
note that the total density of superfluids and the chemical potentials are dependent on the proton fraction but not on the symmetry energy .
we exhibit the results for two cases of @xmath194 and @xmath195 because solutions for other values of @xmath196 are expressed in terms of linear superpositions of those two solutions .
first , let us consider the unperturbed stars . in our unperturbed states ,
two superfluids are assumed to be in the same rotational motion and in chemical equilibrium . as mentioned in the last section ,
the structures of unperturbed stars are , therefore , almost the same as those of uniformly rotating @xmath7 polytropic stars .
thus , there are no new features for the unperturbed stars . for reader
s convenience , however , we exhibit some results for the unperturbed stars . in figures 1 through 4
, we show the non - dimensional fundamental quantities of the unperturbed stars , the axis ratios , @xmath138 , the total masses , @xmath197 , the total moments of inertia , @xmath198 , and the ratios of the rotational energy to the absolute value of the gravitational energy , @xmath199 , as functions of the angular velocity , @xmath200 , where @xmath200 is the non - dimensional angular velocity of the star , defined by @xmath201 , and @xmath202 denotes the stellar radius on the equatorial plane . here , the total mass , @xmath203 is defined by @xmath204 the total moments of inertia , @xmath205 , is given by @xmath206 the rotational energy , @xmath207 , and the gravitational energy , @xmath208 , are , respectively , defined by @xmath209 since @xmath210 for the unperturbed stars , we can write the masses , @xmath211 , @xmath212 , and the moments of inertia , @xmath213 , @xmath214 , for the neutrons and the protons as @xmath215 in order to check our numerical code , we have calculated a solution for a slowly rotating star with the axis ratio @xmath216 , whose angular velocity is @xmath217 , and compared them with the analytical solution obtained by prix et al .
( 2002a ) , in which the structure of a slowly rotating superfluid star was examined analytically with the slow rotation approximation .
the perturbed densities obtained from two methods are shown as functions of @xmath218 in figure 5 .
the solid lines show the analytical result for @xmath219 with the slow rotation approximation , and the solid squares our numerical result without assuming the slow rotation approximation .
the parameters of the model shown in figure 5 have been chosen to be @xmath118 and @xmath194 . as seen from this figure ,
our result is in good agreement with the analytical result of prix et al .
( 2002a ) as long as the angular velocity of the star is small enough . in figures 6 through 11 ,
typical density distributions for the neutrons and the protons , @xmath219 and @xmath220 are represented for three values of the symmetry energy parameter , @xmath193 . for the models shown in figures 6 through 11 , the axis ratio @xmath138 for
the unperturbed star has been taken to be @xmath221 , and the corresponding angular velocity is given by @xmath222 .
figures 6 through 8 show the density distributions for @xmath194 , and figures 9 through 11 for @xmath195 . in those figures ,
the perturbed densities are shown versus the radial coordinate @xmath218 for six different values of @xmath223 , in which the longest curve corresponds to the result on the equator , @xmath224 , and @xmath223 s decrease for each successively shorter curve toward the value @xmath225 on the symmetry axis . comparing figure 6 ( left panel ) with figure 5 , in which distributions of @xmath219 are shown for the same parameter as those in figure 6 except for the axis ratio @xmath138 or for the rotation velocity @xmath200 , we observe that the basic qualitative properties of the density perturbations for two superfluids do not depend on the value of @xmath200 so much .
we do not therefore show the density distributions for other rotation rates @xmath200 in this paper . from figures 6 through 11
, we can see that the amplitudes of @xmath219 are much larger than those of @xmath220 for the solutions of @xmath194 , while the amplitudes of @xmath219 is smaller than those of @xmath220 for the case of @xmath195 . as shown in figures 9 through 11 , however , for the solutions of @xmath195 ,
difference of the amplitude between @xmath219 and @xmath220 is not so large .
this is because the proton fraction is not very large in the interior .
similar behaviors were found by prix et al .
( 2002a ) .
it is also found from figures 6 through 11 that @xmath220 is strongly dependent on the values of @xmath0 for the @xmath194 case , while @xmath219 depends strongly on @xmath0 for the @xmath195 case .
this means that the significant parameters to model a superfluid neutron star are appropriate combinations of the symmetry energy parameter @xmath0 and the velocity difference between two superfluids . in figures 12 and 13 ,
the perturbed axis ratios are plotted as functions of the angular velocity @xmath200 . here ,
the functions , @xmath226 , @xmath227 , and @xmath228 can be interpreted as the perturbations of the axis ratios of the whole star , the neutron superfluid , and the proton superfluid , respectively , and are defined by @xmath229 the results for @xmath194 and @xmath195 are shown in figures 12 and 13 , respectively .
note that in these figures , the perturbed axis ratios are shown for @xmath230 because they diverge as @xmath200 goes to its maximum value .
these figures illustrate how the surface of the star is deform due to the effect of the rotation velocity difference between two superfluids .
it is found that the change of the axis ratio of the rotating component with the slower angular velocity becomes positive when the angular velocity @xmath200 becomes larger than some critical value . in other words ,
the rotating component with the slower angular velocity tends to become prolate when a star rotates very rapidly . in figure 14 ,
the perturbed mass is shown as a function of @xmath200 . here
, the perturbed total mass @xmath231 is defined as @xmath232 the solid curve and the dashed curve show @xmath231 for the solutions with @xmath194 and @xmath195 , respectively . in this paper
, we consider the perturbed states whose central density is the same as that of the unperturbed states .
thus , the mass of the perturbed stars can change . in figures 15 and 16
, we plot the perturbed moments of inertia for the neutron and the proton as functions of the angular velocity @xmath200 . here , the perturbed moments of inertia for the neutron and the proton , @xmath233 and @xmath234 , are defined as @xmath235 figures 15 and 16 show the perturbations of the moments of inertia for @xmath194 and @xmath195 , respectively . in these figures ,
results for three different values of @xmath0 , @xmath236 are displayed .
figures 15 and 16 show that the perturbed moments of inertia for the proton are strongly dependent on the symmetry energy parameter @xmath0 for @xmath194 case , while the perturbed moments of inertia for the neutrons strongly depend on @xmath0 for @xmath195 case .
this is consistent with that of the properties of the density distributions for the neutrons and the protons .
it is noted that amplitudes of @xmath233 and @xmath234 become large as the angular velocity @xmath200 is increased because the perturbations are roughly proportional to @xmath237 due to equation ( [ peq1 ] ) .
in this paper we have developed a formulation for constructing and examining rapidly rotating neutron stars that contain two superfluids , taking account of the effect of the rotation velocity difference between two superfluids .
we assumed neutron stars to be composed of the superfluid neutrons and the mixture of the superfluid protons and the normal fluid electrons . to describe newtonian dynamics of the two superfluids , the newtonian version of the two - fluid formalism developed by prix ( 2002 ) was employed . in this paper , we considered the situation where two superfluids rapidly rotate around the same rotation axis , but the rotation velocity difference between two superfluids is very small .
we then treated the effect of the rotation velocity difference on the equilibrium configurations as a small perturbation to rapidly rotating superfluid stars whose rotation velocities of two superfluids are the same .
we derived basic equations for perturbations on the structure of rapidly rotating superfluid stars due to the rotation velocity difference between two superfluids . assuming the superfluids to obey a simple analytical equation of state used by prix et al .
( 2002a ) , we obtained numerical solutions for the perturbed quantities and found that the density distributions of the superfluids are strongly dependent of the symmetry energy parameter @xmath0 , which appears in the analytical equation of state .
similar properties were found in prix et al .
( 2002a ) .
it was also found that if prix et al .
( 2002a ) s analytical equation of state is assumed , the perturbations can be represented in terms of the universal functions that are independent of the parameters of the equation of state .
although we only considered one special equation of state for superfluids in the present investigation , the formalism we derived in this paper is straightforwardly applicable to general equations of state .
we treated the superfluids inside neutron stars in the framework of newtonian dynamics .
the effect of general relativity must not be however neglected for the structures of neutron stars because neutron stars are quite compact in a sense that a general relativistic effect can be expressed by the factor @xmath238 whose typical value is @xmath239 for neutron stars , where @xmath240 is the speed of light .
it is straightforward to extend the present formulation to general relativistic configurations and we will do it in the future . recently ,
prix et al .
( 2002b ) have obtained rapidly rotating relativistic superfluid stars whose rotation velocity of the neutrons differs from that of the protons .
they did not assume that the rotation velocity difference between two fluids is very small .
in other words , their method can be applied to equilibrium configurations with any rotation velocities . yet
, our perturbation method developed in this investigation has an advantage in studying structures of real neutron stars because it would give results with higher accuracy by taking account of the smallness of the velocity difference directly , as long as the rotation velocity difference between two fluids in real neutron stars is very small .
a solid crust is believed to exist near the surface of cold and old neutron stars , and can have a significant influence on the structure of the equilibrium states if the solid crust is not in a strain - free state in the equilibrium states of rotating neutron stars .
therefore , an investigation of the effect of the solid crust on the equilibrium states remains as a challenging problem in the future too .
are given as functions of @xmath218 .
the solid curves and solid squares are used to indicate @xmath219 obtained from prix et al .
( 2002a ) s analytical formula and our numerical scheme , respectively .
the model parameters are @xmath217 , @xmath118 , and @xmath194.,width=302 ] for six different values of @xmath182 .
the model parameters are @xmath241 , @xmath118 , and @xmath194.,title="fig:",width=302 ] for six different values of @xmath182 .
the model parameters are @xmath241 , @xmath118 , and @xmath194.,title="fig:",width=302 ] for six different values of @xmath182 .
the model parameters are @xmath241 , @xmath118 , and @xmath195.,title="fig:",width=302 ] for six different values of @xmath182 .
the model parameters are @xmath241 , @xmath118 , and @xmath195.,title="fig:",width=302 ] , @xmath227 , and @xmath228 are given as functions of @xmath200 .
the solid , dashed , and dotted curves denote @xmath226 , @xmath227 , and @xmath228 , respectively .
the model parameter is @xmath194.,width=302 ]
|
we develop a formulation for constructing and examining rapidly rotating newtonian neutron star models that contain two superfluids , taking account of the effect of the rotation velocity difference between two superfluids .
we assume neutron stars to be composed of the superfluid neutrons and the mixture of the superfluid protons and the normal fluid electrons . to describe newtonian dynamics of the two superfluids , the newtonian version of the so - called two - fluid formalism
is employed .
the effect of the rotation velocity difference on the structure of equilibrium state is treated as a small perturbation to rapidly rotating superfluid stars whose angular velocities of two superfluids are assumed to be exactly the same .
we derive basic equations for the perturbed structures of rapidly rotating superfluid stars due to the rotation velocity difference between two superfluids . assuming the superfluids to obey a simple analytical equation of state proposed by prix , comer , and andersson
, we obtain numerical solutions for the perturbations and find that the density distributions of the superfluids are strongly dependent on the parameter @xmath0 which appears in the analytical equation of state and characterizes the so - called symmetry energy .
it is also found that if prix et al.s analytical equation of state is assumed , the perturbations can be represented in terms of the universal functions that are independent of the parameters of the equation of state .
stars : neutron stars : rotation hydrodynamics
|
it is evident that the internal structure of the progenitor of a core - collapse supernova ( sn ) is one of the dominant factors that determines the characteristics of the supernova explosion , such as the light - curve and the abundances produced in the supernova ( see e.g. @xcite ) .
it has also been shown that the distribution of the angular velocity can produce a strong asymmetry in the nucleosynthesis during the sn explosion and in its ejecta @xcite .
this makes it necessary to follow the detailed evolution of the abundances and the rotation profile at all stages of the evolution of a massive star before it explodes as a core - collapse sn .
observationally , it is well established that @xmath0% of all massive stars are members of binary systems with orbital periods shorter than 1 year and that at least 25% of these will start to interact by roche - lobe overflow ( rlof ) during the advanced stages of the primary s evolution @xcite .
this implies that a significant fraction of all core - collapse supernova progenitors will have been affected by a previous binary interaction , where one of the most important interactions is the spiral - in of the two binary components inside a common envelope ( ce ) @xcite .
the final result of the spiral - in depends on how much of the released orbital energy has been converted into driving the expansion of the envelope relative to its binding energy . here
we study the situation where the deposited energy is not sufficient to eject the common envelope .
this leads to the complete merger of the secondary with the core of the primary , forming a rapidly rotating single star in the process .
since the timescale of the merger is much longer than the dynamical timescale of the ce , it can not be treated with a purely hydrodynamical code .
mass transfer in the merging phase changes not only the chemical composition profile and the angular velocity distribution , but can also cause the erosion of part of the core , changing the core / envelope ratio .
the evolutionary path of the merger product may also differ from that of a normal single star , e.g. by making a late blue loop more probable .
since the resulting supernovae may also differ from those with single - star progenitors , these merged objects are likely to be responsible for some of the variety observed among type ii supernovae . in this paper
we present the results of the modelling of the complete slow merger of a massive binary within a common envelope . in section 2
we briefly describe the assumptions that were used to model mergers , and in section 3 we present some of the main results .
we used a standard henyey - type stellar evolution code @xcite , updated recently @xcite .
the nuclear reactions rates were taken from thielemann s library reaclib @xcite and updated as in @xcite . in the code
opal opacities @xcite are used , supplemented with contributions from atomic , molecular and grain absorption in the low temperature regime @xcite . to model the merger we implemented a number of modifications to the single stellar evolution code .
these modifications were made mainly to treat the presence of the secondary inside the primary s envelope , including the mass transfer from the secondary to the core and the associated nucleosynthesis and mixing .
a more detailed description of the modifications in the code can be found in @xcite .
we determine how deep the hydrogen - rich stream penetrates into the core of the primary using the prescription developed in @xcite .
we considered binaries consisting of a @xmath1 primary and a @xmath2 secondary . at the start of the spiral - in ,
the primary had already completed core helium burning .
the chemical composition was taken as typical for young stars in the lmc ( @xmath3 and @xmath4 ) .
we adopted the schwarzschild criterion for convection and took a mixing - length parameter @xmath5 and a convective - overshooting parameter equal to 25% of a pressure scale height .
these parameters were chosen since they are most appropriate for merger models of the progenitor of sn 1987a ( see @xcite ) .
the qualitative behaviour of the merger and the temporal evolution of the structure of the primary within the secondary s orbit depend on the interaction of the hydrogen - rich material with the surrounding ambient matter .
in particular , it depends on how deep and how fast the hydrogen - rich material penetrates into the primary s core and its placement with respect to the hot and/or convective zones within the secondary s orbit . according to our prescription for determining the stream penetration depth , the most effective penetration should occur near the end of the merger phase , when the mass - loss rate is high and the exposed material from the secondary has low entropy
. however , at this time the structure of the primary has changed , and a dense hydrogen - enriched region may already have been built up around the helium core , preventing the stream from penetrating deeper .
all these effects combined create a very non - linear picture of the primary s response to mass transfer . in general , as a result of the merger , the primary s core expands , and the central temperature of the core drops ( the degree of core cooling depends on the core expansion , i.e.the merger efficiency ) .
this increases the total evolutionary time before core carbon ignition .
based on our systematic study , we can distinguish three qualitatively different types of merger , where the classification of the mergers can be well explained by considering the temporal evolution of the convective zones during the merger .
+ * @xmath6 the quiet merger . * all of the he affected by the penetrating stream is mixed with the outer envelope during the merger by convection , but since most of the he shell is not disturbed , there is only a moderate change in the surface abundances ( he at the surface increases only by @xmath7 ) . the merger product remains a red supergiant , although the progenitor might be significantly spun up compared to a single supergiant evolved in isolation .
+ this type of merger may happen in systems where the primary is close to carbon ignition at the start of the merger .
this implies high pressure and temperature gradients and a correspondingly high entropy dissipation coefficient @xcite .
in addition , the secondary has to have a mass larger than @xmath8 ( due to the larger entropy ) . + * @xmath6 the moderate merger . * here
the he core expands significantly , but an extensive he - rich shell remains . during the penetration ,
the stream creates a hydrogen - rich zone around the core .
this zone becomes convective at some point and suppresses the bottom helium convective zone ( see fig . 1 ) . during the merger ,
the primary core expands more drastically than in the case of the quiet merger .
the merger product appears as a red supergiant , rotating rapidly and contracting immediately after the merger .
significant rotationally enhanced mass loss in the equatorial direction during the contraction phase is expected .
depending on the core / envelope ratio , the merger product may then perform a blue loop ( if the hydrogen shell source becomes temporarily dominant ) or continues its evolution as a red supergiant . in the first case
, some dredge - up of helium takes place during the merger , and it can be expected that rotationally induced mixing will cause further significant enhancement of the surface helium abundance . in the second case ,
a delayed dredge - up phase ( a few thousand years after the merger , see fig . 1 ) )
takes place , resulting in a large overabundance of he in the envelope ( @xmath9 ) .
+ at the start of the merger , the primary can be either already close to core carbon ignition and has a companion of @xmath10 , or the primary is at the start of he shell burning and has a companion of @xmath11 or larger .
a moderate merger is most appropriate for the progenitor of sn 1987a . + * @xmath6 the explosive merger .
* during the stream - core interaction , the hydrogen - rich material accumulates between two initially convective zones and slowly creates an intermediate zone , which at some point connects to the he burning shell .
this leads to a dramatic nuclear flash , resulting in a drastic expansion of the he shell and complete mixing . the duration of this he shell explosion is about 0.25 year . in the case of a strong
he shell explosion , even the carbon core may be disturbed . in all cases
, there is an immediate significant increase of helium in the envelope , often accompanied by an increase in the carbon abundance .
the stripped - off naked carbon core connects with the hydrogen - rich convective envelope and provides the site for efficient s - processing .
the merger product continues its evolution as a red supergiant .
+ this type of the merger can take place in binaries with low - mass secondaries , not very close to core carbon ignition .
the anomalous carbon star v hydrae may provide an example for this merger channel .
|
we study the complete merger of two massive stars inside a common envelope and the subsequent evolution of the merger product , a rapidly rotating massive supergiant .
three qualitatively different types of mergers have been identified and investigated in detail , and the post - merger evolution has been followed to the immediate presupernova stage .
the `` quiet merger '' case does not lead to significant changes in composition , and the star remains a red supergiant . in the case of a `` moderate merger '' , the star may become a blue supergiant and end its evolution as a blue supergiant , depending on the core to total mass ratio ( as may be appropriate for the progenitor of sn 1987a ) . in the case of the most effective `` explosive merger '' , the merger product stays a red giant .
in last two cases , the he abundance in the envelope is increased drastically , but significant s - processing is mainly expected in the
|
energy type functionals @xmath0 = \iint f\left ( \left| x - y\right| \right ) \xdif { \omega ( x)}\xdif { \omega ( y)}$ ] , where @xmath1 is a measure of compact support , appears in statistical mechanics of systems , topological classification problems of knots , isoperimetric problems , harmonic analysis , discrete energy problems and other areas of pure and applied mathematics ( see , for example , @xcite ) . motivated by problems of shape optimal design @xcite we consider the model domain functional @xmath2 = \int\limits_\omega \int\limits_\omega f\left ( \left| x - y\right| \right )
\xdif x\xdif y,\ ] ] defined on the set of bounded domains @xmath3 with smooth boundaries . here
@xmath4 is known function .
let @xmath5 equation @xmath6 will produce a flow @xmath7 which moves a bounded smooth domain @xmath8 to its new position @xmath9 ( see fig .
[ fig1 ] on p. ) . for all @xmath10 close enough to zero the domain
@xmath11 has been shown to be bounded and smooth @xcite .
shape derivative in the direction of vector field @xmath12 is defined by formula @xmath13 : = \frac d{dt } |_{t=0}\it{f}\left [ \omega _ t^v\right ] .\ ] ] [ critic ] we say nonempty domain @xmath14 is _ critical _ if @xmath15=0\ ] ] for any @xmath16 . if @xmath14 is critical , then ( see section [ shpdrv ] ) equality @xmath17 holds for all @xmath18 with @xmath19 . as a special case equation ( [ eq0 ] )
includes the inverse potential problem @xmath20 this last equation defines the body which has constant gravity potential on its own shape .
the observation based on the `` moving planes '' method was done in @xcite : [ monot ] let @xmath21 function @xmath22 is strictly monotone .
if nonempty @xmath21 domain @xmath14 solves equation ( [ eq0 ] ) , then it is a ball .
thus , if @xmath21 function @xmath22 is strictly monotone , then any critical point of the domain functional ( [ funk ] ) is the ball of fixed radius . if function @xmath23 is not strictly monotone , then critical domains of different shapes are possible .
usually _ stable _ minimum point of general domain functional @xmath24 is defined as the domain @xmath25 at which inequality @xmath26<\it{f}\left [ \omega\right]$ ] holds for all @xmath14 close enough to @xmath25 ( see , for example , @xcite ) . even despite of the difficulty in proper understanding of the closeness , domain functional ( [ funk ] )
has no stable minimum ( and maximum ) points in this sense because it is invariant under the rigid motions . to reach the uniformity with definition of critical point
we shall introduce the notion of _ weakly stable _ minimum point for the domain functional @xmath24 .
[ defstb ] we say the domain @xmath27 is stable minimum point of the domain functional ( [ funk ] ) in the direction of vector field @xmath28 , if either for all @xmath10 , @xmath29 and close enough to zero , inequality @xmath26<\it{f}\left [ \omega _ t^v\right]$ ] holds , either for all @xmath10 close enough to zero , @xmath25 as the rigid body coincides with @xmath11 .
the domain @xmath27 is said to be _ weakly stable _ minimum point of the domain functional ( [ funk ] ) in @xmath30 , if this domain is stable minimum point of @xmath24 in the direction of vector field @xmath28 for any vector field @xmath28 .
[ rmrk1 ] * a ) * inequality @xmath31 > 0,\ ] ] here @xmath32 denotes the second shape derivative of the domain functional @xmath24 , is sufficient for the critical point @xmath27 to be stable minimum in the direction of vector field @xmath12 .
* b ) * if @xmath33 , then any @xmath25 as the rigid body coincides with @xmath34 for all @xmath35 .
* c ) * if normal part of @xmath12 on the boundary @xmath36 is zero , then @xmath37 for all @xmath10 close to zero .
statements * b ) * and * c ) * are easy to understand from geometrical point of view .
they follow from the standard properties of ordinary differential equations .
if @xmath21 function @xmath22 is strictly monotone , then by proposition [ monot ] the domain functional @xmath38 has at most one critical point to be considered as a rigid body which , of course , is weakly stable in @xmath30 . in this paper we consider domain functional ( [ funk ] ) when function @xmath39 is positive definite , that is @xmath40 for all complex @xmath41 and thus inequality @xmath42 \geq 0\ ] ] holds for all @xmath43 bounded .
[ stbl ] let @xmath24 is the domain functional ( [ funk ] ) generated by positive definite function @xmath39 and @xmath44=0\ ] ] for some @xmath14 the ball .
then @xmath14 is weakly stable minimum point of @xmath24 in @xmath30 .
investigation of the domain functional ( [ funk ] ) with the positive definite function @xmath39 is stimulated because of its relation to the pompeiu problem @xcite .
we will prove [ pompeiu ] let @xmath14 be bounded nonempty domain .
* a ) * if @xmath14 satisfies ( [ eq1 ] ) for some positive definite function @xmath39 , then there is @xmath45 : @xmath46 here @xmath47 denotes the fourier transform of indicator function @xmath48 . *
b ) * on the other hand , if condition ( [ eq2 ] ) holds , then @xmath49 = 0 $ ] for some positive definite function @xmath39
. domain @xmath50 which satisfies ( [ eq2 ] ) is known as domain without the _ pompeiu property _ @xcite .
if its boundary @xmath51 is smooth and connected , then overdetermined problem [ schif ] @xmath52 has nontrivial solution @xcite .
it is expected ( schiffer conjecture @xcite ) such the domain to be necessarily a ball . in @xcite
the schiffer conjecture is affirmed for `` small '' perturbations of the ball .
kobayashi @xcite use precise fourier transform estimates of domain indicator function .
agranovsky and semenov @xcite , instead , exploit the overdetermined problem ( [ schif ] ) .
their approach leads to the generalization of results in riemannian spaces .
more on the pompeiu and schiffer problems see the expository paper @xcite .
the proof of the theorem [ stbl ] may be derived from @xcite . instead of
, we consider the pompeiu problem as shape optimization problem and shall give the direct proof of the theorem [ stbl ] based on the shape derivative method .
the paper is organized as follows . in section [ shpdrv ]
shape derivatives of the domain functional ( [ funk ] ) are calculated . the domain equation ( [ eq0 ] ) which describes critical domains
is derived and an expression of the second shape derivative of the domain functional ( [ funk ] ) in a case of positive definite function @xmath39 is examined .
the proof of the theorem [ stbl ] is preceded by [ lem1 ] let @xmath24 is the domain functional ( [ funk ] ) generated by positive definite function @xmath39 and @xmath53=0 $ ] for @xmath14 the ball centered at @xmath54
. then @xmath55 = 0 $ ] for vector field @xmath12 if and only if the corresponding function @xmath56 on the boundary of @xmath14 coincides with first order spherical harmonic .
this lemma is proved in section [ harmonic ] .
based on the lemma [ lem1 ] the proof of theorem [ stbl ] is done in section [ thrm ] and consists of two steps .
first , we consider special choice of vector fields @xmath12 when all domains @xmath34 has the same center of mass @xmath54 ( lemma [ lem2 ] ) .
second , the case when center of mass moves , is reduced to the previous step by suitable choice of non autonomous vector field .
the proof of the proposition [ pompeiu ] about the pompeiu problem equivalence to the shape optimization problem is done in section [ secpmp ] . in section [ sec5 ]
the remark on generalization of theorem [ stbl ] to other domain functionals is done .
then a few numerical examples of shape evolution to the solution of the pompeiu problem are presented .
let @xmath57 be the smooth function .
we consider the domain functional @xmath58 = \int\limits_\omega \int\limits_\omega f\left ( \left| x - y\right| \right ) \xdif x\xdif y.\ ] ] then , according to notions on page ( see also fig . [ fig1 ] ) , @xmath59 = \int\limits_{\omega _ 0}\int\limits_{\omega _ 0}f\left ( \left| t_t^v\left ( x\right ) -t_t^v\left ( y\right ) \right| \right ) \det dt_t^v\left ( x\right ) \det dt_t^v\left ( y\right ) \xdif x\xdif y,\ ] ] here @xmath60 denotes the jacobian of the mapping @xmath7 , @xmath10 considered as parameter
. using equalities @xmath61 ( see @xcite on calculus technique ) and verifying the equality @xmath62 one gets @xmath63 = \int\limits_{\omega _ 0}\int\limits_{\omega _ 0}\{\left [ \left\langle \nabla _ xf\left ( \left| x - y\right| \right ) , v\left ( x\right ) \right\rangle + \left\langle\nabla _ yf\left ( \left| x - y\right| \right ) , v\left ( y\right ) \right\rangle \right ]
\\ + f\left ( \left| x - y\right| \right ) \left[{\mbox{\rm div}\ } v\left ( x\right ) + { \mbox{\rm div}\ } v\left ( y\right ) \right ] \}\xdif
x\xdif y=\\ 2\int\limits_{\omega _ 0}\int\limits_{\omega
_ 0}{\mbox{\rm div}\ } _ x\left [ f\left ( \left| x - y\right| \right ) v\left ( x\right ) \right ] \xdif x\xdif y= 2\int\limits_{\partial \omega _ 0}\left [ \int\limits_{\omega _ 0}f\left ( \left| x - y\right| \right ) \xdif y\right ] \left\langle v,\eta \right\rangle \left ( x\right ) \xdif s_x , \end{split}\ ] ] here @xmath64 is exterior unit normal vector field on @xmath65 if @xmath66 gives an extremum to @xmath24 , then necessarily @xmath67 for all @xmath68 replacement of the initial moment @xmath69 by the arbitrary moment @xmath10 leads to equality @xmath70 = 2\int\limits_{\partial \omega _
t^v}\left [ \int\limits_{\omega _ t^v}f\left ( \left| x - y\right| \right ) \xdif y\right ] \left\langle v,\eta _ t^v\right\rangle \left ( x\right ) \xdif s_x,\ ] ] here @xmath71 denotes exterior unit normal vector field on @xmath72 now let us calculate the second derivative @xmath73 : = \frac{d^2}{dt^2}|_{t=0}\it{f}\left [ \omega _ t^v\right]\ ] ] in case when @xmath66 is critical domain to functional @xmath24 . applying formula ( 5.9 ) in page 1105 of @xcite and because of ( [ eq6 ] ) one gets here @xmath74 denotes mean curvature of @xmath75 at the point @xmath76 .
if function @xmath77 is positive definite and @xmath26 = 0 $ ] , then indicator function @xmath78 minimizes functional @xmath79 = \int \int f\left ( \left| x - y\right| \right ) \varphi \left ( x\right ) \varphi ( y)\xdif x\xdif y\ ] ] defined on the set of compactly supported functions .
then necessarily @xmath80 for all @xmath81 as a consequence , the second term in the last part of ( [ eq8 ] ) is zero .
so if @xmath26 = 0 $ ] , then @xmath82 = 2\int\limits_{\partial \omega _ 0}\int\limits_{\partial \omega _ 0}f\left ( \left| x - y\right| \right ) v\left ( x\right ) v(y)\xdif s_x\xdif s_y,\ ] ] for all @xmath83 here @xmath84 denotes the normal component of vector field @xmath12 on @xmath65
the proof is based on standard properties of fourier transform and of bessel functions . for them
we refer to @xcite and @xcite .
let @xmath85 be a ball of radius @xmath86 centered at @xmath54 and @xmath87=0 $ ] for the domain functional ( [ funk ] ) with positive definite function @xmath88 .
then @xmath89 here @xmath90 denotes the surface measure of the unit sphere .
because function @xmath91 is entire and @xmath92 is positive borel measure of polynomial growth @xcite , equality in ( [ eq9 ] ) is possible only if measure @xmath92 is discrete and @xmath93 now , let us consider equality @xmath94 = 0 $ ] .
we have @xmath95 = \int\limits_{\partial b}\int\limits_{\partial b}f\left ( \left| x - y\right| \right ) v\left ( x\right ) v\left ( y\right ) \xdif s_x\xdif s_y=\\ \int\limits_0^\infty \int\limits_{\left| \xi
\right| = r}\left| \left ( \chi _ { \partial b}v\right ) ^{\wedge } \right| ^2\left ( \xi \right ) \xdif
\mu \left ( r\right ) .
\end{split}\ ] ] function @xmath96 enables decomposition @xmath97 for spherical harmonic @xmath98 of order @xmath99 @xcite .
then @xmath95 = \int\limits_0^\infty r^{n-1}\int\limits_{\left| \xi \right| = 1}\sum\limits_{k=0}^\infty \left| c_k\right| ^2\left| \left ( \chi _ { \partial b}y_k\right ) ^{\wedge } \right| ^2\left ( \xi r\right ) \xdif s_\xi \xdif
\mu \left ( r\right ) = \\
= \int\limits_0^\infty r^{n-1}\int\limits_{\left| \xi \right| = 1}\sum\limits_{k=0}^\infty \left| a_k\right| ^2\frac{j_{\frac{n+2k-2}2}^2\left ( rr\right ) } { r^{n-2}}y_k^2\left ( rr\right ) \xdif s_\xi \xdif
\mu \left ( r\right ) , \end{split}\ ] ] since @xmath100 and @xmath101 @xcite .
bessel functions of different order has nonintersecting sets of zeroes @xcite , so due to relation ( [ eq10 ] ) equality @xmath102 is valid if and only if @xmath103 where @xmath104 is the first order spherical harmonic .
[ rmrk2 ] function @xmath104 has the form @xmath105
i.e. on the surface of the ball @xmath85 it coincides with a normal component of a constant vector field .
in this section we shall prove theorem [ stbl ] . because of lemma [ lem1 ] , remarks [ rmrk1 ] and [ rmrk2 ] it is sufficient to verify stability of the ball @xmath106 centered at @xmath54 in direction of vector field @xmath12 when function @xmath107 on the boundary @xmath51 is the first order spherical harmonic .
[ lem2 ] suppose that point @xmath54 is mass center of the domain @xmath34 for all @xmath10 close enough to zero and function @xmath107 on the boundary of the ball @xmath25 is the first order spherical harmonic .
then @xmath108 on @xmath36 . for all @xmath10 close enough to zero and for any @xmath109 equalities @xmath110 and @xmath111 holds .
consequently , @xmath108 on the boundary @xmath36 . in general case ,
mass center @xmath112 of the domain @xmath11 is not zero .
we consider the flow @xmath113 .
this flow generates the family of domains @xmath114 which coincide as rigid bodies with @xmath11 and are centered at @xmath54 .
of course , the new vector field @xmath115 is non autonomous .
fortunately , shape derivative method is applied also in the case of non autonomous vector fields @xmath116 , @xmath117 , @xmath118 @xcite .
moreover , withdrawal of formula ( [ eq12 ] ) in section [ shpdrv ] remains the same in the case of @xmath119 .
in this section we shall prove proposition [ pompeiu ] .
* proof of a)*let equation ( [ eq1 ] ) has bounded domain @xmath14 as its own solution for some @xmath120 . by bochner - schwartz theorem @xcite and because of spherical symmetry of f@xmath121 there exists @xmath122 positive borel measure on @xmath123 of polynomial growth : @xmath124 = \int\limits_{\omega}\int\limits_{\omega}f\left ( \left| x - y\right| \right ) \xdif
x\xdif y= \int\limits_{\xr^n}\left| \hat \chi _ { \omega}\right| ^2\left ( \xi \right ) \xdif \mu \left ( \left| \xi \right| \right ) = \\
\int\limits_0^\infty \int\limits_{\left| \xi \right| = r}\left| \hat \chi _ { \omega}\right| ^2\left ( \xi \right ) \xdif s_\xi \xdif
\mu \left ( r\right ) , \end{split}\ ] ] where @xmath125 then equality @xmath126 = 0\ ] ] implies the existence of @xmath127 @xmath128 consequently , @xmath129 for all @xmath130 because of @xmath131 one has @xmath132 * proof of b)*suppose that @xmath14 has no the pompeiu property .
then ( @xcite , @xcite ) there exists @xmath45 equality ( [ eq4 ] ) is true and , therefore , @xmath133 \xdif s_\xi = \\ \int\limits_{\omega}\int\limits_{\omega}\left [ \int\limits _ { \left|\xi \right| = \lambda } e^{i\left\langle \xi , x - y\right\rangle } \xdif s_\xi \right ] \xdif x\xdif y= \left ( 2\pi \lambda \right ) ^{\frac n2}\int\limits_{\omega}\int\limits_{\omega}\frac{j_{\frac{n-2}2}\left ( \lambda \left| x - y\right| \right ) } { \left| x - y\right| ^{\frac{n-2}2}}\xdif x\xdif y , \end{split}\ ] ] here @xmath134 denotes the bessel function .
this will end the proof of proposition [ pompeiu ] .
let @xmath24 be a domain functional defined on the set of bounded domains @xmath50 and invariant under the rigid motions of @xmath135 .
suppose that for the ball @xmath85 centered at @xmath54 the condition * 2 . * concerning the schiffer problem .
the result of theorem [ stbl ] does not cover ones stated in @xcite .
this is because spectral parameter @xmath138 in ( [ schif ] ) may vary with the domain @xmath14 .
though this more general situation may be considered on the basis of lemma [ lem1 ] the proof requires additional techniques .
* methods based on the shape derivative allow numerical experiments in the pompeiu problem . in fig .
[ fig2 ] evolution of the long thin ellipse and of the quadrate to the `` figure '' without the pompeiu property in the shape antigradient direction is presented .
|
the pompeiu problem is considered as shape optimization problem . we show stability of the ball which is the minimum point of related domain functional . the proof is based on shape derivative method .
stability of the ball for general domain functionals invariant under the rigid motions is discussed .
|
laser cooling and trapping methods are an important ingredient in the recent achievements in the low - temperature physics of gaseous atoms and molecules @xcite . in alkali atoms
one can reach very low temperatures with sisyphus and polarization gradient techniques , which surpass the doppler cooling method in efficiency , and appear in magneto - optical traps without additional efforts @xcite .
the drawback is that it becomes impossible to test the basic two - state doppler cooling theory with these systems .
the appearance of sub - doppler cooling is based on the hyperfine structure of the alkali atoms .
the same hyperfine structure is reflected in the complicated molecular state structure of the quasimolecule formed by two colliding atoms . in dilute atomic gases binary collisions
dominate the atomic interactions , and the quasimolecule states couple with the cooling and trapping light field .
this leads to inelastic light - assisted collisions that cause either loss or heating of the trapped atoms @xcite .
thus in alkali atoms the hyperfine structure leads to difficulties in testing the doppler theory , as well as in modelling the collisional processes .
the major isotopes of alkaline earth atoms , on the other hand , have no nuclear spin and thus the hyperfine structure is missing .
this makes the doppler limit for temperatures the true laser cooling limit , and allows for testing the basic doppler theory .
experimental studies have been published recently for @xmath5sr @xcite , and have been obtained for @xmath0 mg as well @xcite .
they indicate an intensity - dependent heating rate which prevents reaching the theoretical doppler limit .
inelastic laser - assisted collisions are expected to depend strongly on intensity even when the atomic transition is saturated ( @xmath6 0.444 w/@xmath7 for the @xmath1s@xmath2-@xmath1p@xmath3 laser cooling transition of @xmath0 mg ) @xcite . in this region
the problem is that most theoretical approaches are applicable only to the weak field situation , and can not handle the energy exchange when the energy change spectrum is continuous ( as is the case for radiative collisional heating ) .
an important exception is the monte carlo wave function method ( mcwf ) @xcite , which we apply here to the case of @xmath0 mg , to obtain the average energy increase per collision near the doppler limit ( at 1 mk for @xmath0 mg ) .
although simple in theory , the collisions between the alkaline earth atoms at strong fields are complicated since the partial waves and rotational states of the ground and excited states ( respectively ) of the quasimolecule form a network of two infinite sets of coupled states @xcite . in this article
we report the results of monte carlo simulations of collisional radiative heating , based on a truncated set of partial waves and rotational states .
we derive a heating rate due to these collisions , and compare it to the photon scattering heating rate that sets the doppler limit in the standard collisionless theory .
we also outline a channelling mechanism which is possible due to the specific structure and selection rules of the alkaline earth dimer quasimolecule , and which could lead to the suppression of inelastic processes , and test it with the simulations . here
we concentrate on the results provided by the simulations , and report the multitude of technical simulation details elsewhere .
the lasers for cooling and trapping in magneto - optical traps ( mot ) are normally detuned a few atomic linewidths ( @xmath8 ) below the atomic transition , which means that the attractive quasimolecule states have a finite internuclear distance , the condon point @xmath9 , where they become resonant with the laser field . for @xmath0 mg and other alkaline earth atoms we have two such states , a strongly coupled @xmath10 state , and a weakly coupled @xmath11 state @xcite . in the radiative heating mechanism ,
the colliding atoms on a partial wave @xmath12 are excited at @xmath9 to an attractive quasimolecule state with rotational quantum number @xmath13 .
the kinetic energy of the relative atomic motion increases as the atoms approach each other further , until they return to the ground state via spontaneous emission .
the energy exchange due to such a process forms a continuous distribution below the trap depth energy ( above which the radiative heating becomes radiative escape ) .
since the @xmath11 state is only weakly coupled to the light field , and provides mainly a channel for trap loss , we can concentrate in our modelling of heating on the @xmath10 state only ( see ref .
@xcite for a more detailed discussion ) .
the ground state partial waves and the excited rotational states are coupled by the selection rules @xmath14 , i.e. , the @xmath4 branch for molecular excitation is missing . in the weak field limit
one can consider each partial wave as paired independently to a rotational state , forming a set of independent two - state systems . at strong fields multiple couplings are possible , and one can move back and forth on the `` angular momentum ladder '' shown in fig .
[ fig : branching ] .
note that for the bosonic @xmath0 mg atoms the exchange symmetry allows only even values for @xmath12 .
_ channelling .
_ one should note , as fig .
[ fig : potentials ] illustrates , that each pair ( @xmath15 ) is coupled resonantly at a _
different _ value of @xmath9 .
thus at modest field strengths one could consider a situation where the separation of the condon points is valid while the excitation probability is close to unity . due to the missing @xmath4 branch
one can see that except for the @xmath16-wave , we always have a sequence @xmath17 so that an incoming ground state partial wave @xmath12 is channelled via two crossings into another ground state partial wave @xmath18 , then reflected by the centrifugal barrier , returning the same route back to the original state @xmath12 .
this process would reduce the collisions practically elastic ( the @xmath16-wave still remains `` unchannelled '' , though ) .
the energy increase from a single collision connects to a heating rate and the appropriate rate coefficient in the following way .
the collisional heating rate @xmath19 describes the kinetic energy change per unit volume and unit time . for identical particles it will be equal to @xmath20 , where @xmath21 is the rate coefficient for collisional heating , and
@xmath22 is the atomic density .
the factor of @xmath23 removes the doubling in collision counting for identical particles .
the total collisional heating rate is obtained by integrating @xmath24 over the trap volume
. for trap loss rate coefficient @xmath25 one normally calculates the collision frequency times the probability for loss , i.e. , @xmath26 , where @xmath27 is the relative velocity , and @xmath28 is the cross - section for collisions , including the probability for a loss event to occur .
this simple classical picture is connected to thermodynamics by assuming a thermal equilibrium distribution of velocities , @xmath29 , over which we take an average , and obtain @xmath30 .
the dynamics of the two - body collision on the microscopic level then enters in calculating @xmath28 . to calculate the heating rate coefficient @xmath21 is slightly more complicated , because in addition to the probability of an inelastic collision event to happen , we also need to estimate the amount of kinetic energy increase associated with it . technically , we would have a continuous distribution of final energies corresponding to each initial value of @xmath27 , due to the randomness of the spontaneous emission events . for practical reasons we consider an averaged rate , i.e. , we calculate the average change in kinetic energy per unit time . since we can perform the averaging over final energy states before the averaging over initial states , we define an inelastic ( heating ) cross section @xmath31 ( units energy@xmath32length@xmath33 ) that gives the difference of the average final relative kinetic energy and initial relative kinetic energy . in the partial wave approximation
we can write the quantum mechanical cross - section for identical atoms in a three - dimensional trap as @xcite @xmath34 where @xmath35 is the wave number related to @xmath27 and @xmath36 is the event probability .
thus , as a generalization , we write for heating @xmath37 where @xmath38 is the average single - collision energy increase related to the initial partial wave @xmath12 . in principle one
should take the thermal average over an isotropic maxwell - boltzmann distribution of relative velocities @xmath27 .
due to the complexity of the monte carlo simulations we are , in practice , limited to calculating @xmath38 for a rather narrow range of initial relative velocities @xmath27 . as in ref .
@xcite , we define a single - energy rate coefficient as ( @xmath39 ) @xmath40 with @xmath41 , where @xmath42 is the reduced two - particle mass .
the simulations have shown this to be a reasonable approximation .
we have performed a series of computationally tedious simulations at the detuning @xmath43 for various laser intensities , and the main results are given in table [ table1 ] .
we have used the semiclassical trajectory argument ( turning point should be at @xmath44 for any involved partial wave ) to truncate the sum in eq .
( [ khe ] ) . clearly a strong increase in the energy change with intensity continues above the saturation intensity @xmath45 .
the semiclassical landau - zener theory also indicates that in this intensity region excitation probabilities should be equal to unity .
thus the continued increase can be attributed to the re - excitation of the decayed population , as suggested in ref .
@xcite ; the delayed decay model predicts ( roughly ) @xmath46 .
.[table1 ] results from the multistate single - collision simulations .
the total number of states has been 12 , and the initial state has been @xmath47 .
the energy increase @xmath48 is calculated as a time average of the kinetic energy in the region where it is flat , subtracted by the initial energy .
the error refers to the statistical error of the average kinetic energy at the end of the monte carlo simulation . [
cols=">,>,^,^",options="header " , ] additional simulations showed that pre- and post - collision steady state formation is fast and strong for @xmath49 , and makes the energy exchange independent of the initial partial wave , i.e. , @xmath50 in eq .
( [ khe ] ) . in other words , in a system where the energy states are strongly coupled and remain so even asymptotically , the definition of an initial state is obscured , and we can actually make meaningful studies because the spontaneous emission always leads to the same steady state , independent of initial @xmath12 , well before @xmath9 is reached . for @xmath51
there is some initial state dependence , but on a scale that allows one to use the values given in table [ table1 ] as a reasonable estimate of magnitude . as for the channelling effect
, there was no sign of it in the simulations .
the explanation is rather clear .
the condon points are close to each other both in position and energy , and the true picture of the excitation process is more like adiabatic following of the field - dressed ( adiabatic ) quasimolecule states , shown in fig .
. however , this may not be the case for other values of detuning , or weaker fields than those used in our simulations .
for this reason we want to emphasise the general possibility for channelling , and also because it is a good test for the validity of the reduction into independent two - state models @xcite .
we have confined our computationally demanding simulations to large intensities because then there are suitably many quantum jumps for having reasonable statistics for small ensembles of 64 or 128 members .
finally it should be noted that any truncation of the partial wave / rotational state manifolds must be done so that the total number of states is even .
otherwise one would introduce an artificial dark state as one of the field - dressed eigenstates , into which the strong and fast steady state formation would trap the system well before the actual collision begins .
the heating rate prediction of the simulations must be compared to the photon scattering heating rate that enters the calculation of the doppler limit . in this
we follow ref .
@xcite and obtain the approximate rate @xmath52 where @xmath53 is the photon recoil energy , and @xmath54 is the saturation intensity @xcite . adding the numbers for @xmath0 mg
, we get the heating rate per particle ( @xmath55 ) @xmath56 the above expression has been derived for a one - dimensional two - beam setup , and its extension to the full six - beam three - dimensional situation is an interesting and open question , and one of the strong motivations to study doppler cooling with alkaline earth atoms .
the collisional heating rate coefficient becomes , using eq .
( [ khe ] ) , @xmath57 here @xmath58 , with @xmath59 nm for @xmath0 mg .
this result is obtained if we note that by truncating the partial wave series to the maximum classically allowed @xmath12 , denoted by @xmath60 , we have @xmath61 . the condon point @xmath9 is determined by the @xmath10 state potential , @xmath62 $ ] , being equal to @xmath63 at @xmath9 . finally , @xmath64 . to compare with the single - particle rate ( [ spr ] ) we write ( for @xmath43 ) @xmath65 for @xmath66 atoms/@xmath67 and @xmath55
this gives about 0.1 k / s .
this result implies that radiative collisional heating can not explain the problems in reaching the doppler limit for @xmath0 mg , because the densities are usually in the range of @xmath68 to @xmath69 atoms/@xmath67 .
we have earlier performed two - state studies for cs and found that there the heating rate present in sisyphus cooling should match the collisional heating rate at @xmath70 atoms/@xmath67 @xcite . as for making a comparison to the @xmath5sr case , without performing simulations
, we can make a rough estimate . in eq .
( [ khe ] ) we have written @xmath71 in a scaled form , so we can put in the corresponding values for @xmath5sr as well ( @xmath72 mk , @xmath73 nm , @xmath74k , and taking again @xmath75 ) .
the remaining problem is to estimate @xmath48 . in a toy model that assumes total excitation during a collision , and ignores re - excitation , we can write @xmath76 , where @xmath77 is the slope of the excited state ( ignoring @xmath13 ) potential at @xmath9 , and @xmath78 is the average survival time on the excited molecular state , and @xmath27 is the initial collision velocity .
this leads to an estimate of @xmath79 , which at least for @xmath0 mg gives 230 , i.e. , the correct magnitude near @xmath80 .
with all this , we get for @xmath5sr at the doppler temperature the result @xmath81 k@xmath67/s , to be compared with @xmath82 k / s .
these results meet only at densities of @xmath83 atoms/@xmath67 , which are a few magnitudes higher than those used in the experiment of ref .
this toy model also predicts a detuning dependence of @xmath71 of @xmath84 , i.e. , heating decreases with detuning , which means that while the number of involved partial waves increases with diminishing @xmath85 , the effect is more than countered by the reduction in the steepness of the excited state potential .
this study is the first attempt to describe quantitatively the radiative collisional heating by taking into account the inherent multistate structure due to the partial waves and rotational states .
the truncation of the partial wave manifold has required careful testing of various possibilities .
although it appears that collisional radiative heating does not play a role in the thermodynamics of the trap at realistic densities for @xmath0 mg ( and other alkaline earth atoms ) , our work does not rule it out completely .
although assumed to be irrelevant here , the weakly coupled @xmath11 state may play a role , as well as the two quasimolecule states corresponding to the repulsive potentials .
experimentally the role of collisional processes can be distinguished from other processes by looking at the scaling with density ; this has already been used to separate the cold collision contributions from the background collisions in trap loss .
while we aim to improve our modelling and to use our heating results in estimating radiative trap loss , we also expect that more experimental data will become accessible in the near future .
it should be pointed out that other effects may also lead to heating .
the methods to study e.g. radiative trapping ( reabsorption of scattered photons ) are , however , very different from the ones employed here .
the authors acknowledge the eu network cauac ( contract hprn - ct-2000 - 00165 ) and the academy of finland ( projects 50314 and 206108 ) for financial support .
jp has also been supported by the eu network cocomo ( contract hprn - ct-1999 - 00129 ) and the academy of finland ( project 204777 ) .
discussions with e. rasel , j. thomsen and j. ye are gratefully acknowledged , as well as useful comments by i. mazets and j. vigu .
the finnish it center for science ( csc ) is gratefully acknowledged for providing the supercomputer environment used in the simulations .
|
we report monte carlo wave function simulation results on cold collisions between magnesium atoms in a strong red - detuned laser field .
this is the normal situation e.g. in magneto - optical traps ( mot ) .
the doppler limit heating rate due to radiative collisions is calculated for @xmath0 mg atoms in a mot based on the @xmath1s@xmath2-@xmath1p@xmath3 atomic laser cooling transition .
we find that radiative heating does not seem to affect the doppler limit in this case .
we also describe a channelling mechanism due to the missing @xmath4 branch in the excitation scheme , which could lead to a suppression of inelastic collisions , and find that this mechanism is not present in our simulation results due to the multistate character of the excitation process .
|
the characterization of entanglement in many - body strongly correlated systems has been a very active research area in the last decade ( see e.g. ref .
@xcite for a review ) .
entanglement is expected to be particularly relevant in quantum phase transitions ( qpt ) and , in order to gather valuable insight , ground states of paradigmatic spin chain models have been exhaustively analyzed .
first , it was shown in refs .
@xcite that for the ising chain in a transverse field , pairwise entanglement measured by the concurrence between nearest ( and next - to nearest ) neighbors signals the position of the critical point ( but does not display critical behavior ) while is strictly zero otherwise .
the conjecture that , near criticality , entanglement should be present at all length - scales led to the concept of localizable entanglement , which was introduced in ref .
@xcite . for a given @xmath0-partite state @xmath1
, the localizable entanglement is the maximum average entanglement that can be made available to two pre - determined parties ( say , spins 1 and 2 ) , by performing general local measurements on the rest of the system .
if the correlation length diverges at criticality , so does the length of localizable entanglement .
however , the converse is not necessarily true @xcite .
a quite different approach was considered in ref .
@xcite , where the scaling of the entanglement between a given block of spins and the rest of a chain was analyzed against the size of the block itself . in such cases , the entanglement between two blocks can be fully determined simply by considering the von neumann entropy of a block ( also termed entanglement entropy ) . in this approach ,
contrary to the pairwise case , entanglement is clearly related to critical behavior .
away from criticality , the entanglement entropy reaches a constant value and shows logarithmic divergence with the size of the block when approaching the critical point .
conformal field theory was used to relate such divergence to the central charge of the corresponding effective theory , thus linking entanglement and the universality class of the corresponding qpt .
remarkably , both the entanglement entropy as well as the renyi entropies are functions of the eigenvalues of the reduced density matrix of the block . while all previous studies concern exclusively bipartite entanglement ( either pairwise or block - block ) a full description of many - body strongly correlated systems should include multipartite entanglement .
indeed , multipartite entanglement has been demonstrated in certain spin models @xcite and its role in qpt has been long discussed ( see e.g. ref .
@xcite and references therein ) .
however , the study of multipartite entanglement is presently much less developed in light of its daunting nature and the lack of appropriate tools .
in fact , we even lack general measures of entanglement for mixed states of three spins and it is not possible to extend the concept of entanglement entropy to more than two blocks . in this paper
we analyze long range multipartite entanglement in the vicinity of quantum phase transitions providing a general method that relies only on three - point correlators only .
the paper is organized as follows : in sec .
[ sec : rotinv ] we briefly review rotationally invariant states that will be used as the starting point for the construction of a family of entanglement witnesses ( i.e. , observables ) in sec . [
sec : witness ] .
we provide a sufficient condition to assess genuine tripartite entanglement in the ground state of a many - body system . by relying on the availability of three - point correlation functions only
, our approach does not depend on the hamiltonian properties ( type and range of the interactions , symmetries of the model , lattice geometry ) , its dimensionality or the actual choice of the subset of three spins picked from the lattice . with such tools at hand ,
we demonstrate that genuine multipartite entanglement ( gme ) is highly sensitive to quantum phase transitions . in order to provide a significant context where our formalism can be applied , we focus on the spin-@xmath2 xxz chain . in sec .
[ sec : xxz ] we first shortly review the properties of the model . then we consider the transition from the xy critical phase to the ferromagnetic one .
it is known that as the critical point is approached pairwise bipartite entanglement becomes independent of the distance between the spins @xcite . here
we show that also long range multipartite entanglement emerges in the vicinity of the the critical point . in this sense
we provide a finer - grained entanglement structure than previous analysis based on collective operators @xcite by proving that entanglement extends far beyond the nearest - neighbor scale .
we summarize our results in sec . [
sec : summary ] .
our analysis stems from rotationally invariant states of three qubits , whose entanglement properties have been unambiguously characterized in ref .
@xcite through a set of scalar inequalities . by projecting a generic state onto its rotationally invariant subspace
, we can also address gme in non - rotationally invariant states .
equivalently , one can detect gme by constructing suitable rotationally invariant entanglement witnesses . before proceeding
further we review , for completeness , the characterization of rotationally invariant states .
we then provide a geometrical description of such space and construct a family of entanglement witnesses .
the class of @xmath3 invariant tripartite states @xmath4 is defined as @xmath5=0\right\},\ ] ] where @xmath6 denotes the unitary irreducible representation of the rotations @xmath7 from the @xmath3 group . for three qubits ,
@xmath4 acts trivially in the two subspaces of total angular momentum @xmath2 and the one of @xmath8 .
however , since in general @xmath9\neq 0 $ ] ( with @xmath10 ) the two subspaces with @xmath11 can be mixed . denoting by @xmath12 the two orthogonal projectors onto the mixed @xmath11 subspaces , parametrized by the angles @xmath13 and @xmath14 , and by @xmath15 the projector onto the subspace @xmath16
, the density operator can be decomposed as @xmath17 where @xmath18 .
hence @xmath19 real parameters @xmath20 describe the set of three - qubit @xmath3-invariant states .
the construction can be extended to higher spins , although the number of parameters grows dramatically .
for instance , for spin-@xmath21 particles 13 variables are required .
as the @xmath3 and @xmath22 groups are isomorphic , the above representation can be straightforwardly mapped onto the one for @xmath22 invariant states @xcite @xmath23 where @xmath24 and the factor @xmath25 ensures normalization .
the hermitian operators @xmath26 read @xmath27 where @xmath28 is the permutation ( or swap ) operator acting on qubits @xmath29 and @xmath30 ; @xmath31 and @xmath32 are the two operators which cyclically permute all three particles , and @xmath33 denotes the identity . the operator @xmath34 ( @xmath35 ) is proportional to the projector @xmath15 ( @xmath36 ) .
the three remaining matrices @xmath37 ( @xmath38 ) act on the four - dimensional subspace of total spin-@xmath2 , follow the angular momentum commutation rules and are thus traceless . in order to ensure that eq .
represents a legitimate state , the coefficients @xmath39 must satisfy the conditions @xmath40
the entanglement characterization of three - qubit states distinguishes four classes @xcite : ( i ) separable states @xmath41 of the form @xmath42 , ( ii ) biseparable states @xmath43 belonging to the convex hull of states separable with respect to one of the partitions @xmath44 , @xmath45 or @xmath46 denoted by @xmath47 , @xmath48 , @xmath49 respectively , and two gme classes ( iii ) w - type states , and ( iv ) ghz - type states .
each class embraces those that are lower in the hierarchy , i.e. @xmath50 . the distinction between the w - type states and the ghz - type ones arises from the fact that for , three qubits , there are two non - equivalent classes of genuinely entangled states with representative elements being precisely the w state @xmath51 , and the ghz state @xmath52
. elements of one class can not be inter - converted into elements of the second one using stochastic local operations and classical communications .
therefore , the w and ghz classes are formed by convex combinations of states equivalent to @xmath53 and of combinations of states equivalent to @xmath54 . in ref .
@xcite , the subsets @xmath41 and @xmath43 of the @xmath22 invariant states are fully described in terms of inequalities for the coefficients @xmath39 . in particular ,
the set @xmath41 is constrained by the conditions @xmath55 . & \end{aligned}\ ] ] analogously , the states belonging to the set @xmath56 , can be shown to fulfill the condition @xmath57 ^ 2,&\end{aligned}\ ] ] where @xmath58 .
the corresponding sets @xmath48 and @xmath59 are found by rotating @xmath47 by @xmath60 around the axis @xmath61 .
finally , the set @xmath62 of genuine tripartite entangled states is found as the complement of @xmath43 in the set of all states .
\(a ) ( b ) + . (
a ) separable and ( b ) biseparable states .
( c ) a horizontal cut showing the three sets of biseparable states , their convex hull and the tangent in the point @xmath63 to the set representing the witness .
( d ) same as in ( c ) but a vertical is shown.,title="fig:",scaledwidth=23.0% ] .
( a ) separable and ( b ) biseparable states .
( c ) a horizontal cut showing the three sets of biseparable states , their convex hull and the tangent in the point @xmath63 to the set representing the witness .
( d ) same as in ( c ) but a vertical is shown.,title="fig:",scaledwidth=23.0% ] \(c ) ( d ) + .
( a ) separable and ( b ) biseparable states .
( c ) a horizontal cut showing the three sets of biseparable states , their convex hull and the tangent in the point @xmath63 to the set representing the witness .
( d ) same as in ( c ) but a vertical is shown.,title="fig:",scaledwidth=24.0% ] for real hamiltonians , the above description is further simplified as their ground states and their reductions are represented by real density operators .
this is equivalent to setting @xmath64 in ( [ eq : ourrep ] ) or @xmath65 in ( [ eq : wernerrep ] ) and allows us to visualize the set of rotationally invariant states in the space @xmath66 ( fig .
[ fig : sets ] ) .
the complete set is a cone with symmetry axis parallel to the axis @xmath61 . in fig .
[ fig : sets](a - b ) we depict the set of separable @xmath41 and biseparable @xmath43 states , respectively ; the complementary volume contains genuine tripartite entangled states .
[ fig : sets](c ) show a horizontal section of the cone for a fixed @xmath61 and fig .
[ fig : sets](d ) a vertical section .
we notice that a necessary condition for a state to be tripartite entangled is @xmath67 . below this value ,
biseparable states fill the cone completely .
the above criteria can be extended to all states using the _ twirling map _ that projects each state onto its rotational invariant subspace : @xmath68 where @xmath69 consists of local unitaries @xmath70 .
the following statements now hold : ( i ) @xmath71 is @xmath22 invariant ( ii ) if @xmath4 is separable then @xmath71 is separable @xcite , ( iii ) if @xmath72 is biseparable but not separable then @xmath73 is not separable and ( iv ) if @xmath71 is genuine tripartite entangled so is @xmath4 . the geometrical description of the hilbert space depicted in fig .
[ fig : sets ] facilitates the construction of a multipartite entanglement witness , i.e. , an observable @xmath74 such that @xmath75 @xmath76 , and there exists at least one state @xmath77 such that @xmath78 .
it is sufficient to choose a witness of the form @xmath79 , where @xmath80 , so its expectation value with a rotationally invariant state simplifies to a scalar product @xmath81 .
we determine @xmath82 using the geometric description of the witness as a plane intersecting the cone of rotationally invariant states and tangent to the set @xmath83 at the point @xmath84 with normal vector @xmath85 [ see fig . [
fig : sets](d ) ] .
we find @xmath86 and , from the definition of @xmath82 , we obtain @xmath87 .
the witness plane is calculated with @xmath84 at the midpoint of the line between the biseparable subsets @xmath88 and @xmath89 .
it also clearly depends on the choice of @xmath61 .
in fact @xmath90 the witness can then be rotated about the @xmath61 axis by @xmath91 to obtain witness planes tangential to @xmath43 on the lines between @xmath92 and @xmath93 , or @xmath94 and @xmath89 , respectively .
the explicit derivation of eq . and the demonstration that @xmath74 is a witness can be found in the appendix [ app : witness ] .
let us now focus on the study of gme in an open chain of size @xmath0 described by the xxz hamiltonian @xmath95 , \label{xxz}\ ] ] where @xmath96 is the anisotropic parameter . for lattices with even @xmath0 ( and also in the thermodynamic limit )
the sign of @xmath97 is irrelevant since it is possible to change @xmath98 and @xmath99 for all even ( odd ) sites , thus the hamiltonian is @xmath22 invariant for @xmath100 . for any value of @xmath96 , @xmath101 is @xmath102 invariant . for non - bipartite lattices ( or odd @xmath0 )
geometrical frustration can appear . in
what follows , we assume without loss of generality that @xmath103 . .
for @xmath104 all the concurrences collapse to the same value : @xmath105 close to the isotropic ferromagnetic point ( @xmath106 ) .
the reduced density matrices have been calculated using dmrg simulations with @xmath107 . for convenience
the spin arrangements are denoted as @xmath108 ( corresponding to a pair of nearest - neighbor spins , next to nearest - neighbor spins , etc . ) , however the entanglement is studied close to the middle of the chain . in the chain of length n
the spins correspond to @xmath109 with @xmath110 .
the plotted quantity is dimensionless.,title="fig : " ] + the complete phase diagram of the model is well known : the ground state is ferromagnetic for @xmath111 , xy - critical for @xmath112 , and ising - antiferromagnetic ( nel ) otherwise . at @xmath113
a first order transition separates the ferromagnetic and critical phases .
this point is not conformal and has recently attracted some attention @xcite .
here we focus on the multipartite entanglement content in the vicinity of this phase transition . before proceeding further ,
notice that at @xmath113 the highly degenerate ground state is in the @xmath22 isotropic ferromagnetic multiplet spanned by any state with maximum total angular momentum @xmath97 : @xmath114 for all possible values of @xmath115 .
neither of these states exhibit finite size corrections to the energy per site , nor are they rotationally invariant
. however , each of them corresponds to a symmetric dicke state , i.e. , @xmath116 with @xmath117 subsytems in the state @xmath118 and the remaining in the state @xmath119 ; @xmath120 are the elements of the permutation group , @xmath121 is the binomial coefficient , and @xmath122 is the computational basis ( spin up , spin down ) .
thus , the interchange between any two spins leaves the corresponding dicke state unchanged and by construction any linear combination as well . in the region @xmath112 ,
the total spin , @xmath97 , is not well defined but @xmath123 . in the limit
@xmath124 the ground state was found numerically to be an equally weighted superposition of all the elements of the standard basis within the sector @xmath123 @xcite .
finally , notice that for @xmath125 , the ground state is a rotationally invariant singlet with @xmath126 .
\(a ) ( b ) @xmath127 . as @xmath128
( see the insets ) gme is detected in all the arrangements depicted in ( a ) , and those shown in ( b ) except for @xmath129 .
the reduced density matrices have been calculated using dmrg simulations with @xmath107 . for convenience
the spin arrangements are denoted as @xmath130 , however the entanglement is studied close to the middle of the chain .
the plotted quantities are dimensionless.,title="fig : " ] \(b ) ( b ) @xmath127 .
as @xmath128 ( see the insets ) gme is detected in all the arrangements depicted in ( a ) , and those shown in ( b ) except for @xmath129 .
the reduced density matrices have been calculated using dmrg simulations with @xmath107 . for convenience
the spin arrangements are denoted as @xmath130 , however the entanglement is studied close to the middle of the chain .
the plotted quantities are dimensionless.,title="fig : " ] we now analyze the entanglement content of the model by computing both the bipartite concurrence @xmath131 of two qubits at distance @xmath132 as well as the mean value of the entanglement witness ( [ eq : witness ] ) , @xmath133 , for three qubits @xmath134 .
concurrences can be analytically obtained in the @xmath22 multiplet after realizing that any reduction of a multipartite symmetric dicke state @xmath135 is also symmetric , i.e. independent of @xmath132 and @xmath29 , and read @xcite : @xmath136}}{2n(n-1 ) } \nonumber\end{aligned}\ ] ] since the only non - zero matrix elements @xmath137 are the symmetric ones given by @xmath138 , @xmath139 . as previously noted
, the value of the concurrence for large @xmath0 is very small and close to the exact value @xmath140 achieved for @xmath141 ( @xmath142 ) .
thus , for large @xmath0 , all members of the @xmath22 multiplet ( except the trivially separable @xmath143 ) have equal concurrences which tend to zero in the thermodynamic limit .
let us now discuss our results .
we compute the ground state of @xmath101 for the whole phase diagram using the density matrix renormalization group ( dmrg ) @xcite for open chains of up to 192 sites ( we have checked that the accuracy of the results does not depend on the size of the chain ) . from the ground state , we construct the reduced density matrix of either two or three spins ( not necessarily adjacent ) close to the centre of the chain to avoid edge effects .
we then calculate the corresponding concurrences as well as the tripartite entanglement by means of the entanglement witness ( [ eq : witness ] ) for the whole phase diagram .
there exist suitable multipartite entanglement witnesses that detect gme of adjacent sites for two - body models @xcite or global gme of dicke states @xcite .
our method , however , allows us to choose our reduced system at will without imposing further symmetries .
our results are summarized in fig .
[ fig : conc ] , where we display @xmath144 for different values of @xmath132 as a function of @xmath96 , particularly near the isotropic point @xmath113 . in accordance with previous results based on bipartite measures @xcite , we observe that close to the isotropic ferromagnetic point ( i.e. @xmath145 ) all concurrences for @xmath104 collapse to the same value @xmath105 . to investigate whether or not this feature of bipartite entanglement is also shared by the multipartite structure , we investigated the minimum mean value of the witness @xmath146 for different spin arrangements ( the spin @xmath29 is near the middle of the chain to avoid edge effects ) .
for adjacent sites @xmath147 we recover previous results @xcite , indicating the presence of gme for @xmath148 ( see fig .
[ fig : witness](a ) ) .
long - range gme is also detected by the witness for other spin - arrangements @xmath134 when at least two site indexes have different parity ( see fig . [
fig : witness](b ) ) .
our results provide evidence that ( i ) distant multipartite entanglement is present in the system ; ( ii ) the global @xmath22 symmetry is already broken very close to the qpt , evidencing the sensitivity of gme to the fine ground state structure .
as we demonstrated for the ground state of the xxz model , the method we provide is not restricted to @xmath22 invariant states and can be applied to any 3-qubit states .
it can also be extended to ground and thermal states of spin-1/2 hamiltonians in any lattice geometry , not only 1d .
notice however that while the detection of gme for @xmath22 invariant states is unambiguous , the negativity of the expectation value of the witness @xmath149 is in general only a sufficient condition for gme .
we have constructed a witness for genuine tripartite entanglement that , although explicitly designed for rotationally invariant states , is useful to assess the multipartite quantum correlations of states lying outside such a class .
our method is general and can be used together with analytical or numerical methods , e.g. dmrg , exact diagonalization , quantum monte carlo or , more generally , any technique yielding three point correlation functions .
we have used such a tool to gather insight into the structure of quantum correlations of the many - body ground state of a spin chain close to a first order quantum phase transition , beyond the standard framework of bipartite entanglement ( see also @xcite ) .
we are indebted to o. ghne and m. hofmann for clarifying insights .
we thank r. augusiak , j. calsamiglia , a. luchli , r. muoz - tapia , and j. serra for useful discussions .
we acknowledge financial support from the spanish mineco ( fis2008 - 01236 ) , european regional development fund , generalitat de catalunya grant no .
sgr2009 - 00347 .
js acknowledges the support of fundaci catalunya
la pedrera and spanish micinn ( fis208 - 00784 toqata ) .
mp acknowledges the uk epsrc for a career acceleration fellowship and a grant from the `` new directions for epsrc research leaders '' initiative ( grant no .
ep / g004759/1 ) .
gdc and mp thank the epsrc funding under the scheme : ep / k029371/1 and the john templeton foundation ( grant i d 43467 ) for financial support .
a plane tangent to the surface of the set of biseparable states is uniquely characterised by a point @xmath84 and a normal unit vector @xmath85 .
we choose the point lying on the line between sets @xmath150 and @xmath151 ( later denoted by @xmath152 ) and define a curve parametrized by @xmath61 as @xmath153\end{aligned}\ ] ] with @xmath67 .
then the normal vector @xmath85 is the cross - product of the vector tangent to the curve @xmath154 : @xmath155,\ ] ] and the vector @xmath156 parallel to the @xmath152 .
therefore , we have @xmath157.\end{aligned}\ ] ] substituting eq .
( [ eq : normal ] ) in the expression of the witness : @xmath158 we obtain the unnormalized witness in the following form @xmath159 the other two witnesses obtained by rotation of the witness plane by @xmath91 are @xmath160 by construction the operator @xmath74 is positive for all biseparable rotationally invariant states , even those with an imaginary component since @xmath161 .
furthermore , @xmath74 is an entanglement witness for genuine tripartite entangled states , i.e. , it is positive for all biseparable states . for this purpose
it is enough to show that @xmath162 for all @xmath163 being product vectors with respect to bipartitions @xmath164 , @xmath44 , @xmath165 . without loss of generality
we focus on the specific partition and have : @xmath166 where we use the fact that @xmath74 is rotationally invariant and that the state @xmath167 is rotationally invariant and either biseparable or separable .
note that eq .
( [ eq : wit_twirl ] ) establishes equivalence between the biseparability test based on the witness and the twirling criterion for all real density matrices .
|
we provide a novel insight in the quantum correlations structure present in strongly correlated systems beyond the standard framework of bipartite entanglement . to this aim
we first exploit rotationally invariant states as a test bed to detect genuine tripartite entanglement beyond the nearest - neighbor in spin-1/2 models .
then we construct in a closed analytical form a family of entanglement witnesses which provides a sufficient condition to determine if a state of a many - body system formed by an arbitrary number of spin-1/2 particles possesses genuine tripartite entanglement , independently of the details of the model .
we illustrate our method by analyzing in detail the anisotropic xxz spin chain close to its phase transitions , where we demonstrate the presence of long range multipartite entanglement near the crtical point and the breaking of the symmetries associated to the quantum phase transition .
|
the study of the korteweg de - vries ( kdv ) equation @xmath0 has been an interesting issue since the discovery of soliton .
its exact solution@xcite can be used to describe the overtaking collision of soliton on a uniform layer of water , but the solution is only physically meaningful for the unidirectional soliton .
all the left moving solitons with zero boundary conditions are prohibited . in ref.@xcite , three sets of model equations are derived for modelling nonlinear and dispersive long gravity waves travelling in two horizontal directions on shallow waters of uniform depth .
a good understanding of all solutions of these models are helpful for coastal and civil engineers to apply the nonlinear water wave model in a harbor and coastal design .
therefore , finding more types of solutions of these equations are fundamental interest in fluid dynamics .
for the various different models , a point of central interest is to examine and compare the solitary wave solutions .
for the case of plane motion , the three set of models can be rewritten as the boussinesq class after omitting the higher order terms : ( a ) .
the @xmath1 system , the depth - mean velocity basis @xmath2_x=0 .
\end{aligned}\ ] ] ( b ) .
the @xmath3 system , the bottom variable basis @xmath4_x=\dfrac{1}{6}\underline{u}_{xxx}. \end{aligned}\ ] ] ( c ) .
the @xmath5 system , the surface variable basis @xmath6_x=-\dfrac{1}{3}\hat u_{xxx}. \end{aligned}\ ] ] in the boussinesq types of systems - , the field @xmath7 is wave elevation and @xmath8 and @xmath9 are depth - mean , bottom and surface velocities respectively . to study the systems - uniformly , we can extended them to the following generalized boussinesq system ( gbqs ) : @xmath10_x = c_2u_{xxx}. \end{aligned}\ ] ] the system has been shown to be integrable and has hamiltonian structures@xcite . the exact solitary wave and periodic wave solutions of have also been obtained by many authors@xcite . especially , zhang and li@xcite presented a theory of bidirectional solitons on water by using integrable boussinesq surface - variable equation .
however the integrability of the other two systems and or generally the gbqs is not known . in this paper
, we will show that the gbqs is non - painlev integrable generally except for a special equivalent case of . though the model is non - painlev integrable
, it is still possible and useful to give out some interesting solitary wave solutions . in this paper , we are concentrate on to study the exact solitary wave solutions of the gbqs under the physically significant boundary conditions @xmath11 because the integrability and the soliton solutions of the model have been known in literature , we always assume @xmath12 except for the special cases which will be particularly pointed out if it is necessary . in the section 2 of this paper ,
the non - painlev integrability of the gbqs is studied by using the standard weiss - tabor - carnevale ( wtc ) approach@xcite .
a special explicit solitary wave solution with a specific velocity selection is given by the truncated painlev expansion . in sec .
3 , after reflecting the problem to find the possible solitary waves to the possible motions of the newtonian type quasi - particles moving in some proper potential fields , the velocity prohibition and selection phenomena for the solitary waves of the gbqs are discussed . for the general solitary wave with allowed velocities , an implicit form of the exact solitary wave solutions
some special solitary wave solutions are plotted also according to the implicit expression .
the last section is a short summary and discussions .
* 2.1 non - painlev integrability of the gbqs with @xmath13 * in the modern soliton theory , the study of the painlev property@xcite plays a very important role because it can be used not only to isolated out ( painlev ) integrable models@xcite but also to find many other integrable properties such as the bcklund transformations , lax pair , schwarzian form etc@xcite . furthermore ,
even if the model is non - painlev integrable , the method can still be used to find some useful things such as the special exact explicit solutions@xcite .
because it has been known that the model ( i.e. , with @xmath14 ) is integrable and the general real physical model may require @xmath12 , say , and , we only check the painlev property of the model with @xmath15 in this subsection and find a special solitary wave solution from the next subsection .
the painlev property of a partial differential equation system is defined as all the solutions of the system are free of the essential and branch movable singularities around an arbitrary ( both characteristic and non - characteristic ) singular manifold@xcite . according to the above definition of the painlev property ,
the general solutions of the gbqs should have the following expansion around the arbitrary singular manifold @xmath16 @xmath17 where @xmath18 should be negative finite integers .
furthermore , if the model possesses the painlev property , there should be at least one primary branch ( one possible selection of @xmath19 ) such that enough arbitrary functions ( six for the gbqs , @xmath20 and five of @xmath21 ) can be included in the painlev expansions . by using the standard leading order analysis , i.e. substituting @xmath22 into the gbqs and balancing the leading terms
, we can find that there is only one possible selection @xmath23 for @xmath24 .
for @xmath25 , there is no possible negative integer selection for @xmath26 and @xmath27 . substituting the above expansion with to eq . and
vanishing all the coefficients of the powers @xmath28 , we can find the following recursion relation for determining the expansion functions @xmath29 and @xmath30 @xmath31 where @xmath32 and @xmath33 are all only functions of @xmath34 and the derivatives of @xmath20 .
from the recursion relations and , we know that all the functions @xmath29 and @xmath30 are fixed by the recursion relations except for those related to the resonance points determined by vanishing the coefficient determinant of and : @xmath35 from the expression and by checking the resonance conditions at the resonant points @xmath36 and @xmath37 , we know that there are two facts which destroy the painlev integrability of the gbqs with @xmath12 .
the first one is that there is no primary branch at all because the both equations of the qbqs are three order , so the primary branch should have six resonant points such that six arbitrary functions can be entered into the general painlev expansion solution .
the second one is the resonance conditions at @xmath38 and @xmath39 are not satisfied .
the lack of the primary branch of the model means some kinds of the logarithmic branch and/or essential singularities around the singular manifold @xmath40 may appear for the gbqs with @xmath41 . actually ,
substituting @xmath42 into the gbqs , we can find that there exists a possible non - completely negative selection for the constants @xmath26 and @xmath27 : @xmath43 for both @xmath24 and @xmath25 cases ! the allowance of the selection @xmath44 means that the logarithmic term @xmath45 should be included in the expansion of the function @xmath46 to balance the leading order terms of the first equation of . according to the above discussions
, we can conclude that the gbqs with @xmath12 has no painlev property .
in other words , the model is non - painlev integrable .
though the model is non - painlev integrable , it is still possible to find some useful information from the painlev analysis . in the next subsections we are specially interested to find the possible solitary wave solution from the truncated painlev expansion . * 2.2 truncated painlev expansion and explicit exact solitary waves with special velocity selections * according to the discussion of the last subsection , we know that the standard truncated painlev expansion has the form @xmath47 where @xmath48 is a seed solution of the gbqs and usually is taken as constant solution to get the single travelling solitary wave solution of the model . as usual , substituting the truncated expansion with constants @xmath49 and @xmath50 into the gbqs and vanishing the coefficients for different orders of @xmath20 , we have @xmath51 for the functions
@xmath52 while the function @xmath20 should be determined from the following over - determined system @xmath53 @xmath54 @xmath55 \nonumber\\ & & \quad -2u_0\phi_t+[24c_1(u_{0t}-u_1\phi_{t})_x-2v_0-u_1 ^ 2 - 2u_2u_0)]\phi_x=0,\label{f31}\end{aligned}\ ] ] @xmath56-(u_1+u_1v_2+v_1u_2)\phi_x\nonumber\\ & & \quad ( 1+v_2)u_{0x}+v_{0t}+v_{0x}u_2+(v_1u_1)_x - v_1\phi_t=0,\label{f22}\end{aligned}\ ] ] and @xmath57c_2=0 .
\label{f32}\end{aligned}\ ] ] to find all the possible exact solutions of the over - determined system is quite difficult .
however , it is a quite easy work to find the travelling solitary wave from the system . for the travelling wave solution @xmath58 , we have @xmath59 with @xmath60 being an arbitrary velocity parameter . under the travelling wave solution condition ,
are simplified to @xmath61 substituting into and yields @xmath62\phi_{\tau\tau\tau}=0 .
\label{phi2}\end{aligned}\ ] ] the equation system and is consistent only for @xmath63 where the constant @xmath64 is introduced for convenience later . using the constant relation , the general solution of ( and )
reads @xmath65 with @xmath66 and @xmath67 being arbitrary constants . substituting with into
, we find that the constants appeared in have to be fixed by @xmath68 furthermore , without loss of generality , the constant @xmath69 can be taken as zero because @xmath64 is arbitrary ( can be taken as both positive and negative ) . after finishing some simple direct calculation , one can find that with solves all other remained equations . substituting , , with , and @xmath70 into , we get a special solitary wave @xmath71,\label{41u}\\ v & = & \frac{c_2 ^ 2 + 2c^2c_1 ^ 2(c_2k^2 - 2)}{4c^2c_1 ^ 2}-\frac32 c_2k^2 { \rm sech}^2 \left[\frac k2(\tau-\tau_0)\right].\label{41v}\end{aligned}\ ] ] where @xmath72 .
in general , the boundary conditions expressed in can not be satisfied for the solitary wave solution expressed by and . in order to satisfy the boundary conditions for and
, we have to fix the constants @xmath64 and @xmath60 as @xmath73 and then the solitary wave solution becomes @xmath74,\label{41su}\\ v & = & -\dfrac{3}{4}(c_0 ^ 2 + 2){\rm sech}^2\left[\dfrac{\sqrt{2c_1(c_0 ^ 2 + 2)}}{4c_1c_0 } \left(x\mp \dfrac{c_0 ^ 2}{\sqrt{2-c_0 ^ 2}}t-\tau_0\right)\right ] .
\label{41sv}\end{aligned}\ ] ] from the above discussions , the zero boundary travelling solitary wave solution ( and ) obtained via the truncated painlev expansion is valid only for the gbqs with the condition @xmath75 and @xmath76 for @xmath41 . the solitary wave solution ( and )
possesses only two special isolated velocities @xmath77 .
is there any other travelling solitary wave with other velocities for the gbqs ?
the answer is positive because in the truncated painlev expansion approach we require all the coefficients of the different powers of @xmath20 being zero .
this strong condition leads to the loss of the generality . in the next section
, we discuss this problem generally by casting the problem to the possible motions of a newtonian classical quasi - particle in some possible potentials .
from the last section , we know that by means of the truncated painlev expansion approach , we can only obtain a special exact solitary wave solution with the boundary condition . actually , the travelling solitary waves with different velocities do exist for the gbqs .
though we can not explicitly write down the exact solitary wave solutions with the boundary condition for the different velocities , we can find an uniform and implicit formula for the solitary wave solutions except for some isolated special critical velocities . for a travelling wave solution , @xmath78 ,
the motion equation system becomes an ordinary differential equation system @xmath79 where the both equations have been integrated once with respect to @xmath80 and the integration constants have been fixed as zero because of the boundary conditions given in . from the first equation of , we know that the travelling wave of the field @xmath7 can be simply expressed by @xmath81 substituting it into the second equation of we have @xmath82 the first integration of reads @xmath83 for @xmath15 and @xmath84 for @xmath85 . to get the relations and ,
the integration constants have been fixed appropriately such that the boundary conditions may be satisfied .
now , from the expressions , it is known that to find the possible travelling solitary wave solutions of the gbqs with boundary condition is equivalent to find the possible special motions of a classical quasi - particle moving in the potential fields @xmath86 and/or @xmath87 related to the _ maximum point at @xmath88 with the space " variable @xmath46 and the time " variable @xmath80@xcite .
_ if the existence problem of the solitary wave is solved , except for some special critical cases ( see later ) , the travelling solitary waves with the boundary condition can be uniformly expressed implicitly by @xmath89 where @xmath86 is defined in and the integration constant @xmath90 is related to the location of the solitary wave . to see the possible solitary wave solutions qualitatively , the possible motions of the quasi - particle in the potentials @xmath86 and/or @xmath87 , it is convenient to study the structures of the potentials @xmath86 and @xmath87 at the same time . to study the structures of the potentials @xmath86 , we firstly isolate out some special and/or critical cases .
the first special case is same as that in . when the velocity @xmath60 of the solitary wave is fixed by , the logarithmic term of vanishes and the function @xmath46 can be explicitly integrated out .
the result is reasonably same as that obtained in the last section via the truncated painlev analysis . the potential structure at this special case is plotted in fig . 1a for the special model , i. e. @xmath91 and then @xmath92 and @xmath93 .
the corresponding soliton solution for the field @xmath46 is plotted in fig.1b .
epsf the second special critical case is that if the velocity is determined by @xmath94 then the logarithmic term of is also vanished .
however , in this case there is no solitary wave solution with the boundary condition except for @xmath95 , i.e. , @xmath96 .
in other words , the solitary wave solutions with the velocities @xmath97 are prohibited for the gbqs with @xmath98 .
this velocity prohibition property can be observed more clearly from . after substituting into
, one can directly see that it is impossible to satisfy the boundary condition except for @xmath95 . while if @xmath95 , is same as and then the related solitary wave solution is also given by . the related potential structure of the potential @xmath86 with @xmath99 and
then @xmath100 is plotted in fig .
2 . from fig .
2 , we can see also that when @xmath101 , @xmath88 is not an extremum of the potential while a solitary wave is corresponding to a possible special motion of the quasi - particle related to a maximum of the potential .
epsf the third special critical case is related to the static solitary wave solution .
for the static solitary wave solution of the gbqs , two subcases , @xmath102 and @xmath103 , should be clarified . for the gbqs with @xmath104 ,
is not valid and has to be changed as . the corresponding structure for the potential @xmath87 is plotted in fig .
3a for @xmath105 and the corresponding solitary wave solution is plotted in fig . 3b . from and , it is straightforward to get that the static solitary wave of the gbqs with @xmath102 has the form @xmath106 epsf fig .
4 shows the structure for the potential @xmath87 with @xmath107 which includes the integrable case . from and fig .
4 , it is also easy to see that whence @xmath108 , @xmath88 is a minimum of the potential @xmath87 .
so there is no static solitary wave of the gbqs with @xmath108 and zero boundary conditions because a solitary wave corresponding to the motion of the classical quasi - particle in the potential field related to a _ maximum .
furthermore , from with @xmath85 and @xmath25 , we can also know that there is no static solitary wave at all . in summary ,
the static solitary wave with zero boundary condition is prohibited for the gbqs with @xmath103 .
especially , there is no static solitary wave with zero boundary conditions for the systems and even if is integrable .
_ epsf the fourth type of velocity selection is related to @xmath109 usually , the solitary waves are decay exponentially apart from the solitary wave center . however , for the gbqs with @xmath110 , if the solitary wave possesses the velocities @xmath111 then it decays only algebraically .
this velocity selection property can be seen from the second order differentiation with respect to @xmath46 of the potential @xmath86 at the point @xmath88 .
the second order differentiation of @xmath86 for @xmath112 reads @xmath113}{2c_1(1-cu - c_0 ^ 2)^2},\ ( c^2=1).\ ] ] it is known that the boundary values ( @xmath88 in our case ) of exponentially decayed solitary wave solutions are linked with the maxima of the potential @xmath86@xcite .
however , from we know that when the velocity @xmath114 , @xmath88 is only an inflexion point of the potential @xmath86 .
it is also known that an inflexion point may be related to a rational ( or algebraic ) solitary wave solution@xcite .
this type of the algebraic solitary wave is similar to some what of the solitary wave at the critical point where the phase transition occur for some types of quantum fields and condense matter systems@xcite . to see the algebraic decay property of the solitary waves at this critical case , we take @xmath115 and @xmath116 . under this parameter selection ,
the related potential becomes @xmath117 the corresponding structure for the exact potential @xmath118 and the approximate potential @xmath119 are plotted at the same figure , fig .
5a . from fig .
5a , it is seen that two lines are almost coincide with each other at the plotted region which related to the quasi - particle move in . actually , the approximation is quite well up to @xmath120 while the solitary wave is located only at the region @xmath121 .
so to see the solitary wave structure in this special case we can safely use the approximate potential @xmath119 to calculate the related solitary wave .
the result reads @xmath122 fig .
. shows the structure of the algebraic solitary wave expressed by .
epsf from the analysis of the existence problem of the solitary waves at the critical velocities , we know that for different velocities the solitary waves may possesses different shapes .
this type phenomena occurs also in integrable cases .
for instance , for the bidirectional kaup - kupershmidt equation , its right moving solitons possess single - humped shape while its left moving solitons possess double - humped shape@xcite . in integrable case ,
the velocity prohibition phenomena are also common .
for instance , for the system the static solitary wave with zero boundary condition is prohibited and for the kdv equation , all the left moving and static soliton solutions with zero boundary conditions are prohibited .
now the important problem is in addition to the selections and prohibitions at the critical points whether there exist further prohibitions for some other velocity regions . to solve this problem , we restrict ourselves for @xmath123 ( because most of the real physical systems , say , and , possess this property ) and @xmath124 ( because the symmetry property @xmath125 of ) . to find the existence conditions of the solitary waves of the gbqs with zero boundary conditions for noncritical cases
, we can use the maximum condition , @xmath126 and the real condition of the potential @xmath86 .
the maximum condition of the potential @xmath86 at @xmath88 reads @xmath127 and the real condition of @xmath86 , reads @xmath128 after finishing the detail analysis with help of the conditions and , we find the following six different structures for the potential @xmath86 for @xmath123 : _ case 1.@xmath129 when @xmath130 , the square of the velocity parameter , is located in the range @xmath131 with @xmath132 , the related potential @xmath86 possesses the structure as shown in fig .
6a for @xmath124 . from fig .
6a we know that there exist a special motion for a classical quasi - particle moving in this potential related to the maximum at @xmath88 : at the beginning ( @xmath133 ) , the quasi - particle is located at the peak center @xmath88 , as time
@xmath80 " increases , the quasi - particle roll " down the hill up to the @xmath134 point and then return back to the original point @xmath88 at time " @xmath135 .
this type of special motion of the quasi - particle is just related to the solitary wave solution of the gbqs with . from fig .
6a we know also that the quasi - particle can roll only to right ( the positive @xmath46 direction ) that means the right moving solitary wave ( @xmath124 ) possesses bell shape for the field @xmath46 . by using the invariant transformation of , @xmath136 , we know that the left moving solitary wave ( @xmath137 ) possesses ring shape . the corresponding exact solitary wave with the same parameters as shown in fig .
6a is plotted in fig .
_ epsf _ case 2 .
@xmath138 in this case , the related potential @xmath86 possesses the structure as shown in fig . 7 for @xmath124 . from fig .
7 we know that there is no solitary wave solution with zero boundary condition in this case because @xmath88 is only a minimum of the potential .
if the quasi - particle stay at @xmath88 at the beginning then it can only be stayed there forever .
in other words , the solitary waves of the gbqs with the velocities @xmath139 and zero boundary condition are totally prohibited .
_ epsf _ case 3 .
@xmath140 the structure of the potential @xmath86 for @xmath141 with @xmath142 possesses the form as shown in fig .
similar to the case 1 , there exist the right moving bell shape solitary waves and the left moving ring shape solitary waves for the field @xmath46 .
the right moving bell shape solitary wave solution related to fig .
8a is plotted in fig .
_ epsf in the first three cases , @xmath143 is less than @xmath144 and the special models and satisfy this condition .
when this condition is not satisfied , we have three other different potential structures . _
@xmath145 the corresponding potential structure of the fourth case is plotted in fig .
9a . different from the first and third cases , from fig . 9a we know the quasi - particle located the @xmath88 hill may roll down the hill to both the left and right sides and finally return back to the hill .
so in this case , the right moving solitary waves may have both bell and ring shapes and the left moving solitary waves possess the same property .
the right moving bell shape solitary wave solution related to this case with the same parameters as fig .
9a is plotted in fig .
9b . from fig .
9a , we can know also that for the right moving solitary waves , the bell shape solitary waves possess larger amplitudes and for the left moving solitary waves , the ring shape solitary waves possess larger amplitudes .
_ epsf _ case 5 .
@xmath146 when @xmath130 is located in the range @xmath147 , the corresponding potential structure possesses the form as shown in fig .
10 . from fig .
10 we know that @xmath88 is only a minimum of the potential .
so similar to the discussion of the case 2 , we can conclude that there is no zero boundary solitary wave for the gbqs with the condition .
_ epsf _ case 6 .
@xmath148 for the last case , the related potential structure is shown in fig .
11 . from fig .
11 we know that @xmath88 is really a maximum of the potential , the quasi - particle may roll down the @xmath88 hill in both sides , however , it can not be back !
that means the fast moving ( @xmath149 ) solitary waves for the gbqs with the condition @xmath150 are completely prohibited also . _
epsf combining the critical cases , we summarize the conclusions for the possible solitary waves of the gbqs with zero boundary conditions and @xmath123 in the following table .
.2 in * table 1 .
existent and prohibited regions of the solitary waves . * [ cols="^,^,^,^,^",options="header " , ]
in summary , using the standard wtc s painlev pde test , we prove that the gbqs is non - painlev integrable except for the special case for @xmath14 , i.e. , . for the non - painlev integrable gbqs with zero boundary conditions
, the truncated painlev expansion approach leads to only a special @xmath151 shape solitary wave solution with a special velocity selection . to find all the possible travelling solitary wave solutions of the gbqs with @xmath123 for all the possible model parameter regions , we map the problem to find the possible motions of a newtonian classical quasi - particle moving in some possible potentials . after considering all the possible motions of the classical quasi - particle in the possible potentials related to the maxima of the potential at @xmath88 ,
all the possible travelling solitary wave solutions related to zero boundary conditions are found for all the possible velocities and the model parameter ranges with @xmath123 .
similar to the integrable cases such as the kdv equation , the solitary waves at some special ranges , the travelling solitary waves with zero boundary conditions are completely prohibited . for the kdv equation ,
all the left moving and static solitons with zero boundary conditions are prohibited . for the gbqs with @xmath123 , there are three different cases . for the first case , @xmath152 , both the faster and slower
moving solitary waves are allowed while the solitary waves moving in the middle " velocities @xmath153 are completely prohibited . for the second case , @xmath154 ,
only the faster moving solitary waves are allowed while all the slower moving solitary waves @xmath155 are prohibited . for the third case @xmath156 ,
contrary to the second case , all the slower moving solitary waves are allowed while all the faster moving @xmath141 solitary waves are completely prohibited .
for the first two cases the solitary waves at the critical velocities @xmath157 are prohibited while for the third case , there are algebraic solitary waves at the critical velocities @xmath114 . for the first two types of gbqs ,
the right moving solitary waves possess bell shape and the left moving solitary waves possess ring shape . for the third type of gbqs ,
the right moving bell shape solitary wave possess larger amplitudes than those of the right moving ring shape solitary waves ( with the same value of the velocity ) while for the left moving solitary waves , the ring shape solitary waves possess larger amplitude .
the work was supported by the national outstanding youth foundation of china ( no . 19925522 ) , the national natural science foundation of china ( no . 90203001 ) , the research fund for the doctoral program of higher education of china ( grant . no .
2000024832 ) , the natural science foundation of zhejiang province of china and the arc large grant a00103139 of the university of new south wales university . c. s. gardner , j. m. green , m. d. kruskal , r. m. miura , sl phys .
lett . , 19:1095(1967 )
t. y. wu , and j. e. zhang , on modelling nonlinear long waves : in _ mathematics is for solving problem : a volume in honor of julian cole on his 70th birthday _
l. p. cook , v. roytburd and m. tulin .
sima , 233 - 249 ( 1996 ) d. j. kaup , _ prog .
54:396(1975 ) b. a. kupershmidt , _ comm .
phys . _ 99:51(1985 ) r. conte , m. musette and a. pickering , j. phys .
27 : 2831 ( 1994 ) ; m. musette and r. conte , j. phys .
a : math . gen .
27 : 3895 ( 1994 ) ; m. musette , r. conte and a. pickering , j. phys . a : math
28 : 179 ( 1995 ) ; a. pickering , j. math .
phys . 37 : 1894 ( 1996 ) ; c. l. chen and s. y. lou , chaos , solitons and fractals , 16 : 27 ( 2003 )
. j. e. zhang , and y. s. li , bidirectional solitons on water , in press j. weiss , m. tabor , g. carnevale , _ j. math
_ 24 : 522(1983 ) .
s - y lou , phys . rev .
80 : 5027 ; s - y lou ans j - j xu , j. math .
39 : 5364 ( 1998 ) .
s - y lou , z. naturforsch .
53a : 251 ( 1998 ) .
j. weiss , nato .
b 278 : 225 ( 1992 ) ; s - y lou , phys . lett .
a176 : 96 ( 1993 ) ; x - y tang and h - c hu , chin . phys .
19 : 1225 ( 2002 ) .
r. rajaraman , _ solitons and instantons , ( north - holland , amsterdam , 1982 ) , pp .
y - j zhu and s - y lou , _ commun . theor .
_ , 23 : 83 ( 1995 ) ; s - y lou and g - j ni , _ j. math . phys . _
30 : 1614 ( 1989 )
. j. m. dye and a. parker , _
_ 43 : 4921 ( 2002 ) .
|
the possible solitary wave solutions for a general boussinesq ( gbq ) type fluid model are studied analytically . after proving the non - painlev integrability of the model , the first type of exact explicit travelling solitary wave with a special velocity selection is found by the truncated painlev expansion .
the general solitary waves with different travelling velocities can be studied by casting the problems to the newtonian quasi - particles moving in some proper one dimensional potential fields . for some special velocity selections ,
the solitary waves possess different shapes , say , the left moving solitary waves may possess different shapes and/or amplitudes with those of the right moving solitons .
for some other velocities , the solitary waves are completely prohibited .
there are three types of gbq systems ( gbqss ) according to the different selections of the model parameters .
for the first type of gbqs , both the faster moving and lower moving solitary waves allowed but the solitary waves withmiddle " velocities are prohibit . for the second type of gbqs
all the slower moving solitary waves are completely prohibit while for the third type of gbqs only the slower moving solitary waves are allowed .
.2 in = 0.65 cm [ section ] [ section ] [ section ] [ section ]
|
the theory of gromov - witten invariants has been greatly influenced by its interactions with string theory .
in particular , the mirror symmetry principle has led to completely unexpected predictions concerning gw - invariants of calabi - yau manifolds .
the original prediction of @xcite for the genus - zero gw - invariants of a quintic threefold was verified about ten years ago in a variety of ways in @xcite , @xcite , @xcite , @xcite , and @xcite .
the 1993 prediction of @xcite for the genus - one gw - invariants of a quintic threefold is verified in @xcite , using the results of this paper . + the proof of the genus - zero mirror symmetry for a projective hypersurface @xmath0 essentially consists of computing certain equivariant integrals on moduli spaces @xmath1 of stable degree-@xmath2 maps from genus - zero curves with @xmath3 marked points into @xmath4 .
while the integrals appearing in chapters 29 and 30 of @xcite are over @xmath5 , the integrands involve only one marked point . for this reason , such integrals can be easily expressed in terms of integrals on @xmath6 and determine genus - zero one - point gw - invariants of @xmath0 ; see _ refz1ptdfn_e1 below . in this paper
we compute integrals on @xmath5 with integrands involving both marked points .
these integrals in a sense correspond to arithmetic genus one and indeed constitute a significant portion of the proof of mirror symmetry for genus-_one _ gw - invariants in @xcite .
theorem [ main_thm ] also provides closed mirror formulas for genus - zero _ two_-point gw - invariants of @xmath0 , including with descendants . at the end of subsection [ outline_subs ]
, we describe the issue arising for integrals with more marked points , a potential way of addressing it , and a reconstruction algorithm for genus - zero gw - invariants of @xmath0 with descendants .
+ let @xmath7 be the universal curve over @xmath1 , with structure map @xmath8 and evaluation map @xmath9 : @xmath10^{\pi } \ar[r]^{\ev } & \p^{n-1 } \\
\ov\m_{0,m}(\p^{n-1},d).}\ ] ] in other words , the fiber of @xmath8 over @xmath11 $ ] is the curve @xmath12 with @xmath3 marked points , while @xmath13\big)=f(z ) \qquad\hbox{if}\quad
z\!\in\!\cc.\ ] ] if @xmath14 is a positive integer , the sheaf @xmath15 is locally free .
we denote the corresponding vector bundle by @xmath16 . over a point @xmath11\!\in\!\ov\m_{0,m}(\p^{n-1},d)$ ] is @xmath17 . ]
its euler class , @xmath18 , relates genus - zero gw - invariants of a degree-@xmath14 hypersurface in @xmath4 to genus - zero gw - invariants of @xmath4 ; see section 26.1 in @xcite . for each @xmath19 , there is a well - defined bundle map @xmath20\big)=\big[\xi(x_i(\cc))\big],\ ] ] where @xmath21 is the @xmath22th marked of @xmath12 .
since it is surjective , its kernel is again a vector bundle .
let @xmath23 whenever @xmath24 and @xmath25 , respectively . and
@xmath26 in _ refz1ptdfn_e1 are switched ; the analogues of @xmath16 and @xmath27 over @xmath28 are denoted by @xmath29 and @xmath30 , respectively . ]
+ the standard action of the @xmath31-torus @xmath32 on @xmath4 induces @xmath32-actions on @xmath1 , @xmath7 , @xmath16 , @xmath27 , and @xmath33 ; see subsections [ equivcoh_subs ] and [ local_subs1 ] for details on equivariant cohomology .
in particular , @xmath16 , @xmath27 , and @xmath33 have well - defined equivariant euler classes @xmath34 these classes are related by @xmath35 where @xmath36 is the equivariant hyperplane class . for each @xmath37
, there is also a well - defined equivariant @xmath38-class , @xmath39 the first chern of the vertical cotangent line bundle of @xmath7 pull - backed to @xmath1 by the section @xmath40\lra\big[\cc , f;x_i(\cc)\big].\ ] ] since @xmath1 is a smooth stack ( orbifold ) , there is an integration - along - the - fiber homomorphism @xmath41.\ ] ] for each @xmath42 , let @xmath43\big/(x\!-\!\al_1)\ldots(x\!-\!\al_n)\ ] ] be the equivariant poincare dual of the @xmath22th fixed point @xmath44 .
let @xmath45 denote the field of fractions in @xmath46 . for @xmath47 , an explicit algebraic formula for @xmath48\big)\big[\big[u\big]\big ] \end{split}\ ] ]
is confirmed in chapters 29 and 30 of @xcite .
the equality in _
refz1ptdfn_e1 is a straightforward consequence of the string relation for gw - invariants ; see section 26.3 in @xcite .
+ one of the ingredients in genus-_one _ localization computations is a two - pointed version of _ refz1ptdfn_e1 : @xmath49\big]\big)\big[\big[u\big]\big ] .
\end{split}\ ] ] note that the term of degree zero in @xmath50 above is symmetric in @xmath51 and @xmath52 , just as are the positive - degree terms . in turn
, @xmath53 can be determined from the power series @xmath54\big)\big[\big[u\big]\big]\ ] ] with @xmath55 .
the seemingly unfortunate choice of indexing is partly motivated by the central role played by the power series @xmath56 defined in _ refz1ptdfn_e1 and the simple relation @xmath57 which follows from _ refeulerclass_e , along with _ refabothm_e , _ refphidfn_e , and _ refrestrmap_e . as shown in this paper , @xmath58 where @xmath59 is the @xmath60th elementary symmetric polynomial in @xmath46 ; see theorem [ main_thm ] .
+ _ remark 1 : _ the right - hand side of _ refz2ptform_e is in fact symmetric in @xmath51 and @xmath52 , because @xmath61 the reason for this relation is explained in subsection [ outline_subs ] .
+ _ remark 2 : _ we will see in subsection [ mainthm_subs ] that the power series @xmath62 can be represented by elements of @xmath63 $ ] . the relation _ refz2ptform_e might then suggest that the corresponding element of @xmath64 $ ] representing @xmath53 has a simple pole at @xmath65 .
in fact , there is no pole at @xmath65 , except in degree zero .
this is immediate from the localization formula _ refabothm_e ; see also subsection [ local_subs1 ] .
+ the power series _ refz2wptdfn_e1 encode genus - zero two - point gw - invariants of a degree-@xmath14 hypersurface in @xmath4 with constraints coming from @xmath4 .
thus , theorem [ main_thm ] provides mirror formulas for such invariants ; the coefficients @xmath66 are `` purely equivariant '' and are irrelevant for this purpose . in the table
below , we give the first ten genus - zero two - point bps numbers , defined from gw - invariants by equation ( 2 ) in @xcite , for the degree-@xmath67 hypersurface in @xmath68 .
these numbers are integers as predicted by conjecture 1 in @xcite . in fact
, we have used the first statement of theorem [ main_thm ] , along with a computer program , to confirm this conjecture for all degree-@xmath2 two - point bps counts in a degree-@xmath31 hypersurface @xmath69 in @xmath4 with @xmath70 and @xmath71 .
+ [ cols="<,<",options="header " , ] + the explicit expressions of section 29.1 in @xcite for the power series @xmath56 have very different forms for @xmath72 and @xmath73 .
the @xmath73 case is the most interesting one and corresponds to calabi - yau hypersurfaces . as the power series @xmath56 are central to our computation of @xmath74 and @xmath53 , for the purposes of the explicit expressions preceding theorem [ main_thm ] in the next subsection we consider only the case @xmath73 . for the rest of the paper .
however , the statements of lemmas [ recgen_lmm ] and [ polgen_lmm ] and their proofs are valid for all @xmath14 .
therefore , the proofs of _ refz2ptform_e and _ refsymmrel_e are valid as well .
the same is the case with theorem [ main_thm ] if the power series @xmath75 and @xmath76 are chosen appropriately .
] this is also the case used in @xcite .
the essence of mirror symmetric predictions for gromov - witten invariants is that these invariants ( and relatedly @xmath56 ) can be expressed in terms of certain hypergeometric series . in this subsection , we define these series and then express @xmath56 , @xmath62 , and @xmath53 in terms of them
. + let @xmath31 be a positive integer .
for each @xmath77 , define @xmath78 by @xmath79 each @xmath78 is a degree-@xmath80 polynomial in @xmath81 with coefficients that are power series in @xmath82 .
for example , @xmath83 + for @xmath84 with @xmath85 , let @xmath86 it is straightforward to check that each of the `` diagonal '' terms @xmath87 is a power series in @xmath82 with constant term @xmath88 , whenever it is defined ; see @xcite , for example .
thus , the division in _ reftden_e is well - defined for all @xmath60 .
let @xmath89 by _ refhgexamp_e , the map @xmath90 is a change of variables ; it will be called the mirror map .
if @xmath91 and @xmath92 is a power series in @xmath82 with coefficients that are functions of a complex variable @xmath93 and possibly some other variable @xmath94 , let @xmath95 + we define @xmath96\big]\big/\prod_{k=1}^{k = n}(x\!-\!\al_k).\ ] ] expanding @xmath97 as a power series in @xmath98 , we obtain @xmath99 where @xmath100 is a degree-@xmath101 symmetric polynomial in @xmath46 with coefficients in @xmath102[[e^t]]$ ] .
for example , @xmath103 the main conclusion of section 30.4 in @xcite is that the power series @xmath104 defined in _
refz1ptdfn_e1 is the evaluation of @xmath105\big]\big/\prod_{k=1}^{k = n}(x\!-\!\al_k)\ ] ] at @xmath106 , if @xmath107 and @xmath81 are related by the mirror map _ refmirmap_e .
+ the power series @xmath62 and @xmath53 defined in _ refz2wptdfn_e1 and _ refz2ptdfn_e1 , respectively , are also evaluations of certain power series @xmath108\big ] \big/\prod_{k=1}^{k = n}(x\!-\!\al_k ) \qquad\hbox{and}\\ & \wt\cz(\hb_1,\hb_2,x_1,x_2,u)\in \q_{\al}(\hb_1,\hb_2,x_1,x_2)\big[\big[u\big]\big]\big/ \prod_{k=1}^{k = n}(x_1\!-\!\al_k)(x_2\!-\!\al_k ) \end{split}\ ] ] that have a mirror transform shape analogous to _ refz1ms_e .
let @xmath109\big]\big/\prod_{k=1}^{k = n}(x\!-\!\al_k).\ ] ] [ main_thm ] there exist @xmath110\big]$ ] , with @xmath111 and @xmath112 , such that the coefficients of the powers of @xmath82 in @xmath66 are degree-@xmath101 symmetric polynomials and the power series defined in _ refz2wptdfn_e1 are given by @xmath113 if @xmath107 and @xmath81 are related by the mirror map _ refmirmap_e .
furthermore , the power series defined in _ refz2ptdfn_e1 are given by @xmath114 + _ remark : _ we note that by _ refmain_e1 and _ refmain_e2 @xmath115 in other words , @xmath116 is the same transform of @xmath117 in both @xmath118 and @xmath119 as @xmath120 and @xmath121 are of @xmath76 and @xmath122 in @xmath123 .
+ the only relevant property of the power series @xmath66 for the purposes of the genus - one localization computations in @xcite is that the @xmath82-coefficients of @xmath66 lie in the ideal generated by @xmath124 if @xmath125 .
this is automatic in the case of theorem [ main_thm ] , since each of these coefficients is a symmetric polynomial in @xmath46 of a positive degree @xmath126 .
the approach of @xcite suggests that these coefficients are likely to be irrelevant in many other localization computations as well .
nevertheless , they are described inductively in this paper in the process of proving the first statement of theorem [ main_thm ] .
the proof of _ refz1ms_e in chapter 30 of @xcite essentially consists of showing that 1 .
@xmath127 and @xmath128 satisfy a certain recursion on the @xmath50-degree ; 2 .
@xmath127 and @xmath128 satisfy a certain self - polynomiality condition ( spc ) ; 3 .
[ equalmods_item ] the two sides of _ refz1ms_e , viewed as a powers series in @xmath98 , agree mod @xmath129 ; 4 .
if @xmath130 satisfies the recursion and the spc , so do certain transforms of @xmath130 ; 5 .
[ uniqs_item ] if @xmath130 satisfies the recursion and the spc , it is determined by its `` mod @xmath129 part '' . for the purposes of these statements , in particular [ equalmods_item ] and [ uniqs_item ]
, we assume that @xmath131\big]\big/\prod_{k=1}^{k = n}(x\!-\!\al_k).\ ] ] for example , [ uniqs_item ] means @xmath132 + the proof of _ refmain_e1 in this paper essentially consists of showing that 1 .
[ recm_item ] @xmath133 and @xmath134 satisfy the recursion _ refrecur_dfn_e ; 2 .
[ polm_item ] @xmath135 and @xmath136 satisfy the mutual polynomiality condition ( mpc ) of lemma [ phistr_lmm2 ] ; 3 .
[ equalmodm_item ] the two sides of _ refmain_e1 , viewed as a powers series in @xmath98 , agree mod @xmath98 ; 4 .
[ transm_item ] if @xmath137 satisfies _ refrecur_dfn_e and the mpc with respect to @xmath130 , the transforms of @xmath137 of lemma [ phistr_lmm4 ] satisfy _ refrecur_dfn_e and the mpc with respect to appropriate transforms of @xmath130 ; 5 . [ uniqm_item ] if the @xmath50-constant term of @xmath130 is independent of @xmath138 and nonzero and @xmath137 satisfies _ refrecur_dfn_e and the mpc with respect to @xmath130 , it is determined by its `` mod @xmath98 part '' ; see proposition [ uniqueness_prp ] .
the statements [ equalmodm_item ] and [ uniqm_item ] should be interpreted analogously to [ equalmods_item ] and [ uniqs_item ] . in other words ,
the equalities are modulo @xmath139 , or equivalently after the evaluation at each @xmath106 .
similarly , the requirement on the degree - zero term in @xmath130 in [ uniqm_item ] means that it is nonzero even after the evaluation at each @xmath106 .
+ the claims of [ recm_item ] and [ polm_item ] concerning @xmath134 are special cases of lemmas [ recgen_lmm ] and [ polgen_lmm ] below , since @xmath140 by the string relation ; see section 26.3 in @xcite .
[ recgen_lmm ] for all @xmath141 , @xmath142 , and @xmath143 , the power series @xmath144 defined by @xmath145satisfies the recursion _ refrecur_dfn_e . [ polgen_lmm ] for all @xmath141 , @xmath142 , and @xmath143 , the power series @xmath146 , with @xmath144 as in lemma [ recgen_lmm ] , satisfies the polynomial condition of lemma [ phistr_lmm1 ] with respect to @xmath128 .
our proof of lemma [ recgen_lmm ] is practically identical to the proof in section 30.1 of @xcite that @xmath128 satisfies a certain recursion on the @xmath50-degree . in _
refrecur_dfn_e are `` shifts by one '' of the coefficients in the recursion for @xmath128 . ]
the proof of lemma [ polgen_lmm ] is similar to the proof in section 30.2 of @xcite that @xmath128 satisfies the spc .
however , there are differences in how the key idea for the setup used in @xcite is utilized .
an explanation of the modifications and a complete justification of their appropriateness are not very simple to state . in order to avoid any confusion
, we thus give a full account in subsections [ local_subs2 ] and [ phizstrpf_subs ] .
as it requires most of what constitutes the proof of the recursivity relation _ refrecur_dfn_e , we give a proof of the latter in subsection [ local_subs1 ] .
+ we are now able to justify _ refsymmrel_e . by _ refz2wptdfn_e1 , @xmath147 along with [ uniqm_item ] ,
these three properties of @xmath134 imply _ refsymmrel_e , since @xmath148 the last expression above vanishes at @xmath106 for all @xmath42 .
+ we will check by a direct computation that @xmath149 satisfies _ refrecur_dfn_e and the mpc with respect to @xmath127 ; see subsection [ algcomp_subs ] . by
_ refym1dfn_e , @xmath150 thus , the @xmath151 case of _ refmain_e1 follows from _ refz2wmod_e , [ transm_item ] , and [ uniqm_item ] .
+ we could also verify directly that @xmath152 satisfies _ refrecur_dfn_e and the mpc with respect to @xmath127 .
fortunately , this is an immediate consequence of parts [ deriv_ch ] and [ mult_ch ] of lemma [ phistr_lmm4 ] , since @xmath153 thus , @xmath152 satisfies the two properties because @xmath154 does . on the other hand , by _ refyexpand_e and _ refyexpand_e2 , @xmath155 thus , the @xmath156 case of _ refmain_e1 also follows from _ refz2wmod_e , [ transm_item ] , and [ uniqm_item ] .
+ the differentiation transform [ deriv_ch ] of lemma [ phistr_lmm4 ] is the only one of the five `` admissible '' transforms that has no analogue in chapter 30 of @xcite .
, i.e. transform [ hbmult_ch ] , is not explicitly mentioned in chapter 30 of @xcite , but its admissibility is nearly immediate from the relevant definitions . ]
the admissibility of this transform , along with that of [ mult_ch ] of lemma [ phistr_lmm4 ] , implies that @xmath133 defined by the second equation in _ refmain_e1 satisfies the recursion _ refrecur_dfn_e and the mpc of lemma [ phistr_lmm2 ] with respect to @xmath127 , no matter what the coefficients @xmath157 are . in light of _ refz2wmod_e , the @xmath158 cases of the first equation in _
refmain_e1 thus reduce to showing there exist @xmath157 such that @xmath159 this is proved by induction , using _ refyexpand_e and _ refyexpand_e2 ; see subsection [ algcomp_subs ] .
+ the proof of _
refmain_e2 follows the same principle . by the string relation and _
refeulerclass_e , @xmath160 thus , by lemmas [ recgen_lmm ] and [ polgen_lmm ] , @xmath161 satisfies the recursion _ refrecur_dfn_e and the mpc of lemma [ phistr_lmm2 ] with respect to @xmath128 for @xmath162 and @xmath163 fixed . by symmetry
, it also satisfies the two properties for @xmath164 and @xmath165 fixed .
it is then sufficient to compare the two sides of _ refmain_e2 multiplied by @xmath166 modulo @xmath167 : @xmath168 in order to see that the two resulting expressions are equal , we compare them modulo @xmath169 : @xmath170 from this we conclude that the two sides of _ refmain_e2 multiplied by @xmath166 modulo @xmath167 are equal .
therefore , the two sides _ refmain_e2 are equal by [ uniqm_item ] . + central to this paper are the use of the transforms @xmath171 in conjunction with part [ deriv_ch ] of lemma [ phistr_lmm4 ] and a desymmetrization of the approach of chapter 30 of @xcite to obtain an explicit closed formula for the integrals in _ refz2ptdfn_e1 . the transforms @xmath171 , combined with the transforms [ mult_ch ] and [ hbmult_ch ] of lemma [ phistr_lmm4 ] , make it possible to construct a power series @xmath172 , satisfying the recursion _ refrecur_dfn_e and the mpc with respect to @xmath127 , that agrees with a pre - specified @xmath173-symmetric element of @xmath174\big[\big[u\big]\big]$ ] modulo @xmath98 . on the other hand , for the purposes of [ uniqm_item ] , it is sufficient to assume that the coefficient of each power of @xmath50 in @xmath175 and @xmath176 is a sum of a power series in @xmath98 and a polynomial in @xmath138 .. in fact , the condition on the middle term of the recursion used chapter 30 of @xcite can also be relaxed as in definition [ recur_dfn ] . ] using [ uniqm_item ] and the last two parts of lemma [ phistr_lmm4 ] , a variety of integrals on @xmath177 involving @xmath178 and products of @xmath179 can be reduced to integrals involving @xmath178 and powers @xmath38-classes , with each exponent bounded by @xmath180 .
+ for example , suppose one would like to compute @xmath181 these integrals may be useful for localization computations in ( arithmetic ) genus two . by lemmas
[ recgen_lmm ] and [ polgen_lmm ] and part [ hbmult_ch ] of lemma [ phistr_lmm4 ] , @xmath182 satisfies the recursion and the mpc with respect to @xmath120 for each @xmath183 .
thus , @xmath184 can be reconstructed , one obtains @xmath185 ] from the `` mod @xmath186 '' part of @xmath182 , @xmath187 once one computes @xmath188 similarly , @xmath189 can be reconstructed from the its `` mod @xmath169 '' part , if one can compute @xmath190 unfortunately , the `` mod @xmath167 part '' of @xmath191 is not simple . +
the above approach does , nevertheless , lead to a reconstruction theorem for @xmath178-twisted gw - invariants of @xmath4 , or equivalently for gw - invariants of projective hypersurfaces .
the theorem arising here is different from @xcite and @xcite for example , as the reduction is made to gw - invariants with low powers of @xmath38-classes and without increasing the number of marked points .
furthermore , in may be possible to get a handle on the `` components '' of the `` mod @xmath167 part '' of @xmath192 , i.e. @xmath193 and @xmath192 itself through the approach of subsections [ bcov1-zreg_subs ] and [ bcov1-localprppf_subs ] in @xcite .
along with _ refrecur_dfn_e , this approach leads to an explicit , but complicated , recursion for @xmath192 or ( the components of its `` mod @xmath167 part '' ) .
this subsection describes the extent of rigidity of power series with coefficients in rational functions that satisfy a certain recursion and a polynomiality condition .
it is the analogue of section 30.3 .
+ denote by @xmath194 the set of nonnegative integers and by @xmath195 $ ] , whenever @xmath196 , the set of positive number not exceeding @xmath31 : @xmath197=\big\{1,2,\ldots , n\big\}.\ ] ] whenever @xmath198 is a function of @xmath199 ( and possibly of other variables ) which is holomorphic at @xmath200 ( for a dense subspace of the other variables ) and @xmath201 , let @xmath202 this is a function of the other variables if there are any .
+ let @xmath203 be any field extension of @xmath204 , possibly @xmath204 itself .
given @xmath205\big],\ ] ] we define @xmath206\big ] \qquad\hbox{by}\notag\\ \label{phidfn_e } \phi_{y , z}(\hb , u , z)= \sum_{i=1}^{i = n}\frac{e^{\al_iz}}{\prod_{k\neq i}(\al_i\!-\!\al_k ) } y\big(\hb,\al_i , ue^{\hb z}\big)z(-\hb,\al_i , u).\end{gathered}\ ] ] [ phistr_lmm1 ] if @xmath207 $ ] , there exists a unique collection @xmath208\ ] ] such that the @xmath209-degree of @xmath210 is at most @xmath211 for every @xmath212 and @xmath213 where each path integral is taken over a simple closed loop in @xmath214 enclosing all points @xmath215 with @xmath216 and @xmath217 .
the equality holds for a dense collection of complex parameters @xmath138 .
_ proof : _ it can be assumed that @xmath218 , ~r , r'\!\in\!\bar\z^+ , ~(r , k)\!\neq\!(r',k').\ ] ] note that for every @xmath219 and @xmath220 , @xmath221 where @xmath222 by the residue theorem , @xmath223 on the other hand , since @xmath207 $ ] , @xmath224 for unique @xmath225 . by _ refphidfn_e and _ refphistr_lmm1e3 , @xmath226 by _ refphistr_lmm1e1b and _ refphistr_lmm1e5 , _
refphistr_lmm1e is satisfied if and only if @xmath227 for a dense collection of complex parameters @xmath138 , there exists a unique polynomial @xmath228\ ] ] of @xmath209-degree at most @xmath211 that satisfies _ refphistr_lmm1e7 .
[ phistr_lmm2 ] if @xmath207 $ ] and @xmath229 $ ] are as in lemma [ phistr_lmm1 ] , then @xmath230\big[\big[u , z\big]\big ] \quad&\llra\quad e_{y , z;d}\in\ti\q_{\al}[\hb,\om]~~\forall\ , d\!\in\!\bar\z^+\\ & \llra\quad e_{z , y;d}\in\ti\q_{\al}[\hb,\om]~~\forall\ , d\!\in\!\bar\z^+ \quad\llra\quad \phi_{z , y}\in \ti\q_{\al}[\hb]\big[\big[u , z\big]\big ] .
\end{split}\ ] ] + _ proof : _ the equivalence of the two middle statements in _ refphistr_lmm2e follows from _ refphistr_lmm1e7 , which implies that @xmath231 on the other hand , by the residue theorem on @xmath232 , @xmath233 where @xmath234 $ ] is given by @xmath235 the path integral in _
refphistr_lmm2e1 is again taken over a simple closed loop enclosing all points @xmath215 with @xmath236 .
write @xmath237 by
_ refphistr_lmm1e , _ refphistr_lmm2e1 , and _ refphistr_lmm2e3 , @xmath238 since @xmath234
$ ] , it follows that @xmath239 $ ] if @xmath240 $ ] for all @xmath241 .
conversely , since @xmath242 , @xmath243 \qquad\lra\qquad f_{d,(d+1)n-1},\ldots , f_{d,0}\in \ti\q_{\al}[\hb].\ ] ] these observations imply the two remaining statements of lemma [ phistr_lmm2 ] .
[ recur_dfn ] for @xmath244 $ ] with @xmath245 and @xmath246 , let @xmath247 we will call @xmath248 $ ] @xmath249-recursive if @xmath250 for every
@xmath251 $ ] and for some @xmath252 and @xmath253 .. ] _ remark 1 : _
suppose @xmath259 for some @xmath260 . in the statement of proposition [ uniqueness_prp ] and throughout the rest of the paper , @xmath261 i.e. we drop @xmath262 and higher powers of @xmath98 , instead of higher powers of @xmath138 .
+ _ remark 2 : _ in contrast to the situation in ( * ? ? ?
* chapter 30 ) , the assumptions of proposition [ uniqueness_prp ] , i.e. recursivity and @xmath138-polynomiality with respect to @xmath175 , are both linear conditions on @xmath176 . consequently , the @xmath98-term of @xmath263 is no longer necessary to determine @xmath176 .
+ _ proof : _ suppose @xmath264 and we have shown that @xmath265 with notation as in _ refrecur_dfn_e and by the last assumption on @xmath263 , it follows that @xmath266 if @xmath267 are as in the proof of lemma [ phistr_lmm1 ] , @xmath268 by _
refphistr_lmm1e3 , _
refpolyn_prp_e1 , and _ refpolyn_prp_e2 . since @xmath269 by _ refphistr_lmm1e7 and _ refpolyn_prp_e3 and @xmath270 $ ] by lemma [ phistr_lmm2 ] , @xmath271 for some @xmath272 $ ] .
thus , @xmath273 for some @xmath274 $ ] . on the other hand , by _ refphistr_lmm1e7 and
_ refpolyn_prp_e3 @xmath275 by _ refpolyn_prp_e5 and _ refpolyn_prp_e6 , @xmath276 along with _ refpolyn_prp_e2 , this implies that _ refpolyn_prp_e1 holds with @xmath2 replaced by @xmath277 .
this subsection is the analogue of the beginning of section 30.4 in @xcite .
we describe transforms of @xmath278 that preserve the polynomiality property of lemma [ phistr_lmm2 ] and the recursivity property of definition [ recur_dfn ] .
the statement of lemma [ phistr_lmm4 ] is followed by complete proofs .
the first of the five transforms below has no analogue in @xcite .
[ phistr_lmm4 ] suppose @xmath279 $ ] are such that @xmath176 is @xmath249-recursive and @xmath255[[u , z]]$ ] .
then , 1 . [ deriv_ch ] if @xmath280 , @xmath281 is @xmath249-recursive and @xmath282[[u , z]]$ ] ; 2 .
[ mult_ch ] if @xmath283 $ ] , then @xmath284 is @xmath249-recursive and @xmath285[[u , z]]$ ] ; 3 .
[ hbmult_ch ] if @xmath286 $ ] , then @xmath287 is @xmath249-recursive and @xmath282[[u , z]]$ ] ; 4 .
[ multoverh_ch ] if @xmath283 $ ] , @xmath288 , and @xmath289 , then @xmath290 is @xmath249-recursive and @xmath291[[u , z]]$ ] ; 5 .
[ transf_ch ] if @xmath292 $ ] , @xmath293 , @xmath294 then @xmath295 is @xmath249-recursive and @xmath291[[u , z]]$ ] .
+ _ remark : _ in fact , [ mult_ch ] and [ hbmult_ch ] are special cases of the admissible transform defined by @xmath296[[u]]$ ]
. + [ deriv_ch ] the operator @xmath297 preserves the structure of the first term on the right - hand side of _ refrecur_dfn_e .
the @xmath298-summand in the last term becomes @xmath299 since the last term in _ refderiv_ch_e1 does not depend on @xmath138 and @xmath176 is @xmath249-recursive , it follows that @xmath295 is also @xmath249-recursive . since @xmath300 and @xmath255[[u , z]]$ ] , @xmath282[[u , z]]$ ] by the middle equivalence in lemma [ phistr_lmm2 ] . +
[ mult_ch ] since @xmath176 is @xmath249-recursive and the multiplication by @xmath198 preserves the structure of each of the terms in _ refrecur_dfn_e , @xmath284 is also @xmath249-recursive . since @xmath301 and @xmath255[[u , z]]$ ] , @xmath302[[u , z]]$ ] .
similarly , since @xmath303 , @xmath304[[u , z]]$ ] by lemma [ phistr_lmm2 ] .
+ [ hbmult_ch ] it is sufficient to assume that @xmath305 .
the multiplication by @xmath138 preserves the structure of the first term on the right - hand side of _ refrecur_dfn_e .
the @xmath298-summand in the last term becomes @xmath306 since the last term in _ refhbmult_ch_e1 does not depend on @xmath138 and @xmath176 is @xmath249-recursive , it follows that @xmath295 is also @xmath249-recursive . since @xmath307 and @xmath255[[u , z]]$ ] , @xmath282[[u , z]]$ ] . + [ multoverh_ch ] since @xmath288 , i.e. @xmath198 contains no degree-0 term in @xmath50 , the multiplication by @xmath308 preserves the structure of the first term on the right - hand side of _ refrecur_dfn_e .
the @xmath298-summand in the last term becomes @xmath309 since @xmath176 is @xmath249-recursive and @xmath310\big[\big[u\big]\big],\ ] ] it follows that @xmath295 is @xmath249-recursive as well . on the other hand , @xmath311 since @xmath255[[u , z]]$ ] and @xmath312\big[\big[u , z\big]\big],\ ] ] _ refmultoverh_ch_e2 implies that @xmath291[[u , z]]$ ] as well .
+ [ transf_ch ] since @xmath293 , the operation of replacing @xmath50 with @xmath313 followed by multiplication by @xmath314 preserves the structure of the first term on the right - hand side of _ refrecur_dfn_e .
the @xmath298-summand in the last term becomes @xmath315 since @xmath176 is @xmath249-recursive and @xmath316\big[\big[u\big]\big],\ ] ] it follows that @xmath295 is @xmath249-recursive as well . on the other hand , @xmath317\big[\big[u\big]\big ] .
\end{split}\ ] ] since @xmath255[[u , z]]$ ] , _
reftransf_ch_e2 implies that @xmath291[[u , z]]$ ] as well . in this subsection
we verify the three claim concerning the power series @xmath133 made in subsection [ outline_subs ] : 1 .
[ ymprec_item ] @xmath154 satisfies the @xmath249-recursivity condition of definition [ recur_dfn ] ; 2 .
[ ympphi_item ] @xmath318[[u , z]]$ ] ; 3 .
[ ypcontr_item ] there exist @xmath66 as in theorem [ main_thm ] such that _ refypcond_e is satisfied .
+ _ proof of [ ymprec_item ] : _ note that @xmath319 thus , by the residue theorem on @xmath232 , @xmath320 on the other hand , if @xmath321 is the degree-@xmath2 term of @xmath75 , @xmath322,\\ \res_{z=\i}\bigg\{\frac{1}{\hb\!-\!z}\cy_{-1}(z,\al_i , u)\bigg\}&=1,\end{aligned}\ ] ] where @xmath323 is as in _ refderivdfn_e .
refrecurpf_e1 implies that @xmath75 satisfies the recursion _ refrecur_dfn_e .
+ _ proof of [ ympphi_item ] : _ let @xmath324 by [ mult_ch ] of lemma [ phistr_lmm4 ] , it is sufficient to show that @xmath325[[u , z]]$ ] . note that @xmath326 thus , by the residue theorem on @xmath232 , @xmath327 the @xmath60th summand above is polynomial in @xmath138 .
thus , @xmath325[[u , z]]$ ] . + _ proof of [ ypcontr_item ] : _ suppose @xmath328 , and we have constructed power series @xmath329\big],\ ] ] satisfying the second equation in _ refmain_e1 and _ refypcond_e , so that @xmath330 where @xmath331 is a degree-@xmath101 symmetric polynomial in @xmath46 with coefficients in @xmath102[[e^t]]$ ] such that @xmath332 this assumption is satisfied for @xmath333 by _ refyexpand_e and _ refyexpand_e2 . since @xmath334 $ ] , @xmath335,\quad c_{p-1,1}^{(r)}(t)\in \q_{\al}[[e^t ] ] ~~\forall\,r\!\ge\!1 \qquad\lra\notag\\ \label{psstr_e } \frac{d}{dt}c_{p-1,1}^{(r)}(t)\in \q_{\al}[[e^t ] ] ~~\forall\,r\!\ge\!0.\end{gathered}\ ] ] thus , @xmath336\big ] .
\end{split}\ ] ] by _ reftden_e , _
refyexpand_e5 , and _
refyexpand_e6 are satisfied with @xmath60 replaced by @xmath337 and @xmath338 in particular , @xmath122 satisfies _ refypcond_e . by _ refyexpand_e8 , the coefficients @xmath339
are inductively defined by @xmath340 thus , by _ refpsstr_e and inductive assumptions , @xmath341 $ ] is a degree-@xmath101 symmetric polynomial in @xmath46 with coefficients in @xmath342 $ ] , as required .
in subsection [ local_subs1 ] , we apply the classical localization theorem _ refabothm_e with the standard action of the @xmath31-torus @xmath32 on @xmath4 and on @xmath1 to verify lemma [ recgen_lmm ] . in subsection
[ local_subs2 ] , we apply _ refabothm_e with an action of the @xmath343-torus @xmath344 on @xmath345 and on a subspace of @xmath346 to verify lemma [ polgen_lmm ] . the aim of this subsection is to review the basics of equivariant cohomology and to set notation . throughout this subsection
, @xmath347 denotes an @xmath3-torus , either @xmath348 or @xmath349 . +
the @xmath3-torus @xmath347 acts freely on @xmath350 ( or @xmath351 ) : @xmath352 thus , the classifying space for @xmath347 and its group cohomology are given by @xmath353,\ ] ] where @xmath354 if @xmath355 are the projection onto the @xmath22th component and the tautological line bundle , respectively .
denote by @xmath356 the field of fractions of @xmath357 : @xmath358 + a representation @xmath359 of @xmath347 , i.e. a linear action of @xmath347 on @xmath360 , induces a vector bundle over @xmath361 : @xmath362 if @xmath359 is one - dimensional , we will call @xmath363 the weight of @xmath359 . for example , @xmath364 is the weight of representation @xmath365 more generally , if a representation @xmath359 of @xmath347 on @xmath360 splits into one - dimensional representations with weights @xmath366 , we will call @xmath366 the weights of @xmath359 . in such a case ,
@xmath367 we will call the representation @xmath359 of @xmath32 on @xmath368 with weights @xmath46 the standard representation of @xmath32 . + if @xmath347 acts on a topological space @xmath369 , let @xmath370 be the equivariant cohomology of @xmath369 . the projection map @xmath371 induces an action of @xmath357 on @xmath372 .
let @xmath373 if the @xmath347-action on @xmath369 lifts to an action on a ( complex ) vector bundle @xmath374 , then @xmath375 is a vector bundle over @xmath376 .
let @xmath377 denote the equivariant euler class of @xmath378 .
+ the standard action of @xmath32 on @xmath4 is the action induced by the standard action @xmath359 of @xmath32 on @xmath368 : @xmath379 = \big[e^{\i\th_1}z_1,\ldots , e^{\i\th_n}z_n\big].\ ] ] since @xmath380 , @xmath381 \big/\big(x^n\!+\!c_1(v_{\rho})x^{n-1}\!+\!\ldots\!+\!c_n(v_{\rho})\big),\ ] ] where @xmath382 and @xmath383 is the tautological line bundle . since @xmath384 it follows that @xmath385\big/(x\!-\!\al_1)\ldots(x\!-\!\al_n ) \qquad\hbox{and}\\ \h_{\t}^*(\p^{n-1 } ) & = \q_{\al}[x]\big/(x\!-\!\al_1)\ldots(x\!-\!\al_n ) .
\end{split}\ ] ] + the standard action of @xmath32 on @xmath4 has @xmath31-fixed points : @xmath386 , \qquad p_2=[0,1,0,\ldots,0 ] , \quad\ldots\quad p_n=[0,\ldots,0,1].\ ] ] for each @xmath42 , let @xmath387 by equation _ refphiprop_e below , @xmath388 is the equivariant poincare dual of @xmath389 .
we also note that @xmath390 , where @xmath391 is as in _ refifactorrep_e .
thus , the restriction map on the equivariant cohomology induced by the inclusion @xmath392 is given by @xmath393\big/\prod_{k=1}^{k = n}(x\!-\!\al_k ) \lra h_{\t}^*(p_i)=\q[\al_1,\ldots,\al_n ] , \qquad x\lra\al_i.\ ] ] by _ refrestrmap_e , @xmath394 + the tautological line bundle @xmath395 is a subbundle of @xmath396 preserved by the diagonal action of @xmath32 .
thus , the action of @xmath32 on @xmath4 naturally lifts to an action on @xmath397 and @xmath398 via the exact sequence @xmath399 of vector bundles on @xmath4 , @xmath32 also lifts to an action on @xmath400 . by _ refweightspord_e and _ refgarestr_e , @xmath401 + if @xmath347 acts smoothly on a smooth compact oriented manifold @xmath369 , there is a well - defined integration - along - the - fiber homomorphism @xmath402 for the fiber bundle @xmath371
. the classical localization theorem of @xcite relates it to integration along the fixed locus of the @xmath347-action .
the latter is a union of smooth compact orientable manifolds @xmath403 and @xmath347 acts on the normal bundle @xmath404 of each @xmath403 .
once an orientation of @xmath403 is chosen , there is a well - defined integration - along - the - fiber homomorphism @xmath405 the localization theorem states that @xmath406 where the sum is taken over all components @xmath403 of the fixed locus of @xmath347 .
part of the statement of _ refabothm_e is that @xmath407 is invertible in @xmath408 . in the case of the standard action of @xmath32 on @xmath4 , _
refabothm_e implies that @xmath409 see also _
+ finally , if @xmath410 is a @xmath347-equivariant map between two compact oriented manifolds , there is a well - defined pushforward homomorphism @xmath411 it is characterized by the property that @xmath412 the homomorphism @xmath413 of the previous paragraph corresponds to @xmath414 being a point .
it is immediate from _ refpushdfn_e that @xmath415 the standard @xmath32-action on @xmath4 ( as well as any other action ) induces @xmath32-actions on moduli spaces of stable maps @xmath1 by composition on the right : @xmath416=[\cc , g\circ f ] \qquad\forall~g\in\t , \,[\cc , f]\in\ov\m_{0,m}(\p^{n-1},d).\ ] ] all evaluation maps @xmath417\lra f(y_i),\ ] ] are @xmath32-equivariant . via the natural lift of the @xmath32-action to @xmath418
described in subsection [ equivcoh_subs ] , the @xmath32-action on @xmath1 lifts to @xmath32-actions on the vector bundles @xmath16 , @xmath16 , and @xmath33 , as well as on the universal tangent line bundles . + as described in detail in ( * ? ? ?
* section 27.3 ) , the fixed loci of the @xmath32-action on @xmath1 are indexed by decorated graphs that have no loops .
a graph consists of a set @xmath419 of vertices and a collection @xmath420 of edges , i.e. of two - element subsets of @xmath419 .
a loop in a graph @xmath421 is a subset of @xmath420 of the form @xmath422 neither of the three graphs in figure [ loopgraph_fig ] has a loop .
a decorated graph is a tuple @xmath423 where @xmath421 is a graph and @xmath424 , \qquad \d\ ! :
\edg\lra\z^+ , \quad\hbox{and}\quad \eta\!:[m]\lra\ver\ ] ] are maps such that @xmath425 in figure [ loopgraph_fig ] , the values of the map @xmath426 on some of the vertices are indicated by letters next to those vertices . similarly , the value of the map @xmath427 on one of the edges is indicated by a letter next to the edge .
the three elements of the set @xmath428\!=\![3]$ ] are shown in bold face .
they are linked by line segments to their images under @xmath429 . by
_ refdecorgraphcond_e , no two consecutive vertex labels are the same and thus @xmath430 . + ( -.7,-1.5)(10,2 ) ( 2.5,0)(6,0)(4.2,.7 ) ( 2.5,0).2(2.5,.7 ) ( 6,0).2(6,.75 ) ( 2.5,0)(1,1.5)(1,2 ) ( 6,0)(7.5,1.5)(8,1.5 ) ( 6,0)(7.5,-1.5)(8,-1.5 ) ( 4.2,-1.5)@xmath431 ( 11.5,0)(15,0)(13.2,.7 ) ( 11.5,0).2(11.5,.7 ) ( 15,0).2(15,.75 ) ( 11.5,0)(10,1.5)(10,2 ) ( 15,0)(17,2)(17,2).2 ( 17,2)(20,3)(20,3).2 ( 17,2)(20,1)(20,1).2 ( 15,0)(17,-2)(17,-2).2 ( 17,-2)(18.5,-.5)(19,-.5 ) ( 17,-2)(18.5,-3.5)(19,-3.5 ) ( 26.5,0)(30,0)(30,0).2 ( 26.5,0).2(26.5,.7 ) ( 26.5,0)(25,1.5)(25,2 ) ( 30,0)(32,2)(32,2).2 ( 32,2)(35,3)(35,3).2 ( 32,2)(35,1)(35,1).2 ( 30,0)(32,-2)(32,-2).2 ( 32,-2)(33.5,-.5)(34,-.5 ) ( 26.5,0)(24.5,-2)(24.5,-2).2 ( 26.5,0)(28.5,-2)(28.5,-2).2 ( 26.5,0)(28.5,2)(28.5,2).2 ( 28.5,2)(31.5,3)(31.5,3).2 ( 32,-2)(33.5,-3.5)(34,-3.5 ) the fixed locus @xmath432 of @xmath177 corresponding to a decorated graph @xmath433 consists of the stable maps @xmath198 from a genus - zero nodal curve @xmath434 with @xmath435 marked points into @xmath4 that satisfy the following conditions . the components of @xmath434 on which the map @xmath198 is not constant are rational and correspond to the edges of @xmath433 .
furthermore , if @xmath436 is an edge , the restriction of @xmath198 to the component @xmath437 corresponding to @xmath438 is a degree-@xmath439 cover of the line @xmath440 passing through the fixed points @xmath441 and @xmath442 .
the map @xmath443 is ramified only over @xmath441 and @xmath442 .
in particular , @xmath444 is unique up to isomorphism .
the remaining , contracted , components of @xmath434 are indexed by the vertices @xmath445 such that @xmath446\ ! :
\eta(i)\!=\!v\}\big| \ge 3.\ ] ] the map @xmath198 takes such a component @xmath447 to the fixed point @xmath448 .
thus , @xmath449 where @xmath450 denotes the moduli space of stable genus - zero curves with @xmath451 marked points . for the purposes of this definition , @xmath452 and @xmath453
are one - point spaces .
for example , in the case of the last diagram in figure [ loopgraph_fig ] , @xmath454 is a fixed locus in @xmath455 for some @xmath456 .
+ we now verify lemma [ recgen_lmm ] .
let @xmath457 suppose @xmath433 is a decorated graph as in _ refdecortgraphdfn_e that contributes to _ refrecgen_e , in the sense of the localization formula _ refabothm_e . by _ refphidfn_e and _ refrestrmap_e , @xmath458 where @xmath459 is the kronecker delta function .
thus , by _ refabothm_e , @xmath433 does not contribute to _
refz2wptdfn_e1 unless @xmath460 , i.e. the marked point of the map is taken to the point @xmath44 .
there are two types of graphs that do ( or may ) contribute to _ refz2wptdfn_e1 ; they will be called @xmath461 and @xmath462-types . in a graph of the @xmath461-type
, the marked point @xmath88 is attached to a vertex @xmath463 of valence two which is labeled @xmath22 . in a graph of the @xmath462-type
, the marked point @xmath88 is attached to a vertex @xmath464 of valence at @xmath465 , which is still labeled @xmath22 .
examples of the two types are depicted in figure [ loopgraph_fig ] .
+ suppose @xmath433 is a graph of type @xmath462 and @xmath466 so that @xmath433 contributes to the coefficient of @xmath467 in _ refz2wptdfn_e1 . in this case , the restriction of @xmath468 to @xmath432 is the pull - back of a @xmath38-class from the component @xmath469 in the decomposition _ refzlocus_e .
since the @xmath32-action on the corresponding tautological line bundle is trivial , @xmath470 thus , @xmath433 contributes a polynomial in @xmath98 , of degree at most @xmath471 , to the coefficient of @xmath467 in _ refz2wptdfn_e1 .
therefore , the contributions of the loci of type @xmath462 to _ refz2wptdfn_e1 are accounted for by the middle term in _ refrecur_dfn_e . +
a graph @xmath433 as in _ refdecortgraphdfn_e of type @xmath461 has a unique vertex @xmath472 joined to @xmath464 .
denote by @xmath473 the set of all graphs @xmath433 of type @xmath461 such that @xmath474 and @xmath475 , i.e. the unique vertex @xmath472 of @xmath433 joined to @xmath464 is mapped to @xmath476 and the edge @xmath477 corresponds to the @xmath478-fold cover of @xmath479 branched only over @xmath389 and @xmath480 . by
_ refdecorgraphcond_e , @xmath481 + suppose @xmath482 and @xmath472 is the unique vertex joined to @xmath464 by an edge .
we break at @xmath472 into two graphs : 1 .
@xmath483 consisting of the vertices @xmath464 and @xmath472 , the edge @xmath477 , and marked points @xmath88 and @xmath26 attached to @xmath464 and @xmath472 , respectively ; 2 .
@xmath484 consisting all vertices and edges of @xmath433 , other than the vertex @xmath464 and the edge @xmath477 , with a new marked point attached to @xmath472 ; see figure [ splitgraph_fig ] .
let @xmath485 denote the degree of @xmath484 , i.e. the sum of all edge labels . by _ refzlocus_e , @xmath486 denote by @xmath487 and @xmath488 the two component projection maps .
+ ( -1.2,-1.5)(10,2 ) ( 2.5,0)(6,0)(4.2,.7 ) ( 2.5,0).2(2.5,.7 ) ( 6,0).2(6,.75 ) ( 2.5,0)(1,1.5)(1,2 ) ( 6,0)(8,2)(8,2).2 ( 8,2)(11,3)(11,3).2 ( 8,2)(11,1)(11,1).2 ( 6,0)(8,-2)(8,-2).2 ( 8,-2)(9.5,-.5)(10,-.5 ) ( 8,-2)(9.5,-3.5)(10,-3.5 ) ( 5,-2)@xmath433 ( 17.5,0)(21,0)(19.2,.7 ) ( 17.5,0).2(17.5,.7 ) ( 21,0).2(21,.75 ) ( 17.5,0)(16,1.5)(16,2 ) ( 21,0)(22.5,1.5)(22.5,2) ( 19,-1.5)@xmath483 ( 27,0).2(27,.75 ) ( 27,0)(25.5,1.5)(25.5,2 ) ( 27,0)(29,2)(29,2).2 ( 29,2)(32,3)(32,3).2 ( 29,2)(32,1)(32,1).2 ( 27,0)(29,-2)(29,-2).2 ( 29,-2)(30.5,-.5)(31,-.5 ) ( 29,-2)(30.5,-3.5)(31,-3.5 ) ( 26,-1.5)@xmath484 by section 27.4 in @xcite , @xmath489 where @xmath490 and @xmath491 are the tautological tangent line bundles .
thus , by _ reftangrestr_e , @xmath492 by sections 27.1 and 27.2 in @xcite , on @xmath493 @xmath494 thus , using _ reftangrestr_e and taking into the account the automorphism group , @xmath495 , we obtain @xmath496 by _ refzlocus_e2 , _
refbndlsplit_e3 , and _ refbndlsplit_e5 , the contribution of @xmath433 to _ refrecgen_e is @xmath497 + we next sum _ refbndlsplit_e6 over @xmath482 .
this is the same as summing the expression in the curly brackets over all @xmath3-pointed graphs with the marked point @xmath88 attached to a vertex @xmath472 labeled @xmath498 , i.e. all graphs of types @xmath499 and @xmath500 . by the localization formula _ refabothm_e ,
the sum of the terms in the curly brackets over all such graphs @xmath484 is @xmath501 .
thus , @xmath502 we conclude that @xmath144 is @xmath249-recursive in the sense of definition [ recur_dfn ] : * the middle term in _ refrecur_dfn_e consists of the contributions from the graphs of type @xmath462 ; * the @xmath503-summand in _ refrecur_dfn_e consists of the contributions from the graphs of type @xmath473 .
in this subsection we deduce lemma [ polgen_lmm ] from lemma [ phizstr_lmm ] , which is proved in the next subsection . the argument , in this subsection and the next one , is a modification on the proof of self - polynomiality of @xmath120 in section 30.2 of @xcite .
+ we will denote the weight of the standard action of the one - torus @xmath504 on @xmath214 by @xmath138 .
thus , by subsection [ equivcoh_subs ] , @xmath505 , \quad h_{\wt\t}^*\approx\q[\hb,\al_1,\ldots,\al_n ] \qquad\lra\qquad \h_{\wt\t}^*\approx\q_{\al}(\hb).\ ] ] throughout this subsection , @xmath506 will denote the representation of @xmath504 with the weights @xmath507 and @xmath508 .
the induced action on @xmath509 has two fixed points : @xmath510 , \qquad q_2\equiv[0,1].\ ] ] let @xmath511 be the tautological line bundle . then , @xmath512 + for each @xmath513 , the action of @xmath514 on @xmath515 induces an action on @xmath516 it has @xmath517 fixed points : @xmath518 , \qquad i\in[n],~r\in\{0\}\!\cup\![d],\ ] ] if @xmath519 are the standard coordinates on @xmath378 and @xmath520 is the @xmath22th coordinate vector ( so that @xmath521\!=\!p_i\!\in\!\p^{n-1}$ ] ) .
let @xmath522 denote the equivariant hyperplane class .
+ for all @xmath251 $ ] and @xmath523 $ ] , @xmath524 since @xmath525 the @xmath514-equivariant cohomology of @xmath526 is given by @xmath527\big/ \prod_{s=0}^{s = d}\prod_{k=1}^{k = n}\big(\om-(\al_k\!+\!s\hb)\big)\\ & \approx \q\big[\om,\hb,\al_1,\ldots,\al\big]\big/ \prod_{s=0}^{s = d}\prod_{k=1}^{k = n}\big(\om-\al_k-\!s\hb\big)\\ & \subset \q_{\al}[\hb,\om]\big/ \prod_{s=0}^{s = d}\prod_{k=1}^{k = n}\big(\om-\al_s - r\hb\big ) .
\end{split}\ ] ] + there is a natural @xmath514-equivariant morphism @xmath528 a general element of @xmath529 of @xmath530 determines a map @xmath531 up to an automorphism of the domain @xmath532 .
thus , the map @xmath533 is well - defined and determines an element @xmath534 .
the map @xmath535 extends continuously over the boundary of @xmath530 .
we denote the restriction of @xmath535 to the smooth substack @xmath536 of @xmath537 by @xmath538 , or simply by @xmath539 whenever there is no ambiguity .
+ let @xmath540 be the natural projection map .
similarly section 30.2 in @xcite , this lemma implies that @xmath544\big[\big[u , z\big]\big]\ ] ] for the following reason . with @xmath545 as in _ refetabedfn_e , by _
refxdcohom_e @xmath546 for some @xmath547 $ ] of @xmath209-degree at most @xmath211 .
therefore , by lemma [ phizstr_lmm ] , _ refpushprop_e , _ refabothm_e , and _ refxrestr_e , @xmath548 in the last expression ,
the integral has the same meaning as in lemma [ phistr_lmm1 ] .
we have thus shown that @xmath541 is polynomial with respect to @xmath120 , assuming lemma [ phizstr_lmm ] . in this subsection
we use the localization formula _ refabothm_e to prove lemma [ phizstr_lmm ] .
we show that each fixed locus of the @xmath514-action on @xmath549 contributing to the right - hand side of _
refphizstr_e corresponds to a pair @xmath550 of a graphs , with @xmath551 and @xmath552 contributing to @xmath56 and @xmath553 , respectively , for some @xmath251 $ ] .
+ similarly to subsection [ local_subs1 ] , the fixed loci of the @xmath514-action on @xmath554 correspond to decorated graphs @xmath433 with @xmath3 marked points and no loops .
each edge should be labeled by a pair of integers , indicating the degrees of the corresponding maps in @xmath509 and in @xmath4 .
each vertex should be labeled either @xmath555 or @xmath556 for some @xmath557 $ ] , indicating the fixed point , @xmath558 or @xmath559 , to which the vertex is mapped .
no two consecutive vertex labels are the same , but if two consecutive vertex labels differ in the @xmath435th component , with @xmath560 , the @xmath435th component of the label for edge connecting them must be nonzero .
+ ( -.5,-1.2)(10,1.8 ) ( 17,0)(22,0 ) ( 17,0).2(17,.7 ) ( 22,0).2(22,.7 ) ( 17,0)(14,1)(14,1).2 ( 15.5,1)(14,1.7 ) ( 14,1)(11.5,2.5)(11.5,2.5).2 ( 12.8,2.3)(10.9,2.5 ) ( 14,1)(11.5,-.5)(11.5,-.5).2 ( 12.3,.6)(10.9,-.5 ) ( 17,0)(14.5,-1.5)(14.5,-1.5).2 ( 16.2,-1.1)(14.5,-.8 ) ( 14.5,-1.5)(12,-2)(11.5,-2 ) ( 22,0)(25,1)(25,1).2 ( 23.5,1)(25,1.7 ) ( 25,1)(27,2)(27.5,2 ) ( 25,1)(27.5,-.5)(27.5,-.5).2 ( 26.7,.6)(28.1,-.5 ) ( 22,0)(24.5,-1.5)(24.5,-1.5).2 ( 22.9,-1.1)(24.5,-.8 ) ( 24.5,-1.5)(27,-2)(27.5,-2 ) the situation for the @xmath514-action on @xmath561 is simpler , however .
there is a unique edge of positive @xmath509-degree .
we draw it as a thick horizontal line .
the first component of all other edge labels must be @xmath507 ; so we drop it .
the first components of the vertex labels change only when the thick edge is crossed .
thus , we drop the first components of the vertex labels as well , with the convention that these components are @xmath88 on the left side of the thick edge and @xmath26 on the right . in particular , the vertices to the left of the thick edge ( including the left endpoint ) lie in @xmath562 and the vertices to its right lie in @xmath563 . thus , by _ refxdfn_e , the marked point @xmath88 is attached to a vertex to the left of the thick edge and the marked point @xmath26 is attached to a vertex to the right .
furthermore , by the first identity in _ refhbaract_e , such a graph will not contribute to the right - hand side of _ refphizstr_e unless the remaining marked points are also attached to vertices to the right of the thick edge . finally , both vertices of the thick edge have the same ( remaining , second ) label @xmath251 $ ] .
let @xmath564 denote the set of graphs as above so that the two endpoints of the thick edge are labeled @xmath22 ; see figure [ xlocus_fig ] .
+ if @xmath565 , we break it into three sub - graphs : 1 .
@xmath551 consisting of all vertices and edges of @xmath433 to the left of the thick edge , including its left vertex @xmath566 , and a new marked point @xmath26 attached to @xmath566 ; 2 .
@xmath483 consisting of the thick edge , its two vertices @xmath566 and @xmath567 , and new marked points @xmath88 and @xmath26 attached to @xmath566 and @xmath567 , respectively ; 3 .
@xmath552 consisting of all vertices and edges of @xmath433 to the right of the thick edge , including its right vertex @xmath567 , and a new marked point @xmath88 attached to @xmath567 ; see figure [ xsplit_fig ] .
the fixed locus in @xmath549 corresponding to @xmath433 is then @xmath568 the middle term is a single point .
let @xmath569 , @xmath487 , and @xmath570 denote the three component projection maps .
denote by @xmath571 and @xmath572 be the degrees of @xmath551 and @xmath552 , i.e. @xmath573 the exceptional case for the first statement is @xmath574 , in which case the corresponding moduli space does not exist .
+ ( -.5,-1.2)(10,1.8 ) ( 10,0).2(10,.7 ) ( 10,0)(7,1)(7,1).2 ( 8.5,1)(7,1.7 ) ( 7,1)(4.5,2.5)(4.5,2.5).2 ( 7.8,2.3)(3.9,2.5 ) ( 7,1)(4.5,-.5)(4.5,-.5).2 ( 5.3,.6)(3.9,-.5 ) ( 10,0)(7.5,-1.5)(7.5,-1.5).2 ( 9.2,-1.1)(7.5,-.8 ) ( 7.5,-1.5)(5,-2)(4.5,-2 ) ( 10,0)(11.5,1.5)(11.5,2 ) ( 17,0)(15.5,1.5)(15.5,2 ) ( 22,0)(23.5,1.5)(23.5,2 ) ( 17,0)(22,0 ) ( 17,0).2(17,.7 ) ( 22,0).2(22,.7 ) ( 29,0)(27.5,1.5)(27.5,2 ) ( 29,0).2(29,.7 ) ( 29,0)(32,1)(32,1).2 ( 30.5,1)(32,1.7 ) ( 32,1)(34,2)(34.5,2 ) ( 32,1)(34.5,-.5)(34.5,-.5).2 ( 33.7,.6)(35.1,-.5 ) ( 29,0)(31.5,-1.5)(31.5,-1.5).2 ( 29.9,-1.1)(31.5,-.8 ) ( 31.5,-1.5)(34,-2)(34.5,-2 ) suppose @xmath565 , @xmath571 and @xmath572 are as above , and @xmath575 . similarly to _ refbndlsplit_e1 , @xmath576 where @xmath577 is the normal bundle of @xmath432 in @xmath549 and @xmath578 , @xmath579 , and @xmath580 are the tautological tangent line bundles .
we note that @xmath581 thus , by _ refxbndlsplit_e1 , _
reftangrestr_e , and _ refhbaract_e , @xmath582 the map @xmath539 takes the locus @xmath432 to a fixed point @xmath583 .
it is immediate that @xmath584 . by continuity considerations , @xmath585 .
thus , by the first identity in _
refxrestr_e , @xmath586 combining _ refxlocus_e1 and _ refxbndlsplit_e3 with this observation , we obtain @xmath587 we note that this identity remains valid for @xmath574 if we set the term in the first curly brackets to @xmath88 for @xmath574 . + we now sum up _
refxbndlsplit_e5 , multiplied by @xmath588 , over all @xmath565 .
this is the same as summing over all pairs @xmath550 of graphs such that 1 .
@xmath551 is a 2-pointed graph of a degree @xmath589 such that the marked point @xmath26 is attached to the vertex labeled @xmath22 ; 2 .
@xmath552 is an @xmath3-pointed graph of a degree @xmath590 such that the marked point @xmath88 is attached to the vertex labeled @xmath22 .
|
in this paper we compute certain two - point integrals over a moduli space of stable maps into projective space .
computation of one - point analogues of these integrals constitutes a proof of mirror symmetry for genus - zero one - point gromov - witten invariants of projective hypersurfaces .
the integrals computed in this paper constitute a significant portion in the proof of mirror symmetry for genus-_one _ gw - invariants completed in a separate paper .
these integrals also provide explicit mirror formulas for genus - zero _ two_-point gw - invariants of projective hypersurfaces .
the approach described in this paper leads to a reconstruction algorithm for all genus - zero gw - invariants of projective hypersurfaces .
|
the atlas and cms experiments at the large hadron collider ( lhc ) have recently reported discovery of a new particle consistent with the higgs boson of the standard model @xcite .
each experiment independently reports nearly @xmath4 combined local significance for a mass of @xmath5 , largely fueled by the searches for resonances in diphoton and @xmath6 final states . the discovery of the higgs marks the final chapter in the quest to discover the particle content of the standard model .
it also represents the start of a new era of higgs physics at the lhc . with additional data and further analyses
the experiments will be able to study the properties of this new particle , in particular its production modes and decay branching ratios .
the standard model makes very definite predictions for these properties ( up to calculational uncertainties ) given the mass of the higgs , so any significant deviation from the expected values would be an immediate sign of new physics .
there are many ways that physics beyond the standard model ( bsm ) can manifest itself in higgs events .
one possibility is that the higgs couplings to known sm particles are modified by new physics , such as loop effects of new particles , mixing of the higgs with other new states ( extended higgs sectors ) , and higgs compositeness .
such effects will modify both the decay branching ratios and production modes of the higgs , potentially giving observable deviations from sm predictions .
several works perform general analyses of these possibilities in the context of recent lhc higgs results @xcite .
a second possibility is that on - shell production of bsm particles can lead to higgs production in the decay cascade .
this possibility has been discussed as a discovery mode for an sm - like higgs boson in several models @xcite , for the extra higgs states of the minimal supersymmetric standard model ( mssm ) and its extensions @xcite , and for composite higgs models @xcite .
conversely , the discovery sensitivity for bsm physics through higgs production has been studied in the context of the mssm with a bino lsp @xcite , a gravitino lsp @xcite , and additional feebly coupled lsps @xcite , as well as in the context of vector - like heavy quarks @xcite .
higgs production in cascade decays has also been studied as a way to probe the details of the the spectra and couplings of bsm particles after their discovery @xcite . in this work
we consider a class of susy models with neutralino nlsps which dominantly decay to a higgs boson .
in particular we focus on models with light third generation squarks , which are motivated by naturalness of the electroweak scale .
such models can remain unconstrained by existing susy searches while still having total susy production cross sections at the 7 or 8 tev lhc of up to a few picobarns comparable to the total sm higgs production cross section and significantly larger than the cross section for associated production in the sm
. the sensitivity of susy searches is limited due to the suppression of missing energy when a light nlsp ( @xmath7 ) decays to a higgs and the lsp , as has been explored previously in more general models @xcite .
the dominant decays for a light higgs are @xmath8 and @xmath9 , and much of the existing work on bsm higgs production has focused on these modes .
the @xmath10 mode suggests search channels with many @xmath1 s or searches for ` higgs - tagged ' boosted jets @xcite , while the @xmath9 decay suggests searches in leptonic channels with missing energy .
the many-@xmath1 modes have large qcd backgrounds when the met signature is suppressed , while the leptonic modes have smaller branching fractions and can have significant backgrounds as well .
furthermore , it can be difficult to link a new physics signal with the higgs in these channels , especially in the presence of other hard objects from the susy cascade .
we will focus instead on the potential for observing such models using the rare but distinctive diphoton decay mode of the higgs , which has a branching fraction of 0.22% for a 125 gev sm higgs .
we show that searches requiring a diphoton with invariant mass near 125 gev plus other objects such as @xmath1-tagged jets , missing energy or leptons can target the models we consider while having very low sm backgrounds ( a similar point is made in the context of searching for heavy vector - like quarks in ref .
since such searches look for a narrow resonance in the diphoton mass , the backgrounds can be directly determined from data and the observation of a peak near 125 gev would strongly suggest new physics involving production of an on - shell higgs .
we find that in the @xmath11 tev lhc run the @xmath3 channel could give a strong signal in regions of parameter space where other searches are not sensitive , and that the @xmath12 and @xmath13 channels have sensitivity comparable to and in some regions exceeding that of existing susy searches .
these results are summarized in fig .
[ fig : combinedlimits ] .
although targeted @xmath14 searches have the best potential to probe our models , for sufficiently light squarks even inclusive searches for the sm higgs boson in the resonant @xmath15 and @xmath16 channels can have some sensitivity .
enhancements of 20 - 50% over the sm rates can be accommodated within current bounds from other searches .
searches in non - resonant channels , such as @xmath17 , @xmath18 and @xmath19 , generally include additional cuts that tend to reject susy events for the models we consider , so no excess rate is predicted for these analyses .
our models therefore predict a very particular pattern of excesses in certain higgs searches which can be observable with sufficient data .
such excesses can be consistent with the currently observed rates in the atlas and cms higgs searches , which report a higher than expected rate in the diphoton channels .
we find that in the regions of parameter space where the excess rate in sm higgs searches is significant , strong signals are expected in the @xmath20 modes within the 8 tev run .
we now give an outline of the rest of this work . in sec .
[ sec : models ] we describe models in which a neutralino nlsp can decay dominantly to the higgs and in which the third generation squarks are light and give the dominant susy production cross section . in sec .
[ sec : constraints ] we consider the limits on these models from existing lhc susy searches .
we then consider the signal in the higgs to diphoton channels ( sec .
[ sec : diphoton ] ) , determining the sensitivity of existing searches for @xmath13 and @xmath12 , and suggesting a new resonant search for diphotons in association with a @xmath1-tagged jet .
the possibility that current hints of excesses in the inclusive higgs diphoton channel can be explained within this model is discussed in sec .
[ sec : othersearch ] .
we also briefly discuss the signals from direct electroweak production of neutralinos ( sec .
[ sec : directprod ] ) .
we then conclude in sec .
[ sec : conclusions ] with an outlook for the rest of the 8 tev run and a discussion of more general motivations for searches for new physics in the higgs + @xmath21 channel .
in this section , we describe a set of models in which neutralino nlsps decay dominantly to the lsp and a higgs
. an important consequence of this mechanism for higgs production is that if the lsp is light and the splitting between the nlsp and the higgs is not too large , the met signature in the susy cascade is suppressed .
this can relax the collider constraints on superpartners , as we will discuss in detail for our model in sec .
[ sec : constraints ] and is explored for gluino and degenerate squark simplified models and the c(n)mssm in refs .
in particular , this can allow for large colored production cross sections with only weak constraints from existing searches . amongst the susy particles the squarks and gluinos typically have the largest production cross sections due to their color charge .
if the squarks are degenerate in mass , existing susy searches strongly constrain their mass in the simplest @xmath22-parity conserving scenarios .
squarks decaying directly to light neutralinos must be heavier than @xmath23tev @xcite .
these bounds are in tension with naturalness of the electroweak scale in susy , as if the stops are heavy then they give large corrections to the higgs masses that must be tuned away .
it is possible however to recover susy naturalness if only the third generation squarks are light , with masses @xmath23 few hundred gev , with the other squarks much heavier so - called natural spectra " @xcite . (
some extension of the mssm is then necessary if the lightest higgs is to have a mass near 125 gev ; a number of possibilities have been explored @xcite . )
this scenario is much less constrained than the case of degenerate squarks since the third - generation squarks have smaller production cross sections and can produce less missing energy in their decays @xcite .
combining these two elements , a natural squark spectrum and a neutralino nlsp decaying dominantly to the higgs , we find a set of models that evade constraints from existing searches but give large cross sections for higgs production , potentially giving striking signatures in searches exploiting the higgs resonance .
the key features of these realistic models can be cast into a simplified model ( section [ sec : simpmodel ] ) which will be our main focus for the rest of this work .
if @xmath22-parity is conserved , the two - body decays of a neutralino nlsp to a neutral lsp can produce an on- or off - shell higgs , @xmath24 , or photon .
in this section we describe some types of models in which a wino , bino , or higgsino nlsp decays dominantly to an on - shell higgs and a neutral lsp , making this channel a promising way to probe such a model at the lhc .
_ states uncharged under the @xmath24 _ + at the renormalizable level , one simple scenario for obtaining large branching ratios of the nlsp to higgses is to have both the lsp and nlsp be formed mostly from states that do not couple to the @xmath24 or the photon ( i.e. states with @xmath25 ) .
this is achieved in the mssm with a neutral wino - like nlsp and a bino - like lsp ( as studied in @xcite ) .
when @xmath26 , the mixing of the nlsp and lsp with the higgsino is small .
decays of the nlsp to the higgs are suppressed by one power of this mixing , while decays to the @xmath24 require mixing of both the nlsp and lsp states and are suppressed by two powers of the mixing .
this is illustrated diagramatically in fig .
[ fig : mixingfeyn ] . from an effective operator viewpoint ,
integrating out the higgsinos generates the dimension five operator @xmath27 , while decays to the @xmath24 will arise only from the dimension six operator @xmath28 ( since the dimension five operator does not involve couplings to the neutral goldstone mode of the higgs ) and are therefore suppressed by additional powers of the higgsino mass .
the dimension five operator @xmath29 can be generated by loops involving sfermions and allows decays to the photon or @xmath24 , but the loop factor suppression renders this negligible .
[ fig : brwtoh ] shows the branching ratio of a 175 gev wino to a higgs and a nearly massless bino as a function of @xmath30 and @xmath31 , for a higgs mass of 125 gev in the decoupling limit .
though large @xmath31 suppresses the wino decay to the higgs , it can be highly dominant even for relatively light higgsinos .
in extensions of the mssm there can be additional states uncharged under the @xmath24 but with renormalizable couplings to the higgs and higgsinos .
this can cause the nlsp to decay dominantly to the higgs by the same mechanism as above .
examples include the singlino @xmath32 of the nmssm @xcite and string photini @xcite
. a singlino lsp mixes through the operator @xmath33 , and so unlike the bino has a decay width to the higgs that increases with large @xmath34 and is suppressed by low @xmath34 .
string photini are superpartners of light @xmath35 gauge fields from string theory @xcite , which can have small mixings with the bino . a string photino lsp therefore couples to the higgs exactly as the bino does but with a coupling suppressed by the wino - photino mixing @xmath36 .
a similar enhancement of the higgs branching ratio can also occur for decays of a wino or bino nlsp to the goldstino of an @xmath22-symmetric susy breaking sector @xcite .
since the lsp can be weakly coupled to the mssm fields in these scenarios , we will consider the limit in which all of the other mssm fields decay promptly to the bino or neutral / charged winos , which then decay to the lsp .
we restrict our attention to the case where the decay to the lsp is not displaced on detector scales ( @xmath37 ) , ensuring accurate reconstruction of collider events and avoiding constraints from searches for metastable particles .
_ higgsino nlsp and gravitino lsp _ + another scenario for obtaining large nlsp branching ratios to the higgs can occur for a higgsino - like nlsp decaying promptly to a gravitino lsp in models with low - scale susy breaking .
@xcite discuss this possibility and some of the collider signatures at the lhc and tevatron .
if @xmath38 , then as @xmath31 approaches unity the coupling of the lightest higgsino to the @xmath24 approaches zero @xcite , due to cancellation of the contributions from the @xmath39 and @xmath40 components .
the higgsino lsp can then decay dominantly to a higgs and gravitino , as illustrated in fig .
[ fig : brhtoh ] . since the gravitino lsp can be very weakly coupled , we will again consider the limit when all other mssm fields decay promptly to the nlsp , which then decays promptly to the gravitino ( requiring @xmath41 ) .
_ spectra _ + in the models discussed above the nlsp is often accompanied by other nearly degenerate states that are important for phenomenology .
we identify three classes of spectra ( illustrated in fig .
[ fig : directspectra ] ) that capture the possible signals : a. * bino nlsp * if the bino is the nlsp , then it can decay dominantly to a higgs and singlino ( or string photino etc . ) lsp provided that @xmath30 is large . if the lsp is very weakly coupled , essentially all susy cascade decays will proceed through the nlsp and produce a higgs .
a pure bino has vanishing pair - production cross section , so this spectrum alone is not observable .
b. * wino co - nlsps * if the nlsp is a neutral wino with small mixings ( large @xmath30 and @xmath42 not too close to @xmath43 ) , then the charged winos are nearly degenerate with the nlsp .
the decay of the charginos to the nlsp is then extremely suppressed ; if this is the dominant decay mode then the charginos have a macroscopic decay length .
we instead consider the case where the chargino decays promptly to the lsp , e.g. @xmath44 . in this scenario
not all susy decays will proceed through the true nlsp ( the neutral wino ) and produce a higgs . c. * quasi - degenerate charginos * two types of models can give the spectrum and decays shown in fig .
[ fig : quasi ] .
one is the higgsino nlsp scenario described in sec .
[ sec : chitohiggs ] . for small
mixing , the neutral higgsino nlsp is accompanied by a chargino and a slightly heavier neutralino .
we will consider the minimal case where the splitting between these states and the nlsp is @xmath45 , so that they decay promptly to nlsp but the extra objects produce in the decay are soft and do not contribute significantly to collider signals .
for a higgsino - like nlsp at @xmath46 gev , this is achieved with @xmath47 .
because the gravitino is very weakly coupled , the other higgsinos will always decay to the nlsp .
+ a second source for this spectrum are models where @xmath48 so that the wino and bino are significantly mixed , while the lsp is an additional state such as a singlino or string photino .
the spectrum then includes a chargino and heavier wino - bino state in addition to the mixed wino - bino nlsp . the splitting between these heavier states and
the nlsp can be large enough to allow prompt decays to the latter ( unlike the pure wino case ) . while the wino - bino mixing is suppressed at the larger values of @xmath30 required for the higgs decay to dominate , sufficient mixing
can still be obtained for quasi - degenerate gauginos with @xmath49 gev .
susy naturalness motivates light stops @xmath50 and @xmath51 , possibly with large mixing , which are generally accompanied by a light left - handed sbottom .
when @xmath52 is small and the stop and sbottom mixing are negligible , the lightest squark is @xmath53 , and the tree - level splitting with @xmath54 is given by m^2_t_l - m^2_b_l m_t^2 + ( 1 - ^2_w ) ( 2 ) m_z^2 [ eq : qlmass ] when @xmath55 , this gives a splitting @xmath56 .
a light right - handed sbottom @xmath57 is not directly motivated by naturalness , but can result from uv realizations of natural spectra .
the decay cascades of the lightest third generation squarks to a neutralino nlsp involve the production of a @xmath58 quark ( on or off shell ) or @xmath1 quark .
when the nlsp decay produces a higgs dominantly , the final states of interest will be @xmath59 .
we find that final states with higgses _ and _ tops are substantially more constrained by existing leptonic susy searches , so we will focus on spectra where the decays of the lightest squark(s ) proceed dominantly to the @xmath60 final state . for each of the three nlsp spectra described in sec .
[ sec : chitohiggs ] , we discuss the signals from the various third - generation squarks : a. * bino nlsp * if the bino is the nlsp and the lsp is very weakly coupled , then the squarks will decay to the bino , producing a @xmath1 quark and , for a stop , a @xmath61 boson .
sbottoms therefore give @xmath62 final states , while stops give @xmath63 .
the latter case is strongly constrained by same - sign dilepton searches ( see sec . [
sec : constraintheavierq ] ) , requiring the stops to be heavier than @xmath64 .
a left - handed sbottom is thus generally required to be @xmath65 , though a light right - handed sbottom @xmath66 can remain to give @xmath62 at observable levels .
b. * wino co - nlsps * if the winos are co - nlsps , the squarks can decay to either the charged or neutral wino .
this decay will be dominant if the lsp is a very weakly coupled state such as a string photino .
if the splitting between the winos and the squarks is not too much larger than the top mass , then decays to bottom quarks dominate , so that sbottoms decay to the neutral wino and stops to the charged wino .
sbottom decays through the winos then give the final state @xmath62 , while stops give @xmath67 .
the latter is constrained by direct stop searches ( sec .
[ sec : constraintheavierq ] ) , requiring the stops to be heavier than @xmath68 .
+ if the lsp is a bino , then the squarks can decay directly to the lsp and a quark
. this will be the dominant decay for right - handed squarks , so higgses will not be produced in this case .
left - handed squarks will have a subdominant branching ratio to the bino ( at least @xmath69 ) , but this is enough for searches in @xmath70 final states to be sensitive , as will be shown in sec .
[ sec : constraintbrtolsp ] .
these searches constrain the left - handed sbottom to be heavier than @xmath68 .
the bino lsp scenario therefore does not have as much potential to give observable signals during the 8 tev run . c. * quasi - degenerate charginos * as before
, we will assume that the splitting between the nlsp and the squarks is small enough that the squarks decay dominantly to bottoms rather than tops .
the sbottoms then decay to neutralinos while stops decay to charginos .
all of these decays eventually produce the nlsp , possibly with additional soft decay products , so all squarks give the @xmath62 final state . from the above discussion
, we can see that the signatures of a third generation squark decaying to the @xmath71 final state can in many cases be well described by the simplified model ( illustrated in fig . [ fig : simpmodel ] ) of a sbottom decaying with 100% branching fraction to a @xmath1 quark and neutral nlsp which decays with 100% branching fraction to the higgs and lsp , giving the final state @xmath72 .
some of the above scenarios give additional final state particles that carry very little energy due to small mass splittings .
these will generally not appear as hard objects in collider events and will only slightly affect the kinematics of the hard @xmath1 s and higgses . therefore the simplified model is an appropriate description of the collider signatures in these models as well .
is considered . ]
the pair production cross sections are essentially the same for all sbottoms and stops ( up to small electroweak effects ) , so this model is a good description for either a light stop or sbottom .
the model is described by two parameters , the sbottom / stop mass and the nlsp mass . for simplicity
, we fix the lsp mass at @xmath73 and consider a 125 gev lightest higgs boson in the decoupling limit , where its couplings are equal to those of the sm higgs .
we also note that several other types of models can give enhanced higgs production with similar final states , i.e. higgses + ( @xmath1-quarks ) + ( low met ) .
for example , in models with additional vector - like generations , decays of the vector - like quarks to a higgs and a top or bottom also produce this final state .
in particular , decays of a new bottom - like quark to @xmath74 give the same event kinematics as this simplified model with nlsp mass equal to the higgs mass , though with different production cross section .
the phenomenology of top - like quarks is explored in @xcite .
@xcite showed that for @xmath22-parity violating susy , higgs bosons can be produced preferentially in the decays between third generation squarks , e.g. @xmath75 , giving the same signatures .
another possibility is new physics enhancing the rare decay @xmath76 , which can be observed in top pair production events @xcite .
other recent work studied the possibility that light colored particles in loops could increase the rate of higgs pair production @xmath77 to make its effects comparable to the sm single higgs production rate @xcite , which would have similar signatures in multilepton and diphoton searches .
we expect that some of our conclusions based on this simplified model will also apply qualitatively to these more general models , as discussed in sec .
[ sec : conclusions ] . in the following sections , we describe in detail the constraints and signals of this simplified model in existing susy searches ( sec .
[ sec : constraints ] ) and existing higgs diphoton searches ( sec .
[ sec : diphoton ] ) .
these results are summarized in fig .
[ fig : combinedlimits ] . in short
, we find that a sbottom heavier than @xmath78 with an nlsp lighter than @xmath79 is not currently excluded . in this region there can be strong signals in a proposed search for resonant @xmath80 ( sec .
[ sec : ggb ] ) , and even observable enhancements in the inclusive higgs diphoton searches ( secs .
[ sec : gg ] and [ sec : othersearch ] ) .
we also consider the constraints on deviations from the simplified model , in particular non - zero branching ratios of the nlsp to the @xmath24 ( sec .
[ sec : constraintbrtoz ] ) , non - zero branching ratio for the direct decay of the squark to the lsp ( sec .
[ sec : constraintbrtolsp ] ) , and decays of the heavier third generation squarks ( sec .
[ sec : constraintheavierq ] ) ; we find that the simplified model is a good description of the dominant lhc signals for a wide range of realistic models .
in this section we determine the constraints on models like those discussed above from existing searches for supersymmetry . to do so ( and to determine signal rates in higgs searches in sec .
[ sec : diphoton ] ) , we perform monte carlo ( mc ) simulation of collider events to determine the efficiencies of various lhc searches for these models .
we use madgraph 5 @xcite to generate susy production events , pythia 6.4 @xcite to decay particles , perform parton showering and hadronization and produce initial and final state radiation , and pgs4 @xcite to simulate object reconstruction in detectors ( we use the pythia - pgs package to interface these tools ) .
we use the default pgs parameter cards for the atlas and cms detectors , but to implement @xmath1-tagging efficiencies we use truth - level @xmath1-jets and apply the @xmath81-dependent @xmath1-tagging efficiencies reported by the atlas and cms experiments for each search .
we scan a grid of points in @xmath82 and @xmath83 spaced by 5 to 15 gev , and generate 50,000 events at each point for the diphoton channels and 250,000 events for the non - diphoton channels to obtain adequate statistics for low rate searches .
we obtain the cross sections for production of third - generation squarks from @xcite and for neutralino / chargino direct production from prospino 2.1 @xcite ; in both cases we use the central value of the prediction and do not account for the theoretical uncertainty ( giving more conservative bounds ) .
we take sm higgs boson cross sections and branching ratios from @xcite .
when available we use the @xmath84 visible cross section limits presented in the searches we study ; when these are not available we compute the @xmath84 limit from the observed event rate and backgrounds as described in @xcite .
where not otherwise noted , we estimate the expected limits in @xmath85 by linearly extrapolating the expected background rate and systematic uncertainty from the current results .
where possible , we have validated our results against quoted limits and generally find agreement within @xmath86 .
nonetheless there is the possibility of significant systematic errors since the detector simulation we use does not fully capture the performance of the atlas and cms detector , in particular in analyses that select for large missing energy and are therefore sensitive to the tails of distributions .
for these reasons we emphasize that the results of this section should not be taken as strict experimental bounds but rather as approximate guidelines as to what regions of parameter space are accessible by various searches .
we first consider constraints on the simplified model of section [ sec : simpmodel ] .
we then discuss the realm of validity of the simplified model , considering further constraints on models with nonzero branching ratio of the nlsp to the @xmath24 and of the lightest squark directly to the lsp , as well as constraints on the decays of the heavier 3rd generation squarks .
_ cms @xmath87 @xmath88 _ + the final states from our simplified model contain at least two @xmath1-quarks , a moderate amount of missing energy , and the decay products of the two higgses .
a 125 gev higgs has a branching ratio to bottom quarks of 58% in the sm , so events can easily contain up to six @xmath1-jets .
standard model processes rarely produce more than two bottom quarks in an event , so searches in final states with three or more @xmath1-jets can place strong limits on new physics . both the atlas and
cms collaborations have performed searches in final states with three or more @xmath1-tagged jets and missing energy @xcite .
the cms search in @xmath89 of 8 tev data places the strongest individual limit ; in the region @xmath90 , @xmath91 events were expected and @xmath92 were observed , for which we compute a 95% @xmath93 limit on the visible cross section of @xmath94 .
the results from our mc simulation for the visible cross section ( total bsm cross section times branching ratio to final state times selection / identification efficiency ) in this cms search channel is shown in fig .
[ fig:3bcmsconstraint ] .
the bound is satisfied for nlsp masses @xmath7 .
the atlas @xmath95 search places weaker bounds and is not shown .
( @xmath96 ) : * visible cross section from our simplified model for the cms search in final states with 3 @xmath1-jets and missing energy @xcite .
the thick dashed contour is the observed 95% limit , and dotted is the expected limit extrapolated to @xmath85 .
, width=288 ] _ cms ssdl + @xmath60 _ + another type of analysis that can place relevant constraints are searches in final states with @xmath1-jets and same - sign leptons , such as those performed by cms in data at both 7 and 8 tev @xcite .
this too is a rare final state in the standard model . in our models ,
same - sign leptons can be produced if both higgses in the event produce a lepton in their decay .
this rate is suppressed by small branching fractions , but unlike the 3@xmath1-jet searches this analysis does not require large met or @xmath97 .
the strongest constraint on our model comes from signal region 2 of the 8 tev search @xcite , which required two @xmath1-tagged jets and two positively charged leptons as well as met @xmath98 and @xmath99 . through an apparent downward fluctuation
, zero events were observed in this channel though 11 were observed in the corresponding channel with two negatively charged leptons .
a 95% cl upper limit of 2.8 expected events from non - sm processes was placed using the @xmath100 method , corresponding to a visible cross section of .71 fb at 8 tev , which we interpret as a limit of 1.42 fb in the combined @xmath101 channel ( which would be populated evenly in our simplified model ) .
the predicted cross sections from our simplified model are shown in fig .
[ fig : ssl2bcms8constraint ] . for an nlsp mass @xmath102
, the squark mass is bounded to be @xmath103 .
we validated our implementation of this search by computing efficiencies for the simplified model of @xmath104 ; the efficiencies we calculated corresponded to limits @xmath105 stronger than reported in ref .
. :* visible cross section from our simplified model for the cuts of signal region 2 of the cms search in final states with 2 @xmath1-jets and same - sign leptons at 8 tev @xcite .
the thick dashed contour is the observed 95% limit , and dotted is the expected limit extrapolated to @xmath85 .
, width=288 ] _ cms multilepton _ + multilepton searches without @xmath1-tags also put a strong constraint on our model , especially when identifying tau pairs , which are produced with branching fraction 6.4% in the decay of a 125 gev higgs .
cms has a performed search in three ( @xmath106 ) and four lepton ( @xmath107 ) channels in @xmath108 of 7 tev data @xcite .
although our models have slightly higher rates in the @xmath106 channels , much lower backgrounds make the @xmath107 channels far more constraining .
the most sensitive @xmath107 channels are those vetoing @xmath24 resonances ( ` no-@xmath24 ' ) , which are populated by our model when both higgses decay either to @xmath109 or @xmath110 . the search channels are further separated by the number of hadronic taus ( @xmath111 ) reconstructed .
the branching ratios to the @xmath112 channels are several times larger than the @xmath113 channel while the background remains low .
the dominant constraint over most of our parameter space comes from the @xmath114 , 2@xmath115 , met @xmath116 50 gev , no-@xmath24 channels . combining the low and high @xmath97 channels , in @xmath108 @xmath117 events
were expected and 1 was observed , from which we compute a @xmath118 @xmath93 limit on the visible cross section of 0.8@xmath119 .
[ fig : multileptoncms7constraint ] shows the visible cross section for our simplified model .
the rate depends only weakly on the nlsp mass and the limit is satisfied for squark masses @xmath120 .
we note that the @xmath113 and @xmath121 regions with the same kinematic selections would also set a strong limit but for an observation of 4 events where @xmath122 was expected
. taken together with the @xmath123 channels , the observed rate would be consistent with a @xmath124 gev squark in our simplified model .
we estimate the expected limit in @xmath85 of 8 tev data ( shown in fig .
[ fig : combinedlimits ] ) by applying the same search cuts to the 8 tev signal while assuming the same background rate as in the 7 tev search , scaled only by the luminosity .
@xcite consider the limits from the same multilepton search on sm @xmath125 production , and ref .
@xcite considers the limits set on @xmath125 production in bsm physics with a focus on the rare decay @xmath76 .
compared to @xmath125 production , the multilepton limits on @xmath126 production are stronger due to the increased rates in @xmath107 channels and the importance of hadronic tau channels ( not considered in refs .
@xcite ) . for comparison ,
our simulated efficiencies for the sm higgs contribution to the @xmath113 channels were about 20% higher than reported in ref .
@xcite , likely due to a more detailed modeling of the cms lepton efficiencies in their study .
limits in the @xmath107 channels are statistics dominated and will likely improve significantly with the full @xmath11 tev run , especially with a properly combined analysis of the sensitive channels .
if limits continue to improve , sm higgs production will in fact be an important background in the most sensitive channels .
the @xmath106 channels may also become sensitive if backgrounds can be reduced by @xmath1-tagging , a strategy also suggested in @xcite for identifying higgs production in the @xmath127 and @xmath76 channels . , 2@xmath128 , met @xmath129 , no-@xmath24 channels in the cms 7 tev multilepton search @xcite .
the thick dashed contour is the observed 95% limit .
, width=288 ] an atlas search in 2.05 @xmath130 for @xmath1-jets and a leptonically decaying @xmath24 @xcite places a constraint on the branching ratio of the nlsp to a @xmath24 in our model
. this search targets pair production of stops in gmsb , with the stop decaying through the neutralino to a z and gravitino , @xmath131 .
this is identical to our simplified model but with 100% branching fraction to a @xmath24 instead of to the higgs .
they quote a limit of @xmath132 gev for an nlsp mass @xmath133 190 gev .
we determine efficiencies for this search by generating events in which one or both nlsps decays to the @xmath24 instead of the higgs .
the visible cross section limits can be translated into a 95% constraint on the branching ratio br(@xmath134 ) , as shown in fig .
[ fig : nlsptozconstraint ] . branching ratios less than these values
are easily achieved in the models described in sec .
[ sec : chitohiggs ] .
nominally the contours of @xmath135 should agree with the quoted experimental exclusions for the gmsb stop model , and we see that our simulation is somewhat overestimating the strength of this search . in the simplified model from the atlas search in final states with @xmath1-jets and a leptonically decaying @xmath24 @xcite .
, width=288 ] if the squark can decay directly to the lsp instead of through the nlsp , then the bounds from searches with met cuts and @xmath1-tags become considerably tighter .
figure [ fig : constraintbrtolsp ] show the 95% upper limit on the branching ratio to this mode from the cms @xmath136 search , obtained by simulating events where one or both squarks decay directly to the lsp instead of through the nlsp .
-jets and missing energy in @xmath89 at 8 tev @xcite .
the gray region is ruled out by this search even for 100% branching fraction to the nlsp ( see fig .
[ fig:3bcmsconstraint ] ) .
the dashed contour is the 95% exclusion for the specific model of a left handed sbottom with a wino nlsp and bino lsp.,width=288 ] the limit on this branching ratio is 1 - 5% in the region of interest . for a weakly coupled lsp
( photino / singlino / goldstino / gravitino ) this constraint is easily satisfied .
however , for a left handed sbottom with a wino nlsp and bino lsp , as described in sec .
[ sec : coloredprod ] , @xmath137 is ruled out . while we have focused on the decays of the lightest squark , there are also constraints on the heavier squarks . when the lightest squark is @xmath138 , the constraints on @xmath139
directly constrain the parameter space of allowed sbottom masses ( see eq .
[ eq : qlmass ] ) , while in the other cases the limits simply set a lower bound on the masses of other squarks in the spectra .
the strongest constraints on the heavier squarks in the spectra occur when their decays give the final state @xmath140 , which can arise for example from the decay of a heavier stop to a bino nlsp , @xmath141 with the @xmath142 decaying to a higgs .
in particular the cms ssdl + 2@xmath1 search , discussed in sec . [
sec : constraintsimpmodel ] , places a much stronger limit on this final state than on the @xmath143 state due to leptonic decays of the @xmath61 s produced from the top decay .
we find that a squark decaying with 100% branching fraction to this mode with an on - shell top is excluded at 95% for masses
@xmath144 by the search in @xmath145 at 8 tev ; in @xmath85 the expected reach is @xmath146 .
direct stop searches have very targeted kinematic requirements and do not significantly constrain the @xmath147 final states . however , in the case of @xmath139 decaying through a chargino co - nlsp to the final state @xmath148 ( as in figure [ fig : winonlsp ] ) , the final state is similar enough to the @xmath149 topology to be constrained .
the strongest limit comes from the lowest met signal region of the atlas 7 tev search for stop pair production with a single leptonic @xmath61 decay in @xmath150 @xcite . when validating our method for the stop decay model in ref .
@xcite , we found the ratio of our efficiencies to those presented to be @xmath151 uniformly over the parameter space presented .
we therefore scaled our efficiencies by a factor of @xmath152 to compensate . for stops near @xmath153 ,
the cross - section changes by a factor of two as the mass is changed by @xmath154 , so we expect our bounds on squark mass to be within this margin of error .
the constraints we obtain are shown in fig .
[ fig : directstoplimits ] in terms of both the @xmath54 mass ( top of frame ) and the corresponding @xmath53 mass ( bottom of frame ) for comparison with the direct constraints on @xmath53 production .
the limit is roughly @xmath155 for an nlsp heavier than 150 gev . for heavier @xmath139
the @xmath140 final state also opens up , but we find that the dominant constraint is still from the direct stop search . in the atlas
single lepton direct stop search @xcite .
@xmath156 and @xmath157 are related by eq .
[ eq : qlmass ] , for which we chose @xmath158 .
the thick dashed contour is the 95% limit from @xmath150 at 7 tev.[fig : directstoplimits],width=288 ]
we now discuss the possibility of observing the simplified model of sec .
[ sec : simpmodel ] in searches taking advantage of the higgs diphoton resonance .
we first describe our methods in the simplest case by considering the inclusive diphoton searches .
we then discuss the potential reach of the existing cms searches targeting sm associated production in @xmath13 and @xmath12 channels @xcite and explore possible ways to increase their sensitivity to higgs production in susy processes .
finally , we propose a search in @xmath159 final states that could be much more sensitive to our model than other searches .
both atlas and cms have presented combined searches at 7 and 8 tev for inclusive higgs to diphoton final states , finding respectively best fit values of @xmath160 and @xmath161 for a higgs mass near @xmath162 gev @xcite .
an important point is that this search does not veto on any extra activity in the event , so we expect our model to give a fractional enhancement to the rate of roughly @xmath163 , where @xmath164 is the squark production cross section and the factor of 2 arises because two higgses are produced in each susy event . to obtain a more accurate result reflecting the changes in kinematics and isolation between the sm and susy cascade production modes , we again generated simulated signals as described in sec .
[ sec : constraints ] and applied the kinematic cuts specified by each search . for searches in diphoton final states , the total efficiencies @xmath36
depend on the efficiencies of the photon identification used by atlas and cms .
the pgs4 detector simulation applies only isolation cuts and a cut on the ratio of energy in the hadronic and em calorimeters when identifying photons , while atlas and cms use additional cuts on the shape and quality of the em shower in the calorimeter .
our mc simulation therefore overestimates the photon identification efficiency . to account for this , we simulated @xmath165 in the sm and computed the efficiency for the atlas @xcite and cms @xcite diphoton searches assuming the photon identification reported by pgs4 .
we define variables @xmath166 as the ratios of these quantities ( @xmath167 ) to the actual @xmath168 efficiencies reported by atlas and cms including the tighter photon identification criteria ( @xmath169 , @xmath170 ) : _ at both 7 and 8 tev we find @xmath171 .
we interpret @xmath166 as a constant efficiency for both photons from a higgs decay to pass the atlas or cms photon i d criteria given that they pass the pgs4 photon i d criteria and the photon kinematic cuts , and normalize all efficiencies for diphoton searches computed using our mc simulation by this constant : ^_bsm ^_bsm we chose to normalize to the @xmath172 production mode because the higgs has somewhat similar kinematics to the bsm production modes we are considering . for comparison , applying the same procedure to the @xmath173 production mode we obtain @xmath174 .
the rates from the search due to sm higgs production ( @xmath175 ) and susy higgs production in our model ( @xmath176 ) are thus given by @xmath177 where we assume the experimentally reported efficiencies and theoretically predicted cross section for a 125 gev higgs to determine the sm contribution .
note that we define @xmath36 to include the branching ratio for the @xmath178 decay . using the above method we can ( approximately )
determine the contribution of higgs production in our simplified model to the total rate in the atlas and cms inclusive @xmath179 searches .
we define @xmath30 to be the ratio of the total signal rate including bsm production to the sm rate , @xmath180 .
[ fig : gginclusive ] shows @xmath30 as a function of the squark mass and nlsp mass in the 8 tev cms inclusive diphoton search .
the efficiencies are insensitive to the stop / sbottom and nlsp mass , and @xmath30 therefore depends dominantly on the squark production cross section . because the squark cross section increases more rapidly with @xmath181 then the higgs cross section , at 7 tev @xmath30 is about @xmath182 smaller .
the rates at atlas are similar to those at cms .
for a 250 gev stop / sbottom , the total rate is @xmath183 of the sm rate .
we will discuss the possibility of explaining current hints of excesses in the inclusive diphoton channel in section [ sec : othersearch ] .
cms has published searches for diphoton resonances accompanied by a lepton or missing transverse energy ( met ) , targeting associated @xmath125 and @xmath184 production with a focus on the fermiophobic higgs model @xcite .
both of these final states are promising search channels for observation of bsm higgs production . in our simplified model ,
the signal in the @xmath185 channel occurs when one higgs decays to photons while the other produces leptons through decays to @xmath109 , @xmath186 or @xmath187 .
though the branching ratio to this state is small , the standard model background rate from non - higgs processes and signal rate from @xmath188 or @xmath189 production are both quite low . as before ,
we calculate the contribution to the rate in this search from our model , adjusting our mc efficiencies by @xmath190 , and plot the results against squark mass .
[ fig : leptondiphotonrate ] shows the total rate both in units of cross section and as a ratio to the standard model rate .
the search in @xmath191 of 8 tev data @xcite observed one event with about 1.6 expected background events within @xmath192 gev .
with 0.5 sm higgs events expected , this gives a 95% confidence level bound of @xmath193 on the visible cross section from new physics , excluding squark masses @xmath194 .
because the background is directly measured from the continuum diphoton spectrum , we scale the background uncertainty by @xmath195 to obtain an expected limit of @xmath196 in 20@xmath130 . to the rate from standard model higgs production , as a function of squark mass for a nlsp mass of 160 gev .
the red shaded region is ruled out at 95% cl by the existing @xmath197 of data at 8 tev .
dashed is the _ expected _ limit in @xmath85 .
, width=288 ] we validated our results by comparing our computed efficiencies for sm @xmath198 and @xmath199 events to the efficiencies determined from @xcite for a 120 gev fermiophobic higgs , finding agreement within 10% .
we now turn to the cms search for diphoton + missing transverse energy ( met ) targeting @xmath200 @xcite .
the standard model signal in this search is suppressed by branching ratios and the small @xmath199 production cross section . however
, this search can be quite sensitive to higgs production in nlsp decays due to the guaranteed met from the lsp , as long as the nlsp and higgs are not too degenerate .
the kinematic selections on the photons for this search are similar to the other diphoton channels , except that only photons within the barrel are selected .
additionally , events passing the lepton tag are vetoed .
events with missing transverse energy greater than 70 gev are selected .
the background rate ( fit from data ) was @xmath201 events in the range @xmath192 gev , and 3 events were observed in @xmath191 . with a sm higgs expectation of 0.2 events ,
this gives a 95% limit on the visible cross section from new physics of 1.1 fb .
the expected visible cross section for this search in the simplified model is shown as a function of squark and nlsp mass in fig .
[ fig : ggmet702d ] . in fig .
[ fig : combinedlimits ] , the 95% exclusion are compared to the limits from susy searches , both for current data and extrapolated to @xmath85 . since
the background in this channel is fit from continuum diphoton data , we scaled the systematic uncertainty as @xmath195 to estimate the expected limits in @xmath85 . this search has better reach than the susy searches for nlsp masses of @xmath202 . for nlsps lighter than @xmath203 ,
the signal in our model is suppressed significantly by a cut of met @xmath116 70 gev , which removes 80 - 90% of the signal events passing the other selections .
the sensitivity of such a search can therefore likely be improved by including one or more additional channels with lower met threshold , as was done in the non - resonant cms diphoton + met search @xcite . as an example
, the bin @xmath204 generally contains 2 - 3 times as many signal events as the @xmath205 bin .
if the background scales with the met cut as in the non - resonant search @xcite , it would be a factor of about 20 times larger in this bin , giving @xmath206 about half that of the higher met channel .
as the background increases rapidly with lower met cuts , sub - dividing this region could further increase the sensitivity in lower met channels . as before , we validated our results against the experimental efficiencies for a 120 gev fermiophobic higgs .
our analysis yields a 125 gev sm higgs signal rate of @xmath207 , which agrees within 10% with the efficiency from the expected signal rate in @xcite .
limit from @xmath197 of data at 8 tev to be 1.1 fb.,width=288 ] the standard model background rate for the @xmath178 search in @xmath197 at @xmath209 at cms is 433.8 events / gev near the higgs mass peak @xcite , and consists dominantly of prompt diphoton ( @xmath210 ) events and single prompt photon events with one or more jets ( @xmath211 ) , with one jet faking a photon @xcite .
this background can be considerably reduced by requiring one or two @xmath1-tags , while maintaining very good efficiency for the diphoton signal from our simplified model , which produces copious amounts of @xmath1 quarks in the final state .
we therefore consider a search with the same cuts as the cms diphoton search , but with the additional requirement of one or two @xmath1-tagged jets with @xmath212 .
we consider a counting experiment in which the number of events with diphoton mass within 2 gev of @xmath213 is compared to the background expectation .
the sensitivity of a similar search has been studied in refs .
@xcite in the context of measuring variations in the higgs couplings in higgs pair production events at very high luminosities with the 14 tev lhc , and the sensitivity of a diphoton + jets search at 14 tev without @xmath1-tags has been considered in @xcite for studying fermionic top - partner decays to the final states @xmath214 . to estimate the continuum , non - higgs background from @xmath210 and @xmath211 , we simulated @xmath215 mlm matched samples of @xmath216 and @xmath217 , where @xmath218 is a light flavor or gluon , a charm quark , or @xmath1 quark at 8 tev . as before we simulate the response of the cms detector with the default pgs cms parameter card , but here we use a different modified @xmath1-tagging function which gives a flat tag rate of @xmath219 for light flavor / gluon jets , @xmath220 for charm , and @xmath221 for @xmath1 quarks @xcite .
we separately normalize the @xmath210 and @xmath211 rates to the rate in @xcite with relative proportions of 25% @xmath211 and 75% @xmath210 based on the monte carlo results in @xcite .
the results for the background rates are shown in table [ tab : ggbrates ] .
the first @xmath1-tag reduces the non - higgs background by a factor of 100 to @xmath222 , with the remaining contributions from true @xmath1-jets , light quark fakes , and charm fakes all roughly comparable .
the second @xmath1-tag reduces this background by another factor of 50 to @xmath223 , with the remaining contributions dominantly from true @xmath1-jets or mistagged charm . .
the visible cross section for non - higgs continuum and sm higgs diphoton rates after various cuts in the @xmath80 search , including all detector efficiencies , branching ratios , and fake rates .
the @xmath210 and @xmath211 categories include the non - resonant , non - electroweak contributions to the rate in events with 2 real photons or 1 real photon and 1 jet faking a photon respectively .
@xmath224 is the contribution from top quark pair production with hard radiated photons or electrons / jets faking photons .
@xmath225 is a higgs - strahlung process in the sm , while @xmath226 includes this and also qcd production of a @xmath227 pair in gluon fusion higgs production .
for @xmath228 we use the tree level cross section from madgraph5 .
@xmath184 is the associated production of a sm higgs and @xmath229 .
the cross section for @xmath224 is from @xcite and for @xmath230 and @xmath184 from @xcite .
the @xmath210 and @xmath211 contributions are normalized to the observed background in @xcite as described in the text .
the signal rate in the simplified model for a 300 gev squark and 150 gev nlsp is also shown .
[ tab : ggbrates ] [ cols="^,^,^,^,^,^,^,^,^,^",options="header " , ] for the @xmath60 selection , top pair production with hard radiated photons or electrons / jets faking photons is also an important continuum background . to estimate this ,
we take a jet - photon fake rate of @xmath231 @xcite and an electron - photon fake rate of @xmath232 @xcite .
the estimated rate is @xmath233 , with fake electrons and radiated photons giving the dominant contributions .
sm @xmath230 also gives a resonant contribution to the background with a rate comparable to the continuum background .
@xmath199 and @xmath228 production give a negligible rate , as does non - resonant electroweak production with prompt or fake photons .
the 2@xmath1 channel is therefore essentially backgroundless at @xmath234 and in @xmath85 has @xmath235 expected continuum event falling in the higgs mass window and @xmath236 expected sm @xmath230 event . importantly , because this is a search for a known resonance , the background rate can be measured from data as in the usual higgs diphoton search , so an actual search would be limited only by statistical uncertainties in the background fit . the visible cross section for the @xmath237 and @xmath60 selections are shown for our simplified model in fig .
[ fig : bggvisxs ] . the efficiency for signal events passing the inclusive diphoton selections to also pass the @xmath237 selection is @xmath238 , while for the @xmath60 selection it is @xmath239 . for the @xmath237 channel
we estimate the systematic uncertainty in a background fit to be @xmath105 and @xmath182 at @xmath234 and @xmath85 respectively based on counting statistics in the neighboring diphoton invariant mass bins . with a background rate of @xmath240 ,
we find expected @xmath241 limits on the visible cross of @xmath242 for @xmath234 and @xmath243 for @xmath85 .
the @xmath237 channel has sensitivity significantly exceeding the other diphoton and susy searches we have discussed , as shown in fig .
[ fig : combinedlimits ] .
the near backgroundless @xmath60 channel is even more sensitive to our model and can even lead to a discovery at the 8 tev lhc .
since there are large uncertainties in our background estimate due to the fake rates for photons and @xmath1-tags , we estimate the reach of the @xmath60 channel taking the background rate to be @xmath244 .
[ fig:2bdiscovery ] shows the expected signal significance in @xmath245 and @xmath85 , along with the 95% cl excluded region from the existing susy and diphoton+met searches and their extrapolated limits in 20@xmath130 . for direct comparison with the other searches , the expected 95% exclusions from this channel are also shown in fig .
[ fig : combinedlimits ] .
in 5@xmath130 the @xmath3 search already has significantly better reach than the other searches we have considered , and in @xmath85 it gives an expected significance of @xmath246 for squarks up to @xmath247 , well beyond even the exclusion reach for the other searches , and a potential @xmath4 discovery for squarks up to @xmath248 .
another important advantage of this channel over other susy searches that can probe our models is that by exploiting the higgs diphoton resonance , this search clearly identifies the the involvement of an on - shell higgs boson in the new physics .
in contrast , a signal observed in the multi-@xmath1 or leptonic searches would not immediately suggest excess higgs production .
@xcite presents a combined fit to the atlas and cms 7 and 8 tev signals in the various sm higgs searches .
the only @xmath249 deviations from the sm predictions are excesses in the inclusive and vbf tagged @xmath210 channels and a low rate in the @xmath250 channel , driven primarily by a low rate in the vbf tagged @xmath251 channel at cms @xcite . as shown in fig .
[ fig : gginclusive ] , for a squark mass of 250 - 290 gev , an enhancement of 20 - 50% in the diphoton channel from higgs production in susy cascades is possible while remaining consistent with other searches .
this search is sensitive to our model because it makes use of the clean @xmath210 resonance and does not place cuts on the non - higgs activity in the event .
this is also true for the @xmath252 resonance searches , so we would expect any enhancement in the inclusive diphoton searches by this mechanism to give a comparable enhancement to @xmath252 .
on the other hand , searches in other channels often require additional cuts on the event to reduce sm backgrounds and consider more complicated kinematic variables , resulting in lowered signal strength from the non - standard higgs production modes we consider .
for example , in order to reduce backgrounds from top production , the atlas and cms searches in the @xmath253 @xcite and in most channels for @xmath18 @xcite veto events with any @xmath1-tagged jets . since events from our models typically contain several @xmath1-jets , they give a negligible contribution to the rate in these searches .
furthermore , the kinematic variables considered in these analyses ( such as the transverse mass of leptons and the reconstructed tau - tau invariant mass ) are distorted by the different kinematics and additional missing energy in events from squark production , so those events that are selected will not necessarily increase the signal strength for the 125 gev sm higgs hypothesis .
similar issues limit the sensitivity of searches for @xmath172 production with @xmath254 and the vector boson @xmath255 decaying to leptons and/or neutrinos @xcite .
although our model produces these same final states , the @xmath1-jets selected by the analyses will not necessarily be the higgs decay products , so a mass peak will not be visible .
both analyses place many kinematic cuts on events , such as requiring the @xmath256 momentum to be nearly back - to - back with the reconstructed vector boson momentum when projected into the transverse plane , greatly reducing the efficiency in our case . in the @xmath179 analyses
both cms and atlas define dijet - tagged event classes , selecting jets widely separated in pseudorapidity @xmath257 to obtain sensitivity to vbf higgs production in the sm .
the additional jets present in bsm events from our simplified model rarely satisfy this @xmath258 requirement , and the enhancement relative to the sm rate is at most a few percent in the parameter range we consider . to summarize , if the simplified model of excess higgs production we have considered gives an observable enhancement to inclusive @xmath178 rate , then this is consistent with the lack of observed excesses in @xmath17 , @xmath18 , and @xmath259 , but is in slight tension with an observed rate for @xmath260 which is consistent with the sm and does not explain the excess in the dijet - tagged @xmath179 channel .
therefore , while the enhancement to the inclusive diphoton channel is encouraging , the simplified model alone likely does not give a significantly better fit to current data than does the sm prediction .
thus far we have always assumed that the effective couplings of the higgs are the same as those predicted by the standard model .
however , the light third generation squarks of our models can modify the higgs couplings to gluons and photons through loop effects , particularly if there are light stops ( discussed recently in refs .
large stop mixing ( e.g. from @xmath261-terms ) is required to give the correct sign to enhance the diphoton branching ratios , and extremely large mixing is required to actually enhance the total diphoton rate from gluon fusion higgs production since the @xmath262 rate is at first suppressed as the diphoton branching ratio is enhanced @xcite . by contributing another production mode which is not sensitive to the @xmath262 coupling , higgs production in cascade decays allows an overall enhancement to the diphoton rate at smaller mixings .
the modified @xmath178 branching ratio then also enhances the dijet - tagged diphoton rate to reproduce the observed excess , while the reduced @xmath262 rate prevents an excess in the @xmath263 channel .
we do not attempt a detailed fit to the data , but rather note that in the range of squark masses where these models can give significant enhancements to inclusive sm higgs searches , it is likely that excesses in the @xmath12 search and susy searches discussed in sec .
[ sec : ggmet ] and [ sec : constraints ] will be observed in 20@xmath130 of 8 tev data and that a discovery could be made in a search for @xmath80 as described in sec .
[ sec : ggb ] . as a general point ,
excess higgs production in bsm cascades , possibly with less dramatic exclusive signatures than in our models , is an interesting degree of freedom to consider when fitting rates in higgs searches , giving qualitatively different effects from modifications to higgs couplings .
we also note that to improve sensitivity , both atlas and cms divide the inclusive diphoton search into several event classes based on the kinematics of the photon and the quality of its reconstruction .
our results in sec .
[ sec : gg ] based on the inclusive diphoton count therefore only give an approximation of expected best fit rates , which in refs .
@xcite are actually determined with a fit taking into account the different signal and background expectations in the individual event classes .
of particular note , both atlas and cms have event classes capturing higgs production with higher transverse momentum , which has a smaller background rate . for example
, atlas separates diphoton candidates into high and low @xmath264 event classes , where @xmath264 is the component of the diphoton momentum transverse to the thrust axis @xcite .
the simplified model higgs production leads to a substantially larger signal enhancement in the high @xmath264 category compared to sm @xmath262 events . for a given enhancement to the inclusive rate relative to the sm ,
the expected enhancement in the high @xmath264 categories relative to the sm rate in these channels is generally @xmath265 times greater . in the combined 7 + 8
tev results these channel show roughly equal enhancements compared to the sm rate , but with large enough individual uncertainty to accommodate a substantially higher rate in the high @xmath264 channels .
while other exclusive diphoton channels are more sensitive to our specific model , with more statistics this could be a useful constraint on more general possibilities for enhancing higgs production through bsm cascades .
if the nlsp in our models is a sufficiently light wino or higgsino , then the electroweak direct pair production rates can be large compared to sm associated higgs production rates .
this can mildly constrain the neutralino / chargino spectra possible for the colored production case , and can also enhance the signal in diphoton searches . as discussed in sec .
[ sec : chitohiggs ] , we consider the possibilities of wino co - nlsps with small splittings such that the charginos decay promptly to the lsp , and a mixed wino - bino or higgsino nlsp , with the nearby neutralinos and charginos decaying promptly to the nlsp
. for wino - like co - nlsps , the dominant production modes are @xmath266 and @xmath267 ( here @xmath268 denotes the neutral nlsp and @xmath269 the lsp ) .
the winos decay as @xmath270 and @xmath271 , so the final states of interest are @xmath272 and @xmath273
@xcite discuss the possibility that @xmath267 production can appear in @xmath274 search channels , but for @xmath275 decaying dominantly to @xmath276 there is no significant constraint on direct production in the parameter space of interest .
the discovery potential for heavier charginos and neutralinos in the @xmath125 mode at the 8 and 14 tev lhc has been studied in refs .
however for a lighter nlsp with small splitting with the higgs , the met signature is suppressed and the resulting signals become much more similar to sm associated higgs production . because the production cross section can be many times larger than the sm rate and there is no additional activity in the event ( unlike in the colored production case ) , all of the existing searches for sm - like @xmath125 associated production will directly constrain the @xmath277 production cross section , which is shown relative to the sm @xmath125 cross section at 8 tev in fig .
[ fig : winodirectxs ] .
the most constraining direct limit on @xmath125 is an atlas search is for @xmath278 at 7 tev with a limit of @xmath279 @xcite , requiring @xmath280 ; a similar bound in the same channel is also presented in ref .
@xcite in a search for direct production of gauginos with leptophilic decays .
the limit from the cms @xmath13 search @xcite discussed earlier is @xmath281 .
the strongest combined constraint on @xmath199/@xmath198 production currently comes from a cms search for @xmath282 with leptonically decaying @xmath61 s and @xmath24 s at 7 and 8 tev , which gives a 95% upper limit of @xmath283 for a 125 gev higgs @xcite , but individual bounds on @xmath125 and @xmath184 production were not presented .
@xcite has suggested using jet substructure methods to further improve sensitivity for direct production of gauginos decaying in the @xmath284 channel .
associated production at 8 tev .
dashed is the 95% cl limit on the bsm contribution in the atlas @xmath285 channel , and dotted is the limit from the cms @xmath185 channel .
[ fig : winodirectxs],width=240 ] for a mixed wino - bino nlsp or a higgsino - like nlsp , the charginos can promptly decay to the nlsp which then decays to a weakly coupled lsp , so the dominant signatures are in @xmath286 .
the tevatron and lhc signatures for direct production of higgsinos decaying to gravitinos have been discussed in @xcite , with emphasis on the situations where the branching ratio to @xmath24 dominates or is comparable to the higgs branching ratio .
we focus instead on the case where the @xmath126 mode is dominant , which in the higgsino case at low masses requires @xmath287 .
these modes are even less constrained than the wino co - nlsp because of their smaller branching fraction to leptonic final states .
@xcite found that a dedicated tevatron search in @xmath288 could be sensitive to a higgsino nlsp with masses @xmath289 gev , but to date the most constraining search has been a cdf search @xcite for gluinos decaying to sbottoms requiring 2@xmath1-tags and met @xmath116 70 gev , which does not constrain either case even assuming the same efficiencies as the targeted @xmath290 channel . ref .
@xcite briefly discussed the possibility of a @xmath291 search at the lhc , but again we find that existing searches ( cms @xmath292 and atlas @xmath87 ) do not set a limit on direct production .
the signals in the inclusive diphoton , diphoton+met , and diphoton+1@xmath1 channels are shown in fig .
[ fig : directdiphoton ] for the two wino nlsp possibilities assuming @xmath135 branching fractions for the nlsp to the higgs , and for a higgsino nlsp decaying to a gravitino lsp at @xmath293 and @xmath294 .
only the mixed wino - bino nlsp with prompt decays of the charginos to the nlsp has a potentially interesting signal in these channels , due to its large production cross section compared to a higgsino nlsp and the potential to produce two higgses in each event , unlike the wino co - nlsp scenario .
for masses lighter than 150 gev it can have a signal of 20 - 30% the sm rate in the inclusive @xmath210 channel and can be potentially observed in the @xmath295 channel in 20@xmath130 .
the rate in the @xmath3 channel is too low to afford sensitivity .
the @xmath12 search is just barely sensitive enough to excludes masses around @xmath296 in 20@xmath130 , but for lower masses the met signal is too suppressed .
again , the rate in a lower met bin can be several times higher and may increase the sensitivity for a lighter wino .
although the rates for the other nlsp scenarios are too small to be observable from just direct production , in the context of the simplified squark production model the additional contribution from direct production can slightly enhance the overall sensitivity of the diphoton channels .
we have shown that in a variety of models the decays of a neutralino nlsp can give copious extra higgs boson production while remaining consistent with constraints from bsm searches .
this leads to potentially observable signatures in higgs resonance searches .
well - motivated realizations of this scenario include wino- and bino - like nlsps with a singlino / goldstino / string photino lsp , or a higgsino - like nlsp with a gravitino lsp .
we have focused on pair production of light third generation squarks , motivated by naturalness of the electroweak scale .
a wide class of realistic models can be well described by the simplified model of fig .
[ fig : simpmodel ] , leading to final states with two bottom quarks , two higgses , and missing energy .
existing searches constrain these models to have squark masses @xmath297 and nlsp masses @xmath298 , as summarized in fig .
[ fig : combinedlimits ] . for light squarks
we argue that the most effective search strategy to probe these models is to target the @xmath179 decay mode , which would give a clear signal of a higgs resonance .
a search in @xmath299 final states would have high signal efficiency and very low sm background , and could give high signal significance in 8 tev lhc data for squarks up to @xmath68 ( figure [ fig:2bdiscovery ] ) .
although a @xmath299 search is expected to be the best probe for the models we consider , even existing analyses can have some sensitivity with further 8 tev data .
of particular interest is the possibility that the bsm production rate is high enough to account for the high signal strength reported by atlas and cms in the inclusive diphoton channels .
this would require squarks near the limits placed by susy searches ( see fig .
[ fig : gginclusive ] ) . some excess
are then expected for these searches in 8 tev data , though this would be difficult to link to higgs physics .
more informative would be the searches for associated higgs production , namely the @xmath300 and @xmath301 analyses , which would all have rates well in excess of the standard model prediction . as discussed in sec .
[ sec : othersearch ] , little or no enhancement is expected in most searches in other higgs decay modes such as @xmath109 or @xmath186 .
therefore , observing excesses in only the diphoton channels and particularly the associated production channels would suggest our class of models . in the parameter space where observable excesses can occur in these channels , a search in @xmath299
can likely give a @xmath4 discovery in 20@xmath130 .
although we have focused on a particular susy simplified model in this work , similar phenomenology can arise in many other models .
final states very similar to those in our simplified model can arise for example from decays of new vector - like quarks to a higgs and third generation quarks , especially in the decays of bottom - like quarks , e.g. @xmath302 .
more generally , much of what we have said is relevant for pair production of any new particles that decay dominantly to higgses .
the diphoton decay mode still offers an excellent opportunity for identifying the higgs resonance , and a @xmath303 search can still have good efficiency because of the large higgs branching ratio to @xmath304 .
finally , while this work has explored one well motivated scenario in which non - standard higgs channels can become sensitive , we emphasize the importance of searching for higgs production in as many channels as possible .
abandoning all theoretical prejudice , little is currently known about the production and decay of the newly observed resonance at 125 gev , and there remains open the possibility of observing significant departures from standard model expectations in the very near future .
we would like to thank masha baryakhtar , timothy cohen , nathaniel craig , savas dimopoulos , peter graham , jo - anne hewett , caitlin malone , john march - russell , tom rizzo , surjeet rajendran , ken van tilburg and giovanni villadoro for useful discussions .
we especially thank timothy cohen and peter graham for giving comments on a draft of this work before publication .
is supported by a stanford graduate fellowship , and k.h .
is supported by a nsf graduate research fellowship under grant no .
this work was partially supported by erc grant bsmoxford no .
228169 .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1207.7214 [ hep - ex ] .
[ cms collaboration ] evidence for a new state decaying into two photons in the search for the standard model higgs boson in pp collisions .
cms - pas - hig-12 - 015 [ 2012 ] a. azatov , r. contino and j. galloway , jhep * 1204 * , 127 ( 2012 ) [ arxiv:1202.3415 [ hep - ph ] ] .
r. benbrik , m. g. bock , s. heinemeyer , o. stal , g. weiglein and l. zeune , arxiv:1207.1096 [ hep - ph ] .
d. carmi , a. falkowski , e. kuflik , t. volansky and j. zupan , arxiv:1207.1718 [ hep - ph ] .
j. ellis and t. you , jhep * 1206 * , 140 ( 2012 ) [ arxiv:1204.0464 [ hep - ph ] ] .
j. r. espinosa , c. grojean , m. muhlleitner and m. trott , arxiv:1207.1717 [ hep - ph ] .
giardino , k. kannike , m. raidal and a. strumia , jhep * 1206 * , 117 ( 2012 ) [ arxiv:1203.4254 [ hep - ph ] ] .
giardino , k. kannike , m. raidal and a. strumia , arxiv:1207.1347 [ hep - ph ] .
m. klute , r. lafaye , t. plehn , m. rauch and d. zerwas , arxiv:1205.2699 [ hep - ph ] . i. low , j. lykken and g. shaughnessy , arxiv:1207.1093 [ hep - ph ] .
f. del aguila , g. l. kane and m. quiros , phys .
rev . lett . * 63 * , 942 ( 1989 ) .
f. del aguila , l. ametller , g. l. kane and j. vidal , nucl .
b * 334 * , 1 ( 1990 ) .
m. sher , phys .
d * 61 * , 057303 ( 2000 ) [ hep - ph/9908238 ] . j. a. aguilar - saavedra , jhep * 0612 * , 033 ( 2006 ) [ hep - ph/0603200 ] . c. kilic , k. kopp and t. okui , phys
. rev .
d * 83 * , 015006 ( 2011 ) [ arxiv:1008.2763 [ hep - ph ] ] .
a. carmona , m. chala and j. santiago , jhep * 1207 * , 049 ( 2012 ) [ arxiv:1205.2378 [ hep - ph ] ] . g. d. kribs , a. martin , t. s. roy and m. spannowsky , phys . rev .
d * 81 * , 111501 ( 2010 ) [ arxiv:0912.4731 [ hep - ph ] ] .
p. bandyopadhyay , e. j. chun and j. -c .
park , jhep * 1106 * , 129 ( 2011 ) [ arxiv:1105.1652 [ hep - ph ] ] . m. son , c. spethmann and b. tweedie , arxiv:1204.0525 [ hep - ph ] .
p. bandyopadhyay , bandyopadhyay:2010tv jhep * 1108 * , 016 ( 2011 ) [ arxiv:1008.3339 [ hep - ph ] ] .
a. datta , a. djouadi , m. guchait and y. mambrini , phys .
d * 65 * , 015007 ( 2002 ) [ hep - ph/0107271 ] .
a. datta , a. djouadi , m. guchait and f. moortgat , nucl .
b * 681 * , 31 ( 2004 ) [ hep - ph/0303095 ] .
f. de campos , o. j. p. eboli , m. b. magro , d. restrepo and j. w. f. valle , phys .
d * 80 * , 015002 ( 2009 ) [ arxiv:0809.1637 [ hep - ph ] ] .
a. c. fowler and g. weiglein , jhep * 1001 * , 108 ( 2010 ) [ arxiv:0909.5165 [ hep - ph ] ] .
s. gori , p. schwaller and c. e. m. wagner , phys .
d * 83 * , 115022 ( 2011 ) [ arxiv:1103.4138 [ hep - ph ] ] .
g. d. kribs , a. martin , t. s. roy and m. spannowsky , phys .
d * 82 * , 095012 ( 2010 ) [ arxiv:1006.1656 [ hep - ph ] ] .
o. stal and g. weiglein , jhep * 1201 * , 071 ( 2012 ) [ arxiv:1108.0595 [ hep - ph ] ] .
g. azuelos , k. benslama , d. costanzo , g. couture , j. e. garcia , i. hinchliffe , n. kanaya and m. lechowski _ et al .
_ , eur .
j. c * 39s2 * , 13 ( 2005 ) [ hep - ph/0402037 ] . g. d. kribs , a. martin and t. s. roy , phys .
d * 84 * , 095024 ( 2011 ) [ arxiv:1012.2866 [ hep - ph ] ] .
a. azatov , o. bondu , a. falkowski , m. felcini , s. gascon - shotkin , d. k. ghosh , g. moreau and s. sekmen , arxiv:1204.0455 [ hep - ph ] .
n. vignaroli , jhep * 1207 * , 158 ( 2012 ) [ arxiv:1204.0468 [ hep - ph ] ] .
h. baer , v. barger , a. lessa , w. sreethawong and x. tata , phys .
d * 85 * , 055022 ( 2012 ) [ arxiv:1201.2949 [ hep - ph ] ] .
d. ghosh , m. guchait and d. sengupta , arxiv:1202.4937 [ hep - ph ] .
p. byakti and d. ghosh , arxiv:1204.0415 [ hep - ph ] .
m. asano , h. d. kim , r. kitano and y. shimizu , jhep * 1012 * , 019 ( 2010 ) [ arxiv:1010.0692 [ hep - ph ] ] .
y. kats , p. meade , m. reece and d. shih , jhep * 1202 * , 115 ( 2012 ) [ arxiv:1110.6444 [ hep - ph ] ] .
k. t. matchev and s. d. thomas , phys .
d * 62 * , 077702 ( 2000 ) [ hep - ph/9908482 ] .
p. meade , m. reece and d. shih , jhep * 1005 * , 105 ( 2010 ) [ arxiv:0911.4130 [ hep - ph ] ] .
j. t. ruderman and d. shih , arxiv:1103.6083 [ hep - ph ] .
j. thaler and z. thomas , jhep * 1107 * , 060 ( 2011 ) [ arxiv:1103.1631 [ hep - ph ] ] .
d. das , u. ellwanger and a. m. teixeira , jhep * 1204 * , 067 ( 2012 ) [ arxiv:1202.5244 [ hep - ph ] ]
. m. baryakhtar , n. craig and k. van tilburg , arxiv:1206.0751 [ hep - ph ] .
j. a. aguilar - saavedra , pos top * 2006 * , 003 ( 2006 ) [ hep - ph/0603199 ] .
j. a. aguilar - saavedra , jhep * 0911 * , 030 ( 2009 ) [ arxiv:0907.3155 [ hep - ph ] ] .
k. harigaya , s. matsumoto , m. m. nojiri and k. tobioka , arxiv:1204.2317 [ hep - ph ] .
a. girdhar and b. mukhopadhyaya , arxiv:1204.2885 [ hep - ph ] .
n. vignaroli , arxiv:1207.0830 [ hep - ph ] .
h. baer , m. bisset , x. tata and j. woodside , phys . rev .
d * 46 * , 303 ( 1992 ) .
p. bandyopadhyay , a. datta and b. mukhopadhyaya , phys .
b * 670 * , 5 ( 2008 ) [ arxiv:0806.2367 [ hep - ph ] ] .
p. bandyopadhyay , jhep * 0907 * , 102 ( 2009 ) [ arxiv:0811.2537 [ hep - ph ] ] .
b. bhattacherjee and a. datta , jhep * 1203 * , 006 ( 2012 ) [ arxiv:1104.4529 [ hep - ph ] ] .
i. hinchliffe , f. e. paige , m. d. shapiro , j. soderqvist and w. yao , phys .
d * 55 * , 5520 ( 1997 ) [ hep - ph/9610544 ] .
k. huitu , r. kinnunen , j. laamanen , s. lehti , s. roy and t. salminen , eur .
j. c * 58 * , 591 ( 2008 ) [ arxiv:0808.3094 [ hep - ph ] ] .
[ atlas collaboration ] , `` search for squarks and gluinos with the atlas detector using final states with jets and missing transverse momentum and 4.7 @xmath130 of @xmath305 proton - proton collision data , '' atlas - conf-2012 - 033 ( 2012 ) .
a. g. cohen , d. b. kaplan and a. e. nelson , phys .
b * 388 * , 588 ( 1996 ) [ hep - ph/9607394 ] .
k. agashe and m. graesser , phys . rev .
d * 59 * , 015007 ( 1999 ) [ hep - ph/9801446 ] .
d. e. kaplan , f. lepeintre , a. masiero , a. e. nelson and a. riotto , phys .
d * 60 * , 055003 ( 1999 ) [ hep - ph/9806430 ] .
o. aharony , l. berdichevsky , m. berkooz , y. hochberg and d. robles - llana , phys . rev .
d * 81 * , 085006 ( 2010 ) [ arxiv:1001.0637 [ hep - ph ] ] .
r. barbieri , e. bertuzzo , m. farina , p. lodone and d. zhuridov , jhep * 1012 * , 070 ( 2010 ) [ erratum - ibid .
* 1102 * , 044 ( 2011 ) ] [ arxiv:1011.0730 [ hep - ph ] ] .
n. craig , d. green and a. katz , jhep * 1107 * , 045 ( 2011 ) [ arxiv:1103.3708 [ hep - ph ] ] .
g. larsen , y. nomura and h. l. l. roberts , jhep * 1206 * , 032 ( 2012 ) [ arxiv:1202.6339 [ hep - ph ] ] .
n. craig , s. dimopoulos and t. gherghetta , jhep * 1204 * , 116 ( 2012 ) [ arxiv:1203.0572 [ hep - ph ] ] .
n. craig , m. mccullough and j. thaler , jhep * 1206 * , 046 ( 2012 ) [ arxiv:1203.1622 [ hep - ph ] ] .
h. baer , v. barger , p. huang and x. tata , jhep * 1205 * , 109 ( 2012 ) [ arxiv:1203.5539 [ hep - ph ] ] .
t. cohen , a. hook and g. torroba , arxiv:1204.1337 [ hep - ph ] .
l. randall and m. reece , arxiv:1206.6540 [ hep - ph ] .
r. dermisek , mod .
a * 24 * , 1631 ( 2009 ) [ arxiv:0907.0297 [ hep - ph ] ] .
p. w. graham , a. ismail , s. rajendran and p. saraswat , phys .
d * 81 * , 055016 ( 2010 ) [ arxiv:0910.3020 [ hep - ph ] ] .
s. p. martin , phys . rev .
d * 81 * , 035004 ( 2010 ) [ arxiv:0910.2732 [ hep - ph ] ] . c. brust , a. katz , s. lawrence and r. sundrum , jhep * 1203 * , 103 ( 2012 ) [ arxiv:1110.6670 [ hep - ph ] ] .
m. papucci , j. t. ruderman and a. weiler , arxiv:1110.6926 [ hep - ph ] .
h. m. lee , v. sanz and m. trott , jhep * 1205 * , 139 ( 2012 ) [ arxiv:1204.0802 [ hep - ph ] ] . c. brust , a. katz and r. sundrum , arxiv:1206.2353 [ hep - ph ] .
n. craig , j. a. evans , r. gray , m. park , s. somalwar , s. thomas and m. walker , arxiv:1207.6794 [ hep - ph ] .
g. d. kribs and a. martin , arxiv:1207.4496 [ hep - ph ] .
j. alwall , m. herquet , f. maltoni , o. mattelaer and t. stelzer , jhep * 1106 * , 128 ( 2011 ) [ arxiv:1106.0522 [ hep - ph ] ] .
t. sjostrand , s. mrenna and p. z. skands , jhep * 0605 * , 026 ( 2006 ) [ hep - ph/0603175 ] .
j. conway , pgs - pretty good simulation of high energy collisions .
http://physics.ucdavis.edu/ conway / research / software / pgs / pgs4-general.htm .
w. beenakker , m. klasen , m. kramer , t. plehn , m. spira and p. m. zerwas , phys .
lett . * 83 * , 3780 ( 1999 ) [ erratum - ibid . *
100 * , 029901 ( 2008 ) ] [ hep - ph/9906298 ] .
lhc higgs cross section working group , s. dittmaier , c. mariotti , g. passarino , and r. tanaka ( eds . ) , _ handbook of lhc higgs cross sections : 1 .
inclusive observables _ , cern-2011 - 002 ( cern , geneva , 2011 ) , arxiv:1101.0593 [ hep - ph ] .
https://twiki.cern.ch/twiki/bin/view/lhcphysics/crosssections [ atlas collaboration ] , `` search for gluino - mediated scalar top and bottom quark production in final states with missing transverse energy and at least three @xmath1-jets with the atlas detector , '' atlas - conf-2012 - 058 ( 2012 ) .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1204.6736 [ hep - ex ] .
[ atlas collaboration ] search for direct top squark pair production in final states with one isolated lepton , jets , and missing transverse momentum in sqrt(s ) = 7 tev pp collisions using 4.7 ifb of atlas data .
atlas - conf-2012 - 073 [ 2012 ] g. aad _
[ atlas collaboration ] , phys .
lett . * 108 * , 111803 ( 2012 ) [ arxiv:1202.1414 [ hep - ex ] ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 710 * , 403 ( 2012 ) [ arxiv:1202.1487 [ hep - ex ] ] .
[ cms collaboration ] search for supersymmetry in events with photons and missing energy .
cms - pas - sus-12 - 018 u. baur , t. plehn and d. l. rainwater , phys .
d * 69 * , 053004 ( 2004 ) [ hep - ph/0310056 ] .
r. contino , m. ghezzi , m. moretti , g. panico , f. piccinini and a. wulzer , arxiv:1205.5444 [ hep - ph ] .
[ cms collaboration ] , `` status of b - tagging and vertexing tools for 2011 data analysis '' cms - pas - btv-11 - 002 ( 2012 ) .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1206.0756 [ hep - ex ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 710 * , 91 ( 2012 ) [ arxiv:1202.1489 [ hep - ex ] ] .
g. aad _ et al .
_ [ atlas collaboration ] , arxiv:1206.5971 [ hep - ex ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 713 * , 68 ( 2012 ) [ arxiv:1202.4083 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas collaboration ] , arxiv:1207.0210 [ hep - ex ] .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 710 * , 284 ( 2012 ) [ arxiv:1202.4195 [ hep - ex ] ] .
a. arvanitaki and g. villadoro , jhep * 1202 * , 144 ( 2012 ) [ arxiv:1112.4835 [ hep - ph ] ] .
k. blum , r. t. dagnolo and j. fan , arxiv:1206.5303 [ hep - ph ] .
m. r. buckley and d. hooper , arxiv:1207.1445 [ hep - ph ] .
b. feigl , h. rzehak and d. zeppenfeld , arxiv:1205.3468 [ hep - ph ] .
b. feigl , h. rzehak and d. zeppenfeld , eur .
j. c * 72 * , 1903 ( 2012 ) [ arxiv:1108.1110 [ hep - ph ] ] . m. lisanti and n. weiner , phys . rev .
d * 85 * , 115005 ( 2012 ) [ arxiv:1112.4834 [ hep - ph ] ] .
[ atlas collaboration ] search for the associated higgs boson production in the @xmath307 decay mode using 4.7 fb@xmath308 of data collected with the atlas detector at @xmath181 = 7 tev atlas - conf-2012 - 078 ( 2012 ) [ atlas collaboration ] `` search for direct production of charginos and neutralinos in events with three leptons and missing transverse momentum in sqrt(s)=7 tev pp collisions with the atlas detector '' , atlas - conf-2012 - 077 ( 2012 )
|
the atlas and cms experiments have recently claimed discovery of a higgs boson - like particle at @xmath0 confidence and are beginning to test the standard model predictions for its production and decay . in a variety of supersymmetric models ,
a neutralino nlsp can decay dominantly to the higgs and the lsp . in natural susy models ,
a light third generation squark decaying through this chain can lead to large excess higgs production while evading existing bsm searches .
such models can be observed at the 8 tev lhc in channels exploiting the rare diphoton decays of the higgs produced in the cascade decay .
identifying a diphoton resonance in association with missing energy , a lepton , or @xmath1-tagged jets is a promising search strategy for discovery of these models , and would immediately signal new physics involving production of a higgs boson .
we also discuss the possibility that excess higgs production in these susy decays can be responsible for enhancements of up to 50% over the sm prediction for the observed rate in the existing inclusive diphoton searches , a scenario which would likely by the end of the 8 tev run be accompanied by excesses in the @xmath2 and susy multi - lepton/@xmath1 searches and a potential discovery in a @xmath3 search .
|
for nearly half a century the theory of the weakly interacting bose gases ( wibg ) remained in the realm of purely theoretical investigations @xcite ( for a recent review , see @xcite ) providing insight into the nature of superfluid states of matter but not directly relating to existing experimental systems .
the situation changed with the realization of bose einstein condensation in cold atomic gases @xcite .
the typical values of the interaction parameter for alkali atoms are very small @xmath0 , where @xmath1 is the number density of the gas and @xmath2 is the @xmath3-wave scattering length .
since 1995 both the theory of wibg and experiments have progressed with a close relationship between the two @xcite . until recently , existing mean - field and variational treatments such as the bogoliubov zero - temperature approximation and the finite - temperature quasi - condensate theory @xcite were capable of describing the data within the relatively large experimental uncertainties .
however , the improvements in detection techniques , the studies of optical lattice systems in which the effective gas parameter can be made arbitrarily large , and the need for reliable thermometry , e.g. , through precise entropy matching , all indicate the need for a more accurate and controllable theory .
it is highly desirable , similarly to the well - known corrections to the ground state energy @xcite and condensate density @xcite , to have a theory which accounts for leading corrections to all thermodynamic properties at finite temperature , works in any dimension , and provides estimates for omitted higher order terms .
the accuracy of the description plays an important role in this paper .
we systematically estimate the errors of all approximations , which is crucial for comparing seemingly different schemes proposed for bose gases . indeed , alternative theories can be equivalent to each other within the level of their accuracy , but a definitive conclusion
can only be drawn on the basis of analyzing their systematic errors .
for example , within the same accuracy gap - less and gap - full finite - temperature schemes may often be regarded as equivalent without any preference for one of them , as long as the gap is on the order of or smaller than the terms in the chemical potential which the theory ignores . also , in the fluctuation region all perturbative schemes are equally inaccurate in terms of their treatment of the order parameter , and it makes no sense whatsoever to distinguish between them by the type of the transition they predict . in this work , we provide a rigorous framework for obtaining a consistent description of all thermodynamic finite-@xmath4 properties of the wibg in one , two , and three dimensions .
it is based on beliaev s regularized diagrammatic technique @xcite which allows us to calculate all relevant thermodynamic functions of the system .
the diagrammatic technique also leads to an accurate description of correlation functions up to sufficiently large distances where popov s hydrodynamic description takes over @xcite .
the key results of our calculations are summarized in section [ sec : summary ] , which is accessible to the general reader .
it turns out that all final results for thermodynamic functions perfectly fit into the pseudo - hamiltonian picture , i.e. we prove that the same answers would be obtained if they were calculated in the standard way with bogoliubov s non - interacting quasi - particle picture ( with a self - consistently defined temperature dependent spectrum ) .
in addition , we find an explicit expression for the pressure and give a more accurate expression for the chemical potential .
we obtain the equations of state for all basic thermodynamic quantities in an integral parametric form using temperature , effective chemical potential , and the interaction strength as parameters .
these results include sub - leading corrections .. at low enough temperatures , these quantities lose their ideal - gas leading terms , getting nothing instead as opposed to energy , chemical potential , and pressure , having straight - forward mean - field contributions that start to lead at low temperatures . ]
the analysis of systematic errors shows that the accuracy of our results can not be further improved within the paradigm of non - interacting quasiparticles with a ( temperature - dependent ) bogoliubov spectrum .
moreover , there exists a two - parametric continuum of possibilities for recasting the form of the final answers with the same accuracy due to the freedom of modifying the definitions of approximate notions of `` effective interaction '' and `` effective chemical potential '' . using this freedom
, one can trade the higher order terms in the explicit integral expressions for the thermodynamic quantities for those of the effective interaction vertex .
our analysis shows that the question of the energy and momentum dependence of the interaction vertex is more a matter of taste rather than accuracy .
we show that even in 2d where , in view of the vanishing @xmath3-scattering amplitude , a low - energy cutoff is unavoidable and one could expect that the momentum - dependent effective - interaction vertex is crucially important in any theory taking care of sub - leading corrections the effects associated with the energy and momentum dependence of the effective interaction can be naturally accounted for on equal footing with all the other effects of the same order . within the systematic error bars of our pseudo - hamiltonian approach ,
the effective interaction of the 2d wibg can be safely treated as an energy and momentum independent constant .
such analytical approaches are not supposed to work in the vicinity of the critical point @xmath5 .
this so - called fluctuation region is characterized by scaling functions that are universal for all u(1 ) weakly - interacting systems .
it has already been studied numerically with high accuracy @xcite , so that a combination of analytical and numerical data provides an accurate description at all temperatures . despite its success in treating the weakly interacting bose gas , beliaev
s original technique features a subtlety that has seriously questioned its reliability : the normal and anomalous self energies the fundamental quantities of the theory have been demonstrated to be essentially momentum - dependent , with the anomalous self - energy vanishing in the limit of zero momentum and frequency @xcite . on the other hand the first - order diagrams in the theory of the weakly interacting gas with a contact interaction yield momentum independent self energies . in the literature
, one can find a number of recipes of how to deal with this inherent infrared problem of beliaev s technique . in popov s approach @xcite
, the diagrammatic technique applies only to the higher - momentum part of the bosonic field , while the lower - momentum part is treated separately within the hydrodynamic representation .
similar results can be achieved by working with a finite - size system and taking advantage of the bi - modality of correlation properties of weakly interacting superfluid bosonic systems ( away from the fluctuation region ) @xcite .
the bi - modality allows one to select such a system size that in terms of local thermodynamic properties the system is already well in the macroscopic limit while the effect of the infra - red renormalization of the self energies is still negligibly small .
another option of a controllable microscopic description of weakly interacting bose gas is the renormalization - group treatment @xcite .
similar to the finite - size regularization , our approach is to exploit the bi - modality of correlation properties of the wibg , see figure [ fig : bimodal ] , by cutting off infra - red divergencies using small u(1 ) symmetry breaking terms .
the amplitude of the symmetry - breaking terms is such that their effect on the local thermodynamic properties of the system is negligibly small while all infrared singularities in beliaev s technique are gone .
clearly , the solution to the infrared problem comes at the expense of suppressing long - range fluctuations of the phase of the order parameter ( with opening of a gap in the spectrum at small momenta ) , and thus distorting the long - range behaviour of correlation functions
. nevertheless , this distortion is essentially irrelevant since it takes place at large distances where the ( generic to all superfluids ) behaviour of correlators is governed by hydrodynamic fluctuations of the phase of the order parameter , with a simple and transparent theoretical description @xcite . moreover , in all final answers one can explicitly take the limit of vanishing symmetry breaking terms . in terms of the final results ,
we confine our analysis to the most typical ( and most difficult in two and three dimensions ) case of the dilute gas , when the size of the potential @xmath6 is much smaller than the distance between the particles : @xmath7 in principle , the theory works also at @xmath8 , and becomes even simpler , since in this case the weakness of interaction literally implies born type of the potential ( while in the opposite case the amplitude of potential can be arbitrarily large in two and three dimensions ) .
however , only with the condition ( [ dilute ] ) the final answers become insensitive to the details of the interaction potential .
the second condition which is assumed in our final results is @xmath9 where @xmath10 is the de broglie wavelength .
given the inequality ( [ dilute ] ) , the condition ( [ lambda_t_vs_r_0 ] ) is satisfied _ automatically _ as long as one is interested in essentially quantum properties of the system , implying @xmath11 .
an extension of the theory to the boltzmann high - temperature regime @xmath12 , where answers become sensitive to the particular form of the interaction potential , goes beyond the scope of this paper .
the paper is organized as follows . in section [ sec : summary ]
we first summarize the key results of our approach for thermodynamic quantities and correlation functions in a form easily accessible to the general reader .
the next chapters systematically introduce the technique and derive the results . in section [ sec : diagrams ] we re - derive the diagrammatic technique for the phase with the broken u(1 ) symmetry by introducing explicit ( arbitrary small ) symmetry breaking terms in the hamiltonian .
we keep the discussion as general as possible to cover inhomogeneous , e.g. trapped , systems and states with non - zero average current . in section [ sec : diagrams ] we also derive the central expressions for the proper self energies and chemical potential . in section [ sec : quasicondensate ] we consider thermodynamics ( including the superfluid density ) of a homogeneous system in the superfluid phase . in particular , we introduce the notion of the renormalized chemical potential @xmath13 and show that @xmath13 and temperature form a natural pair of thermodynamic parameters , in terms of which all the other quantities are obtained from the single - particle momentum - space integrals . in section [ sec : estimates ] we discuss , at the order - of - magnitude level , the structure of higher - order corrections to the spectrum of elementary excitations and thermodynamic quantities .
this analysis , in particular , leads to the conclusion that within the pseudo - hamiltonian approximation and bogoliubov ansatz for the spectrum of elementary excitations our results can not be further improved . in section [ sec : off ]
we show that the answers for the asymptotic long - range behaviour of off - diagonal correlation functions are readily obtained on top of beliaev s diagrammatic treatment for medium - range correlations , by employing the generic hydrodynamic description of a superfluid . the short section [ sec : normal_region ] deals with the normal region . in section
[ sec : pseudo ] we show that in both normal and superfluid regimes ( but away from the fluctuation region around the critical point ) the thermodynamic relations can be associated with a pseudo - hamiltonian in the following sense .
the partition function and the entropy of the system turn out to correspond to those of a non - interacting bosonic hamiltonian with temperature - dependent parameters and a temperature - dependent global energy shift . in section [ sec : fluct ]
we render the results for the fluctuation region , which previously were obtained in 2d and 3d , but not in 1d . in section
[ sec : numerics ] we finally compare our solutions with first - principle monte carlo data for wibg both in continuous space and on a lattice .
here we summarize the most important relations which we derive in the main part of the paper , assuming that the conditions ( [ dilute ] ) and ( [ lambda_t_vs_r_0 ] ) are met .
the results of this section work for both continuous - space and lattice systems ; in the latter case , the mass @xmath14 is understood as the effective mass of low - energy motion .
the equilibrium state of the system is conveniently parametrized by two independent variables , the temperature @xmath4 and the effective chemical potential @xmath15 . the latter is always negative , so that @xmath16 .
the interaction is characterized by an ( effective ) coupling constant @xmath17 . in one dimension
@xmath17 is the zero - momentum fourier component of the bare potential @xmath18 . in two and three dimensions , @xmath17
is naturally expressed in terms of the @xmath3-wave scattering length @xmath2 ; the latter being defined as the radius of hard disk / sphere potential @xmath19 with low - energy scattering properties identical to those of the potential @xmath18 .
a representational freedom allows one to use slightly different expressions for @xmath17 , without sacrificing the order of accuracy .
our particular choice is @xmath20 in three dimensions , and @xmath21 in two dimensions , where @xmath22 is euler s constant .
a particular expression for @xmath17 comes from fixing the specific form of an auxiliary function @xmath23 .
the latter enters thermodynamic relations in such a way that the values of those remain insensitive up to irrelevant higher - order corrections to slightly changing @xmath17 by adopting a slightly different @xmath24 .
equations ( [ u_3d_gas ] ) and ( [ u_2d_gas ] ) correspond to @xmath25 @xmath26 respectively , where @xmath27 is the particle dispersion relation . in one dimension @xmath28 . within the representational freedom , different choices of the @xmath29-pair
are straightforwardly connected with each other by ( [ u1_u2])-([c_12 ] ) . in the 2d case , the coupling constant @xmath17 , and , correspondingly ,
function @xmath30 slowly evolve with changing effective chemical potential and temperature : @xmath31 it is _ well _ within the representational freedom to multiply the r.h.s . of ( [ epsilon_0_mu_t ] ) by a constant of order unity ,
provided the same expression for @xmath32 is used in both ( [ u_2d_gas ] ) and ( [ pi_2d_case ] ) .
the only reason for introducing the order - unity factor @xmath33 is to cast the zero - temperature equations of state in the form ( [ 2d_ground_1])-([2d_ground_2 ] ) adopted previously by some authors . in the case
when the problem of finding the scattering length @xmath2 is not straightforward , one can resort directly to the integral equation ( [ eq : def_of_u ] ) defining @xmath17 in terms of @xmath34 and @xmath24 .
numeric solution of this equation can be obtained within bold diagrammatic monte carlo technique @xcite , which gives the scattering length as a by - product of the calculation .
thermodynamic quantities are obtained in parametric form , as functions of the pair of parameters @xmath15 and @xmath4 .
the main relation is for the number density : @xmath35 @xmath36 \ , , \label{n_prime_rel}\ ] ] where @xmath37 } \label{bog_disp}\ ] ] is the bogoliubov - type quasiparticle dispersion relation and @xmath38 is the bose - einstein occupation number . in equation ( [ n_prime_rel ] ) and below throughout the text where we sum over momenta
, we set the system volume equal to unity thus omitting a trivial normalization factor .
the genuine chemical potential is then given by @xmath39 the pressure and entropy density are obtained as @xmath40 \right\ } , ~\label{pressure}\end{aligned}\ ] ] and @xmath41 , \label{entropy}\end{aligned}\ ] ] from which the energy density follows from the general relation @xmath42 .
at zero temperature the integrals can be evaluated analytically , as discussed in subsection [ subsec : practical ] . in continuous space , where @xmath43
, the superfluid density is given compactly by landau s formula @xmath44 due to the condition ( [ dilute ] ) the answer for a lattice differs _ only _ by a global factor equal to the ratio between the bare and effective masses . in the superfluid region
, the single - particle density matrix can be parameterized as @xmath45 with @xmath46 , \label{tilde - rho}\ ] ] @xmath47 \ , .
\label{lambda_rel}\ ] ] by definition , @xmath48 is the condensate density .
the function @xmath49 is finite in any dimensions . in 1d and at finite temperature in 2d
, the function @xmath50 diverges at @xmath51 , consistently with the general fact of absence of condensate in those systems .
nevertheless , at distances at which the function @xmath52 saturates to its asymptotic value @xmath53 , the function @xmath50 is still much smaller than unity .
this characteristic feature of the weakly interacting system allows one to speak of the quasi - condensate with the density given by @xmath53 . up to the distances
at which @xmath54 becomes @xmath55 , the correlation properties of condensed and quasi - condensed systems are essentially the same .
details of the general approach to long - range off - diagonal many - particle correlation functions are described in section [ sec : off ] .
the density @xmath56 in the normal region is given by @xmath57^{-1 } \
, , \label{n_normal_rel}\ ] ] with @xmath58 the pressure is obtained by @xmath59 \ ; . \label{p_normal_rel}\ ] ] for the entropy and energy densities we have the relations @xmath60 \ , , \label{s_nrml_rel}\ ] ] @xmath61 a necessary condition for the above - outlined relations to apply is the smallness of the parameter @xmath62 at @xmath63 , the parameter @xmath64 directly controls the systematic error of the theory , and the condition @xmath65 is sufficient for the theory to be accurate . at higher temperatures ,
the condition ( [ gamma_0_summary ] ) is only _ necessary _ , since there exists the so - called fluctuation region where the fluctuations of the order parameter are essentially non - linear and are not captured by our theory .
the closeness to the fluctuation region is described by the dimensionless parameter @xmath66 in 2d and 3d systems , @xmath67 is the critical value of the chemical potential for a given temperature ( for the explicit expressions , see section [ sec : fluct ] ) , in 1d systems , where there is no finite - temperature phase transition ( see subsection [ subsec : specifics_1d ] for more detail ) , @xmath68 .
the theory applies as long as @xmath69 , getting progressively less accurate with decreasing @xmath70 . at @xmath71
, the theory fails to properly describe condensate and superfluid densities the values of which being defined by the fluctuating classical - field order parameter .
the description of other thermodynamic quantities is better since the fluctuation contributions to them are of the higher order than the leading ( and sometimes even sub - leading ) terms .
the estimates for systematic errors for major thermodynamic quantities away from and within the fluctuation region are given in sections [ sec : estimates ] ( in the superfluid region ) and [ subsec : exp_par_norm ] ( in the normal region ) . in the fluctuation region and its vicinity , the theory can be improved by incorporating accurate description of fluctuation contributions by
universal to all weakly interacting u(1 ) models and available in literature
dimensionless scaling functions .
the description is outlined in section [ sec : fluct ] .
the diagrammatic technique for bosons can be derived in terms of functional integrals over the classical complex - valued fields @xmath72 , propagating in the imaginary time from @xmath73 to @xmath74 ( we set @xmath75 ) , and subject to the @xmath76-periodicity constraint with respect to the variable @xmath77 @xcite .
the classical - field grand - canonical hamiltonian has a form @xmath78 where @xmath79 is the ideal - gas term , with @xmath14 the particle mass , and @xmath80 is the pairwise interaction term .
the @xmath81 term contains linear and bilinear terms associated with the chemical potential , @xmath82 , external potential , @xmath83 , and the field @xmath84 which explicitly breaks the u(1 ) symmetry , @xmath85 |\psi|^2 { \rm d}{\bf r } \ , - \int [ \eta^*\psi + { \rm c.c . } ] \ , { \rm d}{\bf r } \ , .
\label{h_1}\ ] ] for a homogeneous system @xmath86 . strictly speaking , at finite @xmath87
one can not refer to @xmath88 as a chemical potential because the total number of particles is not conserved .
however , we employ @xmath87-terms solely for explicitly breaking the symmetry and stabilizing supercurrent states ; at the end of the calculation we take the limit @xmath89
. we will therefore continue referring to @xmath82 as the chemical potential .
standard finite - temperature treatments of the wibg typically suffer from infra - red divergencies @xcite , especially in lower dimensions .
the ultra - violet divergence is removed by introducing the notion of the pseudopotential which allows one to express final answers in terms of the scattering length alone .
our treatment successfully and systematically deals with divergence issues , and features ( at least ) three important considerations : + - first , the derivation is based on the appropriate definition of the pseudopotential @xmath90 for low momenta [ see the discussion of equation ( [ u_of_k ] ) below ] in any dimension .
+ - second , the u(1)-symmetry breaking field @xmath87 introduces a gap in the spectrum of goldstone modes in a well - controlled way .
this suppresses long - range phase fluctuations of the order parameter and removes infra - red divergencies in the behaviour of correlation functions and self - energies without modifying any of the physically important quantities in the relevant order of approximation . in lower dimensions , finite @xmath87 results in a finite value of the genuine condensate density which dramatically simplifies the description . in this respect ,
a small @xmath87 does essentially the same job as berezinskii s finite - size trick , cf .
@xcite , or popov s special ( hydrodynamic ) treatment of long - wave parts of the fields @xcite . in all final answers
one can set @xmath91 .
+ - third , the effects of _ both _ interaction and chemical / external potential are fundamentally non - perturbative separately below the critical temperature . in the absence of interactions
a positive chemical potential immediately leads to the instability of the ideal bose gas , so that there is no way of consistently introducing a positive @xmath82crucial for describing the system below the critical point before switching on the interaction .
this observation explains the idea behind omitting all terms in ( [ h_1 ] ) from the non - interacting hamiltonian , @xmath92 . at the end of the day
, we will use @xmath93 and @xmath4 as two variables describing the thermodynamic ensemble .
since the non - interacting hamiltonian corresponds to the ideal gas at chemical potential approaching zero from below , i.e. with no bose - einstein condensate , the corresponding ( bare ) matsubara green s function reads @xmath94 @xmath95^{-1}\ , , \label{g_0_k}\ ] ] where @xmath27 is the single particle dispersion relation .
we typically assume a parabolic dispersion relation @xmath43 , but our final answers ( excluding formulae for superfluid properties which explicitly invoke galilean invariance ) are valid for arbitrary @xmath27 with a parabolic dependence on momentum in the long - wave limit , e.g. for the tight - binding spectrum .
we use the matsubara imaginary - frequency representation : @xmath96 ( @xmath97 ) with @xmath4 the temperature . for graphical representation of diagrammatic expansions and relations we introduce a set of objects in figure [ fig : hamiltonian ] depicting the single - particle propagator , the condensate , and various terms in the hamiltonian . the non - perturbative response to @xmath98 and @xmath81
is accounted for by dyson summation .
first , we consider diagrams that feature only one incoming or outgoing line we call them tails . given our starting point with no condensate in the non - perturbed system such particle - number - changing diagrams exist only due to the symmetry breaking field @xmath87 .
the frequency of the line connecting @xmath87 to the rest of the diagram is zero , by frequency conservation .
the dyson summation of all tails attached to a given point replaces them with a single line which we denote as @xmath99 and @xmath100 , if the tails are incoming and outgoing , respectively , see figure [ fig : dyson_psi ] .
below we write expressions in the frequency representation ; if the frequency argument is not mentioned it is implied that its value is zero .
we also adopt a convention of integration over repeated coordinate / momentum / frequency arguments . the dyson equation for @xmath99
then reads : @xmath101\psi_{\rm in}({\bf r}')+g^{(0)}({\bf r } - { \bf r } ' ) \theta_{\rm in}({\bf r}')\ , .
\label{psi_in}\end{aligned}\ ] ] here @xmath102 is the sum of all other diagrammatic elements attached to the first line which are not accounted for by the first two terms , i.e. excluding diagrams with the field @xmath87 and diagrams connected to the first solid line by the @xmath103 $ ] vertex .
the subscript ` in ' reminds that @xmath102 has an extra incoming particle line .
similarly , @xmath104g^{(0)}({\bf r}'- { \bf r } ) + \theta_{\rm out}({\bf r}')g^{(0)}({\bf r}'- { \bf r } ) \ , .
\label{psi_out}\end{aligned}\ ] ] the fact that @xmath105 is a real even function of its argument implies that @xmath106 the diagrammatic expansion for the tail is identical to that for the condensate wave function defined as an anomalous average @xmath107 correspondingly , the condensate density is defined as @xmath108 . from now on we will write @xmath109 for @xmath110 and @xmath111 for @xmath112 , and use the notions of condensate lines and tails on equal footing . of course , in the limit of @xmath113 speaking of the condensate wave - function is meaningful only when the long - range fluctuations of the phase are small .
the next step is to introduce exact , or ` bold ' , particle propagators and work with skeleton diagrams .
there are a couple of differences between our approach and the standard beliaev technique .
our bare ( thin - line ) propagators have zero chemical potential and this , according to ( [ psi_in ] ) , immediately results in a constraint relating the condensate density to the chemical potential and @xmath102 [ see ( [ chem_pot_rel ] ) below ] .
also , the chemical / external potential have to be explicitly introduced into the standard beliaev - dyson equations for the normal and anomalous green s functions , see e.g. , @xcite .
these equations can be defined purely diagrammatically . to this end
proceeding in the frequency representation for the sake of definiteness we introduce the normal green s function , @xmath114 , defined as the sum with a global minus sign , which is a mere convention of all diagrams which have an incoming @xmath115-line with frequency @xmath116 to point @xmath117 and an outgoing @xmath115-line ( with the same frequency ) from point @xmath118 .
the anomalous green s function , @xmath119 , by definition , has an incoming @xmath115-line with frequency @xmath116 to point @xmath120 and another incoming @xmath115-line with frequency @xmath121 , by conservation of frequency , to point @xmath122 .
the anomalous green s function @xmath123 is a counterpart of the function @xmath124 : instead of two incoming it has two outgoing @xmath115-lines ; one with frequency @xmath116 from point @xmath120 and another with frequency @xmath121 from point @xmath118 .
the symmetry under exchanging the end points of the anomalous green s functions immediately implies the following relations @xmath125 @xmath126 since complex conjugation is equivalent to changing the sign of the matsubara frequency and direction of propagation we also have @xmath127^*\ , = \ , g(-\xi , { \bf r}_2 , { \bf r}_1 ) \ ; , \label{sym3}\ ] ] @xmath128 the physical meaning of @xmath129 , @xmath130 , and
@xmath131 follows from the structure of the two - point correlation functions in the imaginary - time coordinate representation : @xmath132 @xmath133 these relations can be readily checked by expanding the averages into diagrammatic series .
the special case of ( [ g_phys ] ) , corresponding to @xmath134 and @xmath135 , relates the local density to the normal green s function and the condensate density : @xmath136 the beliaev - dyson equations then read @xmath137g^{(0)}(\xi,{\bf r}',{\bf r}_2 ) \nonumber \\
+ g(\xi,{\bf r}_1,{\bf r}')\ , \sigma_{11}(\xi,{\bf r}',{\bf r}'')\ , g^{(0)}(\xi,{\bf r}'',{\bf r}_2 ) \nonumber \\ + \ , f_{\rm in}(\xi,{\bf r}_1,{\bf r}')\ , \sigma_{20}(\xi,{\bf r}',{\bf r}'')\ , g^{(0)}(\xi,{\bf r}'',{\bf r}_2 ) \ , , \label{db1}\end{aligned}\ ] ] @xmath138g^{(0)}(-\xi,{\bf r}_2,{\bf r } ' ) \nonumber \\ + \ , f_{\rm in}(\xi,{\bf r}_1,{\bf r}')\ , \sigma_{11}(\xi,{\bf r}',{\bf r}'')\ , g^{(0)}(-\xi,{\bf r}_2,{\bf r } '' ) \nonumber \\
+ \ , g(\xi,{\bf r}_1,{\bf r}')\ , \sigma_{02}(\xi,{\bf r}',{\bf r}'')\ ,
g^{(0)}(-\xi,{\bf r}_2,{\bf r } '' ) \ , , \label{db2}\end{aligned}\ ] ] with the standard definition of self - energies @xmath139 s as sums of diagrams which can not be cut through a single @xmath129 or @xmath140 line . equations ( [ db1 ] ) and ( [ db2 ] ) are shown graphically in figure [ fig : beliaev - dyson ] for a homogeneous system in momentum representation .
the complexity of the theoretical solution is in the evaluation of the @xmath141 and @xmath139 functions . since the bare green s function at zero frequency is identical to the inverse laplacian operator one can cast ( [ psi_in ] ) in the differential form @xmath142\ , \psi_0({\bf r})+ \theta_{\rm in}({\bf r } ) \ , = \ , \eta({\bf r})\ ; . \label{psi_in_dif}\ ] ] this equation reduces to the gross - pitaevskii equation at low enough temperature when the leading term in @xmath102 is @xmath143 . in the homogeneous case at @xmath144 , when @xmath109 and @xmath145 are
coordinate - independent , equation ( [ psi_in_dif ] ) simplifies to @xmath146 generalizing relations ( 21.3)-(21.5 ) , and @xmath147 , in @xcite to finite temperatures .
there is also an exact relation between @xmath148 , @xmath149 , @xmath102 , and @xmath109 ( see also @xcite ) . here
( and only here ! )
we assume that all diagrams for @xmath139 s are in terms of @xmath115 and condensate lines .
let @xmath150 be the sum of diagrams contributing to @xmath102 with @xmath151 incoming and @xmath152 outgoing condensate lines . then ( for @xmath153 ) @xmath154 because each diagram , with @xmath152 incoming condensate lines , contributing to @xmath155 produces upon integration over @xmath156 with the weight @xmath157a diagram contributing to @xmath158 , and there are @xmath151 such diagrams contributing to @xmath148 .
an identical argument leads to @xmath159 by subtracting ( [ rel_sigma_02 ] ) from ( [ rel_sigma_11 ] ) we obtain @xmath160 in the homogeneous case @xmath161 . then , in the @xmath113 limit , equations([combin ] ) and ( [ chem_pot_rel ] ) can be combined to yield the hugenholtz - pines relation @xcite ( and its finite - temperature version @xcite ) @xmath162 in order to discuss the superfluid properties of the homogeneous system , we add a phase factor to the symmetry breaking field @xmath163 which readily translates into phase of the condensate wave function @xmath164 the only difference with the previous discussion is that now we have to associate a finite momentum @xmath165 carried by the condensate lines and modify the momentum conservation laws accordingly . then ( [ chem_pot_rel ] ) and ( [ hp ] ) become @xmath166 @xmath167 where @xmath168 .
it is convenient ( for transparency of expressions which follow ) to combine frequency and momentum into a single `` 4-momentum '' variable @xmath169 and to introduce an auxiliary momentum @xmath170 .
the symmetry between the two ends of the anomalous green s functions and , equivalently , between the two ends of the anomalous self - energies is then expressed by ( accounting for the momentum carried by the condensate lines ) @xmath171 @xmath172 [ in a more comprehensive notation scheme one has to mention momenta of both incoming lines in @xmath173 . ]
complex conjugation of propagators and condensate lines changes the signs of their 4-momenta .
this property can be used to prove the symmetry relation for the green s function @xmath174 similar symmetry relations take place for the anomalous green s functions and all three self - energies . in the momentum representation ,
inverting the direction of @xmath175 does not change the analytical expression for propagators .
hence , by inverting the direction of all the lines , including the condensate ones , it follows that @xmath176 @xmath177 ( for the normal green s function , @xmath178 , and the self - energy , @xmath179 , inverting momentum directions results in the same series , and in this sense is trivial . )
we are ready to formulate the pair of beliaev - dyson equations in the momentum representation .
using symmetry properties into consideration and the shorthand notation @xmath180 we obtain @xmath181g^{(0)}(p ) \nonumber \\
+ f(p)\tilde{\sigma}(p)g^{(0)}(p)\ , , \\
\label{db1_p } f(p ) = f(p)[\sigma(p')\ ! -\ ! \mu]g^{(0)}(p ' ) + g(p)\tilde{\sigma}(p)g^{(0)}(p')\ , .
\label{db2_p}\end{aligned}\ ] ] the solution in terms of self - energies reads @xmath182 @xmath183 where @xmath184[\epsilon(k ) + \sigma(p)-\mu -{\rm i}\xi ] .~ \label{denom}\ ] ] with these relations at hand , one can calculate the current density induced by the phase gradient in the condensate wave - function , see subsection [ subs : sd ] . in two and three dimensions , the expansion in terms of the bare interaction potential can be ( and in most realistic cases , is ) non - perturbative .
the system is regarded as weakly - interacting only because of the low density of particles . in one dimension ,
the physics is perturbative instead in the high - density limit for a given potential .
theoretically , dealing with the strong bare potential implies summation of an infinite sequence of ladder diagrams and produces an effective interaction in the form of the four - point vertex @xcite , @xmath185 , see figure [ fig : ladder ] .
the analytical relation behind figure [ fig : ladder ] reads @xmath186 when the bare - interaction lines are replaced with @xmath185 s , the rest of the series becomes perturbative ( excluding the critical region ) . on the technical side ,
working with @xmath185 s is convenient as long as one does not intend to systematically take into account higher - order corrections , utilizing thus only the ( simple and transparent ) leading - order expression for @xmath187 . for the higher - order corrections ,
equation ( [ eq : gamma ] ) involves three different length scales : the size of the potential , @xmath6 , the healing length of the bose condensed system , and the de broglie wavelength , while only the non - perturbative physics at the scale @xmath6 requires summation of the ladder diagrams .
hence , it makes sense to construct an object with a much simpler structure than @xmath185 that captures the non - perturbative physics ( and thus coincides with the leading - order expression for @xmath187 ) , and then systematically investigate the difference between this object and @xmath185 . to achieve this goal
let us define the _ pseudo - pontential _
@xmath188 by the equation ( see figure [ fig : pseudo_potential ] ) @xmath189 where @xmath190 is such that when @xmath191 is much larger than the inverse healing length or thermal momentum the value of @xmath190 approaches that of the product @xmath192 summed over the frequency of the 4-vector @xmath193 .
that is , @xmath194 in this case , @xmath195 , and one can expand the difference in a perturbative series . as a result
, we arrive at the diagrammatic technique where instead of thin dashed lines standing for the bare interaction potential we have bold dashed lines representing the pseudo - potential @xmath17 , with an additional requirement that whenever two ( normal ) green s function lines are sandwiched between two pseudo - potential lines , the former are supposed to be summed over the frequency difference ( which is always possible in view of the frequency - independence of the pseudo - potential ) , and then @xmath24 has to be subtracted from the result of summation .
specifically , if @xmath196 and @xmath197 are the two external 4-momenta of the above mentioned diagrammatic element , then the following replacement is supposed to take place for internal propagator lines @xmath198 where @xmath199 is the frequency of the 4-momentum @xmath193 . as long as we are not interested in the non - universal ultraviolet corrections
, we can replace @xmath188 with @xmath200 , the systematic error introduced by the replacement being controlled by the following dimensional estimate @xmath201\ , .
\label{u_of_k}\ ] ] one may wonder how this estimate is reconciled with the momentum dependence of @xmath185 , which is of the order @xmath202 in 3d , and of the order @xmath203 in 2d .
the solution is provided by ( [ corrected ] ) explaining that this dependence , which is both universal and perturbative , is taken into account by the second - order ladder - type diagram in @xmath17 . in 3d ,
a natural choice for @xmath204 is @xmath205 in which case we have the usual @xcite ( see also below ) @xmath206 with @xmath2 the @xmath3-wave scattering length . in 2d
one can not use ( [ pi_3d ] ) because of the infra - red divergence of the integral in ( [ eq : def_of_u ] ) .
a reasonable choice here is @xmath207 where @xmath32 is an arbitrary low - energy cutoff .
the particular value of @xmath32 is not important since final answers are not sensitive to it . within the systematic - error bars of the pseudo - hamiltonian description discussed below
, any choice of @xmath32 such that @xmath208 is equally reasonable in terms of accuracy , provided the temperature is not much larger than @xmath209 ; moreover , replacing @xmath32 with @xmath210 is also acceptable .
( an optimal choice of @xmath32 in the regime @xmath211 will be discussed in subsection [ subsec:2d_large_t ] . )
there is also no need to introduce @xmath24 in @xmath212 .
we will assume that formally @xmath213 in @xmath212 in which case our final answers can be used as written in all spatial dimensions .
given that ( [ pi_3d ] ) and ( [ pi_2d ] ) are not unique [ there is a free parameter in ( [ pi_2d ] ) ] , it is instructive to explicitly relate two pseudo - potentials , @xmath214 and @xmath215corresponding to @xmath216 and @xmath217 , respectively to _ each other_. we notice that ( [ eq : def_of_u ] ) implies @xmath218 u_2({\bf k } ) .~ \label{rel_u1_u2}\end{aligned}\ ] ] in view of ( [ u_of_k ] ) this simplifies [ up to @xmath219 terms that we neglect in what follows ] to @xmath220 @xmath221 \ , .
\label{c_12}\end{aligned}\ ] ] if @xmath216 and @xmath217 are defined by ( [ pi_2d ] ) with different cut - offs @xmath222 and @xmath223 , we have @xmath224 the right choice of the functions @xmath216 and @xmath217 implies that @xmath214 and @xmath215 differ only by sub - leading terms .
thus , we can expand the r.h.s . of ( [ u1_u2 ] ) in powers of @xmath225 : @xmath226 formally , the expansion in terms of the pseudo - potential is perturbative only as long as we exclude the high - momentum contributions to the @xmath227-body diagrams generating , upon complete summation , @xmath228-body scattering amplitudes . in a dilute gas ,
the corresponding diagrams are small ( contain extra powers of the gas parameter @xmath229 ) and are neglected in this manuscript .
the expression ( [ u_3d ] ) has the meaning of mapping a dilute three - dimensional system with an arbitrary short - range interaction potential onto a system with the interaction potential ( [ hard_pot ] ) , so that the pseudo - potentials of the two systems coincide .
the same approach is possible ( and popular ) in 2d , the parameter @xmath2 being called the two - dimensional scattering length , since the mapping applies to scattering properties as well . for a given potential @xmath34 , the value of @xmath2 can be obtained either from the asymptotic behaviour of the pseudo - potential @xmath17 at appropriately small @xmath32 , or from the asymptotic behaviour of the scattering amplitude @xmath230 at @xmath231 @xcite . both asymptotic behaviours are closely related to each other since the large-@xmath191 limit of the kernel @xmath190 in ( [ eq : def_of_u ] ) coincides with the large-@xmath191 limit of the scattering kernel @xmath232 in the integral equation for @xmath230 , @xmath233 moreover ,
a direct relationship between @xmath17 and @xmath230 is readily obtained by noticing that ( [ eq : def_of_t ] ) and ( [ eq : def_of_u ] ) imply [ cf .
( [ rel_u1_u2 ] ) ] @xmath234 f({\bf q } , { \bf k}')\ , , \label{rel_u_t}\end{aligned}\ ] ] which dramatically simplifies upon replacing @xmath188 with @xmath200 , in accordance with ( [ u_of_k ] ) . in this case
, @xmath230 is @xmath235-independent , and for @xmath236 we get @xmath237 \ , .
\label{u_f_smpl}\end{aligned}\ ] ] substituting ( [ pi_2d ] ) and ( [ scatt_krnl ] ) into ( [ u_f_smpl ] ) and performing the integral , we find @xmath238 comparing this to the known hard - disk result ( see , e.g. , @xcite ) @xmath239 where @xmath22 is euler s constant , we conclude that @xmath240 in 3d , equation ( [ u_f_smpl])with @xmath241 given by ( [ pi_3d])yields @xmath242 thus leading to ( [ u_3d ] ) by comparison with known expression for the hard - sphere scattering amplitude .
explicitly calculating the lowest - order diagrams shown in figure [ fig : sigma11 ] and utilizing ( [ g_phys_special ] ) one finds @xmath243 @xmath244 we see that within the first approximation , both @xmath139 and @xmath245 turn out to be momentum- and frequency - independent .
[ it is easy to check that the next - order diagrams inevitably introduce momentum and frequency dependence to self - energies and drastically change the structure of the theory . ] at this level of accuracy , the chemical potential equals to @xmath246 according to the hugenholtz - pines relation . as mentioned above , it is extremely convenient to use an effective chemical potential @xmath247 as a thermodynamic variable to characterize properties of the wibg .
[ note that @xmath248 is negative . ] within the same accuracy we can substitute @xmath245 with @xmath249 and simplify expressions for @xmath129 and @xmath140 to @xmath250 @xmath251 @xmath252 \ , .
\label{e2}\ ] ] a note is in order here .
below we will calculate higher - order corrections to the chemical potential which are necessary for the construction of the accurate equation of state , see ( [ mu_dgr ] ) and ( [ rel_delta ] ) .
however , it is still possible to use @xmath15 in the definitions of the propagators and the spectrum of elementary excitations while keeping the accuracy of the entire scheme intact , see , subsection [ sec : estimates ] where we study systematic errors involved in approximations .
here we mention briefly that the anomalous average contribution to thermodynamic properties is always small , and thus further corrections to @xmath140 are negligible .
the same is true for @xmath129 at low temperature . at temperatures @xmath253 ,
on the other hand , one can ignore tiny modifications in the spectrum because of an additional small parameter @xmath254 .
for the total density we have @xmath255 , \label{n_total}\end{aligned}\ ] ] where @xmath256 is the bose - einstein occupation number for the mode with the energy @xmath257 . formula ( [ n_total ] ) includes the leading and sub - leading terms . to get an expression for the chemical potential with the same degree of accuracy , we need to take into account higher - order diagrams and go beyond the leading - order expressions for @xmath139 and @xmath245 .
this is because in the absence of interaction the chemical potential is identically zero and thus the leading term is proportional to @xmath17 .
there are three second - order diagrams contributing to @xmath141 .
the first one is the anomalous green s function convoluted with the bare interaction potential ( see figure [ fig : mu ] .
the second one is the ` sunrise ' diagram with the proper correction ( [ corrected ] ) for the ladder structure and the third one is similar to the sunrise diagram but with propagators connecting different vertices ( see the two diagrams in the second row in figure [ fig : mu ] ) .
the latter two can be safely neglected away from the fluctuation regime because they involve an additional small parameter @xmath64 or @xmath258 , see section [ sec : estimates ] for the analysis and definitions of the fluctuation region and parameters @xmath259 in different spatial dimensions .
it is worth noting that consistently taking into account contributions of the two neglected diagrams in the condensed regime would require simultaneously going to higher orders in the expansions for self - energies , figure [ fig : sigma11 ] .
the latter , however , is impossible without sacrificing the attractive paradigm of independent quasiparticles with bogoliubov dispersion .
keeping the leading and the largest sub - leading diagrams for @xmath141 , we get @xmath260 we have no choice but to use the bare potential here because all ladder diagrams leading to the pseudopotential vertex are already absorbed in the anomalous green s function , by construction of the latter .
this feels unsatisfactory only at first glance since simple formal manipulations allow one to express ( [ mu_dgr ] ) in terms of @xmath17 alone .
let us introduce an auxiliary function @xmath261 defined by the integral equation ( shown graphically in figure [ fig : delta ] ) @xmath262 it can be used , in combination with the definition of the pseudopotential eq.([eq : def_of_u ] ) , to transform the last two terms in ( [ mu_dgr ] ) @xmath263 where @xmath264 the last approximate equality in ( [ rel_delta ] ) comes from @xmath265 and the observation that @xmath261 vanishes at momenta much smaller than @xmath266 .
the smallness of the parameter @xmath267 at momenta @xmath268 allows one to expand @xmath269 [ see ( [ delta ] ) ] in the series : @xmath270 + \ldots \ ; . \label{delta_series}\ ] ] for the effective chemical potential we thus get @xmath271 where within our order of accuracy we can take @xmath272 \ ; . \label{delta_0_actual}\ ] ] we find it convenient to introduce a quantity with the dimension of density , @xmath273 with this quantity the form of certain thermodynamic relations simplifies .
for example , using ( [ qc_def ] ) we can replace @xmath274 with @xmath275 in ( [ n_total ] ) and arrive at the result @xmath276 @xmath277 \ , .
\label{n_prime}\ ] ] this completes the self - consistent theory because we obtain a closed set of relations which define @xmath278 in the parametric form given some @xmath15 , or @xmath275 , one calculates @xmath1 from ( [ n_total_qc ] ) and ( [ n_prime ] ) and then determines @xmath82 from ( [ eff_mu ] ) .
the integral in ( [ n_prime ] ) is convergent not only in 3d , but also in 2d and 1d .
hence , at this point the quantity @xmath87 can be set equal to zero .
we want to emphasize the fact that all specific expressions derived in this section feature a two - parametric representational freedom , within the same order of accuracy .
first , it is possible to use @xmath15 with or without higher - order corrections in the spectrum of elementary excitations and propagators .
second , there is a freedom of choosing the function @xmath23 .
for example , one could be tempted to absorb the first two terms in the integral in ( [ n_prime ] ) into the definition of @xmath17 , such that @xmath279 gets identically equal to zero , and @xmath280 equals to the total number of particles .
we , however , do not see any merit in this protocol , because @xmath17 becomes dependent on @xmath15 and can not be considered as a fixed external parameter .
finally , this and analogous `` improvements '' do not make the theory more / less accurate since the difference is of the same order as omitted diagrams .
a remark is in order here concerning the two - dimensional case , where the value of @xmath17 can not be defined irrespectively of the system density . even in this case
, we can proceed with formally independent parameter @xmath32 in ( [ pi_2d ] ) till we arrive at the final answers along with the estimates of neglected higher - order terms ( the latter being essentially @xmath32-dependent ) .
after that , the value of @xmath32 is selected in such a way that the order - of - magnitude values of neglected terms are minimal . to derive relations for other thermodynamic properties , we start with the pressure as a function of @xmath4 and @xmath15 . using the general thermodynamic formula @xmath281 and adopting throughout this subsection the convention that @xmath4 is treated as a fixed constant , so that partial derivatives with respect to either @xmath82 or @xmath15 can be replaced with ordinary ones , we write @xmath282 where @xmath283 \ ; \label{p_c}\ ] ] is the value of pressure at the mean - field critical density @xmath284^{-1 } \ ;
. \label{n_c}\ ] ] equations ( [ eff_mu ] ) and ( [ n_total_qc ] ) allow one to represent ( [ p1 ] ) as @xmath285 ( after doing straightforward integrations by parts ) .
the integral in ( [ p2 ] ) is readily done by noticing that @xmath286 @xmath287 \ , , \label{obs_2}\ ] ] and thus @xmath288 \nonumber \\ - t { { \rm d}\over { \rm d}\tm } \ , \sum_{\bf k}\ , \ln \left[1-{\rm e}^{-e(k)/t } \right ] \ , , \label{obs_3}\end{aligned}\ ] ] where @xmath27 makes the first term convergent .
the result of the integration is @xmath289 and we finally get @xmath290 \right\ } \ , . \label{p3}\end{aligned}\ ] ] whenever @xmath1 , @xmath82 , and @xmath291 are specified as functions of @xmath292 , where @xmath293 is a quantity of arbitrary nature , the expressions for entropy per unit volume , @xmath3 , and energy per unit volume , @xmath257 , are readily found from the following two generic thermodynamic relations : @xmath294 @xmath295 with @xmath296 we thus find @xmath41 \ , .
\label{s_cnd}\end{aligned}\ ] ] @xmath297 \ , .
\label{e_cnd}\end{aligned}\ ] ] it is useful to note that the following two integrals ( we use @xmath299 ) , @xmath300 , \label{int_1}\ ] ] @xmath301 = { \partial i_1^{(d ) } \over \partial |\tm| } , \label{int_2}\ ] ] entering relations for thermodynamic quantities see subsections [ subsec : basic]-[subsec : entropy]can be explicitly performed .
[ here we assume that in 3d and 2d the kernel @xmath302 is fixed by expressions ( [ pi_3d ] ) and ( [ pi_2d ] ) , respectively , and remind that in 1d it is zero .
] explicitly doing the integrals is especially useful at @xmath298 . in this case , all other integrals in the expressions for thermodynamic functions nullify , and the answers reduce to algebraic relations . in the grand - canonical form , these relations read @xmath303 @xmath304 @xmath305 here we take into account that within our level of accuracy we can replace @xmath306 in the arguments of @xmath307 and @xmath308 , since the integrals are responsible for sub - leading corrections , while the sub - sub - leading terms are beyond our control . performing the integrations , one finds : @xmath309 @xmath310 @xmath311 and differentiating with respect to @xmath312 we find @xmath313 @xmath314 @xmath315 in 2d , we have a freedom of fine - tuning @xmath32 to simplify the form of the answer .
a natural choice , especially convenient for the @xmath298 limit , is to set @xmath316 in which case we have @xmath317 at @xmath298 this translates into very compact grand - canonical expressions @xcite
@xmath318 @xmath319 with @xmath320 it is seen that the simplicity of the form of ( [ 2d_ground_1 ] ) comes at the expense of more sophisticated ( fine - tuned ) form of the effective interaction ( [ u_fine ] ) .
if , instead , one would opt to simplify the form for the effective interaction , getting rid of sub - logarithmic terms : @xmath321 , then ( [ 2d_ground_1 ] ) would acquire the generic form ( [ tm_groud])-([n_groud ] ) .
needless to note that this does not change the _ sum _ of leading and sub - leading terms in the equations of state since , by construction , the result can not depend on the specific choice of @xmath32 .
the standard way of calculating the superfluid density within the quasi - particle picture is based on landau s formula for the normal component density , @xmath322 , and the relation @xmath323 .
here we employ a more general approach based on the current induced by the gradient of the condensate wave - function . by definition ,
the superfluid density is the linear - response coefficient relating the persistent - current density to the gradient of the phase , @xmath324 , of the ( complex ) order parameter field : @xmath325 in the last equality we assumed that @xmath326 . on the other hand
, the average current can be calculated from the microscopic operator expression in terms of the condensate density and the green s function @xmath327 \right\rangle = \frac{{\bf k}_0}{m } n_0 + \sum_{\bf k}\ , \frac{{\bf k}}{m } \ : n({\bf k}_0,{\bf k } ) , \label{ns2}\ ] ] where @xmath328 in the limit of @xmath329 we obtain the required relation @xmath330 at first glance equation ( [ ns3 ] ) does not resemble landau s formula at all , since the first term is given by the condensate density , not @xmath1 . however , the structure of the normal average contribution is such that it has a part which completes @xmath274 to the total density and a part which equals ( with minus sign ) to the normal density component . to derive this result ,
we start with the expression for the green s function @xmath331 [ { \rm i } \xi -\epsilon({\bf k})+\tm ] } \ , , \label{nkk1}\ ] ] and solve for the roots of the denominator in ( [ nkk1 ] ) @xmath332\tm+\left [ \frac{\epsilon({\bf k})-\epsilon({\bf k}\ ! -\ !
2{\bf k}_0)}{2}\right]^2 } \label{nkk2}\end{aligned}\ ] ] to rewrite the green s function identically as @xmath333 @xmath334 now we express frequency sums through the bose functions to arrive at @xmath335 the rest of the calculation is straightforward . in view of the limit to be taken in ( [ ns3 ] )
it is sufficient to keep only terms linear in @xmath336 \approx 2{\bf k}\cdot { \bf k}_0/m$ ] in @xmath337 , @xmath338 , and @xmath339 .
we omit the algebra of doing the expansion which leads to the result @xmath340 . \label{ns4}\ ] ] finally , using ( [ n_total ] ) to exclude @xmath274 and integrating by parts to simplify the expression we arrive at landau s formula @xmath341
let us analyze the structure of small parameters which control the applicability of the diagrammatic technique presented above . as we will see _ a posteriori _
, it is sufficient to consider two characteristic limits : ( i ) the @xmath298 case and ( ii ) the finite-@xmath4 contributions of diagrams with zero frequencies ( i.e. , classical - field contributions ) .
the analysis is based on dimensional estimates of the diagrams , the fact that an extra interaction vertex increases the total number of propagators by two , and that the largest contributions come from small momenta ( and frequencies ) , where the normal and anomalous propagators behave as @xmath342 and we do not need to distinguish between them .
we will further assume that there is an infra - red momentum cutoff @xmath343 where @xmath344 is directly related to the inverse healing length .
the symmetry - breaking field @xmath87 also changes the behaviour of propagators at @xmath345 , but for our purposes we do not need the explicit form of propagators at finite @xmath87 because we will keep @xmath346 large enough to neglect effects originating from finite @xmath87 . at @xmath347
there are two kinds of generic vertices in higher - order diagrams , namely ( i ) full vertices , where four propagators meet , and ( ii ) vertices with one condensate line ( vertices with two and three condensate lines are all accounted for by the lowest - order diagrams for the self - energies and can not be part of the diagrammatic expansion ) . a naive definition of the diagram order as the total number of vertices proves inconvenient , since contributions of condensate vertices turn out to be larger than contributions of full vertices .
below we will see that the difference is exactly compensated by the fact that condensate vertices ( generically ) come in pairs .
this suggests to define the diagram order as the sum of the total number of full vertices and half the total number of condensate vertices ( plus @xmath348 for anomalous diagrams ) . at @xmath298 ,
the structure for the dimensionless factor associated with adding an extra full vertex to a diagram is @xmath349 ^ 2\ , , \label{full_0_rel1}\ ] ] with @xmath350 and @xmath351 representing the @xmath352-functions taking care of momentum and frequency conservation laws .
the dimensional estimate , following from ( [ full_0_rel1 ] ) and ( [ g_f_est ] ) , reads @xmath353 the structure for the dimensionless factor associated with adding an extra pair of condensate vertices is @xmath354 ^ 2 \left[|g|\ , { \rm d}^dk\ , { \rm d}\xi \right]^3\ , , \label{cnd_0_rel1}\ ] ] and from the corresponding dimensional estimate one can see that @xmath355 equations ( [ full_0_rel2 ] ) and ( [ cnd_0_rel2 ] ) clearly show the infrared problem of the theory in dimensions @xmath356 @xcite ( as usual , a zero power of the momentum cutoff should be understood as a logarithm ) , and emphasize the usefulness of the cutoff - enforcing field @xmath87 . for the diagrammatic expansion to be consistent , we need @xmath346 to be large enough in order to guarantee @xmath357 . on the other hand
, we do not want the field @xmath87 to significantly affect the physics of the system , and thus need @xmath358 .
hence , the smallest possible expansion parameter one can afford without distorting the physics corresponds to @xmath359 , in which case @xmath360 here we took into account that @xmath361 , and expressed @xmath17 in terms of @xmath2 in 3d .
it is clear that @xmath64 , equation ( [ gamma_0_rel3 ] ) , is the actual expansion parameter for all local thermodynamic quantities .
these quantities , by referring to large length scales at @xmath298 , should mainly have contributions from wavevectors @xmath362 .
technically , it means that infrared divergencies of individual diagrams have to cancel each other in the final answers and the resulting integrals become convergent at the wavevectors @xmath363 , leading to a well defined expansion in powers of @xmath64 ( [ gamma_0_rel3 ] ) .
[ since dimensional estimates do not distinguish between pure powers and powers with logarithmic pre - factors , we extend the meaning of the term ` in powers ' to the powers with logarithmic pre - factors . ]
note that all our relations between the thermodynamic quantities that do not vanish in the @xmath298 limit ( specifically , @xmath1 , @xmath82 , @xmath291 , @xmath257 , and @xmath274 ) , are accurate up to the first - order corrections in @xmath64 , so that their systematic errors are of the order of @xmath364 ( with possible logarithmic pre - factors ) .
the relation for the entropy is accurate only up to the leading term , which actually has the same quasi - particle origin as the _ sub - leading _ terms in all non - vanishing at @xmath365 quantities .
the same is true for the heat capacity that behaves similarly to entropy at @xmath365 .
the dimensionless factor associated with adding an extra full vertex to a diagram consisting of zero - frequency propagators scales as @xmath366 ^ 2\ , , \label{full_t_rel1}\ ] ] yielding the dimensional estimate @xmath367 for the pair of condensate vertices we have @xmath368 ^ 2 \left [ |g|\,t\ , { \rm d}^dk\ , \right]^3\ , .
\label{cnd_t_rel1}\ ] ] similarly to the @xmath298 case ( assuming that @xmath369 , we get @xmath370 in complete analogy with the @xmath298 case , the largest possible @xmath346 ( enforced by @xmath87 ) can not exceed @xmath371 . assuming that @xmath372 results in @xmath373 this expression shows that at @xmath374 the parameter @xmath64 dominates over @xmath258 , and thus this temperature regime is equivalent to @xmath298 . in the crossover regime ,
@xmath375 , both @xmath64 and @xmath258 are of the same order . in the regime @xmath376 , the classical - field parameter @xmath258 dominates , and becomes of order unity when @xmath274 gets sufficiently small .
the latter situation corresponds to the fluctuation region , where the perturbative theory becomes inadequate , since it fails to properly describe the non - linear long - wave fluctuations of the classical component of the quantum field . before hitting the region @xmath377
, one has to switch to the description in terms of scaling functions @xcite which are universal for all @xmath378 weakly - interacting theories , see section [ sec : numerics ] .
at @xmath376 , all our relations for the thermodynamic quantities are accurate up to the first - order corrections in @xmath258 , so that their systematic errors are of the order of @xmath379 ( with possible logarithmic pre - factors ) .
note that at @xmath376 the leading terms for all the thermodynamic quantities , except for the chemical potential ( and with a reservation for the density in 1d where the corresponding condition is @xmath380 ) , are the same as for the ideal gas . to summarize our analysis , we present below an order - of - magnitude estimate of systematic uncertainties due to omitted higher - order terms for each thermodynamic quantity .
it is convenient to start with @xmath82 and recall that the omitted term in ( [ mu_dgr ] ) comes from the ` sunrise ' diagram for @xmath141 .
a dimensional analysis of this diagram gives @xmath381\ ; .
\label{delta_mu}\ ] ] which transforms into a similar estimate for the total and superfluid density @xmath382\ ; .
\label{delta_n}\ ] ] @xmath383 this can be seen from the dimensional analysis of the second - order diagrams contributing to the self - energies .
relation ( [ p1 ] ) for the pressure implies @xmath384\ ; . \label{delta_p}\ ] ] in the case of entropy , it is convenient to start with the regime @xmath385 when using ( [ s2 ] ) to relate the uncertainty in @xmath3 to @xmath386 , @xmath387 , and @xmath388we find @xmath389 we see that at @xmath375 , in contrast to the behaviour of other thermodynamic functions , the uncertainty in @xmath3 scales only as the first power of @xmath64 ( simply because @xmath3 itself gets of order @xmath390 ) .
this scaling persists down to @xmath365 : @xmath391 because , at @xmath374 , the thermodynamics of the system corresponds to the generic low - temperature behaviour of superfluids , where the leading temperature - dependent contributions are due to phonons .
hence , the first - order correction to the sound velocity , which is of the order of @xmath64 as is seen from the expression for the energy , translates into the order-@xmath64 correction to the entropy . at @xmath374 ,
when thermodynamics is exhausted by dilute non - interacting phonons , the correction to the sound velocity can be found directly from the zero - temperature compressibility , and the accuracy of the expression for @xmath3 ( and other thermodynamic quantities ) can be improved . at @xmath375
, however , the order-@xmath64 correction to the quasiparticle dispersion law goes beyond the bogoliubov ansatz @xcite .
note also that improving the value of the sound velocity in the @xmath392 , while rendering the answers for thermodynamic quantities more accurate , is inconsistent with retaining the bogoliubov form of the spectrum for _ all _ the momenta .
finally , equation ( [ e_cnd ] ) in combination with the above results yields the estimate for the uncertainty of energy : @xmath393 \ ; . \label{delta_e}\ ] ] in all spatial dimensions , there is an interval ( in terms of temperature , if density is kept fixed , or in terms of chemical potential at fixed temperature , etc . ) where @xmath258 becomes of order unity , and the systematic perturbative description breaks down . in 3d and 2d
this happens in the vicinity of the superfluid phase transition point . in 1d
there is no superfluidity in the strict sense of the word , and no phase transition occurs , but the picture remains very similar to that of 2d and 3d case for a weakly interacting gas , as explained in the next subsection .
a detailed discussion of the fluctuation region , including accurate expressions for the critical points , will be presented later , in section [ sec : fluct ] . meanwhile , here we want to utilize the fact that the results ( [ delta_mu])-([delta_e ] ) allow one to make order - of - magnitude estimates of the non - perturbative fluctuation contributions to the thermodynamic functions .
these estimates are quite important as they show the degree to which the perturbative results of the previous sections are inaccurate in the fluctuation region . as we will see , for some quantities the fluctuation corrections are smaller than the leading ideal - gas contributions .
by continuity , the order - of - magnitude estimates for fluctuation contributions can be obtained from ( [ delta_mu])-([delta_e ] ) by simply setting @xmath394 , while for the quantities themselves we can use their mean - field critical expressions . this way we arrive at the following results @xmath395 @xmath396 for the superfluid and condensate densities we get an obvious answer that in both cases the fluctuation contributions are on the order of 100% . as long as a 1d system is weakly interacting , i.e. @xmath397
, the notion of the order parameter field with well defined amplitude and the two - component ( normal + superfluid ) description remain physically meaningful .
since superfluidity is a topological phenomenon , it can be destroyed only by topological defects
phase slips .
at @xmath397 the phase slips are rare events which do not contribute significantly to the local thermodynamic quantities ; in correlation functions , phase slips show up only at length scales much larger than @xmath398 , where their effect can be described at the hydrodynamic level . when the temperature reaches the characteristic scale of @xmath399 the parameter @xmath258 becomes of order unity . apart from the lack of a genuine phase transition ,
the physics in the temperature range @xmath400 is close to that of the fluctuation region in 2d and 3d . only the long - wave classical - field subsystem of the original quantum - field experiences strong non - linear fluctuations .
this leads to non - perturbative contributions to the system thermodynamics which are universal for all weakly interacting one - dimensional u(1 ) systems , and can be described by universal scaling functions in direct analogy with the fluctuation regions in 2d and 3d @xcite . at @xmath401 ,
the low - momenta part of the classical - field component gets depleted , so that the non - linearity of interactions becomes weak and accurately accounted for within the normal - gas mean - field picture .
beliaev s diagrammatic technique allows one to calculate correlation functions up to distances much larger than the correlation radius @xmath402 .
however , addressing the asymptotic long - range behaviour of off - diagonal correlation functions requires special techniques properly accounting for long - wave fluctuations of the goldstone mode , i.e. the phase of the superfluid order parameter .
since the diagrammatic and hydrodynamic description overlap , one can straightforwardly extend the diagrammatic calculation of off - diagonal correlators obtained up to large enough distances , @xmath403 to arbitrarily large @xmath404 s .
we proceed in the spirit of popov s hydrodynamic approach @xcite , but with a significant simplification that the hydrodynamic treatment takes place on top of beliaev s techniques , and thus does not require any additional modifications .
the simplification is possible due to the fact that at large distances only phase fluctuations are important and their effect can be factored out as @xmath405 where , under coarse - graining up to momentum @xmath346 , @xmath406 is the long - wave part of the phase field . as before
, the momentum cutoff @xmath407 should be small enough to guarantee ( i ) the statistical independence of the fluctuations of @xmath408 from the fluctuations of @xmath409 , and ( ii ) the gaussian character of phase fluctuations in the hydrodynamic regime .
( a particular value of @xmath346 is not important and does not appear in final expressions . ) in view of ( i ) and ( ii ) , we have for an @xmath14-particle correlator , @xmath410 where the subscripts label the space - time variables . using ( [ phi ] ) it can be written as @xmath411 with @xmath412 where @xmath413 is obtained from @xmath414 by substituting @xmath415 , for all @xmath416 s . at distances @xmath417
the correlator @xmath413 remains constant , and the remaining dependence on distance is exclusively due to the fluctuations of @xmath408 .
[ the correlator @xmath413 does depend on the coarse - graining momentum @xmath346 and this dependence is crucial for compensating the @xmath346-dependence of the function @xmath54 ; in fact its origin is directly related to @xmath54 when coarse - graining is stopped at larger momenta and then taken further to @xmath346 . ]
the independence of @xmath413 on coordinates and times in the asymptotic limit immediately leads to the relation @xmath418 between correlators at different sets of variables , @xmath419 and @xmath420 , provided both are in the asymptotic region . now
if @xmath420 is within reach of diagrammatic expansions , then ( [ k_k ] ) allows one to `` extrapolate '' @xmath421 to @xmath422 at _ arbitrary _ @xmath419 in the asymptotic domain . by the structure of ( [ k_k ] ) ,
the difference @xmath423 is independent of @xmath346 , as opposed to individual @xmath54 s .
here we assumed for simplicity that all variables in @xmath419 and @xmath420 are large enough .
the same analysis is readily adapted for cases when only some of the coordinates in the @xmath14-particle correlation function are in the asymptotic regime ; only these coordinates have to be mentioned in ( [ k_k ] ) .
the gaussian character of the field @xmath408 implies , see ( [ lambda ] ) , that @xmath424 where @xmath425 , @xmath426 ( the same for primed variables ) , and @xmath427 \left\langle \ , |\phi({\bf k},\xi)|^2\right\rangle\ , .
\label{xi_def}\end{aligned}\ ] ] we see that for the extrapolation of correlation functions to arbitrarily large distances one needs to know only @xmath428 which is defined through the average @xmath429 . the latter is readily found from popov s hydrodynamic action @xmath430 |\phi({\bf k},\xi)|^2 \ , , \label{hydro}\ ] ] leading to @xmath431^{-1 } . \label{phase_fluct}\ ] ] in fact ,
non - zero frequencies ( accounting for the quantization of the fluctuations of the phase ) are relevant only at @xmath374 , meaning that the parameter @xmath432 in ( [ phase_fluct ] ) can be taken in its @xmath298 limit , @xmath433 at @xmath385 , only the @xmath153 term should be left in ( [ xi_def ] ) . within
the relevant orders , equation ( [ spdm_2 ] ) corresponds to the expansion of the exponential in ( [ spdm_extr ] ) in powers of @xmath442 , up to the term @xmath443 included .
this immediately suggests that within the same accuracy one can exponentiate @xmath442-terms in ( [ spdm_2 ] ) to extend its domain of applicability to much larger distances .
the other advantage of proceeding this way is having a physical definition of the quasi - condensate density as the amplitude of the order parameter field in the long - wave limit @xcite .
the equations that follow have the same accuracy as ( [ spdm_2 ] ) though they do contain artificial higher - order terms which arise from factorizing the correlation function .
more specifically , we single out terms in ( [ spdm_2 ] ) which on large distances reproduce @xmath442 , and then exponentiate them : @xmath444 \ \longrightarrow \
\tilde{\rho}({\bf r})\ , { \rm e}^{-\lambda ( { \bf r } ) } \ ; , \label{spdm_3}\ ] ] where @xmath445 , and @xmath446 \ ; , \label{ntilde}\ ] ] @xmath447 \;. \label{lambda}\ ] ] we have added and subtracted @xmath448\ [ \tm \epsilon /2e^2 ] $ ] to ( [ spdm_2 ] ) to ensure that ( i ) both @xmath449 and @xmath450 are free from ultra - violet and infra - red divergencies in all dimensions , and ( ii ) that only small momenta @xmath451 contribute to the phase correlator @xmath450 . by comparing ( [ lambda ] ) and ( [ xi_low ] ) we see that they coincide at large distances up to leading terms .
it means that in 2d and in 1d at zero temperature , equation ( [ spdm_3 ] ) can be trusted up to exponentially large scales , while in 1d at finite temperature @xmath452 it works at least up to distances @xmath453 .
we are now in position to define the quasi - condensate density as the limiting value of @xmath454 : @xmath455 \ , .
\label{nquasi}\ ] ] the physical meaning of this relation is the amplitude squared of the order parameter field at large distances .
there is a certain degree of freedom in attributing terms which do not result in the power - law or exponential decay of @xmath456 to @xmath457 or to @xmath450 .
the idea behind our choice is three - fold : ( i ) equations([spdm_3 ] ) , ( [ ntilde ] ) , and ( [ lambda ] ) are valid as written in all spatial dimensions ; ( ii ) at large distances @xmath450 has the structure of hydrodynamic phase correlations ; ( iii ) the final expressions have the simplest form possible within the same accuracy .
it is also worth mentioning that @xmath458 is a quantity which controls all physical processes happening in the wigb at short distances at low temperature , e.g. @xmath14-body recombination rates . in all spatial dimensions it plays the same role as the condensate density in 3d system as long as one is interested in length scales not much larger than the healing length @xcite .
finally , the condensate density is defined from @xmath459 it is not accidental that the ultimate long - wave length property of the superfluid system is determined last .
it was always an unpleasant feature of numerous mean - field treatments that the crucial parameter determining physics at short scales was linked to @xmath274 .
as was mentioned in the introduction , we assume that the temperature is not very high , so that the condition ( [ lambda_t_vs_r_0 ] ) is preserved and the quantity @xmath17 remains the only parameter characterizing the interaction between the particles . in the normal region , where there are no anomalous correlators , we have to deal only with the dyson equation for the green s function @xmath129 in terms of self - energy @xmath148 . within the leading - order approximation , @xmath460 ( in 2d this formula implies an adequate choice of the parameter @xmath461 for the effective interaction @xmath17 , discussed in subsection [ subsec:2d_large_t ] ) , and the expression for @xmath129 yields the self - consistent relation for the number density , @xmath57^{-1 } \
, , \label{n_normal}\ ] ] with @xmath462 for the pressure we find @xmath463 using @xmath464 \ ; , \label{obs}\ ] ] we finally arrive at @xmath59 \ ; . \label{p_normal}\ ] ] equations ( [ n_normal ] ) , ( [ eff_mu ] ) , and ( [ p_normal ] ) specify @xmath1 , @xmath82 , and @xmath291 as functions of @xmath465 . utilizing ( [ s2 ] ) , with @xmath466 , and then ( [ e1 ] ) , we find for the entropy @xmath467 \right\ } , \label{s_nrml}\ ] ] and energy @xmath468 in the quasi - condensate region , the parameter @xmath32 defining the effective interaction @xmath17 can vary over a wide range of values thanks to the retained second - order correction that produces counter - terms compensating for the arbitrary choice . since in the normal region
we confine ourselves to the first - order expression for the self - energy , @xmath469 , we need to choose the value of @xmath32 which minimizes the omitted second - order contributions originating from the sunrise diagram , see figure [ fig : sunrise ] .
due to momentum independence of the pseudo - potential line , the third ( fourth ) diagram is identical to the first ( second ) one ; we thus consider only the first two diagrams and multiply the result by a factor of 2 .
the parameter @xmath32 can be much smaller or much larger than @xmath4 . in either case
the leading term comes from the logarithmic ultraviolet contribution from the product of two propagators , yielding the result @xmath470 + nu{\cal o}(mu)\ , , \label{sec_order_sig}\ ] ] where @xmath471 is the thermal momentum .
this leads to the expression @xmath472 \ , .
\label{sec_order_sig2}\ ] ] comparing the first two terms in the brackets with ( [ c_12_2d ] ) and ( [ u1_u2_expd ] ) , we see that if we take the value of @xmath32 substantially away from @xmath4 , then the leading second - order correction to the expression @xmath473 amounts to renormalizing the value of the pseudo - potential @xmath17 in such a way that it corresponds to @xmath474 .
this brings us to the conclusion that @xmath475 is the optimal choice for @xmath32 , in which case omitting the diagrams shown in figure [ fig : sunrise ] is justified by the parameter @xmath476 . at temperatures above the fluctuation region where the renormalized chemical potential remains small @xmath477 the expansion parameter
is defined by the infra - red behaviour of the zero - frequency lines .
the dimensionless factor , @xmath258 , associated with adding an extra full interaction vertex to a diagram is given by ( [ full_t_rel1 ] ) , but now with @xmath478 , resulting in the estimate @xmath479 where @xmath480 is the size of the fluctuation region in terms of the chemical potential . at temperatures
much higher than the degeneracy temperature @xmath481 where @xmath482 , all corrections are small in the parameter @xmath483
in section [ sec : quasicondensate ] we have obtained the thermodynamics of the system in the quasi - condensate region , see ( [ n_total_qc])-([n_prime ] ) , ( [ p3 ] ) , ( [ s_cnd ] ) , and ( [ e_cnd ] ) which specify @xmath1 , @xmath82 , @xmath291 , @xmath3 , and @xmath257 as functions of two independent variables , @xmath15 and @xmath4 , in the form of integrals involving the @xmath15-dependent quasiparticle spectrum , @xmath484 , given by ( [ e2 ] ) .
so far , we did not discuss the physical meaning of the function @xmath484 , since this was not necessary for our derivations . on the other hand , equation ( [ s_cnd ] ) for
the entropy clearly suggests that , within our approximation , the system is equivalent to a system of non - interacting bosonic quasiparticles with dispersion @xmath484 , described by the hamiltonian @xmath485 where @xmath486 is the quasiparticle occupation number operator and @xmath487 is some constant .
the natural question is then whether the hamiltonian ( [ h_ps ] ) can be used for calculating other thermodynamic quantities within the non - interacting quasiparticle gas picture . by direct comparison with ( [ n_total_qc])-([n_prime ] ) , ( [ p3 ] ) , ( [ s_cnd ] ) , and ( [ e_cnd ] )
, it can be shown that this is indeed the case and one can write @xmath488 @xmath489 it is important , however , to remember that @xmath487 is a function of both @xmath15 and @xmath4 , satisfying @xmath490 in view of this temperature dependence , we refer to ( [ h_ps ] ) as a _
pseudo_-hamiltonian , rather than an effective hamiltonian .
the pseudo - hamiltonian description can be applied to the normal region as well . in this case , @xmath491 were @xmath492 is specified by ( [ n_normal ] ) .
the @xmath15-dependent spectrum of quasi - particles is now given by ( [ eff_e ] ) .
in the fluctuation region the long - wave part of the classical - field component of the quantum field experiences strong ( non - perturbative ) fluctuations . it is exclusively due to these fluctuations that the diagrammatic technique for the quantum field loses an expansion parameter .
the special role of the classical - field component is evident from the fact that on approach to the fluctuation region the leading contribution to the expansion parameter @xmath258 is associated with zero - frequency propagators , see subsections [ subsec : finite_t ] and [ subsec : exp_par_norm ] . the diagrammatic technique built on zero - frequency propagators ( with perturbative parts of self - energies included ) and the interaction represented by pseudo - potential @xmath17 corresponds to the diagrammatic expansion of the gibbs distribution @xmath493/t } { \cal d}\psi \label{z_classical}\ ] ] for the in momentum space truncated classical field @xmath494 with the hamiltonian functional @xmath495 = \int \left [ { 1\over 2 m } |\nabla \psi|^2 + { u\over 2}|\psi|^4 -\mu ' |\psi|^2 \right]\ , { \rm d}{\bf r}\ , .
\label{h_classical}\ ] ] here @xmath496 is a truncation momentum ( to be discussed later ) , and @xmath497 is the reduced chemical potential obtained from the bare one by subtracting relevant perturbative contributions from all the harmonics with @xmath498 , including the quantum ones .
the bottom line is that describing the quantum gas in the fluctuation region reduces to solving the classical - field problem ( [ z_classical])-([h_classical ] ) , in its fluctuation region . in @xmath499 ,
numerical solutions to the problem ( [ z_classical])-([h_classical])in terms of scaling functions for thermodynamic quantities are available in the literature and have already been used to accurately describe weakly interacting 2d and 3d quantum gases in the fluctuation region @xcite .
the same is possible in 1d , but to the best of our knowledge it has not been done so far . in the fluctuation region ,
an important quantity is the momentum @xmath500 separating weakly and strongly coupled classical modes . for a quantitatively accurate description of the fluctuating classical - field sub - system
, the momentum @xmath501 separating classical modes of interest from the rest of the modes has to be much larger than @xmath502 .
however , as long as we are interested in the order - of - magnitude estimates , it is safe and convenient to set @xmath503 , so that all the modes we are dealing with in ( [ z_classical])-([h_classical ] ) are strongly coupled . to estimate @xmath502 we note that with @xmath503 all three terms in the hamiltonian ( [ h_classical ] ) have to be of the same order of magnitude , that is @xmath504 where @xmath505 is the long - wavelength contribution to the total density , and @xmath506 .
for @xmath507 we have @xmath508 , and since @xmath502 is separating strongly coupled long - wave harmonics from slightly perturbed short - wave ones , the order - of - magnitude estimate for @xmath509 follows by continuity from the ideal system formula : @xmath510 substituting this back into ( [ est1])-([tilde_n ] ) yields @xmath511 @xmath512 @xmath513 the estimates ( [ est2_n])-([est2_mu ] ) imply the following parameterization of the grand - canonical equation of state in the fluctuation region and its vicinity : @xmath514 @xmath515 here @xmath516 is a dimensionless scaling function of a dimensionless scaling variable @xmath293 . in 2d and 3d , the quantities @xmath517 and @xmath67 are the critical values of density and chemical potential at a given temperature . in 1d , where the phase transition is absent , one can set , without loss of generality , @xmath518 , and , correspondingly , @xmath519 .
similarly , superfluid and condensate densities in the dimensions in which they are meaningful are parameterized as @xmath520 and @xmath521 the functions @xmath522 , @xmath523 , and @xmath524 are universal for all weakly interacting u(1)-symmetric systems in the given dimension ( no matter quantum or classical , continuous space or lattice ) ; the numerical data for them is available for @xmath499 in @xcite . by numerically solving the problem ( [ z_classical])-([h_classical ] ) , it was established that , up to higher - order corrections in the parameter @xmath64 , @xcite @xmath525 @xmath526 with @xmath527 the critical density for bose - einstein condensation in an ideal 3d gas .
analogous relations for @xmath67 read @xcite @xmath528 @xmath529 note that the leading terms in the above relations , as well as order - of - magnitude estimates for the values of sub - leading terms , are readily obtained from the condition @xmath377 ; the accurate numerical treatment of the problem ( [ z_classical])-([h_classical ] ) is necessary to fix the values of the dimensionless constants . by re - writing ( [ n_c_2d ] ) , ( [ n_c_3d ] ) in terms of the critical temperature as a function of density , we get @xmath530 @xmath531 where @xmath532 is the critical temperature of the ideal 3d gas .
now it is easy to estimate @xmath502 .
the relevant temperature for having strong fluctuations is given by @xmath533 the estimate ( [ est2_k ] ) at @xmath534 then yields @xmath535 we see that @xmath536 in all cases , which justifies the statement that the physics of the fluctuation region is that of a ( strongly interacting ) classical field .
putting aside @xmath537 and @xmath274 , which are most sensitive to non - perturbative fluctuations of the classical field @xmath538 , the next two quantities which are sensitive to fluctuations , especially in 1d and 2d , are the density and the chemical potential
see estimates ( [ n_tilde_mu_tilde ] ) .
for other quantities as is seen from ( [ p_e_s_fluct])neglecting fluctuation corrections will not result in a significant error .
it is also important that up to a few dimensionless constants characterizing sub - leading contributions to critical values of @xmath291 , @xmath257 , and @xmath3 , the fluctuation corrections to these quantities are expressed in terms of the same function @xmath516 and its integral . indeed ,
equation ( [ rel1 ] ) implies the following relation for @xmath291 : @xmath539\nonumber \\ + \ , ( m^{2d } t^4 u^d)^{1\over 4-d}\int_0^x \lambda^{(d)}(x')\ , { \rm d } x ' \ , , \label{param_p}\end{aligned}\ ] ] and then with ( [ s2])-([e1 ] ) one obtains similar relations for @xmath3 and @xmath257 .
to see how accurate our description is for the density matrix we performed monte carlo simulations of weakly interacting two- and one - dimensional systems where phase fluctuations are the strongest .
we deliberately aim at systems with substantial decay of the density matrix to see the difference between the perturbative treatment and exact solution . in figure
[ fig : rho2d ] we present data for the two dimensional square - lattice system described by the bose - hubbard hamiltonian @xmath540 \ ; , \label{bh}\ ] ] where @xmath541 is the hopping amplitude between the nearest neighbor sites . model ( [ bh ] ) was simulated for @xmath542 , @xmath543 and filling factor @xmath544 on a lattice with @xmath545 points by using the worm algorithm approach @xcite .
formally , all expressions in this manuscript relating energy spectrum , particle density , and density matrix remain valid for an arbitrary dispersion relation with @xmath546 at small momenta .
moreover , the theory is supposed to work as written for systems with periodic boundary conditions provided the sums over momenta are understood as @xmath547 , where @xmath548 and @xmath549 is a @xmath550-dimensional integer . to suppress statistical noise ,
simulation data for @xmath551 were collected to spherically symmetric bins .
we are using exactly the same procedure to present theoretical data .
thus , the only difference between the essentially exact ( up to statistical errors ) simulation data and the theory is due to the finite value of @xmath17 . even for relatively large values of @xmath17 and temperature @xmath552
we observe a remarkable accuracy of ( [ spdm_2 ] ) .
one can get an idea of the systematic theoretical error by comparing curves derived from ( [ spdm_2 ] ) and ( [ spdm_3 ] ) . [ the latter is obtained by exponentiating the phase correlator at large distances ] . in figure [
fig : rho1d ] we apply our theory to the lieb - liniger model of the one - dimensional bose gas with contact interactions . since exact correlation functions at finite temperature
are not known , we performed monte carlo simulations using recently developed techniques for continuous space systems @xcite .
it is convenient to introduce the on - dimensional characteristic length @xmath553 as the unit of length .
the limit of weak interactions is obtained then by considering large densities @xmath554 , and we have chosen @xmath555 corresponding to the power law - decay of off - diagonal correlations at zero temperature with the exponent @xmath556 .
the upper curve in figure [ fig : rho1d ] shows results for low temperature @xmath557 , or @xmath558 when thermal phase fluctuations are barely influencing the data for the simulated system size @xmath559 .
when temperature is increased to @xmath560 ( lower curve in figure [ fig : rho1d ] ) we clearly see the bimodal decay of the density matrix and the crossover from the power - law to exponential decay .
contrary to the 2d case , equation ( [ spdm_3 ] ) captures the actual behaviour better than ( [ spdm_2 ] ) which is not so surprising given the large suppression of @xmath456 at large distances .
finally , figures [ fig : energyn_comparison ] and [ fig : pressuren_comparison ] show the agreement of monte carlo results for the energy ( squares ) and the pressure ( triangles ) at temperatures @xmath561 with theoretical curves from ( [ e_nrml ] ) and ( [ p_normal ] ) ( solid lines ) . in the following we compare thermodynamic functions obtained in sections [ sec : quasicondensate ] and [ sec : normal_region ] with exact quantum monte carlo results of a weakly interacting three - dimensional system .
let us start with the case of a homogeneous system with a small parameter @xmath562 . in figures [ fig : energyqc_comparison ] and
[ fig : pressureqc_comparison ] we compare monte carlo results for the energy ( squares ) per particle and pressure ( triangles ) at temperatures @xmath347 , with the theoretical curves from ( [ e_cnd ] ) and ( [ p3 ] ) ( solid lines ) .
the analytical expressions are in very good agreement with numerical results up to temperatures @xmath563 and there is no need to switch to the description in terms of universal functions @xcite .
this is consistent with the estimate ( [ p_e_s_fluct ] ) showing that fluctuation contributions to energy and pressure are very small .
the same does not apply to the superfluid and condensate densities for which the perturabtive expansion in @xmath17 is not valid as fluctuations of the order parameter can not be neglected on approach to the critical temperature . in figure ( [ fig : nsf_ncnd_comparison ] ) squares and triangles are the results of quantum monte carlo simulations for the superfluid and condensate fraction , respectively .
solid and dashed lines are the theoretical expressions from ( [ nslandau ] ) for the superfluid density and ( [ n_total ] ) for the condensate density , respectively , combined with preexisting classical monte carlo results @xcite for the fluctuation region . at a temperature @xmath564 we switch to the universal description , as @xmath565 [ see ( [ gamma_t_rel3 ] ) ] becomes of the order of unity . for @xmath563
the remaining discrepancy between classical and quantum monte carlo results is due to finite interaction strength .
one should not be misled by the small value of the gas parameter in the fluctuation region since the proper parameter controlling the size of the fluctuation region in temperature , @xmath566 , and the accuracy of the universal description involves a large numerical prefactor and is rather @xmath567 , see @xcite .
we have shown that beliaev s diagrammatic technique regularized by adding a small term explicitly breaking the u(1 ) symmetry to the hamiltonian yields a simple and controllable description of the weakly interacting bose gas in any dimension .
this approach is especially convenient for obtaining thermodynamic functions that are not sensitive ( up to higher - order corrections ) to the long - range phase fluctuations .
the symmetry breaking term introduces a small gap in the otherwise gapless goldstone mode and suppresses long - range fluctuations of the phase of the order parameter , thereby introducing a genuine condensate density and thus substantially simplifying the description of lower - dimensional systems . in this sense ,
the symmetry - breaking trick is an alternative to berezinskii s finite - size trick and popov s special ( hydrodynamic ) treatment of long - wave parts of the fields . at the end of the calculation ,
the symmetry breaking term is set to zero . another important feature of our approach is that it completely avoids such notions as seed- or quasi - condensate for its construction .
the quasi - condensate density defined as the modulus squared of the order parameter field at distances larger then the healing length is calculated as the long - wave property of the theory , rather than serving as an input parameter .
it is only natural to have the theory which handles the short - range physics first and uses it for dealing with the low - energy physics next .
as far as long - range off - diagonal correlations are concerned , the approach produces accurate answers for the off - diagonal correlation functions up to distances where further evolution of the correlators is controlled by generic hydrodynamic relations , and thus can be accurately extrapolated to arbitrarily large distances .
we confined our analysis to the most typical and non - trivial ( in 2d and 3d ) case of dilute gas , when the size of the potential @xmath6 is much smaller than the distance between the particles .
in the ( boltzmann ) high - temperature regime , we have restricted ourselves to low enough temperatures at which de broglie wavelength remains much larger than @xmath6 and the answers do not depend on the details of the pair potential . under the two above - mentioned conditions ,
the effective inter - particle interaction is described by a single parameter @xmath17 .
technically , generalization of the theory to the case of @xmath8 is straightforward . in this case
, the weakness of interaction literally means born interaction potential , and complete summation of ladder diagrams in 2d and 3d becomes irrelevant .
the description of a normal gas at @xmath568 is a well - studied topic that goes beyond the scope of the present paper . the pair of variables @xmath569 and @xmath4 form the most convenient set of parameters determining the rest of thermodynamic quantities , including the condensate density , if any .
the structure of the answers corresponds to a picture of non - interacting bosonic quasiparticles with the bogoliubov spectrum ( [ e2 ] ) , on top of the ` vacuum ' with temperature - dependent energy .
this allows one to introduce the bilinear bosonic pseudo - hamiltonian ( [ h_ps ] ) with bogoliubov spectrum .
we argue that within the bilinear pseudo - hamiltonian ansatz our results can not be further improved at least at @xmath385 . in the @xmath392 limit ,
when thermodynamics is exhausted by dilute non - interacting phonons , using the improved value of the sound velocity obtained directly from the zero - temperature compressibility , yields more accurate answers for thermodynamic quantities , but is inconsistent with retaining the bogoliubov form of the spectrum for _ all _ the momenta .
we also show that any attempt to improve the self - consistent mean - field description of the superfluid region , aimed at getting rid of the spurious first - order phase transition , is senseless as long as the fluctuation - induced non - perturbative shift of the critical temperature is not accounted for .
in fact , there is no need to extrapolate the pseudo - hamiltonian description to the fluctuation region around the critical temperature in view of the availability of the numerical answers for universal scaling functions describing all the weakly interacting u(1 ) systems at @xmath570 . with these data , the two analytical descriptions in the normal and superfluid phases are readily connected across the fluctuation region . by comparing our results to first - principles mote carlo data , we observe that the most vulnerable region of parameters is @xmath570 , where the description is accurate only when the gas parameter @xmath571 is as small as @xmath572 .
we acknowledge support from nsf ( grant no .
phy-0653183 ) , the swiss national science fund , itamp , and the darpa ole program .
|
aiming for simplicity of explicit equations and at the same time controllable accuracy of the theory we present results for all thermodynamic quantities and correlation functions for the weakly interacting bose gas at short - to - intermediate distances obtained within an improved version of beliaev s diagrammatic technique . with a small symmetry breaking term beliaev s diagrammatic technique becomes regular in the infrared limit .
up to higher - order terms ( for which we present order - of - magnitude estimates ) , the partition function and entropy of the system formally correspond to those of a non - interacting bosonic ( pseudo-)hamiltonian with a temperature dependent bogoliubov - type dispersion relation . away from the fluctuation region , this approach provides the most accurate in fact , the best possible within the bogoliubov - type pseudo - hamiltonian framework description of the system with controlled accuracy .
it produces accurate answers for the off - diagonal correlation functions up to distances where the behaviour of correlators is controlled by generic hydrodynamic relations , and thus can be accurately extrapolated to arbitrarily large scales . in the fluctuation region , the non - perturbative contributions are given by universal ( for all weakly interacting u(1 ) systems ) constants and scaling functions , which can be obtained separately by simulating classical u(1 ) models and then used to extend the description of the weakly interacting bose gas to the fluctuation region .
the theory works in all spatial dimensions and we explicitly check its validity against first - principle monte carlo simulations for various thermodynamic properties and the single - particle density matrix .
|
crystals of rare - earth ( re ) doped quasi - one - dimensional double bromides cscdbr@xmath2 are widely studied , mainly , because of their property to incorporate re@xmath8 ions in pairs , even at low re concentrations .
this makes them a promising material for up - conversion lasers .
the structure of cscdbr@xmath2 belongs to the @xmath10 space group and consists of linear chains of face - sharing [ cdbr@xmath11@xmath12 octahedra along the @xmath13-axis .
positional symmetry for cd@xmath14 is @xmath15 .
re@xmath8 ions substitute for cd@xmath14 forming centers with different mechanisms of charge compensation .
the main center consists of two re@xmath8 ions placed in the chain on each side of an adjacent cadmium vacancy , [ re@xmath8cd@xmath14 vacancy
re@xmath1 ] @xcite .
both re@xmath8 ions in such a center are equivalent , their positional symmetry lowers from @xmath15 to @xmath17 .
spectra of many other centers with different configurations were reported in epr and optical studies of cscdbr@xmath2:re crystals .
in particular , positional symmetry for re@xmath8 ions in minor axial centers which are supposed to appear at higher re concentrations , namely , [ re@xmath8re@xmath8cd@xmath18 vacancy ] or [ re@xmath8 cd@xmath14 vacancy cd@xmath14re@xmath8 ] , is again @xmath17 but two re@xmath8 positions are not equivalent , while in [ re@xmath8cs@xmath20 vacancy ] centers positional symmetry of re@xmath8 ion is @xmath21 .
though the spectra of cscdbr@xmath2:pr@xmath8 were intensively studied before @xcite , crystal field energies of the lowest excited multiplets were not investigated in absorption .
the positions of crystal field levels for the main center , as reported by different authors , differ by 214 @xmath22 .
typical spectral resolution used in the mentioned works was 0.82 @xmath22 , while the only high resolution study ( it concerned the @xmath23 level ) revealed spectral lines as narrow as 0.1 @xmath22 @xcite .
peculiar doublet line shape was observed for the @xmath24 transition and explained by the combined effect of unresolved hyperfine structure and non - axial crystal strains @xcite .
the very first analysis of the crystal field energies of pr@xmath8 in cscdbr@xmath2 @xcite showed that the crystal field affecting the symmetric dimer centers consists of a strong cubic and a weak trigonal components and that the last one is determined mainly by the quadrupolar ( b@xmath25 ) term .
this conclusion was confirmed in the studies of the crystal field energies of tm@xmath26 and ho@xmath8 symmetric pair centers where data obtained from both optical and submillimeter epr investigations were taken into account @xcite .
the sets of crystal field parameters obtained in refs .
@xcite from the fitting of the simulated energy level schemes to the measured 40 energy levels of pr@xmath8 are close to one another and provide small enough root - mean square deviations ( 22.411.1 @xmath22 ) between calculated and observed energy level data .
however , some of the crystal field parameters in cscdbr@xmath2:re@xmath8 presented in the literature @xcite , in particular b@xmath27 and b@xmath28 , vary nonmonotonically with the occupation number of the 4@xmath29 electron shell .
we have undertaken the low temperature high resolution infrared absorption study of cscdbr@xmath2:pr@xmath8 with the aims ( i ) to determine directly positions of crystal field levels in the region between 2000 and 7000 @xmath3 and ( ii ) to study the line widths and shapes and thus to obtain information on relaxation processes and hyperfine interactions .
we also performed crystal field calculations for the main center in the framework of the semiphenomenological exchange charge model ( ecm ) @xcite , based on the analysis of the local lattice structure .
parameters of the ecm were corrected by fitting the calculated hyperfine splittings of crystal field levels to the experimental data from our measurements .
the electron - phonon interaction effects , in particular , the low temperature relaxation of excited states of the ( pr@xmath30-dimer due to one- and two - phonon transitions between the crystal field sublevels of the @xmath31 @xmath32 , @xmath6 and @xmath7 multiplets , are studied in the framework of the recently derived rigid ion model of lattice dynamics of doped cscdbr@xmath2:re crystals @xcite .
cscdbr@xmath2 crystals containing 0.2@xmath0 of pr@xmath8 were grown by the method described in ref .
crystals easily cleave along the @xmath13 -axis .
we have prepared a 1.3 mm thick sample ( sample a ) with the @xmath13-axis parallel to the cleaved face .
the sample was upheld at a controlled temperature of 5 k in a helium vapor cryostat .
unpolarized absorption spectra were measured with a bomem da3.002 fourier transform spectrometer in the spectral range between 2000 and 7000 @xmath22 with a resolution from 1.0 to 0.05 @xmath22 . in fourier
transform spectroscopy , the quantity res@xmath33 , where @xmath34 is the maximum optical path difference , is indicated as the resolution .
the full width at half hight ( fwhh ) of the instrumental function depends on a particular apodization function used .
we used no apodization and , thus , worked with the narrowest possible instrumental function ( fwhh@xmath35 res ) .
it is easy to show that , in this case , the shape of a spectral line is essentially unchanged provided that fwhh @xmath36 res .
precision of the experimental line positions was 0.050.3 @xmath22 , depending on a particular line .
precise absolute wave number scale is an intrinsic property of fourier transform spectroscopy .
line widths were determined from absorbance spectra calculated with a zero - line taken at a half of transmitted intensity . in such a way we tried to take into consideration the fact that the incident light is unpolarized but a majority of lines is 100@xmath0 polarized .
the lowest frequency region of the @xmath37 spectral transition has also been measured with the resolution of 0.005 @xmath22 using another sample , 3.1 mm thick and oriented approximately along the @xmath13-axis ( sample b ) .
it allowed almost pure @xmath38 geometry and , consequently , @xmath39 polarization of the incident light .
1 shows low temperature transmittance spectra of cscdbr@xmath2:pr@xmath8 ( 0.2@xmath0 ) corresponding to different infrared transitions .
the first excited state of the ground multiplet @xmath5 lies at about 170 @xmath22 @xcite and is not populated at helium temperatures .
so , low temperature absorption spectra display crystal field levels of the excited multiplets directly .
they are listed in table 1 , together with data from other publications . as the ground state is the @xmath5(@xmath40 ) singlet , excited @xmath41 states manifest themselves in @xmath42-polarization but @xmath43 states in @xmath44-polarization , while @xmath45 states are silent in electric dipole approximation ( see table 2 ) .
magnetic dipole transitions are usually very weak , with a possible exception for the @xmath46 transition that is allowed for the free pr@xmath8 ion .
the symmetries of the @xmath7 crystal field levels have been determined from a comparison between the spectra of the two differently oriented samples ( samples a and b ) .
the assignement of the 6483.3 @xmath22 level as a @xmath43 level is strongly supported by the well resolved hyperfine structure ( hfs ) observed in high resolution spectra ( see inset 1 of fig .
1d and fig .
the point is that while @xmath43 levels exhibit the magnetic hfs , @xmath47 and @xmath45 do not .
the next @xmath43 level also shows the traces of unresolved hfs , namely , the 6503.5 @xmath22 line has a flat bottom ( see inset 2 of fig .
expansion of the 6483.3 @xmath22 line into six components yields 0.040 @xmath22 wide gaussians .
the gaussian shape of hfs components gives evidence that an inhomogeneous broadening due to random crystal fields exceeds a homogeneous broadening due to nonradiative transitions to the nearest @xmath48 level lying at about 10 @xmath22 lower @xcite ( we did not observe @xmath49 ) transitions forbidden as electric dipole ones ) .
the lineshapes of the 6496.5 @xmath22 ( @xmath50 ) and 6503.5 @xmath22 ( @xmath51 ) lines reveal homogeneous broadening of 0.17 and 0.31 @xmath22 , respectively .
such a broadening comes , evidently , from electron - phonon interaction and implies a noticeable density of phonon states in the region between 20 and 30 @xmath22 corresponding to the distances from the considered levels to lower crystal field levels .
other examples of level broadening due to electron - phonon interaction can be found in the @xmath4 multiplet which is situated in the region of about 2400 @xmath22 ( see fig .
the lowest level of this multiplet lying at about 2234 @xmath22 was reported to be the @xmath48 level , from polarized fluorescence measurements @xcite . in that case , a strong very narrow ( with the width of 0.05 @xmath22 ) line observed at 2234.8 @xmath22 corresponds to the magnetic dipole transition @xmath52 . as we have already mentioned , such a situation is possible just for the @xmath53 spectral multiplet .
the next level lies at 26 @xmath22 from the bottom of the multiplet and , though suffers from a phonon relaxation , exhibits hfs as it follows from the shape of 2261.4 @xmath22 line .
this is in accordance with the previously made assignments of this level symmetry as @xmath43 .
three closely spaced levels at 2316.6 , 2331.9 and 2347.8 @xmath22 are broad .
their widths are 3.0 , 3.3 and 3.7 @xmath22 respectively , and it is not possible to determine their symmetries .
calculations described below show that an overall hyperfine splittings of @xmath43 levels do not exceed 0.5 @xmath22 . consequently , the level widths of 34 @xmath22 are , mainly , due to rapid phonon - assisted decay to lower levels . in accordance with the crystal field calculations ( see below ) , at least one of these three broad absorption lines is to be assigned to another center .
the next level is more narrow again .
according to the shape of the line at 2546.7 @xmath22 it contains unresolved hfs and thus corresponds to the @xmath43 level .
the distances from this level to lower levels exceed the length of the phonon spectrum ( @xmath54 180 @xmath22 @xcite ) . because of
that , one - phonon relaxation is not possible .
the highest frequency line observed at 2572.1 @xmath22 could be attributed to the @xmath55 magnetic dipole transition .
but the @xmath56 level lying at 26 @xmath22 above the @xmath57 level ( 2546 @xmath22 ) should be broadened due to one- and two - phonon transitions to the lower crystal field sublevels , and the width of the respective spectral line should be about 0.30.5 @xmath22 ( see table 3 ) , not 0.05 @xmath22 as observed .
probably , this line belongs to some other pr@xmath8 center .
weak narrow lines observed at 2153.6 and 2167.3 @xmath22 ( see fig .
1a ) also have to be assigned to some other centers .
if these levels belonged to the main center , the level at 2234.8 @xmath22 would be broadened by the nonradiative transitions to these levels .
it is not the case experimentally .
we were able to find only two levels in the @xmath6 multiplet and the lowest sublevel in each of @xmath58 and @xmath32 .
the level at 5073 @xmath22 which lies far ( 250 @xmath22 ) from a lower level is , evidently , broadened by unresolved hfs and is thus a @xmath43 level .
the measured hyperfine splittings and widths of hyperfine sublevels found from the experimental line shapes are presented in table 3 , to compare with the calculated hyperfine splittings and the estimated one - phonon decay rates ( see the next section ) .
many of the observed levels coincide with those of refs .
@xcite found from selectively excited fluorescence measurements .
we did not observe any additional structure due to nonequivalence of pr@xmath8 positions in asymmetric pair centers .
thus , our results support the conclusion of refs .
@xcite that the main center in cscdbr@xmath2:pr@xmath8 is a symmetric pair [ pr@xmath8cd@xmath14 vacancy
pr@xmath8 ] .
the energy level pattern of the @xmath59pr@xmath8 ion ( the nuclear spin @xmath60 , 100@xmath0 abundant , the 4@xmath61 electron shell ) in the trigonal crystal field can be represented by eigenvalues of the effective parametrized hamiltonian @xmath62 where @xmath63 is the two - body coulomb energy ; @xmath64 corresponds to the spin - orbit interaction ( we use in the present work @xmath65
@xmath22 and slater integrals @xmath66 ; @xmath67 ; @xmath68 @xmath22
@xcite ) ; @xmath69 is the crystal field hamiltonian @xmath70 ( @xmath71 are the stevens operators @xcite , the quantization axis @xmath72 coincides with the symmetry axis @xmath73 of the crystal lattice ) , and @xmath74 is the nuclear energy responsible for the hyperfine structure of the optical spectra . because we neglect many minor interactions ( three - body terms , spin - other orbit interaction , two - body correlations in the crystal field terms and etc . ) , shifts of the multiplet centers of gravity @xmath75 ) are introduced to fit the calculated spectrum to the experimental data .
only projections of the electron - nuclear magnetic and quadrupolar interactions on multiplet manifolds with fixed orbital ( @xmath76 ) , spin ( @xmath77 ) and total ( @xmath78 ) angular moments were considered : @xmath79- \nonumber\\ \frac{3e^2q \langle r^{-3 } \rangle \alpha _ j } { 4i(2i-1)}\biggl[\frac 13\left(3j_z^2-j(j+1)\right)\left(3i_z^2-i(i+1)\right)+\frac 12\left(j_{+}^2i_{-}^2+j_{-}^2i_{+}^2\right)+ \\
\label{eq3 } \frac 12\left(j_zj_{+}+j_{+}j_z\right)\left(i_zi_{-}+i_{-}i_z\right ) + \frac 12\left(j_zj_{-}+j_{-}j_z\right)\left(i_zi_{+}+i_{+}i_z\right)\biggr ] .
\nonumber\end{aligned}\ ] ] here @xmath80 is the magnetic dipole hyperfine constant ( in the free ion @xmath81 ghz @xcite , the corresponding hyperfine constants for other multiplets are obtained using this value and neglecting the core electron polarization contributions which do not exceed 1.5@xmath0 of the 4@xmath29 electron contributions @xcite ) , @xmath82 m@xmath83 @xcite is the nuclear quadrupole moment , @xmath84 is the sternheimer antishielding factor @xcite , @xmath85 are reduced matrix elements of the second rank spherical operators , @xmath86 are moments of the radial wave function of the 4@xmath29 electrons calculated in @xcite , and the electric field gradient at the pr@xmath8 nucleus equals @xmath87 where @xmath88(@xmath89 ) is the charge ( in units of the proton charge @xmath90 ) of a lattice ion with spherical coordinates @xmath91(@xmath89 ) , @xmath92(@xmath89 ) and @xmath93(@xmath89 ) in the system of coordinates having its origin at the pr@xmath8 nucleus . as we have already mentioned in the introduction , the crystal field parameters obtained earlier from the fitting to the measured energy levels of pr@xmath8 @xcite and nd@xmath8 @xcite in cscdbr@xmath2 , though providing small enough deviations between the calculated and observed energy levels , exhibit nonmonotonic variation within the re@xmath8 family ( see table 4 ) .
moreover , there are large discrepancies between the measured in this work hfs of the absorption lines and the hfs , calculated with the crystal field parameters from ref . @xcite .
in particular , the 6483 @xmath22 line exhibits well resolved hfs of total width 0.21 @xmath22 , while the calculated width is 0.11 @xmath22 .
the formal fitting procedure employed in @xcite evidently brings about overestimated 6-rank parameters of the trigonal crystal field component . to clear up relations between the crystal field parameters and their origin , we estimated them in the framework of the exchange charge model @xcite .
the crystal field is represented as a sum ( @xmath94 ) of the electrostatic field of the lattice ions and the exchange charge field defined by the parameters @xmath95 and @xmath96 respectively . here
@xmath97 are numerical factors @xcite , @xmath98 are the shielding factors , @xmath99 , @xmath100 , @xmath101 , @xmath102 and @xmath103 are dimensionless parameters of the model .
the expression ( 6 ) involves the sum over only the nearest neighbors ( six bromine ions ) of the re@xmath8 ion .
the overlap integrals @xmath104 , @xmath105 have been computed using the radial 4@xmath29 wave function of the pr@xmath8 ion from ref .
@xcite and the 4@xmath106 , 4@xmath107 wave functions of the br@xmath108 ion given in @xcite .
dependencies of the overlap integrals on the interionic distance @xmath91 ( in angstroms ) are approximated by the following expressions : @xmath109 equations ( [ eq5 ] , [ eq6 ] , [ eq7 ] ) present the crystal field parameters as explicit functions of relative positions of the impurity pr@xmath8 ion and the lattice ions .
we analyzed distortions in the relaxed cscdbr@xmath2 lattice containing the substitutional dimer [ pr@xmath8cd@xmath110 vacancy
pr@xmath8 ] in the framework of the same quasimolecular model as the one already used in @xcite to calculate the structure of symmetric tm@xmath8 dimers .
values of ion displacements from their equilibrium positions in the perfect lattice were obtained by minimizing the potential energy of the cluster which involved the nearest 50 ions around the symmetric dimer center .
these ions belong to the first ( 12 br@xmath108 ) , second ( 2 cd@xmath14 ) , third ( 12 cs@xmath20 ) , and fourth ( 24 br@xmath108 ) coordination shells of the impurity pr@xmath8 ions ( see table 5 ) .
the energy of the interionic interaction is assumed to be a sum of coulomb and non - coulomb terms .
the effective ion charges ( @xmath88(cd@xmath14)@xmath111 , @xmath88 ( cs@xmath20)@xmath112 , @xmath88(br@xmath108)@xmath113 , @xmath88(pr@xmath8)@xmath114 ) and the non - coulomb force constants , corresponding to interactions between the nearest neighbors , pr@xmath8br@xmath108 , cd@xmath14br@xmath108 , cs@xmath20br@xmath108 , and br@xmath108br@xmath108 , have been obtained from the study of the cscdbr@xmath2 lattice dynamics @xcite . according to results of simulating the relaxation energy minimum with respect to the ion displacements ( see table 5 ) , the distance between the pr@xmath8 ions attracted by the cd@xmath14 vacancy diminishes from the value of the lattice constant @xmath115 nm @xcite down to 0.5915 nm .
this value is very close to the distance of 0.593 nm between the gd@xmath8 ions @xcite and 0.5943 nm between the tm@xmath8 ions @xcite in the symmetric pair centers in cscdbr@xmath2 , as determined from the epr spectra . in the perfect lattice ,
the first coordination shell of a cd@xmath14 ion has the radius of 0.277 nm ; in the symmetric dimer center six br@xmath108 ions nearest to the cd@xmath14 vacancy are forced outwards , but the displacements of the pr@xmath8 ions along the dimer axis are so large that the corresponding pr@xmath8br@xmath108 distance decreases down to 0.275 nm . on the contrary , the terminal triangles of the br@xmath108 ions lag behind the impurity ions and the corresponding interionic distance increases up to 0.290 nm .
these values agree with the distances ( 0.264 and 0.286 nm ) between the yb@xmath8 and the nearest br@xmath108 ions in the intrinsic dimer units in cs@xmath2yb@xmath116br@xmath117 @xcite .
taking into account the local lattice deformation , we obtained from eqns .
( [ eq5 ] , [ eq6 ] , [ eq7 ] ) the following crystal - field parameters ( values are in @xmath22 ) for pr @xmath8 ion in the symmetric dimer center ( the electrostatic contribution to the quadrupolar crystal field component @xmath118 was computed exactly with the ewald method ) : @xmath119 values of the model parameters @xmath120 ; @xmath121 have been determined by comparing the calculated and measured total splittings of the ground state @xmath5 and of the excited multiplet @xmath32 , the shielding factors have been fixed at @xmath122 , @xmath123 in accordance with the theoretical estimations @xcite . thus obtained crystal field parameters ( column @xmath124 of table 4 ) were used as starting values for the fitting procedure constrained by the additional conditions that the calculated hfs had to agree with the experimental data .
the final values of the crystal field parameters for the pr@xmath8 centers as compared with the corresponding parameters of tm@xmath8 , ho@xmath8 and nd@xmath8 centers are given in column @xmath125 of table 4 . despite the large trigonal distortion of the nearest bromine octahedron ,
the main features of the crystal field splittings are determined by the cubic component of the crystal field .
the @xmath126 and @xmath127 parameters involve the largest changes from the initial ( calculated ) to the final ( fitted ) values .
these differences are most probably caused by the overestimated displacements of the six br@xmath128br@xmath129 ions in the basis plane ( see table 5 ) .
the calculated energy levels of the pr@xmath8 ions presented in table 1 were obtained by diagonalizing the hamiltonian ( 1 ) in the space of 516 states ( terms @xmath130 , @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 were taken into account ) with fitted barycenters of the free - ion multiplets .
the measured splittings of the @xmath5 , @xmath31 @xmath32 , @xmath6 , @xmath7 , @xmath58 , @xmath136 , @xmath137 , @xmath23 multiplets of pr@xmath8 are satisfactorily described by our final set of crystal - field parameters ( column @xmath125 in table 4 ) .
the computed hfs of @xmath47 and @xmath45 crystal field singlets consists of three doublets with the total width @xmath138 of no more than 0.02 @xmath22 which can not be resolved in the optical spectra .
the quadrupolar contributions to the hfs are very small ( @xmath139 @xmath22 ) , the calculated hfs of @xmath43 doublets is actually equidistant with intervals of no more than 0.1 @xmath140 an example of the simulated envelope of the absorption line @xmath141 where individual transitions between different electron - nuclear states are presented by gaussians of 0.04 @xmath22 width , is given in fig .
the observed irregular shape of the hfs of some absorption lines ( see fig . 2a and ref .
@xcite ) is most probably caused by crystal fields of low symmetry .
we discussed this point in more details in another publication @xcite .
the most essential difference between results of our calculations and fitting procedures performed in @xcite is connected with assignments of symmetry types of some crystal field energy levels . in particular , symmetries of the third ( @xmath142 ) and fourth ( @xmath47 ) sublevels of the @xmath4 multiplet and of the second ( @xmath143 ) and third ( @xmath45 ) sublevels of the ground @xmath5 multiplet are inverted in @xcite ( it should be noted that our scheme of crystal field energies of the ground state agrees with the assignment of ref .
additional measurements of polarized spectra are to be carried out to check our predictions .
even at the liquid helium temperature the hfs of the most of excited crystal field sublevels is masked by the spontaneous relaxation broadening .
strong electron - phonon interaction effects in cscdbr@xmath2:re@xmath8 crystals originate from the specific density of phonon states which has large maxima in the low frequency region ( 2040 @xmath22 ) in the perfect crystal lattice ( see fig . 3 and ref .
the hamiltonian of the electron - phonon interaction , expanded in a power series in ion displacements from their equilibrium positions , can be written as follows @xmath144 + \nonumber \\
+ \frac 12{\sum\limits_{s\alpha \beta } } v_{\alpha \beta } ( s)[u_\alpha ( s)-u_\alpha ( \rm re)][u_\beta ( s)-u_\beta ( \rm re ) ] ... ; \label{eq8 } \\ v_\alpha ( s)={\sum\limits_{pk}}b_{p,\alpha } ^k(s)o_p^k , v_{\alpha \beta } ( s)={\sum\limits_{pk}}b_{p,\alpha \beta } ^k(s)o_p^k , \nonumber\end{aligned}\ ] ] where @xmath145 $ ] is the difference between dynamic displacements of the ligand ion @xmath106 and the re ion , and @xmath146 @xmath147 are the coupling constants . the probability of the one - phonon transition between the initial @xmath148 and final @xmath149 states of the re ion with the energy gap @xmath150 can be represented as @xmath151 where @xmath152 is the phonon occupation number , and @xmath153 are spectral representations of the green s functions for differences between the ion displacements @xcite .
we performed a calculation of the relaxation rates for all the crystal field sublevels within the manifolds of @xmath4 , @xmath6 and @xmath7 multiplets using the phonon green s functions of the perfect ( cscdbr@xmath2 ) and locally perturbed ( impurity dimer centers in cscdbr@xmath2:pr@xmath8 ) crystal lattices obtained in @xcite .
the formation of a dimer leads to a strong perturbation of the crystal lattice ( mass defects in the three adjacent cd@xmath14 sites and large changes of force constants ) .
as it has been shown in @xcite , the local spectral density of phonon states essentially redistributes and several localized modes corresponding to different representations of the dimer point symmetry group @xmath15 appear near the boundary of the continuous phonon spectrum of the unperturbed lattice . as an example of changes in the phonon spectrum , we present in fig .
3 spectral densities @xmath154 of displacement - displacement autocorrelation functions @xmath155 in the perfect ( @xmath91 = cd@xmath14 ) and perturbed ( @xmath91 = pr@xmath8 ) lattices at zero temperature .
the contributions to @xmath156 from the localized modes of @xmath157 ( 196.5 @xmath22 ) and @xmath158(190 @xmath3 ) symmetry , and to @xmath159 from the localized modes of @xmath160 ( 191.7 @xmath22 ) and @xmath161(191.2 @xmath22 ) symmetry are presented by lorentzians with the proper weights and the full width of 2 @xmath22 .
we took into account interactions of pr@xmath8 ion with its nearest neighbors ( six br@xmath108 ions ) only .
values of coupling constants were computed in the framework of the ecm with the same parameters as those used in the crystal field calculations ( explicit expressions of the @xmath162 @xmath147 are given in @xcite ) .
matrix elements of electron operators @xmath163 were calculated with the eigenfunctions of the hamiltonian ( 1 ) .
calculated inverse lifetimes @xmath164 ( the sum is over states @xmath29 which belong to the same multiplet as the crystal field level @xmath89 does , with lower energies ) are given in table 3 .
it is seen that despite many simplifying approximations there is a good correlation between the measured linewidths and the relaxation rates obtained with the perturbed green s functions .
calculations with green s functions of the perfect lattice gave overestimated ( up to an order of magnitude higher ) values of relaxation rates .
thus , we conclude that the increased pr@xmath8-ligand elastic interaction , as compared to the cd@xmath14-br@xmath108 interaction , and corresponding enhancement of correlations between displacements of the impurity re@xmath8 ion and its neighbors strongly suppresses the electron - phonon coupling .
it should be noted that the measured width of the doublet @xmath165 is five times larger than the estimated total hfs width , though the one - phonon relaxation broadening of this level is not possible at low temperatures .
we suppose that this level , and two upper levels @xmath166 @xmath167 are essentially broadened due to the two - phonon relaxation .
the probability of the spontaneous emission of two phonons contains three terms which correspond to the first order contribution from the nonlinear electron - phonon interaction : @xmath168 to the second order contribution from the linear electron - phonon interaction : @xmath169 where @xmath170 and to the combined action of the first and second order transition amplitudes : @xmath171 the calculated sum of two - phonon relaxation rates @xmath172 corresponding to spontaneous transitions from the doublet @xmath173 to all lower sublevels of the @xmath4 multiplet , induced by the nonlinear terms in eqn .
( [ eq8 ] ) equals @xmath174 s@xmath175 and is almost exactly cancelled by the contribution @xmath176 s@xmath175 from crossed
terms of eqn .
( [ eq14 ] ) .
we have obtained the remarkably larger contribution to the inverse lifetime of this doublet ( @xmath177 s@xmath175 ) , that agrees satisfactorily with its measured linewidth , from the second order terms given in eqn .
( [ eq13 ] ) .
it should be noted that relaxation processes that involve excitations of localized modes of @xmath158 and @xmath157 symmetry dominate yielding more than a half of the total relaxation rate .
high resolution ( 10.005 @xmath22 ) infrared ( 20007000 @xmath22 ) absorption spectra of cscdbr@xmath2:pr@xmath8 ( 0.2@xmath0 ) were taken at low temperatures .
crystal field levels of the @xmath4 , @xmath32 , @xmath6 , and @xmath7 multiplets for the symmetric pair center [ pr@xmath8cd@xmath18 vacancy pr@xmath8 ] were found directly in absorption .
hyperfine structure , inhomogeneous and relaxation - induced widths of several crystal field levels were measured .
we performed crystal field calculations in the framework of the semiphenomenological exchange charge model and took into account local lattice deformation around an impurity center .
only two parameters of the model had to be determined by a comparison of calculated and experimental energy levels . thus obtained crystal field parameters were corrected by fitting the calculated hyperfine structure to the measured one .
it follows from our results that despite a large trigonal distortion of the bromine octahedra in the nearest surroundings of impurity pr@xmath8 ions , crystal field splittings are determined , mainly , by the cubic component of the crystal field . using the results of the crystal field calculations and of the lattice dynamics analysis we were able , for the first time , to calculate exactly ( without any additional fitting parameters ) one- and two - phonon spontaneous relaxation rates .
the calculated total relaxation rates reveal the linewidths which are in reasonable agreement with the measured ones .
the obtained information on relative efficiency of linear and nonlinear terms in the hamiltonian of the electron - phonon interaction in stimulation of the phonon emission is important for the thorough derivation of the theory of multiphonon relaxation in rare - earth compounds .
the peculiarities of the electron - phonon interaction effects in cscdbr@xmath2:pr@xmath8 crystals originate from the specific density of phonon states that extends by only @xmath178 @xmath22 in the perfect cscdbr@xmath2 lattice , and from the localized modes induced by the impurity dimer centers .
another important feature and , probably , the most interesting result of this study , is a strong suppression of the effective electron - phonon coupling due to a local increase of elastic forces in the activated crystal and the corresponding enhancement of correlations between displacements of the impurity re ion and its neighbors .
an encouraging interest of g. n. zhizhin is acknowledged .
this work was supported in part by the rfbr ( grant n 9902 16881 ) and the russian ministry of science ( grant n 08.02.28 of the program `` fundamental spectroscopy '' ) .
f. ramaz , r. m. macfarlane , j. c. vial , j. p. chaminade , and f. madeore , j. lumin . * 55 * , 173 ( 1993 ) .
e. antic - fidancev , m. lemaitre - blase , j. p. chaminade , and p. porcher , j. alloys compd . *
225 * , 95 ( 1995 ) .
k. m. murdoch and n. j. cockroft , phys .
b * 54 * , 4589 ( 1996 ) .
j. neukum , n. bodenschatz , and j. heber , phys .
b * 50 * , 3536 ( 1994 ) .
j. p. chaminade , r. m. macfarlane , f. ramaz , and j. c. vial , j. lumin . * 48 & 49 * , 531 ( 1991 ) .
j. heber , m. lange , m. altwein , b. z. malkin , and m. p. rodionova , j. alloys compd . *
275 - 277 * , 181 ( 1998 ) .
b. z. malkin , a. i. iskhakova , v. f. tarasov , g. s. shakurov , j. heber , and m. altwein , j. alloys compd . *
275 - 277 * , 209 ( 1998 ) .
g. w. burdick and f. s. richardson , j. alloys compd . *
275 - 277 * , 379 ( 1998 ) .
j. r. quagliano , n. j. cockroft , k. e. gunde , and f. s. richardson , j. chem .
phys . * 105 * , 9812 ( 1996 ) .
b. z. malkin , in spectroscopy of solids containing rare - earth ions , edited by a. a. kaplyanskii and r. m. macfarlane ( north - holland , amsterdam , 1987 ) , p.13 . b. z. malkin , a. i. iskhakova , s. kamba , j. heber , m. altwein , g. schaack , submitted to phys
b. r. b. barthem , r. buisson , f. madeores , j. c. vial , and j. p. chaminade , j. phys .
( paris ) * 48 * , 379 ( 1987 ) .
g. gorller - walrand and k. binnemans , in handbook on the physics and chemistry of the rare - earths , edited by k. a. gschneidner and leroy eyring ( north - holland , amsterdam , 1996 ) , vol .
23 , p. 121 .
m. a. h. mccausland and i. s. mackenzie , adv .
* 28 * , 305 ( 1979 ) .
g. netz , z. phys .
b * 63 * , 343 ( 1986 ) .
v. c. das and a. k. ray chaudhuri , j. phys .
c * 6 * , 1385 ( 1973 ) .
a. j. freeman and r. e. watson , phys . rev . *
127 * , 2058 ( 1962 ) .
e. clementi and c. roetti , atomic data and nuclear data tables ( academic press , new york , 1974 ) , vol .
14 , n 34 .
g. l. mcpherson , a. m. mcpherson , and j. a. atwood , j. phys .
* 41 * , 495 ( 1980 ) .
l. m. henling and g. l. mcpherson , phys .
b*16 * , 4756 ( 1977 ) .
v. f. tarasov , g. s. shakurov , b. z. malkin , a. i. iskhakova , j. heber , and m. altwein , jetp lett.*65 * , 559 ( 1997 ) .
m. p. hehlen and h. u. gudel , j. chem . phys . *
98 * , 1768 ( 1993 ) .
e. p. chukalina , m. n. popova , e. antic - fidancev , and j. p. chaminade , phys .
a , * 258 * , 375 ( 1999 ) .
m. n. popova , e. p. chukalina , b. z. malkin , s. k. saikin , phys .
b * 61 * , 7421 ( 2000 ) .
.positions ( @xmath22 ) and symmetries of energy levels for the symmetric dimer centers , and widths ( in brackets , @xmath22 ) of absorption lines from the ground state in cscdbr@xmath2:pr@xmath8 .
[ cols="^ , < , < , < , < , < , < " , ]
|
high resolution low - temperature absorption spectra of 0.2@xmath0 pr@xmath1 doped cscdbr@xmath2 were measured in the spectral region 20007000 @xmath3 .
positions and widths of the crystal field levels within the @xmath4 , @xmath5 , @xmath6 , and @xmath7 multiplets of the pr@xmath8 main center have been determined .
hyperfine structure of several spectral lines has been found .
crystal field calculations were carried out in the framework of the semiphenomenological exchange charge model ( ecm ) .
parameters of the ecm were determined by fitting to the measured total splittings of the @xmath5 and @xmath9 multiplets and to the observed in this work hyperfine splittings of the crystal field levels . one- and two - phonon relaxation rates
were calculated using the phonon green s functions of the perfect ( cscdbr@xmath2 ) and locally perturbed ( impurity dimer centers in cscdbr@xmath2:pr@xmath8 ) crystal lattice .
comparison with the measured linewidths confirmed an essential redistribution of the phonon density of states in cscdbr@xmath2 crystals doped with rare - earth ions .
|
graph embedding is one of the most important topics in topological graph theory . in particular ,
@xmath0-cell embeddings of graphs ( loops and multiple edges allowed ) have been widely studied . a @xmath0-cell embedding or map of a given graph @xmath1 on a closed surface of genus @xmath2 , @xmath9 , is an embedding on @xmath9 such that the complement of any face is homeomorphic to an open disk .
the closed surfaces could be either orientable or unorientable . in this paper
, we restrict ourselves to the orientable case .
let @xmath10 and @xmath11 denote the minimum and the maximum genus @xmath2 of the embeddings of @xmath1 , respectively .
there are many studies on determining these quantities and related problems @xcite .
assume @xmath1 has @xmath12 edges and @xmath13 vertices , and that @xmath1 is embedded in @xmath9 via @xmath3 .
in view of euler s characteristic formula , @xmath14 where @xmath15 is the number of faces of @xmath3 and @xmath16 is the betti number of @xmath1 .
thus , the largest possible value of @xmath2 is @xmath17 .
any embedding of @xmath1 in a closed orientable surface can be equivalently represented by a fatgraph , generated by @xmath1 @xcite .
a fatgraph generated by @xmath1 is the graph @xmath1 with a specified cyclic order of edges around ( i.e. , incident to ) each vertex of @xmath1 , i.e. the topological properties of the embedding are implied in the cyclic orderings of edges .
any variation of the local topological structure around a vertex , i.e. , the cyclic order of edges around the vertex , may change the topological properties of the whole embedding , e.g. the genus of the embedding .
plane permutations @xcite were recently used to study hypermaps , in particular to enumerate hypermaps with one face .
since maps are particular hypermaps , we shall employ plane permutations in order to study graph embeddings .
the paper is organized as follows : in section @xmath0 , we establish some basic facts on hypermaps . in section @xmath18 , we study embeddings with one face .
we ask which local embeddings ( reembeddings ) of a fixed vertex do not affect the topological genus . by changing the local embedding , we mean changing the cyclic order of edges around the vertex .
we shall show that the probability of preserving the genus is at least @xmath19 for reembedding any vertex @xmath8 of degree ( i.e. , valence ) @xmath7 .
in addition we show that there exists at least one alternative way to reembed a vertex @xmath8 preserving the genus if @xmath20 . in section @xmath21 , we study embeddings with more than one face : given a genus @xmath2 embedding @xmath3 of the graph @xmath1 and a vertex of @xmath1 , how many different ways of reembedding the vertex such that the resulting embedding @xmath4 is of genus @xmath5 ? we give a formula to compute this quantity and the local minimal genus achieved by reembedding , as well as an easy - to - check necessary condition for an embedding of @xmath1 to be of minimum genus .
let @xmath22 denote the group of permutations , i.e. the group of bijections from @xmath23=\{1,\dots , n\}$ ] to @xmath23 $ ] , where the multiplication is the composition of maps .
we shall use the following two representations of a permutation @xmath24 : + _ two - line form : _ the top line lists all elements in @xmath23 $ ] , following the natural order .
the bottom line lists the corresponding images of elements on the top line , i.e. @xmath25 _ cycle form : _ regarding @xmath26 as a cyclic group , we represent @xmath24 by its collection of orbits ( cycles ) .
the set consisting of the lengths of these disjoint cycles is called the cycle - type of @xmath24 .
we can encode this set into a non - increasing integer sequence @xmath27 , where @xmath28 , or as @xmath29 , where we have @xmath30 cycles of length @xmath31 .
a cycle of length @xmath32 will be called a @xmath32-cycle . for a permutation @xmath24 on @xmath23 $ ]
, we denote its total number of cycles by @xmath33 , and we denote @xmath34 the partition of @xmath23 $ ] induced by the cycles of @xmath24 , i.e. , every set of elements in a same cycle of @xmath24 contributes a part ( or block ) in @xmath34 .
a map or fatgraph having @xmath35 edges is a triple of permutations @xmath36 on @xmath37 $ ] where @xmath38 is a fixed - point free involution and @xmath39 .
this can be seen as follows : we denote the two ends of an edge as half edges and label the half edges using the labels from the set @xmath37 $ ] so that each label appears exactly once .
this induces two permutations @xmath38 and @xmath40 , where @xmath38 is a fixed point free involution , whose cycles consist of the labels of the two half edges of the same ( untwisted ) edge and @xmath40-cycles represent the counterclockwise cyclic arrangement of all half edges incident to the same vertex .
@xmath39-cycles are called the boundaries components .
the topological genus of a map @xmath36 satisfies @xmath41 any map induces a unique graph @xmath1 , via @xmath42 , where each block corresponds to a @xmath1-vertex and each @xmath38-cycle determines a @xmath1-edge .
hypermaps represent a generalization of maps by allowing hyper - edges , i.e. , triples @xmath43 , where @xmath39 and @xmath38 is not necessarily fixed point free .
we can also define genus @xmath2 of a hypermap @xmath36 on @xmath23 $ ] , e.g. , @xcite , by @xmath44 we will also call the pair @xmath45 the underline graph of the hypermap although it does not induce a conventional graph .
furthermore , any hypermap @xmath46 having @xmath1 as the underlying graph is called an embedding of @xmath1 .
a cyclic plane permutation on @xmath23 $ ] is a pair @xmath47 where @xmath48 is an @xmath35-cycle and @xmath24 is an arbitrary permutation on @xmath23 $ ] .
the permutation @xmath49 is called the diagonal of @xmath50 .
[ 2def1 ] given @xmath51 , a cyclic plane permutation @xmath47 can be represented by two aligned rows : @xmath52 indeed , @xmath53 is determined by the diagonal - pairs ( cyclically ) in the two - line representation here , i.e. , @xmath54 for @xmath55 , and @xmath56 .
w.l.o.g .
we shall assume @xmath57 and by `` the cycles of @xmath47 '' mean the cycles of @xmath24 .
we will refer to the blocks in @xmath34 as @xmath50-vertices or vertices for short , and elements in a vertex half edges .
hypermaps having one face , @xmath36 , can be represented as cyclic plane permutations @xmath58 . in particular , one - face maps are cyclic plane permutations where the diagonals are fixed - point free involutions .
any cyclic plane permutation having a fixed - point free involution diagonal encodes a one - face embedding of some graph .
a fatgraph with one boundary component is displayed in figure [ fig - fig1 ] .
the two presentations are showing the same one - face map , whose corresponding cyclic plane permutation reads @xmath59 edges . ] given a cyclic plane permutation @xmath47 on @xmath23 $ ] and a @xmath50-vertex @xmath8 , the localization of @xmath50 at @xmath8 , @xmath60 is the cyclic plane permutation @xmath61 which is obtained by deleting all columns not containing half edges incident to @xmath8 in the two - line representation of @xmath47 .
let @xmath62 denote the diagonal of @xmath63 , i.e. , @xmath64 .
note that even if @xmath65 is a map , @xmath66 is not necessarily a fixed - point free involution .
for example , given @xmath65 @xmath67 where @xmath68 , let @xmath69 .
then , @xmath70 we arrive at @xmath71 a set of consecutive diagonal - pairs in @xmath47 is called a diagonal block . in
above example , @xmath72 is a diagonal block .
it is completely determined by its corners , in this case , the lower left corner , @xmath73 , as well as the upper right corner , @xmath21 .
the diagonal block is denoted by @xmath74 . given a cyclic plane permutation @xmath47 on @xmath23 $ ] and a sequence @xmath75 on @xmath76 $ ] , let @xmath77 , i.e. @xmath78 is acted upon by @xmath79 via translation its indices , and @xmath80 .
this induces the new cyclic plane permutation @xmath81 having by construction the same diagonal as @xmath65 .
equivalently , the two - line representation of @xmath81 can be obtained by permuting the diagonal - pairs of @xmath65 . in the following
, a cyclic plane permutation written in the form like @xmath81 , always means that it is obtained from @xmath65 by permuting diagonal - pairs by @xmath79 .
[ 2lem2 ] let @xmath82 be cyclic plane permutations such that @xmath83 and @xmath24 and @xmath84 exclusively differ at the vertex @xmath8 .
then , there exists some @xmath79 such that @xmath85 and furthermore @xmath86 .
assume @xmath47 and @xmath87 are respectively @xmath88 & \cdots & s_{i_{k-2 } } & { \cdots } & s_{i_{k-1}}\quad \ar@{--}[dl ] \cdots\\ \cdots \pi(s_{i_0})\ar@{--}[ur ] & { \cdots } & \pi(s_{i_1 - 1 } ) & \pi(s_{i_1 } ) & \cdots & \pi(s_{i_{k-2}})\ar@{--}[ur ] & { \quad \cdots \quad } & \pi(s_{i_{k-1 } } ) \cdots } } \right),\\ & \left ( \vcenter{\xymatrix@c=0pc@r=1pc { \cdots s'_{i_0 } & s'_{i_0 + 1 } & { \cdots } & s'_{i_1}\ar@{--}[dl ] & \cdots & s'_{i_{k-2 } } & { \cdots } & s'_{i_{k-1}}\quad \ar@{--}[dl ] \cdots\\ \cdots \pi'(s'_{i_0})\ar@{--}[ur ] & { \cdots } & \pi'(s'_{i_1 - 1 } ) & \pi'(s'_{i_1 } ) & \cdots & \pi'(s'_{i_{k-2}})\ar@{--}[ur ] & { \quad \cdots \quad } & \pi'(s'_{i_{k-1 } } ) \cdots } } \right ) , \end{aligned}\ ] ] where we assume @xmath89 and @xmath90 .
since by assumption @xmath24 and @xmath84 only differ at the vertex @xmath8 , we have @xmath91 for @xmath92 . furthermore @xmath83 implies @xmath93 .
+ _ claim . _ @xmath94 .
+ to show this we observe that for fixed @xmath95 , each diagonal block @xmath96 equals the diagonal block @xmath97 for some @xmath98 , i.e. , @xmath99 & s'_{i_j+2 } & \cdots & s'_{i_{j+1}}\ar@{--}[dl]\\ \pi'(s'_{i_j } ) & \pi'(s'_{i_j+1 } ) & { \cdots } & \pi'(s'_{i_{j+1}-1 } ) & } } \right)= \left ( \vcenter{\xymatrix@c=0pc@r=1pc { & s_{i_l+1}\ar@{--}[dl ] & s_{i_l+2 } & \cdots & s_{i_{l+1}}\ar@{--}[dl]\\ \pi(s_{i_l } ) & \pi(s_{i_l+1 } ) & { \cdots } & \pi(s_{i_{l+1}-1 } ) & } } \right ) \end{aligned}\ ] ] to prove this , we observe that by construction for fixed @xmath95 there exists some index @xmath100 such that @xmath101 holds .
this implies , @xmath102 in case of @xmath103 , we have @xmath104 and derive @xmath105 iterating we arrive at @xmath106 , whence the two diagonal blocks are equal , the claim follows and the proof of the lemma is complete .
[ 2def3 ] given a cyclic plane permutation @xmath47 on @xmath23 $ ] and its localization at @xmath8 , @xmath107 .
suppose @xmath108 is such that @xmath109 .
then the inflation of @xmath108 w.r.t .
@xmath50 is the cyclic plane permutation @xmath110 , obtained from @xmath108 by substituting each diagonal - pair with the diagonal block in @xmath50 having the diagonal - pair as its corners .
let @xmath111 then the inflation of @xmath108 w.r.t .
@xmath65 is @xmath112 [ lem211 ] let @xmath113 .
then we have @xmath114 and @xmath84 differs from @xmath24 only at the vertex @xmath8 . by construction , @xmath115 . any half edge not contained in @xmath8 , is located inside the respective diagonal blocks , whence @xmath24 and @xmath84 are equal on these half edges .
all @xmath8-half edges contribute exactly one block in @xmath34 and @xmath116 , since @xmath109 .
accordingly , we have @xmath83 , completing the proof of the lemma . combining lemma [ 2lem2 ] and [ lem211 ]
, we obtain [ 2thm2 ] let @xmath47 be a cyclic plane permutation .
let @xmath117 and let @xmath118 .
then there is a bijection between @xmath119 and @xmath120 and we have the commutative diagram : @xmath121^{\text{\rm loc}_\nu } \ar[r]^{h:\quad par_{\pi}=par_{\pi^h } } & { ( s',\pi')}\\ ( s_{\nu},\pi_{\nu})\ar[r]_{h:\quad par_{\pi_{\nu}}=par_{\pi_{\nu}^h } } & ( s_{\nu}^h,\pi_{\nu}^h)\ar[u]^{\text{\rm inf}_{\mathfrak{p } } } } \ ] ]
let @xmath47 be a cyclic plane permutation with the underline graph @xmath122 . then , @xmath123 is an embedding of @xmath1 iff @xmath124 for some @xmath79 and @xmath125 .
if @xmath123 is an embedding of @xmath122 , then , by definition , @xmath126 .
thus , @xmath127 . clearly , there exists some @xmath79 such that @xmath128 . by construction
we have @xmath129 .
hence , @xmath124 for some @xmath79 .
the converse is clear , whence the lemma .
this lemma shows that any one - face embedding is originated by the action of some @xmath79 on a cyclic plane permutation .
explicitly , by permuting diagonal - pairs based on a fixed one - face embedding @xmath65 of the graph such that @xmath130 .
let @xmath47 be a one - face map with the underline graph @xmath131 .
fixing the ( local ) embedding of all vertices of @xmath1 as in @xmath50 but the vertex @xmath8 , each local embedding of @xmath8 leading to a one - face map @xmath123 corresponds to a @xmath132 such that @xmath133 and @xmath84 differs from @xmath24 only at the vertex @xmath8 . according to the bijection between @xmath132 and @xmath79 in theorem [ 2thm2 ] ,
the number of different embeddings of @xmath8 keeping one face is equal to the number of different @xmath79 such that @xmath108 and @xmath134 having the same underlying graph .
we denote this number by @xmath135 .
moreover , the half edges contained in @xmath8 split the cyclic plane permutation into @xmath136 diagonal blocks .
we can view these diagonal blocks as they are arranged in a circular fashion , as displayed in figure [ fig - circ ] .
to reembed @xmath8 means to permute these diagonal blocks circularly . ]
let @xmath137 denote the set of cyclic plane permutations having diagonal @xmath138 , where @xmath138 is a fixed permutation .
note @xmath139 iff @xmath140 .
thus , @xmath141 enumerates the ways to write @xmath138 as a product of an @xmath35-cycle with another permutation .
a partition @xmath142 of @xmath35 is written as @xmath143 .
let @xmath144 .
we write @xmath145 if @xmath146 can be obtained by splitting one @xmath147-block into @xmath148 non - zero parts .
let furthermore @xmath149 denote the number of different ways to obtain @xmath147 from @xmath146 by merging @xmath150 @xmath146-blocks into one , where @xmath151 and @xmath152 denote the number of blocks in the partitions @xmath146 and @xmath147 , respectively .
@xcite[2cor1 ] let @xmath153 denote the number of @xmath154 having @xmath32 cycles , where @xmath138 is of cycle - type @xmath142 .
let @xmath155 denote the number of permutations of cycle - type @xmath142 . then
, @xmath156}.\ ] ] note that the localizations @xmath108 and @xmath134 have the same underlying hyper - graph @xmath1 , iff both of them belong to @xmath137 where @xmath157 and @xmath158 . therefore ,
if @xmath62 has cycle - type @xmath142 , @xmath159 . as a ressult we obtain [ 3thm2 ] let @xmath3 be a one - face embedding of @xmath1 , and @xmath8 be a vertex of @xmath1 with @xmath20 .
then there exists at least one additional way to reembed @xmath8 such that the obtained embedding @xmath4 has the same genus as @xmath3 .
assume @xmath160 and @xmath3 is localized at @xmath8 @xmath161 where @xmath162 .
firstly , if @xmath163 and @xmath164 , i.e. case @xmath18 of ( * ? ? ?
* lemma @xmath18 ) , then there exists at least one additional way to reembed @xmath8 preserving genus .
otherwise , we have @xmath165 . in this case , @xmath166 whence @xmath167 it remains to show that if @xmath160 we have @xmath168 in all cases . to this end
, we apply a formula for @xmath169 due to stanley @xcite . if @xmath170 , then @xmath171 where @xmath172 ranges over all non - negative integer solutions of the equation @xmath173 . applying stanley s formula we can compute @xmath135 ,
if @xmath174 as @xmath175 the simplification of the summation is stems from the following formula @xcite @xmath176 it is not hard to see that @xmath168 if @xmath160 .
similarly , if @xmath177 and @xmath160 , we have @xmath178\\ & = & \frac{2(d-1)!}{d+1}(1-{d\choose \frac{d}{2}}^{-1 } ) . \end{aligned}\ ] ] if @xmath179 and @xmath180 , we have @xmath181\\ & = & \frac{2(d-1)!}{d+1}(1+{d\choose \frac{d}{2}}^{-1 } ) .
\end{aligned}\ ] ] in both cases , if @xmath160 , it is straightforward to show that @xmath168 , since both @xmath182 and @xmath183 are increasing functions of @xmath184 .
accordingly , in all cases , if @xmath160 , then @xmath168 , completing the proof .
note in case of a vertex @xmath8 having degree @xmath185 or @xmath0 , the situation is clear .
thus it remains to consider the case @xmath186 .
for such a vertex , a reembedding preserving genus can be impossible .
for example , @xmath187 is the unique decomposition of @xmath62 .
[ 3cor3 ] any even permutation on @xmath23 $ ] with @xmath188 has at least two different factorizations into two @xmath35-cycles .
since @xmath189 and both @xmath190 as well as @xmath191 have only one cycle , @xmath62 is an even permutation .
theorem [ 3thm2 ] implies that @xmath62 has at least @xmath0 factorizations into two @xmath35-cycles .
[ 3cor4 ] let @xmath1 be a graph having @xmath192 vertices of degree no less than @xmath21 .
if there exists a one - face embedding of @xmath1 , then there are at least @xmath193 one - face embeddings of @xmath1 .
let @xmath65 be a one - face embedding of @xmath1 and @xmath134 its localization at @xmath8 .
suppose @xmath62 has cycle - type @xmath170 where @xmath194 , then the probability @xmath195 of a reembedding of @xmath8 to be one - face satisfies @xmath196 in particular , for any vertex @xmath8 , @xmath197 . in zagier @xcite , it was proved that @xmath198 since there are @xmath199 different ways to embed @xmath8 , eq
. immediately follows in view of @xmath200 . clearly , we have @xmath201 , whence the second assertion .
if there exists a one - face embedding of @xmath1 , then the probability of a random embedding of @xmath1 to have one face is at least @xmath202 .
in this section , we generalize cyclic plane permutations to general plane permutations .
this puts us in position to study graph embeddings having @xmath32 faces .
although it is hard to determine @xmath203 and @xmath204 , as well as the genus distribution for a given graph @xmath1 , we will show that locally these quantities can be more easily obtained .
[ 4def1 ] a plane permutation on @xmath23 $ ] is a pair , @xmath50 , of permutations @xmath78 and @xmath24 on @xmath23 $ ] .
the permutation @xmath49 is called the diagonal of @xmath50 . if @xmath78 has @xmath32 cycles , we write @xmath205 .
we call blocks @xmath214 the cycles of the plane permutation .
if the face @xmath215 is incident to a @xmath50-vertex @xmath8 , the corresponding cycle is said to be incident to @xmath8 . since every embedding
having @xmath32 faces can be represented by a triple @xmath43 , where @xmath216 and @xmath217 has @xmath32 cycles , any embedding can be expressed via a plane permutation @xmath218 .
let @xmath219 denote the set of half edges contained in the face @xmath220 .
[ 4lem1 ] let @xmath8 be a vertex of the graph @xmath1 and @xmath3 be an embedding of @xmath1 , where @xmath8 is incident to @xmath221 faces , @xmath222 , for @xmath223 .
let @xmath4 be an embedding , obtained by reembedding @xmath8 such that @xmath8 is incident to @xmath224 faces , @xmath225 , for @xmath226 .
then we have @xmath227 let @xmath228 be two embeddings represented by @xmath205 and @xmath229 , respectively , such that @xmath230 and @xmath83
. note that @xmath3 and @xmath4 only differ w.r.t .
the cyclic order of the half edges around @xmath8 .
thus , for @xmath231 , we have @xmath232 .
clearly , any face @xmath220 of @xmath3 can be expressed as the sequence @xmath233 for any @xmath234 .
the lemma is implied by the following _ claim .
_ any face @xmath235 of @xmath4 either intersects some @xmath3-face @xmath222 for @xmath223 and is incident to @xmath8 or it coincides with an @xmath3-face @xmath220 not incident to @xmath8 .
+ suppose @xmath235 does not intersect any @xmath222 for @xmath223 .
for any @xmath236 , @xmath231 holds and by construction @xmath237 . as a result , @xmath238 , i.e. @xmath235 coincides with an @xmath3-face @xmath220 that is not incident to @xmath8 .
if @xmath235 intersects some @xmath3-face @xmath222 for @xmath223 , we shall prove that @xmath235 is incident to @xmath8 .
assume the half edge @xmath239 is contained in the @xmath3-face @xmath240 as well as in the face @xmath235 of @xmath4 .
then , @xmath241 where @xmath242 is the first half edge of @xmath8 that appears in @xmath240 .
since @xmath232 if @xmath231 , we have @xmath243 , whence the entire subsequence from @xmath244 to @xmath242 in @xmath240 appears also in @xmath235 .
in particular we have @xmath245 , which means that @xmath235 is incident to @xmath8 and the claim follows .
let @xmath3 be an embedding of the graph @xmath1 and @xmath8 be a vertex of @xmath1 , where @xmath8 is incident to @xmath221 faces in @xmath3 .
assume @xmath3 is represented by @xmath205 .
similar to the situation of one - face maps , we can define the localization at @xmath8 which is a plane permutation having @xmath221 cycles , @xmath246 , and that is obtained as follows : the @xmath221 cycles of @xmath246 are obtained from the @xmath221 cycles of @xmath50 incident to @xmath8 by deleting all columns having no half edges of @xmath8 .
let @xmath62 denote the diagonal of @xmath246 . by construction
, we have @xmath247 , having @xmath221 cycles . given a plane permutation @xmath248 , where @xmath249 , we can inflate w.r.t .
@xmath50 into an embedding of @xmath1 as in the case of cyclic plane permutations .
namely , we substitute each diagonal - pair with the corresponding diagonal block in @xmath50 and then add any @xmath50-cycles containing half edges not incident to @xmath8
. fix an embedding @xmath3 , represented by the plane permutation @xmath250 , of genus @xmath251 .
we compute in the following the distribution of genera resulting from reembedding the vertex @xmath8 .
let @xmath252 denote the number of different embeddings , @xmath4 , coming from reembedding @xmath8 such that @xmath253 and denote the cycle - type of @xmath62 as @xmath254 .
then we have according to lemma [ 4lem1 ] , we have the following situation : @xmath259 is reorganized into @xmath260 @xmath4-faces , @xmath261 and any other @xmath4-face coincides with some @xmath3-face not incident to @xmath8 . _ claim @xmath185 .
_ given @xmath3 represented by @xmath250 , any reembedding of @xmath8 , @xmath4 represented by @xmath256 satisfies @xmath266 .
+ suppose @xmath8 is incident to @xmath221 @xmath3-faces , @xmath267 .
furthermore , suppose the @xmath4-cycle of face @xmath225 reads : @xmath268 where @xmath269 .
then , by the same argument as in the proof for the lemma [ 2lem2 ] , the diagonal block @xmath270 is also a diagonal block in @xmath3 , which in turn implies @xmath266 .
+ _ claim @xmath0 .
_ suppose @xmath262 is a localization such that @xmath266 and @xmath271
. then @xmath262 can be inflated into an embedding @xmath4 such that @xmath253 , holds .
+ suppose @xmath262 is given by : @xmath272 inflating every diagonal - pair into a diagonal block w.r.t .
@xmath3 and adding the @xmath3-cycles which are not incident to @xmath8 , we obtain an embedding @xmath4 with @xmath258 more faces than @xmath3 , i.e. , @xmath253 . by construction ,
@xmath3 and @xmath4 only differ by cyclic rearrangement of the half edges around @xmath8 . for @xmath279 ,
the assertion is clear , whence we can assume w.l.o.g .
@xmath280 . for any permutation @xmath38 on @xmath23 $ ] of cycle type @xmath142 and @xmath281
, we have @xmath282 where @xmath283 is the identity permutation on @xmath23 $ ] which obviously has @xmath35 cycles .
therefore , in case of @xmath281 , @xmath284 suppose for any @xmath142 with @xmath285 holds @xmath286 let @xmath287 be a permutation on @xmath23 $ ] of cycle type @xmath288 and @xmath289 . since @xmath290 , we can always find @xmath291 and @xmath292 such that @xmath291 and @xmath292 are in different cycles of @xmath287 .
let @xmath293 .
thus , @xmath38 must be of cycle type @xmath146 for some @xmath146 such that @xmath294 . by assumption
, there exists a relation @xmath295 such that @xmath78 has only one cycle and @xmath24 has @xmath296 cycles .
then , @xmath297 note @xmath298 has the number of cycles either @xmath299 or @xmath300 .
the latter is impossible because it would contradict the bound established in proposition [ 2pro1 ] .
hence , for any @xmath288 with @xmath289 , @xmath301 which completes the proof of the proposition .
[ 4cor2 ] let @xmath3 be a fixed embedding of the graph @xmath1 .
then for any vertex @xmath8 with localization @xmath302 , there exists for any @xmath303 an embedding @xmath4 of @xmath1 such that @xmath253 .
according to corollary [ 2cor1 ] , we have @xmath304 as long as @xmath305 for some @xmath306 holds .
furthermore , proposition [ 2pro2 ] implies @xmath307 .
therefore , for any @xmath308 reembedding @xmath8 can lead to an embedding where @xmath8 is incident to @xmath184 faces .
accordingly , euler s characteristic formula , implies @xmath309 completing the proof of the corollary .
this result is similar to the result in @xcite , where it was shown for any @xmath310 , there exists an embedding of @xmath1 on @xmath9 .
however , while it is very hard to obtain @xmath203 and @xmath204 , we obtain easily the local minimum and the local maximum .
suppose we are given two vertices , such that there exists no face of @xmath3 incident to both , then we call these two vertices @xmath3-face disjoint . in view of lemma [ 4lem1 ] ,
corollary [ 4cor2 ] has the following implication .
[ 4cor22 ] let @xmath3 be an embedding of the graph @xmath1 .
if the vertices @xmath311 , @xmath312 , are mutually @xmath3-face disjoint , then there exists an embedding @xmath4 of @xmath1 for any @xmath313 such that @xmath314 .
the following corollary provides a necessary condition for an embedding of @xmath1 to be of maximum genus as well as an easy - to - check necessary condition for an embedding of @xmath1 to be of minimum genus .
[ 4cor3 ] if @xmath3 is an embedding of the graph @xmath1 with genus @xmath11 , then every vertex is incident to at most @xmath0 faces in @xmath3 .
furthermore , if @xmath3 is an embedding of the graph @xmath1 with genus @xmath10 , and @xmath315 is a localization at @xmath8 , then @xmath316 the fact that if there exists a vertex incident to at least @xmath18 faces in an embedding , an embedding with higher genus always exists , is well known , see e.g. , in @xcite .
however , to the best of our knowledge , given an embedding @xmath3 , there is no simple characterization in order to determine if there exists an embedding of lower genus .
corollary [ 4cor3 ] gives a sufficient condition , i.e. , if @xmath317 for some vertex @xmath13 , then there exists an embeding of lower genus .
a. duke , the genus , regional number , and betti number of a graph , canad .
j. math . 18
( 1966 ) , 817 - 822 .
j. edmonds , a combinatorial representation for polyhedral surfaces , notices amer .
, vol . 7 , ( 1960 ) a646 .
m. jungerman , a characterization of upper - embeddable graphs , trans .
241 ( 1978 ) , 401 - 406 .
m. kotrbk , martin koviera , locally - maximal embeddings of graphs in orientable surfaces , the seventh european conference on combinatorics , graph theory and applications crm series volume 16 , 2013 , pp 215 - 220 .
lando , a.k .
zvonkin , graphs on surfaces and their applications , encyclopaedia math .
141 , springer- verlag , berlin , 2004 .
liu , the maximum non - orientable genus of a graph ( in chinese ) , scientia sinica ( special issue on math ) , i ( 1979 ) , 191 - 201 .
e. nordhaus , b. stewart , a. white , on the maximum genus of a graph , j. combin .
theory b , 11 ( 1971 ) , 258 - 267 .
l. nebesky , a new characterizations of the maximum genus of graphs , czechoslovak math .
j. , 31 ( 106 ) ( 1981 ) , 604 - 613 .
stanley , factorization of permutation into @xmath35-cycles , discrete math .
37 ( 1981 ) , 255 - 262 .
r. sprugnoli , riordan array proofs of identities in gould s book , 2006 . c. thomassen
, the graph genus problem is np - complete , j. algorithms 10 ( 1989 ) 568 - 576 .
|
the @xmath0-cell embeddings of graphs on closed surfaces have been widely studied .
it is well known that ( @xmath0-cell ) embedding a given graph @xmath1 on a closed orientable surface is equivalent to cyclically ordering the edges incident to each vertex of @xmath1 . in this paper , we study the following problem : given a genus @xmath2 embedding @xmath3 of the graph @xmath1 and a vertex of @xmath1 , how many different ways of reembedding the vertex such that the resulting embedding @xmath4 is of genus @xmath5 ? we give formulas to compute this quantity and the local minimal genus achieved by reembedding . in the process
we obtain miscellaneous results .
in particular , if there exists a one - face embedding of @xmath1 , then the probability of a random embedding of @xmath1 to be one - face is at least @xmath6 , where @xmath7 denotes the vertex degree of @xmath8 .
furthermore we obtain an easy - to - check necessary condition for a given embedding of @xmath1 to be an embedding of minimum genus .
* keywords : * graph embedding ; genus ; plane permutation ; hypermap mathematics subject classifications 2010 : 05c30 ; 05c10 ; 97k30
|
clusters and superclusters of galaxies have been studied extensively both for their intrinsic properties and to investigate the dark matter in the universe , the baryon content of the universe , large - scale structure , evolution , and cosmology . for previous reviews
see zwicky ( 1958 ) , bahcall ( 1977 , 1988 , 1996 ) , oort ( 1983 ) , dressler ( 1984 ) , rood ( 1988 ) , and peebles ( 1993 ) . in these lectures
i discuss the following topics and their implications for structure formation and cosmology .
section 2 : : : optical properties of galaxy clusters section 3 : : : x - ray properties of galaxy clusters section 4 : : : the baryon fraction in clusters section 5 : : : cluster masses section 6 : : : where is the dark matter ?
section 7 : : : the mass function of clusters section 8 : : : quasar - cluster association section 9 : : : superclusters section 10 : : : the cluster correlation function section 11 : : : peculiar motions of clusters section 12 : : : some unsolved problems a hubble constant of @xmath1 is used throughout .
clusters of galaxies are bound , virialized , high overdensity systems of galaxies , held together by the clusters self gravity .
rich clusters contain , by traditional definition ( abell 1958 ) , at least 30 galaxies brighter than @xmath2 ( where @xmath3 is the magnitude of the third brightest cluster member ) within a radius of @xmath4 of the cluster center .
this galaxy count is generally defined as the richness of the cluster .
the galaxies in rich clusters move with random peculiar velocities of typically @xmath5 ( median line - of - sight velocity dispersion ) .
this motion corresponds to a typical rich cluster mass ( within @xmath6 ) of @xmath7 .
in addition to galaxies , all rich clusters contain an intracluster medium of hot plasma , extending as far as the main galaxy concentration ( @xmath8 ) .
the typical temperature of the hot intracluster gas is @xmath9 kev , with a range from @xmath10 to 14 kev ; the central gas density is @xmath11 electrons @xmath12 .
the hot plasma is detected through the luminous x - ray emission it produces by thermal bremsstrahlung radiation , with @xmath13 .
like mountain peaks on earth , the high density rich clusters are relatively `` rare '' objects ; they exhibit a spatial number density of @xmath14 clusters @xmath15 , as compared with @xmath16 galaxies @xmath17 for the density of bright galaxies .
5inlll & & + & & + richness@xmath18 & 30300 galaxies & 330 galaxies + radius@xmath19 & ( 12 ) @xmath20 mpc & ( 0.11 ) @xmath20 mpc + radial velocity dispersion@xmath21 & 4001400 km@xmath22 & 100500 km@xmath22 + radial velocity dispersion@xmath23 & @xmath24 km@xmath22 & @xmath25 km@xmath26 + ( median ) + [ 5pt ] mass @xmath27 mpc)@xmath28&(10@xmath292@xmath30 @xmath31 & ( 10@xmath3210@xmath33 @xmath31 + luminosity @xmath34@xmath35 & @xmath366@xmath37 @xmath38 & ( 10@xmath3910@xmath40 @xmath38 + @xmath41 mpc ) & & + @xmath42@xmath43 & @xmath44 @xmath45 & @xmath46 @xmath45 + x - ray temperature@xmath47 & 214 kev & @xmath48 kev + x - ray luminosity@xmath47 & ( 10@xmath4910@xmath50 @xmath51 & @xmath52 @xmath51 + cluster number density@xmath53 & ( 10@xmath5410@xmath55 mpc@xmath56 & ( 10@xmath5710@xmath58 mpc@xmath57 + cluster correlation scale@xmath59 & @xmath60 mpc @xmath61 & @xmath62 mpc + fraction of galaxies in & @xmath63 & @xmath64 + clusters or groups@xmath65 & & + * in most entries , the typical range in the listed property or the median value is given .
groups and poor clusters are a natural and continuous extension to lower richness , mass , size , and luminosity from the rich and rare clusters . *
cluster richness : the number of cluster galaxies brighter than @xmath66 + 2@xmath67 ( where @xmath66 is the magnitude of the third brightest cluster galaxy ) , and located within a 1.5@xmath68 mpc radius of the cluster center ( [ sec:2.2 ] ) .
* the radius of the main concentration of galaxies ( where , typically , the galaxy surface density drops to @xmath69 of the central core density ) .
many clusters and groups are embedded in larger scale structures ( to tens of mpc ) .
* typical observed range and median value for the radial ( line - of - sight ) velocity dispersion in groups and clusters ( [ sec:2.9 ] ) .
* dynamical mass range of clusters within 1.5@xmath68 mpc radius ( [ sec:2.10 ] ) .
* luminosity range ( blue ) of clusters within 1.5@xmath68 mpc radius ( [ sec:2.10 ] ) .
* typical mass - to - light ratio of clusters and groups ( median value ) ( [ sec:2.10 ] ) . *
typical observed ranges of the x - ray temperature and 210-kev x - ray luminosity of the hot intracluster gas ( [ sec:3 ] ) . *
the number density of clusters decreases sharply with cluster richness ( [ sec:7 ] ) . * the cluster correlation scale for rich @xmath70 , @xmath71 , @xmath72@xmath73 mpc@xmath57 ) and poor @xmath74 , @xmath75 @xmath73 mpc@xmath57 ) clusters ( [ sec:2.3 ] , [ sec:7 ] ) . *
the fraction of bright galaxies @xmath76 ) in clusters and groups within @xmath77 mpc .
the main properties of clusters and groups of galaxies are summarized in table [ tab : one ] ( bahcall 1996 ) .
the table lists the typical range and/or median value of each observed property .
groups and poor clusters , whose properties are also listed , provide a natural and continuous extension to lower richness , mass , size , and luminosity from the rich and rare clusters . in the following subsections
i discuss in more detail some of these intrinsic cluster properties .
c@rrclr & + @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 & @xmath83 + 1 & 0.0283 & 9 & ( 0)@xmath18 & ( 3049 ) & @xmath84 + 2 & 0.0400 & 2 & 1 & 5079 & 1224 + 3 & 0.0577 & 33 & 2 & 80129 & 383 + 4 & 0.0787 & 60 & 3 & 130199 & 68 + 5 & 0.131 & 657 & 4 & 200299 & 6 + 6 & 0.198 & 921 & 5 & @xmath85 & 1 + & & @xmath86 & & & @xmath86 + & + & & & + & 104 & & & + & 71 & & & + & 33 & & & + & 82 & & & + & 22 & & & + * statistical sample .
@xmath87 boundaries as given in table 1 of abell ( 1958 ) .
notation : @xmath88 distance group ( defined by the estimated redshifts of the clusters ) ; @xmath89 average estimated redshift from the magnitude of the tenth brightest galaxy ; @xmath90 number of clusters ; @xmath91 richness class ; @xmath92 number of galaxies brighter than @xmath93 within @xmath94 mpc ( richness count ) .
* @xmath95 clusters are not part of the statistical sample .
* redshifts by hoessel et al .
( 1980 ) . *
this sample is limited to @xmath96 in addition to the @xmath87 boundaries of the statistical sample .
an illustration of the distribution of rich clusters with richness and distance is summarized in table [ tab : two ] .
the table refers to the statistical sample of the abell ( 1958 ) cluster catalog ( i.e. , richness class @xmath97 , corresponding to a richness threshold count of @xmath98 galaxies within @xmath99 mpc radius and magnitude @xmath100 ( see table [ tab : one ] ) ; redshift range @xmath101 to 0.2 ; and sky coverage @xmath102 and @xmath103 ) .
the abell catalog covers @xmath104 of the entire sky to @xmath105 .
recent smaller automated surveys from digitized plates are reported by the edinburgh - durham catalog ( edcc ; lumsden et al . 1992 ) , and the apm survey ( dalton et al .
a large automated cluster catalog will be available in the near future from the sloan digital sky survey ( sdss ) .
this and other planned surveys will allow a more accurate determination of the distribution of rich clusters with richness and distance and other statistical studies of clusters . the number density of clusters is a strong function of cluster richness .
integrated cluster densities , @xmath106 ( @xmath107 ) , representing the number density of clusters above a given richness threshold , and the associated mean cluster separation , @xmath108 ( @xmath109 ) , are listed in table [ tab : three ] ( bahcall and cen 1993 ) .
the fraction of galaxies in rich @xmath110 clusters is @xmath63 ( within the abell radius @xmath111 mpc ) .
the fraction of all galaxies that belong in clusters increases with increasing radius @xmath112 and with decreasing cluster richness threshold .
the average number of galaxies per cluster for @xmath113 clusters within @xmath99 mpc radius and @xmath114 is @xmath115 , or @xmath116 . for @xmath97 clusters the average is @xmath117 , or @xmath118 .
the number of galaxies increases to fainter luminosities following the schechter ( 1976 ) luminosity function .
4.5in[ht]c@ccc @xmath81 & @xmath82 & & @xmath108 ( @xmath119 ( mpc ) + @xmath120 & @xmath121 & 13.5 @xmath122 & 42 + @xmath123 & @xmath124 & 6.0 @xmath122 & 55 + @xmath125 & @xmath126 & 1.2 @xmath122 & 94 + @xmath127 & @xmath128 & 1.5 @xmath129 & 188 + * approximate uncertainties are 10@xmath130 for the @xmath131 densities and 10@xmath132 for @xmath133 .
the average number density of bright @xmath134 galaxies in @xmath110 clusters ( within @xmath111 mpc ) is @xmath135 the average overall ( field ) number density of bright @xmath136 galaxies is @xmath137 the average galaxy overdensity in rich @xmath138 clusters ( within @xmath99 mpc radius ) is thus @xmath139 the typical threshold galaxy overdensity in clusters ( within @xmath99 mpc radius ) is @xmath140 @xmath141 the galaxy overdensity increases at smaller radii from the cluster center .
the galaxy overdensity in the cores of typical compact rich clusters is approximately @xmath142 the radial density distribution of galaxies in a rich cluster can be approximated by a bounded emden isothermal profile ( zwicky 1957 ; bahcall 1977 ) , or by its king approximation ( king 1972 ) in the _ central _ regions . in the central regions ,
the king approximation for the galaxy distribution is @xmath143 @xmath144 @xmath145 and @xmath146 are , respectively , the space and projected profiles ( of the number density of galaxies , or brightness ) , @xmath147 and @xmath148 are the respective central densities , and @xmath149 is the cluster core radius [ where @xmath150 $ ] .
typical central densities and core radii of clusters are listed in the following subsection .
the projected and space central densities relate as @xmath151 a bounded emden isothermal profile of galaxies in clusters yields a profile slope that varies approximately as ( bahcall 1977 ) @xmath152 @xmath153 therefore @xmath154 the galaxy cluster cross - correlation function ( peebles 1980 ; lilje and efstathiou 1988 ) also represents the average radial density distribution of galaxies around clusters . for @xmath155 clusters , and @xmath156 in @xmath68 mpc ,
these references suggest , respectively @xmath157 or @xmath158 the average density distribution profile of galaxies in clusters thus follows , approximately , @xmath159 @xmath160 some substructure ( subclumping ) in the distribution of galaxies exists in a significant fraction of rich clusters ( @xmath161% ) ( geller 1990 ) .
central number density of galaxies in rich compact clusters ( bahcall 1975 , 1977 ; dressler 1978 ) is ( for galaxies in the brightest 3 magnitude range ) @xmath162 the central density reaches @xmath163 galaxies@xmath164mpc@xmath57 for the richest compact clusters .
the typical central _ mass _ density in rich compact clusters , determined from cluster dynamics is @xmath165^{\rm 2 } h^{\rm 2}\nonumber \label{eq:18}\end{aligned}\ ] ] where @xmath166 is the radial ( line - of - sight ) central cluster velocity dispersion ( in km@xmath26 ) and @xmath149 is the cluster core radius ( in mpc ) .
core radii of typical rich compact clusters , determined from the galaxy distribution ( bahcall 1975 ; dressler 1978 ; sarazin 1986 ) are in the range @xmath167 core radii of the x - ray emitting intracluster gas ( [ sec:3 ] ) are @xmath168 the core radius of the mass distribution determined from gravitational lensing observations of some clusters may be smaller , @xmath169 kpc , than determined by the galaxy and gas distribution .
the typical central density of the hot intracluster gas in rich clusters ( [ sec:3 ] ) is @xmath170 the fraction of elliptical , s0 , and spiral galaxies in rich clusters differs from that in the field , and depends on the classification type , or density , of the cluster ( see [ sec:2.11 ] ) ( bahcall 1977 ; oemler 1974 ; dressler 1980 ) .
see table [ tab : four ] .
the fraction of elliptical ( e ) and s0 galaxies increases and the fraction of spirals decreases toward the central cores of rich compact clusters . the fraction of spiral galaxies in the dense cores of some rich clusters ( e.g. , the coma cluster ) may be close to zero .
the galactic content of clusters as represented in table [ tab : four ] is part of the general density morphology relation of galaxies ( dressler 1980 ; postman and geller 1984 ) ; as the local density of galaxies increases , the fraction of e and s0 galaxies increases and the fraction of spirals decreases . for local galaxy densities
@xmath171 galaxies mpc@xmath172 , the fractions remain approximately constant at the average field " fractions listed above .
lcccc & e & s0 & sp & ( e+s0)/sp + regular clusters ( cd ) & 35% & 45% & 20% & 4.0 + intermediate clusters ( spiral - poor ) & 20% & 50% & 30% & 2.3 + irregular clusters ( spiral - rich ) & 15% & 35% & 50% & 1.0 + field & 10% & 20% & 70% & 0.5 + the typical radial ( line - of - sight ) velocity dispersion of galaxies in rich clusters ( median value ) is @xmath173 the typical range of radial velocity dispersion in rich clusters ( struble and rood 1991 ) is @xmath174 a weak correlation between @xmath175 and richness exists ; richer clusters exhibit , on average , larger velocity dispersion ( bahcall 1981 ) .
the observed velocity dispersion of galaxies in rich clusters is generally consistent with the velocity implied by the x - ray temperature of the hot intracluster gas ( [ sec:3.5 ] ) , as well as with the cluster velocity dispersion implied from observations of gravitational lensing by clusters ( except possibly in the central core ) .
velocity dispersion and temperature profiles as a function of distance from the cluster center have been measured only for a small number of clusters so far .
the profiles are typically isothermal @xmath176 $ ] for @xmath177 mpc , and drop somewhat at larger distances .
the typical dynamical mass of rich clusters within @xmath178 mpc radius sphere ( determined from the virial theorem for an isothermal distribution ) is @xmath179 the approximate range of masses for @xmath110 clusters ( within 1.5@xmath20 mpc ) is @xmath180 comparable masses are obtained using the x - ray temperature and distribution of the hot intracluster gas ( hughes 1989 ; bahcall and cen 1993 ; lubin and bahcall 1993 ; [ sec:3 ] ) .
the typical ( median ) blue luminosity of rich clusters ( within 1.5@xmath20 mpc ) is @xmath181 the approximate range of rich cluster blue luminosities is @xmath182 the typical mass - to - luminosity ratio of rich clusters is thus @xmath183 the inferred mass - density in the universe based on cluster dynamics is @xmath184 ( if mass follows light , @xmath185 , on scales @xmath186 mpc ) .
@xmath187 corresponds to the critical mass - density needed for a closed universe and @xmath188 @xmath189 .
rich clusters are classified in a sequence ranging from early- to late - type clusters , or equivalently , from regular to irregular clusters .
many cluster properties ( shape , concentration , dominance of brightest galaxy , galactic content , density profile , and radio and x - ray emission ) are correlated with position in this sequence .
a summary of the sequence and its related properties is given in table [ tab : five ] .
some specific classification systems include the bautz
morgan ( bm ) system ( bautz and morgan 1970 ) , which classifies clusters based on the relative contrast ( dominance in extent and brightness ) of the brightest galaxy to the other galaxies in the cluster , ranging from type i to iii in decreasing order of dominance ; and the rood - sastry ( rs ) system ( rood and sastry 1971 ) which classifies clusters based on the distribution of the ten brightest members ( from cd , to binary ( b ) , core ( c ) , line(l ) , flat ( f ) , and irregular ( i ) ) .
5inllll & & & + & & & + zwicky type & compact & medium - compact & open + bm type & i , i ii , ii & ( ii ) , ii
iii & ( ii iii ) , iii + rs type & cd , b,(l , c ) & ( l),(f),(c ) & ( f),i + shape symmetry & symmetrical & intermediate & irregular shape + central concentration & high & moderate & low + galactic content & elliptical - rich & spiral - poor & spiral - rich + e fraction & 35% & 20% & 15% + s0 fraction & 45% & 50% & 35% + sp fraction & 20% & 30% & 50% + e : s0:sp & 3:4:2 & 2:5:3 & 1:2:3 + radio emission & @xmath190 detection & @xmath190 detection&@xmath191 detection + x - ray luminosity & high & intermediate & low + fraction of clusters & @xmath104 & @xmath104 & @xmath104 + examples & a401 , coma & a194 & virgo , a1228 +
all rich clusters of galaxies produce extended x - ray emission due to thermal bremsstrahlung radiation from a hot intracluster gas ( jones and forman 1984 ; sarazin 1986 ; david et al .
1993 ; edge et al . 1990 , 1991 ;
henry et al . 1991 , 1992 ; burg et al .
the cluster x - ray luminosity emitted in the photon energy band @xmath192 to @xmath193 by thermal bremsstrahlung from a hot @xmath194 degrees ) intracluster gas of uniform electron density @xmath195 and a radius @xmath196 is @xmath197 the gaunt factor correction @xmath198 ( of order unity ) is a slowly varying function of temperature and energy .
the bolometric thermal bremsstrahlung luminosity of a cluster core can be approximated by @xmath199 some of the main properties of the hot intracluster gas are summarized below .
some of the main properties of the x - ray emission from rich clusters of galaxies are summarized in table [ tab : six ] .
5inll@c & & notes + @xmath200 ( 210 kev ) & @xmath201 ergs@xmath202 & @xmath203 + @xmath204 & @xmath205^{-3\beta + 1/2}$ ] & @xmath206 + @xmath207 & @xmath208 & @xmath209 + @xmath210 & @xmath211^{-3\beta /2}$ ] + & @xmath212^{-1}$ ] & @xmath108 + @xmath213 & @xmath1014 kev & @xmath214 + @xmath215 & @xmath21610@xmath217 k & @xmath214 + @xmath218 & @xmath219 & @xmath220 + @xmath221 & @xmath2220.3)@xmath20 mpc & @xmath198 + @xmath195 & @xmath223 cm@xmath172 & @xmath225 + @xmath226 @xmath227 mpc ) & @xmath228 [ range : ( 10@xmath22910@xmath230 h@xmath231 $ ] & @xmath232 + @xmath233 @xmath234 mpc ) & @xmath235 ( range : 0.030.15 h@xmath236 ) & @xmath232 + iron abundance & @xmath237 solar ( range : 0.20.5 ) & @xmath238 + * the x - ray luminosity of clusters ( 210 kev band ) .
@xmath239 increases with cluster richness and with cluster type ( toward compact , elliptical - rich clusters ) ( bahcall 1977a , b ; sarazin 1986 ; edge et al .
1991 ; jones and forman 1992 ; david et al .
1993 ; burg et al .
* x - ray surface brightness distribution , @xmath204 ; @xmath149 is the cluster core radius . *
mean @xmath207 from observations of x - ray brightness profiles ( jones and forman 1984 ; sarazin 1986 ) .
* implied spatial density profile of the hot gas in the cluster [ from @xmath204 ; isothermal ] .
* range of observed x - ray gas temperature in rich clusters ( edge et al . 1990 ; henry and arnaud 1991 ; arnaud et al .
* @xmath240 is the ratio of galaxy to gas velocity dispersion : @xmath241 is mean molecular weight in amu @xmath242 , @xmath243 is mass of the proton , @xmath175 is radial velocity dispersion of galaxies in the cluster , and @xmath215 is the x - ray temperature of the gas ( lubin and bahcall 1993 ) .
* cluster core radius determined from the x - ray distribution in the cluster ( jones and forman 1992 ) . * typical intracluster gas density in rich cluster cores ( jones and forman 1992 ) . * typical mass ( and range of masses ) of hot gas in rich clusters and its fraction of the total ( virial ) cluster mass @xmath244 within @xmath245 mpc of the cluster center ( edge and stewart 1991 ; jones and forman 1992 ; white and fabian 1995 ; lubin et al . 1996 ) . *
typical iron abundance ( and range ) of the intracluster gas ( in solar units ) ( edge and stewart 1991 ; jones and forman 1992 ) .
some of the fundamental questions about the intracluster gas relate to its origin , evolution , metal enrichment , hydrodynamical state in the cluster , and its relation to the distribution of galaxies and mass in the cluster .
i list below some of the relevant questions and discuss a selection of these topics in the following subsections
. some of the questions posed do not yet have sufficient observational constraints to suggest a possible solution . * what is the hydrodynamical state of the hot gas in clusters ?
is it in approximate hydrostatic equilibrium with the cluster potential ? * what is the relation of the intracluster gas to the galaxies and mass in the cluster ?
for example ( the subscripts below refer to gas , galaxies , and mass , respectively ) : + : : density profiles : @xmath246 vs. @xmath247 vs. @xmath248 ? : : temperature - velocity relation : @xmath249 vs. @xmath250 vs. @xmath251 ? : : mass : @xmath252 vs. @xmath253 vs. @xmath254 ? : : profiles of the above properties:.5 in : : core radii : @xmath255 vs. @xmath256 vs. @xmath257 ? : : luminosity versus galaxy - type relation : @xmath258 vs. spiral fraction ?
* what is the origin of the hot intracluster gas ?
* what is the origin of the metal enrichment of the gas ?
* what is the evolution of the intracluster gas ?
the standard model of clusters assumes that both the gas and the galaxies are in approximate hydrostatic equilibrium with the binding cluster potential ( bahcall and sarazin 1977 ; forman and jones 1984 ; sarazin 1986 ; evrard 1990 ; bahcall and lubin 1994 ) . in this model
the gas distribution obeys @xmath259 where @xmath260 and @xmath261 are the gas pressure and density , and @xmath262 is the total cluster binding mass within a radius @xmath156 .
the cluster mass can thus be represented as @xmath263 where @xmath249 is the gas temperature and @xmath264 is the mean particle mass of the gas .
the galaxies in the cluster respond to the same gravitational field , and they satisfy @xmath265 where @xmath266 is the radial velocity dispersion of galaxies in the cluster , @xmath267 is the galaxy density profile , and @xmath268 represents a possible anisotropy in the galaxy velocity distribution [ @xmath269 , where @xmath270 and @xmath156 represent the tangential and radial velocity components ] .
the above two relations yield @xmath271 where the @xmath272 parameter , defined by the left side of the above relation , can be determined directly from observations of cluster velocity dispersions and gas temperatures .
the @xmath273 parameter represents the ratio of energy per unit mass in the galaxies to that in the gas .
observations of a large sample of clusters yield a mean best - fit value of @xmath274 ( lubin and bahcall 1993 ; see also [ sec:3.5 ] ) .
this suggests that , on average , the gas and galaxies follow each other with comparable energies @xmath275 .
the observed mean value @xmath276 is consistent with the value of @xmath277 determined from the right - hand side of the @xmath277 relation ( referred to as @xmath278 , and determined from the gas and galaxy density profile fits ) . using @xmath279 ( [ sec:3.2 ] ) and @xmath280 ( [ sec:2.6 ] ) , one finds @xmath281 ( for an isothermal distribution ) ( bahcall and lubin 1994 ) .
the above consistency supports the assumption that the gas is in an approximate hydrostatic equilibrium with the cluster potential , and suggests that the galaxies and gas approximately trace each other in the clusters .
the hot intracluster gas in rich clusters appears to trace reasonably well the galaxies in the clusters , and with larger uncertainty also the cluster mass .
-2.0 in 2.0 in [ [ velocity - temperature - relation ] ] velocity - temperature relation + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the galaxy velocity dispersion in clusters is well correlated with the temperature of the intracluster gas ; it is observed ( fig .
[ fig : one ] ) that @xmath282 ( lubin and bahcall 1993 ) .
the best - fit @xmath283@xmath249 relation is listed in [ sec:3.7 ] .
the observed correlation indicates that , on average , the energy per unit mass in the gas and in the galaxies is the same .
figure [ fig : one ] shows that , unlike previous expectations , the galaxy velocities ( and therefore the implied cluster mass ) are not biased low with respect to the gas ( and , by indirect implications , with respect to the cluster mass ; see also [ sec:3.4 ] ) .
results from gravitational lensing by clusters also suggest that no significant velocity bias exists in clusters , and that the gas , galaxies , and mass provide consistent tracers of the clusters .
cosmological simulations of clusters ( lubin et al .
1996 ) produce @xmath283@xmath249 correlations that match well the data in figure [ fig : one ] .
[ [ density - profiles ] ] density profiles + + + + + + + + + + + + + + + + the gas density profile in clusters follows @xmath284^{-1}\ , \label{eq:36}\ ] ] with core radii in the range @xmath285@xmath286 mpc ( [ sec:3.2 ] ) .
this implies @xmath287 for @xmath288 .
the galaxy density profile in clusters follows approximately ( [ sec:2.6 ] ) @xmath289 with core radii @xmath290 ( [ sec:2.7 ] ) .
the mass density profile in clusters is less well established , but initial results from gravitational lensing distortions of background galaxies by foreground clusters suggest that the mass profile is consistent with the galaxy density profile ( tyson and frische 1996 ) . in the small central core regions of some clusters ( @xmath291 kpc ) ,
the mass distribution may be more compact than the gas or galaxies , with a small mass core radius of @xmath292 kpc .
the results for the overall cluster , however , suggest that the distributions of gas , galaxies , and mass are similar ( with the gas distribution possibly somewhat more extended than the galaxies , as seen by the mean density slopes above ) .
[ [ beta - discrepancy ] ] beta - discrepancy + + + + + + + + + + + + + + + + the mean @xmath293 result discussed above , combined with the similarity of the gas and galaxy density profile slopes ( that yields @xmath294 ; [ sec:3.3 ] ) show that the long claimed @xmath277-discrepancy for clusters ( where @xmath295 was claimed ) has been resolved ( bahcall and lubin 1994 ) . the gas and galaxies trace each other both in their spatial density distribution and in their energies , as expected for a quasi - hydrostatic equilibrium .
[ [ gas - mass - fraction ] ] gas mass fraction + + + + + + + + + + + + + + + + + the ratio of the mass of gas in clusters to the total virial cluster mass ( within @xmath296 mpc ) is observed to be in the range @xmath297 with a median value of @xmath298 ( jones and forman 1992 ; white et al .
1993 ; white and fabian 1995 ; lubin et al .
. the implications of this result , which shows a high fraction of baryons in clusters , is discussed in [ sec:4 ] .
the total gas mass in clusters , @xmath299@xmath300 , is generally larger than the total mass of the luminous parts of the galaxies ( especially for low values of @xmath225 ) .
with so much gas mass , it is most likely that a large fraction of the intracluster gas is of cosmological origin ( rather than all the cluster gas being stripped out of galaxies ) .
additional optical - x - ray correlations of clusters are summarized in [ sec:3.7 ] . the iron abundance in the intracluster gas
is observed to be @xmath301 solar , with only small variations @xmath302 from cluster to cluster ( e.g. , jones and forman 1992 ) .
a strong correlation between the total iron mass in clusters and the total luminosity of the cluster galaxies is observed ( jones and forman 1992 ) .
the metal enrichment of the intracluster gas is likely caused by gas stripped out of the elliptical galaxy members .
the iron abundance profile as a function of radius from the cluster center is generally flat , i.e. , a constant abundance at all radii ( except for some poor , low - mass clusters dominated by a single massive galaxy ) .
no evolution is observed in the overall iron abundance of clusters to @xmath303 ( mushotzky 1996 ) .
recent results using asca observations ( mushotzky 1996 ) of different element abundances in nearby clusters ( o , ne , mg , si , s , ar , ca ) suggest a snii origin for the metals ( resulting from early massive stars ) rather than the expected snia . these new results will be expanded in the near future as additional accurate x - ray data become available and will provide further clues regarding the origin of the metallicity of the intracluster gas .
some observed correlations between x - ray and optical properties are listed in table [ tab : seven ] ( bahcall 1977a , b ; edge and stewart 1991 ; david et al .
1993 ; lubin and bahcall 1993 ) .
3.5inll & + @xmath175-@xmath249 & @xmath175 ( kms@xmath304^{0.6\pm 0.1}$ ] + + @xmath249-@xmath305 & @xmath306 ( kev)@xmath307 + + @xmath200-@xmath308 & @xmath200(bol)@xmath309 + + @xmath310-@xmath311 & @xmath200(bol)@xmath312 + + @xmath313-@xmath249 & @xmath314^{-0.94\pm 0.38}$ ] + + @xmath249-@xmath200 & @xmath306 ( kev)@xmath315^{0.297\pm 0.004}$ ] + * @xmath175 is the galaxy line - of - sight velocity dispersion in the cluster ( kms@xmath202 ) .
@xmath249 is the temperature of the intracluster gas @xmath316 $ ] .
@xmath308 is the central galaxy density in the cluster [ number of galaxies brighter than @xmath93 , within @xmath317 of the cluster center ( bahcall 1977a ; edge and stewart 1991 ) ] .
@xmath200(bol ) is the bolometric x - ray luminosity of the cluster ( ergs@xmath202 ) .
@xmath313 is the fraction of spiral galaxies in the cluster ( @xmath318 mpc ) ( bahcall 1977b ; edge and stewart 1991 ) .
typical uncertainties of the coefficients are @xmath190 ( see references ) .
the x - ray luminosity function of clusters ( the number density of clusters with x - ray luminosity @xmath200 to @xmath319 is approximately ( edge et al .
1990 ) @xmath320 for @xmath321 , where @xmath200 is the 210-kev x - ray luminosity in units of ergs@xmath202 ( for @xmath322 ) .
the luminosity function can also be approximated as a power law ( edge et al .
1990 ) @xmath323 the number of x - ray clusters with x - ray luminosity brighter than @xmath200 is approximately @xmath324 the observed evolution of the x - ray cluster luminosity function suggests fewer high - luminosity clusters in the past @xmath325 ( edge et al .
1990 , henry et al .
additional data is required , however , to confirm and better assess the cluster evolution .
cooling flows are common at the dense cores of rich clusters ; x - ray images and spectra of @xmath190 of clusters suggest that the gas is cooling rapidly at their centers ( sarazin 1986 ; fabian 1992 ) .
typical inferred cooling rates are @xmath326 .
the gas cools within @xmath327 kpc of the cluster center ( generally centered on the brightest galaxy ) .
the cooling flows often show evidence for optical line emission , blue stars , and in some cases evidence for colder material in hi or co emission , or x - ray absorption .
the sunyaev zeldovich ( 1972 ) effect is a perturbation to the spectrum of the cosmic microwave background radiation as it passes through the hot dense intracluster gas .
it is caused by inverse compton scattering of the radiation by the electrons in the cluster gas . at the long - wavelength side of the background radiation spectrum ,
the hot gas lowers the brightness temperature seen through the cluster center by the fractional decrement @xmath328 where @xmath329 k is the microwave radiation temperature , @xmath330 is the thomson scattering optical depth through the cluster ( @xmath331 , where @xmath332 is the thomson scattering cross section and @xmath333 is the distance along the line of sight ) , @xmath215 is the intracluster gas temperature , and @xmath334 is the electron mass . for typical observed rich cluster parameters of @xmath335 ergs@xmath202 , @xmath336 mpc , and @xmath337 kev , the bremsstrahlung relation @xmath338 , [ sec:3.1 ] ) implies a central gas density of @xmath339 electrons cm@xmath172 , thus yielding @xmath340 @xmath341 $ ] .
therefore @xmath342 this temperature decrement remains constant over the cluster core diameter @xmath343 and decreases at larger separations .
the effect has been detected in observations of some rich , x - ray luminous clusters ( e.g. , coma , a665 , a2163 , a2218 , cl 0016 + 16 ) ( birkinshaw et al . 1984 ; jones et al . 1993 ; wilbanks et al .
1994 ; herbig et al .
rich clusters of galaxies provide the best laboratory for studying the baryon fraction ( i.e. , the ratio of the mass in baryons to the total mass of the system ) on relatively large scales of @xmath344 . the mass of baryons in clusters is composed of _ at least _ two components : the hot intracluster gas ( [ sec:3 ] ) and the luminous parts of the galaxies .
the baryon fraction in clusters is therefore @xmath345 where the first term on the right - hand side represents the gas mass ratio ( [ sec:3.5 ] ) and the second term corresponds to the stellar ( luminous ) contribution .
the baryon density required by big - bang nucleosynthesis is ( walker et al . 1991 ) @xmath346 comparison of the above relations indicates that if @xmath347 then there are many more baryons observed in clusters than allowed by nucleosynthesis . in fact , combining the two relations yields an @xmath348 value for the mass - density of @xmath349 for the observed range of @xmath350 .
therefore , the baryon density given by nucleosynthesis and the high baryon content observed in clusters ( mainly in the hot intracluster gas ) suggest that @xmath351 .
this assumes , as expected on this large scale ( and as seen in simulations ) , that the baryons are not segregated relative to the dark matter in clusters ( white et al .
1993 ) .
-3.0 in 2.0 in figure [ fig : two ] compares the observed gas mass fraction in clusters ( @xmath352 baryon fraction ) with expectations from cosmological simulations of @xmath347 and @xmath353 cold - dark - matter ( cdm ) flat models ( lubin et al .
the results show , as expected , that the @xmath347 model predicts a much lower gas mass fraction than observed , by a factor of @xmath354 . a low - density cdm model with @xmath355 ( in mass ) best matches the data , as expected from the general analysis discussed above ( see white et al .
1993 ; white and fabian 1995 ; lubin et al .
1996 ) . in summary ,
the high baryon fraction observed in clusters suggests , independent of any specific model , that the mass density of the universe is low , @xmath356 .
this provides a powerful constraint on high - density @xmath357 models ; if @xmath358 , a resolution of this baryon problem needs to be found .
the masses of clusters of galaxies within a given radius , @xmath359 , can be determined from three independent methods : a ) : : optical : galaxy velocity dispersion ( and distribution ) assuming hydrostatic equilibrium ( [ sec:3.3 ] ) ; b ) : : x - rays : hot gas temperature ( and distribution ) assuming hydrostatic equilibrium ( [ sec:3.3 ] ) ; c ) : : lensing : gravitational distortions of background galaxies ( tyson et al .
1990 , 1996 ; kaiser and squires 1993 ) .
this method determines directly the surface mass overdensity .
the first two methods were discussed in [ sec:2 ] and [ sec:3 ] ; the last ( and newest ) method is discussed in this book by narayan .
( see also tyson et al .
1990 , 1996 ; kaiser and squires 1993 ; smail et al .
1995 , and references therein ) .
the galaxy and hot gas methods yield cluster masses that are consistent with each other as discussed in [ sec:3 ] ( with @xmath282 ) .
gravitational lensing results , which provide direct information about cluster masses , are available only for a small number of clusters so far ( with data rapidly increasing ) . for all but one of the clusters the masses determined from lensing are consistent , within the uncertainties , with the masses determined from the galaxies and hot gas methods ( see bahcall 1995 for a summary ) .
some differences in masses between the different methods are expected for individual clusters due to anisotropic velocities , cluster orientation ( yielding larger lensing surface mass densities for clusters elongated in the line - of - sight , and vice versa ) , and sub - structure in clusters . on average , however , all three independent methods yield cluster masses that are consistent with each other .
this triple check on cluster masses provides strong support for the important cluster mass determinations .
the masses of rich clusters range from @xmath360 to @xmath361 within @xmath178 mpc radius of the cluster center .
when normalized by the cluster luminosity , a robust mass - to - light ratio is determined for nearby clusters , with only small variations from cluster to cluster ( [ sec:2.10 ] ) @xmath362 this result is similar to the one obtained from the baryon fraction in
[ sec:4 ] .
if , as desired by theoretical arguments , the universe has critical mass density @xmath347 , than most of the mass in the universe _ can not _ be associated with galaxies , groups , and clusters ; the mass distribution in this case would be strongly biased ( i.e. , mass does not follow light , with the mass distributed considerably more diffusely than the light ) .
-2.0 in 2.0 in
a recent analysis of the mass - to - light ratio of galaxies , groups and clusters by bahcall , lubin and dorman ( 1995 ) suggests that while the @xmath363 ratio of galaxies increases with scale up to radii of @xmath364@xmath365 mpc , due to the large dark halos around galaxies ( see fig .
[ fig : three ] ; also ostriker et al .
1974 ) , this ratio appears to flatten and remain approximately constant for groups and rich clusters , to scales of @xmath366 mpc , and possibly even to the larger scales of superclusters ( fig .
[ fig : four ] ) .
the flattening occurs at @xmath367@xmath368 , corresponding to @xmath369 .
this observation suggests that most of the dark matter is associated with the dark halos of galaxies . unlike previous expectations
, this result implies that clusters do _ not _ contain a substantial amount of _ additional _ dark matter , other than that associated with ( or torn - off from ) the galaxy halos , and the hot intracluster medium .
bahcall et al . ( 1995 ) suggest that the relatively large @xmath370 ratio of clusters @xmath371 results mainly from a high @xmath370 ration of elliptical / so galaxies .
they show ( fig .
[ fig : three ] ) that ellipticals have an @xmath370 ratio that is approximately 3 to 4 times larger than typical spirals at the same radius @xmath372 within @xmath373 $ ] . since clusters are dominated by elliptical and so galaxies , a high @xmath370 ratio results .
-3.0 in 2.0 in unless the distribution of matter is very different from the distribution of light , with large amounts of dark matter in the voids " or on very large scales , the above results suggest that the mass density in the universe may be low , @xmath369 ( or @xmath374 for a small bias of @xmath375 , where the bias factor @xmath206 relates the overdensity in galaxies to the overdensity in mass : @xmath376 ) .
the observed mass function ( mf ) , @xmath377 , of clusters of galaxies , which describes the number density of clusters above a threshold mass @xmath378 , can be used as a critical test of theories of structure formation in the universe .
the richest , most massive clusters are thought to form from rare high peaks in the initial mass - density fluctuations ; poorer clusters and groups form from smaller , more common fluctuations .
bahcall and cen ( 1993 ) determined the mf of clusters of galaxies using both optical and x - ray observations of clusters .
their mf is presented in figure [ fig : five ] .
the function is well fit by the analytic expression @xmath379 with @xmath380 , ( where the mass @xmath378 represents the cluster mass within @xmath178 mpc radius ) .
-3.0 in 2.0 in the observed cluster mass function is compared in figure [ fig : five ] with expectations from different cold - dark - matter cosmologies using large - scale simulations ( bahcall and cen 1992 ) .
the comparison shows that the cluster mf is indeed a powerful discriminant among models .
the standard cdm model @xmath381 can not reproduce the observed mf for any bias parameter ; when normalized to the cobe microwave background fluctuations on large scales , this model produces too many massive clusters , unseen by the observations .
a low - density cdm model on the other hand , with @xmath382 ( with or without a cosmological constant ) , appears to fit well the observed cluster mf ( fig .
[ fig : five ] ) .
imaging and spectroscopic data ( yee et al . 1987 ; ellingson et al . 1991 ; yee et al .
1992 ) indicate that quasars are found in environments significantly richer than those of average galaxies .
the data show a positive association of quasars with neighboring galaxies .
optically selected quasars to @xmath383 exhibit a quasar
galaxy cross - correlation function amplitude , @xmath384 , that is approximately 2.3 times stronger than the galaxy - galaxy correlation amplitude ( to separations @xmath385 mpc ) : @xmath386 this excess correlation suggests that the quasars are typically located in groups of galaxies with a _ mean _
richness @xmath387 ( where @xmath388 mpc@xmath172 is the mean density of galaxies ) .
the range of individual group richnesses is , however , wide .
radio - loud quasars at @xmath389 are found in similar environments to those of the optical quasars above . at @xmath390 ,
the radio quasars appear to be located in richer environments , with @xmath391 this cross - correlation amplitude corresponds to a _
mean _ environment of rich clusters @xmath392 , @xmath393 .
radio quasars at these redshifts are thus typically found in rich clusters . the average galaxy velocity dispersion of the parent clusters associated with the quasars ( ellingson et al .
1991 ; yee et al .
1992 ) is @xmath394 the observed auto - correlation function of optically selected quasars is approximately ( shaver 1988 ) @xmath395^{-1.8 } \label{eq:55}\ ] ] the quasar correlation strength is intermediate between the correlation of individual galaxies and the correlation of rich clusters .
this correlation strength is consistent with the quasars location in groups of the above mean richness , as would be suggested by the richness - dependent cluster correlation function ( [ sec:10 ] ) .
the quasars may thus trace the correlation function of their parent clusters ( bahcall and chokshi 1991 ) .
similar results are observed for the association of radio galaxies with groups and clusters .
this association explains the observed increase in the strength of the radio galaxy correlation function over the general galaxy correlations ( bahcall and chokshi 1992 ) .
rich clusters of galaxies are powerful tracers of the large - scale structure of the universe ( bahcall 1988 , peebles 1993 ) .
i summarize in the sections below the use of clusters in tracing the large - scale structure ; i include superclusters ( [ sec:9 ] ) , statistics of the cluster correlation function ( [ sec:10 ] ) , and peculiar motions on large scales ( [ sec:11 ] ) .
redshift surveys of galaxies reveal that superclusters are very large , high - density systems of galaxies that are flattened or filamentary in shape , extending to tens of mpc .
the superclusters appear to surround large under - dense regions ( `` voids '' ) of comparable sizes creating a `` cellular - like '' morphology of the universe on large scales ( gregory and thompson 1978 ; gregory et al .
1981 ; chincarini et al .
1981 ; giovanelli et al .
1986 ; de - lapparent et al .
1986 ; da costa et al .
1988 ; rood 1988 ; schectman et al .
1996 ; landy et al .
1996 ) .
large scale superclusters have been identified very effectively by rich clusters of galaxies ( bahcall and soneira 1984 ) , like high mountain peaks tracing mountain chains .
superclusters are generally defined as clusters of rich clusters of galaxies above a given spatial density enhancement @xmath220 .
here @xmath396 , where @xmath397 is the number density of clusters in a supercluster and @xmath398 is the mean number density of clusters .
the observed superclusters are large flattened systems , extending to @xmath399 in size .
the superclusters typically contain several clusters .
the high - density superclusters generally surround low - density regions .
a complete catalog of superclusters was constructed by bahcall and soneira ( 1984 ) to @xmath400 .
the catalog identifies all superclusters that have a spatial density enhancement @xmath401 .
the mean number density of the bahcall - soneira superclusters is @xmath402 , with a mean separation between superclusters of @xmath403 .
a summary of the main properties of the superclusters is presented in table [ tab : eight ] .
the superclusters trace well the structure observed in the more detailed , but smaller , galaxy redshift surveys .
4inll & + number density of scs & @xmath404mpc@xmath57 + number of clusters per sc & 215 clusters + fraction of clusters in scs & 54% + size of largest sc & @xmath405mpc + sc shape & flattened + volume of space occupied by scs & @xmath406 + observations of the redshift distribution of galaxies in narrow @xmath407 pencil - beam surveys to @xmath408 ( broadhurst et al . 1990
; hereafter beks ) reveal a highly clumped and apparently periodic distribution of galaxies .
the distribution features peaks of galaxy counts with an apparently regular separation of 128 mpc , with few galaxies between the peaks .
what is the origin of this clumpy , periodic distribution of galaxies ?
what does it imply for the nature of the large - scale structure and the properties discussed above ?
bahcall ( 1991 ) investigated these questions observationally , by comparing the specific galaxy distribution with the distribution of known superclusters .
bahcall showed that the observed galaxy clumps originate from the tails of large superclusters ( [ sec:9.1 ] ) . when the narrow - beams intersect these superclusters , which have a mean separation of @xmath409 , the beks galaxy distribution is reproduced .
the redshift distribution of the superclusters is essentially identical to the galaxy redshift distribution , i.e. , it reproduces the observed peaks in the beks survey , for @xmath410 .
this indicates that the galaxy clumps observed in the pencil - beam survey originate from these superclusters as the beam crosses the superclusters surface .
the main superclusters that contribute to the clumps were identified .
for example , the first northern clump originates from the coma - hercules supercluster (= the great - wall ) ; the second northern clump is mostly due to the large corona borealis supercluster .
the narrow - beam survey of beks is directed toward the north and south galactic poles .
some of the bahcall - soneira superclusters coincident with the beks peaks are located at projected distances of up to @xmath411100 mpc from the poles .
this suggests that the high - density supercluster regions are embedded in still larger halo surfaces , @xmath412 mpc in size , and that these large structures surround large underdense regions .
the observed number of clumps and their mean separation are consistent with the number density of superclusters and their average extent ( [ sec:9.1 ] ) .
the narrow widths of the beks peaks are consistent with , and imply , flat superclusters . from simulations of superclusters and pencil - beams ,
bahcall , miller , and udomprasert ( 1996 ) find that the observed peak - widths distribution is consistent with that expected of randomly placed superclusters with @xmath413 mpc width and @xmath414 mpc extent
. the bahcall - soneira superclusters may exhibit weak positive correlations on scales @xmath415150 mpc ( bahcall and burgett 1986 ) .
this implies that the superclusters , and thus their related galaxy clumps , are not randomly distributed but are located in some weakly correlated network of superclusters and voids , with typical mean separation of @xmath415 mpc .
this picture is consistent with statistical analyses of the beks distribution as well as with the observational data of large - scale structure .
the apparent periodicity in the galaxy distribution suggested by beks is expected to be greatly reduced when pencil - beams in various directions are combined ; the scale reflects the typical _ mean _ separation between large superclusters , @xmath415@xmath416 mpc , but with large variations at different locations .
the correlation function of clusters of galaxies efficiently quantifies the large - scale structure of the universe .
clusters are correlated in space more strongly than are individual galaxies , by an order of magnitude , and their correlation extends to considerably larger scales ( @xmath417 mpc ) .
the cluster correlation strength increases with richness ( @xmath352 luminosity or mass ) of the system from single galaxies to the richest clusters ( bahcall and soneira 1983 ; bahcall and west 1992 ) .
the correlation strength also increases with the mean spatial separation of the clusters ( szalay and schramm 1985 ; bahcall and burgett 1986 ; bahcall and west 1992 ) .
this dependence results in a universal " dimensionless cluster correlation function ; the cluster dimensionless correlation scale is constant for all clusters when normalized by the mean cluster separation . empirically , two general relations have been found ( bahcall and west 1992 ) for the correlation function of clusters of galaxies , @xmath418 : @xmath419 @xmath420 where @xmath421 is the amplitude of the cluster correlation function , @xmath422 is the richness of the galaxy clusters of type @xmath232 ( [ sec:2.2 ] ) , and @xmath423 is the mean separation of the clusters .
here @xmath424 , where @xmath425 is the mean spatial number density of clusters of richness @xmath422 ( [ sec:2.3 ] ) in a volume - limited , richness - limited complete sample . the first relation , eq .
( [ eq:56 ] ) , states that the amplitude of the cluster correlation function increases with cluster richness , i.e. , rich clusters are more strongly correlated than poorer clusters . the second relation , eq .
( [ eq:57 ] ) , states that the amplitude of the cluster correlation function depends on the mean separation of clusters ( or , equivalently , on their number density ) ; the rarer , large mean separation richer clusters are more strongly correlated than the more numerous poorer clusters . eqs .
( [ eq:56 ] ) and ( [ eq:57 ] ) relate to each other through the richness function of clusters , i.e. , the number density of clusters as a function of their richness . equation ( [ eq:57 ] ) describes a universal scale - invariant ( dimensionless ) correlation function with a correlation scale @xmath426 ( for @xmath427 ) .
there are some conflicting statements in the literature about the precise values of the correlation amplitude , @xmath421 .
nearly all these contradictions are caused by not taking account of eq .
( [ eq:56 ] ) . when apples are compared to oranges , or the clustering of rich clusters
is compared to the clustering of poorer clusters , differences are expected and observed . figure [ fig : six ] clarifies the observational situation .
the @xmath428 relation for groups and clusters of various richnesses is presented in the figure .
the recent automated cluster surveys of apm ( dalton et al . 1992 ) and edcc ( nichol et al . 1992 ) are consistent with the predictions of eqs .
( [ eq:56 ] ) and ( [ eq:57 ] ) , as is the correlation function of x - ray selected rosat clusters of galaxies ( romer et al .
bahcall and cen ( 1994 ) show that a flux - limited sample of x - ray selected clusters will exhibit a correlation scale that is smaller than that of a volume - limited , richness - limited sample of comparable apparent spatial density since the flux - limited sample contains poor groups nearby and only the richest clusters farther away . using the richness - dependent cluster correlations of eqs .
( [ eq:56 ] ) and ( [ eq:57 ] ) , bahcall and cen ( 1994 ) find excellent agreement with the observed flux - limited x - ray cluster correlations of romer et al .
( 1994 ) .
-3.0 in 2.0 in the strong correlation amplitude of galaxy clusters , and the large scales to which clusters are observed to be positively correlated @xmath429 , complement and quantify the superclustering of galaxy clusters discussed in [ sec:9 ] .
clusters of galaxies are strongly clustered in superclusters of large scales ( [ sec:9 ] ) , consistent with the strong cluster correlations to these scales ( [ sec:10 ] ) .
this fundamental observed property of clusters of galaxies the cluster correlation function can be used to place strong constraints on cosmological models and the density parameter @xmath348 by comparison with model expectations .
bahcall and cen ( 1992 ) contrasted these cluster observations with standard and nonstandard cdm models using large n - body simulations ( @xmath430 box , @xmath431 particles ) .
they find that none of the standard @xmath347 cdm models can fit consistently the strong cluster correlations . a low - density ( @xmath356 ) cdm - type model ( with or without a cosmological constant ) , however , provides a good fit to the cluster correlations ( see figs .
[ fig : seven][fig : nine ] ) as well as to the observed cluster mass - function ( [ sec:7 ] , fig .
[ fig : five ] ) .
this is the first cdm - type model that is consistent with the high amplitude and large extent of the correlation function of the abell , apm , and edcc clusters .
such low - density models are also consistent with other observables as discussed in this paper
. the @xmath348 constraints of these cluster results are model dependent ; a mixed hot + cold dark matter model , for example , with @xmath347 , is also consistent with these cluster data ( see primack s chapter in this book ) .
the cdm results for clusters corresponding to the rich abell clusters @xmath432 with @xmath433 mpc are presented in figure [ fig : seven ] together with the observed correlations ( bahcall and soneira 1983 ; peacock and west 1992 ) .
the results indicate that the standard @xmath347 cdm models are inconsistent with the observations ; they can not provide either the strong amplitude or the large scales ( @xmath434 mpc ) to which the cluster correlations are observed .
similar results are found for the apm and edcc clusters .
-3.0 in 2.0 in the low - density , low - bias model is consistent with the data ; it reproduces both the strong amplitude and the large scale to which the cluster correlations are detected .
-3.0 in 2.0 in -3.0 in 2.0 in the dependence of the observed cluster correlation on @xmath108 was also tested in the simulations .
the results are shown in figure [ fig : eight ] for the low - density model .
the dependence of correlation amplitude on mean separation is clearly seen in the simulations . to compare this result directly with observations , i plot in figure [ fig : nine ] the dependence of the correlation scale , @xmath435 , on @xmath108 for both the simulations and the observations .
the low - density model agrees well with the observations , yielding @xmath436 , as observed . the @xmath347 model , while also showing an increase of @xmath435 with @xmath108 , yields considerably smaller correlation scales and a much slower increase of @xmath437 . what causes the observed dependence on cluster richness [ eqs .
( [ eq:56][eq:57 ] ] ?
the dependence , seen both in the observations and in the simulations , is most likely caused by the statistics of rare peak events , which kaiser ( 1984 ) suggested as an explanation of the observed strong increase of correlation amplitude from galaxies to rich clusters .
the correlation function of rare peaks in a gaussian field increases with their selection threshold . since more massive clusters correspond to a higher threshold , implying rarer events and thus larger mean separation , eq .
( [ eq:57 ] ) results . a fractal distribution of galaxies and clusters would also produce eq .
( [ eq:57 ] ) ( e.g. , szalay and schramm 1985 ) .
how is the mass distributed in the universe ?
does it follow , on the average , the light distribution ? to address this important question , peculiar motions on large scales are studied in order to directly trace the mass distribution .
it is believed that the peculiar motions ( motions relative to a pure hubble expansion ) are caused by the growth of cosmic structures due to gravity .
a comparison of the mass - density distribution , as reconstructed from peculiar velocity data , with the light distribution ( i.e. , galaxies ) provides information on how well the mass traces light ( see chapter by dekel , and 1994 ) . the basic underlying relation between peculiar velocity and density
is given by @xmath438 where @xmath439 is the mass overdensity , @xmath440 is the galaxy overdensity , and @xmath441 is the bias parameter discussed in [ sec:6 ] .
a formal analysis yields a measure of the parameter @xmath442 .
other methods that place constraints on @xmath277 include the anisotropy in the galaxy distribution in the redshift direction due to peculiar motions ( see strauss and willick 1995 for a review ) .
measuring peculiar motions is difficult .
the motions are usually inferred with the aid of measured distances to galaxies or clusters that are obtained using some ( moderately - reliable ) distance - indicators ( such as the tully - fisher or @xmath443 relations ) , and the measured galaxy redshift .
the peculiar velocity @xmath444 is then determined from the difference between the measured redshift velocity , @xmath445 , and the measured hubble velocity , @xmath446 , of the system ( the latter obtained from the distance - indicator ) : @xmath447 .
-3.0 in 2.0 in a summary of all measurements of @xmath277 made so far is presented in strauss and willick ( 1995 ) .
the dispersion in the current measurements of @xmath277 is very large ; the various determinations range from @xmath448 to @xmath219 , implying , for @xmath449 to @xmath219 .
no strong conclusion can therefore be reached at present regarding the values of @xmath277 or @xmath348 .
the larger and more accurate surveys currently underway , including high precision velocity measurements , will likely lead to the determination of @xmath277 and possibly its decomposition into @xmath348 and @xmath206 ( e.g. , cole et al .
1994 ) .
clusters of galaxies can also serve as efficient tracers of the large - scale peculiar velocity field in the universe ( bahcall , gramann and cen 1994 ) . measurements of cluster peculiar velocities are likely to be more accurate than measurements of individual galaxies , since cluster distances can be determined by averaging a large number of cluster members as well as by using different distance indicators .
using large - scale cosmological simulations , bahcall et al . ( 1994 ) find that clusters move reasonably fast in all the cosmological models studied , tracing well the underlying matter velocity field on large scales .
the clusters exhibit a maxwellian distribution of peculiar velocities as expected from gaussian initial density fluctuations .
the model cluster 3-d velocity distribution , presented in figure [ fig : ten ] , typically peaks at @xmath450 and extends to high cluster velocities of @xmath451 .
the low - density cdm model exhibits lower velocities ( fig .
[ fig : ten ] ) .
approximately 10% of all model rich clusters ( 1% for low - density cdm ) move with @xmath452 .
a comparison of model expectation with recent , well calibrated cluster velocity data ( giovanelli et al . 1996 ) is presented in figure [ fig : eleven ] ( bahcall and oh 1996 ) .
the comparison between models and observations suggests that the cluster velocity data is consistent with a low - density cdm model , and is inconsistent with a standard @xmath347 cdm model , since no high velocity clusters are observed .
-3.0 in 2.0 in cen , bahcall and gramann ( 1994 ) determined the expected velocity correlation function of clusters in different cosmologies .
they find that close cluster pairs , with separations @xmath453 mpc , exhibit strong attractive motions ; the pairwise velocities depend sensitively on the model .
the mean pairwise attractive cluster velocities on @xmath454 mpc scale ranges from @xmath455 for @xmath347 cdm to @xmath456 for @xmath457 cdm .
the cluster velocity correlation function , presented in figure [ fig : twelve ] , is negative on small scales indicating large attractive velocities , and is positive on large scales , to @xmath458 mpc indicating significant bulk motions in the models .
none of the models reproduce the very large bulk flow of clusters on @xmath459 mpc scale , @xmath460 , recently reported by lauer and postman ( 1994 ) .
the bulk flow expected on this large scale is generally @xmath461 for all the models studied ( @xmath347 and @xmath462 cdm , and pbi ) .
-3.0 in 2.0 in
considerable progress has been made over the last two decades in the study of clusters and superclusters of galaxies , as described in these lectures . however
, many problems still remain open .
i highlight some of the unsolved problems in this field that are likely to be solved in the coming decade .
currently planned large redshift surveys of galaxies and clusters such as the sloan digital sky survey and the 2df survey , deep optical and x - ray surveys using hst , keck , rosat , asca , and axaf , among other , should allow a considerable increase in our understanding of the nature and evolution of these fundamental systems .
at the same time , state of the art cosmological simulations to be available in the next decade ( e.g. , ostriker , this book ) should greatly enhance our ability to compare the observations with detailed expectations from various cosmologies and hopefully narrow down the correct cosmological model of our universe .
here is a partial list of some of the interesting unsolved problems in the field of clusters and superclusters of galaxies .
[ [ clusters - of - galaxies ] ] clusters of galaxies + + + + + + + + + + + + + + + + + + + + * what is the mass distribution and its extent in clusters of galaxies ? using gravitational lensing distortions , one can determine the mass density profile , @xmath248 , and total cluster mass , @xmath463 , of clusters and compare it with the distribution of galaxies and gas for a large sample of clusters .
* does mass follow light on these scales ? if not , what is the bias factor as a function of scale , @xmath464 ?
* what is the implied density parameter from clusters , @xmath465 ?
* what is the accurate baryon - fraction in clusters and groups of galaxies , as a function of scale , @xmath466 ?
* what is the origin of the hot intracluster gas and its metallicity ?
* what is the evolution of clusters in the optical and in x - rays ?
* what are the cosmological implications from studies of clusters ?
[ [ superclusters ] ] superclusters
+ + + + + + + + + + + + + * what is , quantitatively , the morphology of superclusters and large - scale structure ( superclusters , filaments , and void network ) ? * what is the dependence of the superclustering properties on galaxy luminosity , surface brightness , type ( e , s ) , and system ( galaxies versus clusters ) ? * what are the peculiar motions in superclusters and on large scales ?
* what is the mass , and mass distribution , in superclusters and on large scales ? does mass follow light ? * what is in the `` voids '' ? * what is @xmath348 on large scales ?
* what is the baryon fraction in superclusters ?
* what is the time evolution of superclusters ?
* what are the constraints made by the observed superclusters and large - scale structure on cosmology and galaxy formation models ?
i expect that many of these questions will be addressed and possibly solved in the coming decade .
i thank the organizers of the jerusalem winter school 1995 , j. p. ostriker and a. dekel , for an outstanding , productive , and fun school . the work by n. bahcall and collaborators
is supported by nsf grant ast93 - 15368 .
100 abell , g. o. 1958 , apjs , 3 , 211 arnaud , m. , hughes , j. p. , forman , w. , jones , c. , lachieze - rey , m. , yamashita , k. , & hatusukade , i. 1992 , apj , 390 , 345 bahcall , j. n. , & sarazin , c. 1977 , apj , 213 , l99 bahcall , n. a. 1975 , apj , 198 , 249 bahcall , n. a. 1977 , ara&a , 15 , 505 bahcall , n. a. 1977a , apj , 217 , l77 bahcall , n. a. 1977b , apj , 218 , l93 bahcall , n. a. 1981 , apj , 247 , 787 bahcall , n. a. 1988 , ara&a , 26 , 631 bahcall , n. a. 1991 , apj , 376 , 43 bahcall , n. a. 1995 , in dark matter , aip conf .
proceedings 336 , ed .
s. holt and c. bennet ( new york : aip ) , 201 bahcall , n. a. 1996 , in astrophysical quantities , ed .
a. cox ( new york : aip ) bahcall , n. a. , & burgett , w. 1986 , apjl , 300 , l35 bahcall , n. a. , & cen , r. 1992 , apj , 398 , l81 bahcall , n. a. , & cen , r. 1993 , apj , 407 , l49 bahcall , n. a. , & cen , r. y. 1994 , apj , 426 , l15 bahcall , n. a. , & chokshi , a. 1991 , apj , 380 , l9 bahcall , n. a. , & chokshi , a. 1992 , apjl , 385 , l33 bahcall , n. a. , gramann , m. , & cen , r. 1994 , apj , 436 , 23 .
bahcall , n. a. , & lubin , l. m. 1994 , apj , 426 , 513 bahcall , n. a. , lubin , l m. , & dorman , v. 1995 , apj , 447 , l81 bahcall , n. a. , miller , n. , & udomprasert , p. 1996 , in preparation bahcall , n. a. , & oh , p. 1996 , apj , 462 , l43 bahcall , n. a. , & soneira , r. m. 1983 , apj , 270 , 20 bahcall , n. a , .
& soneira , r.m . , 1984 , apj , 277 , 27 bahcall , n. a. , & west , m. 1992 , apjl , 392 , 419 bautz , l. p. , & morgan , w. w. 1970 , apj , 162 , l149 birkinshaw , m. , gull , s. f. , & hardebeck , h. e. 1984 , nature , 309 , 34 broadhurst , t. j. , ellis , r. , koo , d. , & szalay , a. 1990 , nature , 343 , 726 burg , r. , giacconi , r. , forman , w. , jones , c. 1994 , apj , 422 , 37 cen , r. , bahcall , n. a. , & gramann , m. 1994 , apj , 437 , l51 chincarini , g. , rood , h. j. , & thompson , l. a. 1981 , apj , 249 , l47 cole , s. , fisher , k. b. , & weinberg , d. h. 1994 , mnras , 267 , 785 da costa , l. n. , et al .
1988 , apj , 327 , 544 dalton , g. b. , efstathiou , g. , maddox , s. j. , & sutherland , w. 1992 , apj , 390 , l1 david , l. p. , slyz , a. , jones , c. , forman , w. , vrtilek , s. , & arnaud , k. 1993 , apj , 412 , 479 de lapparent , v. , geller , m. , & huchra , j. 1986 , apj , 302 , l1 dekel , a. 1994 , ara&a , 32 , 371 dressler , a. 1978 , ap.j , 226 , 55 dressler , a. 1980 , apj , 236 , 351 dressler , a. 1984 , ara&a , 22 , 185 edge , a. , & stewart , g. c. , 1991 , mnras , 252 , 428 edge , a. , stewart , g. c. , fabian , a. c. , & arnaud , k. a. 1990 , mnras , 245 , 559 ellingson , e. , yee , h. k. c. , & green , r. f. 1991 , apj , 371 , 45 evrard , a. e. 1990 , apj , 363 , 349 fabian , a. c. 1992 , in clusters and superclusters of galaxies , nato asi series no .
366 , edited by a. c. fabian ( dordrecht : kluwer academic ) , p.151 geller , m. j. 1990 , in clusters of galaxies , stsci symposium no .
4 , ed . w. r. oegerle et al .
( cambridge : cambridge univerity press ) , p. 25
giovanelli , r. , & haynes , m. 1996 , in preparation giovanelli , r. , haynes , m. , & chincarini , g. 1986 , apj , 300 , 77 gregory , s. a. , & thompson , l. a. 1978 , apj , 222 , 784 gregory , s. a. , thompson , l. a. , & tifft , w. 1981 , apj , 243 , 411 henry , j. p. , & arnaud , k. a. 1991 , apj , 372 , 410 henry , j. p. , gioia , i. m. , maccacaro , t. , morris , s. l. , stocke , j. t. , & walter , a. 1992 , apj , 386 , 408 herbig , t. , lawrence , c. r. , readhead , a. c. s. , & gulkis , s. 1995 , apjl , 449 , l1 hoessel , j. g. , gunn , j. e. , & thuan , t. x. 1980 , apj , 241 , 486 hughes , j. p. 1989 , apj , 337 , 21 jones , c. , & forman , w. 1984 , apj , 276 , 38 jones , c. , & forman , w. 1992 , in clusters and superclusters of galaxies , nato asi series , no . 366 , ed .
a. c. fabian ( dordrecht : kluwer academic ) , p. 49
jones , m. , saunders , r. , alexander , p. , birkinshaw , m. , & dillon , n. 1993 , nature , 365 , 320 kaiser , n. , & squires , g. 1993 , apj , 404 , 441 king , i. 1972 , apjl , 174 , l123 landy , s. , shectman , s. , lin , h. , kirshner , r. , oemler , a. , & tucker , a. 1996 , apj , 456 , l1 lauer , t. , & postman , m. 1994 , apj .
, 425 , 418 lilje , p. b. , & efstathiou , g. 1988 , mnras , 231 , 635 lubin , l. , & bahcall , n. a. 1993 , apjl , 415 , l17 lubin , l. , cen , r. , bahcall , n. a. , & ostriker , j. p. 1996 , apj , 460 , 10 lumsden , s. l. , nichol , r. c. , collins , c. a. , & guzzo , l. 1992 , mnras , 258 , 1 mushotzky , r. 1996 , preprint nichol , r. , collins , c. a. , guzzo , l. , & lumsden , s. l. 1992 , mnras,255 , 21p oemler , a. 1974 , apj , 194 , 1 oort , j. 1983 , ara&a , 21 , 373 ostriker , j. p. , peebles , p. j. e. , & yahil , a. 1974 , apj , 193 , l1 peacock , j. , & west , m. 1992 , mnras , 259 , 494 peebles , p. j. e. 1980 , the large scale structure of the universe ( princeton : princeton university press ) peebles , p. j. e. 1993 , principles of physical cosmology ( princeton : princeton university press ) postman , m. , & geller , m. 1984 , apj , 281 , 95 romer , a.k . , collins ,
c. , bhringer , h. , cruddace , r. , ebeling , h. , macgillivray , h. , & voges , w. 1994 , nature , 372 , 75 rood , h. j. 1988 , ara&a , 26 , 245 rood , h. j. , & sastry , g. n. 1971 , pasp , 83 , 313 sarazin , c. l. 1986 , rev . mod .
phys . , 58 , 1 schechter , p. l. 1976 , apj , 203 , 297 shaver , p. 1988 , in large - scale structure of the universe , iau symposium no .
j. audouze et al .
( dordrecht : reidel ) , 359 shectman , s. , landy , s. , oemler , a. , tucker , a. , lin , h. , & kirshner , r. 1996 , apj , in press ( preprint astro - ph/9604167 ) smail , j. , ellis , r. , fitchett , m. , & edge , a. 1995 , mnras , 273 , 277 strauss , m. , & willick , j. 1995 , phys . reports , 261 , 271 struble , m. , & rood , h. 1991 , apjs , 77 , 363 sunyaev , r. a. , & zeldovich , ya . b. , 1972 , comments astrophys .
space phys . , 4 , 173 szalay , a. , & schramm , d. n. 1985 , nature , 314 , 718 tyson , j. a. , & fischer , p. 1996 , apjl , in press tyson , j. a. , valdes , f. , & wenk , r. a. 1990 , apj , 349 , ll walker , t. p. , et al .
1991 , apj , 376 , 51 white , d. , & fabian , a. 1995 , mnras , 273 , 72 white , s. d. m. , navaro , j. f. , evrard , a. e. , & frenk , c. s. 1993 nature , 366 , 429 wilbanks , t. m. , ade , p. a. r. , fischer , m. l. , holzapfel , w. l. , & lange , a. 1994 , apj , 427 , l75 yee , h. k. c. 1992 , in clusters and superclusters of galaxies , nato asi series no .
a. c. fabian ( dordrecht : kluwer academic ) , 293 yee , h. k. c. , & green , r. f. 1987 , apj , 319 , 28 zwicky , f. 1957 , morphological astronomy ( berlin : springer )
|
rich clusters of galaxies are the most massive virialized systems known .
even though they contain only a small fraction of all galaxies , rich clusters provide a powerful tool for the study of galaxy formation , dark matter , large - scale structure , and cosmology .
superclusters , the largest known systems of galaxies , extend to @xmath0 mpc in size and highlight the large - scale structure of the universe .
this large - scale structure reflects initial conditions in the early universe and places strong contraints on models of galaxy formation and on cosmology .
some of the questions that can be addressed with clusters and superclusters of galaxies include : how did galaxies and larger structures form and evolve ? what is the amount , composition , and distribution of matter in clusters and larger structures ? how does the cluster mass density relate to the matter density in the universe ?
what constraints can the cluster and supercluster data place on cosmology ?
i will discuss in these lectures some of the properties of clusters and superclusters of galaxies that can be used to investigate these topics .
|
g0.9 + 0.1 is the only composite snr known in the direction of the gc ( d@xmath6 10kpc , @xmath7@xmath8 ) : a bright centrally condensed synchrotron nebula ( with relatively flat spectral index ) , powered by the loss of rotational energy from the neutron star , and a radio shell with a steeper radio spectrum ( @xcite ) .
@xcite were able to fit the _ bepposax _ x - ray spectrum of the pwn with a power - law and interpret the x - ray emission as non - thermal in origin .
the small angular extent of the x - ray emission ( radius @xmath6 1 ) , combined with an estimated age of the remnant of a few thousand years , is further evidence that the central radio core is powered by a young pulsar ( @xmath6 2,700yr ; @xcite ) .
recently , g0.9 + 0.1 was observed with _
chandra _ by @xcite .
they found that the synchrotron - emitting pulsar wind nebula ( pwn ) has a clear axial symmetry and a faint x - ray point source lying along the symmetry axis possibly corresponds to the pulsar itself .
they argued that the nebular morphology can be explained in terms of a torus of emission in the pulsar s equatorial plane and a jet directed along the pulsar spin axis , as is seen in the x - ray nebulae powered by other young pulsars ( crab , vela ) .
here , we present the observations of g0.9 + 0.1 obtained with _ xmm - newton _ and we focus on the first spectral analysis of the large and small - scale structures of the pwn ( for more details see @xcite ) .
we present in this section the spectral analysis of the pwn using combined fits of the mos and pn data .
the observation time after screening rejection of the flares are respectively for mos1 and mos2 about 15.5ksec and 15.4ksec , and 10ksec for pn .
we use the following camera response matrices : m1@xmath9medv9q20t5r6@[email protected] , m2@xmath9medv9q20t5r6@[email protected] , epn@xmath9ef20@[email protected] .
the local background used corresponds to the central mos1 ccd region excluding the bright point sources and the central part within a radius of 2.6. we subtract from the source and the local background in our pointing , a blank - field observation ( kindly provided by david lumb ) at the same position in order to take into account the particle background ( see e.g. @xcite ) .
the normalization between our pointing and the blank - field files is determined respectively by the count rate ratio in the 1012kev range for the mos and 1214kev for the pn .
the epic spectra were binned to give s / n@xmath103 .
xspec v11.1.0 was used for the spectral analysis .
all subsequent errors are quoted at 90@xmath11 confidence .
abundances are those of @xcite .
the _ xmm - newton _ observation of the supernova remnant g0.9 + 0.1 shows emission from the pwn , but not from the surrounding shell observed in radio .
the soft x - ray emission expected from the shell is likely not seen due to the large interstellar absorption towards the source .
the epic image of the pwn in g0.9 + 0.1 is presented in figure [ fig : image ] .
it exhibits a very bright central core , surrounded by extended diffuse emission .
the extent of this emission is about twice larger than observed with _
chandra _ , thanks to the high sensitivity of _ xmm - newton_. for comparison with previous analysis , we fit the spectrum of the overall nebula .
the region considered corresponds to the sum of the core with regions 1 , 2 and 3 , as is shown in figure [ fig : image ] .
the data are well fitted by either a power - law or a thermal bremsstrahlung ( table [ table : table1 ] ) . for the photo - electric absorption ,
we use the cross - sections of @xcite .
the parameters ( @xmath12 , @xmath2 ) related to the power - law are compatible , within the error bars , with the _ bepposax _ values ( @xcite : @xmath12=10.9@xmath13@xmath8 , @xmath2=1.95@xmath14 ) , though slightly higher but more tightly constrained . in the following , we use the updated cross - sections for x - ray absorption by the ism ( tbabs in xspec ) from @xcite .
the parameters although in good agreement with the previous ones ( see table [ table : table1 ] ) , give systematically slightly lower @xmath12 values using tbabs absorption model .
cccccc model & @xmath15 & @xmath2 or @xmath16 & @xmath17 & f@xmath18 + & ( @xmath19@xmath8 ) & ( kev ) & ( 446 dof ) & + wabs*pl & 1.46@xmath20 & 2.00@xmath21 & 376.4 & 5.74 + wabs*brems & 1.33@xmath22 & 10.6@xmath23 & 376.1 & 5.11 + wabs*bb & 0.83@xmath24 & 1.73@xmath25 & 389.0 & 3.63 + tbabs*pl & 1.39@xmath26 & 1.94@xmath27 & 375.2 & 5.69 + tbabs*brems & 1.27@xmath28 & 11.5@xmath29 & 375.0 & 5.12 + tbabs*bb & 0.80@xmath24 & 1.75@xmath24 & 387.9 & 3.65 +
+ the _ xmm - newton _ data enables us to study the variation of the spectral index within the pwn . in other plerions ,
a softening of the spectrum has been observed toward the outer regions like in 3c 58 ( @xcite , @xcite ) , g21.5 - 0.9 ( @xcite , @xcite ) and ic 443 ( @xcite ) .
this softening could be explained by the shorter lifetime of high energy electrons than for the lower energy electrons . + to search for this effect , we extracted the spectra in four regions , which are displayed in figure [ fig : image ] .
the observed spectra and best - fit models for the regions 1 ( top panel ) and 3 ( bottom panel ) are shown , as examples , in figure [ fig : spectra ] .
the spectrum extends up to almost 12kev for the pn data , and to about 9kev in the mos data allowing strong constraints on the determination of the photon index .
.combined fits of the regions of the pwn in g0.9 + 0.1 .
the cross - sections of the interstellar absorption are from wilms et al .
unabsorbed x - ray flux ( 2 - 10kev ) are expressed in 10@xmath30ergs@xmath8s@xmath31 . _
top _ : @xmath32 is frozen to the value obtained for the overall pwn ( see table [ table : table1 ] ) .
_ bottom _ : @xmath32 is a free parameter . [ cols="^,^,^,^,^ " , ] [ table : table2 ] we fit the spectra of these four regions by absorbed power - laws , fixing @xmath33 to the value obtained for the overall pwn , i.e @xmath33= 1.4310@xmath34@xmath8 ( see table [ table : table1 ] ) .
the best - fit parameters are given in table [ table : table2 ] .
letting the absorption column densities free do not afford significantly better fits , and the values of column density @xmath33 are compatible to within 15@xmath11 for the different regions of the nebula .
+ a clear steepening of the spectrum is observed from the inner part toward the outer part of the pwn as is shown in figure [ fig : gamma - r ] .
the observed spectral softening with radius in g0.9 + 0.1 from the core to the outskirts of the pwn , is most likely due to the effect of synchrotron radiation losses on the highest energy electrons as they progress through the nebula .
for comparison we have reported on figure [ fig : gamma - r ] the variation of the spectral index inferred from the observation of ic443 by @xcite .
the axis unit used ( r@xmath35 , as defined in fig .
[ fig : gamma - r ] and in @xcite ) gives a measure independent on the distance to the nebulae .
our data appear to show a stronger softening from the core to the outer part of the nebula of g0.9 + 0.1 , compared to ic443 ( see fig .
[ fig : gamma - r ] ) , and also to other pwne such as 3c58 and g21.5 - 0.9 ( see for comparison fig . 5 in @xcite ) .
this could be consistent with a stronger magnetic field in the galactic center region compared to other snr environments .
the contours ( in black ) corresponds to the _ chandra _ observation in the same energy range .
the white boxes represent different small - scale structures surrounding the bright core ( dark blue circle ) , for which the spectral index and 90@xmath11 confidence uncertainties are given .
the position of the structures were determined according to the _ chandra _ observation ( see gaensler et al .
2001 ) : `` east arc '' ( @xmath36 1.2 ) , `` jet '' ( @xmath36 1.8 ) , and `` south - west arc '' ( @xmath36 2.8 ) . ] the brightest region of the nebula corresponds to a core of angular diameter 8 as observed with _
xmm - newton_. the _ chandra _ observation has resolved it as an elliptical clump of size 5 @xmath37 8 ( @xcite ) . the faint point source proposed to be the pulsar by @xcite is outside this region ( about 8 from the center of the clump ) .
the fact that the hardest spectral index is obtained for the core would likely indicate that it hosts the pulsar . as proposed by @xcite
, this clump may also be an analog to the knots observed in the crab nebula ( @xcite ) and close to the vela pulsar ( @xcite ) , which exhibit variability in brightness , position and morphology .
the position and the morphology are however very similar in both _
xmm - newton _ and _ chandra _ observations which are separated from only about one month .
the nature of the bright core is still unclear : a clump , which could be a region of separate particle acceleration as in ( @xcite ) , or a very compact region in which the pulsar is embedded ? .
tracing the spectral variations in the small - scale structures around the core is then essential for understanding this particular pwn .
figure [ fig : gamma ] displays a closer view of the inner region and shows the small - scale structures present around the core of pwn g0.9 + 0.1 . for comparison
, the x - ray contours of the _ chandra _ data are superposed . in our data ,
the small - scale structures are not so clearly detected as with _
( jet - like and arc - like features ) .
however using the position of the structures seen by _
chandra _ and the high sensitivity of _ xmm - newton _ , we were for the first time able to carry out spectral analysis of these small - scale structure . determining their spectral index gives an indication of the geometry and orientation of the nebula .
figure [ fig : gamma ] reports the photon spectral index of the corresponding structures obtained with combined fits ( mos and pn ) .
the region corresponding to the eastern part of the arc - like feature in @xcite , shows the hardest spectral index ( @[email protected] ) among all the structures together with the core region ( @xmath41.1@xmath5 0.5 ) . on the contrary the region corresponding to
the south - west part of the arc - like feature , appears to have the steepest photon index ( @[email protected] ) over all the pwn .
a possible explanation is that the east arc is pointing towards the observer and that its spectral hardness is due to the relativistic beaming or doppler boosting of the electrons , while the south - west arc is the opposite part of the torus .
the region associated to the jet - like feature in @xcite does not exhibit a harder spectrum as suggested by their hardness ratio .
we report _ xmm - newton _ observations of the x - ray pulsar wind nebula ( pwn ) inside the composite supernova remnant g0.9 + 0.1 ( @xcite ) , which was convincingly detected for the first time in x - rays by _ bepposax _ ( @xcite ) .
+ the large - scale image of the pwn g0.9 + 0.1 shows a very bright central core , surrounded by relatively large - scale diffuse emission with a much spatial extent than is apparent for the recent _ chandra _ images .
thanks to its high sensitivity , _ xmm - newton _ allows for the first time a spectral analysis of the large and small - scale structures ( see also @xcite ) .
our spectral analysis combining the mos and pn data offers a clear indication of a softening of the spectrum with distance from the centroid ( related to the finite lifetime of the synchrotron emitting electrons ) , as observed in other known x - ray plerions .
even if in our data , the small - scale structures are not so clearly detected as in the _ chandra _ data ( jet - like and arc - like features ) , we were able for the first time to carry out their spectral analysis .
two of the structures show distinct spectral indexes from the general trend ( from @[email protected] to @[email protected] ) : the eastern part of the arc - like feature has a very hard photon index ( @[email protected] ) , while the south - west part of the arc - like feature has a softer photon index ( @[email protected] ) .
+ the study of g0.9 + 0.1 provides further clues concerning the processes by which pulsars connect with their environment and also illustrates the impact of a strong ambient magnetic field as in the galactic center region .
anders e. & grevesse n. , 1989 , geochimica et cosmochimica acta , 53 , 197 bocchino f. & bykov a. m. , 2001 , a&a , 376 , 248 bocchino f. , warwick , r. s. , marty , p. , lumb , d. , becker , w. , pigot , c. 2001 , a&a , 369 , 1078 gaensler , b.m .
, pivovaroff m.j .
, garmire g.p .
2001 , apj , 556 , l107 gaensler , b.m . ,
arons , j. , kaspi , v.m . , pivovaroff , m.j . ,
kawai , n. , tamura , k. 2002 , apj , 569 , in press helfand d.j . & becker r.h . , 1987 ,
apj 314 , 203 hester j. j. , scowen p. a. , sankrit r. , et al . , 1995 , apj , 448 , 240 majerowicz , s. & neumann d. m. 2001 , in the `` xxi moriond conference '' , eds .
d. neumann , f. durret , & j. tran thanh van , in press ( http://www-dapnia.cea.fr/conferences/morion@xmath9astro@xmath92001/index.html ) mereghetti s. , sidoli l. , israel g. l. 1998 , a&a , 331 , l77 morrison r. & mccammon d. , 1983 , apj , 270 , 119 pavlov , g. g. , kargaltsev , o. y. , sanwal , d. , garmire , g. p. 2001
, apj , 554 , l189 porquet , d. , decourchelle a. , warwick r.s . , 2002 , in preparation sidoli l. , mereghetti s. , israel g. l. , bocchino f. , 2000 , a&a , 361 , 719 slane p. , chen , y. , schulz , n. s. , seward , f. d. , hughes , j. p. , gaensler , b. m. 2000 , apj , 533 , l29 torii k. , slane , p. o. , kinugasa , k. , hashimotodani , k. , tsunemi , h. 2000 , pasj , 52 , 875 warwick r. s. , bernard , j .-
bocchino , f. , decourchelle , a. , ferrando , p. , griffiths , r. g. , haberl , f. , la palombara , n. , lumb , d. , mereghetti , s. , read , a. m. , schaudel , d. , schurch , n. , tiengo , a. , willingale , r. 2001 , a&a , 365 , l248 wilms j. , allen a. , mccray r. , 2000 , apj , 542 , 914
|
we report the results of the _ xmm - newton _ observation of the pulsar wind nebula ( pwn ) in the composite supernova remnant g0.9 + 0.1 located in the galactic center region .
thanks to the sensitivity of the epic cameras , we focus on the first spectral analysis of large and small - scale structures of this pwn using mos and pn data in combination .
our spatially resolved spectral analysis at large - scale offers a clear indication of a softening of the photon index with distance from the centroid of the nebula ( from @xmath0=1.12@xmath1 to @xmath2=2.42@xmath3 ) , as observed in other known x - ray plerions .
a spectral analysis of the small - scale structures in the central region of this pwn , delineates variations of the spectral index within the arc - like feature observed with _
chandra _ : the eastern part has a hard photon index ( @[email protected] ) , while the south - west part has a softer photon index ( @[email protected] ) .
|
in the early 1970 s , philippe delsarte showed that there is a remarkable connection between three objects in mathematics : 1 .
strongly regular graphs ; 2 .
linear codes with two weights ; 3 .
sets of points in a projective space with two intersection sizes with respect to hyperplanes . to illustrate these connections ,
consider the five - dimensional projective space over the field of order @xmath0 .
there are some captivating examples of two - intersection sets in this space , one of which was discovered by ray hill @xcite in 1973 and it is known as _ hill s @xmath2-cap _ since it consists of @xmath2 points ( of @xmath3 ) , no three lying on a common line .
this cap has the property that any hyperplane of @xmath3 has only two possible ways of intersecting the cap : in @xmath4 or @xmath5 points .
equivalently , we can associate the points of this cap with the columns of a generator matrix for a ternary linear code with parameters @xmath6 $ ] and weights @xmath7 and @xmath8 . using a beautiful geometric construction known as _ linear representation _ , hill s @xmath2-cap gives rise to a geometry known as a _ partial quadrangle _ , and it therefore gives rise to a strongly regular graph on @xmath9 vertices .
today , we also know that hill s @xmath2-cap is intimately linked to of the hermitian variety @xmath10 and to an interesting imprimitive cometric @xmath11-antipodal @xmath12-class association scheme @xcite .
there are other examples in this space found by gulliver @xcite , and these examples have something in common ; they each admit a common symmetry of order @xmath1 , which happens to be @xmath13 when @xmath14 . in general ,
when @xmath15 , we can _ factor out _ this symmetry and identify these two - intersection sets with subsets of the classical algebraic variety known as the _ cubic segre variety _
( see section [ segre ] ) .
this paper details the search for , and enumeration of , a family of strongly regular graphs arising from two - intersection sets of @xmath16 .
of the three mathematical objects above , two - intersection sets are the most readily computable , and hence effort was focused on finding two - intersection sets using linear programming techniques . in particular , the projective space @xmath16 was examined as a successor to @xmath17 , since the binary projective codes of dimension at most @xmath18 have been classified by bouyukliev @xcite .
over 6000 new strongly regular graphs were discovered as a result of enumerating two - weight ternary codes of dimension @xmath18 , many of which overlap the list of codes on kohnert s webpage . up to projective equivalence , there are 6635 such two - intersection sets of @xmath3 invariant under a collineation of order @xmath1 .
we begin with the elementary definitions of all three objects , and the relationships between them . an _
@xmath19 $ ] linear code _ is a @xmath20-dimensional subspace of @xmath21 , and its vectors are known as _ codewords_. the weight of a codeword is the number of non - zero values in its coordinate vector .
every @xmath19 $ ] linear code is the rowspace of some @xmath22 generator matrix .
a linear code @xmath23 is _ projective _ if no two columns of its generator matrix are linearly dependent .
equivalently , @xmath23 is projective if the minimum weight of the dual code @xmath24 of @xmath23 is at least @xmath0 .
a two - weight linear code is simply a linear code in which all non - zero codewords have weight one of two fixed values @xmath25 or @xmath26 .
the family of two - weight linear codes is an important one ; a @xmath27-error - correcting linear code is perfect if and only if its dual is a two - weight code , and a @xmath28-error - correcting linear code is uniformly packed if and only if its dual is a two - weight code ( * ? ? ?
* theorem 4.2 , theorem 4.3 ) .
a two - intersection set @xmath29 with parameters @xmath30 is a set of points in @xmath31 such that every hyperplane in @xmath31 is incident with either @xmath32 or @xmath33 points in @xmath29 .
for example , the points lying on a line form a simple two - intersection set with parameters @xmath34 : every hyperplane either lies completely on the line ( and hence intersects it at @xmath35 points ) , or meets it at one point .
the complement of a two - intersection set with parameters @xmath30 is again a two - intersection set ( but with different parameters ) .
every projective two - weight linear code arises from a two - intersection set and gives rise to a strongly regular graph .
a strongly regular graph with parameters @xmath36 is a graph containing @xmath37 vertices in which every vertex has degree @xmath20 , every two adjacent vertices share @xmath38 common neighbours , and every two non - adjacent vertices share @xmath39 common neighbours .
the following results were originally proved by delsarte @xcite , and provide the necessary details for converting a two - intersection set into a two - weight linear code and for converting a two - weight linear code into a two - intersection set or a strongly regular graph . if @xmath40 is a two - intersection set with parameters @xmath41 that spans @xmath42 , then @xmath43 is the generator matrix of a projective two - weight @xmath19 $ ] linear code with weights @xmath44 and @xmath45 , where it is understood that the @xmath46 are vectors representing the homogeneous coordinates of points of @xmath47 .
conversely , if @xmath48 is the generator matrix of a projective two - weight @xmath19 $ ] linear code with weights @xmath25 and @xmath26 , then @xmath49 are the homogeneous coordinate vectors for a two - intersection set with parameters @xmath50 that spans @xmath42 .
delsarte defined the associated graph @xmath51 of a projective two - weight linear code @xmath52 with weights @xmath25 and @xmath26 as follows : let the vertices of @xmath51 correspond to the codewords of @xmath52 .
two distinct vertices in @xmath51 are adjacent if and only if the weight of the difference between their corresponding codewords is @xmath25 .
@xmath51 is a strongly regular graph for any projective two - weight linear code @xmath52 .
our motivation stemmed from the fact that many of the most interesting two - intersection sets of @xmath16 had a stabiliser divisible by @xmath1 , and @xmath1 is a primitive prime divisor of @xmath54 ( and so an element of order @xmath1 must act irreducibly ) .
the hill @xmath2-cap and the examples found by gulliver could be constructed by taking unions of orbits of this cyclic subgroup . restricting the search to two - intersection sets that could be constructed in this manner prevented the scope of the problem from becoming infeasible and allowed for a full enumeration of possible solutions . to classify all two - weight ternary codes of dimension
@xmath18 seems to be out of the current range of computational power .
let us model @xmath55 simply as the field extension @xmath56 over @xmath57 . if @xmath58 is a primitive root of @xmath56 , then multiplication by @xmath58 yields a cyclic subgroup acting regularly on the points of @xmath55 ; the so - called _ singer cycle _ of @xmath55 .
if we raise @xmath58 to a suitable power , namely @xmath59 , then we will obtain a field element of order @xmath60 in @xmath56 , which will induce a collineation @xmath61 of @xmath55 of order @xmath13 .
alternatively , we can write the orbits of @xmath61 down by taking the equivalence classes of the following relation : @xmath62 where @xmath63 . a two - intersection set invariant under @xmath61
will automatically span @xmath55 since @xmath61 acts irreducibly on this space .
the map @xmath64 defined by @xmath65 , [ y_{1 } , y_{2 } , y_{3 } ] ) = [ x_{1 } y_{1 } , x_{1 } y_{2 } , x_{1 } y_{3 } , x_{2 } y_{1 } , x_{2 } y_{2 } , x_{2 } y_{3}]\ ] ] is a _ segre embedding _ , and the image of this map is a cubic _
segre variety _
@xmath66 ( see ( * ? ? ?
* chapter 25 ) ) . in order to work with this segre embedding
, it became necessary to find an equivalent operation over elements of @xmath67 .
the map @xmath68 from @xmath69 to @xmath56 is equivalent to the segre embedding @xmath70 .
we model @xmath71 as @xmath72 over @xmath73 with basis @xmath74 , and @xmath75 as @xmath76 over @xmath73 with basis @xmath77 .
now suppose we have @xmath78 \in { \mathsf{pg}}(1 , q)$ ] and @xmath79 \in { \mathsf{pg}}(2 , q)$ ] ; that is , @xmath80 and @xmath81 .
+ clearly , @xmath82 so it suffices to show that @xmath83 is a basis for @xmath67 ( over @xmath73 ) . to this end , suppose @xmath84 .
further suppose , for the sake of contradiction , that @xmath85 .
we may then rearrange the equation as follows : @xmath86 noting that the left hand side of the equation is an element of @xmath72 , and the right hand side of the equation is an element of @xmath76 , we conclude that both sides of the equation are elements of @xmath87
. then we must have @xmath88 , for some @xmath89 a contradiction .
hence , we must have @xmath90 and consequently , @xmath91 . as @xmath92 , @xmath93 and @xmath94 are linearly independent , we see that @xmath95 .
this confirms that @xmath96 is a basis for @xmath67 .
it was discovered that , for certain values of @xmath97 , the image of @xmath70 is a transversal of the orbits discussed above . for @xmath98 ,
every equivalence class of @xmath99 contains exactly one element of the segre variety arising from the products @xmath100 consider the following subgroup of @xmath56 of order @xmath101 : @xmath102 we will first show that the subgroup @xmath103 is contained in @xmath57 . the order of the intersection @xmath103 must divide the order of both subgroups , and hence must also divide @xmath104 noting that @xmath105 and @xmath106 are coprime , we may simplify the expression to @xmath107 this can be further simplified to @xmath108 , with one step of the euclidean algorithm . clearly , when @xmath98 , the above expression evaluates to @xmath109 .
in these cases , the order of @xmath103 divides the order of @xmath57 , and hence @xmath103 is a subgroup of @xmath57 .
now we may proceed to proving the stated result .
suppose @xmath110 and @xmath111 .
we show that if @xmath112 and @xmath113 are in the same equivalence class of @xmath114 , they represent the same element of the segre variety . to do so
, we suppose that @xmath115 . this equation can be rearranged into the form @xmath116^{(q^2 - q + 1)(q - 1 ) } = 1 $ ] .
so @xmath117 is an element of @xmath103 , and is therefore in @xmath57 , using the fact shown above .
hence , we have @xmath118 , for some @xmath119 , which may be rearranged into the form @xmath120 . therefore , @xmath112 and @xmath113 are projectively equivalent .
we have shown , then , that the segre variety arising from @xmath121 is a transversal of the given equivalence classes , as there are an equal number of equivalence classes and elements of @xmath121 .
this relationship between the segre variety and the orbits of a collineation group allows for the expression of conforming two - intersection sets as geometric entities in the segre variety . every orbit included in such a two - intersection set corresponds to exactly one point in the segre variety .
fix a primitive root @xmath58 of @xmath122 .
then we can describe the eight orbits comprising the hill @xmath2-cap by single elements of the segre variety : @xmath123 these elements of the segre variety were obtained by taking the products @xmath124 , where @xmath125 note that the first coordinates of these pairs are each divisible by @xmath126 , and the second coordinates are each divisible by @xmath127 ( and so @xmath128 and @xmath129 for each @xmath130 ) .
we will catalogue each of our examples in section [ results ] by points of the segre variety in this way .
the following is an extension of the results presented by penttila and royle @xcite , and was used to quickly identify implausible parameter sets in the computational search .
suppose @xmath29 is a two - intersection set with parameters @xmath30 .
then there must exist an integer solution to : @xmath131 if @xmath132 and @xmath133 are the number of hyperplanes that are incident with @xmath32 and @xmath33 points in @xmath29 , respectively , then counting arguments reveal the following equations : l|p7 cm + @xmath134 & counting the number of hyperplanes + + @xmath135 & counting incident point - hyperplane pairs + + @xmath136 & counting triples @xmath137 where @xmath138 and @xmath139 are two different points incident with the hyperplane @xmath140 + if we take the left - hand sides of these three equations , and sum them with coefficients @xmath141 , @xmath142 , and @xmath28 accordingly , we will obtain @xmath143 .
the result then follows .
suppose @xmath29 is a two - intersection set with parameters @xmath30
. then @xmath144 must divide @xmath145 .
suppose that @xmath146 and @xmath147 are the number of hyperplanes through a point @xmath148 that are incident with @xmath32 and @xmath33 points in @xmath29 , respectively .
similarly , suppose @xmath149 and @xmath150 are the number of hyperplanes through a point @xmath151 that are incident with @xmath32 and @xmath33 points in @xmath29 , respectively .
then counting arguments show that : @xmath152 @xmath153 @xmath154 @xmath155 by solving this system of equations , we see that @xmath156 must be an integer .
a significant advantage of searching for two - intersection sets , rather than the other mathematical objects discussed in this paper , is the relative ease with which two - intersection sets can be found computationally .
the task of finding two - intersection sets in a projective space can be naturally represented as a linear programming problem , as evinced by kohnert @xcite .
this section details a method of framing the task in such a way .
let @xmath157 be a group of collineations of @xmath55 .
we let @xmath158 be the @xmath157-quotient of the point - hyperplane incidence matrix wherein the @xmath159 entry is the number of elements from the @xmath160th @xmath157-orbit on points incident with the @xmath161th @xmath157-orbit representative on hyperplanes .
alternatively , we can use a duality arising from a bilinear form , say , so that the hyperplanes are replaced by points and incidence is replaced by orthogonality . in our application , where @xmath157 is a cyclic group of order @xmath13 acting on @xmath55 , we can simply construct @xmath158 by indexing its rows and columns by the cubic segre variety @xmath66 . the relative trace map ( from @xmath67 to @xmath73 ) yields a bilinear form when we model @xmath55 by the field @xmath67 : @xmath162 hence , we may instead calculate @xmath163 by calculating the number of elements @xmath164 such that @xmath165 , where @xmath166 are the @xmath160-th and @xmath161-th elements of the segre variety modelled in @xmath67 .
an _ orbit inclusion vector _ is a 0 - 1 vector in which the @xmath160th entry is @xmath28 or @xmath143 , dependent on whether the @xmath160-th orbit is included or excluded from the current point set ( resp . ) .
clearly , premultiplying the inclusion vector with the incidence matrix yields a vector wherein the @xmath160-th entry contains the number of points in the current point set incident with the @xmath160-th hyperplane . in order to find a two - intersection set with intersection values @xmath32 and @xmath33
, we need to find solutions to the system of equations : @xmath167 where @xmath168 is a 0 - 1 vector .
while this is not a linear system , it can be converted into one with a minor modification . for a given solution @xmath168 ,
we introduce a vector @xmath169 of length @xmath37 , such that the @xmath160th entry of @xmath169 is @xmath143 if the @xmath160th hyperplane is incident with @xmath32 points , or @xmath28 if the @xmath160th hyperplane is incident with @xmath33 points .
this allows the problem to be framed as a linear system of equations : @xmath170 which can then be solved with software such as _ gurobi _ @xcite .
this also has the benefit of immediately revealing which hyperplanes are incident with @xmath32 points and which are incident with @xmath33 .
an initial search for two - intersection sets , filtered only by fulfilment of the necessary conditions discussed above , was able to return hundreds of thousands of results without exhausting the problem space . to prevent the re - computation of redundant data , the action of the full collineation group was used to impose a number of constraints on the search .
we computed valuable information regarding the symmetries of the projective space using the free software package gap @xcite .
firstly , it was noted that the normaliser of our element @xmath61 of order @xmath1 in the full collineation group acted transitively on the point - orbits of @xmath61 , meaning any solution was equivalent to one that included the first orbit .
hence , the first orbit was selected in every one of the reported solutions .
next , the subgroup that stabilised the first orbit was examined .
this group had nine orbits itself , represented by the orbits with indices in the set @xmath171 .
hence , any solution that included two or more orbits was equivalent to one that included either orbits 1 and 2 , or orbits 1 and 3 , or orbits 1 and 5 etc .
the solution space was then restricted to ensure that one of these orbit pairs was always included .
this method was reapplied to those subproblems that proved computationally difficult .
for instance , including orbits 1 and 2 in a solution was shown to be equivalent to including one of nine triples of orbits . in this manner , a tree of computations was formed and traversed , sometimes reaching seven levels deep .
one optimisation that proved particularly efficient was facilitated by the use of the branching method discussed above . as noted ,
there are many sets of orbits that are equivalent to one another .
the search was accelerated by precluding the consideration of any set equivalent to one already encountered .
for example , there exist 52 pairs of orbits whose inclusion is equivalent to the inclusion of the orbits 1 and 2 . hence , after all solutions containing orbits 1 and 2 have been found , any of these 52 equivalent pairs of orbits can be excluded from appearing again in any future results .
these optimisation techniques allowed for the enumeration of all @xmath61-invariant two - intersection sets in @xmath172 within a viable time - frame ( approximately 2 weeks ) . in order to have confidence in the correctness of the search
, we used the constraint satisfaction solver minion @xcite to simulate some of our results .
in particular , we enumerated all the two - intersection sets with parameters @xmath173 , @xmath174 , @xmath175 , @xmath176 , or @xmath177 , invariant under a group with order divisible by @xmath1 , and we obtained the same results which we outline in table [ smallparams ] : .examples for small parameter sets . [ cols="^,<,<",options="header " , ] the above examples give rise to strongly regular graphs with parameters @xmath178 .
the example with stabiliser of order @xmath179 arises by taking the derived subgroup of the stabiliser of the hill - cap .
this group has orbit lengths @xmath2 , @xmath2 , @xmath180 , @xmath180 on the points of @xmath3 .
taking a union of an orbit of size @xmath2 with one of size @xmath180 yields a two - intersection set with parameters @xmath181 .
the five examples in the table above that have an asterisk in the last column arise from partial spreads of @xmath3 ; the ` su2 ' construction in @xcite .
two of these examples arise from the ` cy2 ' construction in @xcite , denoted by two asterisks .
first the second author would like to acknowledge pablo spiga for invaluable discussions on research related to this work . in particular , there are some amorphic @xmath0-class association schemes related to some of the codes found in our search .
we would like to thank gordon royle and michael giudici for their invaluable time and for the use of their computational facilities .
finally , we also acknowledge andrew bassom for supporting the first author in a summer vacation scholarship .
|
we detail the enumeration of all two - intersection sets of the five - dimensional projective space over the field of order @xmath0 that are invariant under an element of order @xmath1 , which include the examples of hill ( 1973 ) and gulliver ( 1996 ) . up to projective equivalence
, there are 6635 such two - intersection sets .
|
it is well known that if we consider a natural classical hamiltonian system on the @xmath2-dimensional ( @xmath2d ) euclidean space = t()+u ( ) , [ uno ] the harmonic oscillator potential @xmath12 and the coulomb potential @xmath13 define two _ maximally superintegrable _ ( ms ) systems ( in the liouville sense ) , since both systems are endowed with @xmath14 functionally independent and globally defined integrals of the motion . in the first case
such integrals are provided by the components of the demkov
fradkin tensor @xcite , and in the second one by the angular momenta together with the @xmath2 components of the runge
lenz vector ( see e.g. @xcite and references therein ) . at the classical dynamical level , the footprint of superintegrability consists in the fact that all bounded trajectories of these two systems are closed ones , a fact which is diretly related with bertrand s theorem @xcite .
moreover , when the quantization of these systems is performed it is found that such superintegrability implies that their spectrum exhibits maximal degeneracy due to a superabundance of quantum integrals of the motion . in this paper
we review two spherically symmetric deformations of the oscillator and coulomb systems that define two new ms systems @xcite . as a consequence
, their quantization @xcite is shown to present maximal degeneracy in the spectra . at a first sight
, the existence of such deformations could seem impossible since the only spherically symmetric potentials on the euclidean space that are ms are just the oscillator and the coulomb ones .
therefore , the addition of any radial perturbation on these systems leads to superintegrability breaking and thus to a lack of maximal degeneracy in the spectra , a fact that is very well known in quantum perturbation theory .
however , as we shall see , such superintegrable perturbations can be obtained if _ both the potential and the kinetic energy _ are simultaneously deformed in a very precise way .
explicitly , the hamiltonian will be smoothly deformed into _ ( , ) = t_(,)+u _ ( ) , [ dos ] where @xmath15 can be regarded as a ( generic ) _ deformation parameter _ in such a manner that we will be no longer working on the _ flat _ euclidean space , but on a suitable curved space with metric and kinetic energy given by @xmath16 this fact will provide additional interesting geometric features to the systems we will deal with . in particular , we will see that the curved / deformed generalization of the demkov fradkin tensor and of the runge lenz vector do exist , and will be the essential tool to prove the ms property of the deformed systems .
we recall that the quantization problem on curved spaces is clearly a non - trivial one , since the kinetic energy term @xmath17 is a function of both positions and momenta that creates severe ordering ambiguities .
nevertheless , we shall explicitly show that a quantization of @xmath18 ( [ dos ] ) that preserves the ms property is achieved through the _ conformal laplacian quantization _
( see @xcite and references therein ) : _
c,=_c , + u_=-2 _
c,+u_= -2 ( _ lb , - r_)+u _ , [ conf ] where @xmath19 is the scalar curvature on the underlying @xmath2d curved manifold @xmath20 , the operator @xmath21 is the conformal laplacian @xcite and @xmath22 is the usual laplace
beltrami operator on @xmath20 , _
i.e. _ , @xmath23 where @xmath24 is the inverse of the metric tensor @xmath25 and @xmath26 is the corresponding determinant
. the limit @xmath27 gives rise to the quantization of the flat hamiltonian ( [ uno ] ) with @xmath28 since @xmath29 we also recall that the quantization ( [ conf ] ) can be related through a similarity transformation to the hamiltonian obtained by means of the so - called _ direct schrdinger quantization _ prescription on conformally flat spaces @xcite @xmath30 where @xmath31 is the conformal factor of the metric on @xmath20 written as @xmath32 . in this case
, the scalar curvature reads r_=-(n-1 ) ( ) .
[ aac ] in the next two sections , we review the exactly solvable deformations of the @xmath2d isotropic oscillator @xmath33 @xcite and the coulomb system @xmath34 @xcite , correspondingly .
new results are sketched in the last section by presenting the only possible multiparametric spherically symmetric generalizations of the above systems which are ms with _ quadratic _
integrals of motion , that is , the most generic deformations that can be endowed , respectively , with a generalized demkov fradkin tensor and with a runge lenz @xmath2-vector .
the @xmath2d classical hamiltonian system given by _ ( , ) = t_(,)+u_()= + , [ ac ] where @xmath3 and @xmath35 are real parameters and @xmath36 are canonical coordinates and momenta , was proven in @xcite to be ms .
the kinetic energy @xmath37 can be interpreted as the one generating the geodesic motion of a particle with unit mass on a conformally flat space with metric and ( non - constant ) scalar curvature ( [ aac ] ) given by ds_^2= ( 1+^2)d^2 , r_()=- .
[ ad ] such a curved space is , in fact , an @xmath2d spherically symmetric generalization @xmath38 @xcite of the darboux surface of type iii @xcite .
the limit @xmath39 of the above expressions leads to the well known results concerning the ( flat ) @xmath2d isotropic harmonic oscillator with frequency @xmath35 : = 12 ^2 + 12 ^2 ^ 2 , ds^2= d^2 , r=0 .
[ ae ] the remarkable point is that @xmath0 is a ms hamiltonian , a fact that can be stated as follows
. * proposition 1 .
* @xcite _ ( i ) the hamiltonian @xmath0 ( [ ac ] ) is endowed with the following constants of motion ( @xmath40 ) : _
@xmath41 @xmath42 angular momentum integrals : c^(m)= _
1i < jm ( q_ip_j - q_jp_i)^2 , c_(m)= _ n - m < i < jn ( q_ip_j - q_jp_i)^2 , c^(n)=c_(n ) . [ af ] @xmath41 @xmath43 integrals which form the nd curved / deformed demkov fradkin tensor ( @xmath44 ) : i _ , ij = p_ip_j-(2h_(,)-^2 ) q_iq_j , _
= 12 _ i=1^n i_,ii .
[ ag](ii ) each of the three sets @xmath45 , @xmath46 ( @xmath40 ) and @xmath47 ( @xmath48 ) is formed by @xmath2 functionally independent functions in involution .
\(iii ) the set @xmath49 for @xmath40 with a fixed index @xmath50 is constituted by @xmath14 functionally independent functions .
let us now consider the standard definitions for the quantum positions @xmath51 , momenta @xmath52 and @xmath53 operators ( @xmath54 : @xmath55=\rmi \hbar\ , \delta_{ij } , \qquad \mathbf{q}\cdot\nabla = \sum_{i=1}^n q_i\frac{\partial}{{\partial}q_i } \ , .\ ] ] if we now apply the conformal laplacian quantization prescription ( [ conf ] ) to @xmath0 ( [ ac ] ) we find : * proposition 2 . * @xcite _ let @xmath56 be the quantum hamiltonian given by @xmath57 ( i ) @xmath56 commutes with the @xmath42 quantum angular momentum operators @xmath58 ^(m)= _ 1i < jm ( _ i _ j- _ j _ i)^2 , _ ( m)=
_ n - m < i
< jn ( _ i _ j- _ j _ i)^2 ,
^(n)=_(n ) , [ cb ] as well as with the @xmath43 curved / deformed demkov fradkin operators given by & & _ c , , ij= _
i_j -(n-2 ) ( _ i_j + _
j_i ) + ( 1- ) + & & - _ ij - 2_i_j ( , ) + ^2 _ i_j , [ kp]with @xmath59 and such that @xmath60 .
_ \(ii ) each of the three sets @xmath61 , @xmath62 ( @xmath40 ) and @xmath63 ( @xmath48 ) is formed by @xmath2 algebraically independent commuting observables .
\(iii ) the set @xmath64 for @xmath40 with a fixed index @xmath50 is formed by @xmath14 algebraically independent observables .
\(iv ) @xmath56 is formally self - adjoint on the space @xmath65 , associated with the underlying darboux iii space , defined by | _
c , = _ m _ ( ) ( 1+^2)^n/2d . [ product2 ] the complete solution to the eigenvalue problem along with the corresponding eigenfunctions for the case of _ positive _ deformation parameter @xmath3 is summarized in the following statement . * theorem 3 .
* @xcite _ let @xmath56 be the quantum hamiltonian ( [ kn ] ) with @xmath66 . then : _
\(i ) the continuous spectrum of @xmath56 is given by @xmath67 .
there are no embedded eigenvalues and its singular spectrum is empty .
\(ii ) @xmath56 has an infinite number of eigenvalues , all of which are contained in @xmath68 .
their only accumulation point is @xmath69 which is the bottom of the continuous spectrum .
\(iii ) all the discrete eigenvalues of @xmath70 are of the form e_,n & = & - ^2(n + ) ^2 + ( n+n 2 ) , n .
[ lan ] ( iv ) the eigenfunction @xmath71 of @xmath70 with eigenvalue @xmath72 is given by @xmath73 where @xmath74 are hermite polynomials with @xmath75 such that @xmath76 and the deformed frequency @xmath77 is defined by @xmath78 ( [ lan ] ) for @xmath79 , @xmath80 , @xmath81 and @xmath82 starting from the upper dot ( straight ) line corresponding to the isotropic harmonic oscillator with @xmath83 , that is , @xmath84 .
in the same order , @xmath85 and @xmath86 .
[ figure1],width=415,height=302 ] moreover , the bound states of this system satisfy @xmath87 the discrete spectrum ( [ lan ] ) is depicted in figure [ figure1 ] as a function of @xmath88 for several values of @xmath3 .
as it was remarked in the introduction , the spectrum turns out to be maximally degenerate since it can be described as a function of just one quantum number @xmath89 . by taking into account the definition of @xmath88 , the number of degenerate states @xmath90 for a given energy level
@xmath91 corresponds to all the possible combinations of @xmath92 obeying to the constraint @xmath93 , namely @xmath94 which for @xmath80 reduces to the well known expression @xmath95 for the degeneracies of the isotropic oscillator .
therefore , the curved system has a similar set of integrals of the motion as the undeformed one and , as a consequence , it exhibits the same degeneracy .
now we consider the @xmath2d hamiltonian system given by _ = t_(,)+u_()=^2- , [ otro ] where @xmath4 and @xmath96 are real parameters .
the metric and scalar curvature of the underlying manifold @xmath97 turns out to be ds_^2=(1 + ) d^2,r_=(n-1 ) , r=||= .
[ metr]we remark that the system ( [ otro ] ) is directly related to a reduction @xcite of the geodesic motion on the taub nut space @xcite .
in fact , this system can be regarded as an @xmath4-deformation of the @xmath2d euclidean coulomb problem with coupling constant @xmath96 , since the limit @xmath98 yields @xmath99 remarkably enough , the hamiltonian @xmath100 turns out to be a ms classical system , and this result can be summarized as follows .
* proposition 4 .
* @xcite _
( i ) the hamiltonian @xmath101 ( [ otro ] ) is endowed with the @xmath42 angular momentum integrals ( [ af ] ) and poisson - commutes with the @xmath102 components ( @xmath48 ) of the runge lenz @xmath2-vector given by @xmath103 ( ii )
the set @xmath104 with @xmath40 and a fixed index @xmath50 is formed by @xmath14 functionally independent functions .
_ we also recall that the classical system @xmath100 has been fully solved in @xcite .
next the quantum counterpart of ( [ otro ] ) can be obtained by applying ( [ conf ] ) , and reads : * proposition 5 . * @xcite _
( i ) the quantum hamiltonian @xmath105 given by & & = -2_lb , - + ^2 ( n-2 ) , + & & _ lb,= - ( ) , [ kn2 ] commutes with the @xmath42 quantum angular momentum operators ( [ cb ] ) as well as with the following @xmath2 runge lenz operators @xmath106 : & & + & & + _ j=1^n ( _ j _ i - _ i _ j ) ( _ j + ) + ( _ c,(,)+k ) . _ \(ii ) each of the three sets @xmath107 , @xmath108 ( @xmath40 ) and @xmath109 ( @xmath48 ) is formed by @xmath2 algebraically independent commuting operators .
\(iii ) the set @xmath110 for @xmath40 with a fixed index @xmath50 is formed by @xmath14 algebraically independent operators .
\(iv ) @xmath105 is formally self - adjoint on the hilbert space @xmath111 with the scalar product | _ c , = _ m _ ( ) ( 1 + ) ^n/2d . for a _ positive _ value of the deformation parameter @xmath4 , the complete solution of the eigenvalue problem for this quantum mechanical deformed coulomb problem is the following . * theorem 6 . *
@xcite _ let @xmath105 be the quantum hamiltonian ( [ kn2 ] ) with @xmath112 and @xmath113 . then
\(i ) the continuous spectrum of @xmath105 is given by @xmath114 .
there are no embedded eigenvalues and the singular spectrum is empty .
\(ii ) @xmath105 has an infinite number of eigenvalues @xmath115 , depending only on the sum @xmath116 and accumulating at @xmath117 .
\(iii ) the eigenvalues @xmath118 of @xmath105 are of the form @xmath119 such that the radial eigenfunction @xmath120 of @xmath105 with eigenvalue @xmath118 reads @xmath121 where @xmath122 are generalized laguerre polynomials and the deformed coupling constant @xmath123 reads @xmath124 since @xmath105 is a hamiltonian with radial symmetry , its complete eigenfunction is so given by @xmath125 where @xmath126 denotes the usual hyperspherical harmonics , @xmath127 and @xmath128 is a vector of @xmath129 quantum numbers @xmath130 such that ( see ( [ cb ] ) ) @xmath131 ) for the fundamental and the three first excited states of the hamiltonian @xmath105 ( [ kn2 ] ) when @xmath132 with @xmath133 and @xmath134 .
note that the effect of the @xmath4 deformation is quite strong for the fundamental state , since it comes from the shift @xmath135 in the usual coulomb potential .
[ figure2],height=302 ] notice also that the bound states of this system satisfy _
n , le_,n , l= 0 , _ ( e_,+1-e _ , ) = 0 , = n+ l. as expected , the limit @xmath136 of @xmath118 provides the well known formula for the standard coulomb eigenvalues @xmath137 @xmath138 and we find that the perturbative series for the eigenvalues of the deformed system @xmath105 ( [ kn2 ] ) reads @xmath139 in figure 2 the eigenvalues of the fundamental and of the first three excited states are plotted for different values of the deformation parameter @xmath4 . as we can see from ( [ eq13 ] )
the spectrum is maximally degenerate as , again , it depends on a unique principal quantum number @xmath140 .
the degeneracy @xmath141 of a given energy level @xmath142 can be computed straightforwardly by taking into account that the cardinality @xmath143 given by the set of the hyperspherical harmonics @xmath144 having the same quantum number @xmath145 and such that @xmath146 reads @xcite @xmath147 from it we obtain that @xmath148 in particular , for @xmath80 we obtain @xmath149 , which coincides with the degeneracy of the energy levels of the undeformed coulomb problem .
so far we have reviewed some _ specific _ exactly solvable deformations of the oscillator and coulomb potentials , which can be regarded as the most natural ms deformations beyond constant curvature .
nevertheless , there are more possible generalizations within this framework that preserves the classical ms property and that would lead to other exactly solvable deformed oscillator and coulomb systems .
these arise within the classification of bertrand hamiltonians formerly introduced in @xcite and further developed in @xcite .
such systems are ms and their underlying bertrand spaces are spherically symmetric ones . if we require to keep _
integrals of motion , so generalizing the demkov
fradkin tensor and the runge
lenz @xmath2-vector , it can be shown that there only exists _
one _ possible generalization of the deformations of the oscillator and coulomb systems here studied that depends on _ two _ deformation parameters .
in particular , the two - parameter ms deformation of the oscillator system turns out to be _ , ( , ) = t_,(,)+u_,()= + , where @xmath150 is a real parameter . obviously , the limit @xmath151 gives rise to the hamiltonian @xmath152 ( [ ac ] ) the underlying manifold @xmath153 is endowed with a conformally flat metric given by @xmath154 and the corresponding scalar curvature ( [ aac ] ) reads & & r_,(r)=-\ { n(2 + 3r^2 + 6 r^4+r^6 ) ( + 3 r^4 + 2 r^2 ( 3+r^4 ) ) + & & -6r^2(^2- 4 ) ( 1-r^4)^2 } , where recall that @xmath155 .
as far as the coulomb system is concerned , the resulting two - parameter ms deformation is given by @xmath156 which generalizes the one - parameter hamiltonian @xmath157 ( [ otro ] ) .
hence the metric of the underlying spherically symmetric space @xmath158 and its scalar curvature are found to be @xmath159 & & r _ , ( r)=-\ { 6 ( 1-r^2)^2(+ r(2 + r(6 + r [ 2+r ] ) ) ) + & & -n(3 + r ( 4 + r [ 6-r^2 ] ) ) ( -r^2 ( 6 + r [ 4 + 3r ] ) ) } .
it is worth stressing that @xmath158 turns out to be the @xmath2d spherically symmetric generalization of the darboux surface of type iv @xcite constructed in @xcite .
consequently , by applying the conformal laplacian quantization ( [ conf ] ) to the above two - parameter hamiltonians , new exactly solvable systems , @xmath160 and @xmath161 , would be obtained as deformations of the oscillator and coulomb systems .
their solution would generalize the results presented in theorems 3 and 6 .
work on this line is currently in progress .
this work was partially supported by the spanish mineco through the ramn y cajal program ( a.e . ) and under grants mtm2013 - 43820-p ( a.b and f.j.h . ) and fis2011 - 22566 ( a.e . ) , by the spanish junta de castilla y len under grant bu278u14 ( a.b .
, a.e . and f.j.h .
) , by the icmat severo ochoa under grant sev-2011 - 0087 ( a.e . ) , and by a postdoctoral fellowship from the laboratory of mathematical physics of the crm , universit de montral ( d.r . )
|
we present two maximally superintegrable hamiltonian systems @xmath0 and @xmath1 that are defined , respectively , on an @xmath2-dimensional spherically symmetric generalization of the darboux surface of type iii and on an @xmath2-dimensional taub nut space .
afterwards , we show that the quantization of @xmath0 and @xmath1 leads , respectively , to exactly solvable deformations ( with parameters @xmath3 and @xmath4 ) of the two basic quantum mechanical systems : the harmonic oscillator and the coulomb problem . in both cases
the quantization is performed in such a way that the maximal superintegrability of the classical hamiltonian is fully preserved . in particular , we prove that this strong condition is fulfilled by applying the so - called conformal laplace
beltrami quantization prescription , where the conformal laplacian operator contains the usual laplace
beltrami operator on the underlying manifold plus a term proportional to its scalar curvature ( which in both cases has non - constant value ) . in this way
, the eigenvalue problems for the quantum counterparts of @xmath0 and @xmath1 can be rigorously solved , and it is found that their discrete spectrum is just a smooth deformation ( in terms of the parameters @xmath3 and @xmath4 ) of the oscillator and coulomb spectrum , respectively
. moreover , it turns out that the maximal degeneracy of both systems is preserved under deformation .
finally , new further multiparametric generalizations of both systems that preserve their superintegrability are envisaged .
ngel ballesteros@xmath5 , alberto enciso@xmath6 , francisco j. herranz@xmath7 + orlando ragnisco@xmath8 and danilo riglioni@xmath9 @xmath5 departamento de fsica , universidad de burgos , e-09001 burgos , spain @xmath6 instituto de ciencias matemticas , csic , nicols cabrera 13 - 15 , e-28049 madrid , spain @xmath10 dipartimento di matematica e fisica , universit di roma tre and istituto nazionale di fisica nucleare sezione di roma tre , via vasca navale 84 , i-00146 roma , italy @xmath11 centre de recherches mathmatiques , universit de montral , h3 t 1j4 2920 chemin de la tour , montreal , canada e - mail :
|
it is believed that in the early universe a series of vacuum phase transitions led to several types of topological defects @xcite .
topological defects are relics of the early universe and are expected to convey some information on physics of very high energy scales beyond our reach with ground - based accelerators . on the other hand , topological defects are candidates for the seed of the observed large scale structure of the universe such as sheet - like or filamentary structures or voids .
thus topological defects are attractive examples which connect the high energy physics and cosmology and we might be able to have some information on high energy physics through cosmological observations such as gravitational wave detection in future .
topological defects , if they existed , would interact with other strong gravitational sources such as black holes , and then they might have experienced large deformation and emitted some information on themselves as gravitational waves . if we succeed in detecting such gravitational waves and identifying them , we will be able to confirm the existence of topological defects and hence the occurrence of a vacuum phase transition .
the topological defects with finite extent such as the cosmic string or the domain wall are known to produce an unusual gravitational field @xcite and their dynamics are slightly complicated .
there are some studies on the interaction between a cosmic string and a black hole with the hope of detecting gravitational waves from cosmic strings .
de villier and frolov studied @xcite the dynamics of the scattering and capturing process of an infinitely thin test string by a schwarzschild black hole .
however less is known on the interaction between a domain wall and a black hole ; only the static configurations have been studied so far .
morisawa et al .
@xcite showed the existence of static configurations of a thick domain wall intersecting a schwarzschild black hole in its equatorial plane .
christensen et al .
@xcite numerically found a series of static configurations of a domain wall as a nambu - goto membrane in the schwarzschild spacetime , some of which represent an intersecting pair of a domain wall and a black hole . when arranged in sequence , these configurations seem to represent a scattering and capturing process of a domain wall by a black hole . in most situations of cosmological interest
, the thickness of a domain wall will be much smaller than the horizon radius of a solar - mass or primordial black hole , so that the infinitely thin wall ( membrane ) approximation will be valid in such cases .
when we further neglect the gravitational effect of a domain wall , its spacetime history , i.e. , the world sheet is governed by the nambu - goto action and it is called the nambu - goto membrane . a membrane lying in the equatorial plane of the black hole spacetime has the highest symmetry among the configurations which represent an intersecting pair of a membrane and a black hole .
this simple configuration is a possible candidate for a final state of the scattering and capturing process provided such a state stably exists . having this in mind
, we study the stability of a nambu - goto membrane at the equatorial plane of the reissner - nordstrm - de sitter ( rnds ) spacetime .
the charge of the black hole and the positive cosmological constant are included to see their effect on the stability .
this mimics the situation where a domain wall resulting from a vacuum phase transition is intersecting a charged primordial black hole in the early universe .
we consider the linear perturbation of the membrane by means of the covariant perturbation formalism @xcite .
the perturbation equation reduces to a wave equation on the world sheet with the mass term , which is negative in the presence of the charge of the black hole or the positive cosmological constant .
accordingly instability is naively expected .
solving the perturbation equation reduces to solving a two point boundary value problem .
we numerically solve it by the shooting and the relaxation methods .
we find that a membrane intersecting a rnds black hole is unstable .
we find the instability at least in the range of @xmath2 for @xmath3 , and @xmath4 for @xmath5 , whereas the membrane in the schwarzschild background is stable .
while our treatment of a domain wall is simple with both its thickness and gravitational effect neglected , our model of the membrane - black hole system is all the more elementary and relevant in other contexts .
the interaction between black holes and extended objects is important as an elementary physical process .
indeed much attention has recently been paid to the interaction between black holes and extended objects such as membranes or strings in the string / m - theory @xcite , the brane world scenario @xcite etc .. for instance in the context of the brane world models , our universe is described as a four - dimensional domain wall in the five dimensional bulk spacetime .
some people argue that a black hole on the gravitating membrane is realized as a `` black cigar '' in the bulk spacetime which intersects the membrane @xcite .
the rest of the paper is organized as follows . in sec.[eqs ] we review the covariant perturbation formalism and specify the perturbation equation and the boundary conditions .
the numerical algorithm is also explained there . in sec.[results ] we present the results of numerical and analytical considerations .
finally we summarize and discuss the results in sec.[discussion ] .
the history of a membrane ( world sheet ) is described by the timelike hypersurface @xmath6 embedded in the four - dimensional spacetime @xmath7 .
the embedding is given by @xmath8 where @xmath9 s are the spacetime coordinates and @xmath10 s the world sheet coordinates .
the induced metric @xmath11 on @xmath12 is given by @xmath13 the dynamics of the membrane is described by the nambu - goto action @xmath14=-\sigma\int_{\sigma } d^3\xi \sqrt{-\gamma } \label{ngaction},\ ] ] where @xmath15 represents the surface energy density of the membrane in its rest frame .
we consider the variation of the action with respect to @xmath16 , @xmath17 where @xmath18 is the spacetime affine connection and @xmath19 is the dalembertian on the world sheet @xmath20 recall the gauss - weingarten equation @xmath21 where @xmath22 is the extrinsic curvature defined as @xmath23 @xmath24 is the normal to the world sheet , @xmath25 is the world sheet covariant derivative and @xmath26 is the spacetime covariant derivative .
contracting eq .
( [ gausswein ] ) with @xmath27 , we obtain @xmath28 with this equation eq .
( [ variation ] ) becomes @xmath29 eq .
( [ variation2 ] ) has only the component perpendicular to the world sheet .
hence the variation parallel to the world sheet has no physical meaning .
finally the equation of motion of the membrane is @xmath30 this is the equation of minimal surfaces . in general , eq .
( [ eom ] ) can not be solved analytically . as noted above , the physically meaningful measure of the perturbation
@xmath31 is the scalar @xmath32 taking the variation of eq .
( [ eom ] ) with respect to @xmath33 and setting it to zero , we obtain the perturbation equation @xmath34 where @xmath35 is the projection operator onto the world sheet .
a membrane lying in the equatorial plane of the black hole spacetime has the highest symmetry among the configurations found by @xcite which represent an intersecting pair of a membrane and a schwarzschild black hole .
this simple configuration is a possible candidate for a final state of the scattering and capturing process of a membrane and a black hole provided such a state stably exists .
in general , spherically symmetric black hole spacetimes have the discrete symmetry about the equatorial plane , i.e. , the equatorial plane consists of fixed points of this symmetry .
hence the equatorial plane is totally geodesic and @xmath36 .
the most general spherically symmetric black hole spacetime is the reissner - nordstrm - de sitter spacetime @xmath37 where @xmath38 @xmath39 is the cosmological constant and @xmath40 and @xmath41 are the mass and charge of the black hole , respectively .
the relation between the spacetime coordinates @xmath42 and the world sheet coordinates @xmath43 of the membrane at the equatorial plane is simply @xmath44 then the induced metric ( [ indmetric ] ) is @xmath45 this means that the geometry of the membrane is a three - dimensional black hole spacetime . the perturbation equation ( [ perturb ] )
becomes @xmath46 the mass term becomes negative when the charge of the black hole or the positive cosmological constant is present , which implies instability .
this is why we consider the equatorial plane of the reissner - nordstrm - de sitter spacetime .
in addition , it mimics a cosmologically interesting situation where a domain wall resulting from a vacuum phase transition is intersecting a charged primordial black hole in the early universe . by a separation of variables , @xmath47 , eq .
( [ perturb2 ] ) is transformed into a schrdinger type equation in the three - dimensional black hole spacetime @xmath48 @xmath49\label{potential } , \ ] ] where @xmath50 is the usual tortoise coordinate .
for the cases @xmath51 we normalize @xmath52 as @xmath53 the present problem reduces to solving a schrdinger type equation , and so does the problem of the metric perturbation of a four - dimensional black hole spacetime .
hence we will follow the stability analysis of the latter .
we consider the cases @xmath51 .
then the tortoise coordinate @xmath54 goes from @xmath55 to @xmath56 .
@xmath57 corresponds to the event horizon and @xmath58 to infinity ( @xmath5 ) or the cosmological horizon ( @xmath59 ) .
the potential ( [ potential ] ) approaches zero as @xmath54 goes to @xmath60 .
therefore the asymptotic solution to eq .
( [ schroedinger ] ) is a linear combination of @xmath61 and @xmath62 . with the time dependence of the solution of the form @xmath63 , the solutions @xmath61 and @xmath62 represent an ingoing wave and an outgoing wave , respectively . at infinity or the cosmological horizon
we admit an outgoing wave , and at the event horizon an ingoing wave .
this is because we assume no external source of perturbation of the membrane during its evolution .
hence we set the boundary conditions as follows @xmath64 with @xmath65 constant .
this form of boundary conditions appears in the standard analysis of a quasi - normal mode of a black hole @xcite . with the time dependence @xmath63
, @xmath66 corresponds to unstable modes .
then from the boundary conditions ( [ bc ] ) , an unstable solution to eq .
( [ schroedinger ] ) decays as @xmath67 when @xmath68 goes to @xmath56 and as @xmath69 when @xmath70 goes to @xmath55 .
multiplying eq .
( [ schroedinger ] ) by @xmath71 ( upper bar means the complex conjugate ) and integrating by parts , we obtain @xmath72 } _ { -\infty}^{+\infty } = { \omega}^{2}\int_{-\infty}^{+\infty}{\left|\chi\right| } ^{2}d r _ { * } .\ ] ] as long as we consuder unstable modes , the surface term on the l.h.s .
vanishes and the integrals on both sides converge .
therefore we obtain @xmath73 since the r.h.s .
( [ omegasq ] ) is real , @xmath74 is real or pure imaginary . here
we seek for unstable modes ( @xmath66 ) .
so we can set @xmath75 .
then the boundary conditions ( [ bc ] ) read @xmath76 we shall examine the eigenvalue problem of eq .
( [ schroedinger ] ) subject to the boundary conditions ( [ bc2 ] ) .
the eigenvalue problem considered is reduced to a two point boundary value problem .
we define four dependent variables as @xmath77 where all @xmath78 s are functions of @xmath54 , and @xmath79 and @xmath80 are constants .
@xmath80 is the ratio of @xmath81 to @xmath82 ( or @xmath82 to @xmath81 ) .
the evolution equations are @xmath83 y_1 , \ ; y_{3 } ' = 0 , \ ; y_{4 } ' = 0 , \label{numevol}\ ] ] where the prime denotes the differentiation with respect to @xmath54 .
we impose two sets of boundary conditions .
the set 1 ( @xmath84 ) is @xmath85 @xmath86 the set 2 ( @xmath87 ) is @xmath88 @xmath89 the condition @xmath90 is used for checking the reliability of calculations .
note that the boundary conditions ( [ exbc1]-[exbc4 ] ) leave @xmath91 arbitrary .
in general an arbitrary choice of @xmath91 at one boundary and the subsequent integration of the dependent variables do not ensure that the boundary conditions at the other boundary are satisfied .
we solve this problem by the shooting method .
we also perform the calculation by the relaxation method to confirm the results of the shooting method .
for pure reissner - nordstrm cases the potential ( [ potential ] ) is positive semi - definite for @xmath92 : @xmath93 \ge 0 \hspace{2em}\mbox{($r\ge r_+ $ ) } , \ ] ] where @xmath94 is the radius of the outer event horizon .
the equality is satisfied only if @xmath95 .
therefore , by the argument of the energy integral @xcite , there is no unstable mode subject to the boundary conditions ( [ bc2 ] ) when @xmath92 . for the schwarzschild case with @xmath96 the radial function @xmath97 satisfies @xmath98 in terms of a new variable @xmath99
this equation becomes @xmath100 the solution to this equation is regarded as the zero energy eigenfunction for the potential @xmath101 .
this potential is positive definite since @xmath79 is real and negative .
therefore we have no solution which decays when @xmath102 approaches @xmath60 ( the event horizon and infinity ) as required by the boundary conditions ( [ bc2 ] ) @xmath103 hence there exists no unstable mode for the schwarzschild case . before describing the numerical results , we comment on the upper bound of @xmath104 for which unstable modes may exisit in the reissner - nordstrm - de sitter case . the potential ( [ potential ] ) takes the form @xmath105 where @xmath106 is defined by eq .
( [ f ] ) , @xmath107 and @xmath108 firstly we consider @xmath109 case .
@xmath110 is a monotonously decreasing function of @xmath111 and has the only one root for @xmath112 .
therefore , by the arguement of the energy integral @xcite , the sufficient condition for the stability of perturbation is @xmath113 where @xmath114 is the radius of the cosmological horizon .
( [ stconde0 ] ) reads @xmath115 where @xmath116 is used . when @xmath109 , @xmath117 .
therefore eq . ( [ stcond2e0 ] ) shows that at least for @xmath118 the potential ( [ potential ] ) is positive definite and that a membrane is stable against such perturbation .
next we consider @xmath119 case . for @xmath112 , @xmath120 either stays negative , or becomes positive for an interval @xmath121 but otherwise negative .
then the sufficient condition for the stability of perturbation is @xmath122 which excludes the possibility that @xmath120 stays negative , and ensures @xmath123 .
when @xmath124 increases , @xmath125 and @xmath126 decrease and @xmath127 and @xmath114 increase .
hence it is sufficient to consider @xmath128 case . in this case eq .
( [ stcondene0 ] ) reads @xmath129 where again @xmath130 are used . when @xmath3 , @xmath131 . for @xmath132 @xmath133
therefore eq . ( [ stcond2ene0 ] ) shows that at least for @xmath134 there exists no unstable mode .
the condition that the reissner - nordstrm - de sitter spacetime has a static region is @xmath135 , where @xmath136 this is a monotonously increasing function of @xmath137 and has the minimum value @xmath138 .
so we numerically investigate the cases @xmath139 .
we found unstable modes for various parameter sets @xmath140 .
the eigenvalues of @xmath96 modes are shown in fig .
[ plots ] .
the results of the shooting method and the relaxation method agree very well , the error being within 5% .
[ plots ] shows that when @xmath141 becomes larger , we have a larger @xmath142 and hence a higher growth rate of the perturbations .
therefore the charge of the black hole indeed destabilizes the membrane at the equatorial plane .
since the membrane is electrically neutral , this instability should be understood as the curvature effect of the charge of the black hole . as to the effect of the cosmological constant @xmath143 , we find no simple tendency : for a given @xmath141 , a larger @xmath143 does not necessarily lead to a larger @xmath142 .
however with a non - vanishing cosmological constant the magnitude of the eigenvalue is typically @xmath144 for @xmath145 in contrast to the rn cases , in which the magnitude of the eigenvalue almost exponentially decreases with decreasing specific charge and is far below @xmath144 for @xmath146 .
therefore we can at least say that the cosmological constant also destabilizes the membranes .
this expectation is supported by the consideration of the de sitter background case ( see appendix [ desitter ] ) .
we plotted some of the eigenfunctions for @xmath5 cases in fig .
[ profiles ] . some features of the profiles are listed in table [ features ] .
when we decrease the value of the specific charge of the black hole , the peak of the corresponding profile gets higher and approaches the event horizon .
as noted above we found that the charge of the black hole destabilizes the membrane in general .
however , in the present analysis , we could not show the existence of unstable modes for small @xmath141 ( @xmath147 ) when @xmath5 due to the difficulty in the numerical calculation as follows . as @xmath141 becomes smaller
, @xmath142 decreases approximately in powers of @xmath141 for @xmath148 ( see fig .
[ loglog ] ) .
judging by the extrapolation of this relation to the regime @xmath149 , @xmath142 will be infinitesimal when @xmath150 . in order for the asymptotic forms @xmath151 of @xmath152 to be justified
, @xmath153 must be much smaller than @xmath154 at the boundaries .
then the coefficients in the evolution equations ( [ numevol ] ) get extremely small and we suffer from underflows of numerical calculations . by the same numerical difficulty , we leave open the possibility that unstable modes with @xmath92 exist in the presence of the positive cosmological constant .
in this paper we numerically studied the stability of a nambu - goto membrane at the equatorial plane of the reissner - nordstrm - de sitter spacetime and found that in general such a membrane is unstable when the black hole is charged .
a membrane is a two - dimensional extended object in the three - space .
so when it moves towards a black hole , it is expected to be inevitably captured by the black hole .
a series of configurations found by christensen et al . seem to represent such situations . among these configurations
a membrane lying in the equatorial plane of the black hole spacetime has the highest symmetry and is most likely to be the final state of the scattering and capturing process . on seeing our result , however
, we expect that it is not the case .
in particular , fig . [ profiles ]
allows us to imagine that the membrane moves away from the equatorial plane
. then what eventually happens to that membrane ?
causality prohibits the membrane from escaping from the event horizon . by analogy with the scattering process of a cosmic string off a black hole @xcite
the membrane might experience large deformation and the topology of the membrane might change . here
are some speculations on the fate of the membrane ; it may settle down to some other configuration than the equatorial plane , or it may break up into two parts with one swallowed by the black hole and the other escaping to infinity .
we need to perform a full dynamical computation to resolve this issue .
however that is beyond the scope of this paper .
we ignored the gravitational effect of a domain wall whereas a gravitating domain wall is known to make a repulsive gravitational field @xcite , which is opposite to the strong attractive gravity of a black hole .
it is quite intriguing to find out the consequences of the competition of these opposite forces .
we would like to thank profs .
h. sato , t. nakamura , h. kodama and t. chiba for many useful comments .
we also thank y. morisawa and r. yamazaki for fruitful discussion .
we are grateful to dr .
t. harada for the discussion , especially , on the boundary conditions .
finally we thank the referee for constructive comments .
a.i . and d.i .
were supported by the jsps .
this work was supported in part by the grant - in - aid for scientific research fund ( d.i .
to understand the destabilizing effect of the positive cosmological constant , we compare the de sitter and minkowski cases . in these cases
( [ schroedinger ] ) is rewritten in terms of a new function @xmath155 as @xmath156 where @xmath104 is set to zero for simplicity .
the reason we use @xmath157 is that in the de sitter and minkowski cases the potential ( [ potential ] ) for @xmath158 diverges at the center @xmath159 and that numerical calculations become unstable . in the minkowski case @xmath160 and eq .
( [ besseleq ] ) becomes bessel s differential equation of order @xmath161 @xmath162 the solution of which is a linear combination of the modified bessel functions @xmath163 and @xmath164 .
however @xmath163 and @xmath164 diverge at @xmath165 and @xmath159 , respectively and do not satisfy the regularity conditions .
hence membranes are stable in the minkowski case .
( the stability for @xmath92 cases is also verified . )
when the positive cosmological constant is present , the situation changes . in the limit of @xmath166 @xmath167 , eq .
( [ besseleq ] ) reduces to @xmath168 unlike eq .
( [ mink ] ) , the solution to eq .
( [ ds ] ) differs depending on the sign of @xmath169 . when @xmath170 , the solution is a linear combination of the bessel function @xmath171 and the neumann function @xmath172 , where @xmath173 .
when @xmath174 , the solution is a linear combination of the modified bessel functions @xmath175 and @xmath176 .
the regularity condition at the center excludes @xmath177 and @xmath178 as the asymptotic solutions .
when @xmath170 , the asymptotic behavior of @xmath179 near the center changes from the minkowski case and we expect the emergence of unstable modes since @xmath180 is a bounded function .
in fact we found unstable modes by numerical calculations .
the method is similar to the one for the rnds cases .
the asymptotic form of @xmath179 near the cosmological horizon @xmath181 is determined to be @xmath182 by the boundary condition similar to eq .
( [ bc2 ] ) .
the eigenvalues are @xmath183 for @xmath184 , respectively .
though we performed numerical calculations just for three values of the cosmological constant , unstable modes are expected as long as the positive cosmological constant is present .
we calculated the energy spectra for a one - dimensional square - well potential to check the reliability of our numerical code .
we consider an eigenvalue problem @xmath185 for a square - well potential @xmath186 the problem reduces to a two point boundary value problem at the boundaries @xmath187 ( @xmath188 ) .
the dependent variables are @xmath189 where @xmath79 and @xmath80 are constants .
the evolution equations and the boundary conditions are similar to eq .
( [ numevol ] ) and eqs .
( [ exbc1],[exbc2 ] ) or eqs .
( [ exbc3],[exbc4 ] ) , respectively .
we can analytically show that the eigenvalues for the potential ( [ wellpot ] ) satisfy the equation @xmath190 where @xmath191 .
the eigenvalues @xmath192 of eq . ( [ testeq ] )
are labeled by @xmath193 and @xmath194 .
we found good agreements of eigenvalues @xmath192 obtained by solving the two point boundary value problem , with those obtained by directly solving eq .
( [ wellexact ] ) ( table [ testtable ] ) .
in this paper we present the results for which the shooting and relaxation methods agree within 5 % error .
the agreement is , however , far better than 5% in most cases .
the comparison of the two methods is summarized in table [ comp ] for some cases .
99 a. chamblin , s.w . hawking , and h.s .
reall , hep - th/9909205 .
s. chandrasekhar , _ the mathematical theory of black holes _ ( oxford university press , new york , 1992 )
. m. christensen , v.p .
frolov , and a.l .
larsen , phys .
d * 58 * , 085008 ( 1998 ) .
j.p . de villier and v.p .
frolov , phys .
d * 58 * , 105018 ( 1998 ) .
de villier and v.p .
frolov , class .
quantum grav .
* 16 * , 2403 ( 1999 ) .
r. emparan , g.t .
horowitz , and r. myers , jhep * 01 * , 007 ( 2000 ) . j. garriga and m. sasaki , hep - th/9912118 . j. garriga and a. vilenkin , phys .
d * 44 * , 1007 ( 1991 ) .
r. gregory , hep - th/0004101 .
r. gregory and r. laflamme , phys .
70 * , 2837 ( 1993 ) .
r. gregory and r. laflamme , nucl . phys . *
b428 * , 399 ( 1994 ) .
r. gregory and r. laflamme , phys .
d * 51 * , r305 ( 1995 ) .
j. guven , phys .
d * 48 * , 4604 ( 1993 ) .
j. ipser and p. sikivie , phys . rev .
d * 30 * , 712 ( 1984 ) .
for example , e.w . kolb and m.s .
turner , _ the early universe _ ( addison - wesley , redwood city , california , 1990 ) ; a. vilenkin and e.p.s .
shellard , _ cosmic strings and other topological defects _
( cambridge university press , cambridge , 1994 ) . y. morisawa , r. yamazaki , d. ida , a. ishibashi , and k. nakao , phys .
d * 62 * , 084022 ( 2000 ) .
l. randall and r. sundrum , phys .
* 83 * , 4690 ( 1999 ) .
l. randall and r. sundrum , phys .
. lett . * 83 * , 3370 ( 1999 ) .
a. vilenkin , phys .
d * 23 * , 852 ( 1981 ) . @xmath195 & peak location @xmath196 & peak amplitude @xmath197 & @xmath198 + 1 & 1 & 1.202 & 0.22 & 1.202 + 0.99 & 1.141 & 1.281 & 0.57 & 1.123 + 0.90 & 1.436 & 1.468 & 0.79 & 1.022 + 0.80 & 8/5 & 1.603 & 0.79 & 1.002 + 0.75 & 1.661 & 1.662 & 0.78 & 1.001 +
|
the stability of a nambu - goto membrane at the equatorial plane of the reissner - nordstrm - de sitter spacetime is studied .
the covariant perturbation formalism is applied to study the behavior of the perturbation of the membrane .
the perturbation equation is solved numerically .
it is shown that a membrane intersecting a charged black hole , including extremely charged one , is unstable and that the positive cosmological constant strengthens the instability .
* instability of a membrane intersecting a black hole * susumu higaki@xmath0 , akihiro ishibashi@xmath1 and daisuke ida@xmath0 + @xmath0_department of physics , kyoto university , kyoto 606 - 8502 , japan _
+ @xmath1_yukawa institute for theoretical physics , kyoto university , kyoto 606 - 8502 , japan _
|
the gravitational instability paradigm assumes that large - scale structure in the universe evolves in a hierarchical fashion , forming larger and larger systems via the assembly of smaller sub - units .
it is natural to think that the development of this structure should have a significant effect on the properties of galaxies that we observe today and that a strong correlation between these and the surrounding environment should be in general observed .
one of the most evident correlations of galaxy properties with the environment in the present - day universe is the morphology - density ( md ) relation : early - type galaxies , i.e. @xmath11 and @xmath12 ( _ spheroidal _ ) gas - poor objects are preferentially found in dense environments , such as groups and clusters ( oemler 1974 , dressler 1980 , giovanelli , haynes & chincarini 1986 ) .
measurements now exist also at redshifts up to unity , that show a similar relationship to be already in place at earlier times ( dressler et al .
1997 , andreon 1998 , smith et al .
2005 , postman et al .
2005 ) .
two competing scenarios can explain the md relation .
on one side , it is tempting to associate its origin to the build - up of the hierarchical dark - matter skeleton of the universe .
if mergers played a role in the early construction of the hubble sequence , with ( some or all ) early - type galaxies plausibly produced from the merging of gas - rich sub - systems , we would expect the cross - section for this process to depend strongly on the typical velocity dispersion of the hosting structure , which in turn depends on the mass scale of the environment where the galaxy lives .
equally important effects on the evolution of the galaxies stellar and gaseous components may also be provided by other environmental effects in rich clusters , like ram - pressure stripping by the dense intra - cluster medium ( icm ) ( e.g. gunn & gott 1972 , quilis et al .
2000 ) , close - encounters ( barnes 1992 ) and `` harassment '' via multiple weak encounters ( moore et al . 1996 ) , or `` strangulation '' ( larson et al . 1980 ) .
this seemingly natural picture would point to a _ nurture _ origin for the md relation : that it is the product of one or more evolutionary processes acting on already formed galaxies .
the alternative picture is that morphological properties might be set early on at galaxy formation , and are due to the very _ nature _ of the object , such as the mass of its hosting dark - matter halo .
a direct systematic investigation of these processes at different cosmic epochs has been so far limited by the relatively small size of the samples for which hst morphological information is available ( e.g. smith et al .
2005 ) , and thus by the difficulty to adequately sample the full range of environments . the cosmic evolution survey ( cosmos ) , the largest contiguous survey ever with hst , provides for the first time a combined data set capable of simultaneously exploring large - scale structure and detailed galaxy properties ( luminosity , size , color , morphology , nuclear activity ) out to a redshift approaching @xmath13 . in this paper
, we use deep , panoramic multi - color imaging from subaru suprime - cam , extended x - ray imaging from xmm - newton and the superb resolution hst data to characterise an outstanding large - scale structure at @xmath14 within the cosmos field and explore in detail the relationship between galaxy morphology and the environment ( local density ) . in a second
, parallel paper ( cassata et al . 2007 ) , we instead discuss for the same sample the environmental dependence and morphological composition of the galaxy color - magnitude diagram . additionally , capak et al .
( 2007b ) use the full cosmos data set to explore the evolution of the md relation for @xmath15 $ ] over the whole range of environments .
the paper is organised as follows : in 2 we provide a brief description of the cosmos data sets used in this work ; in 3 we discuss the properties of the large - scale structure , in terms of galaxy surface density , weak lensing projected mass distribution and x - ray surface brightness ; here we study in particular the lensing and x - ray properties of dominant cluster of galaxies in the structure ; in
4 we introduce our estimates of galaxy morphology , presenting the measurement of the md relation ; in 5
we show how galaxy morphologies are affected by large - scale features detected in the intra - cluster medium of the dominant cluster ; finally , we summarise our results in 6 .
we adopt throughout the paper a cosmological model with @xmath16 , @xmath17 and , in general , parameterise explicitly the hubble constant as h = h@xmath18 when quoting linear quantities , densities and volumes . when computing absolute magnitudes and x - ray luminosities we specialize it to @xmath19 .
the cosmos project is centred upon a complete survey in the near - infrared f814w ( @xmath20 ) band using the advanced camera for surveys ( acs ) on board hst of a 1.7 square - degree equatorial field ( @xmath21(j2000 ) = @xmath22 , @xmath23(j2000 ) = @xmath24 , scoville et al .
2007 ) . with an acs field of view of 203 arcsec on a side , this required a mosaic of 575 tiles , corresponding to one orbit each , split into 4 exposures of 507 seconds dithered in a 4-point line pattern .
the final 575 `` multidrizzled '' ( koekemoer et al .
2002 ) images have absolute astrometric accuracy of better than 0.1 arcsec .
a version with 0.05 arcsecond pixels was used to measure the morphology of galaxies in the structure ; a more finely sampled version with 0.03 arcsecond pixels was required to measure the shapes of more distant galaxies behind the cluster , which are needed for the weak lensing analysis .
more details and a full description of the acs data processing are provided in koekemoer et al .
( 2007 ) .
the cosmos field is also the target of further observations in several regions of the electromagnetic spectrum ( see scoville et al
. 2007 for an overview ) .
extensive ground - based imaging has been performed using suprime - cam ( kaifu et al .
2000 ) on the 8.2 m subaru telescope on mauna kea , in six bands , @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 , and @xmath30 , with a @xmath31 sensititivity in ab magnitudes corresponding to 27.3 , 26.6,27.0,26.8 , 26.2 , 25.2 , respectively ( taniguchi et al .
2007 , capak et al .
these data were complemented with further imaging in @xmath32 @xmath33 and @xmath34 from cfht - megacam , kpno and ctio respectively .
the resulting global photometric catalogue contains around 796,000 objects to @xmath35 .
more details on the observations and data reduction are given in taniguchi et al .
( 2007 ) , capak et al .
( 2007a ) and mobasher et al .
( 2007 ) .
photometric redshifts were derived using a bayesian photometric redshift method ( bpz ) ( benitez 2000 ; mobasher et al .
2007 ) , which uses six basic sed types , together with a loose prior distribution for galaxy magnitudes .
the output of the code includes , in addition to the best estimate of the redshift , the 68 and 95% confidence intervals , the galaxy sed type and stellar mass and its absolute magnitude in several bands .
the latter is particularly important for the current analysis , where the morphology - density relation has been studied for different luminosity thresholds .
based on nearly 1200 objects for which spectroscopic observations are available ( not considering catastrophic failures ) , the difference between spectroscopic and photometric redshifts has a typical _ rms _ value @xmath36 .
this figure is consistent with the restricted statistical sample of spectroscopic redshifts available within the large - scale structure at @xmath37 studied here ( [ proj_dens ] ) . to further improve on this ,
for the present analysis we restrict the photometric redshift sample to @xmath38 , corresponding roughly , at the structure mean redshift and for a mean sed type , to an absolute magnitude @xmath39 in the adopted cosmology .
more details on the quality of the photometric redshifts can be found in mobasher et al .
( 2007 ) .
relevant for the present work is the extensive x - ray survey that is nearing completion using the xmm - newton satellite , for a total granted time of 1.4 msec ( hasinger et al .
2007 , cappelluti et al . 2007 ) .
the mosaiced data used in this paper correspond to the first 0.8 msec , providing a median 40 ksec effective depth , taking vignetting into account .
this in turn results in a typical flux limit for extended sources of @xmath40 erg s@xmath3 @xmath41 , an unprecedented value for a survey this size . a first catalogue of x - ray selected clusters in the cosmos field is also being presented in this same volume ( finoguenov et al .
2007 ) . a full description of the the xmm observations can be found in hasinger et al .
( 2007 ) . in this paper we present a pilot study that concentrates on a small subset of the cosmos data , covering an area of about @xmath42 in right ascension and @xmath43 in declination .
this corresponds to 30 acs tiles and is covered by all the multi - wavelength data sets described above . for several of our computations ( as e.g. to estimate the overall mean galaxy background )
, we shall also make use of the full cosmos catalogue over the entire @xmath44 square degrees .
a search for large - scale structures within the cosmos field based on adaptive - smoothing within photometric redshift `` slices '' , finds among a number of structures found at different redshifts one conspicuous peak in the redshift interval @xmath45 $ ] ( scoville et al .
2007c ) . as we shall show in the following sections , the overall structure is evident in both a weak - lensing analysis of the whole cosmos field ( massey et al . 2007 ) and in the distribution of x - ray emission from the xmm mosaic ( finoguenov et al .
specifically , the strongest peak in the lensing map is also one of the most prominent extended x - ray sources in the cosmos field ( massey et al .
2007 ) . in the following sections
, we shall probe the reliability of these different observations in locating structures and characterize the environment and its correlation to galaxy morphologies .
scoville et al . (
2007c ) use an adaptive kernel technique to map the surface density of galaxies in intervals of photometric redshift . here
, we aim at using an estimate of local density that would allow us to study , in particular , the morphology - density relation within the structure and compare that consistently with previous measurements . following dressler ( 1980 ) and similar works in the literature ,
we have thus computed the @xmath10 estimator @xmath46 where @xmath47 is the projected distance in h@xmath3 mpc ( computed from the angular separation using the galaxy corresponding angular diameter distance ) to the 10-th nearest neighbour .
this is an adaptive estimate of density , with the major advantage over a fixed aperture measurement of avoiding being dominated by shot noise in low - density regions .
dressler s original work was limited to fairly luminous galaxies within rich abell clusters , i.e. well - defined , local structures of high contrast where the two - dimensional over - density signal produced by cluster galaxies strongly dominates over the background .
the estimator @xmath10 is thus in this case a fairly direct ( projected ) probe of the 3d density around the selected galaxy , with negligible contribution from the background / foreground objects . when the same definition is applied to the general `` field '' , it is necessary to specify a typical depth over which the integration is performed , as to remain a meaningful proxy for the true 3d environment .
a sensible choice is a size of the order of the galaxy correlation length or the typical group / cluster velocity dispersion ( @xmath48 ) .
the ten nearest neighbours around each galaxy are then counted within such a redshift cylinder and used to estimate @xmath10 .
this kind of approach has been used at high redshift by smith et al .
( 2005 ) , who extend their study also out of rich clusters .
the problem is further exacerbated when only photometric redshifts are available . with typical errors of the order of @xmath49 in redshift ( corresponding to moving a galaxy along the line of sight by @xmath50 or more ) , the choice of the redshift interval over which integrating galaxy counts becomes crucial .
this is essentially a compromise between a range large enough to recover all true companions scattered along the line of sight by the redshifts errors ( completeness ) and small enough to minimize fake projections ( contamination ) .
the situation in our case , with a sample centred on a well - defined main structure located at @xmath51 , is slightly simplified with respect to a more general case of estimating local density for galaxies within a generic redshift slice ( cooper et al . 2005 ) . following similar work in the literature ( e.g. kodama et al .
2001 ; smith et al . 2005 ; postman et al . 2005 ) , we looked first at the dispersion in the differences between spectroscopic and photometric redshifts .
the overall standard deviation in the cosmos photometric redshift catalogue , based on the larger set of @xmath52 spectroscopic redshifts from the first runs of the zcosmos redshift survey ( not covering this area yet ) is @xmath53 ( mobasher et al .
2007 , lilly et al . 2007 ) .
in addition , vlt - fors1 spectroscopy of 15 red galaxies around the central cluster of this structure ( comastri , private comm . ) , give a comparable value for this redshift , @xmath54 , with a median redshift for the cluster of @xmath55 . in the following , we shall assume that the bulk of our structure within the @xmath56 area under study , lies at the mean redshift @xmath37 .
this assumption is fully justified by the distribution of photometric redshifts and allows us to simplify our computations . on the basis of these indications ,
we use the simple approach of selecting a single slice centered at z=0.73 ( similar to kodama et al .
2001 ) and with appropriate thickness , and then correct statistically the measured @xmath10 of each galaxy for the background contamination .
we obtain a first estimate the robustness of the method and of the most appropriate thickness to be chosen for the redshift slice , by exploring three different sizes , @xmath57 , 0.12 , 0.18 corresponding to @xmath58 , 2 and @xmath59 s ( note that these values comfortably include the velocity dispersion of any known system , being this about 20 times smaller than the redshift errors ) . additionally , in appendix
a we present a more direct test of how well we are able to recover the true density , given the photometric redshift errors , based on the analysis of realistic mock samples of the cosmos survey constructed from the millennium simulation ( kitzbichler & white 2006 ) .
a more sophisticated approach to measure local densities from the photometric sample has been adopted in our parallel work on the evolution of the morphology density relation from cosmos ( capak et al .
2007b ) ; in that paper , we use a moving @xmath60 cylinder centred at the redshift of each galaxy ( as in smith et al . , 2005 ) , and describe each galaxy not as a dirac delta function in redshift space , but as a broader probability distribution corresponding to its specific photometric redshift likelihood function .
as shown by our tests , the simpler approach adopted here is more than adequate for the specific geometry of our sample , producing results fully consistent with capak et al .
( 2007b ) .
one further source of uncertainty is related to the background correction to the measured density . in principle , with a sufficiently thin slice and spectroscopic redshifts , one might reasonably assume that , outside filaments and clusters , the density of galaxies is practically zero , i.e. that no background correction is needed . in the case of photometric redshifts , however , where a much thicker slice has to be used , a number of background and foreground galaxies not belonging to the inner structure will be included , shifting the background surface density to a non - zero value .
we assume that this term is dominant and estimate the background correction as the median value of @xmath10 over the full cosmos 2 square degrees , in the corresponding redshift slice .
as shown in the appendix , this correction works very well and corrects for the bias in @xmath10 that would tend to systematically overestimate local densities . finally , another choice needed to make @xmath10 fully consistent with previous works , is that of the range of absolute magnitudes to include in the computation . in the original dressler ( 1980 )
paper , the limit was set to @xmath61 , using @xmath62 .
considering a brightening of the luminosity function of 0.6 mag to @xmath63 ( e.g. zucca et al .
2006 ) , which is comparable to values adopted in smith et al .
( 2005 ) or treu et al .
( 2003 ) , and converting to @xmath19 , this corresponds to a value @xmath64 for our data .
we note that there is little agreement in the literature on the appropriate amount of brightening to be applied .
part of the difficulty arises because the actual average brightening depends significantly on the relative contributions of different spectral types : the luminosity function of late - type galaxies evolves more rapidly than that of early types , as shown by the vvds survey results ( zucca et al .
2006 ) . as we shall show in [ morph ] , the resulting md relation does not depend significantly on either the choice of the redshift interval or a brightening between 0 and 1 magnitudes . in figure [ dens_map ] ,
we plot contours of constant projected density @xmath10 , as measured from the @xmath65 slice .
luminous galaxies ( @xmath64 ) with photometric redshifts in the same interval are indicated by circles , with symbol sizes proportional to the value of local density .
we remind that this analysis uses only 30 acs tiles , out of the whole 575 cosmos - acs images .
these tiles cover an inclined rectangle over this area , as indicated by the lack of objects in the corners of the figure .
at least six galaxy concentrations of different strength are evident in this figure , some of which aligned along filaments .
a zoom over the most prominent density peak is displayed in figure [ photoz_overlay ] .
galaxy positions for the luminous galaxies are over - plotted on a section of the @xmath30-band subaru image .
note the `` s''-shaped filament of galaxies seemingly running se - nw , with a sharp bend near the center .
the observed shapes of background galaxies are slightly distorted as light from them is deflected around the gravitational field of foreground structures .
the high resolution hst coverage of the cosmos survey allows us to measure these distortions and to reconstruct the weak lensing `` shear field '' with unprecendented precision . to reconstruct the distribution of mass in this field , we first measured the shapes of galaxies using the `` rrg '' algorithm by rhodes , refregier & groth ( 2000 ) .
the specific application of this method to the cosmos data is fully described in leauthaud et al .
( 2007 ) and massey et al .
the galaxy shapes were then corrected for the effects of instrumental point - spread function ( psf ) convolution using the psf models described in rhodes et al .
( 2007 ) .
photometric redshifts from mobasher et al .
( 2007 ) were used to identify 37.6 galaxies per square arcminute behind the cluster , with a mean ( median ) photometric redshift of 1.50 ( 1.25 ) .
reaching this number density for such a high - redshift structure required a slight reduction in the size and magnitude thresholds used for the analysis of the entire cosmos field by massey et al .
this is possible with a minimum impact in signal to noise because of the strong signal in this region .
however , we have increased the estimate of noise due to possible shear calibration error to @xmath66 to include the additional uncertainty in the shapes of very faint galaxies .
the shear calibration error represents uncertainty in a shear measurement method s ability to deconvolve the shapes of faint galaxies from the instrumental psf and to accurately measure their ellipticities .
the accuracy of our shear measurement method has been estimated using simulated images that resemble cosmos data but which contain a known shear signal .
in fact , as popularised by the shear testing programme ( step ; heymans et al .
2006 ; massey et al .
2007b ) , shear measurement errors can actually be parameterized as two numbers : the shear calibration error @xmath67 ( multiplicative error ) , and the residual shear offset @xmath68 ( additive error ) , although in the cluster lensing regime , where the signal is large , the latter source of error is negligible . both are estimated and tabulated in the parallel paper by leauthaud et al .
( 2007 ) .
the observed shear field from the entire cosmos field was transformed into a convergence @xmath69 map using the method of kaiser & squires ( 1993 ) , filtered by a gaussian of rms width @xmath70 .
the convergence field is related to the newtonian gravitational potential @xmath71 as @xmath72 where @xmath73 @xmath74 is the comoving derivative perpendicular to the line of sight and @xmath75 is the radial sensitivity function @xmath76 in this expression , @xmath77 is the cosmological scale factor , @xmath78 s are angular diameter distances and @xmath79 is the distribution of background source galaxies , normalised such that @xmath80 as shown in figure [ lensing - sensitivity - function ] , in our analysis @xmath75 peaks between @xmath81 and @xmath82 .
lensing is less sensitive to structure closer or more distant than this . in figure
[ morph_map_bright ] , we have plotted the contours of constant snr in the reconstructed projected mass distribution over our study field ( see [ structure_overview ] for more discussion and comparison to the galaxy and x - ray distributions ) . figure [ lensing_overlay ] , instead , shows a zoom on the strongest peak , centred at ra=149.922 , dec=2.515 , and coinciding with the dominant cluster of galaxies within the structure already shown in figure [ photoz_overlay ] ( cf . box in figure [ dens_map ] ) .
it is of interest to estimate the mass of this cluster . to this end
, we first need to fix the geometry of the lens .
we assume that all of the foreground mass lies at @xmath63 and that the background galaxies lie at exactly the redshifts indicated by their photo-@xmath83 estimates .
we then first apply the mass aperture statistic @xmath84 of schneider et al .
( 1998 ) @xmath85 which calculates the value of the surface mass inside the radius @xmath86 on the sky , convolved with a compensated filter @xmath87 where @xmath88 and @xmath89 is the heaviside step function ( see clowe et al .
2006 , for a similar application ) . using a @xmath90 window
, we find @xmath91 where the three sets of @xmath92 error bars in equation ( [ eqn : maperrors ] ) respectively correspond to statistical noise in galaxy shape measurement , including their intrinsic ellipticities ; background and foreground redshift uncertainty ; and shear calibration uncertainty .
these have been combined in quadrature in equation ( [ eqn : map ] ) .
the @xmath84 statistic allows us to assess the significance of our measurements : gravitational lensing by a single cluster produces a shear field that is aligned tangentially around the mass overdensity ( a void would produce a radial shear field ) .
an independent component of the shear field can be separated : by analogy with electromagnetism , this is known as a `` curl '' or @xmath93-mode , and can be measured by rotating galaxies by @xmath94 in advance .
since it is unphysical , it is expected to be zero in the absence of systematics , and provides a good estimate of both systematics plus statistical errors . around this cluster ,
we measure the @xmath93-mode equivalent of @xmath84 to be @xmath95 where the error bars are the same as before . that this is an order of magnitude below the mass signal confirms that systematic effects have been successfully purged from the lensing analysis . however , as evident from fig .
[ dens_map ] the cluster is in a complex large - scale environment , shows significant substructure in the process of merging , and is at the nexus of several filaments ( see also scoville et al .
2007 , massey et al . in preparation ) . to estimate the total mass from its extended dark matter halo ,
we have then fitted a circularly symmetric nfw profile ( navarro , frenk & white 1997 ) to the radial shear field , using the technique of king & schneider ( 2001 ) . as shown in figure [ nfw - fit - to - shear ] , this finds a much larger mass .
the best - fitting model has @xmath96 .
as we shall show in the next section , this cluster is also one of the most prominent extended x - ray sources in the cosmos field .
the grey - scale levels in figure [ morph_map_bright ] shows the distribution of x - ray surface brightness ( in units of standard deviations from the background ) over the area under study , obtained from the reduced xmm mosaic described in hasinger et al .
( 2007 ) and finoguenov et al .
figure [ xray_overlay ] shows a more detailed zoom over the same cluster sub - area of figures [ photoz_overlay ] and [ lensing_overlay ] .
the galaxy cluster emerges as a powerful extended x - ray source , with an observed flux in the @xmath97 $ ] kev band of @xmath98 erg @xmath41 s@xmath3 , corresponding at @xmath63 to an x - ray luminosity ( in the @xmath4 $ ] kev band ) of @xmath99 erg s@xmath3 ( finoguenov et al .
2007 ) .
in addition to the central cluster , a number of fainter sources are detected .
comparison to figure [ photoz_overlay ] clearly shows some of them obviously coinciding with galaxy concentrations . despite the different selection functions , the structure seen in the galaxy distribution and in the lensing and x - ray maps
is remarkably similar .
note in particular the extension toward south - east in the lensing map , coinciding with the secondary galaxy condensation along the `` s''-shaped filament and with an extended source in x - rays .
cll zone & @xmath100 & notes + 1 & @xmath101 & cluster outskirts + 2 & @xmath102 & cluster global icm + 3 & @xmath103 & possible shock region / point source ? + 4 & @xmath104 & cluster core + 5 & @xmath105 & sub - structure + 6 & @xmath106 & `` infalling '' group + 7 & @xmath107 & intermediate region 6 - 1 + combining these complementary data sets allows us to start building up a coherent picture of large - scale structure in this region , which appears to be dynamically very active .
for example , looking at the x - ray map of figure [ xray_overlay ] , one notices a clear asymmetry in the surface brightness distribution of the central cluster , which also shows some significant ( at xmm resolution ) substructure . the main feature is a significantly steeper gradient in surface brightness toward north - east , with respect to a looser profile in the opposite direction , possibly suggesting the presence of a shock front .
this edge coincides also with the sharp bend visible in the `` s''-shaped galaxy filament of figure [ photoz_overlay ] , another indication that it might correspond to the main interface of an ongoing major merging event between two sub - structures along the sw - ne direction .
the depth of the xmm observations allows us to look for the fingerprint of a shock , by studying in some detail the x - ray temperature map over the cluster area , and in particular near the ne rim .
we have thus divided the area into seven regions , estimating for each of them the x - ray temperature .
this has been performed by fitting an emission spectrum from collisionally - ionized diffuse gas , computed using the apec code ( v1.3.1 . ,
see http://hea-www.harvard.edu/apec/ for more information ) .
the location and numbering of each region is reported in the sketchy map of figure [ tx_regions ] , and the corresponding meaning and resulting temperature measurements are reported in table [ table : tx ] .
the cluster icm is found to have a global x - ray temperature ( integrated over a radius @xmath108 arcmin , corresponding to 0.43 h@xmath3 mpc ) of @xmath109 kev .
using the relation derived by pacaud ( private comm . ) from the data of finoguenov et al .
( 2001 ) @xmath110 where @xmath111 is the usual cosmological function specialized to our cosmology , we obtain a mass @xmath112 . alternatively , using the @xmath113 relation from vikhlinin et al .
2006 , and accounting for evolution in the scaling relations as in kotov & vikhlinin ( 2005 ) , we obtain @xmath114 .
the difference between these two estimates is well within the intrinsic uncertainties .
most interestingly , the measurements show a temperature jump corresponding to the `` edge '' seen in the surface brightness ( region 3 ) : in this region we have @xmath115 kev .
this is a strong indication of the presence of a shock front in this area error due to
the limited number of counts available from such a restricted area corresponds in fact to a significant deviation ( @xmath116 confidence , i.e. @xmath117 ) of the value of @xmath118 with respect to the overall cluster icm ] , as speculated from the x - ray surface brightness alone . a residual uncertainty on this interpretation is related to the tentative presence of a point - source within region 3 , which if confirmed would pollute our spectral estimate of the temperature .
however , an hardness - ratio map also shows that the ne side of the cluster core is systematically hotter than the sw one , on a scale significantly larger than the size of a point source .
finally , we note that the observed factor of @xmath119 overestimate in our mass measurements based on x - ray scaling relations is consistent with being a consequence of the hotter core of the cluster ( that moves the cluster off the scaling relation ) , caused by the observed merging event . in the not too distant future
, it will be possible to study this system with much more spatial detail , benefiting of the forthcoming chandra observations of the cosmos field .
having characterised in some detail the large - scale distribution of galaxies and mass within this particular sub - region at @xmath0 , we turn now to exploring the connection of galaxy morphologies to the large - scale environment .
figure [ morph_map_bright ] summarises and compares the information gathered so far .
as already mentioned , this figure shows a remarkable coincidence of the contours in the surface mass ( convergenge ) map from lensing with the diffuse x - ray emission .
this is even more impressive if we consider that the lensing reconstruction shows mass integrated over a broad redshift range , with a different sensitivity function to the @xmath121-ray and galaxy density analyses ( see figure [ lensing - sensitivity - function ] ) .
due to projection , therefore , one expects some real structure in the lensing map that does not appear in the other analyses , and vice versa .
this makes the coincidence of the strongest galaxy concentrations with the peaks in the x - ray emission and in the lensing mass distribution even more fascinating .
in addition to the structures within our study field , note also the two strong x - ray / lensing blobs partly entering at the top and bottom of the figure , which are found to be foreground clusters , with mean photometric redshift of 0.35 and 0.22 respectively ( finoguenov et al .
2007 ) . finally , in this figure we have also indicated galaxy morphological types , estimated from the acs data as detailed in the following section .
the distribution of early - type ( red circles ) and late - type galaxies ( blue asterisks ) provides an immediate qualitative evidence for a significant morphology - density relation within this structure . in the parallel paper by cassata et al .
( 2007 ) , we have measured automatically non - parametric galaxy morphologies for all galaxies in the cosmos field . in brief , morphologies have been characterized on the basis of the so - called concentration , asymmetry , clumpiness , m20 and gini index ( see cassata et al .
2007 for details and references ) and the sample has been split into two broad morphological classes : early- and late - type galaxies .
cassata et al .
( 2005 , 2007 ) show that it is possible to define a precise region in the space of these 5 parameters , to perform this accurately , with high completeness and little contamination with respect to a visual classification .
stars were removed using sextractor according to their class - star and flux - best parameters re - computed on the acs data , where this method proves to be reliable , much more than on ground - based images . using this combined data set
, we shall explore here in detail the behaviour of the md relation within the redshift slice including this large - scale structure .
with only few spectroscopic redshifts available over the full area , working on a structure well - confined in redshift space ( while still spanning a significant range of over - densities ) allows us to be reasonably confident that redshift errors have a negligible impact on our conclusions .
in addition to the experiments with mock samples discussed in the appendix , we will test the robustness of our conclusions explicitly by varying the size of the sample and its limiting luminosity . in this analysis
we shall concentrate on the overall md relation observed at this redshift , its dependence on galaxy luminosity and in particular to its possible relation with large - scale structure as described by x - ray emission .
this study is complemented by the work of cassata et al .
( 2007 ) , who explore in detail the morphological composition and environmental dependence of the color - magnitude relation for galaxies within this same structure , and that of capak et al .
( 2007b ) , who investigate the global evolution of the md relation out to @xmath122 using the whole cosmos field data . the md relation , i.e. the fraction of early - type galaxies as a function of @xmath10 for our @xmath37 sample , computed within each of the three redshift slices defined in [ proj_dens ]
and for @xmath64 is shown in figure [ morph - dens ] . as discussed , the surface density @xmath10 has been normalised within the three different redshift slices , by subtracting the median background value estimated over the whole cosmos 2-square - degrees field .
error bars are here simply the poisson s errors on the number of objects in each bin . however , the scatter among the three estimates is comparable to the the variance one obtains by propagating the ensemble errors obtained from the simulations in the appendix .
these errors on the derived density can move galaxies horizontally among the bins , but have a significant effect only at the lowest densities ( @xmath123 gal h@xmath124 mpc@xmath125 ) . note also that the background correction for the three slices ranges between 20 and 40 h@xmath124 mpc@xmath125 , and it is shown to be crucial to reduce the bias introduced by the photo - z errors below @xmath126 where the uncertainties become dominant . overall , however , the result is remarkably stable , with the three curves being very consistent with each other .
this reassures us that the measurement is not strongly dependent on the size of the redshift slice chosen , neither does it depend on the details of the background correction . in the following ,
therefore , we shall use only the intermediate - size slice ( @xmath127 ) as our reference sample . the corresponding background correction in this case
is of 28 galaxies h@xmath124 mpc@xmath125 ( at @xmath64 ) . in figure [ morph - dens - comparison ]
we compare our measurement at @xmath128 to results at different redshifts from the literature .
our last density bin is centred on @xmath129 and includes objects up to @xmath130 , i.e. @xmath131 galaxies h@xmath124 mpc@xmath125 .
with respect to other studies centred exclusively on rich clusters ( e.g. dressler 1980 , dressler et al .
1997 , postman et al .
2005 ) , we sample less well the @xmath52 mpc@xmath125 regime , but have a fairly good signal in intermediate - density regions .
this is indeed expected , looking at the density map of figure [ dens_map ] .
overall , our measurement at @xmath128 is steeper than the local estimate ( note that the abscissa scale is linear ) : the fraction of early - type galaxies within this structure at @xmath37 goes from about half of that observed at @xmath132 at low densities ( @xmath133 ) , to @xmath134 of the local value in the highest density regime . in other words , comparing similar environments at the current epoch and at @xmath51 we find a smaller percentage of early - type galaxies , with this difference being more severe in the lowest density regime .
interestingly , benson et al .
( 2001 , cf . their figure 6 ) , find a similar behaviour within their n - body plus semi - analytical simulations . in figure
[ morph - dens - comparison ] , the agreement of our result with other high - redshift measurements ( smith et al .
2004 , postman et al . 2005 ) is very good for high densities , @xmath135 , while we seem to find a steeper relation at low densities .
smith et al .
( 2004 ) in particular , who also use photometric redshifts , seem to have in general a flatter md relation .
this result is also in good agreement , again for densitites larger than @xmath126 , with the parallel md analysis of the whole cosmos field using totally independent estimators of galaxy morphology and local density ( cf .
capak et al .
2007b , figure 6 ) . additionally , the approach adopted here is more sensitive to high densities , nicely extending the capak et al .
measurement for @xmath136 to densities @xmath137 mpc@xmath125 .
conversely , with respect to that measurement our estimated fraction of early - type galaxies tends to be low for densities below @xmath126 , either indicating a poor sampling of low - density environments in our region ( as it is indeed the case , being the sample deliberately centered on a known structure ) , or being related to our limits in estimating accurate local densities in these regimes .
a quite general prediction of semi - analytical models of galaxy formation is that the md relation should show a dependence on luminosity ( e.g. benson et al .
figure [ morph_map_faint ] replicates figure [ morph_map_bright ] but now plotting only galaxies fainter than dressler s limit , i.e. @xmath138 , where the fainter limit corresponds approximately to our apparent magnitude cut ( @xmath38 ) at @xmath63 . comparing the two figures we clearly see that the md relation is much stronger for luminous galaxies than for faint ones . to make this observation quantitative , we re -
compute the md relation for this fainter galaxy population .
however , for the comparison between the bright and faint samples to be meaningful , we need to have a common estimate of local density @xmath10 for the two samples .
what we need is a measurement based on the same galaxy background .
we therefore re - estimated the md relation for the bright @xmath64 sample , but now using the surface density @xmath10 computed using _ all _ @xmath38 galaxies within the usual @xmath127 redshift slice .
the result is shown as a dashed line in figure [ morph - dens - faint ] , with the solid line repeating the original relationship from figure [ morph - dens ] .
the comparison of these two curves shows that a similar md relation for bright galaxies is consistently detected even if galaxies of all luminosities are used in the computation of the background density .
we can therefore compute the md relation of the faint sample with respect to this same background , that is is shown by the bottom dotted line .
its shallower slope indicates how the md relation ( at @xmath0 ) is weakened for galaxies fainter than dressler s limit of @xmath139 .
the simplest interpretation of this plot is that luminous ( massive ) early - type galaxies tend to dominate high - density regions much more than less massive spheroidal objects .
the latter constitute not more than 30% of the whole galaxy population in the highest density bin , about half of the fraction of luminous early - type galaxies at similar density regimes .
it is also interesting to test the sensitivity of the md relation to the level of brightening assumed for dressler s absolute magnitude threshold .
figure [ brightening ] shows the result of selecting samples which are respectively 0.6 ( our standard choice ) , 1.0 and 1.4 magnitudes brighter than dressler s local value . as we have discussed in [ proj_dens ]
, there is some level of ambiguity in choosing how this luminosity cut - off evolves with redshift and our choice of @xmath140 is based on the observed mean brightening of @xmath141 from the vvds ( zucca et al .
figure [ brightening ] shows that in fact even for a brightening of 1 magnitude , the measured relation does not change significantly .
secondly , the plotted curves show that only when considering galaxies brighter than @xmath142 , the fraction of early - type objects at densities @xmath143 mpc@xmath125 equals that measured at @xmath144 for @xmath145 .
the results obtained in the previous section confirm a strong md relation within this large - scale structure at @xmath0 , with a lower fraction of early - type galaxies than at @xmath144 , in agreement with other recent work at high redshift ( smith et al .
2005 , postman et al . 2005 ) . in figure [ acs_zoom ]
we show the acs image of the central region of the main cluster of galaxies , centred on the x - ray emission peak seen in figure [ xray_overlay ] .
a sufficiently precise location can be recognized by comparing the x - ray contours in the two figures ( allowing for small changes in their shape due to differently calculated backgrounds in the two plots ) .
immediately at first sight , one notices a surprising number of disk galaxies dominating the galaxy population in this area .
most of these objects have indeed a photometric redshift that places them within the cluster and their distribution is clearly asymmetric with respect to the cluster x - ray surface - brightness .
rather , they tend to concentrate on the eastern side , in the same region where the steep gradient in the x - ray profile and the negative jump in the temperature map ( [ x - ray ] ) are found , suggesting a shock in the icm . over this area , clearly a high - density region
, the md relation seems to have reversed . the visual impression is confirmed by our quantitative morphological classification , that confirms the high fraction of late - type galaxies in this specific sub - area of the structure , also when considering objects above dressler s absolute - magnitude cut ( about two - thirds are brighter than @xmath146 ) .
this area is not large enough , however , to significantly impact the md relation measured over the entire structure .
the results obtained in this paper can be summarised as follows : * we have identified a large - scale structure at @xmath0 in the cosmos survey area , based on a number of independent probes of the density field .
the surface density distribution of galaxies selected by photometric redshifts , weak - lensing mass reconstruction and x - ray surface - brightness distribution show a remarkable agreement in the overall structure , despite their different selection functions .
* we have produced different mass estimates of the dominant cluster of galaxies , using both weak lensing and x - ray emission . over an aperture of 100 arcsec , the mass from weak
lensing is @xmath147 m@xmath7 , while the x - ray mass to @xmath148 arcsec is @xmath149 .
the similarity of these two values within similar apertures is encouraging , especially considering the non - full equilibrium status of the icm suggested by the evidence for ongoing merging .
we have also estimated the total cluster mass by fitting a nfw profile to the radial shear field , obtaining a best - fitting mass @xmath96 . *
the existence of an ongoing major merger between two sub - clusters is strongly suggested by the temperature map that shows a significant `` ridge '' , with @xmath118 rising to @xmath150 kev in the internal region , with respect to a global measurement of @xmath151 kev .
this scenario is consistent with the observed x - ray surface brightness distribution , where a steeper edge is visible on the eastern side of the cluster .
* we have measured the morphology - density relation at @xmath152 for a sample of more than 1200 galaxies brighter than @xmath153 .
the resulting relation is quite robust with respect to our choice of the photometric - redshift slice including the structure .
it shows a steeper slope with respect to the @xmath144 relation , with about half the fraction of early - type galaxies in low - density environments , but more than 70% of the local value within high - density regions .
this agrees with general predictions of hierarchical clustering models ( e.g. benson et al .
2001 ) , in which the population of early - type galaxies in the highest density peaks is established at proportionally older epochs . however , our measurements for densities smaller than @xmath154 mpc@xmath125 have to be treated cautiously , given the comparable value of the background correction ( see appendix ) . for larger values , however , the good agreement with recent independent high - redshift estimates is remarkable , indicating a reduction in the fraction of early - type galaxies at @xmath120 to @xmath134 of the value seen in the local universe . *
we have explored the general behaviour of the md relation at magnitudes fainter than the standard value ( @xmath138 ) , finding a significantly weaker relationship that corresponds to a reduction by a factor of more than two in the fraction of early - type galaxies in the highest density bin , between the luminous and faint sample .
this trend is generally consistent with predictions from semi - analytic models ( benson et al .
conversely , our measured `` standard '' md relation for luminous galaxies is shown not to be too sensitive to increasing the assumed intrinsic brightening of galaxies up to @xmath155 ( i.e. for a sample brighter than @xmath156 ) . to actually match the fraction of early - type galaxies measured at @xmath132 for a density @xmath143 mpc@xmath125 , one needs to select only very luminous objects with @xmath157 , i.e. @xmath158 magnitudes brighter .
* we detect a clear example of how large - scale environmental processes such as the merger between groups and sub - clusters indeed affects the galaxy population , at least in its apparent morphology .
in fact , slightly offset with respect to the centre of the main cluster , an unusual number of disk galaxies is found , coinciding with the colder side of the possible shock front detected in x - rays .
several of these galaxies are brighter than dressler s limit and thus do display an `` inverted '' md relation over this region . due to the relatively small area of the shock , however , they do not affect in a significant way the global morphology - density trend measured over the whole structure .
the natural interpretation of these combined optical and x - ray observations is that star formation has been switched - on ( or enhanced ) in these galaxies during the large - scale merger between two sub - units , currently mixing together to form the central cluster . to our knowledge , this is the first time that a direct link is observed in a high - redshift cluster between the actual morphology of galaxies ( not only the color or spectroscopic properties ) and a large - scale shock happening in the icm surrounding them .
it is particularly fascinating to see so many clear , bright disks inside a cluster at this redshift .
this implies that disks with a significant reservoir of fresh , unprocessed molecular hydrogen had to be present in these galaxies before the merging event , and have become visible due to the burst of star - formation . a connection between the color or star - formation rate of galaxies and a merging event has been advocated in the past to explain the properties of galaxies in merging clusters . for example , in the core of the shapley supercluster , where galaxies at the interface of two merging clusters are found to have bluer colors ( e.g. bardelli et al . 1998 ) .
similarly , ferrari et al . ( 2005 ) observe an increased fraction of emission - line galaxies in the region separating the two sub - clusters of a3921 , and interpret them as related to the ongoing merger .
even closer to the situation observed here , where we find evidence for a merger between two sub - clusters that seems to have consequences on the galaxy population , is the case discussed in sakai et al .
( 2002 ) , gavazzi et al .
( 2003 ) and cortese et al .
these works discuss the discovery and properties of a conspicuous group infalling into the cluster a1367 at 1800 km s@xmath3 , in which twelve members ( two giant and ten dwarf galaxies ) are simultaneously undergoing a conspiquous burst of star formation .
this is interpreted as having been produced either by tidal interaction among the member galaxies or by ram - pressure by the icm related to the high - speed infall into the cluster . + in the near future , new multi - wavelength observations of the cosmos field will allow us to explore in more detail the nature of the spiral galaxy population in this distant cluster and their interaction with the icm .
in particular , chandra imaging will reveal the details of the shock in the hot gas , while data from irac and mips on board of the spitzer satellite , will explore directly the star formation properties and stellar ages in the spiral galaxies .
we gratefully acknowledge the contributions of the entire cosmos collaboration consisting of more than 70 scientists .
the hst cosmos treasury program was supported through nasa grant hst - go-09822 . the cosmos science meeting in may 2005
was supported in part by the nsf through grant oise-0456439 .
this work is based on observations obtained with xmm
newton , an esa science mission with instruments and contributions directly funded by esa member states and the us ( nasa ) . in italy , the cosmos project is supported by inaf under prin-2005/1.06.08.10 and xmm - cosmos is supported by inaf and miur under grants prin/270/2003 and cofin-03 - 02 - 23 and by asi under grant asi / inaf i/023/05/0 . in germany , the xmm newton project is supported by the bundesministerium fr bildung und forschung / deutsches zentrum fr luft und raumfahrt , the max
planck society , and the heidenhain stiftung . part of this work was supported by the deutsches zentrum fr luft und raumfahrt , dlr project numbers 50 or 0207 and 50 or 0405 .
lg thanks claudio firmani and marco scodeggio for enlightening discussions on galaxy evolution , davide lazzati for invaluable help with idl routines and stefania giodini for help with the mock samples .
andreon , s. , 1998 , apj , 501 , 533 bardelli , s. , zucca , e. , zamorani , g. , et al . , 1998 , mnras , 296 , 599 barnes , j. , 1992 , , 393 , 484 benitez , n. , 2000 , apj , 536 , 571 benson , a. j. , frenk , c. s. , baugh , c.m . , 2001 ,
mnras , 327 , 1041 capak , p. , et al . , 2007 , , submitted ( this volume ) capak , p. , abraham , r.g . , ellis , r.s.,et al . , 2007 , , submitted ( this volume ) cappelluti , n. , hasinger , g. , brusa , m. , et al . , 2007 , , in press ( this volume , also astro - ph/0701196 ) cassata , p. , cimatti , a. , franceschini , a. , et al . , 2005 , mnras , 357 , 903 cassata , p. guzzo , l. , franceschini , a. , et al . , 2007 , , in press ( this volume ) clowe , d. , et al .
2006 , , 451 , 395 , c. , 2003 , apjs , 147 , 1 cooper , m. c. , newman , j. a. , madgwick , d. s. , et al .
2005 , , 634 , 833 cortese , l. , gavazzi , g. , boselli , a. , et al .
, 2006 , , 453 , 847 dressler , a. , 1980 , , 236 , 351 dressler , a. , 1997 , , 490 , 577 ferrari , c. , benoist , c. , maurogordato , s. , cappi , a. & slezak , e. , 2005 , , 430 , 19 finoguenov , a. , reiprich , t. h. , bhringer , h. 2001 , , 368 , 749 finoguenov , a. , guzzo , l. , hasinger , g. , et al .
, 2007 , , in press ( this volume , also astro - ph/0612360 ) gavazzi , g. , cortese , l. , boselli , a. , et al . , 2003 , , 597 , 210 giovanelli , haynes , m. p. , & chincarini , g. l. , , 300 , 77 gunn , j.e . , & gott , j.r.i . , 1972 , , 176 , 1 hamana t. , takada m. & yoshida n. 2004 , , 350 , 893 hasinger , g. , cappelluti , n. , brunner , h. , et al .
, 2007 , , in press ( this volume , also astro - ph/0612311 ) heymans , c. , et al . , 2006 , , 368 , 1323 kaifu , n. , et al .
2000 , pasj , 52 , 1 kaiser n. & squires g. , 1993 , apj , 404 , 441 king l. & schneider p. , 2001
, , 369 , 1 kitzbichler , m.g . & white , s.d.m . , 2006 , mnras , submitted ( astro - ph/0609636 ) kotov , o. , & vikhlinin , a. , 2005 , , 633 , 781 kodama , t. , smail , i. , nakata , f. , okamura , s. , bower , r.g . , 2001 , apj , 562 , l9 koekemoer , a. , et al
. , 2007 , , submitted ( this volume ) larson , r.b . ,
tinsley , b.m , caldwell , c.n . , 1980 , , 237 , 692 leauthaud a. et al . 2007 , , submitted ( this volume)lilly , s. , le fevre , o. , renzini , a. , et al .
2007 , , in press ( this volume , also astro - ph/0612291 ) massey r. et al . 2007b , in press ( astro - ph/0608643 ) massey r. et al .
2007 , , submitted ( this volume ) mobasher , b. , capak , p. , et al . , 2007 , , in press ( this volume , also astro - ph/0612344 ) moore , b. , katz , n. , lake , g. , dressler , a. , oemler , a. , 1996 , nature , 379 , 613 navarro j.f . ,
frenk c.s .
& white s.d.m .
, 1996 , , 462 563 navarro j.f .
, frenk c.s . &
white s.d.m .
, 1997 , , 490 508 postman , m. , franx , m. , cross , n. j. g. , 2005 , , 623 , 721 quilis , v. , moore , b. , bower , r. , 2000 , science , 288 , 1617 rhodes j. , refregier a. , groth e. , 2000 , apj , 536 , 79 rhodes , j. et al . , 2007 , , submitted ( this volume ) sakai , s. , kennicutt , r.c . , van der hulst , j.m . , & moss , c. , 2002 , , 578 , 842 schneider , p. , van waerbeke , l. , jain , b. , & kruse , g. 1998 , , 296 , 873 scoville , n.z . , aussel , h. , brusa , m. , et al .
, 2007 , , in press ( this volume , also astro - ph/0612305 ) scoville , n.z .
, benson , a. , blain , a.w . , et al . , 2007 , , in press ( this volume , also astro - ph/0612306 ) scoville , n.z . ,
aussel , h. , benson , a. , et al . ,
2007 , , in press ( this volume , also astro - ph/0612384 ) smith , g.p . , treu , t. , ellis , r.s . ,
moran , s.m . , dressler , a. , 2005 , apj , 620 , 78 treu , t. , ellis , r.s .
, kneib , j .-
p . , et al .
, 2003 , apj , 591 , 53 taniguchi , y. , scoville , n. , murayama , t. , et al .
, 2007 , , in press ( this volume , also astro - ph/0612295 ) vikhlinin , a. , kravtsov , a. , forman , w. , jones , c. , markevitch , m. , murray , s. s. , & van speybroeck , l. , 2006 , , submitted , astro - ph/0507092 zucca , e. , ilbert , o. , bardelli , s. , et al .
( vvds team ) , 2006 , , 455 , 879
to assess how well we are able to reconstruct local densities using our photometric redshift catalogue , we performed a series of tests using one of the mock cosmos surveys constructed by kitzbichler & white ( 2006 ) from the millennium simulation .
the sample we used was tailored as to reproduce the geometry and selection function of the subaru data used for the present paper , i.e. essentially @xmath38 for the global sample and @xmath159 for the volume - limited sub - sample with @xmath160 $ ] .
we obtained very similar results for both the volume - limited and full magnitude - limited samples .
we show here the results from the latter , where we have better statistics .
we simulate photometric redshift errors by smoothing the true redshifts with a gaussian kernel with @xmath161 , where @xmath162 , i.e. consistent with our estimates from the actual cosmos survey ( mobasher et al .
we are aware that the redshift error distribution is probably not very well described by a gaussian , but based on a number of tests using different @xmath163 , we have seen that this assumption does not affect our conclusions . at the position of each galaxy
, we shall compare the surface density value estimated from this simulated photometric - redshift sample to the corresponding spatial density measured in 3d .
in figure [ app : fig1 ] we plot the normalized differences between our projected density @xmath10 , defined as in [ proj_dens ] , and estimated from the `` photo - z '' sample and the `` true '' density measurement one would obtain with perfect knowledge of galaxy positions in space .
the latter is measured analogously to @xmath10 , but in three - dimensions , i.e. as @xmath164 , and then renormalized to obtain the corresponding 2d quantity , @xmath165 , where @xmath166 is the geometrical factor that accounts for the projection , assumuing that only galaxies within the original 3d sphere are involved .
the left panel uses the measurements without subtracting the mean background , while the right panel shows the complete background - subtracted estimates , as done on the real data . as evident ,
this operation is crucial to reduce the large background offset introduced at small values of the density by the very thick slice used to compensate for the large redshift - space blurring .
the solid line gives the expectation values of the measured statistics , computed as the median value within bins of size @xmath167 , spaced by the same amount along the x axis .
the two dotted lines describe the 68.3% confidence interval around this .
these plots indicate first of all that the subtraction of the background is important to reduce the bias of our estimator : after the subtraction , the systematic error of our measurements is consistent with zero down to local density as small as @xmath168 gal h@xmath124 mpc@xmath125 ( although the probability distribution would still favour over- rather than under - estimates ) .
the scatter is clearly very large at these low densities , however already a galaxy living in an environment characterized by 10 gal h@xmath124 mpc@xmath125 , has a 68.3% probability of being assigned a density between 0 and 30 gal h@xmath124 mpc@xmath125 .
even better , at densities of 50 gal h@xmath124 mpc@xmath125 , the 68.3% error corridor extends between -0.5 and @xmath169 , so the estimated density will range between half this value , 25 gal h@xmath124 mpc@xmath125 , and 1.8 times it , i.e. 90 gal h@xmath124 mpc@xmath125 . the consequence of these errors on our estimate of the md relation ( fig .
[ morph - dens - comparison ] ) , is basically that of making the three lowest - density bins at @xmath170 h@xmath171 mpc@xmath125 strongly correlated , shuffling galaxies among them ( consistent with the nearly flat values found at these regimes ) .
however , above this value the errors are smaller than the bin size used to estimate the md relation . these results confirm and corroborate the indications derived in [ proj_dens ] from the stability of our estimates as a function of the slice thickness . together with the similar analysis in the companion paper by capak et al .
( 2007 ) , they indicate that current cosmos photometric redshifts can be safely used to estimate local projected densities for densities @xmath172 h@xmath124 mpc@xmath125 . albeit limited in their scope , they also provide in general a significantly more optimistic view of the ability of photometric - redshift surveys to study enviromental effects , at least in the regime of densities where @xmath173 h@xmath171 mpc@xmath125 , with respect to the general conclusions of previous works ( e.g. cooper et al . 2005 ) .
|
we have identified a large - scale structure at @xmath0 in the cosmos field , coherently described by the distribution of galaxy photometric redshifts , an acs weak - lensing convergence map and the distribution of extended x - ray sources in a mosaic of xmm observations .
the main peak seen in these maps corresponds to a rich cluster with @xmath1 kev and @xmath2 erg s@xmath3 ( @xmath4 $ ] kev band ) .
we estimate an x - ray mass within @xmath5 corresponding to @xmath6 m@xmath7 and a total lensing mass ( extrapolated by fitting a nfw profile ) @xmath8 .
we use an automated morphological classification of all galaxies brighter than @xmath9 over the structure area to measure the fraction of early - type objects as a function of local projected density @xmath10 , based on photometric redshifts derived from ground - based deep multi - band photometry .
we recover a robust morphology - density relation at this redshift , indicating , for comparable local densities , a smaller fraction of early - type galaxies than today .
interestingly , this difference is less strong at the highest densities and becomes more severe in intermediate environments .
we also find , however , local `` inversions '' of the observed global relation , possibly driven by the large - scale environment .
in particular , we find direct correspondence of a large concentration of disk galaxies to ( the colder side of ) a possible shock region detected in the x - ray temperature map and surface brightness distribution of the dominant cluster .
we interpret this as potential evidence of shock - induced star formation in existing galaxy disks , during the ongoing merger between two sub - clusters .
our analysis reveals the value of combining various measures of the projected mass density to locate distant structures and their potential for understanding the physical processes at work in the transformation of galaxy morphologies .
|
supernova remnants ( snrs ) are believed to be the most promising accelerators of ultrarelativistic particles ( electrons , protons and nuclei ) in our galaxy @xcite .
they are responsible for the majority of cosmic rays ( crs ) with energies up to @xmath6 ( lorentz factor of proton @xmath7 ) .
classical observations of synchrotron radiation from snrs , mainly in the radio band , give experimental confirmation of the presence of relativistic electrons in these cosmic objects @xcite .
recently , the presence of ultrarelativistic electrons in snrs has been confirmed by optical , x - ray and soft @xmath0-ray observations @xcite .
there are expected two main cr accelerators in snr .
one of them is a young pulsar ( if there is any ) , i.e. , a spinning neutron star with a strong magnetic field .
the second one is a strong shock wave as the result of an interaction of supernova ejecta with the interstellar medium ( ism ) .
it is the main object of our investigation .
the theory of cr acceleration at the shock front via the first order fermi mechanism is quite well elaborated @xcite .
its main prediction is a considerably higher efficiency of the proton acceleration in comparison with an electron case , i.e. , up to @xmath8 of kinetic energy of the incoming flow may be transformed into the energy of newly accelerated crs .
experimental evidences of the proton acceleration at shock fronts are limited to the solar system region ( a bow shock at the boundary of the earth s magnetosphere , interplanetary shocks and solar flares ) .
therefore , it is of great importance to find additional evidences of the proton and nucleus component acceleration at the fronts of cosmic shock waves . such a new possibility is offered by the recent observations of considerable @xmath0-ray fluxes in directions on some radio bright snrs @xcite
. among the 32 unidentified galactic plane @xmath0-ray sources detected by the energetic gamma - ray experimental telescope ( egret ) aboard the compton gamma - ray observatory ( cgro ) five sources are coincident with snrs without known pulsars .
four of these snrs ( w28 , w44 , @xmath0 cygni and ic 443 ) exhibit evidences of a strong interaction with interstellar clouds .
therefore , it is believed that the observed @xmath0-ray emission is the result of an interaction of the accelerated by snr shock wave protons with the snr matter itself @xcite or with the matter of the adjacent molecular clouds @xcite . in this paper
we check this hypothesis for the case of well - studied snr ic 443 and show that according to the atypically low initial energy of a supernova explosion such simple explanations do not hold here . in order to explain the @xmath0-ray emission of ic 443 we elaborate a hydrodynamical model of ic 443 as a result of a supernova explosion in the medium with a large scale density gradient , what allows us to explain observational data concerning x - ray surface brightness distribution and obtain the parameters of remnants , especially shock wave characteristics ( section [ sect2_paperv ] ) . using these parameters , we consider in detail the snr shock interaction with a nearby molecular cloud ( section [ sect3_paperv ] ) and
show that the reverse shock enhances the energy and number density of newly accelerated crs ( section [ sect4_paperv ] ) .
then we show that caused by the rayleigh - taylor instability , effective mixing of snr and molecular cloud matter leads to the generation of an observable flux of @xmath0-rays ( section [ sect5_paperv ] ) .
conclusions are given in section [ sect6_paperv ] .
snr ic 443 is well studied at different energy bands from radio to @xmath0-rays @xcite ( figure [ ic433_red ] and [ ic433_xrays ] ) .
it is believed to be the result of a supernova explosion which happened about @xmath9 yr ago near a dense molecular cloud at the distance about @xmath10 kpc from us .
its unusual nonspherical morphology should reflect the inhomogeneity of the surrounding ism , but until now a self - consistent model of ic 443 is absent .
( 397,397 ) ( 0,0 ) + we suppose that the main reason of this is the absence of a detailed numerical modelling of ic 443 .
so , despite the evident influence of inhomogeneity of the surrounding ism , ic 443 models are mainly based on the use of the one - dimensional ( 1d ) self - similar sedov solution @xcite for snr evolved in a uniform medium . the role of the density gradient in the surrounding ism
was taken into account only for the explanation of shape asymmetry within the framework of the approximate kompaneetz @xcite approach @xcite .
therefore , we carry out here a detailed 2d calculation of the ic 443 evolution in a nonuniform medium using the proposed in @xcite new approximate method of a complete hydrodynamical description of a point - like explosion in an arbitrary nonuniform medium .
besides geometrical sizes , we took the total flux and the surface brightness distribution in an x - ray band as free characteristics which give the best possibility for the estimation of plasma characteristics inside the snr .
( 310,310 ) ( 0,0 ) + taking into account the generally accepted distance to ic 443 @xmath11 kpc @xcite , we have reproduced the following picture of the supernova explosion . at @xmath5 yr ago , a sn explosion with the total energy @xmath12 erg took place in the transition zone between a warm ism with the hydrogen number density @xmath13 and an interstellar hi cloud with the number density @xmath14 and the high scale @xmath15 so the initial density distribution in this transition zone was @xmath16 with @xmath17 @xmath3 , @xmath18 @xmath3 , @xmath19 pc , the sn position moved on @xmath20 pc from the cloud centre and the initial hydrogen number density in the point of explosion @xmath2 @xmath21 calculations show the following best fitting parameters of the ic 443 model .
depending on the direction from the explosion centre , the shock radius @xmath22 its velocity @xmath23 the postshock temperature @xmath24 and the initial hydrogen number density at shock front position vary in the ranges : @xmath25 ( pc ) @xmath26 @xmath27 ( km / s ) @xmath28 @xmath29 ( k ) @xmath30 @xmath31(r ) ( @xmath32 .
the total swept up mass of the ism is @xmath33 masses of the sun ( @xmath34 ) , the volume of the disturbed by a shock wave region is @xmath35 @xmath36 , the mean snr radius is @xmath37 pc . in figure
[ ic433_xrays_models ] the x - ray surface brightness of the ic 443 model is presented .
the total equilibrium flux in an x - ray band is calculated for x - ray emissivity data from @xcite and is obtained to be @xmath38 kev)=1.4@xmath39 erg / s ( no absorption correction is made ) which corresponds to the experimental result @xmath40 kev)=1.5@xmath39 erg / s @xcite . in figure
[ ic433_xrays_models](b ) we show the effect of the absorption of interstellar gas with the column density @xmath41 @xmath42 and an elongated cylindrical - like molecular cloud in front of the snr with the radius @xmath43 pc and the cloud column density @xmath44 @xmath45 the data for the absorption cross - section from @xcite were used .
this absorption is mainly responsible for the unusual nonspherical x - ray shape of ic 443 .
( 397,177 ) ( 0,0 ) +
our calculations show that the visible x - ray shape of ic 443 is close to spherical , although with different radii in the south - west and the north - east , while the rosat x - ray image provides a strong evidence of two half - spheres in the ic 443 structure @xcite .
this difference may be naturally explained by the enhanced absorption of x - rays in the adjacent to ic 443 elongated molecular cloud which is visible in a radio band @xcite . the same molecular cloud ( or ,
more correctly , the part of the cloud disturbed by the shock wave of snr ) should be responsible for the observable @xmath0-rays .
observations in @xmath46 and @xmath47 emission radiolines @xcite reveal that the total mass disturbed by the shock gas in the molecular cloud reaches @xmath48 although this molecular cloud covers practically the entire remnant , the interaction region is localized in the southeastern part and coincides with the @xmath0-ray source position ( figure [ ic433_xrays ] ) .
it means that the interaction region is somewhat displaced from the snr symmetry axis ( on the angle @xmath49 , as viewed from the supernova progenitor position ) , i.e. , ic 443 has essentially a 3d shape . from the hydrodynamical point of view ,
the snr molecular cloud interaction is a complicated event accompanied by the arising of two shocks : a forward shock in the cloud matter and a reverse shock in the snr matter .
the boundary between the ism and the molecular cloud becomes a contact discontinuity between the shocked plasmas of the snr and the cloud @xcite .
according to the complexity of the resulting flow we can not calculate it within the framework of the used above method for one shock flow calculations .
but simple analytical estimates are possible .
so , the velocity of the forward shock wave in the molecular cloud @xmath50 is determined by the snr shock velocity @xmath51 at the cloud position and the ratio of hydrogen number densities of the molecular cloud @xmath52 and the ism @xmath53 @xcite : @xmath54 where @xmath55 depends on the number density contrast and @xmath56 for @xmath57 if a snr shock wave penetrates into the molecular cloud at the time @xmath58 yr after the sn explosion , when the shock radius is @xmath59 pc , @xmath60 km / s and , according to the sedov solution @xcite , the effective thickness of the postshock gas shell is @xmath61 the typical time scale of the reverse shock wave existence is @xmath62 the forward shock penetrates into the molecular cloud up to the distance @xmath63 and the shock - excited mass of the cloud is of order @xmath64 where @xmath65 is the molecular cloud density ( the typical number density of helium atoms @xmath66 is taken ) , @xmath67 is the proton mass , @xmath68 is the effective surface of the snr cloud interaction .
we note that the obtained mass of the shocked matter of the molecular cloud is close to the observed one , if we take typical values @xmath69 @xmath3 . to summarise , the proposed here 3d model of ic 443 explains in a consistent way the main characteristics of the remnant , including its x - ray properties and interaction with the adjacent molecular cloud .
we use this model for the estimation of the efficiency of cosmic ray acceleration and for the explanation of @xmath0-ray production in ic 443 .
crs , i.e. , high energy relativistic particles ( electrons , protons , nuclei ) with the power - law energy spectrum @xmath70 and the mean energy density @xmath71 ev/@xmath72 erg/@xmath36 are believed to be accelerated by different mechanisms in our galaxy ( crs with energies @xmath73 ev ) and in powerfull extragalactic sources ( crs with energies @xmath74 ev , up to the maximal observable energies @xmath75 ev ) , such as active galactic nuclei , quasars , radiogalaxies etc . by now
the most promising mechanism for the galactic cr generation is the first order fermi mechanism at shock wave fronts @xcite .
in this mechanism fast particles gain energy during their diffusive motion in the shock region as a result of scattering on different types of magnetic field fluctuations ( magnetohydrodynamical waves , especially @xmath76 waves , turbulent pulsation etc . ) ahead and behind the shock front .
the main reason for effective energy gain is the convergent character of hydrodynamical flows in the shock front frame .
the main characteristics of a shock wave acceleration are the following @xcite .
the acceleration time is @xmath77 where @xmath78 ( @xmath79 for upstream and 2 for downstream ) is the plasma velocity in the shock frame , @xmath80 is the diffusion coefficient , @xmath81 is the mean free path , @xmath82 is the larmor radius of a particle with the energy @xmath83 and the charge @xmath84 in the magnetic field @xmath85 @xmath86 is the light velocity , @xmath87 is the shock velocity in the observer s frame .
parameter @xmath88 is believed to be in the range @xmath89 for snr shocks .
the theoretically predicted spectrum of accelerated particles has the power - law type @xmath90 so @xmath91 for plasma with the adiabatic index @xmath92 both electrons and protons are involved in the acceleration process but the number of accelerated electrons is considerably less than the number of protons @xmath93 further we shall consider only a proton component of snr crs .
there are different estimates of the efficiency of energy transformation from the kinetic energy of a hydrodynamical flow into the energy of accelerated particles @xcite .
the reasonable result is that about 10% of the kinetic energy of a flow may be transformed at the shock front into the energy of crs.the maximal energy of crs is restricted by the processes of energy loss , particle escape from the region of acceleration , finite time of the shock existence etc . in the considered here snr ic 443 case , taking into account the calculated ic 443 shock characteristics , we should expect the following characteristics of cr acceleration .
the total energy of cosmic rays in ic 443 is @xmath94 ( in order to provide the necessary amount of crs in our galaxy due to snrs we need @xmath95 ) .
so , for our case @xmath1 erg we obtain @xmath96 for the spectral index of protons we take @xmath97 which is close to theoretical predictions ( [ spectum_of_cr ] ) and corresponds to the observable @xmath0-ray spectrum ( see below )
. then the energy spectrum of ic 443 crs ( mainly protons ) is @xmath98 the process of the pion creation in subrelativistic proton proton at rest collisions is effective beginning from the proton kinetic energy @xmath99 mev for which the pion decay dominates over other mechanisms of @xmath0-ray production @xcite .
the same proton energy is needed to generate @xmath0-photons with the energy @xmath100 mev .
the total number of such protons in ic 443 is @xmath101 the maximal energy of crs is restricted in the ic 443 case by the snr age @xmath5 yr .
then , from @xmath102 we obtain @xmath103 the mean energy density of crs inside ic 443 with volume @xmath104 is @xmath105 exactly speaking , crs should not be uniformly distributed inside ic 443 .
rather they will be concentrated in a thin shell of thickness @xmath106 @xmath107 near the shock front , where the majority of snr mass and about half of its thermal energy are concentrated @xcite , since the necessary time for diffusion on the distance of the order of a shock radius@xmath108 considerably exceeds the ic 443 age .
it means that at least half of the total energy of crs @xmath109 is concentrated in this shell , so , the cr energy density near the shock front is @xmath110 as we shall see below , this value is too low for the generation of the necessary gamma - flux from ic 443 .
therefore , we propose and consider in detail the promising mechanism of the cr energy density enhancement in ic 443 , connected with the reverse shock action . in the reverse shock the plasma of snr
is subjected to another shock and its final values of the number density @xmath111 and magnetic field @xmath112 are 2.5 times enhanced for @xmath113 plasma @xcite , which results in the increase of cr number density @xmath114 meantime the energy of individual ultrarelativistic particles increases according to adiabatic invariant conservation ( @xmath115 where @xmath116 is the normal to the magnetic field orientation component of particle s momentum ) @xmath117 .
therefore , we expect that behind the reverse shock front the energy density of crs will be @xmath118 for our model we will have @xmath119 ev/@xmath120 .
the maximal energy of crs in the reverse shock region will be of the order of maximal energy in the main shock , i.e. about @xmath121 ev , while acceleration conditions in the both shocks are similar .
the @xmath0-ray flux from source 2eg j0618 + 2234 which coincides with the snr ic 443 position is well approximated by the power - law dependence @xcite @xmath123 in the egret band @xmath124 gev .
for the distance to ic 443 @xmath11 kpc the @xmath0-ray spectrum of the whole snr is @xmath125 the total @xmath0-ray luminosity in @xmath126 band is @xmath127 erg / s and the rate of photon creation is @xmath128 phot / s . in @xcite it was shown that in the case of a pulsar absence , the most promising mechanism for @xmath129 @xmath0-ray production is inelastic interaction of relativistic protons with protons at rest resulting in the creation of pions and their consequent decay into @xmath0-rays .
the @xmath0-ray luminosity in this case is equal to the rate of energy transformation from relativistic protons to neutral pions @xcite @xmath130 where @xmath131 is the mean number density of target nuclei in the region of interaction , @xmath132 mev is the minimal proton kinetic energy of the effective pion creation ( with the cross - section @xmath133 close to the mean value @xmath134 @xmath135 ) , @xmath136 is the mean energy transformed into the pion . substituting in ( [ gamma_lum_ic443 ] ) the cosmic ray spectrum ( [ en_spectrum_cr_ic443 ] )
we obtain : @xmath137 so , as it was mentioned earlier in @xcite , it is necessary to have the initial number density in the ic 443 region @xmath138 @xmath3 and the total energy of crs of @xmath139 ergs in order to explain the observable @xmath0-ray flux in such a simple model .
but , as we showed in section [ sect2_paperv ] in the ic 443 case supernova exploded in a low density region with @xmath140 @xmath3 and the total energy of crs is expected to be only of the order of @xmath141 erg . in @xcite , the @xmath0-flux from ic 443
is explained as the result of the interaction of accelerated by snr crs with the matter of the adjacent molecular cloud .
the role of the molecular cloud consists in the increase of the number density of target nuclei @xmath142 in the interaction region .
namely , we can rewrite ( [ gamma_lum_theor ] ) in the form @xmath143 where @xmath144 is the total mass of the interacting with crs gas , @xmath145 is the proton mass . in @xcite , this total mass is taken to be @xmath146 and the deduced mean cr energy density inside ic 443 is @xmath147 ev/@xmath120 .
but , as we can see from ( [ mean_energy_dens_cr_ic443 ] ) , this value of @xmath148 is considerably higher than the one expected for ic 443 .
another problem here is a high total mass of clouds @xmath149 in the region of interaction , while the cloud mass outside snr is included ( as we mentioned above , only @xmath150 of gas is inside the snr ) .
the energy density of crs outside the snr should be considerably lower than that inside it @xcite , and , therefore , the pion production should be depressed .
we suppose that the observable @xmath0-ray flux from ic 443 is caused by enhancement of the cr energy density in the reverse shock from the snr molecular cloud interaction . indeed , using ( [ energy_dens_cr_ic443 ] ) we obtain from ( [ gamma_lum_theor2 ] ) the following @xmath0-ray luminosity in the case of enhancement of the cr energy density : @xmath151 therefore , we obtain the theoretical @xmath0-ray luminosity @xmath152 erg / s for ic 443 despite the low explosion energy @xmath1 erg and , accordingly , the low total energy of cosmic rays @xmath153 erg . additional support for this model
follows from the existence of an effective mixing mechanism of a hot containing cr plasma and dense shocked plasma of the molecular cloud the well known rayleigh - taylor instability of a contact discontinuity between the both media which should lead to a decay of contact discontinuity and fragmentation of the shocked cloud material onto a set of cloudlets @xcite .
another possibility of the effective cr cloud material interaction takes place in the case of a strongly nonuniform cloud , in which dense molecular cores of mass @xmath154 and number density @xmath155 @xmath3 are embedded in a relatively less dense intercloud medium with the number density @xmath156 @xmath21 as a result of displacement of a contact discontinuity in the molecular cloud direction , such separate cores penetrate into the shocked by a reverse shock plasma region and should be the sites of an intense @xmath0-ray production .
radio observations of the shocked molecular gas in the ic 443 region reveal the existence of few such cloudlets @xcite .
if the association of some unidentified egret sources with snrs is confirmed , it will be a new argument for the cr acceleration in snrs .
it is of great importance that contrary to a radio band , @xmath0-rays give unique information about the acceleration of a proton component of crs in snrs .
in this work we have proposed a self - consistent model of ic 443 which explains a total set of its main parameters , including shape , size , x - ray and @xmath0-ray fluxes .
we have shown that , despite the low supernova explosion energy , account for the snr molecular cloud interaction , especially for the role of the reverse shock wave , is the main reason of an intense @xmath0-ray flux from ic 443 .
further observational and theoretical efforts are needed for checking the theory of crs acceleration by snrs and ic 443 is a prominent candidate for such investigations .
24 berezinskii v. , bulanov s. , dogiel v. et al .
astrophysics of cosmic rays .
north holland , amsterdam , 1990 .
gaisser t. k. cosmic rays and particle physics .
cambridge university press , cambridge , 1990 .
lozinskaya t. supernovae and stellar winds in the interstellar medium .
moscow , nauka , 1986 ( in russian ) [ english translation : aip , new york , 1992 ] .
reynolds s. p. synchrotron models for x - rays from the supernova remnant sn 1006 . //
apj , 1996 , vol .
459 , p. l13-l16 . heavens a. f. upper limits for shock - accelerated particle energies . //
mnras , 1984 , vol . 207 , no 1 , p. 1p-5p .
sturner s. j. , dermer c. d. association of unidentified , low latitude egret sources with supernova remnants . //
a&a , 1995 , vol .
293 , p. l17-l20 .
esposito j. a. , hunter s. d. , kanbach g. , sreekumar p. egret observation of radio - bright supernova remnants . //
apj , 1996 , vol .
461 , p. 820 - 827 .
sturner s. j. , dermer c. d. , mattox j. r. are supernova remnants sources of @xmath157 mev @xmath0 rays ? //
a&asuppl , 1996 , vol . 120 , p. 445 - 448 . green d. a. observations of ic 443 at 151 and 1419 mhz . //
mnras , 1986 , vol .
221 , p. 473 - 482 .
braun r. , strom r. g. the structure and dynamics of evolved supernova remnants .
the ic 443 complex . //
a&a , 1986 , vol .
164 , p. 193 - 207 .
mufson s. l. ,
mccollough m. l. et al . a multiwavelenght investigation of the supernova remnant ic 443 . //
aj , 1986 , vol .
92 , no . 6 , p.
1349 - 1357 .
white g. l. , rainey r. , hayashi s. s. , kaifu n. molecular line observations of ic 443 .
the interaction of a molecular cloud and an interstellar shock . //
a&a , 1987 , vol .
173 , p. 337 - 346 .
petre r. , szymkoviak a. e. , seward f. d. , willingale r. a comprehensive study of the x - ray structure and spectrum of ic 433 . //
apj , 1988 , vol .
335 , p. 215 - 238 .
asaoka i. , aschenbach b. an x - ray study of ic 443 and the discovery of a new supernova remnant by rosat . //
a&a , 1994 , vol .
284 , p. 573 - 582 .
sedov l. similarity and dimensional methods in mechanics .
moscow , nauka , 1977 ( in russian ) [ english translation : academic , new york , 1959 ] .
kompaneets a. //
nauk sssr , 1960 , vol
. 130 , p. 1001
( in russian ) .
hnatyk , b. evolution of supernova remnants in interstellar medium with a large scale density gradient .
// pisma azh , 1988 , vol .
14 , p. 725 - 736
( in russian ) [ english translation : sov . astron .
lett . , 1988 , vol .
14 , p. 309- ] .
hnatyk b. , petruk o. the new approximate analytical method for calculation of a point explosion in the nonuniform medium and its application to modelling x - ray radiation from 3d supernova remnants . //
kinematika i fizika nebesnykh tel , 1996 , vol . 12 , p. 44 - 64
( in ukrainian ) [ english translation : kinematics and physics of celestial bodies , 1996 , vol . 12 , p. 35 - 55 ] .
hnatyk b. , petruk o. evolution of supernova remnants in interstellar medium with a large scale density gradient .
i. general properties of shape evolution and x - ray emission . //
a&a , 1998 , _ in press_. raymond j. , smith b. // apjs , 1977 , vol . 35 , p. 419- .
morrison r. , mccammon d. interstellar photoelectric absorption cross sections , 0.003 - 10 kev . //
apj , 1983 , vol . 270 , p. 119 - 122 . dickman r. l. , snell r. l. , ziurys l.m .
, huang y .- l .
structure and kinematics of dense gas associated with the supernova remnant ic 443 . //
apj , 1992 , vol .
400 , p. 203 - 213 .
landau l. d. , lifshitz e. m. theoretical physics , vol .
moscow , nauka , 1986 ( in russian ) .
merck m. , bertsch d. l. , et al .
study of the spectral characteristics of unidentified galactic egret sources . //
a&asuppl , 1996 , vol .
120 , p. 465 - 469 .
|
we examine the hypothesis that some supernova remnants ( snrs ) may be responsible for some unidentified @xmath0-ray sources detected by egret instrument aboard the compton gamma ray observatory .
if this is the case , @xmath0-rays are produced via pion production and decay from direct inelastic collisions of accelerated by snr shock wave ultrarelativistic protons with target protons of the interstellar medium .
we develop a 3-d hydrodynamical model of snr ic 443 as a possible cosmic @xmath0-ray source 2eg j0618 + 2234 .
the derived parameters of ic 443 : the explosion energy @xmath1 erg , the initial hydrogen number density @xmath2 @xmath3 , the mean radius @xmath4 pc and the age @xmath5 yr result in too low @xmath0-ray flux , mainly because of the low explosion energy .
therefore , we investigate in detail the hydrodynamics of ic 443 interaction with a nearby massive molecular cloud and show that the reverse shock wave considerably increases the cosmic ray density in the interaction region .
meantime , the rayleigh - taylor instability of contact discontinuity between the snr and the cloud provides an effective mixing of the containing cosmic ray plasma and the cloud material .
we show that the resulting @xmath0-ray flux is consistent with the observational data .
supernova remnants , individual : ic 443 , x - rays , cosmic rays , @xmath0-rays 98.38.mz , 98.70.qy , 98.70.rz , 98.70.sa
|
nuclear reactors emit a calculable flux of electron antineutrinos ( @xmath5s ) in all directions . for standard particle propagation ,
one expects a detector located a distance @xmath6 from the reactor to measure a flux that decreases as @xmath7 .
but if @xmath5s are massive , they may `` oscillate '' into undetectable flavors on the way to the detector , leading to an apparent `` disappearance '' of the @xmath5s .
neutrinos are produced and detected in weak interactions , which couple to the weak eigenstates @xmath8 , where @xmath9 . for massive neutrinos
, the weak eigenstates may be expressed as a linear combination of three mass eigenstates @xmath10 , @xmath11 , with mass @xmath12 : @xmath13 @xmath14 is a @xmath15 unitary `` mixing '' matrix and is analogous to the ckm matrix in the quark sector . as a neutrino propagates through vacuum with momentum @xmath16 , the phase of each mass eigenstate will change at different rates according to @xmath17 where @xmath18 . at times
@xmath19 the neutrino will be in a superposition of the weak eigenstates .
the probability of detecting flavor @xmath20 at a distance @xmath21 from the source is then @xmath22 in neutrino oscillation experiments , it is common to simplify to two neutrino flavors , say @xmath23 and @xmath24 , in which case @xmath14 is parameterized by a single mixing angle @xmath25 . in this picture , the probability of a neutrino emitted as @xmath23 to be detected as @xmath23 is then @xmath26 where @xmath27 . here
, @xmath28 is taken to be the mass - squared difference between two `` effective '' mass eigenstates relevant to @xmath29 mixing .
note that @xmath30 is a sinusoidal function of l , hence the name `` neutrino oscillation '' .
the situation becomes slightly more complicated for neutrinos propagating through matter . as first recognized by wolfenstein@xcite , mikheyev , and smirnov@xcite ,
while all three weak states participate in neutral - current interactions with normal matter , only electron neutrinos have additional charge - current interactions with electrons in the material being traversed .
this results in an effective index - of - refraction for @xmath23 different from that for @xmath31 by @xmath32 where @xmath33 is the number - density of electrons in the material , and @xmath34 is the forward scattering amplitude .
@xmath35 changes the phase of the @xmath23-component relative to the other components by @xmath36 over a distance @xmath37 given by @xmath38 where @xmath39 is the fermi constant .
this phenomenon , called the msw effect , alters the two flavor oscillation probability as follows : @xmath40 @xmath41 @xmath42^{- \frac{1}{2 } } \label{emsw3}\ ] ] for neutrinos created in and propagating out of the sun , @xmath43 km @xmath44 , and matter effects are significant .
the result is that @xmath23 disappearance experiments that use the sun as a source , such as homestake@xcite , gallex@xcite , sage@xcite , kamiokande@xcite , super - kamiokande@xcite , and sno@xcite , individually allow values of @xmath45 and @xmath46 over many orders of magnitude .
the overlap of allowed regions from multiple experiments with different thresholds leaves a few small patches , with values around @xmath47 and @xmath48 ev@xmath3 highly favored .
this region is called the `` large mixing angle '' ( lma ) msw solution . prior to kamland , another patch at lower @xmath49 ev@xmath3 , known as the `` low '' msw solution , survived at the 99.73% c.l.@xcite for neutrino experiments with artificial sources on the earth , matter effects are typically much less significant .
this is because for matter with density similar to rock @xmath50 km , larger than the radius of the earth .
thus for these sources the modifications to the oscillation parameters can be viewed as perturbations on the vacuum parameters , and for short enough oscillation lengths can be neglected .
artificial @xmath51 sources for oscillation experiments typically take one of two forms : neutrino beams or nuclear reactors .
these two sources comprise a highly complementary experimental program in neutrino oscillation .
neutrino beams in general require very long baselines for good @xmath46 sensitivity , but the collimation of these sources lessens the severity of this obstacle considerably .
moreover , beams produce higher energy neutrinos and multiple flavors , allowing for not only disappearance but also appearance measurements . reactors , on the other hand , are a non - collimated source , and produce exclusively electron ( anti)neutrinos , restricting the experimenter to disappearance - only measurements and thus giving limited sensitivity to @xmath52 . however , due to the lower energies of the neutrinos ( 1 - 10 mev ) , reactor experiments are sensitive to very small @xmath46 .
note that both neutrino beams and reactors are `` laboratory - style '' experiment , in which both the source and the detector are controlled .
the physics of neutrino propagation can then be cleanly tested separately from the mechanisms involved in the production of the neutrinos .
more than 99.9% of @xmath5s emitted by nuclear reactors are produced in the decay chains of the fission products of only four isotopes : @xmath53u , @xmath54u , @xmath55pu , and @xmath56pu .
the number of fissions of each of these isotopes over the data - taking period is estimated from a monte carlo simulation of the reactor core that tracks fuel burn - up and u / pu production over reactor fuel cycles .
the performance of such simulations has been verified by comparing the simulated fuel composition at the end of a fuel cycle with measurements of isotopic abundances in the actual spent fuel rods .
the inputs to the simulation include the initial fuel composition and periodic measurements of the secondary calorimetric power , the pressure and flow rate in the primary cooling system , and various other operational parameters . however , the results of the simulations depend very weakly on most of the inputs .
accuracy is required only for the fuel composition and the power , on which the fission rates depend linearly .
the power is measured rather precisely : reactors are regulated to operate at about a standard deviation or so below their rated power outputs . the economic incentive to produce as much power as possible pushes this error to be small ; uncertainties @xmath57 are common .
the emitted @xmath5 spectrum is then calculated by summing the spectrum emitted by each isotope weighted by the number of fissions of that isotope during the data taking period . for @xmath53u , @xmath55pu , and @xmath56pu , @xmath5 spectra
are derived from @xmath58-spectra measurements.@xcite@xmath59@xcite extracting the @xmath5 spectra from the @xmath58-spectra is non - trivial ; it involves complex bookkeeping of the decay chains for each daughter nucleus and summing the contributions .
the spectra are normalized to the number of @xmath5s emitted per fission , and have uncertainties of a few percent . for @xmath54u ,
no measurements are available , so we must rely on a calculation@xcite of the @xmath5 spectrum .
the uncertainty in the calculation is @xmath6010% , but since fissions from @xmath54u make up only @xmath6010% of the signal , this results in a 1% uncertainty in the full @xmath5 flux .
reactor experiments typically detect @xmath5s via inverse-@xmath58-decay , @xmath61 , because the 2-particle final state distinguishes the @xmath5 from most other particles and dramatically reduces backgrounds .
the positron carries away almost all of the energy of the incident neutrino , so that the detected positron spectrum is to a good approximation the incident neutrino spectrum folded with the detection cross section and shifted by a constant energy .
the cross section has a threshold of 1.8 mev , and has been calculated@xcite with uncertainties on the level of a percent .
the most recent generation of reactor experiments , palo verde@xcite and chooz@xcite , showed that uncertainties on the level of a few percent can be achieved for these calculations .
the fact that @xmath5 disappearance has not been detected in reactor experiments with baselines up to @xmath601 km@xcite@xmath62@xcite allows one to view these experiments as a verification the reactor flux calculations to within a few percent over most of reactor neutrino spectrum .
kamland uses the entire japanese nuclear power industry as a 180 gw@xmath63 long - baseline source .
the experiment is located underneath mt .
ikenoyama in gifu prefecture in central japan .
80% of the @xmath5 flux comes from reactors at baselines of 140 - 210 km .
this range of baselines makes kamland particularly sensitive to values of @xmath46 corresponding to the lma msw solution to the solar neutrino problem .
matter effects in rock can be neglected at this distance scale .
the results presented here are for 145.1 live - days of data taking between march and september 2002 .
calculating the incident flux at kamland during this period requires summing the flux from every reactor in japan .
reactor data , including instantaneous power and fuel burn - up , is provided by the japanese power companies .
the flux from reactors in korea and the rest of the world is estimated from reported power generation , and makes up only a few percent of the overall flux .
as stated earlier , kamland detects @xmath5s via inverse-@xmath58-decay , @xmath61 , with the energy of the positron @xmath64 related to the energy of the incident @xmath5 by essentially a constant offset of 1.8 mev .
the positron quickly deposits this energy in kamland s liquid scintillator before annihilating with an electron , giving a prompt energy deposit of @xmath65 mev
. meanwhile , the neutron thermalizes and eventually captures on a proton , releasing a 2.2 mev gamma in the process .
the characteristic capture time of @xmath60 210 @xmath66s between the prompt and delayed events provides a powerful background reduction .
kamland s liquid scintillator is composed of 80% dodecane , 20% pseudocumene , and 1.52 g / liter of ppo .
one kton of scintillator is held inside a 6.5 m radius transparent balloon made of a 135 @xmath66 m thick sandwich of nylon and ethylene vinyl alcohol copolymer films .
1879 17- and 20-inch photomultiplier tubes ( pmts ) mounted on the inner surface of an 18 m stainless steel sphere view events inside the scintillator .
for the results reported here , only the 1325 17-inch tubes are used , corresponding to 22% coverage .
a buffer of mineral oil between the pmts and the balloon provide buoyancy for the balloon and shielding from radioactivity in the pmt glass .
outside the steel sphere is a water cherenkov outer detector ( od ) serving as a muon veto .
a chimney and deck at the top of the detector provide access for calibration devices to be deployed along the z - axis .
a schematic of the detector is given in figure [ fig : detector ] .
= 5.0 in an event is recorded whenever 200 or more pmts report a signal within 125 ns . a delayed - event window
is then opened for 1 ms with a lower threshold of 120 tubes .
the pulses are digitized by analog transient waveform digitizers ( atwds ) .
there are two atwds per pmt to reduce dead time during digitization .
the visible energy and position of each event is reconstructed from the collected charge and the hit pattern .
the reconstruction algorithms are developed and tested using uniformly distributed spallation products following muons and using calibration data along the z - axis with @xmath67ge , @xmath68co , and @xmath69zn gamma sources , and an ambe neutron / gamma source .
the calibration data are also used to estimate the conversion from visible energy to kinematic energy , taking into account particle - dependent non - linear effects due to quenching and cherenkov light production .
the systematics on the energy scale is shown in the upper panel of figure [ fig : eandr ] .
= 3.0 in antineutrino events are selected according to the following criteria : time correlation 0.5 @xmath66s @xmath70 660 @xmath66s ; vertex correlation @xmath71 1.6 m ; prompt event energy @xmath72 2.6 mev ; delayed event energy 1.8 mev @xmath73 2.6 mev ; spherical fiducial volume @xmath74 5 m ; cylindrical radius @xmath75 1.2 m. the cut on @xmath76 eliminates @xmath5 backgrounds from geological u / th sources .
the cylindrical radius cut removes backgrounds from thermometers deployed along the z - axis to monitor the scintillator .
the total efficiency of these cuts was determined to be 78.3 @xmath77 1.6% , and the residual accidental backgrounds , estimated using an off - coincidence time correlation window 20 @xmath66s @xmath70 20 s , is negligible at @xmath78 events per day .
the size of the fiducial volume is determined by counting the ratio of uniformly distributed spallation products that pass the fiducial volume cut .
the distribution for spallation neutron capture gammas is given in the bottom panel of figure [ fig : eandr ] .
the correlated backgrounds passing the @xmath5 cuts are produced by cosmogenic spallation .
an overburden of 2700 m.w.e . reduces the muon rate to @xmath60 0.3 hz .
muons are identified by their large energy deposit and outer detector activity , and the muon track is reconstructed from the pmt hit pattern .
a 2 ms veto following muons eliminates backgrounds due to spallation neutrons and short lifetime spallation products .
long lifetime spallation products that mimic the @xmath5 signal , such as @xmath79he and @xmath80li , are removed by a 2 second veto in a 3 m radius around the muon track , plus a 2 second veto over the entire volume following muons with very high energy deposits ( @xmath813 gev ) .
the residual correlated backgrounds , estimated from spallation event studies , time - since - last - muon distributions , and , for fast neutrons following muons not tagged by the od , monte carlo simulations , is 1 @xmath77 1 event in 145.1 days .
figure [ fig : eped ] shows the distribution of the @xmath5 candidates prompt and delayed energies .
the delayed 2.2 mev gamma cleanly separates @xmath5 from the accidental coincidences clustered at small energies .
the one event with @xmath82 5 mev is consistent with neutron capture on carbon ; such events are not considered in this analysis at a small cost in efficiency with negligible uncertainty . = 4.0 in the systematic uncertainties are listed in table [ tab : sys ] . the largest contribution to
the systematics comes from the estimate of the fiducial mass ratio .
the total uncertainty is 6.4% .
.estimated systematic uncertainties ( % ) .
[ cols= " < , < , < , < " , ] for no oscillations , in 145.1 live - days we expect 86.8 @xmath77 5.6 events with 1 @xmath77 1 of these coming from backgrounds . the observed number of events is 54 , giving a ratio @xmath83 0.085 ( stat ) @xmath77 0.041 ( syst ) .
this ratio is plotted vs. baseline for kamland and previous reactor experiments in figure [ fig : reacratios ] .
the shaded region corresponds to the lma msw solution to the solar neutrino problem , with which the reactor experiments are in excellent agreement .
`` standard '' @xmath5 propagation is excluded at @xmath4 99.95% c.l .
= 5.0 in the positron energy spectrum is histogrammed in figure [ fig : spectrum ] .
an unbinned maximum likelihood fit was performed and included terms for changes in the overall rate and shape of the spectrum at different points in @xmath84 parameter space .
the best fit solution is plotted as the shaded histogram , and corresponds to @xmath85 ev@xmath3 and @xmath1 .
the regions in parameter space allowed at 95% c.l .
are drawn in figure [ fig : shapeinc ] .
also shown is the exclusion region in @xmath86 parameter space based on the kamland rate alone .
assuming cpt invariance ( i.e. that @xmath51 and @xmath87 masses are identical ) , all neutrino oscillation solutions to the solar neutrino problem are excluded except lma .
the kamland result divides lma into two regions of higher and lower @xmath46 .
the sensitivity in @xmath52 is rather poor .
values of @xmath46 between the inclusion regions in general predict large distortions in the energy spectrum that were not observed and are hence disfavored .
constant suppression of the non - oscillated spectrum is consistent with the data at 53% c.l .
the inclusion region extends to high @xmath46 where spectral distortions can not be distinguished with current statistics and energy resolution .
an upper limit can be placed on @xmath46 using the non - observation of @xmath5 disappearance by palo verde and chooz .
= 4.0 in = 4.0 in
figure [ fig : futflux ] shows the @xmath5 flux at kamland from march 2002 through may 2003 . in early 2003 , japanese utilities powered down their reactors for inspections and maintenance , resulting in a drop in the non - oscillated flux at kamland by about a factor of two
. such a drastic time variation will provide a more precise estimate of backgrounds , and depending on the `` true '' values of @xmath46 and @xmath52 , may also provide a time - varying spectral shape that will further restrict the inclusion regions in parameter space .
= 5.0 in the expected sensitivity of kamland for 5 years of data taking is shown in figure [ fig : futsens ] for two sets of parameters expected from kamland s first results .
kamland stands to provide exquisite sensitivity in @xmath46 .
however , further restrictions on @xmath52 will probably have to come from future solar neutrino experimental results .
with the latest generation of solar and reactor experiments dawns a new age in precision neutrino oscillation physics .
the next step in this exciting field will be to improve on current measurements and to measure some of the remaining unknown parameters in full three - flavor mixing .
in addition to the three masses @xmath12 , there are 6 free parameters in the mixing matrix @xmath14 .
it is possible to parameterize @xmath14 as follows : @xmath88 @xmath89 @xmath90 and the value of @xmath91 has been measured by the super - kamiokande experiment.@xcite @xmath92 and the overall mass scale will require kinematical and neutrino - less double-@xmath58 decay measurements .
kamland is playing a central role in determining @xmath93 as well as the value of @xmath94 .
reactor experiments stand to play another role in neutrino oscillation , this time in the measurement of @xmath95 . because @xmath96 , it must be the case that @xmath97 .
a new generation of reactor experiments has been proposed to search for @xmath5 disappearance at baselines of 1 km corresponding to this value of @xmath46 . to improve on the mixing angle sensitivity achieved by palo verde and chooz , proposals for reactor @xmath98 experiments
include a large detector to reduce the statistical error , and also a second detector positioned very close ( @xmath60100 m ) to the reactor .
the near detector would precisely measure the incident flux , allowing many of the flux calculation systematics to drop out .
this also requires that the detectors be made identical and/or moveable .
sensitivity down to @xmath99 seems within grasp .
such experiments were first discussed by mikaelyan and sinev @xcite ; for a comprehensive list of references on this topic , please see the web site compiled by heeger.@xcite
kamland has observed , for the first time , disappearance of electron antineutrinos in a laboratory - style experiment . assuming cpt invariance , this result excludes solar neutrino oscillation solutions except lma at @xmath4 99.95% c.l .
recall that the lma solution to the solar neutrino problem is a region in oscillation parameter space allowed from the overlap of many experimental results using different techniques and with different thresholds .
moreover , lma is an msw solution for neutrinos , in which matter effects inside the sun drive the oscillations .
it is significant that kamland , which uses yet another detection technique and is sensitive to _ vacuum _ oscillations of _ anti_neutrinos , gives oscillation parameters in agreement with lma .
these experiments are different in so many aspects , yet the physics of neutrino oscillation ties them together in a beautifully consistent theoretical framework .
the kamland experiment is supported by the center of excellence program of the japanese ministry of education , culture , sports , science and technology , and the united states department of energy .
the reactor data are provided courtesy of the following electric associations in japan : hokkaido , tohoku , tokyo , hokuriku , chubu , kansai , chugoku , shikoku , and kyushu electric power companies , japan atomic power co. , and japan nuclear cycle development institute .
kamioka mining and smelting company provided services for activities in the mine .
10 k. eguchi _ et al .
lett . * 90 * , 021802 ( 2003 ) .
j.n . bahcall and r. davis ,
science * 191 * , 264 ( 1976 ) ; j.n .
bahcall , _ neutrino astrophysics _
( cambridge university press , cambridge , uk , 1989 ) ; j.n .
bahcall , astrophys .
j * 467 * , 475 ( 1996 ) .
l. wolfenstein , phys .
d * 17 * , 2369 ( 1978 ) . s.p .
mikheev and a.yu .
smirnov , sov .
. phys . * 42 * , 913 ( 1985 ) .
cleveland _ et al .
_ , astrophys .
j * 496 * , 505 ( 1998 ) .
gallex collaboration , phys .
b * 447 * , 127 ( 1999 ) .
abdurashitov _ et al .
c * 60 * , 055801 ( 1999 ) .
y. fukuda _ et al .
lett . * 77 * , 1683 ( 1996 ) .
super - kamiokande collaboration , phys .
b * 539 * , 179 ( 2002 ) .
ahmad _ et al .
lett . * 89 * , 011302 ( 2002 ) .
hahn _ et al .
b * 218 * , 365 ( 1989 ) .
k. schreckenbach _ et al .
b * 160 * , 325 ( 1985 ) . p. vogel _ et al .
c * 24 * , 1543 ( 1981 ) .
p. vogel and j. f. beacom , phys .
d * 60 * , 053003 ( 1999 ) ; radiative correction from a. kurylov , _ et al .
_ , hep - ph/0211306
. f. boehm _ et al .
d * 64 * , 112001 ( 2001 ) .
m. apollonio _ et al .
_ , epjc * 27 * , 331 ( 2003 ) . h. kwon _ et al .
d * 24 * , 1097 ( 1981 ) .
g. zacek _ et al .
d * 34 * , 2621 ( 1986 ) .
vidyakin _ et al .
_ , jetp * 59 * , 390 ( 1994 ) . b. achkar _ et al .
b * 434 * , 503 ( 1995 ) ; b. achkar _ et al .
b * 374 * , 243 ( 1996 ) .
super - kamiokande collaboration , phys .
b * 467 * , 185 ( 1999 ) .
mikaelyan and v.v .
sinev , phys . atom
* 63 * , 1002 ( 2001 ) .
|
the kamioka liquid - scintillator anti - neutrino detector ( kamland ) has detected for the first time the disappearance of electron antineutrinos from a terrestrial source at the 99.95% c.l.@xcite interpreted in terms of neutrino oscillations @xmath0 , the best fit to the kamland data gives a mixing angle @xmath1 and a mass - squared difference @xmath2 ev@xmath3 , in excellent agreement with the large mixing angle ( lma ) solution to the `` solar neutrino problem''.@xcite assuming cpt invariance , this result excludes other solutions to the solar neutrino problem at @xmath4 99.95% c.l . epsf
|
the discovery of an exotic @xmath0 pentaquark by leps at spring-8,@xcite subsequently confirmed by many other groups , marked a new era for testing our understanding of the hadron spectroscopy and promoted a re - examination of the qcd implications for exotic hadrons .
the mass of the @xmath0 is of order 1535 mev and its width is less than 10 mev from direct observations and can be as narrow as 1 mev or even lower.@xcite many null results for the pentaquark search mostly from high energy experiments have also been reported .
therefore , if the @xmath0 pentaquark is real , it must be established beyond any doubt . if the @xmath0 pentaquark exists , its minimum quark content is @xmath1 . to understand those experimental measurements with positive results
, we are facing three puzzles : ( i ) the doubly charged partner of @xmath0 , namely , @xmath2 with the quark content @xmath3 , should be easily detected via the decay @xmath4 .
the puzzle is why is it not seen so far while stringent limits have been set ?
( ii ) the naive constituent quark model in which quarks are uncorrelated implies a @xmath0 pentaquark mass of order 1900 mev .
why is it so anomalously light ?
( iii ) the fall - apart strong decay @xmath5 is ozi super - allowed .
hence , if quarks in @xmath0 are uncorrelated , the width will be of order several hundred mev .
why is the observed @xmath0 width so narrow ?
these puzzles hint at a possible correlation among various quarks ; two or three quarks could form a cluster .
a popular correlated quark model has been advocated by jaffe and wilczek @xcite in which the @xmath0 is a bound state of an @xmath6 quark with two @xmath7 diquarks .
the diquark is a highly correlated spin - zero object and is in a flavor anti - triplet and color anti - triplet state .
bose statistics of the scalar diquarks requires that the diquark pairs be in an orbital @xmath8 wave state .
the parity of @xmath0 is predicted to be positive owing to the diquark correlation .
consider the strong decay @xmath9 .
it can proceed through two processes : fall - apart decay with quark annihilation and the kaon emission process .
the latter is expected to be severely suppressed ( except in the infinite momentum frame where quark annihilation and annihilation are prohibited ) as it involves the transition between @xmath0 and the 5-quark component of the nucleon .
the fall - apart process is quite unique to the pentaquark and it can not occur in the ordinary baryon decays . as mentioned before , if quarks are uncorrelated , the fall - apart mechanism will yield a width of several hundred mev .
therefore , quark correlation is needed to suppress the ozi supper - allowed process . in the jaffe - wilczek model , this suppression can be understood as follows .
first , one has to break the @xmath10 diquark pair so that one of the light quarks will combine with the @xmath6 quark to form a kaon .
second , one has to overcome the @xmath8-wave potential barrier so that the rest three quarks form a @xmath11-wave nucleon .
hence , the @xmath0 decay width is narrow due to the breaking of diquarks and the transformation of the @xmath12-wave into @xmath11-wave configurations . in a typical pentaquark decay to a meson and a baryon ,
the anti - quark is common to both pentaquark and the final state meson . to the leading order of the spectator approximation ,
the anti - quark can be considered as a spectator in the decay process depicted in fig . 1 .
in this picture , there is a @xmath13 subprocess with @xmath14 being a diquark pair and @xmath15 a baryon .
we use the effective hamiltonian @xcite [ eq : heff ] h_eff&= & _ 5 q^c + i^_5(q^c ) _ to model the @xmath16 subprocess , where @xmath17 is a characteristic scale of the system .
it turns out that only the @xmath18 term contributes to even - parity pentaquark decays .
although we do not know how to calculate @xmath19 from first - principles calculations , we can determine it from the @xmath20 width , and use it to predict the decay widths of other light and heavy pentaquarks .
1.2 in for the strong decay @xmath21 , we have calculated the matrix element @xmath22 .
the matrix elements can be expressed in terms of form factors .
we first computed the form factors in the spacelike region and then extrapolated them to the timelike region .
taking 1 mev as a benchmark for the @xmath20 width , it is found that @xmath23 .
we have estimated the strong decays @xmath24 , @xmath25 , @xmath26 and @xmath27 by normalizing to the @xmath0 width .
we found that @xmath24 , @xmath26 and @xmath28decay rates are of the order of a few mev , while @xmath29 decay rates are of order tens of mev.@xcite if we take @xmath30 mev as observed by na49,@xcite we will have @xmath31 which is consistent with the observed width of @xmath32 mev.@xcite
if the light @xmath0 is real , it is expected the existence of the stable heavy pentaquark @xmath33 by replacing @xmath6 by @xmath34 .
whether the heavy pentaquark lies above or below the strong decay threshold is rather controversial . a narrow resonance in @xmath35 mass distribution with mass @xmath36 mev and width
@xmath37 has been reported by h1,@xcite but it has not been confirmed by many other groups . if the heavy pentaquarks lie below the strong decay threshold , they can be searched for only through weak or electromagentic decays .
we apply the relativistic light - front approach to calculate the pentaquark to pentaquark transition form factors.@xcite in the heavy quark limit , heavy - to - heavy pentaquark transition form factors can be expressed in terms of three isgur - wise functions : two of them are found to be normalized to unity at zero recoil , while the third one is equal to 1/2 at the maximum momentum transfer , in accordance with the prediction of the large-@xmath38 approach @xcite or the quark model.@xcite therefore , the light - front model calculations are consistent with the requirement of heavy quark symmetry .
we have calculated form factors and isgur - wise functions in @xcite .
decay rates of the weak decays @xmath39 , @xmath40 , @xmath41 and @xmath42 with @xmath33 , @xmath43 and @xmath44 being the heavy anti - sextet , heavy triplet and light octet pentaquarks , respectively , are obtained . for weakly decaying @xmath45 and @xmath46 , the branching ratios of @xmath47 , @xmath48 are estimated to be at the level of @xmath49 and a few percents , respectively .
i am grateful to chun - khiang chua and chien - wen hwang for fruitful collaboration .
|
strong and weak decays of pentaquarks are studied in the framework of the light - front approach .
|
understanding the non equilibrium statistical properties of driven particles in disordered media is a challenging question relevant to many experimental situations .
a prominent example are the moving phases of driven vortex lattices in superconductors @xcite .
a key feature of these systems is that the disorder induces anisotropic response and fluctuations which are strongly controlled by the velocity @xcite . in spite of its relevance to understand situations of incoherent or plastic vortex flow
the simple case of an isolated vortex driven in a @xmath0-dimensional random potential with @xmath1 has been tackled analitically only by perturbation theory @xcite , valid at high velocities , or by mean field theory @xcite , valid for @xmath2 . in this paper
we propose a simple model whose long time behaviour can be computed analitically at @xmath3 for any finite velocity .
let us consider the equation of motion , in two dimensions , of a driven isolated vortex at zero temperature , @xmath4 where @xmath5 is the instantaneous velocity of the vortex located at @xmath6 , @xmath7 is the driving force and @xmath8 the friction coefficient .
we model the disorder as a random arrangement of hard disks with center @xmath9 and radius @xmath10 . outside the disks the vortex has a free motion and inside it feels a pinning force , @xmath11 where @xmath12 is the amplitude , and @xmath13 the step function .
this disorder models a diluted distribution of pinning centers separated at a distance @xmath14 . in the following we use
adimensionalized variables : @xmath10 is the length unit , @xmath12 the energy unit , and @xmath15 the time unity . above the depinning transition
the motion of the vortex consists in straight segments of free motion interrupted by the collisions with the different pinning centers . at each collision
the vortex is delayed and deflected with respect to the free motion .
the equation describing the motion of the vortex inside the trap centered at @xmath16 is , @xmath17 the collision starts with the vortex at some initial position @xmath18 on the border of the trap , @xmath19 .
the solution of eq.([eq : eqcol ] ) is @xmath20 after a time interval @xmath21 the vortex will exit from the trap , therefore @xmath22 . using this condition in equation [ eq : soltrap ]
we obtain the following expression for @xmath21 , @xmath23 ^ 2- ( f^2 - 1)({\bf f}_p^0)^2}}{({\bf f}_p^0)^2 } \label{eq : expdt}\end{aligned}\ ] ] where @xmath24 .
the displacement @xmath25 induced by the collision is then given by , @xmath26 due to the random distribution of pinning centers , the motion can be considered as a random walk in the long time limit .
the fluctuations of @xmath27 are thus induced by the uncorrelated sequence of collisions .
assuming identical pinning centers , the randomness comes exclusively from the random initial conditions of each collision , described by eq.([eq : soltrap ] ) . in a long time interval
@xmath28 the vortex collides with a large number @xmath29 of pinning centers . since the core covers an area @xmath30 , where @xmath31 is the mean velocity , we get , @xmath32 by symmetry , the long time displacement is along the direction of @xmath33 , @xmath34 , and @xmath35 where @xmath36 denotes an average over a random distribution of initial conditions @xmath37 in equations [ eq : expdt ] and [ eq : dxi ] . from eq .
( [ eq : dxi ] ) we get the mean velocity , @xmath38 } \label{eq : v}\end{aligned}\ ] ] we can now define longitudinal @xmath39 and transverse @xmath40 diffusion constants , @xmath41 ^ 2 \rangle /t \\
d_{\perp } & \equiv & \langle \delta r_{\perp}^2(t ) \rangle /t
= \frac{n_c(t)}{t}\langle \delta r_{\perp}^2 \rangle \label{eq : difutrans}\end{aligned}\ ] ] to calculate the longitudinal diffusion constant in terms of the single collision displacement we use that @xmath42 is the random longitudinal displacement with respect to the average longitudinal motion in a single collision , with @xmath43 the longitudinal mean distance between pinning centers of two given consecutive collisions .
we thus get , @xmath44 ^ 2 \rangle \label{eq : difulong}\ ] ] we also define the longitudinal mobility as @xmath45 $ ] . in terms of single collision quantities
we get , @xmath46 to define the transverse mobility we need to introduce a small perturbative force @xmath47 .
the velocity induced by this force is , @xmath48 + f_{\perp } \label{eq : vtrans}\end{aligned}\ ] ] and thus we can define the transverse mobility @xmath49_{f_{\perp}\rightarrow 0}$ ] . in terms of single collision quantities
we get , @xmath50 finally we can define effective temperatures using generalized einstein relations , @xmath51 in order to calculate the transport properties defined above ( @xmath31 , @xmath40 , @xmath39 , @xmath52 , @xmath53 , @xmath54 , and @xmath55 ) we need to calculate the first moments of the distribution of @xmath21 , @xmath56 , @xmath25 , @xmath57 , @xmath58 , @xmath59 by performing simple integrals .
in fig . [ fig : vf](a ) we show the velocity - force characteristics ( vf ) of our model , calculated from eq.([eq : v ] ) .
the critical depinning force is @xmath60 . at zero temperature , for @xmath61
, the particle is trapped after a transient , and thus @xmath62 . at low velocities , @xmath63 ,
the vf curve is strongly nonlinear with @xmath64^{-1}$ ] @xcite and at large velocities free flux flow is approached with corrections that scale as @xmath65 , as shown in fig.[fig : vf](b ) . ) and transverse ( @xmath66 ) directions.(b ) mobility in the transverse ( @xmath66 ) and longitudinal ( @xmath67 ) directions .
lines show the small and large velocity asymptotic forms in both figures .
, height=302 ] in fig .
[ fig : difyres](a ) we show the longitudinal and transverse diffusion constants . @xmath39 and @xmath40 are both non monotonous functions of the velocity @xmath31 . at small velocity diffusion constants
grow linearly with @xmath31 while at large velocity @xmath68 and @xmath69 .
let us note that @xmath39 and @xmath40 cross at a characteristic velocity @xmath70 .
this crossing means that the long time diffusion front changes aspect ratio at @xmath70 . for @xmath71 the diffusion front
is elongated in the driven direction , while for @xmath71 is elongated in the transverse direction . at @xmath70 diffusion is isotropic .
interestingly , the same behavior is observed in numerical simulations of interacting vortices in two dimensions @xcite . in fig . [
fig : difyres](b ) we show the longitudinal and transverse mobilities .
we observe that both are velocity dependent and approach the free flux response , @xmath72 , at large velocities .
since the @xmath53 is the differential resistance , @xmath73 , the divergence observed at small velocity is a signature of the depinning transition .
@xmath53 near depinning is high since any small force can reduce strongly the waiting time @xmath74 inside the trap . on the contrary , @xmath75 at small velocity .
a small transverse force has a small effect in the trap time @xmath21 compared with the linear @xmath31 dependence ( eq.([eq : nrocollisions ] ) ) of the number of collisions per unit time . in fig .
[ fig : teff ] we show the effective temperatures in the longitudinal and transverse directions .
since at large velocities the mobilities saturate , the velocity dependence of the effective temperatures are dominated by the diffusion constants , and thus @xmath76 and @xmath77 . at low velocities the effective temperatures are instead strongly determined by the velocity dependence of the mobilities .
@xmath78 reaches a maximum at a intermediate velocity and decreases quickly to zero with decreasing @xmath31 . on the other hand
, @xmath79 saturates at a finite value as the velocity vanishes , due to the linear dependence with velocity of both the diffusion constant and the mobility .
it is worth noting that only at large velocities and in the transverse direction , the effective temperature is found to be identical , except from numerical factors , to the shaking temperature @xcite .
this confirms that the shaking temperature is equivalent to an effective temperature defined from a generalized fluctuation relation only in the limit of large velocity of non interacting vortices @xcite .
it is important to point out here that the thermodynamic nature of the effective temperatures of this model is still unclear , since the system is non interacting and strongly driven @xcite .
we have studied the pinning induced anisotropic diffusion of driven non interacting vortices .
we find that the diffusion front is elongated in the direction of the driving force at low velocities and in the transverse direction at large velocities .
this implies the existence of isotropic diffusion at a characteristic velocity @xmath70 .
the analysis of the anisotropic low frequency voltage noise in superconductors could be a possible experimental probe of this result , since diffusion constants and velocity fluctuations are related by generalized green - kubo relations @xcite . even if the depinning transition does depend on the peculiarities of our model , we find that the main features of the long time fluctuations we obtain at intermediate and large velocities are in agreement with perturbation theory predictions @xcite and with numerical simulations of non interacting vortices for the non simplified model @xcite .
furthermore , it is easy to show that our model can be solved at any dimension @xmath1 and generalized to more complicated short - range random potentials @xcite .
00 le doussal , p. , and giamarchi , t. , _ phys .
b _ * 57 * , 11356 ( 1998 ) ; balents , l. , marchetti , m. c. , and radzihovsky , l. , _ phys . rev . b _ * 57 * , 7705 ( 1998 ) ; scheidl , s. , and vinokur , v. m. , _ phys . rev .
e _ * 57 * , 2574 ( 1998 ) ; _ phys .
b _ * 57 * , 13800 ( 1998 ) .
|
we use a simple model to study the long time fluctuations induced by random pinning on the motion of driven non interacting vortices . we find that vortex motion seen from the co moving frame is diffusive and anisotropic , with velocity dependent diffusion constants .
longitudinal and transverse diffusion constants cross at a characteristic velocity where diffusion is isotropic .
the diffusion front is elongated in the direction of the drive at low velocities and elongated in the transverse direction at large velocities .
we find that the mobility in the driven direction is always larger than the transverse mobility , and becomes isotropic only in the large velocity limit .
random pinning , diffusion , mobility 74.25.qt
|
an economic description of pairing correlations in finite nuclei is provided by hartree - fock - bogoliubov ( hfb ) theory @xcite making use of phenomenological interactions like e.g. the finite range gogny force @xcite or density dependent zero - range forces combined with appropriate energy cut - offs ( cf .
, e.g. , @xcite ) .
such a description leads to values of the pairing gap which are in overall agreement with the experimental findings .
note that in these calculations , the density of levels at the fermi energy @xmath22 is controlled by the so called @xmath23mass ( i.e. @xmath24 ) which , as a rule , is smaller than the bare nucleon mass ( e.g. @xmath25 in the case of the gogny force )
. on the other hand , a number of studies have shown that the superfluid properties of nuclear systems , ranging from infinite nuclear and neutron matter to finite atomic nuclei , are strongly influenced by polarization phenomena @xcite . in these calculations
one starts from a bare nucleon - nucleon potential ( argonne , bonn , paris , etc . ) adding afterwards the renormalization processes .
recently , the pairing gap , the quasiparticle spectrum and the collective modes of @xmath1sn have been calculated solving the dyson - gorkov equation @xcite , in a single - particle space characterized by @xmath26 , allowing the nucleons to interact in the @xmath27 channel through a @xmath28 argonne
n - n potential taking into account the variety of renormalization processes ( self - energy , fragmentation , induced interaction and vertex corrections ) arising from the coupling of the particles with surface vibrations . while the bare n - n interaction accounts for about half of the pairing gap , overall agreement with the experimental findings is achieved by including medium polarization effects . in the present paper we want to shed light into the physics of these results , by studying the magnitude of the different pairing ( @xmath29 ) matrix elements as well as their dependence on the energy of the single - particle states lying in the vicinity of the fermi energy in section 2 , and by discussing the calculation of the pairing gaps associated with these matrix elements in section 3 .
the matrix element of the induced interaction can be written as @xcite @xmath30 where @xmath31 , @xmath32 and @xmath33 denote the quantum numbers of the exchanged collective vibrations , @xmath34 is the associated radial formfactor ( @xmath35 is the deformation parameter associated with the @xmath36th mode of multipolarity @xmath31 , @xmath37 is the ground - state radius and @xmath38 is the average potential ) , @xmath39 is the vibrational energy of the @xmath36th mode , while @xmath40 and @xmath3 are the single - particle and fermi energies .
@xmath41 is the pairing correlation energy per cooper pair , which is of the order of @xmath42 . in practice
we have used @xmath41= - 2 mev .
the pre - factor 2 in eq .
( [ vind ] ) comes from the two possible time orderings associated with the one - phonon exchange . in fig .
[ fig_qm ] we show the value of the diagonal matrix elements @xmath43 of the induced interaction for the nucleus @xmath1sn as a function of the single - particle energy .
the single - particle levels entering eq .
( [ vind ] ) have been calculated making use of a standard saxon - woods potential , whose parameters are provided in the caption of fig .
[ fig_qm ] . in most of our calculations we use this potential and an associated effective mass @xmath44 equal to the bare mass . in the last section
we shall use instead an effective mass @xmath45 , which simulates the typical outcome of a hartree - fock calculation ( with , e.g. , the gogny interaction ) . in those cases ,
the woods - saxon parameters are changed in such a way that the fermi energy remains close to the experimental value ( @xmath46 @xmath47 @xcite , see e.g. fig .
[ fig_m07 ] ) .
the parameters @xmath35 and @xmath39 have been determined by diagonalizing a separable multipole - multipole interaction in the quasiparticle - random phase approximation ( qrpa ) adjusting the coupling constants so as to reproduce the energies and transition probabilities of the lowest - lying states of each spin and parity .
also shown in fig .
1 are the diagonal matrix elements of the gogny interaction . here , and in the rest of the paper , we shall employ the d1 parametrization .
the matrix elements of @xmath9 , which display an average value equal to -0.15 mev , are peaked at the fermi energy , within an energy range of essentially 5 mev around @xmath3 .
this behaviour reflects the fact that @xmath48 is controlled , through the energy denominator appearing in eq.([vind ] ) , by the energy of the low - lying collective vibrations ( @xmath49 5 mev ) . outside this energy range ,
the values of @xmath48 become very small , less than 50 kev in absolute value .
the matrix elements of @xmath2 decrease on average smoothly in magnitude all the way from the deepest levels up to the continuum threshold , displaying a value of about -0.3 mev at @xmath3 .
the induced interaction matrix elements depend on specific properties of finite nuclei , namely the single - particle quantum numbers and energies , and the energy , the zero - point amplitude and the transition density of the vibrational states .
to study the global features of the bare , of the induced , and of the gogny interactions as well as to provide a more transparent form for their matrix elements , we shall present , in the rest of this section , also the matrix elements obtained using the semiclassical thomas - fermi ( tf ) approximation . in fact , this approximation averages out shell effects typical of finite nuclei and provides the overall energy dependence of the matrix elements . within this context
we follow ref .
@xcite where semiclassical expressions for two - body matrix elements have been presented . to obtain a smooth behaviour of the diagonal matrix elements as a function of the continuous energy variable @xmath50
, we compute the fourier transform of a two - body interaction @xmath51 using plane waves of momenta @xmath52 and @xmath53 .
the average of the quantal matrix elements is carried out using the normalized semiclassical density matrix , @xmath54 in the above relation the semiclassical level density @xmath55 is @xmath56 @xmath57 being the classical turning point .
we assume a constant value for the effective mass @xmath44 , neglecting its possible dependence on position .
the expression for the semiclassical diagonal matrix element of @xmath58 as a function of the single - particle energy , is consequently @xcite @xmath59 for an interaction that depends only on the magnitude of the relative coordinate @xmath60 , like the gogny interaction in the s=0 , t=1 channel , one can integrate first over @xmath61 and then over the momenta @xmath62 and @xmath63 , obtaining the expression @xmath64 where @xmath65 indicates the center of mass , @xmath66 is the on - shell pairing matrix element of the interaction @xmath67 in the @xmath68 channel and in a plane wave basis , evaluated at the local momentum @xmath69 , and @xmath70 denotes the quantity @xmath71 in the next section we shall calculate the pairing gap associated with the different interactions , and therefore we shall need the matrix elements between all pairs of particles moving in time reversal . for this purpose
we shall employ a formula analogous to that given in eq .
( [ simple ] ) , namely @xmath72 @xmath73 in this relation @xmath50 and @xmath74 indicate the energies of the single - particle states of each pair .
we now turn to the case of the induced interaction , associated with the matrix elements of eq .
in this case the matrix elements depend separately on @xmath75 and @xmath76 .
we then integrate over the momenta first , obtaining @xmath77 where @xmath78 is the bessel function and @xmath79 . the index @xmath33 labels the different phonons of multipolarity @xmath31 ( the phonons included in the present calculation are specified at the end of this section ) .
note that one can obtain numerical results from eq .
( [ vsemi ] ) by avoiding the six - dimensional integration .
this is achieved by performing a multipole expansion of the quantity @xmath80 which leads to @xmath81 in ref .
@xcite the reliability of the tf approximation in reproducing the quantum mechanical ( qm ) matrix elements has been checked in the particular case of the pairing matrix elements of a @xmath82-interaction acting among particles moving in a harmonic oscillator ( ho ) potential .
note that the tf approach implies an averaging over the quantum numbers associated with the different states belonging to the major shells and displaying an energy @xmath83 .
the tf matrix elements have been compared with the averaged qm matrix elements ( taking degeneracies into account ) .
overall agreement was observed ( cf . table ii of ref .
@xcite ) .
a similar comparison for the pairing matrix elements of the gogny force in the @xmath68 channel is reported in fig .
[ fig_semicl ] .
it is seen that the tf approach provides a good approximation to the average quantum matrix elements
. however , the qm diagonal matrix elements , in particular those associated with @xmath84-states , are systematically larger ( in absolute value ) than the non - diagonal ones .
one should be aware of the fact that the semiclassical density matrix can be considered to be the analogue to the one obtained from a strutinsky smoothing .
therefore , eq . ( [ dos ] ) , implicitly , represents a function of about 1@xmath85 width .
thus , the matrix element in eq .
( [ veve ] ) contains at the quantum level cross terms corresponding to at least one major shell .
taking for example the 2s-1d shell , we have to weigh the two diagonal matrix elements with the 2s and 1d wavefunctions with a factor 1 and 25 , respectively .
the non - diagonal 2s-1d element obtains the weight factor 10 . performing the arithmetic average yields the crosses of fig .
[ fig_semicl ] . in this way , we see that the tf expression ( [ veve ] ) reproduces very well the quantal average .
the more realistic case of the gogny matrix elements in the woods - saxon potential ( including spin - orbit ) is illustrated in fig .
[ fig_ws ] .
the general pattern is similar to that shown in the previous figure .
having assessed its validity , we now employ the semiclassical approximation to compare the matrix elements of the induced interaction , the effective gogny interaction and a bare nucleon - nucleon interaction . concerning the latter , the matrix elements of bare nucleon - nucleon interactions which contain a repulsive core display a qualitatively different momentum dependence than that displayed by the gogny interaction .
we shall instead consider here the @xmath0 interaction , which has been devised to reproduce the low - momentum properties of the nucleon - nucleon force , including the experimental phase shifts , without the introduction of a repulsive core @xcite .
in particular , we shall employ a parametrization of @xmath0 devised to reproduce the properties of @xmath28 argonne potential .
this parametrization requires a momentum cutoff @xmath86 equal to 2.1 @xmath87 .
the matrix elements of @xmath0 in a plane wave basis have been calculated in ref .
@xcite and are reported in fig .
[ figure_sedra ] .
it is seen that , while @xmath2 is less attractive than @xmath0 for small momenta , it becomes more attractive for @xmath5 larger than @xmath881 @xmath87 .
the semiclassical matrix elements of the gogny d1 interaction ( eq .
( [ simple ] ) ) and of the @xmath0 interactions calculated as a function of the single - particle energy for the nucleus @xmath1sn , are displayed in fig .
[ fig_lowk1 ] .
we remark that , close to the fermi energy , the matrix elements of the gogny interaction calculated with the d1s parametrization are about 10 kev less attractive .
it is seen that the general trend is the same as that obtained in infinite matter , with the difference that the matrix elements of the two interactions look somewhat more similar to one another .
insight into this difference can be obtained with the help of eq.([simple ] ) .
this relation is essentially an average of @xmath89 calculated at the various local momenta in the nucleus . from fig .
[ figure_sedra ] we see that in the interior of the nucleus ( at larger local momenta ) @xmath2 is more attractive than @xmath0 , but on the surface it tends to be less attractive than @xmath0 .
the semiclassical matrix elements of the induced interaction are also shown in fig .
[ fig_lowk1 ] .
these matrix elements are strongly peaked around the fermi energy and reproduce rather well the average of the quantal results already shown in fig .
[ fig_qm ] . in fig .
[ fig_ind1 ] we have separated the contributions arising from the exchange of phonons of multipolarities @xmath31=2 , 3 and 4 .
we include all the phonons with energy below 30 mev in the calculation .
the importance of the contributions with @xmath31=2 and 3 can be understood in terms of the high collectivity and low energy of the associated collective modes .
the energies and deformation parameters of the low - lying modes are @xmath90 mev , @xmath91 mev , @xmath92 mev and @xmath93 , @xmath94 , @xmath95 .
we also note that at the fermi energy the matrix elements of @xmath9 are rather large , of the order of those associated with the bare nucleon - nucleon potential @xmath0 ( in this connection , cf . the estimate in ref .
in this section , we shall discuss the results of calculations of the pairing gap which employ the matrix elements discussed above . in the nucleus under study , namely @xmath1sn
, we obtain the state - dependent pairing gaps by solving the generalized bcs equations @xcite : this means that we solve the hfb equations by treating the pairing sector self - consistently , while the mean field is described in terms of the woods - saxon potential already used in the calculation of the matrix elements . in this way
, we take into account scattering processes between nucleons lying on orbits having different number of nodes / energies @xcite , associated with non - diagonal matrix elements of the type @xmath96 , with different energies @xmath97 and @xmath98 ( or @xmath99 and @xmath100 ) .
this is at variance with the usual bcs method in which the pairing gap receives contributions only from scattering processes between time - reversal states .
we shall refer in the following either to quantal or to semiclassical calculations ; in the latter case , this is done by solving the same equations as in the quantal case , but replacing the qm matrix elements with the tf ones , using eq .
( [ offdiag ] ) and the prescription @xmath101 the positive energy states are obtained by setting the system in a spherical box .
we have checked that convergence is achieved using @xmath102=12 fm and including all the states from the bottom of the potential up to the positive energy @xmath103=50 mev . in infinite neutron matter ,
the pairing gaps calculated with the interactions @xmath0 and @xmath2 are very similar up to fermi momenta of the order of 0.6 @xmath87 , becoming increasingly different from each other at higher momenta , as can be seen from fig .
[ fig_deltaf ] ( cf . ref .
@xcite , where the d1s parametrization of the gogny interaction has been used ) .
the pairing gap obtained with @xmath0 goes to zero , around saturation , as a function of @xmath104 much faster than that associated with @xmath2 ; it can also be seen that the pairing gap obtained with @xmath0 reproduces quite accurately the result found allowing particles to interact through the argonne @xmath28 bare n - n potential @xcite . in the case of @xmath1sn , the pairing gap with the @xmath28 argonne interaction and with @xmath105 ,
was calculated previously in ref .
@xcite , obtaining a value of about 2.2 mev close to the fermi energy .
we have verified that essentially the same value is obtained with @xmath0 , in keeping with the infinite matter case ( cf .
[ fig_deltaf ] ) . in the following
, however , we shall adopt the value @xmath106 , which is the typical value associated with hartree - fock calculations performed with gogny or skyrme forces . in fig .
[ fig_m07 ] we compare the state - dependent pairing gap obtained with the gogny force and with the @xmath0 interaction .
we also compare the pairing gaps obtained inserting the semiclassical pairing matrix elements in the hfb equations .
we remark that the semiclassical approach takes out the scatter from shell effects which is a desirable feature when one wishes to reveal generic trends .
the gogny interaction leads to a pairing gap of 1.4 mev close to the fermi energy , in good agreement with the experimental value deduced from the odd - even mass difference .
this value is about two times larger than that obtained with @xmath0 .
note also that the pairing gap obtained with @xmath0 reproduces the results obtained with the bare @xmath28 interaction @xcite .
we remark that using the d1s parametrization of the gogny force , instead of d1 , one obtains a gap at the fermi energy of about 1.2 mev . in fig .
[ fig_ind ] we show the state - dependent pairing gap obtained with the pairing matrix elements of @xmath107 , resulting from the sum of the matrix elements of the bare interaction @xmath0 and of the induced interaction @xmath9 .
the average value of the resulting pairing gap @xmath108 at the fermi energy is about 2 mev .
this value is about 25% higher than that obtained in ref .
@xcite , where we solved the dyson - gorkov equation , taking properly into account all polarization effects ( induced interaction , self - energy and vertex corrections ) @xcite .
this is consistent with the fact that according to nuclear field theory ( nft ) @xcite , if one considers the effects of the exchange of phonons between pairs of nucleons , one has to consider at the same time processes where the phonon is absorbed by the same nucleon which has virtually excited it .
such processes lead to self - energy ( @xmath109-mass and single - particle splitting ) as well as to vertex correction phenomena . of these processes and for the nucleus under discussion ( @xmath1sn ) , @xmath11mass effects @xcite are the most important . taking such effects into account is equivalent to use a residual interaction @xmath110 , where @xmath111 is the quasiparticle strength @xcite at the fermi energy ( cf .
appendix a ) .
consequently @xmath112 .
because @xmath9 is proportional to @xmath113 , that is to the square of the deformation parameters associated with the low - lying collective vibration and inversely proportional to the energy denominator @xmath114 appearing in eq.([vind ] ) , a typical error of 20% on @xmath115 and @xmath114 implies a 60% error in @xmath9 . making use of this fact and of the results of figs .
[ fig_qm ] and [ fig_ind ] one obtains , as discussed in appendix b , @xmath116 mev .
it is seen that the 60% error in @xmath9 has been reduced to a 10% in @xmath117 .
this is because the stronger the particle - vibration coupling is , the larger @xmath118 but the smaller the amount of single - particle content of levels around @xmath119 ( and thus the smaller @xmath12 ) , and viceversa .
the self consistency between collectivity of the modes , strength of the particle - vibration coupling , and quasi - particle residue at the pole , typical of nft , allows theory to make predictions which are more accurate than the basic parameters entering the calculation , as a result of a delicate process of cancellation of errors .
summing up , we have found that the small difference existing between the matrix elements of the bare @xmath0 interaction and those of the gogny interaction leads to important differences between the pairing gaps associated with the two forces ( @xmath120 , for @xmath25 ) .
this difference is removed and eventually overwhelmed by including the attractive contribution coming from the induced interaction @xmath9 arising from the exchange of surface vibrations , which acts only on a rather small energy range around @xmath3 .
we note that the pairing gap in @xmath1sn is not much changed , including also the effect of spin fluctuations @xcite , which instead give the dominant ( repulsive ) contribution to the induced interaction in neutron matter .
the 25% excess displayed by the resulting total pairing gap with respect to the experimental value is corrected by considering the corresponding quasiparticle strength at the fermi energy .
an important question which remains to be addressed quantitatively is the ability @xmath121 as well as @xmath2 have to describe in detail the isotopic and nuclear structure dependences displayed by @xmath8 .
this subject will be addressed in a future publication . within this context , we note that the dependence of the pairing gaps generated by @xmath2 and by @xmath122 on temperature and rotational frequency is expected to be quite different , as @xmath2 and @xmath0 are essentially independent of these parameters , while @xmath9 is quite sensitive to them .
it is also worth mentioning that a recent study in symmetric nuclear matter revealed a very strong extra attraction coming from the induced interaction @xcite .
nuclear matter and finite nuclei may , however , show only qualitative similarities in this case and a quantitative link may be difficult to establish .
we thank l. coraggio for providing us with the matrix elements of the @xmath0 interaction .
blaizot and g. ripka , _ quantum theory of finite systems _ , the mit press ( 1985 ) . p. ring and p. schuck , _ the nuclear many - body problem _ , springer - verlag ( 1980 ) .
j. decharg and d. gogny , phys . rev .
* c21 * ( 1980 ) 1668 .
t. duguet , p. bonche , p .- h .
heenen , j. meyer , phys . rev .
* c65 * ( 2002 ) 014311 ; j. dobaczewski , w. nazarewicz , t.r .
werner , j.f .
berger , c.r .
chinn , j. decharg , phys . rev .
* c53 * ( 1996 ) 2809 .
e. garrido , p. sarriguren , e. moya de guerra , p. schuck , phys .
* c60 * ( 1999 ) 064312 ; e. garrido , p. sarriguren , e. moya de guerra , u. lombardo , p. schuck , h.j .
schulze , phys .
* c63 * ( 2001 ) 037304 .
u. lombardo and h .- j .
schultze , in _ physics of neutron star interiors _ , eds .
d. blaschke , n.k .
glendenning and a. sedrakian , springer ( 2001 ) , p.30 .
f. barranco , r.a .
broglia , g. col , e. vigezzi and p.f .
bortignon , eur .
j. * a21 * ( 2004 ) 57 .
f. barranco , r.a .
broglia , g. gori , e. vigezzi , p.f .
bortignon , j. terasaki , phys . rev
* 83 * ( 1999 ) 2147 .
c.mahaux , p.f .
bortignon , r.a .
broglia , c. h. dasso , phys . rep .
* 120 * ( 1985 ) 1 .
x. vias , p. schuck , m. farine and m. centelles , phys . rev .
* c67 * ( 2003 ) 054307 .
l. coraggio , a. covello , a. gargano , n. itaco , t.t.s .
kuo , phys .
* c66 * ( 2002 ) 064311 .
a. sedrakian , t.t.s .
kuo , h. mther and p. schuck , phys .
* b576 * ( 2004 ) 68 .
a. bohr and b.r .
mottelson , _ nuclear structure _ , vol .
ii , benajmin ( 1975 ) , p. 432 . f. barranco , r.a . broglia , h.esbensen , e. vigezzi , phys.rev . * c58*(1998)1257 .
a. schwenk , b. friman , g.e .
brown , nucl . phys . * a713*(2003 ) 191 .
f. barranco , r.a .
broglia , h. esbensen and e. vigezzi , phys . lett .
* b390 * ( 1997 ) 13 .
j. terasaki , f. barranco , r.a .
broglia , e. vigezzi , p.f .
bortignon , nucl . phys . *
a697*(2002 ) 127 .
d. r. bes , r.a.broglia , g. g. dussel , r. j. liotta and h. m. sofia , nucl . phys .
* a260 * ( 1976 ) 1 ; ibid . ,
27 ; d. r. bes , r.a.broglia , g. g. dussel , r. j. liotta and r. j. perazzo , nucl . phys .
* a260 * ( 1976 ) 77 ; d. brink and r.a .
broglia , _ nuclear supefluidity : pairing in finite systems _ , cambridge university press , cambridge ( 2005 ) .
a. bohr and b.r .
mottelson , _ nuclear structure _ ,
vol i , benjamin ( 1969 ) .
schrieffer , _ superconductivity _ , benjamin , new york ( 1964 ) .
g. gori et al .
, to be published . c. shen et al .
, to be published .
in what follows we consider some of the consequences the particle - vibration coupling has on the pairing correlations of particles moving in a single j - shell interacting through a bare nucleon - nucleon pairing potential with constant matrix elements @xmath123 @xcite . for this simple model , the value of the occupation numbers @xmath124 and @xmath125 must be the same for all the @xmath126 orbitals . in particular , the occupation probability for the case when the system is occupied with @xmath127 particles ( half filled shell ) , where @xmath128 is @xmath129 @xmath130 consequently , the pairing gap is given by the relation @xmath131 because the density of levels is proportional to the @xmath11mass ( @xmath132 ) , the effective ( dressed ) degeneracy can be written @xmath133 in terms of the quasiparticle strength at the fermi energy @xmath134 where @xmath135 , @xmath136 being the mass enhancement factor .
in the case of nuclei this dimensionless quantity ( which measures the strength with which nucleons couple to low - lying collective vibrations ) is of the order of 0.5 , closer to the strong than to the weak coupling situation of bcs ( in which case @xmath137 ) @xcite . because of their coupling to vibrations , nucleons spend part of the time in more complicated configurations than pure single - particle states . the quantity ( quasiparticle strength @xcite ) @xmath138 measures the content of single - particle strength present in levels around the fermi energy available for nucleons to interact through a two - body interaction , in particular through through a pairing force . consequently , due to the self - energy processes arising from the particle - vibration coupling phenomenon , the pairing strength becomes @xmath139 .
the exchange of vibrations between pairs of nucleons moving in time reversal states close to the fermi energy gives rise to an effective pairing interaction of strength @xmath140 , where @xmath141 stands for the induced pairing interaction controlled by the particle - vibration coupling vertices .
taking these effects into account , one can write @xmath142 identifying @xmath123 with @xmath0 and @xmath141 with @xmath9 , one finds from fig .
[ fig_ind ] that @xmath143 as seen from eq .
( [ the16 ] ) , the number to be compared with experimental findings is @xmath144 note that in the above discussion we have not considered the errors to be expected in estimating both @xmath141 and @xmath12 .
this subject is taken up in appendix b.
from fig . [ fig_qm ]
it is seen that the average values of @xmath9 and @xmath0 at the fermi energy are @xmath145 = -0.15 mev and @xmath146 = -0.16 mev . because @xmath9 is of the order of @xmath147
, the matrix elements of @xmath9 will display an error equal to twice the average error displayed by @xmath115 plus that displayed by @xmath114 . assuming this error to be 20% for both quantities
( a rather extreme negative situation ) one concludes that @xmath1480.09 mev . consequently @xmath1490.09 mev . because @xmath108 = 2 mev for @xmath150 = -0.31 mev , one obtains @xmath151 mev .
we now proceed to the calculation of @xmath152 , where @xmath153 is the density of single - particle levels around the fermi energy for one spin orientation and one type of nucleon . making use of the value @xmath154 3.4 mev@xmath155 appropriate for @xmath1sn
, one gets @xmath156 .
consequently @xmath157 = 1.35 @xmath158 0.15 mev , where @xmath159 .
we thus note a conspicuously smaller error for @xmath13 than for @xmath9 .
this is because the larger is @xmath136 ( induced pairing contribution ) , the smaller is the content of single - particle strength and thus of @xmath12 , associated with levels lying close to the fermi energy , and viceversa .
diagonal pairing matrix elements of the induced interaction ( upper panel , solid diamonds ) and of the gogny force ( lower panel , solid circles ) , displayed as a function of the single - particle energy @xmath40 of the state @xmath160 calculated using the bare nucleon mass and the single - particle wavefunctions of a woods - saxon potential with standard parameters ( depth @xmath161= -49 mev , diffusivity @xmath1620.65 fm , and radius @xmath163 6.16 fm ) , including the spin - orbit term , parametrized according to ref . @xcite . also shown by means of vertical lines
is the position of the fermi energy , @xmath164 mev .
note the different scale in the two figures . ]
sn . the same as fig .
[ fig_semicl ] ( i.e. also with @xmath165 ) but for the fact that the single - particle wavefunctions have been calculated making use of the woods - saxon potential including the spin - orbit term already used for fig .
[ fig_qm ] , where the diagonal matrix elements have already been shown .
the non - diagonal matrix elements are plotted here at an energy @xmath166 which is the average between the energies of the initial and final states . ]
the semiclassical matrix elements of the induced interaction , calculated according to eq .
( [ vsemi1])(dash - dotted curve ) , are compared with the matrix elements of the gogny force ( solid curve , cf .
[ fig_ws ] ) and with those of the @xmath0 interaction ( dashed curve ) .
calculations are performed with @xmath105 and with the same woods - saxon potential used in figs .
[ fig_qm ] and [ fig_ws ] . ]
the semiclassical matrix elements of the induced interaction , as a function of the energy of the single - particle levels ( calculated with @xmath105 ) associated with the multipolarities @xmath167 2 , 3 , and 4 of the formfactor ( see eq.([vind ] ) ) , are shown by a dashed , dotted and solid line respectively .
the sum of the three contributions is displayed by means of a dash - dotted line ( cf .
[ fig_lowk1 ] ) . ] sn calculated with a woods - saxon potential ( with depth @xmath168 = -64 mev , diffusivity @xmath169 0.65 fm and radius @xmath170 6.17 fm ) , as a function of the single - particle energy .
the @xmath23mass @xmath44 was set equal to 0.7@xmath171 .
the fermi energy is @xmath172- 8.6 mev .
solid triangles ( open squares ) display the results of a hfb calculation with the gogny interaction , with quantal ( semiclassical ) matrix elements .
the solid diamonds refer instead to a hfb calculation using the semiclassical matrix elements of the @xmath0 potential . ]
sn obtained making use of the woods - saxon potential used in calculating the results displayed in fig .
[ fig_m07 ] ( i.e. with @xmath26 ) and different interactions . diamonds and triangles show again the gaps obtained with the @xmath0 and @xmath2 interactions , respectively .
stars show the gap associated with the matrix elements obtained summing the matrix elements of @xmath0 and @xmath9 .
the latter have been evaluated using eqs .
( [ vsemi1 ] ) and ( [ offdiag1 ] ) ]
|
the dependence on the single - particle states of the pairing matrix elements of the gogny force and of the bare low - momentum nucleon - nucleon potential @xmath0 - designed so as to reproduce the low - energy observables avoiding the use of a repulsive core - is studied in the semiclassical approximation for the case of a typical finite , superfluid nucleus ( @xmath1sn ) .
it is found that the matrix elements of @xmath0 follow closely those of @xmath2 on a wide range of energy values around the fermi energy @xmath3 , those associated with @xmath0 being less attractive .
this result explains the fact that around @xmath3 the pairing gap @xmath4 associated with the gogny interaction ( and with a density of single - particle levels corresponding to an effective @xmath5-mass @xmath6 ) is a factor of about 2 larger than @xmath7 , being in agreement with @xmath8= 1.4 mev .
the exchange of low - lying collective surface vibrations among pairs of nucleons moving in time - reversal states gives rise to an induced pairing interaction @xmath9 peaked at @xmath3 .
the interaction @xmath10 arising from the renormalization of the bare nucleon - nucleon potential and of the single - particle motion ( @xmath11mass and quasiparticle strength @xmath12 ) due to the particle - vibration coupling leads to a value of the pairing gap at the fermi energy @xmath13 which accounts for the experimental value .
an important question which remains to be studied quantitatively is to which extent @xmath4 , which depends on average parameters , and @xmath13 , which explicitly depends on the parameters describing the ( low - energy ) nuclear structure , display or not a similar isotopic dependence , and whether this dependence is born out by the data . * pairing matrix elements and pairing gaps with bare , effective and induced interactions * f. barranco@xmath14 , p.f .
bortignon@xmath15 , r.a .
broglia@xmath16 , g. col@xmath15 , + p. schuck@xmath17 , e. vigezzi@xmath18 and x. vias@xmath19 + @xmath14 departamento de fisica aplicada iii , escuela superior de ingenieros , + universidad de sevilla , camino de los descubrimientos s / n , 41092 sevilla , spain + @xmath20 dipartimento di fisica , universit degli studi , via celoria 16 , 20133 milano , italy + @xmath18 infn sezione di milano , via celoria 16 , 20133 milano , italy + @xmath21 the niels bohr institute , university of copenhagen , blegdamsvej 17 , 20100 copenhagen , denmark + @xmath17 institut de physique nuclaire , 15 rue georges clmenceau , 91406 orsay cedex , france + @xmath19 departament destructura i constituents de la matria , facultat de fsica , + universitat de barcelona , diagonal 647 , 08028 barcelona , spain
|
for decades it has been well established that the population of star clusters in the milky way is strongly bimodal , with a population of old , massive , and compact globular clusters in the galactic spheroid and a physically distinct population of less populous open clusters in the disk .
however a growing body of observations suggests that this distinction is less pronounced in other types of galaxies , and that the nature of the young and intermediate - age cluster populations differs dramatically in different galactic environments .
early studies of the star cluster populations in the magellanic clouds revealed the presence of young and intermediate age `` populous blue clusters '' or `` blue globular clusters '' that are absent in the milky way @xcite , and similar populations have been found in m33 and other nearby late - type galaxies @xcite .
populations of very massive `` super star clusters '' have been found in starburst galaxies and strongly interacting galaxies @xcite , leading to the speculation that many of these objects are progenitors of globular clusters @xcite .
other observations obtained over the past 15 years show that the specific frequencies and mass spectra of young stellar associations vary systematically along the hubble sequence .
the most complete data of this kind come from the luminosity functions of giant hii regions , which roughly trace ( for luminous objects ) the ultraviolet luminosity functions of the embedded stellar associations .
the luminosities of the brightest hii regions increase sharply from s0/sa sc and irr galaxies @xcite , due to the combination of an increase in both the total numbers of star forming regions per unit galaxy mass , and a change in the shape of the association luminosity function @xcite .
reliable measurements of the continuum luminosity functions of the associations themselves are much more difficult to measure , due to the increased difficulty of isolating the associations against the background light of the disk , but studies by @xcite and @xcite are consistent with the results derived from hii region statistics , namely that the frequency and characteristic luminosities of the associations increase in systems with higher star formation rates ( sfrs ) . taken together these studies show that the clustering spectrum of star formation is sensitive to the local physical conditions in the interstellar medium .
such variations are of critical importance for understanding galaxy formation and evolution , because the feedback effects of star formation on the ism via radiation , supernovae , and stellar winds , can vary dramatically depending how concentrated the massive star formation is in space and time @xcite .
a critical missing link in this picture is a systematic study of how the populations of massive , bound star clusters vary as functions of galaxy type , ism properties , or star formation rates . although the existing observations of selected local group galaxies , starburst galaxies , and merger remnants show strong differences in the properties of their young massive cluster populations , the data are fragmentary , and fundamental questions remain about the physical nature and origin of these differences .
for example it is not at all clear whether the `` super star clusters '' or `` blue globular clusters '' in starburst galaxies are the result of a distinct mode of cluster formation ( analogous to the differences between the open and globular cluster systems in the milky way ) , or whether they simply are the extension of a continuous trend in disk cluster populations along the hubble sequence .
likewise it is unclear whether the strong trends seen in the properties of ob / hii associations in galaxies are also present in the bound cluster populations .
the absence of a substantial body of data on clusters in nearby galaxies is not surprising , because a typical object ( @xmath0 , @xmath1 pc ) becomes effectively stellar in appearance beyond a distance of @xmath21 mpc when observed from the ground , and reliable identification and measurement of such clusters is challenging even with the hubble space telescope ( hst ) .
however hst does offer the capability to obtain reliable statistics on at least the upper end of the luminosity function of young clusters out to distances of @xmath220 mpc , if high s / n images are analyzed with precision psf - fitting techniques , and extensive artificial star and cluster simulations are used to quantify the effects of spatial resolution , structured background , and stellar crowding on the cluster identification and photometry .
we have undertaken such a survey of the massive young star cluster populations in nearby spiral and irregular galaxies , using archival imaging from hst .
although we are using a familiar tool ( psf - fitting stellar photometry ) for this project , we are devoting this first paper of the series to our photometry techniques because of this application to slightly - extended objects .
although our methods are optimized for the problem of detecting compact star clusters against a background of resolved stars , many aspects of the technique should be applicable to other problems where semi - resolved objects need to be detected against a complex background of point sources .
section 2 is used to outline the theory behind the technique , and section 3 shows a sample reduction of ngc 3627 ( m66 ) .
paper ii will describe the statistical analysis we have developed to analyze the color - magnitude - radius information .
subsequent papers in this series will apply these techniques to a sample of 30 nearby ( @xmath3 mpc ) spiral and irregular galaxies , in order to quantify the trends in cluster properties along the hubble sequence , and to a more distant sample of starburst and other peculiar galaxies , to better quantify any differences in the cluster populations in these objects .
even with the superb resolution possible with the hubble space telescope , the accurate detection and analysis of distant star clusters is a difficult process . at 20 mpc ,
a cluster with a core radius of 2 pc has a size smaller than the wfpc2 psf .
after convolving the cluster and a stellar point source with the wfpc2 psf , the cluster appears less than 20% broader .
this difficulty in detection of clusters is further compounded by the two factors .
first , the relatively large pixels on wfpc2 ( especially the wide field cameras ) means that both the star and the cluster have psfs smaller than one pixel ; thus measurement of the fwhm of an object is thus impossible and size can only be estimated by the fraction of light falling outside the central pixel .
second , blends of supergiants and distant background galaxies will also appear slightly broader than stars , and thus additional care must be taken when identifying cluster candidates .
a very basic question that must first be addressed is the photometry method that we will employ .
there are three basic options : aperture photometry , stellar psf - fitting photometry , and cluster psf - fitting photometry .
aperture photometry is the simplest ; it involves the measurement of the number of counts within a radius of the center of the aperture . as such
, it is the least sensitive of the three to systematic errors , as no _ a priori _ assumptions are made regarding size or shape of the object being measured .
however , aperture photometry is the most sensitive to variable background ( as psf - fitting photometry of any kind gives the highest weight to the central pixels ) , and most routines do not provide any measurement of star size or shape .
thus aperture photometry is not the best algorithm for use in this project .
the other two possibilities are psf - fitting routines of two flavors stellar and cluster .
stellar psf - fitting routines attempt a simple three - parameter fit for each object ; the three parameters being position ( @xmath4 , @xmath5 ) and number of counts .
in addition to measuring these values for each object , sharpness and roundness values are also returned , generally for the purpose of rejecting non - stellar objects from the photometry list .
when searching for clusters , however , these values can also be used to reject non - clusters stars , blends , background galaxies , etc .
in contrast , cluster psf - fitting photometry attempts a more complex fit of each object .
the routine described by @xcite , for example , fits six parameters per object : position , counts , and three size parameters related to major axis , minor axis , and orientation . while the cluster - fitting approach has the advantage that it directly provides the size and shape of each object , the same information is also contained in the psf - fitting algorithm output .
the only significant differences between the two approaches are the following .
stellar psf - fitting photometry has the disadvantage that there is a systematic error in the magnitudes caused by assuming a stellar psf . while this can be easily corrected , it does add an extra step in the reduction process .
cluster psf - fitting photometry , however , has the disadvantage that the full uncertainties are extremely difficult to measure generally only the standard error of the brightness measurement is reported .
again , this problem can be solved by adding an extra step to the reduction process .
while we conclude that aperture photometry is not optimal for this project because of its large random errors , we find that either stellar psf - fitting photometry or cluster psf - fitting photometry will work equally well . because of our familiarity with the stellar psf - fitting package hstphot @xcite
, we will take that route for our primary reduction procedure , along with an updated version of its cte corrections @xcite available at aed s web site .
the process of photometering the clusters will thus be a two - step process
stellar photometry with hstphot , followed by determination of the systematic errors caused by use of the stellar psf .
a final note involves the determination of the background .
the typical procedure is to measure the image values in an annulus around the object being measured and use a mode or robust mean algorithm to determine the background value .
however , in the galaxies in this survey , we face the difficulty of a rapidly - varying background in which the default hstphot background annulus is much too large to provide an accurate measurement of the background value at the position in question . instead , we require a very crude technique in which the background is calculated by averaging the image values in a 2-pixel wide annulus immediately outside the photometry annulus .
this technique also avoids the bias inherent in most sky algorithms , which are based on the assumption that the lower values are more indicative of the true sky level , while higher values are caused by stars ( which is generally true in uncroweded fields ) .
this also largely eliminates the effect of associations , as any association with a diameter of 5 wfc pixels or more ( 20 pc at a distance of 10 mpc ) will be subtracted out as background and any cluster within the association photometered properly .
regardless of whether one uses stellar or cluster psf - fitting photometry , knowledge of the point source ( stellar ) psf is the critical first step , as the observed image of any cluster is its intrinsic image convolved with the psf . for the data we are analyzing in this project ,
the determination of the stellar psf is not a trivial process and the process should be described in detail .
hstphot s base psf library is computed using tiny tim psfs @xcite .
the exact algorithm can be found in @xcite ; the end result is a library of many psfs per chip per filter .
for the planetary camera , a total of 1600 psfs per filter are measured
@xmath6 subpixel centerings at @xmath7 chip positions .
each wide field camera has a factor of four more psfs , as there are @xmath8 subpixel centerings at each chip position to compensate for the lager pixels ( and thus larger differences between psfs separated by a fraction of a pixel ) .
after reduction of many data sets , it was seen that the psfs showed small systematic errors as a function of position on each chip .
this is assumed to be an error on the part of the tiny tim models , which while excellent , are obviously not expected to be perfect . in order to compensate , each psf taken from the library is expanded or contracted slightly ( a scale difference of order 2 % , independent of chip ) , using a quadratic correction term in addition to the expansion necessary to compensate for geometric distortion .
( tiny tim does not account for geometric distortion . ) the 34th row error is also corrected here , according to the prescription given by @xcite .
these factors do not significantly affect our photometry ( at the @xmath9% level ) , but will significantly affect our cluster discrimination .
the result of this procedure is our best initial guess of the stellar psf at any position on any chip .
however , small focus changes and small pointing errors will modify the stellar psfs further , and it is necessary to add a correction image to the psf .
this correction image is usually measured from bright , isolated stars , but such objects do not exist in many of our galaxy fields . instead , we frequently find that the only objects sufficiently bright for psf determination are clusters ; use of these to set the stellar psf would clearly be disastrous if we are attempting to locate clusters by their size .
instead , we adopt a two - iteration process in our photometry . in the first iteration
, we simply use the library psfs , plus mean correction images determined from well - resolved local group dwarf galaxies , to photometer all stars . as hst pointing errors ( when returning to the same position on the sky ) are generally much more significant than focus changes , this provides a good first - order approximation .
we then plot sharpness vs. magnitude for all recovered objects .
the locus of stars is usually obvious in such a plot ; and a revised correction image can be calculated such that the mean star sharpness is zero .
( the hstphot sharpness parameter is defined such that a clean star should have a sharpness of zero . )
the adjustment to the correction image is modeled as a difference between two psf functions ; one corresponding to the mean fwhm of stars in the image and the other a free parameter . the functional form of the psf used for this fit was @xmath10 for the pc and @xmath11 for the wfc ; these functions provided the best fits to the psfs . in both equations , @xmath12 is the parameter being fit and @xmath13 is the distance from the center of the psf .
thus this is essentially a one - parameter fit that can be made for any galaxy in which even a small number of individual stars are resolved .
the photometry is then re - measured , using the adjusted psf correction image , producing our final stellar photometry for the galaxy . as mentioned in section [ sec_phottype ] ,
the principal weakness of stellar psf - fitting photometry in this application is that the attempt to fit a slightly extended object using a stellar psf will cause a systematic underestimation of the brightness .
this effect can be measured ( and corrected ) in two ways . in this section
we discuss an analytical approach ; the next section discusses a semi - empirical approach .
given the hstphot equations given by @xcite , we can calculate the expected photometric consequences of an object being slightly extended .
hstphot uses @xcite equation 8 , reproduced here in equation [ eq_s ] , to measure the signal contained in an object . to eliminate gain from the equations ,
all values will be given in adu .
@xmath14 where @xmath15 is the weighting factor ( generally 1 ; between 0 and 1 in the outermost pixels to set a circular aperture ) , @xmath16 is the residual image ( original data minus sky and other nearby stars ) , @xmath17 is the stellar psf .
the expected scatter in a pixel , @xmath18 , is given by @xcite equation 10 , @xmath19 where @xmath20 is the mean sky level and @xmath21 is the read noise in electrons .
for any object , the expected residual in any pixel is given by @xmath22 where @xmath23 is the actual number of counts contained while @xmath24 is the object s psf , a convolution of the stellar psf and the object s intrinsic shape .
substituting this into equation [ eq_s ] , we obtain @xmath25 the magnitude correction is simply @xmath26 , so the systematic error ( as a function of actual psf , brightness , and background ) can be calculated using equation [ eq_sratio ] .
likewise , the measured sharpness ( and indication of object size ) is given by @xcite equations 14 and 15 , and is a function of actual psf , brightness , and background that can be easily evaluated . because @xmath16 is randomly distributed with an uncertainty of @xmath18 , the magnitude error and sharpness measured this way are expectation values only .
the full distribution can most easily be measured using a monte carlo test , as @xmath16 appears in both the numerator and the denominator of equation [ eq_sratio ] .
the more direct approach , in terms of simplicity , is to place a large number of artificial clusters on the image and photometer them a procedure analogous to artificial star tests in stellar population work . by placing many artificial clusters with a given magnitude and size on each frame
, one finds not only the systematic difference between input and output magnitudes as a function of input parameters , but also the completeness fraction , photometry scatter , and mean sharpness .
we have developed a very detailed artificial cluster algorithm .
we began by attempting to reconstruct the young lmc cluster ngc 1818 , using @xcite isochrones , a binary fraction of 35% , the wfpc2 psf , and the profile measured by @xcite , @xmath27 where @xmath13 is the distance from the cluster center in arcsec and @xmath28 is the central surface brightness . using these inputs
, we were able to quantitatively reproduce the observed cmd and spatial distribution of stars seen in archival wfpc2 images ( proposal go-6277 ) .
confident in this approach , we applied the same technique to add artificial clusters to our ngc 3627 images . to simplify the problem ( and because clusters at the distances we are observing are too small to determine the shape of the profile ) , we reduced the two structural parameters of @xcite equation 1 to one parameter by noting that most of the scatter in their measured @xmath29 values ( their table 1 ) is from measurement uncertainties ; the values are consistent at the @xmath30 level with @xmath31 .
thus we will assume a universal cluster profile of @xmath32 converting to core radii ( @xmath33 ) using @xcite equation 22 , this becomes @xmath34 and their equation 2 becomes @xmath35 where @xmath36 is the luminosity falling within a radius and @xmath37 is the integrated luminosity .
the artificial cluster tests are thus made by placing clusters with a variety of ages , initial masses , and core radii on the image and photometering them .
we tests @xmath38 such clusters in our ngc 3627 reduction , described below .
we will use these two techniques in very different ways , as each has its strengths and weaknesses .
the artificial cluster approach is less prone to error , as one is actually photometering objects on the image . by building up a large library of artificial cluster results , therefore
, one can reproduce the observational effects with great accuracy
. this will be critical in our statistical analysis , the topic of paper ii .
however , orders of magnitude more artificial clusters are required in order to determine the integrated magnitudes and core radius of an observed cluster , a task that is done much more easily through our analytical technique . thus the analytical technique is used to convert observed properties to physical properties ; the artificial cluster technique to do the reverse .
a comparison will be made in our ngc 3627 analysis .
armed with hstphot photometry and an understanding of how extended psfs affect the photometry , we now attempt to determine which objects in our photometry list are likely clusters .
a library of artificial star tests ( placed on the image with a distribution based on the distribution of light on the images in order to correctly sample the observed data ) will help to identify the distribution of single stars and blended stars on a sharpness vs. magnitude plot ; we can then identify outlying objects on the broad ( negative sharpness ) side of the diagram as clusters . as we are primarily interested in finding compact clusters ( which are round ) , we can also use roundness information , when available , to help reduce the number of spurious detections caused by blended stars and background galaxies .
each candidate cluster is characterized by photometry ( typically @xmath39 ) , standard errors in the magnitudes , sky brightness in each frame , and a sharpness measurement . as sharpness is primarily a function of angular size ,
this value can be determined with reasonable accuracy , which in turn allows a measurement of the integrated @xmath39 magnitudes ( as described in the previous section ) . in terms of studying the properties of the cluster population , however , it is more accurate to statistically model the observed data using theoretical models of the cluster population and observational errors as determined by `` artificial cluster '' tests .
the reason is one of accuracy ; it is far simpler to use the artificial cluster results ( completeness , systematic errors , and scatter ) to create a simulated luminosity function and/or color distribution from a model population than it is to deconvolve these corrections from observed data .
this approach is identical in concept to the cmd - synthesis approach for studying stellar populations described by @xcite , and will be explored in paper ii .
we will employ both types of corrections in this study .
in general , the properties of a single cluster are best described by the former ; while statistical properties can be most easily described by the latter .
to illustrate our reduction and analysis procedures , we present a detailed description of our study of ngc 3627 , using data studied previously by @xcite and @xcite . one of the reasons for this selection is that @xcite found unusual aperture corrections when studying this galaxy , one possible explanation for which was a large cluster population .
thus , while we expect to examine a large number of objects with no _ a priori _ knowledge of cluster properties , ngc 3627 seemed an intriguing prospect for a first test of our cluster - detection methods .
wfpc2 observations of ngc 3627 were taken over a 60-day period between 12 november 1997 and 10 january 1998 for proposal go-6549 ( pi sandage ) .
the data consist of 24 exposures in f555w ( totaling 58800s of integration ) and 10 exposures in f814w ( totaling 25000s of integration ) , all images taken in pairs to allow the creation of 17 cosmic ray cleaned ( 12 f555w and 5 f814w ) single - epoch images .
unfortunately ( for the purposes of creating a deep image ) , the images were split between two pointings separated by a few wfc pixels .
we therefore created two deep images in each filter using the _ crclean _ routine of hstphot @xcite , in addition to the 17 single - epoch frames .
one of the deep f555w images is shown in figure [ fig_image ] . because of the lack of bright , isolated stars in the data , we have used the iterative procedure described in section [ sec_psf ] to determine the stellar psf .
an initial photometric pass was made using the _ multiphot _ routine from hstphot , producing stellar photometry ( f555w and f814w magnitudes and sharpness ) from the default psfs .
the default options for hstphot were used , except that the local sky value was calculated before each photometry measurement to compensate for the variable background , object classification was turned off to produce more uniform photometry , and aperture corrections and psf corrections were turned off because of a lack of adequate stars .
the `` local sky '' algorithm is described in section 2 ; we are confident of its accuracy because we are able to recover artificial stars to a systematic accuracy of better than 1% ( systematic ) down to @xmath40 in wfc2 and @xmath41 in the other chips .
figure [ fig_psf0 ] shows all pc detections from one of the deep f555w images , with sharpness on the @xmath5-axis and @xmath42 on the @xmath4-axis .
( we choose @xmath42 because the amount of exposure time in @xmath43 was 2.4 times that in @xmath44 . ) for comparison , figure [ fig_psfngc121 ] shows a similar plot for an image centered on the smc globular cluster ngc 121 , which has more severe crowding but no clusters .
it is clear from a comparison of the two figures that a large population of extended objects exists in ngc 3627 , which are distinguished by the negative sharpness values .
that this is not simply the result of a very broad stellar psf in these data is demonstrated by the sharpness values of cepheids found by @xcite and @xcite ( assuming that most of the objects they classified as cepheids are single stars ) , which are marked with asterisks in figure [ fig_psf0 ] and largely fall near a median sharpness @xmath45 ( compared with a median sharpness of @xmath46 for all objects in the pc ) .
similar plots for the other three chips , as well as those for the other images , are qualitatively identical , with a stellar population similar to that in figure [ fig_psfngc121 ] plus a population of extended objects . eliminating the cepheids that are outliers and using the remainder as psf stars , we measure the mean values of sharpness , background , and counts , which are given in table [ tab_psf0 ] . from these values
we calculate the psf adjustments necessary to produce zero mean sharpness , given as the final column of table [ tab_psf0 ] . comparing the corrections for the two pointings ( assuming the actual psfs are the same in each pointing ) , we estimate the rms error in the fwhm corrections to be @xmath47 pixels . after determining an appropriate psf adjustment ,
the photometry was redone .
figure [ fig_psff ] shows the revised plot of sharpness values , with the crosses again indicating our psf stars ( the cepheids found by previous studies ) .
although there is some scatter , the psf stars have a median sharpness of @xmath48 , significantly improved from the first iteration .
( that the sharpness is not exactly zero after applying the corrections from [ sec_error ] is a result of the interdependence of the background level measurement and psf in the photometry algorithm .
an additional iteration would reduce the median sharpness further , but @xmath48 is adequate for our purposes . ) one of the selection criteria mentioned in the introduction is that we require the deepest possible images so that cluster sizes can be measured with reasonable signal - to - noise .
a majority of the data sets meeting this criterion were taken in cepheid searches , which require of order 30 deep images in one filter .
because of this , there are two by - products of this cluster survey .
first is the opportunity to make a consistency check on our stellar photometry by comparison with that of the previous groups to have analyzed these data .
second is that these objects have well - measured distances and mean extinctions as a result of the cepheid searches , which will thus improve the accuracy of our cluster measurements .
thus , although we are searching for clusters , a by - product of our ngc 3627 stellar photometry is photometry of the cepheids found by @xcite and @xcite .
we obtained light curves for the previously - discovered cepheids using techniques described by @xcite .
mean magnitudes and periods are given in table [ tab_cepheid ] .
for the six cepheids well - recovered in all three studies ( using a cutoff of @xmath49 for our photometry ) , we find excellent agreement in the reddening - free magnitudes ( @xmath50 ) the present photometry and that of @xcite agree to @xmath51 magnitudes ; @xcite is @xmath52 magnitudes fainter .
we note that our magnitudes are systematically fainter than the other studies ( @xmath53 magnitudes in @xmath42 and @xmath54 in @xmath55 ) ; the cause of this is likely our technique for determining sky values over a rapidly - varying background .
we can determine the distance and extinction of ngc 3627 using the @xmath42 and @xmath56 magnitudes and a cepheid period - luminosity ( p - l ) relation . adopting the @xcite p - l relations
, we find a true distance modulus of @xmath57 and an apparent distance modulus of @xmath58 ( and thus a mean extinction of @xmath59 ) .
this is consistent with the literature values of @xmath60 @xcite and @xmath61 @xcite ; both literature values have been corrected to the zero points used in the present work . adopting the @xcite p - l relations ( adjusted to an assumed lmc distance modulus of 18.50 )
, we find a true distance modulus of @xmath62 and an apparent distance modulus of @xmath63 ( a mean extinction of @xmath64 ) .
we find this distance inconsistent with the value of @xmath65 reported by @xcite ; the reason for the discrepancy is an unexplained subtraction of 0.18 mags from the @xcite distance .
we believe this is because of the aperture correction discrepancy mentioned by @xcite ; if this is indeed the cause , the shorter distance of @xcite is in error .
for either relation , the scatter in @xmath66 is mostly accounted for by observational errors , implying that effects of differential reddening are present but not large .
we conservatively adopt the average values and increase the uncertainties accordingly , producing a distance modulus of @xmath67 ( corresponding to a distance of @xmath68 mpc ) and mean extinction of @xmath69 .
we finally turn our attention to the problem of locating the clusters .
we will again use the sharpness vs. @xmath42 plots to make this discrimination , beginning with the artificial star tests created by hstphot s _ multiphot _ routine . from these tests
, we find the @xmath70 lower limit of the artificial star population to be given by @xmath71 for pc1 , wfc3 , and wfc4 ; and @xmath72 for wfc2 .
similar limits can be given for the hstphot roundness parameter , which is zero for a perfectly - round object and 1 or higher for an elongated object .
the @xmath70 limits for the artificial star populations are given by @xmath73 for pc1 ; @xmath74 for wfc2 ; and @xmath75 for wfc3 and wfc4 .
figure [ fig_psffake_envelope ] shows the @xmath76 sharpness envelopes of artificial stars in four chips ; figure [ fig_psffake_round ] shows the @xmath77 roundness envelopes .
an examination of figure [ fig_psffake_envelope ] uncovers two notable features .
first , that the plot is largely ( though not completely ) symmetric between positive and negative sharpness values indicates that blending of individual stars ( thus creating the appearance of clusters ) is not a major factor in these data .
we can quantify the contribution of single star blends and photometry errors to our cluster sample by measuring the ratio of the number of artificial stars recovered below the sharpness envelope but inside the roundness envelope to the number recovered within the both envelopes . limiting the photometry to @xmath78 in wfc2 and @xmath79 in the other chips to reduce contamination , we find contamination fractions of 0.14% in pc1 , 0.18% in wfc2 , 0.17% in wfc3 , and 0.23% in wfc4 .
a comparison of figures [ fig_psfngc121 ] and [ fig_psffake_envelope ] indicates that , for real data , we should add a constant width of 0.05 to the envelope .
( observed data are unfortunately never as well - behaved as simulated data . )
this is done for our observed ngc 3627 data , which is plotted in full in figures [ fig_psf_envelope ] and [ fig_psf_round ] . in total , we have 923 `` stars '' ( objects with sharpness and roundness values within the stellar envelopes ) , which , based on our contamination fractions from the previous paragraph , produce an estimated 1.7 false detections of clusters . in contrast , we have 553 objects falling below the sharpness envelope but within the roundness envelope ; the contamination is thus negligible .
it is clear that there is also a large population of extended objects at fainter magnitudes , although discrimination is impossible with the accuracy that we have obtained for brighter objects and thus we have chosen to omit them from our sample .
as described in section [ sec_error ] , we have two methods of measuring cluster sizes and integrated magnitudes from hstphot photometry theoretically or semi - empirically .
the theoretical approach attempts to convert observational quantities to physical properties ; the artificial cluster approach is the opposite .
it is important that we test the consistency of these two techniques . in order to do so
, we have added nearly 20,000 artificial clusters to our ngc 3627 images , of which roughly a quarter were recovered with @xmath42 magnitudes and sharpness values that met our selection criteria for clusters .
( the low recovery rate is because our input clusters included many that were saturated , too faint to be detected , or too small to be classified as clusters . )
we show the sharpness and roundness values for artificial clusters with core radii of 2.5 pc in figures [ fig_psffclus_envelope ] and [ fig_psffclus_round ] .
the relatively flat and sharp trends in figure [ fig_psffclus_envelope ] demonstrate that the hstphot sharpness value is principally a function of cluster radius .
the completeness , as a function of both input magnitude and core radius , is shown in figure [ fig_complete_pc ] .
we note that there are limited ranges of both cluster radii and magnitudes to which we are sensitive .
the radii are limited by the 1-pixel limit on the large end ( to minimize false detections of associations ) , and by the sharpness limit ( a function of magnitude ) on the small end .
the brightness range is restricted by our magnitude cut on one end and pixel saturation on the other .
the @xmath42 corrections and core radii from our analytical corrections , as functions of measured @xmath42 magnitude and sharpness , are shown ( for typical pc background levels ) in figures [ fig_sampledv ] and [ fig_samplesize ] , respectively .
the key feature of both of this figures is that both the magnitude correction and core radius are primarily functions of sharpness , allowing a relatively straightforward calculation of cluster properties using the measured magnitudes and sharpness values from hstphot .
combining these two approaches , we can determine integrated magnitudes of the recovered artificial clusters using the analytical corrections and compare the clusters true magnitudes to their measured magnitudes ; this is shown in figure [ fig_fakeclus ] .
we found that the @xmath42 magnitudes are well - recovered ( the mean error is @xmath80 , with rms scatter of 0.12 magnitudes ) , giving us confidence that this technique will produce reliable luminosity functions .
we then applied the analytical cluster measurement technique to our ngc 3627 objects , determining core radii and integrated magnitudes for the 553 candidate clusters .
we find that nine of the objects have recovered core radii greater than 1 pixel ; these are eliminated , as we do not trust our stellar psf - fitting photometry for objects larger than one pixel .
this cutoff also eliminates any small associations from the sample .
we also eliminated two objects that , from visual inspection , were not clusters ( the nucleus of a background galaxy and part of a diffraction spike of a bright star ) .
no other obvious false detections were present . because of the low contamination level measured ( @xmath81 false detections )
, we will henceforth refer to these 528 objects as `` clusters '' rather than `` candidate clusters '' . as a sanity check ,
we calculate the contribution of these clusters to the total light of ngc 3627 .
we measure the combined light of the clusters to have magnitudes of @xmath82 and @xmath83 . in comparison ,
the integrated light in the wfpc2 field is @xmath84 and @xmath85 , implying that the fractional contribution of the clusters is 1.8% in @xmath42 and 1.4% in @xmath55 . in order to compare with the results of larsen & richtler ( 2000 )
, we must count only the clusters with @xmath86 , which account for 1.3% of the @xmath42 luminosity and 1.0% of the @xmath55 luminosity .
while this value is higher than those of the larsen & richtler ( 2000 ) sample , we note that ngc 3627 is an interacting system and thus should be compared with the literature values they present for interacting systems .
comparing with these systems , we find the cluster light fraction we find in ngc 3627 to fall comfortably within the range of values they give ( @xmath42 fractions from 0.11% for ngc 3921 to 15% for ngc 3256 ) .
the properties of our detected clusters are shown in figures [ fig_cluscmd ] and [ fig_clussize ] .
it is clear from figure [ fig_cluscmd ] that unambiguous interpretation of these objects in terms of determining age and initial mass ( even if there is no differential reddening ) is not possible , as @xmath87 is not a monotonic function of cluster age .
rather , a cluster begins its life extremely blue , becomes redder when the first red supergiants form , moves slowly back to the blue , and eventually begins moving redward again .
additionally , for young clusters smaller than @xmath88 , the number of red supergiants at any given age is quite small , leading to a large intrinsic color dispersion for clusters of a single age .
finally , we note that differential extinction is clearly present , as we see red clusters preferentially in and near the dust lanes .
for example , clusters a heavily - extincted region on wfc3 near the wfc2 border have a median @xmath87 color of 1.21 , while those in a relatively unobscured region of wfc4 have a median color of @xmath89 .
if this color difference is caused solely by extinction , the reddening difference is @xmath90 , or @xmath91 .
( the cepheids presumably showed a smaller amount of differential extinction because they are much less luminous , and thus we are more biased against detecting highly - reddened cepheids than we are against detecting highly - reddened clusters . )
although one generally shows a cluster luminosity function , this is also difficult to interpret , as the rapid evolutionary dimming of a cluster makes it difficult to interpret the luminosity function in terms of cluster masses . between ages of 10 and 40 myr , for example , the cluster dims by over a factor of 10 ( the exact number depending , of course , on the assumed imf ) .
because of these difficulties in interpretation , the primary focus of our survey will be the measurement statistical properties of cluster populations of many systems , rather than an attempt to measure an accurate star formation history or cluster mass function of any one system .
the simplest population to isolate and study is the set of 13 clusters falling beyond the visual extent of the galaxy , which we define to be wfc4 objects with @xmath92 .
the positions and characteristics of these objects are shown in table [ tab_halo ] . interestingly , despite suffering little or no internal reddening in ngc 3627 , these clusters have a median color of @xmath93 , much _ redder _ than that of the other ngc 3627 clusters ( 0.69 ) .
we thus conclude that these are predominantly old globular clusters , a conclusion strengthened by their having mean absolute magnitude ( @xmath94 ; rms scatter of 1.0 magnitudes ) comparable to that of galactic globular clusters @xcite . very faint or small clusters
would not have met our detection criteria . without a statistical cmd - synthesis approach ( the subject of paper ii ) , the difficulties mentioned above , as well as incompleteness as a function of @xmath33 in addition to the usual @xmath42 and @xmath55 , mean that we can only make a semi - quantitative study of figures [ fig_cluscmd ] and [ fig_clussize ] .
we first note that there is a very strong correlation between positions of the clusters ( figures [ fig_chart0]-[fig_chart3 ] ) and the spiral structure of ngc 3627 .
the spiral arm passing through the pc and wfc4 , for example , contains 171 ( 32% ) of the detected clusters in just @xmath95% of the wfpc2 field of view . adding a ring in wfc2 just outside the bulge and an inner spiral arm on the wfc2/wfc3 border , we find that a total of 80% of the clusters are contained in 30% of the image .
the concentration is highest for blue ( @xmath96 ) clusters , of which 92% fall in these three regions .
the color ranges @xmath97 , @xmath98 , and @xmath99 all have 75% to 77% of the clusters in the three regions .
the simplest conclusion from these data is that the clusters with @xmath100 are a young population with minimal reddening , while the other clusters are a mixture of @xmath101% reddened young clusters and @xmath102% older clusters .
figure [ fig_lfall ] shows the raw luminosity functions for all clusters recovered ; figure [ fig_lfblue ] shows only the 126 clusters with @xmath103 . because of the shallower photometry limit in wfc2 we have split both diagrams into wfc2 and the other chips . using the deeper pc+wfc3/4 luminosity functions down to @xmath104 , we fit the luminosity functions to @xmath105 for all clusters , and @xmath106 for blue clusters only .
these fits correspond to luminosity function slopes of @xmath107 for all clusters and @xmath108 for blue clusters .
although these values are more shallow than those found in ngc 4038/9 by @xcite , they are similar to the @xcite slopes before completeness corrections .
completeness corrections , however , depend significantly on the assumed distribution of cluster sizes . using large ( @xmath109 pc @xmath110 pc ) clusters ,
we are essentially 100% complete below @xmath104 , thus leaving the lf slopes unchanged . using small ( @xmath111 pc @xmath112 pc ) clusters , we are only @xmath113 complete at @xmath104 , and thus the measured lf slope becomes steeper by 0.25 .
the actual value is likely somewhere in between ; we will adopt a preliminary lf slope of @xmath114 for this data set .
a detailed determination of the cluster initial mass function employing population synthesis techniques will be the subject of paper ii .
the question of modes of cluster formation is presently an open one .
while massive young clusters that are potentially the progenitors of globular clusters have been observed only in galaxies with very high star formation rates ( interacting and starburst systems ) , it is unclear whether this is caused by a different mode of star formation than that observed in our galaxy , or whether it is simply an increase in cluster formation rates of clusters of all masses . in order to answer this question ,
we have begun a survey of nearby galaxies in which we will study the cluster properties of each , as well as trends that appear as functions of galaxy properties . because robust photometric techniques for slightly - extended objects ( core radius smaller than the psf ) are not yet established , we have detailed the difficulties and methods that we have developed for this task .
although our techniques are optimized for the specific task of measuring slightly - extended clusters on top of a rapidly - varying background , we believe the principles used here can be applied to other similar problems .
specifically , our sky measurement algorithm can be used for any photometry ( stellar or extended - object ) on a rapidly - varying background , while our use of sharpness to measure the image shape can be used to psf determination as well as object size determination .
we have applied these techniques to ngc 3627 , for which deep observations ( 34 orbits in f555w and f814w ) are available in the hst archive .
we find that the stellar psf - fitting package hstphot works extremely well for our task , both in its ability to accurately photometer previously - detected cepheids and in its usefulness in discriminating and photometering clusters . for the cepheids , we recover 28 cepheids found in previous studies @xcite .
measuring their periods and mean magnitudes , we find a distance consistent with those measured by those authors , but inconsistent with the final value used by @xcite .
we have located 528 clusters in ngc 3627 , which account for @xmath115% of the galaxy s total light in @xmath42 and @xmath55 .
we find that the blue cluster population traces ngc 3627 s spiral pattern extremely closely , while redder clusters less so .
this is likely because of a population of halo globular clusters , seen clearly beyond the visual extent of the disk but likely appearing throughout the image as foreground objects .
red clusters are also preferentially seen in and near dust lanes ; we interpret these as reddened young clusters .
we measure a preliminary cluster luminosity function of @xmath114 ; a definitive result will come from the statistical analysis that will be presented in paper ii .
support for this work was provided by nasa through grant number ar-09196.01 from the space telescope science institute .
all of the data presented in this paper were obtained from the multimission archive at the space telescope science institute ( mast ) .
stsci is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
ashman , k. m. & zepf , s. e. 1992 , , 384 , 50 bresolin , f. & kennicutt , r. c. , jr .
1997 , , 113 , 975 caldwell , n. , kennicutt , r. c. , jr . , phillips , a. c. , & schommer , r. a. 1991 , , 370 , 526 christian , c. a. & schommer , r. a. 1982 , , 49 , 405 djorgovski , s. 1993 , in structure and dynamics of globular clusters , eds . s. djorgovski and g. meylan , aspcs , 50 , 373 dolphin , a. e. 2000a , , 112 , 1383 dolphin , a. e. 2000b , , 112 , 1397 dolphin , a. e. 2001 , in prep dolphin , a. e. , saha , a. , skillman , e. d. , tolstoy , e. , cole , a. a. , dohm - palmer , r. c. , gallagher , j. s. , mateo , m. , & hoessel , j. g. 2001 , , 550 , 554 elson , r. a. w. , fall , s. m. , & freeman , k. c. 1987 , , 323 , 54 freedman , w. l. , madore , b. f. , gibson , b. k. , ferrarese , l. , kelson , d. d. , sakai , s. , mould , j. r. , kennicutt , r. c. , jr . , ford , h. c. , graham , j. a. , huchra , j. p. , hughes , s. m. g. , illingworth , g. d. , macri , l. m. , & stetson , p. b. 2001 , , 553 , 47 gibson , b. k. , stetson , p. b. , freedman , w. l. , mould , j. r. , kennicutt , r. c. , jr . , huchra , j. p. , sakai , s. , graham , j. a. , fassett , c. i. , kelson , d. d. , ferrarese , l. , hughes , s. m. g. , illingworth , g. d. , macri , l. m. , madore , b. f. , sebo , k. m. , & silbermann , n. a. 2000 , , 529 , 723 girardi , l. , bressan , a. , bertelli , g. , & chiosi , c. 2000 , , 141 , 371 heiles , c. 1990 , , 354 , 483 hodge , p. w. 1961 , , 133 , 413 holtzman , j. a. , faber , s. m. , shaya , e. j. , lauer , t. r. , grothe , j. , hunter , d. a. , baum , w. a. , ewald , s. p. , hester , j. f. , light , r. m. , lynds , c. r. , oneil , e. j. , jr .
, & westphal , j. a. 1992 , , 103 , 691 ikeuchi , s. & norman , c. a. 1991 , , 375 , 479 kennicutt , r. c. , jr .
1988 , , 334 , 144 kennicutt , r. c. , jr . & chu , y .- h .
1988 , , 95 , 720 kennicutt , r. c. , jr . , edgar , b. k. , & hodge , p. w. 1989 , , 337 , 761 krist , j. 1993 , adass , 2 , 536 larsen , s. s. 1999 , , 139 , 393 larsen , s. s. & richtler , t. 2000 , , 354 , 836 mac low , m .-
& mccray , r. 1988 , , 324 , 776 madore , b. f. & freedman , w. l. 1991 , , 103 , 933 oconnell , r. w. , gallagher , j. s. , iii , & hunter , d. a. 1994 , , 433 , 65 saha , a. , sandage , a. , tammann , g. a. , labhardt , l. , macchetto , f. d. , & panagia , n. 1999 , , 522 , 802 schlegel , d. j. , finkbeiner , d. p. , & davis , m. 1998 , , 500 , 525 schweizer , f. , miller , b. w. , whitmore , b. c. , & fall , s. m. 1996 , , 112 , 1839 udalski , a. , szymanski , m. , kubiak , m. , pietrzynski , g. , soszynski , i. , wozniak , p. , & zebrun , k. 1999 , aca , 49 , 201 whitmore , b. c. , zhang , q. , leitherer , c. , fall , s. m. , schweizer , f. , & miller , b. w. 1999 , , 118 , 1551 williams , j. p. & mckee , c. f. 1997 , , 476 , 166 youngblood , a. j. & hunter , d. a. 1999 , , 519 , 55 zepf , s. e. , ashman , k. m. , english , j. , freeman , k. c. , & sharples , r. m. 1999 , , 118 , 752 llrrrr f555w-1 & pc1 & @xmath116 & @xmath117 & @xmath118 & @xmath119 + f555w-1 & wfc2 & @xmath120 & @xmath121 & @xmath122 & @xmath123 + f555w-1 & wfc3 & @xmath124 & @xmath125 & @xmath126 & @xmath127 + f555w-1 & wfc4 & @xmath128 & @xmath129 & @xmath130 & @xmath131 + f555w-2 & pc1 & @xmath132 & @xmath133 & @xmath134 & @xmath135 + f555w-2 & wfc2 & @xmath136 & @xmath137 & @xmath138 & @xmath139 + f555w-2 & wfc3 & @xmath140 & @xmath141 & @xmath142 & @xmath143 + f555w-2 & wfc4 & @xmath144 & @xmath145 & @xmath146 & @xmath147 + f814w-1 & pc1 & @xmath148 & @xmath149 & @xmath150 & @xmath151 + f814w-1 & wfc2 & @xmath152 & @xmath153 & @xmath154 & @xmath155 + f814w-1 & wfc3 & @xmath156 & @xmath157 & @xmath158 & @xmath159 + f814w-1 & wfc4 & @xmath160 & @xmath161 & @xmath162 & @xmath163 + f814w-2 & pc1 & @xmath164 & @xmath165 & @xmath166 & @xmath167 + f814w-2 & wfc2 & @xmath168 & @xmath169 & @xmath170 & @xmath171 + f814w-2 & wfc3 & @xmath172 & @xmath173 & @xmath174 & @xmath175 + f814w-2 & wfc4 & @xmath176 & @xmath177 & @xmath178 & @xmath179 + llrrrr c1-v7 & & @xmath180 & @xmath181 & @xmath182 & 4 + c1-v6 & & @xmath183 & @xmath184 & @xmath185 & 3 + & c15 & @xmath186 & @xmath187 & @xmath188 & 3 + c2-v34 & c23 & @xmath189 & @xmath190 & @xmath191 & 4 + c2-v4 & c04 & @xmath192 & @xmath193 & @xmath194 & 4 + c2-v33 & & @xmath195 & @xmath196 & @xmath197 & 4 + & c20 & @xmath198 & @xmath199 & @xmath200 & 3 + c2-v8 & c10 & @xmath201 & @xmath202 & @xmath203 & 3 + c2-v31 & c11 & @xmath204 & @xmath205 & @xmath206 & 2 + c2-v26 & c05 & @xmath207 & @xmath208 & @xmath209 & 3 + c2-v12 & & @xmath210 & @xmath211 & @xmath212 & 4 + c2-v21 & c19 & @xmath213 & @xmath214 & @xmath215 & 4 + c2-v9 & & @xmath216 & @xmath217 & @xmath218 & 2 + c2-v35 & & @xmath219 & @xmath220 & @xmath221 & 2 + c2-v32 & & @xmath222 & @xmath223 & @xmath224 & 2 + c2-v15 & & @xmath225 & @xmath226 & @xmath227 & 3 + c3-v10 & c24 & @xmath228 & @xmath229 & @xmath230 & 4 + c3-v3 & c28 & @xmath231 & @xmath232 & @xmath233 & 4 + c3-v6 & & @xmath234 & @xmath235 & @xmath236 & 4 + c3-v17 & c26 & @xmath237 & @xmath238 & @xmath239 & 3 + c3-v4 & & @xmath237 & @xmath240 & @xmath241 & 4 + & c32 & @xmath242 & @xmath243 & @xmath244 & 3 + c3-v12 & c30 & @xmath245 & @xmath246 & @xmath215 & 4 + & c36 & @xmath247 & @xmath248 & @xmath249 & 2 + c4-v9 & c34 & @xmath250 & @xmath251 & @xmath252 & 4 + c4-v13 & & @xmath253 & @xmath254 & @xmath255 & 4 + c4-v4 & & @xmath256 & @xmath257 & @xmath258 & 3 + c4-v10 & c33 & @xmath259 & @xmath260 & @xmath261 & 3 + rrrrrr wfc4 & 56.78 & 777.82 & @xmath262 & @xmath263 & @xmath264 + wfc4 & 86.17 & 607.53 & @xmath265 & @xmath266 & @xmath267 + wfc4 & 121.05 & 772.64 & @xmath268 & @xmath269 & @xmath270 + wfc4 & 150.78 & 473.85 & @xmath271 & @xmath272 & @xmath273 + wfc4 & 152.78 & 731.71 & @xmath274 & @xmath275 & @xmath276 + wfc4 & 172.78 & 532.45 & @xmath277 & @xmath278 & @xmath279 + wfc4 & 383.22 & 551.29 & @xmath280 & @xmath281 & @xmath282 + wfc4 & 393.84 & 771.25 & @xmath283 & @xmath284 & @xmath285 + wfc4 & 420.08 & 581.83 & @xmath286 & @xmath275 & @xmath287 + wfc4 & 428.31 & 567.79 & @xmath288 & @xmath289 & @xmath290 + wfc4 & 437.08 & 499.18 & @xmath291 & @xmath292 & @xmath264 + wfc4 & 532.17 & 569.17 & @xmath293 & @xmath294 & @xmath295 + wfc4 & 735.99 & 511.08 & @xmath296 & @xmath297 & @xmath298 +
|
we have developed photometric techniques that can be applied to images with highly - variable backgrounds , as well as to slightly - extended objects ( object size comparable to or smaller than psf size ) .
we have shown that ordinary stellar psf - fitting photometry can be applied to slightly - extended objects , provided that one applies a systematic correction to the photometry that is a function primarily of the observed sharpness . applying these techniques to the cepheid target ngc 3627 , we find that we are successfully able to photometer the stars and clusters , as well as discriminating the cluster population with a negligible number of false detections .
|
numerical analyses of quantum chromodynamics ( qcd ) on a discrete space - time lattice ( lattice qcd ) , indicate that the transition from confined to deconfined matter at finite temperature @xmath0 and vanishing quark chemical potential @xmath1 is a crossover @xcite .
on the other hand , model studies @xcite predict a first - order transition to occur for @xmath1 of the order of @xmath2 of the baryon mass and @xmath3 . in between these two regimes , a second - order critical point is expected in the @xmath4 plane at some intermediate values of @xmath0 and @xmath1 .
the existence and the exact location of the critical point is still a matter of dispute @xcite and has been under intense theoretical study using effective field theory models of qcd @xcite ( see also the recent analysis performed in ref .
unfortunately , a direct application of lattice qcd at finite @xmath1 is , at present , still quite problematic . only relatively recently , new theoretical developments and technical improvements allowed to circumvent in various ways the fermion determinant problem and start performing monte - carlo calculations ( see ref .
@xcite for a review ) .
although most of the results obtained up to now seem to support the qcd critical point , an interesting observation against its existence comes from refs .
@xcite where , from numerical simulations of qcd at imaginary chemical potential , one observes that the region of quark masses where the transition is presumably of the first order ( for quark masses smaller than the physical ones ) , tends to shrink for small positive values of the chemical potential , @xmath1 .
conversely , according to models supporting the critical point , the first order region should expand when @xmath1 increases , so that the physical quark mass point hits the critical line at some finite value of @xmath0 and @xmath1 .
a possible explanation for this discordance has been given in ref .
@xcite , where it was pointed out that a strong ( repulsive ) vector coupling may account for the initial shrinkage of the first order region , that would then start expanding again at larger values of @xmath1 . as a result , two critical points might appear for a given range of ( small ) quark masses , as argued in ref.@xcite .
however , it has been remarked @xcite that this does not necessarily imply the existence of the qcd critical point since a too strong repulsive potential may in fact provoke the disappearance of the first order line ( and thus of the critical point ) for physical quark masses .
if the vector coupling is too small , instead , the initial shrinkage of the first order region is not clearly seen .
a recent estimate @xcite of the vector coupling from flavor susceptibilities evaluated with lattice qcd seems to support the latter scenario .
based on the analysis of the linear sigma model ( l@xmath5 m ) with two flavor quarks , it was shown in refs .
@xcite that the inclusion of thermal fluctuations of the mesonic fields leads to the appearance of two critical points for a small finite vacuum pion mass , @xmath6 mev , without the need for a vector interaction . for physical values of the pion mass the model predicts only one critical point as one would naively expect for qcd .
also in this case , the initial shrinkage of the first order region at small @xmath1 was proved to be a not uncommon feature . however , as we will discuss , the direct transposition of these arguments to qcd should be done with special care . in the chiral limit , in fact , the phase diagram of the two flavor l@xmath5 m is divided in two parts by a continuous first order transition , whereas in qcd universality arguments @xcite suggest a second order phase transition of the o(4 ) universality class , at least at @xmath7 .
very recently , it was found ( see ref .
@xcite ) that the correct treatment of the fermion vacuum fluctuations ( that were neglected in most l@xmath5 m applications ) can change the order of the transition in the chiral limit from first to second order , depending on the coupling constants . in this case
the phase diagram of the l@xmath5 m would resemble the nambu jona - lasinio model ( njl ) one with a second order line starting at @xmath7 and high-@xmath0 and terminating at a tricritical point ( at intermediate @xmath0 and @xmath1 ) where the first order line starts ending at @xmath8 .
interestingly enough , these findings of ref .
@xcite suggest the possibility of a rich structure for the qcd phase diagram in a situation which is similar to the one which arises in metamagnetic systems whose phase diagram may display two critical points in the magnetic field versus temperature plane @xcite . a multicritical point structure induced by charge neutrality and
vector interaction has been recently discussed by zhang and kunihiro in the context of the 2 + 1 flavor njl model @xcite .
it is worth pointing out that the presence of strong magnetic fields ( @xmath9 ) , may change the order of the phase transition as shown by fraga and mizher @xcite who considered the two flavor l@xmath5 m , at @xmath7 , obtaining that the usual crossover can turn into a first order phase transition in this regime .
the aim of the present work is to explore in some details the phase diagram of two of the most important effective field theory models of qcd with two quark flavors represented by the l@xmath5 m and the njl model when small finite pion ( and quark current ) masses are considered .
the thermodynamics of both models has been compared , with standard parametrization in ref .
@xcite where the mfa has been used .
as we shall see , going beyond the mfa allows for the appearance of more than one critical point in both models for certain parameter values . for the l@xmath5 m we will use the same approximation of refs .
@xcite and will closely follow the methods therein to map the phase diagram for various values of the pion mass , @xmath10 .
as it will be shown , by varying @xmath10 , two and also three critical points ( two of them are actually very close to each other ) may appear . as a further step , we will explore the nature of these critical points by analyzing the susceptibilities and the correlations of the net - quark number density , the entropy density and the scalar density .
the second part of the paper will be dedicated to the njl model , in its simplest form , which will be treated in the so - called optimized perturbation theory ( opt ) .
the opt method ( which also goes by different names , or has many variants , e.g. `` delta - expansion '' @xcite , order - dependent mapping @xcite , etc . )
is well known for allowing evaluations beyond mfa due to the way it modifies ordinary perturbative expansion , giving a non trivial ( non - perturbative ) coupling dependence .
examples of successful applications include the most precise analytical values of the critical temperature for non interacting bose gases @xcite as well as the precise location of the tricritical point and the mixed liquid - gas phase within the gross - neveu model in 2 + 1 dimensions @xcite .
the latter illustrates how this method can be a powerful tool beyond standard perturbation theory , since these important effects were missed by the mfa and could not be precisely determined by monte carlo simulations @xcite . the opt version adopted here is mainly indicated to non - gauge theories , which ( at finite temperature ) require the method to be extended , e.g. by adding and subtracting a hard thermal loop improvement that modifies the propagators and vertices in a self - consistent way , in the so - called hard - thermal - loop perturbation theory ( htlpt ) @xcite . regarding its use within renormalizable theories
the opt has just been substantially improved by its combination with renormalization group properties @xcite .
this method has been recently applied to the njl model in the evaluation of the thermodynamical potential beyond mfa using standard parametrization @xcite .
the same type of application will be considered here , with a different set of parameters , in order to verify the possibility of multiple critical points as found in the l@xmath5 m .
our investigation , as already anticipated , shows that in the very strong coupling limit this situation ( which would be missed by the mfa ) arises .
it is interesting to remark that the opt brings @xmath11 corrections to the mfa effective potential which are proportional to the scalar density , @xmath12 , as well as to the net - quark number density , @xmath13 , whose contribution to the pressure goes as @xmath14 where @xmath15 is the usual njl coupling , @xmath16 is the number of colors and @xmath17 the number of flavors .
this means that a type of @xmath11 suppressed vector term whose strength is twice its scalar counterpart will contribute to the pressure so that when the interaction is sufficiently strong the results seem to support the findings of ref .
@xcite where the su(3 ) njl with an explicit repulsive vector interaction , such as the one suggested in ref .
@xcite , was used .
we recall that , contrary to the l@xmath5 m , the njl is a non renormalizable theory which is often regularized by a non covariant ultra violet cut off , @xmath18 .
then , from the quantitative point of view our whole njl application must be taken with care since to generate exotic phase diagrams similar to the l@xmath5 m one needs very high values for @xmath15 which in turn generate high effective quark masses at zero temperature and density .
although the effective quark mass value generated by those parameters becomes larger than @xmath18 , the values of relevant observables , such as the quark condensate and the pion decay constant , remain well within reasonable values . also , as already emphasized , one of the goals of the present njl application is to check whether this model , like the l@xmath5 m , supports the existence of more than one critical point in its phase diagram .
after obtaining the opt effective potential ( or free energy density ) we derive the pressure and many thermodynamical quantities of interest including susceptibilities and critical exponents .
the results obtained with both models indicate that the first order line observed at low chemical potential and high temperature has a more `` chiral '' character while its low temperature and high chemical potential counterpart displays characteristics typical of a `` liquid - gas '' phase transition .
finally , it will be shown that with both models our results seem to support the back - bending of the critical line in the @xmath19 plane ( where @xmath20 is the quark current mass ) which , as first discussed in refs .
@xcite , could reconcile the actual lattice findings with most models predictions .
it is also shown that the mfa completely misses the possibility of more than one critical point in the @xmath4 plane , for both models .
the paper is organized as follows . in the next section
the l@xmath5 m is reviewed with the inclusion of thermal fluctuations .
after obtaining the phase diagram in the @xmath4 plane for small pion masses the characteristics of each different critical point is examined by a careful analysis of the densities and susceptibilities . in sec .
iii the recent opt application @xcite to the njl model is quickly reviewed .
we then obtain a set of parameter values which leads to the emergence of a second critical point , analogous to the one found in the l@xmath5 m .
we perform a comprehensive thermodynamical analysis to investigate how multiple critical points appear in the @xmath4 plane as well as in phase coexistence diagrams in the @xmath21 and @xmath22 planes .
we also numerically investigate the behavior of quantities such as the interaction measure , the equation of state parameter , bulk viscosity , and susceptibilities at , and around , each critical end point .
the latter quantities allow us to estimate some relevant critical exponents in order to distinguish the physical nature of both critical points which is also done by considering the free energy in terms of two ordering densities , @xmath23 and @xmath24 .
finally , in sec iv we present our main conclusions .
in standard notation , the density lagrangian of the l@xmath5 m with quarks reads @xmath25\psi \ ; , \label{lsm1}\ ] ] where @xmath26 is the flavor isodoublet spinor representing the quarks ( @xmath27 and @xmath28 ) , and @xmath29 is the classical potential energy density . in the chiral limit
( obtained by setting @xmath30 in the previous equation ) the chiral symmetry @xmath31 is spontaneously broken at the classical level , and the pion is the associated massless goldston boson . for @xmath32 , the chiral symmetry is explicitly broken by the term @xmath33 in eq .
( [ lsm2 ] ) which gives the pion a finite mass at @xmath8 and @xmath7 .
the scalar field @xmath5 has a finite vacuum expectation value @xmath34 determined by the classical equation of motion @xmath35 accordingly , the @xmath5 field is conveniently expressed as a sum of the condensate plus fluctuations , @xmath36 . in the l@xmath5 m lagrangian given by eq .
( [ lsm1 ] ) there is no explicit mass term for the quark field , the quark mass being given only by the condensate , @xmath37 .
the parameters of the model are given by the set of equations @xmath38 where @xmath39 mev is the pion decay constant .
consistently with refs .
@xcite we set the vacuum @xmath5 mass to @xmath40 mev and the quark mass to @xmath41 mev ( i.e. one - third of the nucleon mass ) .
the last parameter that one needs to fix is the vacuum pion mass , @xmath42 , which will be varied from @xmath43 mev to @xmath44 mev . to calculate the equation of state of the model
we closely follow the self - consistent method first proposed in ref .
@xcite and used also in ref .
accordingly , we write the grand canonical partition function as a functional integral in the euclidean space with the imaginary time @xmath45 @xmath46=\int \mathcal{d}\psi \mathcal{d}{\bar \psi } \mathcal{d}\sigma \mathcal{d}{\boldsymbol{\pi}}\exp\left\{\int_{0}^{\beta } \d \tau \int_{v } \d^3 x \left(\mathcal{l}+\mu { \bar \psi}\gamma_0 \psi \right)\right\ } \ ; , \label{lsm5}\ ] ] where @xmath47 and @xmath48 are the hamiltonian and the net quark number operator , respectively . in eq .
( [ lsm5 ] ) , @xmath49 is the inverse temperature , @xmath1 is the quark chemical potential .
the baryochemical potential being @xmath50 . ] and @xmath51 is the system volume .
following the same steps as in refs .
@xcite , we now integrate away the quark degrees of freedom .
this amounts to calculate the partition function of the quark sector @xmath52 where @xmath53 the gaussian integral in eq .
( [ lsm5.0 ] ) can be solved with standard techniques @xcite and yields @xmath54 in ref .
@xcite an expression such as eq .
( [ lsm5.1 ] ) has been expanded in a series of commutators involving the derivatives of the mesonic fields . in our analysis
we will discard these terms , implicitly assuming that the meson mode amplitudes vary slowly in space and time ( the same approximation was introduced also in refs .
the determinant in eq .
( [ lsm5.1 ] ) can then be evaluated as for the free case . after performing a fourier transformation in momentum - frequency space and using the property @xmath55
we finally obtain @xmath56 \right\ } + \ln\left\{\beta^2\left[\omega^2_{n } + \left ( \varepsilon + \mu \right)^2\right ] \right\}\ ; , \label{lsm5.3}\ ] ] where @xmath57 , @xmath58 , and @xmath59 are the matsubara frequencies taking the values @xmath60 because of the antiperiodicity condition on the fermionic functional integral @xmath61 . in eq . ( [ lsm5.3 ] )
@xmath62 is the energy , with the quark effective mass @xmath63 we note that eq .
( [ lsm5.3 ] ) is formally identical to the standard result except for the dependence of @xmath64 on the mesonic fields .
performing the summation over the matsubara frequencies in eq .
( [ lsm5.3 ] ) one then obtains @xmath65 + \ln\left[1+\e^{-\beta\left(\varepsilon+\mu\right)}\right ] \right\ } \ ; . \label{lsm8}\end{aligned}\ ] ] since we are interested in the low - energy properties of the model , we will ignore ( as was done in @xcite ) , for simplicity , the shift in the zero point energy .
the effect of such a contribution at finite temperature has been analyzed in ref .
@xcite for @xmath66 mev . for physical values of the pion mass , this is not expected to change the qualitative behavior of the model . as was pointed out in the introduction ,
however , the effect of the fermion vacuum loop , could actually play an important role , changing the order of the transition in the chiral limit at @xmath7 from first to second order @xcite .
these contributions are taken into account in the njl model where they are responsible of the dynamical breaking of the chiral symmetry . using the result in eq .
( [ lsm8 ] ) we can now write an effective lagrangian that includes only the mesonic degrees of freedom @xmath67 where @xmath68 the euler - lagrange equations then read : @xmath69 we now proceed linearizing the mesonic action by taking the average @xmath70 of the equations of motion over the field fluctuations
. first we decompose , as before , the @xmath5 field as @xmath36 where @xmath71 and @xmath72 is the fluctuation . of course ,
@xmath73 when @xmath74 is odd , and the same is true for @xmath75 . therefore , since the pion fluctuations always occur as @xmath76 , the average @xmath77 , whereas from the first one of the eqs .
( [ lsm11 ] ) we get the condition for the condensate @xmath78 the meson masses are then identified with the average of the second derivative of the effective potential @xmath79 and the effective potential is linearized as @xmath80 the two terms @xmath81 and @xmath82 on the right - hand side of the last equation are the mass terms to be added to the kinetic energy in the mesonic lagrangian to give the mesonic partition function @xmath83\right\ } \nonumber \\ & = & \exp\left[-\frac{v}{t}\left(\omega_{\sigma}+\omega_{\pi}\right ) \right ] \ ; , \label{lsm5m}\end{aligned}\ ] ] where @xmath84\,\,,\nonumber \\ \omega_{\pi}&= & \frac{3 t}{2 \pi^2 } \int \d p \ ;
p^2\left[\frac{1}{2 } \beta \varepsilon_{\pi } + \ln\left(1-\e^{-\beta\varepsilon_{\pi}}\right ) \right]\ ; \label{lsmq}\end{aligned}\ ] ] and @xmath85 stemming from eq .
( [ lsm5 m ] ) , we have two self - consistency relations between the meson masses and the corresponding fluctuations @xmath86 finally , we write the thermodynamic potential density as and @xmath76 . ] @xmath87 the average over the field fluctuations is performed by using the techniques of refs .
given an arbitrary functional @xmath88 we can write down the taylor expansion around @xmath89 and take the average term by term @xmath90 by using the relation derived in ref .
@xcite , we then relate the terms @xmath91 to powers of @xmath92 and @xmath93 , i.e. @xmath94 and @xmath95 , which amounts of summing up the infinite series of daisy and superdaisy diagrams in the hartree approximation .
the resulting expression , turns out to be equivalent to an integration over a gaussian distribution @xcite @xmath96 where @xmath97 in the following , we will need to perform derivatives of the average value of some quantities such as the thermodynamic potential density , @xmath98 .
after two integration by parts , using eqs .
( [ aver2 ] ) and eqs .
( [ aver3 ] ) one can obtain the following useful relation @xcite @xmath99 with eq .
( [ average ] ) , and the eqs .
( [ eq0 ] ) and ( [ eq1 ] ) it is not difficult to see that @xmath100 in a similar way , using the eqs .
( [ eq2 ] ) one can also easily show that @xmath101 eqs .
( [ lsm12 ] ) and ( [ lsm13 ] ) guarantee the consistency of the approach and the standard connection between thermodynamic and statistical mechanics @xcite .
the equation of state of the system is given by the simultaneous solutions in the variables @xmath102 of eqs .
( [ eq0 ] ) and ( [ eq1 ] ) , with the field fluctuations given by eq .
( [ lsm13 ] ) . in ref .
@xcite this was done numerically and the authors found a rich phase structure with one or two critical points , depending on the value of vacuum pion mass @xmath103 . in the next subsection
, we will repeat the same calculation as in ref .
@xcite and we will show that another critical point ( very difficult to detect ) appears in the phase diagram .
we now proceed and map the phase diagram of the model for various values of the vacuum pion mass ranging from @xmath104 mev to @xmath105 mev . as a first step ,
we need to find a way to localize the first order line(s ) and the critical end point(s ) .
this is usually done by looking at the change in shape of the thermodynamic potential across a transition line .
this method ( that will be adopted for the njl model in sec . [ secnjl ] ) , allows to see quite clearly the onset ( and usually the order ) of a phase transition , but unfortunately can not be used for the present analysis .
our approximation is , in fact , based on an expansion of the thermodynamic potential around a minimum and our equations are defined only there so that a different method must be employed .
a typical signature of the onset of a first order transition is the presence of two minima of the thermodynamic potential , corresponding to the high and low temperature phases . in turn
, we should see two distinct solutions for @xmath102 at the same @xmath0 and @xmath1 .
the transition happens when these two solutions have the same potential ( pressure ) , i.e. the point when the system switches from one minimum to the other . for each value of @xmath1
, we solve the system of equations starting from low temperatures ( the broken phase ) and following the line of minima increasing the temperature by a small amount @xmath106 at each step , using the latest solution as initial point for the solving routine .
once we are sure to be in the high @xmath0 symmetric phase , we solve the system of equations backwards ( going from high to small @xmath0 ) until we are sure to be below the transition temperature .
if the transition is continuous , we will find a unique solution for each value of @xmath0 .
conversely , if the transition is of the first order , we will have a region where two solutions exist at the same @xmath0 . in fig .
[ dsol ] this is shown for the condensate @xmath34 around the transition line for @xmath107 mev . for this value of the pion mass ,
the system undergoes a first order transition at @xmath108 mev and @xmath7 .
as one can see , the solution corresponding to the symmetry - broken phase ( @xmath109 ) exists at low @xmath0 and shortly above the transition temperature , whereas the symmetric solution ( @xmath110 ) exists at high @xmath0 and below the transition . the actual solution ( the one corresponding to the bigger pressure ) is shown by black dots and the dashed line . with this method , we now map the phase diagram of the model in the @xmath4 plane .
our results are consistent with those in ref .
@xcite , where the authors performed the same calculation and found that for sufficiently small values of the vacuum pion mass ( @xmath111 mev ) , the system has two critical points . in fig .
[ f3 ] we plot the phase digram of the model in the mfa ( fig . [ f3]a ) and with the inclusion of mesonic fluctuations ( fig .
[ f3]b ) . in fig .
[ f3]b we have analyzed various values of @xmath112 mev . for @xmath113 mev the phase diagram
is divided in two parts by a continuous first order line .
for @xmath107 mev and @xmath114 mev the first order line is interrupted by a continuous region , and for some @xmath115 mev the branch at low @xmath1 disappears and we have the usual continuous transition at low @xmath1 and first order at high @xmath1 .
the leftmost branch of the first order line ( when it exists ) ends in a critical point . for the rightmost one , instead
, the situation seems to be different .
a closer look revealed the presence of a double critical point .
this is shown in fig .
[ f4]a for @xmath116 mev . as one can see , the first order line bifurcates and ends in _ two _ critical points that have been labeled @xmath117 and @xmath118 .
their existence is revealed by the presence of two first order transitions that have been detected , as before , looking at the double solutions of our system of equations . in fig .
[ f4]b we show the parameter @xmath34 along the dashed line crossing the two first order lines in fig .
as one can see , there are two distinct solutions ( that we have checked to correspond to minima of the thermodynamic potential ) for two slightly different ( @xmath119 mev ) values of the temperature , indicating a double first order .
a similar behavior can be observed for all the explored values of the pion mass , i.e. @xmath120 mev .
also for @xmath121 mev , where the transition is of the first order everywhere , a very short appendix of the the first order line appears around @xmath122 mev and @xmath123 mev producing another very short branch ending in a critical point at @xmath124 mev and @xmath125 mev .
the situation is qualitatively similar to the one in fig .
[ f4]a except for the fact that the leftmost critical point @xmath117 does not exist as the first order line continues until @xmath7 . in fig .
[ ufine ] one can see the multiple solutions for @xmath34 , indicating the two first order lines at @xmath126 mev and @xmath127 mev , respectively . for the rightmost transition ,
the discontinuity in @xmath34 is very small , but its presence is proved by the existence of the metastable solution . at present the reason for the this unusual bifurcation of the high-@xmath1 first order line is not understood and requires further studies .
it may simply be an artifact of the approximation ( mean - field plus fluctuations ) used for the treatment of the linear sigma model .
however , as we will see in the next section , despite their vicinity in the phase digram , the two first order lines appear to have slightly different qualitative features .
we now go one step further and analyze the susceptibilities in the l@xmath5 m with the aim of characterizing the critical point(s ) by studying the fluctuations of the net - quark number density @xmath24 , scalar density @xmath128 , and entropy density @xmath129 in the vicinity of the critical regions . by definition ,
the quark density is @xmath130 by using the eq .
( [ average ] ) and eqs .
( [ lsm12 ] ) and ( [ lsm13 ] ) one then has @xmath131 to calculate the scalar density , it is convenient to introduce a mass term @xmath132 for the quarks in the lagrangian in eq .
( [ lsm1 ] ) , where @xmath133 is a fictitious bare quark mass to be set to zero afterwards .
this amounts to a thermal quark mass @xmath134 from the partition function written as in eq .
( [ lsm5 ] ) one then sees that @xmath135 that is , as before , @xmath136 unlike the quark density in eq .
( [ qdens2 ] ) , the scalar density can be directly evaluated , once @xmath34 and the meson masses are known , without the need to solve any further integral . indeed , from eq .
( [ eq0 ] ) one gets @xmath137 noting that @xmath138 and using eqs .
( [ sdens2 ] ) and ( [ sdens3 ] ) we finally obtain @xmath139 \ ; . \label{sdens5}\ ] ] the net - quark number density and the scalar density are the thermodynamic variables conjugate to @xmath1 and the fictitious quark mass @xmath133 , respectively . the last density that we need to evaluate is the the entropy density ( conjugate to the temperature @xmath0 ) , i.e. @xmath140 in what follows , we will always assume the same number of @xmath27 and @xmath28 quarks in the system .
the isospin density is therefore always vanishing for any value of @xmath0 , @xmath1 and @xmath133 , and , of course , the same is true for its derivatives with respect to these quantities . the isospin density ,
therefore , do not mix with the other degrees of freedom ( @xmath24 , @xmath23 and @xmath129 ) and will be not considered in our analysis at this time .
we can now go ahead and calculate the @xmath141 covariance matrix as @xmath142 this matrix is symmetric and can be diagonalized .
we call @xmath143 the three eigenvalues of @xmath144 and @xmath145 the corresponding eigenvectors .
each one of the three eigenvectors can be related to a different orthogonal combination of the three original densities @xmath146 and @xmath129 @xmath147 these three new densities are now independent . at the critical point
, only one of them will have divergent fluctuations , whereas the other two will remain finite .
we chose , as a convention , to order the eigenvalues from the biggest @xmath148 to the smallest @xmath149 . at the critical point
, we then expect the first eigenvalue , @xmath148 , to diverge . according to eq .
( [ divdens ] ) , the three components of the corresponding eigenvector @xmath150 will then give us the expression of the critical density in terms of the original densities @xmath146 and @xmath129 .
it s worth to mention that in our analysis , the correlations between @xmath24 and @xmath129 are positive , whereas the correlations between @xmath23 and @xmath129 , and @xmath23 and @xmath24 are negative .
this is due to the fact that , at the transition , the thermal contribution to the scalar density drops from its maximum value ( in the broken phase ) to a very small value ( exactly zero in the chiral limit ) in the symmetric phase ( see ref .
@xcite ) , whereas both @xmath129 and @xmath24 exhibit the opposite behavior going from smaller values ( in the broken phase ) to higher values ( in the symmetric phase ) . unlike @xmath129 and @xmath24
the scalar density is , in fact , dominated by the zero - point contributions . with their inclusion one recover the physically expected behavior @xcite .
one must then bear in mind that the following results are relevant for the thermal part of the model .
our general conclusions , however , are in agreement with the results obtained with the njl in sec . [ secnjl ] where the zero - point contributions are included .
we analyze two different values for the vacuum pion mass : @xmath116 mev ( figs .
[ s1],[s2 ] and [ s3 ] ) and @xmath151 mev ( figs .
[ s4 ] and [ s5 ] ) . for clarity ,
it is convenient to chose a label for the various critical points .
we will call @xmath152 the critical endpoint of the leftmost first order line ( if there is any ) that starts at @xmath7 and @xmath153 mev in the phase diagram . with @xmath117 and @xmath118
we will refer to the two critical endpoints of the first order line ( that ultimately splits in two ) that starts at @xmath8 and @xmath154 mev , @xmath118 being the rightmost of the two as in fig . [ f4 ] . as we have seen in the last section , for @xmath113 mev
there is only the critical point @xmath118 .
we will not consider that case here , however .
we begin with @xmath152 for @xmath116 mev . in fig .
[ s1 ] , panels ( a - b ) and ( c - d ) we show the three diagonal components of the covariance matrix ( @xmath155 , @xmath156 , @xmath157 ) , and the three off - diagonal components ( @xmath158 , @xmath159 , @xmath160 ) for @xmath161 mev and @xmath162 mev . for @xmath161 mev ,
the system undergoes a first order phase transition at @xmath163 mev , whereas for @xmath162 mev the transition is continuous ( but still very sharp ) and takes place at @xmath164 mev . in between
, one finds the unusual critical point @xmath152 . by looking at the elements of the covariance matrix for these two values of
@xmath1 one sees that the dominant fluctuations are given by the entropy density and the scalar density .
this is shown in fig .
[ s1]a and [ s1]b , where the dominant diagonal terms in the vicinity of the transition are @xmath165 , @xmath166 . from the off - diagonal terms in fig .
[ s1]c and [ s1]d one notices that the scalar and the entropy density are the most strongly anticorrelated ( the magnitude of @xmath167 is much larger than the other off - diagonal coefficients ) . this fact can be observed in a more direct way by looking at the eigenvalue @xmath148 and the corresponding eigenvector @xmath168 . in fig .
[ s1]e and [ s1]f the three eigenvalues of the covariance matrix are plotted for the same range of values of @xmath0 and @xmath1 .
the largest eigenvalue @xmath148 is the only one showing a peak ( it should actually diverge at the critical point ) . the eigenvector @xmath168 ( see eq .
( [ divdens ] ) ) gives us the three components of the critical density @xmath169 .
as one can see , from fig .
[ s1]g and [ s1]h , the critical density @xmath170 is a mixture of almost solely the scalar and the entropy density ( the first component @xmath171 does not appear in the plot as it is always very close to zero ) .
as we have already discussed , in addition to this critical point the system exhibits two more critical points ( @xmath117 and @xmath118 ) in the high-@xmath1 region ( see fig .
the analysis of these two points is shown in fig .
[ s2 ] and fig .
[ s3 ] for @xmath172 mev . for @xmath173 mev ( panels a and e in fig .
[ s2 ] and [ s3 ] ) the transition is continuous , but one can already clearly distinguish that the two peaks associated with the two critical points have already developed . the first one on the left ( the one corresponding to @xmath117 ) is very sharp , while the other ( a little bit on the right , corresponding to @xmath118 ) looks still like a bump . for @xmath174 mev
( panels b and f in fig .
[ s2 ] and [ s3 ] ) we have a first order and a continuous transition .
the first peak on the left is now a discontinuity , whereas the bump becomes sharper as the critical point @xmath118 is approached . for @xmath175 mev
we have a double first order line ( panels c and g in fig . [ s2 ] and [ s3 ] ) and for @xmath176 mev the two first order lines have merged together to form a single transition line .
looking at the diagonal terms of the covariance matrix in the upper panels of fig . [ s2 ] one immediately sees that the density @xmath24 is now playing a relevant role and its fluctuations are of the same order of those of the scalar density .
the off - diagonal terms ( lower panels in fig . [ s2 ] ) show a positive correlation between @xmath24 and @xmath129 and a negative correlation between @xmath23 and @xmath129 , and @xmath23 and @xmath24 .
the two critical points @xmath117 and @xmath118 , indeed , show different features . from the analysis of the diagonal matrix ,
one sees ( lower panels in fig .
[ s3 ] ) that the critical density @xmath170 is in both cases dominated by the entropy density .
however , for the point @xmath117 ( see fig .
[ s3 ] panels a and e ) the scalar component is rather small ( fig .
[ s3]e ) , and overcomes the net - quark number density component only at a temperature immediately higher than the one corresponding to the peak of the eigenvalue @xmath148 ( compare to fig .
[ s3]a ) . for the critical point @xmath118 , instead
, the critical density @xmath170 is @xmath177 entropy density and the remnant @xmath178 is almost equally split between net - quark and scalar density ( see fig . [ s3]f ) .
the same analysis has been repeated for @xmath151 mev in figs .
[ s4 ] and [ s5 ] in the vicinity of the two critical points @xmath117 and @xmath118 ( the point @xmath152 does not exist ) for @xmath179 mev . once again , for @xmath180 mev the transition is continuous , whereas for @xmath181 mev we have a first order and a continuous transition ( corresponding to the rightmost spike ) , for @xmath182 mev we have a double first order line and finally , for @xmath183 mev the two first order lines have merged together to form a single one . by looking at the diagonal terms of the covariance matrix ( upper panels in fig .
[ s4 ] ) one can see that now the fluctuations of both critical points are dominated by the net - quark density and the entropy density .
in addition , one notices that in all the lower panels of fig .
[ s4 ] , the correlation of the scalar density with the entropy density and the net - quark density are almost identical .
even though in this case the scalar density plays only a minor role , it behaves in an even more visibly different way in the two critical points @xmath117 and @xmath118 . in fig .
[ s4 ] ( panels a and e ) at @xmath184 mev we are very close to the critical point @xmath117 . in correspondence to this temperature , the scalar fluctuations @xmath156 are suddenly suppressed . at the same time , the correlations between @xmath128 and @xmath24 ( @xmath158 ) and the correlations between @xmath128 and @xmath129 ( @xmath160 ) cross zero ( this can be seen , with some difficulty , in fig . [ s4]e ) .
indeed , in correspondence to the leftmost critical point @xmath117 , the scalar density seems to behave as a `` spectator '' of the critical phenomenon as it does not mix with the other densities .
this is not the case for the critical point @xmath118 . even if smaller than the entropy and net - quark number density fluctuations ,
now the scalar density fluctuations show a peak and the correlations @xmath158 and @xmath160 are negative . increasing the mass of the pion the critical points @xmath117 and @xmath118 move to the right of the phase digram towards higher values of the chemical potential and lower temperature . as a result ,
the fluctuations of the net - quark number density becomes more important .
in contrast , the critical point @xmath152 ( when it exists ) , has only a minor component of the net - quark number density , and the transition is dominated by the entropy density and the scalar density , reflecting a more `` chiral '' behavior .
this is in part due to the fact that the critical point @xmath152 exists only when the vacuum pion mass is small , i.e. when the chiral symmetry is only slightly ( explicitly ) broken .
let us now consider the standard version of the two flavor njl model which is described by @xcite @xmath185 \psi + g\left [ ( { \bar \psi}\psi)^{2}+({\bar{\psi } } i\gamma _ { 5}{{\boldsymbol{\tau}}}\psi ) ^{2}\right ] , \label{njl2}\ ] ] where @xmath26 ( a sum over flavors and color degrees of freedom is implicit ) represents a flavor isodoublet ( u and d type of quarks ) @xmath186-plet quark fields while @xmath187 are isospin pauli matrices .
the lagrangian density ( [ njl2 ] ) is invariant under ( global ) @xmath188 and , when @xmath189 , the theory is also invariant under chiral @xmath190 groups .
note that , as emphasized in refs .
@xcite , the introduction of a vector interaction term of the form @xmath191 in eq .
( [ njl2 ] ) is also allowed by the chiral symmetry and that such a term can become important at finite densities , generating a saturation mechanism depending on the vector coupling strength that provides better matter stability . regarding analytic nonperturbative evaluations
, one can consider one loop contributions dressed up by a fermionic propagator , whose effective mass is determined in a self - consistent way .
this approximation is known under different names , e.g. , hartree , large-@xmath16 or mean - field approximations ( mfa ) . to obtain the effective potential ( or landau free energy density ) for the quarks , @xmath192 , it is convenient to consider the bosonized version of the njl , which is easily obtained by introducing auxiliary fields ( @xmath193 ) through a hubbard - stratonovich transformation . then , to introduce the auxiliary bosonic fields and to render our results more suitable to compare with the large-@xmath186 approximation it is convenient to use @xmath194 and to formally treat @xmath186 as a large number , which is set to the relevant value , @xmath195 , at the end of the evaluations
one then has @xmath196 the euler - lagrangian equations show that @xmath197 and @xmath198 .
the basic idea of the opt method is to deform the original lagrangian density by adding a quadratic term like @xmath199 to the original lagrangian density as well as multiplying all coupling constants by @xmath200 .
the new parameter @xmath200 is just a bookkeeping label and @xmath133 represents an _ arbitrary _ mass parameter .
used here has as completely different role than the one used previously in the l@xmath5 m . ]
perturbative calculations are then performed in powers of the dummy parameter @xmath200 which is formally treated as small and set to the original value , @xmath201 , at the end on performs an expansion in powers of @xmath11 where @xmath16 is formally treated as large but set to the original value ( @xmath58 in our case ) at the end . ] .
therefore , the fermionic propagators is dressed by @xmath133 which may also be viewed as an infra red regulator in the case of massless theories .
after a physical quantity , such as @xmath192 , is evaluated to the order-@xmath202 and @xmath200 set to the unity a residual @xmath133 dependence remains .
then , optimal non perturbative results can be obtained by requiring that @xmath203 be evaluated where it is less sensitive to variations of the arbitrary mass parameter .
this requirement translates into the criterion known as the principle of minimal sensitivity ( pms ) @xcite @xmath204 in general , the solution to this equation implies in self consistent relations generating a non perturbative @xmath15 dependence . in most cases
non perturbative @xmath11 corrections appear already at the first non trivial order while the mfa results can be recovered at any time simply by considering @xmath205 .
finally , note that the opt has the same spirit as the hartree and hartree - fock approximation in which one also adds and subtracts a mass term .
however , within these two traditional approximations the topology of the dressing is fixed from the start : direct ( tadpole ) terms for hartree and direct plus exchange terms for hartree - fock . on the other hand , within the opt
@xmath206 acquires characteristics which change order by order progressively incorporating direct , exchange , vertex corrections , etc
. effects . to implement the opt within the njl model one follows the prescription used in refs .
@xcite by first interpolating the original four - fermion version .
then , in terms of the auxiliary fields , the deformed lagrangian density becomes @xmath207 { \psi } -\delta \frac{n_{c } } { 2\lambda } \left ( \sigma ^{2}+{{\boldsymbol{\pi}}}^{2}\right ) \ , , \label{delta12}\ ] ] which shows that the yukawa vertices have weight @xmath200 while the meson `` propagators '' are proportional to @xmath208 .
one is then ready to perform a _ perturbative _ evaluation of landau s free energy in powers of @xmath200 . in the @xmath5 direction , the first non trivial order
the relevant contributions are represented in fig .
[ effpot ] .
then , at finite temperature and finite chemical potential , the order-@xmath200 result for the njl free energy density is ( see ref .
@xcite for a more detailed discussion ) @xmath209 where we have replaced @xmath210 . in the above equation we have defined , for convenience , the following basic relevant integrals : @xmath211 + t\ln \left [ 1+e^{-\left ( \varepsilon-\mu \right ) /t}\right ] \right\ } \;,\ ] ] @xmath212\;,\ ] ] and @xmath213\ ; , \label{basicint}\ ] ] where @xmath214 .
notice also that @xmath215 only survives at @xmath216 . here
, we impose a sharp non - covariant cut - off , @xmath18 , only for the vacuum term , since the finite temperature has a natural cut - off in itself specified by the temperature .
this choice of regularization , which allows for the stefan - boltzmann limit to be reproduced at high temperatures , is sometimes preferred in the literature @xcite .
moreover , in the present application it assures that the temperature integrals appearing in the l@xmath5 m and in the njl are integrated over the same momentum range .
also , as will be further discussed , this regularization choice appears to be a crucial condition in order for the njl model to reproduce the phase diagram with two critical points .
the divergent integrals occurring at @xmath8 and @xmath7 are @xmath217 - 2\sqrt{\lambda^{2}+(\eta+m_c)^{2}}\left[2\lambda^{3}+ \lambda ( \eta + m_c)^{2}\right ] \frac{}{}\right \}\ , , \label{i1div}\end{aligned}\ ] ] and @xmath218^{2}}{(\eta+m_c)^{2}}\right ] \right\}\;. \label{i2div}\ ] ] if needed , the @xmath219 limit of those integrals can be readily obtained ( see ref .
then , by applying the pms relation to @xmath192 one gets @xmath220 \left [ 1 + ( \eta+m_c ) \frac { d}{d \eta } \right ] i_2 + 4 g \ , i_3 \frac{d}{d \eta } i_3 \right\}_{\eta = { \bar \eta}}=0 \ , . \label{simplepms}\ ] ] notice that if one ignores the terms proportional to @xmath15 ( which are of order @xmath11 ) the optimal result is simply @xmath221 . in this situation , eq .
( [ landauenergy ] ) shows that the mfa result is exactly reproduced .
since we are mainly interested in the thermodynamics , one basic quantity of interest is the thermodynamical potential , @xmath98 , whose relation to the the free energy is given by @xmath222 .
the order parameter , @xmath223 is determined from the gap equation generated by minimizing @xmath224 with respect to @xmath5 . from eq .
( [ landauenergy ] ) we obtain @xcite @xmath225 in order to further discuss the relation of some of our results with those obtained by fukushima @xcite it is interesting to note , at this stage , that since @xmath226 one can write the optimum thermodynamical potential as @xmath227 where @xmath228 has the mathematical structure of a mfa like thermodynamical potential whose effective mass given by @xmath206 . since @xmath229 one sees that the opt introduces a correction like @xmath230 which ,
although suppressed by @xmath11 , is of the same form as the one considered by the mfa when a vector interaction term like @xmath231 , as proposed in ref .
@xcite , is added to the original njl lagrangian density .
when @xmath205 , @xmath232 and the mfa result ( for the standard njl model ) with no @xmath24 dependence is recovered . usually , the three njl parameters , @xmath15 , @xmath20 and @xmath18 are fixed by fitting the pion decay constant , @xmath233 , the quark condensate @xmath234 $ \sim$}}~}-\langle { \bar \psi } \psi \rangle^{1/3 } { \raisebox{-0.13cm}{~\shortstack{$<$ \\[-0.07 cm ] $ \sim$}}~}260 \ , { \rm mev}$ ] as well as the pion mass , @xmath235 leading , in the mfa , to values such as @xmath236 , @xmath237 and @xmath238 @xcite . with these values
one obtains satisfactory predictions for the quark vacuum effective mass , @xmath239 , and for the quark condensate , @xmath240 given by @xmath241 and @xmath242 .
although in general the non covariant cut - off lies within the range @xmath243 while @xmath244 the coupling can be further increased producing much higher values for @xmath239 without affecting too much the @xmath15 independent quark condensate . for example , another mfa set also presented in ref .
@xcite is given by @xmath245 , @xmath246 and @xmath247 predicting @xmath248 and @xmath249 .
as one can see the value of @xmath239 doubles while that of the observable @xmath240 remains within the bounds @xmath234 $ \sim$}}~}-\langle { \bar \psi } \psi \rangle^{1/3 } { \raisebox{-0.13cm}{~\shortstack{$<$ \\[-0.07 cm ] $ \sim$}}~}260 \ , { \rm mev}$ ] set by sum rules @xcite and the value @xmath250 which corresponds to lattice estimates @xcite .
regarding applications at finite temperature and density one sees that also the values of the critical temperature ( at @xmath7 ) and of the critical chemical potential ( at @xmath8 ) increase with @xmath15 . in general , the size of the first order transition line , which originates at high @xmath251 and @xmath8 , increases with @xmath15 approaching the @xmath0 axis for very high coupling strengths .
this observation , together with the l@xmath5 m results that the appearance of two critical points becomes possible for @xmath252 gives us the hint to use the opt at high @xmath15 and small @xmath20 ( since @xmath253 ) while setting @xmath18 to usual values .
for consistency we must use the opt two loop relations for @xmath254 and @xmath255 recently found in ref .
@xcite and which predict some deviations from the mfa result for the gell mann - oakes - renner relation .
taking @xmath256 , and @xmath257 with @xmath258 one obtains the very reasonable values @xmath259 , @xmath260 , and @xmath261 .
however , now @xmath15 has an extremely high value which is reflected in the vacuum effective quark mass value , @xmath262 .
next , one can keep those values of @xmath18 and @xmath15 decreasing @xmath20 so as to make contact with the l@xmath5 m results .
for example , @xmath263 lead to @xmath264 while @xmath254 and @xmath265 remain very stable . with this aim
let us analyze the phase transition pattern at high @xmath0 and @xmath266 in search for a first order transition taking @xmath267 ( with @xmath268 and @xmath269 ) . within the opt this choice predicts @xmath270 , @xmath271 and @xmath272 with the expected high quark mass value , @xmath273 . in principle
, one could object to the fact that @xmath274 ( apart from having a numerical value much higher than 1/3 of the baryonic mass ) .
however , as emphasized in the introduction , the goal of our njl investigation has a more qualitative character and , at the same time , it will be shown that the relevant temperature and chemical potential for our purposes of finding evidence for a second critical point fall well below @xmath18 .
finally , note that with our parameter choice the large value of @xmath15 mostly affects @xmath239 while physical observables such as @xmath254 , @xmath255 , and @xmath265 remain realistic . with these unusual parameter values
one indeed obtains a first order transition at @xmath7 and @xmath275 as shown by the top panel of fig .
[ freenergy ] while the usual first order transition is also observed at @xmath8 and @xmath276 starting a line of first order transitions as shown by the bottom panel of fig .
[ freenergy ] which shows the degenerate minima at @xmath277 and @xmath278 .
the differences in latent heat , @xmath279 , between the two transitions can be further observed by looking at fig .
[ latheat ] which shows the normalized @xmath280 versus the dimensionless ration @xmath281 .
the values are @xmath282 for @xmath7 and @xmath283 and @xmath284 for @xmath277 and @xmath285 ( and of course higher for @xmath286 but we shall refrain from numerically exploring the @xmath287 region ) .
the situation can be also observed by analyzing the thermal behavior of the order parameter represented by the quark condensate , @xmath288 , which is given in fig .
[ orparam ] for @xmath7 and @xmath277 .
this figure clearly indicates that the first order happening at high @xmath0 and vanishing @xmath7 is much softer than the one happening in the reversed situation of small @xmath0 and high @xmath1 , a fact which is well illustrated by the three dimensional plot of fig .
[ pd3d ] which already shows that these two first order transition lines are indeed separated by a cross over region at intermediate @xmath0 and @xmath1 . this fact can be more clearly appreciated by projecting the first order lines of fig . [ pd3d ] on the @xmath4 plane as in fig .
[ phdianjl ] .
this phase diagram is qualitatively very similar to the one obtained in fig .
[ f3 ] for the l@xmath5 m ( as mentioned above our value @xmath289 leads to @xmath290 ) and constitutes our first important result concerning the njl model .
namely , that by going beyond the mfa and choosing the parameters so as to reproduce small pion masses one may obtain a second critical end point in the phase diagram .
in addition , the choice of regularization procedure appears to be crucial for the appearance of @xmath152 in the njl - model .
if one uses a cut off also in the @xmath0 dependent integrals the critical point @xmath152 does not emerge .
however , some authors ( see ref.@xcite ) have recognized that a three dimensional cut off is only needed at zero temperature so that the presence of high momentum quarks in the @xmath0 dependent feynman loops is required to ensure that the entropy density scales as @xmath291 at high temperature .
in contrast , by being renormalizable the l@xmath5 m always allows for high momentum quarks in the @xmath0 dependent loops . therefore , with our regularization choice the temperature dependent integrals within the njl are treated in the same footing as their l@xmath5 m counterparts .
a comprehensive discussion about how parameters and regularization affect njl has recently been carried out by costa et al .
@xcite . for our parameter values
the location of @xmath152 happens at the point ( @xmath292 ) and of @xmath117 at the point ( @xmath293 ) . from the quantitative point of view
these values are certainly high as one would expect from the fact that @xmath239 is also high .
nevertheless , the phase diagram resembles the one obtained with l@xmath5 m and which displays more reasonable numerical values .
note also how the first order line associated with @xmath152 is almost parallel to the @xmath1 axis .
the next step is to understand the physical nature of both critical first order lines .
our study of the susceptibilities and densities in the l@xmath5 m has revealed that the high @xmath0 small @xmath1 line terminating at @xmath152 has a more `` chiral '' character while the small @xmath0 and high @xmath1 line terminating at @xmath117 has a more hydrodynamical character .
one can then map the @xmath4 points to form a phase coexistence diagram such as the @xmath0 versus @xmath294 shown in fig .
[ coetrho ] where @xmath295 is the normal nuclear density .
the top figure shows that the coexistence region associated with the traditional @xmath117 point goes from @xmath8 to @xmath296 covering very high baryonic densities .
the unusual region , on the other hand , is very tiny going from @xmath297 to @xmath298 being restricted to a very narrow temperature range as shown by the bottom panel .
the critical points @xmath152 and @xmath117 are located at @xmath299 and @xmath300 respectively . at @xmath297 the small coexistence region
terminates into a point located at @xmath283 .
both coexistence regions have very different shapes , especially near the critical points , from which one may expect that the associated critical exponent @xmath301 , to be evaluated later , acquires different values in each situation .
let us now map the @xmath4 results into the @xmath302 plane as shown by fig .
[ phdiapv ] which , again , clearly indicates that the region associated with @xmath117 in fact looks like the mixed liquid - gas phase appearing in @xmath303 type of phase diagram for a van der waals fluid ( which does not contain an analogous of the mixed phase associated with @xmath152 ) .
the opt result for @xmath304 allows us to obtain @xmath305 for the njl with the inclusion of finite @xmath11 corrections which makes it possible to analyze the critical behavior near each one of the two critical points . with this aim
we will numerically evaluate the following thermodynamical quantities : the interaction measure ( or trace anomaly , @xmath72 ) , the equation of state parameter ( @xmath306 ) , the bulk viscosity over entropy density ( @xmath307 ) , the quark number and chiral susceptibilities ( @xmath308 and @xmath309 ) as well as some critical exponents .
let us start by considering the interaction measure @xmath310 which is plotted in fig .
[ delta ] for regions near @xmath152 and @xmath117 .
this quantity is expected to peak near a phase transition or cross over and can be helpful in locating the critical line . fig .
[ delta ] indicates that the rise near @xmath152 looks more uniform around @xmath117 and happens for temperatures near @xmath311 which is not surprising since there is little variation in @xmath0 along the associated first order line as already emphasized .
although the peaks associated with @xmath117 look more pronounced one has to recall that @xmath72 is normalized by @xmath312 and that the @xmath117 region is associated with lower temperatures .
next , let us investigate the eos parameter , @xmath313 , as represented by fig .
[ w ] for the two critical points . in both situations
one observes a downward cusp at the critical temperatures ( very much like in the case of the squared speed of sound , @xmath314 ) .
below the critical temperature our results show a bump which is also observed in the usual l@xmath5 m , polyakov l@xmath5 m and lattice studies ( see ref.@xcite ) . for high values of @xmath0 the quantity @xmath306 converges to approximately 1/3 as expected .
the bulk viscosity , @xmath315 , is an intrinsic dynamical quantity which , however , can be expressed in terms of the static thermodynamical quantities derived from the free energy as @xcite @xmath316\ ; , \label{bulk}\ ] ] where @xmath317 is a scale which will be set to @xmath18 , as in ref.@xcite , while @xmath318 represents the vacuum part of the energy density .
since @xmath315 is proportional to the specific heat , @xmath319 , the bulk viscosity over entropy density , @xmath307 , behaves as @xmath320 near @xmath321 in this approximation and peaks at the critical end points as indeed shown by our fig .
[ zetas ] where the divergences at ( @xmath322 ) and at ( @xmath323 ) indicate that the energy density has a sudden change at the critical points typical of a first order transition . for our purposes this is an interesting quantity since it has been pointed out that one can distinguish whether the system experiences a first order phase transition or a crossover from observables which are sensitive to the bulk viscosity in rhic type of experiments .
on expects that a sharp rise of bulk viscosity near a phase transition induces an instability in the hydrodynamic flow of the plasma , and this mode will blow up tearing the system into droplets @xcite .
let us now examine numerically the behavior of the quark susceptibility , @xmath308 , as well as of the chiral susceptibility , @xmath309 , near the two critical points .
these quantities are respectively given by @xmath324 and @xmath325 figures [ chiq ] and [ chim ] show @xmath308 and @xmath309 respectively as functions of @xmath0 for relevant values of @xmath1 .
as expected , one observes that these two quantities peak for both @xmath152 and @xmath117 .
however , for @xmath308 , the magnitude of the peak associated with @xmath117 seems to be much larger while for @xmath309 the difference is not so dramatic . in principle
these results could be interpreted as showing that the quark density , @xmath24 , plays a very minor role within @xmath152 while @xmath23 seems to dominate that critical point . at the same time
, the liquid - gas type of critical point , @xmath117 , seems to receive contributions from both types of density in accordance with our results for the l@xmath5 m .
these findings can be further appreciated if one investigates the free energy in terms of the two ordering densities , @xmath23 and @xmath24 , by using the techniques of fujii and ohtani @xcite to legendre transform @xmath326 to @xmath327 .
the result of this type of manipulation is shown in fig .
[ contour ] for the two critical points , @xmath152 and @xmath117 , as well as at an intermediate point ( @xmath328 ) where a cross over takes place .
the contour plots displayed by fig .
[ contour ] indicate that @xmath152 is indeed dominated by the scalar interaction , @xmath23 , while the quark ( vector ) density also plays an important role at @xmath117 in accordance with the covariance matrix results for the l@xmath5 m .
finally , the data used in the coexistence phase diagram @xmath21 allow us to evaluate the critical exponent @xmath301 defined as @xmath329 where @xmath330 represents @xmath152 or @xmath117 while @xmath331 and @xmath332 represent the two corresponding densities for a given temperature . having the quark number susceptibility allows for the numerical evaluation of the critical exponent @xmath333 , @xmath334 , for which one may define a chiral counterpart , @xmath335 , given by @xmath336 .
this procedure is illustrated by the top panel of fig .
[ ce ] in which @xmath333 is obtained by approaching the critical point in a path parallel to the @xmath1 axis from @xmath337 .
the values @xmath338 for @xmath117 is close to the one obtained in ref.@xcite while for @xmath152 the value is @xmath339 .
the bottom panel of fig .
[ ce ] shows the same procedure for @xmath335 with the critical points being approached from the @xmath340 `` wing '' of the @xmath4 plane , with @xmath341 .
interestingly enough the numerical values get approximately inverted , when compared to @xmath333 , and one gets @xmath342 for @xmath117 and @xmath343 for @xmath152 .
a similar type of procedure gives @xmath344 for @xmath152 and @xmath345 for @xmath117 .
having @xmath301 and @xmath333 allows us to determine @xmath200 , @xmath346 and @xmath347 using @xmath348 , @xmath349 and @xmath350 .
the values for the remaining exponents have been obtained after approximating our numbers for @xmath301 and @xmath333 by the ratios shown in table i. note that although our numerical estimates are very crude , since we do not cover many orders of magnitude nor try possible different paths leading to the critical points , they support our previous discussion regarding the nature of the two critical points found in this work . in particular
, the values for @xmath301 listed in table i support the liquid - gas character of @xmath117 @xcite . for this critical point
also the exponent @xmath333 , associated with @xmath308 , is greater than the one associated with the critical point @xmath152 .
note also that the values of @xmath347 obtained by using our approximate ratios for @xmath301 and @xmath333 in the scaling relations show that @xmath351 for both critical points which is consistent with the universal arguments presented in refs .
@xcite since it is expected that @xmath308 and @xmath319 should be the same near the critical points .
a detailed and highly precise determination of the associated critical exponents is beyond the scope of the present application and the interested reader is referred to refs .
@xcite .
.approximate ratios for the critical exponents associated with @xmath152 and @xmath117 in the njl model .
the values in between brackets are the results which have been numerically obtained . [ cols="^,^,^,^,^,^,^",options="header " , ] [ table - results - p4 ] it will be interesting to explicitly show that , at least from a qualitative perspective , our model predictions can be relevant for the lattice results obtained by de forcrand and philipsen @xcite who observed a shrinkage of the first order transition region when considering higher values of @xmath1 . for completeness let us also compare the opt with the mfa results when the model parameters are tuned so that the latter approximation also generates a first order transition at @xmath7 . within the njl model ,
the mfa can predict a first order transition at @xmath7 and for @xmath352 when the higher coupling , @xmath353 , is used with @xmath354 . in this case , the mfa also predicts @xmath355 , @xmath356 , and @xmath357 , and @xmath358 while , for the same set of parameters , the opt predicts @xmath359 , @xmath360 , and @xmath361 , and @xmath362 . in the @xmath19 plane
the njl model then generates fig .
[ mcmu ] in which the top panel , corresponding to the mfa , shows only a one branch first order line . when the opt is applied to the njl with the stronger @xmath353 considered in this subsection one
observes a first order line for @xmath363 which roughly corresponds to @xmath364 .
starting at the point @xmath365 one may follow a second order transition line which goes left touching the @xmath1 axis at @xmath366 and then going up to @xmath367 where it bends back to the right hand side for finite values of @xmath20 .
the situation is illustrated by the bottom panel of figure [ mcmu ] .
therefore , when one goes beyond mfa , the njl also predicts a bending back behavior of the critical line in a way which is consistent with the negative curvature of the lattice simulations of refs .
this result also supports fukushima s suggestion @xcite regarding the eventual back bending .
although this author has considered the su(3 ) version of the njl ( being able to reproduce a critical surface in the @xmath368 space ) with standard parametrization it is interesting to note that the bending was obtained by adding a vector interaction which generated a @xmath369 contribution to the pressure . by looking at our eq .
( [ omegarhos ] ) one sees that the opt also brings a term like @xmath230 to @xmath370 .
finally , note that a back bending behavior is also implied by our l@xmath5 m results if one considers that @xmath117 and @xmath371 almost coincide .
in recent publications @xcite it was shown that for small values of the vacuum pion mass ( @xmath372 mev ) the phase diagram of the l@xmath5 m has two distinct first order lines terminating in two critical points .
one , as predicted by most models , starts at finite @xmath1 and @xmath8 , whereas the other is an unusual first order line which starts at high @xmath0 and @xmath7 . in between the critical endpoints ( whose exact location depends on the vacuum pion mass ) of these two lines
there is a crossover transition . at the origin of this behavior ,
that is not observed in mfa , are the thermal fluctuations of the mesonic fields , that have been taken into account adopting a self consistent method first proposed in @xcite . inspired by this finding
we have performed a careful analysis of the l@xmath5 m with the aim of understanding in deeper details the nature of the critical points . at the same time , we have considered another popular effective quark model , the njl model , to investigate under which conditions a phase diagram with at least two distinct critical points could be reproduced . the analysis of the l@xmath5 m has been performed by using the same method of ref .
@xcite and our results agree with the ones obtained therein : the unusual first order line which starts at high @xmath0 and @xmath7 ending in the `` new '' critical point ( @xmath152 ) was also found . at the same time , a closer look in the vicinity of the `` usual '' critical endpoint revealed that , right before to end , the first order line that starts at finite @xmath1 and @xmath8 bifurcates in _ two _ critical endpoints ( @xmath117 and @xmath118 ) separated by few mev in @xmath0 and @xmath1 .
in addition to the phase diagram we have studied susceptibilities and correlations of the net - quark number density , the entropy density , and the scalar density . our results for the covariant matrix show that for @xmath152 the dominant fluctuations are given by the scalar density and the entropy density , whereas the quark number density plays only a minor role .
for the two critical points @xmath117 and @xmath118 , instead , also the fluctuations of the quark number density are important .
their contribution becomes larger and larger as the vacuum pion mass is increased , probably also due to the fact that @xmath117 and @xmath118 move towards higher values of @xmath1 . despite their vicinity in the phase diagram ,
@xmath117 and @xmath118 seem to exhibit distinct features . the main difference being the fluctuations of the scalar density , that are more important in @xmath118 than in @xmath117 .
we have then considered the njl thermodynamical potential , recently evaluated with the opt @xcite , with parameter values which simulate small pion masses in order to generate a @xmath4 phase diagram where the existence of two critical points , @xmath152 and @xmath117 , separated by a cross over region has been observed .
this type of phase diagram is similar to the one found , in the temperature - magnetic field plane , for the compressible metamagnetic ising model @xcite we have performed an extensive thermodynamical analysis in order to find the essential physical features that distinguish both critical points concluding , in agreement with the l@xmath5 m case , that the usual one ( @xmath117 ) , located at intermediate values of @xmath0 and @xmath1 has a more hydrodynamical character with the quark number density playing an essential role .
this density has little influence at lower values of @xmath1 where the new , unusual , critical point @xmath152 region is dominated by the scalar density .
although we have not considered a vector term in the version of the njl model considered here it is rather interesting to note that the opt pressure has the term @xmath14 which by being @xmath11 suppressed does not contribute to the mfa which completely misses out the possible existence of @xmath152 and the associated first order line . from these results it is clear that the inclusion of contributions beyond mfa can have dramatic consequences on the phase diagram of quark models and can not be neglected . with the combined analysis of the l@xmath5 m and the njl model
, we have shown that the appearance of multiple critical points is not an uncommon feature of effective quark models , even without explicitly introducing a vector interaction as was done in @xcite . before exporting these notions to qcd , however , some caution is advisable .
as has been very recently shown in ref .
@xcite , based on the analysis of the l@xmath5 m in mfa , the inclusion of vacuum fluctuations can change the transition in the chiral limit at @xmath7 from first to second order so that this model would behave as the njl model in the same regime . however , our results for the latter suggest that the consideration of vacuum contributions should not influence the appearance of the critical point @xmath152 , at least qualitatively , when an appropriate tuning of model parameters is carried out beyond the mfa . in the minimal version of the njl model considered here
@xmath152 appears only if the coupling is larger than usual ( see discussion in sec .
[ nonstandard ] ) while a sharp cut off is introduced only in the divergent integrals .
however , having a large parameter space , more sophisticated versions of the model could also allow for the appearance of a critical point point like @xmath152 generating , at the same time , more plausible values for the vacuum effective quark mass . concerning the other two critical points found in the l@xmath5 m ( @xmath117 and @xmath118 ) , instead
, vacuum fermion loops are not expected to modify dramatically the qualitative behavior of the model , especially for values of the vacuum pion mass close to the physical one @xcite .
the presence of @xmath118 in the l@xmath5 m with thermal fluctuations , however , has not been confirmed by the opt analysis of the njl model .
this might be due to various reasons including the fact that the two models have been evaluated under different approximations , the fact that the njl does not have true mesonic degrees of freedom , etc .
the clarification of these points calls for further investigations .
nevertheless , our results allow us to conclude that it is possible to use these effective models to generate metamagnetic like phase diagrams at the expense of using non standard parameters for @xmath20 and @xmath255 which weakly break chiral symmetry in approximations which go beyond the mfa . having two critical points
, it then becomes possible to observe the shrinkage of first order region , in the @xmath373 plane , as one considers higher values of @xmath1 so that model predictions could be conciliated with the lattice results by de forcrand and philipsen @xcite .
then , as we have shown , the first order region should increase again eventually intercepting the physical mass point as argued in ref.@xcite .
the `` chiral '' like first order transition observed in the line associated with @xmath152 seems to be much softer and the coexistence regions much smaller than in the traditional `` liquid - gas '' type of line associated with the @xmath117 which perhaps would make it harder to be detected in lattice qcd calculations ( see for example ref .
although quantities such as the trace anomaly , eos parameter , bulk viscosity and susceptibilities display the expected behavior associated with first order phase transitions it seems that the different critical points belong to distinct universality classes .
m. b. p. thanks the nuclear theory group at lbnl for the hospitality during the sabbatical year .
we thank j .-
kneur , r. ramos , i. n. mishustin , w. figueiredo , p. costa , y. hatta , h. hansen , and a. delfino for discussions .
this work was supported by the director , office of energy research , office of high energy and nuclear physics , divisions of nuclear physics , of the u.s .
department of energy under contract no .
de - ac02 - 05ch11231 , by the helmholtz international center for fair within the framework of the loewe program ( landesoffensive zur entwicklung wissenschaftlich - konomischer exzellenz ) launched by the state of hesse , and by coordenao de aperfeioamento de pessoal de nvel superior ( capes , brazil ) .
y. aoki , g. endrodi , z. fodor , s. d. katz and k. k. szabo , nature * 443 * , 675 ( 2006 ) [ arxiv : hep - lat/0611014 ] . m. asakawa and k. yazaki , nucl .
a * 504 * , 668 ( 1989 ) .
a. barducci , r. casalbuoni , s. de curtis , r. gatto and g. pettini , phys .
b * 231 * , 463 ( 1989 ) .
a. barducci , r. casalbuoni , g. pettini and r. gatto , phys .
d * 49 * , 426 ( 1994 ) .
a. m. halasz , a. d. jackson , r. e. shrock , m. a. stephanov and j. j. m. verbaarschot , phys .
d * 58 * , 096007 ( 1998 ) .
k. fukushima , j. phys .
g * 35 * , 104020 ( 2008 ) .
j. berges and k. rajagopal , nucl .
b * 538 * , 215 ( 1999 ) . p. de forcrand and o. philipsen , jhep * 0811 * , 012 ( 2008 ) ; p. de forcrand and o. philipsen , jhep * 0701 * , 077 ( 2007 ) ; p. de forcrand , s. kim and o. philipsen , pos * lat2007 * , 178 ( 2007 ) ; p. de forcrand and o. philipsen , jhep * 0811
* , 012 ( 2008 ) .
r. seznec and j. zinn - justin , j. math . phys . *
20 * , 1398 ( 1979 ) ; j. c. le guillou and j. zinn - justin , ann . phys . * 147 * , 57 ( 1983 ) ; v. i. yukalov , moscow univ .
phys . bull . *
31 * , 10 ( 1976 ) ; w. e. caswell , ann .
( n.y ) * 123 * , 153 ( 1979 ) ; i. g. halliday and p. suranyi , phys . lett . *
b85 * , 421 ( 1979 ) ; j. killinbeck , j. phys .
* a14 * , 1005 ( 1981 ) ; r. p. feynman and h. kleinert , phys .
a * 34 * , 5080 ( 1986 ) ; h. f. jones and m. moshe , phys . lett .
* b234 * , 492 ( 1990 ) ; a. neveu , nucl .
( proc . suppl . )
* b18 * , 242 ( 1990 ) ; v. yukalov , j. math .
phys * 32 * , 1235 ( 1991 ) ; c. m. bender et al . ,
phys . rev . * d45 * , 1248 ( 1992 ) ; h. yamada , z. phys .
* c59 * , 67 ( 1993 ) ; a. n. sissakian , i. l. solovtsov and o. p. solovtsova , phys .
* b321 * , 381 ( 1994 ) ; c. arvanitis , f. geniet , m. iacomi , j .- l . kneur and a. neveu ,
* a12 * , 3307 ( 1997 ) ; h. kleinert , phys . rev .
d * 57 * , 2264 ( 1998 ) ; phys .
b434 * , 74 ( 1998 ) ; for a review , see h. kleinert and v. schulte - frohlinde , _ critical properties of @xmath374-theories _ , chap .
19 ( world scientific , singapure 2001 ) ; k. g. klimenko , z. phys . * c50 * , 477 ( 1991 ) ; _ ibid . _ * c60 * , 677 ( 1993 ) ; mod . phys . lett . * a9 * , 1767 ( 1994 ) ; s. chiku and t. hatsuda , phys .
rev . d*58 * , 076001 ( 1998 ) .
f. f. souza cruz , m. b. pinto and r. o. ramos , phys .
b * 64 * , 014515 ( 2001 ) ; j .-
kneur , m. b. pinto and r. o. ramos , phys .
lett . * 89 * , 210403 ( 2002 ) ; phys .
a * 68 * , 043615 ( 2003 ) ; e. braaten and e. radescu , phys .
lett . * 89 * , 271602 ( 2002 ) , phys .
a * 66 * , 063601 ( 2002 ) ; j .-
kneur , a. neveu and m. b. pinto , phys .
a * 69 * , 053624 ( 2004 ) ; b. kastening , phys .
a * 70 * , 043621 ( 2004 ) ; j .-
kneur and m. b. pinto , phys .
a * 71 * , 033613 ( 2005 ) .
i. j. r. aitchison and c. m. fraser , phys .
d * 31 * , 2605 ( 1985 ) .
g. w. carter , p. j. ellis and s. rudaz , nucl .
a * 618 * , 317 ( 1997 ) .
g. w. carter , o. scavenius , i. n. mishustin and p. j. ellis , phys .
c * 61 * , 045206 ( 2000 ) .
m. i. gorenstein and s. n. yang , phys .
d * 52 * , 5206 ( 1995 ) .
y. nambu and g. jona - lasinio , phys . rev . * 122 * , 345 ( 1961 ) ; * 124 * , 246 ( 1961 ) .
|
we consider the two flavor version of the linear sigma model as well as of the nambu jona - lasinio model , at finite temperature and quark chemical potential , beyond the mean field approximation .
using parameter values for the pion and quark current masses which weakly break chiral symmetry we show that both models can present more than one critical end point .
in particular , we explicitly show that the appearance of a new critical point associated with a first order line at high temperature and low densities could help to conciliate some lattice results with model predictions . using different techniques , we perform an extensive thermodynamical analysis to understand the physical nature of the different critical points . for both models ,
our results suggest that the new first order line which starts at vanishing chemical potential has a more chiral character than the usual line which displays a character more reminiscent of a liquid - gas phase transition .
|
given a finite metric space @xmath0 with metric @xmath1 we define the spread @xmath2 by @xmath3 this is one of a family of metric space ` sizes ' as we will see in section [ section : leinstercobbold ] where we will also see the motivation for this definition .
this generalizes easily from finite metric spaces to arbitrary metric spaces with a measure , as will be seen in section [ section : measurespaces ] .
one of the things that we will be interested in is how this measure of size alters as the metric is scaled , so we need to define some notation . for @xmath4
let @xmath5 denote the metric space @xmath0 with the metric @xmath1 scaled up by a factor of @xmath6 , so that the distance in @xmath5 between @xmath7 and @xmath8 is @xmath9 .
we can consider the spread profile of the space @xmath0 which is just the graph of @xmath10 for @xmath4 .
@xmath11 \node [ point ] ( one ) at ( 0,0 ) { } ; \node [ point ] ( two ) at ( 4.5,0.4 ) { } ; \node [ point ] ( three ) at ( 4.5,-0.4 ) { } ; \path [ line,<- > ] ( one ) -- node[above , pos=0.5 ] { $ 1000t$ } ( two ) ; \path [ line,<- > ] ( one ) -- node[below , pos=0.5 ] { $ 1000t$ } ( three ) ; \path
[ line,<- > ] ( three ) -- node[right , pos=0.5 ] { $ t$ } ( two ) ; \end{tikzpicture}$ ] + file picsdata / blockofdotsmagthreepoints.dat ; an example of a profile is given in figure [ figure : threepointprofile ] .
we consider the space @xmath12 , for @xmath13 , having three points , two of which are a distance @xmath6 apart and are both a distance @xmath14 from the third point .
this family of metric spaces can be thought of as having three ` regimes ' : where @xmath6 is very small and there looks likes there is one point ; where @xmath6 is smallish and there looks like there are two points ; and where @xmath6 is very large and it looks like there are three points .
this way of thinking is reflected in the values plotted .
so we wish to think of the spread @xmath15 as akin to an ` effective number of points ' .
the basic properties of spread can now be given .
for @xmath0 a finite metric space with @xmath16 points , the spread has the following properties : * @xmath17 ; * @xmath10 is increasing in @xmath6 ; * @xmath18 as @xmath19 ; * @xmath20 as @xmath21 ; * @xmath22 .
these results all follow easily from the definition . in this paper
we consider further properties of the spread which are summarized in the following synopsis . in the first section we recall leinster and cobbold s diversity measures @xcite and
show how that gives rise to the spread @xmath15 of a metric space as the order - zero diversity of the metric space equipped with the uniform probability distribution .
this means that the spread can be thought of as an analogue of the number of species in an ecosystem .
we all see that there is a spread @xmath23 of order @xmath24 for all @xmath25 and relate these to generalized means . in the second section
we compare the spread with leinster s notion @xcite of the magnitude of a metric space .
we show that if the space has a positive definite ` similarity matrix ' then the magnitude is an upper bound for the spread , and if the space is homogeneous then the magnitude is equal to the spread .
we then go on to consider a space with no magnitude , or rather , a space whose magnitude profile is discontinuous and see that the spread profile is very similar , but much better behaved . finally in this section
we see that two spaces with the same magnitude profile can have different spread profiles . in the third section
we show that the definition of spread generalizes immediately to infinite metric spaces which are equipped with a measure .
we calculate the spread of @xmath26 the straight line interval of length @xmath27 with the usual lebesgue measure and show that for large @xmath27 the spread is approximately @xmath28 .
then we consider the spread of compact riemannian manifolds , giving the spread of the @xmath29-sphere explicitly , and we show that asymptotically , as the manifold is scaled up , the leading order terms in the spread are determined by the volume and the total scalar curvature of the manifold . in the final section we consider the growth rate of the spread as the metric space is scaled .
this gives a scale dependent notion of ` dimension ' .
for instance if millions of points are formed into the shape of a square , then at small scales it will look like a point , i.e. zero dimensional , at medium scales it will look two dimensional , and at very large scales it will look like a collection of isolated points , i.e. zero dimensional , again .
we look at numerical calculations for some simple approximations to fractals and see that at the medium scales the ` spread dimension ' is related to the hausdorff dimension of the fractal .
in this section we recall the notion of leinster - cobbold diversity measures and show how this gives rise to the spread of a metric space .
we also see how generalized notions of spread relate to generalized means .
in @xcite leinster and cobbold defined certain ` diversity measures ' .
these are numbers associated to any finite set equipped with a probability distribution and a ` similarity matrix ' we will see that a metric on a finite set gives rise to a similarity matrix in a canonical way . these numbers are supposed to measure the biodiversity of a community where the points represent the different species , the similarity matrix represents the similarity between the species ( a metric represents distance between the species ) and the probability distribution represents the relative abundances of the species . before defining the diversity measures we need to define the notion of a similarity matrix .
if @xmath0 is a finite set with @xmath16 points @xmath30 then a similarity matrix @xmath31 is an @xmath32 matrix with @xmath33 and @xmath34 .
if @xmath35 then this represents @xmath36 and @xmath37 being completely dissimilar and if @xmath38 then this represents @xmath36 and @xmath37 being completely identical .
a metric @xmath1 on @xmath0 gives rise to a similarity matrix @xmath31 by setting @xmath39 , so that nearby points are considered very similar and far away points dissimilar .
given a finite set @xmath0 with @xmath16 points @xmath30 equipped with a probability distribution @xmath40 , so that @xmath41 , and a similarity matrix , @xmath31 , for @xmath42 $ ] , define the leinster - cobbold diversity of order @xmath24 by @xmath43 \displaystyle\prod_{i : p_i>0 } ( z\mathbf{p})_i^{-p_i } & q= 1,\\[2em ] \displaystyle \min_{i : p_i>0 } \frac{1}{(z\mathbf{p})_i } & q=\infty . \end{cases}\ ] ] for fixed @xmath0 , @xmath44 and @xmath31 , the graph of @xmath45 against @xmath24 is known as the diversity profile . as a function of @xmath24 , @xmath45 is monotonically decreasing .
we also have @xmath46 and the order @xmath24 diversity can be thought of as an ` effective number of species ' .
the leinster - cobbold diversity measures generalize a classic family of diversity measures known as the hill numbers @xcite .
the hill number of order @xmath24 , @xmath47 , for @xmath42 $ ] is defined for a finite set @xmath0 with probability distribution @xmath44 on it , so it requires no metric or similarity matrix .
this hill number can be obtained as the leinster - cobbold diversity of the identity similarity matrix , or , equivalenty , of the ` discrete ' metric where all of the points are infinitely far apart from each other ; so all species are considered to be completely dissimilar .
symbolically , we have @xmath48 the hill numbers at the values @xmath49 give , respectively , the following classical diversity measures : the number of species , the exponential shannon index , the simpson index and the reciprocal berger - parker diversity .
we can think of this specialization to hill numbers as using the leinster - cobbold diversity measure to get measures of a finite probability space by equipping the space with a canonical metric , namely the discrete metric . on the other hand we could use the leinster - cobbold diversity measures to get measures of the size of a finite _ metric _ space by equipping the space with a canonical _ probability distribution _
, namely the uniform distribution .
this gives rise to the spread .
symbolically , for a metric space @xmath0 with @xmath16 points , define the @xmath24-spread @xmath50 , for @xmath51 , by @xmath52 explicit formulas are given in table [ table : formulasforeq ] . by the monotonically decreasing nature of the leinster - cobbold diversity measures we have that @xmath53 implies that @xmath54 . in this paper
we have generally concentrated on the greatest of the these values , @xmath2 which we just call the spread ; this is the analogue of the ` number of species ' in an ecosystem .
@xmath55 \displaystyle \sum_{i=1}^n \frac{1}{\sum_{j=1}^n z_{ij}}&q=0,\\[2em ] \displaystyle n\cdot\prod_{i=1}^n \biggl ( \frac{1}{\sum_{j=1}^n z_{ij}}\biggr)^{1/n } & q=1 , \\[2em ] \displaystyle\frac{n^2}{\sum_{i , j=1}^n z_{ij } } & q=2,\\[2em ] \displaystyle\min_{i=1,\dots , n } \biggl ( \frac{n}{\sum_{j=1}^n z_{ij}}\biggr ) & q=\infty .
\end{cases}\ ] ] fundamental to the definition of the leinster - cobbold diversity measures is the idea of generalized mean @xcite .
here we give a description of the @xmath24-spread in those terms .
suppose that @xmath0 is a finite metric space with @xmath16 points @xmath30 , then each point @xmath36 has a reciprocal mean similarity denoted by @xmath56 and defined , as the name suggests , as follows : @xmath57 we have @xmath58 and think of the reciprocal mean similarity as being a measure of how different the space is from the point @xmath36 , with @xmath56 being nearly @xmath59 if all the points are close to @xmath36 and nearly @xmath16 if all of the points are far from @xmath36 . in order to get a measure of the whole space we can take an average of these reciprocal mean similarities .
there are many different averages we could take . for a set of numbers @xmath60 and a number @xmath61 ,
the @xmath62-mean @xmath63 is defined , when @xmath64 as @xmath65 and as a limit when @xmath66 .
this includes many standard means : @xmath67 is the maximum , @xmath68 is the quadratic mean , @xmath69 is the arithmetic mean , @xmath70 is the geometric mean , @xmath71 is the harmonic mean , and @xmath72 is the minimum .
these have various nice properties , but the interesting one to note here is that if @xmath73 then @xmath74 with equality if and only if all of the numbers in @xmath75 are equal . for - spread
can also be defined for negative @xmath24 , but the properties are slightly different and we do not consider that case here . ] @xmath76 $ ] , the @xmath24-spread @xmath50 of the metric space @xmath0 is by definition the @xmath77-mean of the individual reciprocal mean similarities : @xmath78 we have @xmath79 with @xmath50 being near to @xmath59 if all of the points are close to each other and @xmath50 being near to @xmath16 if all of the points are far away from each other .
in this section we recall leinster s notion of magnitude and show how it relates to the spread .
we look at examples of a metric space with no magnitude and two metric spaces with the same magnitude .
magnitude was introduced by leinster in @xcite .
it is defined for ` most ' metric spaces in the following way . for @xmath0 a metric space a weighting on @xmath0 consists of a weight @xmath80 for each @xmath81 such that @xmath82 if a weighting exist then @xmath83 the magnitude of @xmath0 is defined to be the sum of the weights : @xmath84 if it exists then the magnitude is independent of any choice in the weighting .
the definition of magnitude comes from enriched category theory , although it had previously appeared in the biodiversity literature @xcite . for an example of a space without a magnitude see section [ section : naughtyexample ] below .
there are large classes of spaces for which the magnitude is known to exist : one class of spaces on which it is defined is the class of ` positive definite spaces ' .
a positive definite finite metric space is a finite metric space for which the similarity matrix @xmath31 is positive definite .
examples of positive definite spaces include subspaces of euclidean space .
one nice property of the magnitude of positive definite spaces is the following . [
* corollary 2.4.4 ) ] if @xmath0 is a positive definite space then @xmath83 is well defined , furthermore if @xmath85 then @xmath86 is also positive definite and @xmath87 .
the magnitude is related to the leinster - cobbold diversity via the ` maximum diversity ' @xmath88 . before defining that we say that a space @xmath0 has a non - negative weighting if there is a weighting for @xmath0 in which all of the weights are non - negative .
the maximum diversity is defined to be the maximum of the magnitudes of subsets of @xmath0 with a non - negative weighting : @xmath89 for instance , from the positive definite subset bound it follows that if @xmath0 is positive definite with non - negative weighting then @xmath90 .
the connection with diversity is given by the following theorem .
for @xmath0 a metric space and for any @xmath42 $ ] the maximum value of the leinster - cobbold diversity of order @xmath24 , over all probability distributions on @xmath0 , is given by the maximum diversity : @xmath91 this explains the name .
the maximum diversity is certainly in some sense much better behaved than the magnitude , however it is considerably harder to calculate in general .
we can now look at some basic comparisons .
suppose that @xmath0 is a finite metric space . 1 .
the spread of @xmath0 is bounded - above by its maximum diversity : @xmath92 2 .
if @xmath0 is positive definite then its maximum diversity is bounded above by its magnitude , and thus so is its spread : @xmath93 1 .
this follows from immediately leinster s maximizing theorem and the interpretation of @xmath2 as the order - zero leinster - cobbold diversity of @xmath0 with the uniform probability distribution .
[ part : two ] by definition , the maximum diversity of @xmath0 is the magnitude of a subset @xmath86 of @xmath0 , so by the positive definite subset bound , if @xmath0 is positive definite then @xmath94 .
note that the positive definite condition in part [ part : two ] of the above theorem can not simply be removed as we will see in section [ section : naughtyexample ] that there is non - positive definite space with magnitude smaller than spread .
we can show that the spread is actually equal to the magnitude in the special case of a homogeneous metric space . recall that a homogeneous space is a space in which the points are all indistinguishable , or ,
more precisely , a homogeneous metric space is a space with a transitive action by a group of isometries . if @xmath0 is a homogeneous finite metric space then the magnitude and the spread coincide : @xmath95 more generally , the magnitude is equal to the @xmath24-spread for all @xmath42 $ ] : @xmath96 if @xmath0 has @xmath16 points , then the speyer s formula ( * ? ? ?
* theorem 1 ) for the magnitude of a homogeneous space , we have for any @xmath81 that @xmath97 on the other hand , every point in @xmath0 has the same mean reciprocal similarity @xmath98 , with @xmath99 the @xmath24-spread @xmath50 is just the @xmath77-mean of the set of mean reciprocal similarities .
but the @xmath77-mean of @xmath16 copies of @xmath98 is just @xmath98 , thus @xmath100 as required .
there is slightly interesting notational coincidence when @xmath31 is invertible : @xmath101 here we look at an example , given by leinster in @xcite , which has a discontinuity in its magnitude profile and look at its continuous spread profile .
consider the five - point space @xmath102 illustrated in figure [ figure : badmagnitude ] , equipped with the metric induced by the pictured graph , so that points on the same side are a distance @xmath103 from each other and points on opposite sides are a distance @xmath59 from each other .
as we scale this space , considering @xmath104 for @xmath4 , we see that when @xmath105 this has no magnitude .
however , @xmath106 is defined for all values of @xmath4 , and this seems to be a well - behaved version of the magnitude .
this example also shows that magnitude is not always an upper bound for the spread .
\(k ) at ( -1,2 ) @xmath107 ; ( a ) at ( 0,0 ) ; ( b ) at ( 0,2 ) ; ( c ) at ( 0,4 ) ; ( d ) at ( 2,1 ) ; ( e ) at ( 2,3 ) ; ( a ) node[above , pos=0.5 ] ( d ) ; ( a )
node[above , pos=0.5 ] ( e ) ; ( b )
node[above , pos=0.5 ] ( d ) ; ( b )
node[above , pos=0.5 ] ( e ) ; ( c )
node[above , pos=0.5 ] ( d ) ; ( c ) node[above , pos=0.5 ] ( e ) ; file picsdata / blockofdotsmagbadfivepointsi.dat ; ; file picsdata / blockofdotse0badfivepoints.dat ; ; file picsdata / blockofdotsmagbadfivepointsii.dat ; it is straightforward to generate examples of metric spaces with many points on the magnitude profile not defined , for instance you can use maple to take a random graph with say @xmath108 vertices and an expected valency of @xmath109 at each vertex .
we summarize here some of the good properties that the spread has when compared with the magnitude .
* the spread @xmath15 is defined for all metric spaces . * as an @xmath16-point space is scaled up , the spread @xmath110 increases from @xmath59 to @xmath16 . *
it is much easier to calculate @xmath2 than @xmath83 .
a further class of metric spaces to consider is that of trees . given a tree , that is a graph with no cycles , we get a metric space consisting of the set of vertices and the edge - length .
we can generalize an example of leinster ( * ? ? ?
* example 2.3.5 ) to show that all trees with the same number of vertices have the same magnitude .
suppose that @xmath111 is a tree with @xmath16 vertices for @xmath112 , then the magnitude function is given by @xmath113 observe that if @xmath114 then @xmath115 and the result holds .
suppose that @xmath116 .
pick a leaf @xmath117 , i.e. a univalent vertex , of the tree @xmath111 .
let @xmath118 be the metric space @xmath119 and let @xmath86 be the submetric space of @xmath111 consisting of @xmath117 and its adjacent vertex .
then by ( * ? ? ?
* corollary 2.3.3 ) we have @xmath120 and as @xmath118 is a tree with @xmath121 vertices the result follows by induction .
let s look in particular at two extreme examples ; these are pictured in figure [ fig : linearandcorona ] .
on the one hand we have @xmath122 the linear tree with @xmath16 vertices ; on the other hand we have @xmath123 the corona with @xmath16 vertices , that is the tree with one `` central '' vertex which has an edge to each of the other vertices , and there are no other edges .
the corona @xmath123 can be thought of as the complete bipartite graph @xmath124 .
both of these @xmath16-trees give positive - definite metric spaces . by the above theorem these two spaces have the same magnitude function .
however , they have various different properties .
for instance , their diameters are distinct , @xmath125 and @xmath126 ; and @xmath127 always has a positive weighting , whereas @xmath123 has a negative weight on the central point if @xmath128 ( and @xmath129 has a negative weight on the central point if @xmath130 ) .
\(a ) at ( 0,0 ) ; ( b ) at ( 1,0 ) ; ( c ) at ( 2,0 ) ; ( d ) at ( 3,0 ) ; ( e ) at ( 4,0 ) ; ( f ) at ( 5,0 ) ; ( a ) node[above , pos=0.5 ] t ( b ) ; ( b ) node[above , pos=0.5 ] t ( c ) ; ( c ) node[above , pos=0.5 ] t ( d ) ; ( d ) node[above , pos=0.5 ] t ( e ) ; ( e )
node[above , pos=0.5 ] t ( f ) ; \(m ) at ( 0,0 ) ; ( n ) at ( 0:1 ) ; ( o ) at ( 72:1 ) ; ( p ) at ( 144:1 ) ; ( r ) at ( -144:1 ) ; ( s ) at ( -72:1 ) ; ( m ) node t ( n ) ; ( m ) node t ( o ) ;
( m ) nodet ( p ) ; ( m ) nodet ( s ) ; ( m ) node t ( r ) ; the spread distinguishes these spaces .
an easy calculation gives the following .
@xmath131 it is not too hard to calculate the maximum diversity function either . as the linear tree is positive definite and
has a positive weighting the maximum diversity is precisely the magnitude .
the corona does not always have a positive weighting , and one finds that the central point needs to be ` switched off ' when the corona is scaled down sufficiently ; this gives the following maximum diversity function .
@xmath132 expression [ domain=0.01:1.2]exp(9*x ) ; ; expression ( 2 + 10*(exp(x)-1))/(exp(x)+1 ) ; ; file picsdata / e0lineartree10.dat ; ; expression [ domain=0.01:1.2]exp(2*x ) ; ; expression ( 2 + 10*(exp(x)-1))/(exp(x)+1 ) ; ; expression ( 9)/(1+(8)*exp(-2*(x ) ) ) ; ; expression [ samples=100,domain=0.01:10 ] 1/(1+(10 - 1)*exp(-x))+ ( 10 - 1)/(1+exp(-x)+(10 - 2)*exp(-2*(x ) ) ) ; ; expression ( 2 + 10*(exp(x)-1))/(exp(x)+1 ) ; the linear tree and the corona with @xmath109 points are compared in figure [ figure : corona ] .
whilst they have the same magnitude they clearly do not have the same maximum diversity nor spread : the linear tree has greater spread than the corona .
the magnitude of both spaces grows essentially linearly to start with ( see section [ section : growthrate ] ) , which is not really the right thing for the corona .
the spread ( and maximum diversity ) grows linearly for the linear tree and exponentially for the corona , reflecting the geometry of these spaces somewhat more .
the spread of a finite metric space was defined by using the canonical uniform probability measure on the underlying finite set .
the definition generalizes immediately to any metric space equipped with a finite mass measure .
if @xmath133 is a metric space equipped with a measure @xmath134 such that @xmath135 , then we can define the spread of @xmath0 by @xmath136 this is really the spread with respect to the associated probability measure @xmath137 , but the two factors of @xmath138 cancel in the numerator and denominator . for @xmath50 with @xmath139
the total mass @xmath138 makes an appearance in the definition .
we can now look at the following examples : the line interval with the lebesgue measure ; the @xmath29-sphere with its intrinsic metric and standard measure ; and , asymptotically , any compact riemannian manifold .
we can quite straightforwardly calculate the spread of the length @xmath27 line interval @xmath26 equipped with the standard lebesgue measure .
we have @xmath140 and asymptotically , as @xmath141 , @xmath142 this is just a case of calculating the integral .
first observe that for @xmath143 $ ] , @xmath144_{y = 0}^{x } - \left[{e^{-y+x}}\right]_{y = x}^{\ell } \\ & = 2- \left({e^{-x}}+{e^{-(\ell - x)}}\right ) .
\end{aligned}\ ] ] thus @xmath145_{x=0}^\ell \\&= \frac{\operatorname{arctanh}(\sqrt{1-e^{-\ell}})}{\sqrt{1-e^{-\ell } } } .
\end{aligned}\ ] ] now to consider the asymptotic behaviour as @xmath141 , observe @xmath146 thus @xmath147 whence , as @xmath148 decays exponentially to @xmath59 , @xmath149 as required .
this result should be compared with the magnitude for the interval of length @xmath150 @xcite : @xmath151 so asymptotically they have the same leading order term but different sub - leading terms . as an aside
, we can easily calculate the integral theoretic versions of @xmath152 and @xmath153 for the interval .
again , asymptotically these have the the same leading order terms , but different sub - leading order terms . for @xmath150
we have the following results for the length @xmath27 interval . 1 .
2 . @xmath155 as @xmath141 .
3 . @xmath156 .
4 . @xmath157 as @xmath141 . 1 .
this is obtained from the integral version of the order two spread : @xmath158 2 .
this follows from the above .
this is obtained from the integral version of the order - infinity spread : @xmath159 } \frac{\ell } { 2- ( e^{-x}+e^{-(\ell - x ) } ) } \\ & = \frac{\ell}{2(1-e^{-\ell/2})}. \end{aligned}\ ] ] 4 .
this follows from the above . a riemannian
manifold is a smooth manifold equipped with a riemannian metric , so in particular has an inner - product on each tangent space .
this structure gives rise to both a metric and a measure on the manifold .
the metric comes about because the riemannian metric can be used to define a length for each rectifiable path in the manifold and the distance between two points is defined to be the infimum of the lengths of all the paths between the two points .
the measure comes about because the riemannian metric can be used to define a volume form which leads to a density and a measure .
this means that every riemannian manifold has a well - defined spread given by the formula @xmath160 in the case of homogeneous riemannian manifolds this coincides with the formula for the magnitude that was examined in @xcite . in particular
this tells us that the spread of @xmath161 the @xmath29-sphere of radius @xmath162 with its intrinsic metric , for @xmath163 is given by @xmath164 \displaystyle\frac{\pi r } { 1-e^{-\pi r}}\prod_{i=1}^{(n-1)/2}\bigl(\bigl(\tfrac{r}{2i}\bigr)^2 + 1\bigr)\quad & \text{$n$ odd}. \end{cases}\ ] ] moreover , the methods employed in @xcite to calculate the asymptotics carry over essentially unchanged but work for _ all _ closed riemannian manifolds and not just homogeneous ones .
[ thm : riemannianasymptotics ] if @xmath0 is an @xmath29-dimensional riemannian manifold ( without boundary ) , with @xmath165 denoting the volume of the unit @xmath29-ball , @xmath166 denoting the volume of @xmath0 and @xmath167 denoting the total scalar curvature of @xmath0 then as @xmath0 is scaled up the asymptotics of the spread are as follows : @xmath168 this is almost identical to the proof of theorem 11 in @xcite except that now the scalar curvature is not a constant and should be written as @xmath169 .
this simplifies in the case @xmath170 as follows .
for @xmath171 a riemannian surface , the spread is asymptotically given in terms of the area and the euler characteristic by @xmath172 this follows from the theorem above as @xmath173 and the gauss - bonnet theorem says that @xmath174 .
in this section we define the notion of spread dimension of a metric space which is the instantaneous growth rate of the spread of the space .
this notion of dimension is scale dependent .
for instance , we will see that a long , thin rectangular array of points can have spread dimension close to zero , one , or two , depending on the scale .
finally , we look at the spread dimension of some finite approximations to simple fractals and see that the spread dimension is close to the hausdorff dimension at some scales . now that we have a notion of size of a metric space
, we can look at the growth rate of this size as a measure of the ` dimension ' of the space .
typically one looks at the asymptotic growth rate as a measure of dimension , but it is interesting here to look at the instantaneous growth rate .
the size is very scale dependent in a non - obvious way so looking at how the growth rate varies is very interesting .
for a real - valued function @xmath175 defined on some subset of the reals , we can define the growth rate at @xmath6 by @xmath176 for example , if @xmath177 then @xmath178 .
another way of writing this is as @xmath179 the instantaneous growth rate is the gradient in a @xmath180-@xmath180 plot of the function . we define @xmath181 the instantaneous spread dimension , or just spread dimension , of a metric space @xmath0 to be @xmath182 where @xmath183 , in other words , @xmath184 it is then informative to look at examples of this instantaneous spread dimension as the space is scaled .
the following examples were calculated using ` maple ` on a processor with 16 gb of ram . file picsdata / linenomatrixe0derivn30000.dat ; ; file picsdata / rectanglenotplote0derivx220y220.dat ; ; file picsdata / rectanglenotplote0derivx10y4900.dat ; ; expression 1 ; expression 2 ; as a first set of examples we can look at three types of rectangular grids with equally spaced points .
the spread - dimension profiles are shown in figure [ figure : rectangulargriddimensions ] .
starting first with the grid of @xmath185 points , or , in other words , a line of @xmath186 points , we see that when the points are very close together the spread dimension is close to zero , reflecting the fact that the ` line ' at that scale is point - like . as the line of points
is scaled up , it looks more and more like a line , so when the interpoint distance is @xmath187 units , meaning the length is @xmath188 units , the spread dimension is close to one . as the line of points
is scaled up further and further , so that the interpoint distance is @xmath109 units , say , the point - like nature is apparent and the spread dimension drops to zero . considering the square grid of @xmath189 points
, we see that this starts off looking like a point at small scales , with the spread dimension being close to zero , then as the square grid is scaled up to about @xmath190 units by @xmath190 units , with an interpoint distance of about @xmath191 units , it looks more like a genuine square and has an spread dimension of just under two
. then as the square grid is scaled up further , the point - like nature is apparent and the spread dimension drops to zero .
the most interesting case shown is where we consider the rectangular grid of @xmath192 points .
again , at small scales the spread dimension is close to zero whilst the grid looks like a small point .
then as it is scaled up there is a regime , around where the rectangle is of the order of @xmath191 units by @xmath193 units , where the space looks ` line - like ' and the dimension is approximately one . as it is scaled up further to around @xmath109 units by @xmath194 units , the width is apparent and the spread dimension heads towards two .
finally , as it is scaled up further , the point - like nature becomes apparent and the spread dimension descends to zero . from this
we deduce that the spread @xmath15 is clearly measuring something geometric .
we now look at the spread dimension of certain finite approximations to fractal sets in euclidean space , namely to the ternary cantor set , the koch curve and the sierpinski triangle .
we can look at the spread dimension profile and see that at certain scales the spread dimension is roughly the hausdorff dimension of the corresponding fractal , indicating that spread is a reasonable measure of the size of these fractals , and , indeed , of these approximations to these fractals .
file picsdata / cantore0derivdepth9n1024.dat ; ; file picsdata / cantore0derivdepth10n2048.dat ; ; expression[domain=-3:6 ] 0.6309 ; file picsdata / cantore0derivdepth9n1024.dat ; ; file picsdata / cantore0derivdepth10n2048.dat ; ; expression 0.6309 ; + file picsdata / kochnotplote0derivdepth6n4097.dat ; ; file picsdata / kochnotplote0derivdepth7n16385.dat ; ; expression 1.262 ; file picsdata / sierpinskinotplote0derivdepth8n9843.dat ; ; file picsdata / sierpinskinotplote0derivdepth9n29526.dat ; ; expression 1.585 ; + file picsdata / kochnotplote0derivdepth6n4097.dat ; ; file picsdata / kochnotplote0derivdepth7n16385low.dat ; file picsdata / kochnotplote0derivdepth7n16385high.dat ; ; expression 1.262 ; file picsdata / sierpinskinotplote0derivdepth8n9843.dat ; ; file picsdata / sierpinskinotplote0derivdepth9n29526.dat ; ; expression 1.585 ; in the first case we look at the ternary cantor set .
this is approximated by starting with two points a distance @xmath27 apart on a line .
we use the two contractions of the line by a factor of a third which respectively leave the two points fixed . by applying these two contractions successively up to @xmath109 times , starting at the initial points ,
we obtain @xmath195 points .
the spread dimension at various lengths @xmath27 can then be computed numerically . in figure
[ fig : fractaldimensions ] we see that at small scales the spread dimension is close to zero corresponding to the fact that the space looks like a point at those scales .
similarly , at very large scales , the space looks like a collection of distant points and the spread dimension is again zero . at intermediate scalings ,
roughly for @xmath196 the spread dimension is roughly the hausdorff dimension of the cantor set , namely @xmath197 , indicating that the space looks more ` cantor set - like ' at those scales .
the top - right picture in figure [ fig : fractaldimensions ] is an enlargement of the cantor set profile , and shows that things are apparently more intriguing than one might guess . at intermediate scales ,
the spread dimension seems to oscillate around the hausdorff dimension , with the oscillations being of _ multiplicative _ period @xmath198 .
such small oscillations were observed for the magnitude of the cantor set in @xcite .
i have no good explanation for these oscillations at the moment . in the next case
we look at the koch curve . again
, this is approximated by starting with a couple of points and iteratively applying one of four contractions , to obtain a finite metric space contained in the koch curve .
the graph shows that as the approximation is scaled up from very small , the spread dimension increases to roughly the hausdorff dimension , @xmath199 , where it remains over a range of scales , before descending to zero as the approximating space is scaled up sufficiently so that it s discrete , point - like nature is apparent .
the final example of the sierpinski triangle is generated in the same way using an iterated function system , and shows the same behaviour , namely , of having roughly the same spread dimension as its hausdorff dimension at certain scales .
these examples should serve to show that there is something interesting going on which has yet to be examined fully .
it is a pleasure to thank tom leinster and mark meckes for various helpful conversations , comments and terminological assistance .
similarly i would like to thank the centre de recerca matemtica at the universitat autnoma de barcelona where some of this work was carried out and where i had opportunity to talk about this work during the exploratory programme on the mathematics of biodiversity ; i would also like to thank the participants of that programme for their input and enthusiasm .
finally i thank sam marsh and neil dummigan for integral inspiration .
99 m. o. hill , _ diversity and evenness : a unifying notation and its consequences _ , ecology * 54 * ( 1973 ) 427432 .
t. leinster , _ a maximum entropy theorem with applications to the measurement of biodiversity _ , arxiv preprint .
t. leinster , _ the magnitude of metric spaces _ , arxiv preprint .
t. leinster and c. a. cobbold , _ measuring diversity : the importance of species similarity _ , ecology * 93 * ( 2012 ) 477489 .
t. leinster and s. willerton , _ on the asymptotic magnitude of subsets of euclidean space _ , geometriae dedicata , to appear .
m. meckes , _ positive definite metric spaces _ , positivity , to appear .
a. solow , s. polasky , _ measuring biological diversity _ , environmental and ecological statistics * 1 * ( 1994 ) , 95107 .
s. willerton , _ heuristic and computer calculations for the magnitude of metric spaces _ , arxiv preprint .
s. willerton , _ on the magnitude of spheres , surfaces and other homogeneous spaces _ , arxiv preprint .
|
motivated by leinster - cobbold measures of biodiversity , the notion of the spread of a finite metric space is introduced .
this is related to leinster s magnitude of a metric space .
spread is generalized to infinite metric spaces equipped with a measure and is calculated for spheres and straight lines . for riemannian manifolds the spread
is related to the volume and total scalar curvature . a notion of scale - dependent dimension
is introduced and seen , numerically , to be close to the hausdorff dimension for approximations to certain fractals .
|
more than 180 molecules have been detected in the interstellar medium and circumstellar shells of evolved stars , and approximately 40% of them are classified into carbon - chain molecules .
it is therefore of fundamental importance for astrochemistry to study their formation processes .
however , these molecules are so reactive due to unsaturated chemical bonds and/or unpaired electrons that laboratory measurements of their reaction rates are not routine experiments . although formation mechanisms of the carbon - chain molecules have mainly been studied by chemical model calculations ( e.g. * ? ? ?
* ; * ? ? ?
* ) , it is still difficult to reproduce molecular abundances derived by observations , and to determine formation mechanisms of carbon - chain molecules , because of uncertain rate coefficients and of poor knowledge of elementary reactions involving carbon - chain molecules .
another method to study major formation pathways of molecules is based mainly on observations .
recent developments of radio astronomical instruments allow us to detect low - abundance species , including rare isotopologues , with a reasonable observation time . formation mechanisms of some representative carbon - chain molecules have been investigated by observing their @xmath2c isotopic fractionation , such as hc@xmath12n @xcite , cch @xcite , ccs @xcite , c@xmath12s , and c@xmath13h @xcite , toward the cyanopolyyne peak in taurus molecular cloud-1 ( tmc-1 cp ; @xmath14 pc ) , and _ cyclic_-c@xmath12h@xmath15 @xcite toward the low - mass star - forming region l1527 .
tmc-1 cp is a representative cold dark cloud where various carbon - chain molecules are abundant .
for instance , @xcite carried out spectral line survey observations in the @xmath16 ghz region toward tmc-1 cp with the 45-m radio telescope of nobeyama radio observatory ( nro ) , and demonstrated that this source is rich in carbon - chain molecules .
cyanoacetylene , hc@xmath12n , is a very abundant carbon - chain molecule in tmc-1 cp .
@xcite observed three @xmath2c isotopologues of hc@xmath12n toward tmc-1 cp , and the relative abundance of hcc@xmath2cn is significantly higher than that of the other isotopologues : @xmath7h@xmath2cccn@xmath8 : [ $ ] hc@xmath2ccn@xmath8 : [ $ ] hcc@xmath2cn@xmath8 $ ] is @xmath17 .
these results imply that hc@xmath12n is mainly formed by the reaction between a hydrocarbon molecule with two equivalent carbon atoms and a molecule with one carbon atom , whose @xmath10c/@xmath2c ratios are different from each other . according to _ ab initio _ calculations , the reaction of c@xmath15h@xmath15 + cn is exothermic , and has no significant energy barrier @xcite .
the laboratory experiments show that the rate coefficient for this reaction is sufficiently large ( @xmath18 @xmath19 molecule@xmath20 s@xmath20 ) at low temperatures @xcite . considering these results
, @xcite suggested that hc@xmath12n is formed by the reaction between c@xmath15h@xmath15 and cn .
it can naturally be predicted that hc@xmath3n is formed by the similar mechanism of hc@xmath12n , the reaction of c@xmath13h@xmath15 + cn .
@xcite conducted _ ab initio _
calculations on the reactions of c@xmath21h@xmath15 + cn ( @xmath22 ) , which form hc@xmath23n , and found that these reactions are all exothermic and have no energy barriers .
in addition , @xcite measured the rate constant for the reaction c@xmath13h@xmath15 + cn to be ( @xmath24)@xmath2510@xmath26 @xmath19 molecule@xmath20 s@xmath20 at the room temperature .
these results suggest that hc@xmath3n can be formed by the reaction of c@xmath13h@xmath15 + cn .
if so , we would be able to observe different abundances of the five @xmath2c isotopologues of hc@xmath3n .
observations of the @xmath2c isotopologues of hc@xmath3n were carried out toward tmc-1 cp by @xcite .
although they detected the five @xmath2c isotopologues of hc@xmath3n , they could not determine well their relative abundances due to low signal - to - noise ratios ( @xmath27 ) .
@xcite also reported the spectra of the two @xmath2c isotopologues , hc@xmath2ccccn and hcccc@xmath2cn , with slightly higher signal - to - noise ratios ( @xmath28 ) .
however , they could not discuss its main formation pathway , because of the lack of data for the other three @xmath2c isotopologues . in the present paper
, we observed the five @xmath2c isotopologues of hc@xmath3n toward tmc-1 cp with the nro 45-m telescope with high enough signal - to - noise ratios to constrain mechanisms responsible for the formation of hc@xmath3n .
we carried out observations of the normal species and the five @xmath2c isotopologues of hc@xmath3n with the nro 45-m radio telescope in 2014 march , april ( 2013 - 2014 season ) , december , and 2015 january ( 2014 - 2015 season ) .
the observed position was ( @xmath29 , @xmath30 ) = ( 04@xmath3141@xmath324249 , 2541270 ) for tmc-1 cp .
the z45 receiver @xcite and the high electron mobility transistor ( hemt ) amplifier receiver ( h22 ) , both of which can obtain dual - polarization data simultaneously , were used for the observations of the _ j _ = @xmath1 and _ j _ = @xmath0 transitions at 42 ghz and 23 ghz , respectively .
the exact frequencies of the observed lines are given in table [ tab1 ] .
the system temperatures of the z45 and the h22 receivers were from 100 to 130 k , and from 90 to 110 k , respectively .
the beam sizes and the main beam efficiencies ( @xmath33 ) were 37 `` and 0.72 for 42 ghz @xcite , and 72 '' and 0.8 for 23 ghz .
the telescope pointing was checked every 1.5 hours by observing the sio maser line ( _ j _ = 1 - 0 ) from nml tau and gl 5134 , and the pointing error was less than 3 " .
we used the sam45 fx - type digital correlator in frequency settings whose bandwidths and resolutions are 125 mhz and 30.52 khz , respectively , for observations at 42 ghz , and 63 mhz and 15.26 khz , respectively , for those at 23 ghz .
these frequency resolutions correspond to the velocity resolution of about 0.2 km s@xmath20 .
we employed the position - switching mode , where the off - source position was set to be + 30 away in the right ascension .
the smoothed bandpass calibration method @xcite was adopted in the analysis , which allows us to greatly improve the signal - to - noise ratios , compared with the standard position - switch observations .
in fact , the scan pattern of this method was a set of 20 seconds and 5 seconds for on - source and off - source positions , respectively .
we applied 60 and 32 channel - smoothing for the off - source spectra at 23 ghz and 42 ghz , respectively .
the spectra of the five @xmath2c isotopologues of hc@xmath3n were taken with signal - to - noise ratios of @xmath34 , as shown in figures [ fig1 ] .
we also obtained high quality data for the normal species . since the spectra show a symmetric single peak structure , we fitted the spectra with a gaussian profile , and obtained the spectral line parameters , as summarized in table [ tab1 ] .
the line widths ( @xmath35 [ km s@xmath20 ] ) in the 23 ghz region ( @xmath37 km s@xmath20 ) are broader than those in the 42 ghz region ( @xmath38 km s@xmath20 ) .
although the origin of their difference is unclear , the following possibility can be considered .
it is well known that the tmc-1 cp region has a complex velocity structure @xcite . since the telescope beam samples a wider area at 23 ghz than at 42 ghz
, the line width could be larger at 23 ghz .
however , this difference does not affect our discussion on the relative abundance ratios among the five @xmath2c isotopologues seriously , and hence , we do not discuss the line width further .
the values of _
v_@xmath39 are consistent with one another , and are in good agreement with the _ v_@xmath39 value reported for this source ( 5.85 km s@xmath20 , @xcite ) .
the ratio of the integrated intensities ( @xmath40 [ k km s@xmath20 ] ) of the five @xmath2c isotopologues is derived to be @xmath41 ( @xmath42 ) ( 1@xmath6 ) and @xmath43 ( @xmath5 ) ( 1@xmath6 ) for @xmath7h@xmath2cccccn@xmath8 : [ $ ] hc@xmath2ccccn@xmath8 : [ $ ] hcc@xmath2cccn@xmath8 : [ $ ] hccc@xmath2ccn@xmath8 : [ $ ] hcccc@xmath2cn@xmath8 $ ] for the _ j _ = @xmath0 and _ j _ = @xmath1 lines , respectively .
although hcccc@xmath2cn is slightly brighter than the other species , there are no significant differences in intensities among the five @xmath2c isotopologues .
the result contrasts with that for hc@xmath12n , which shows clear differences in intensities of the three @xmath2c isotopologues @xcite .
the rotational population of hc@xmath3n is not completely thermalized to the gas kinetic temperature , because the h@xmath15 density of tmc-1 cp ( @xmath44 @xmath45 @xcite ) is equal to or lower than the critical densities of the hc@xmath3n lines .
however , @xcite showed that the observed brightness temperatures of the hc@xmath3n lines are well approximated by the rotation temperature of 6.5 k. then , we calculated the column densities of the normal species and the five @xmath2c isotopologues of hc@xmath3n using the local thermodynamic equilibrium ( lte ) as shown the following formulae @xcite : @xmath46,\ ] ] where @xmath47 and @xmath48 in equation ( [ tau ] ) , @xmath49 denotes the antenna temperature , _ f _ the beam filling factor , @xmath50 the main beam efficiency ( see section [ obs ] ) , and @xmath51 the optical depth . we used 0.8 and 1 for _ f _ in the 23 ghz and 42 ghz band data , respectively .
we employed these _ f _ values , because a size of the emitting region of carbon - chain molecules in tmc-1 cp is approximately 2.5 according to the mapping observations by @xcite .
@xmath52 is the excitation temperature , @xmath53 the cosmic microwave background temperature ( @xmath54 k ) , and _ j_(_t
_ ) in equation ( [ tem ] ) the planck function .
we derived @xmath55 and @xmath51 simultaneously from the observed intensities of the 23 ghz and 42 ghz lines by using the formula ( 1 ) . in equation ( [ col ] ) , _ n _ is the column density , @xmath35 the line width ( fwhm ) , @xmath56 the partition function , @xmath57 the permanent electric dipole moment of hc@xmath3n ( @xmath58 [ esu cm]@xcite ) , and @xmath59 the energy of the lower rotational energy level . in the energy level calculations , we used the rotational constants of the normal species and the @xmath2c isotopologues @xcite .
the calculated column densities and the @xmath10c/@xmath2c ratios are summarized in table [ tab2 ] . for the normal species , the excitation temperature and the column density
are determined to be @xmath60 k and ( @xmath61)@xmath62 @xmath63 , respectively , where the errors are estimated from the gaussian fits , and represent the confidence level of one standard deviation .
the optical depths for the _ j _ = @xmath0 and @xmath1 lines are @xmath64 and @xmath65 ( 1@xmath6 ) , respectively .
the obtained excitation temperature and column density of the normal species are consistent with the previous result ( @xmath60 k and ( @xmath66)@xmath62 @xmath63 ) @xcite within the error ranges .
we calculated the column densities of the five @xmath2c isotopologues using equation ( [ col ] ) and the excitation temperature derived for the normal species ( 6.5 k ) .
the calculated column densities of the normal species and the five @xmath2c isotopologues , and the @xmath10c/@xmath2c ratios are summarized in table [ tab2 ] .
for the five @xmath2c isotopologues , the column densities were derived to be ( @xmath67)@xmath68 , ( @xmath69)@xmath68 , ( @xmath70)@xmath68 , ( @xmath71)@xmath68 , and ( @xmath72)@xmath68 @xmath63 , for h@xmath2cccccn , hc@xmath2ccccn , hcc@xmath2cccn , hccc@xmath2ccn , and hcccc@xmath2cn , respectively .
the abundance ratio of the @xmath2c isotopologues of hc@xmath3n is then @xmath4 ( @xmath5 ) ( 1@xmath6 ) for @xmath7h@xmath2cccccn@xmath8 : [ $ ] hc@xmath2ccccn@xmath8 : [ $ ] hcc@xmath2cccn@xmath8 : [ $ ] hccc@xmath2ccn@xmath8 : [ $ ] hcccc@xmath2cn@xmath8 $ ] .
this result agrees with those obtained from integrated intensities .
the averaged @xmath10c/@xmath2c ratio of hc@xmath3n is @xmath9 ( 1@xmath6 ) , which is slightly higher than the elemental ratio of @xmath11 in the local interstellar medium ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the intensities are also affected by the calibration errors ( main beam efficiency and filling factor ) .
however , we obtained the data of the normal species and the five @xmath2c isotopologues simultaneously .
then , these calibration errors do not affect the @xmath10c/@xmath2c ratios .
we discuss these ratios in section [ dil ] .
we investigated the sensitivities of the assumed @xmath52 values on the column densities and the @xmath10c/@xmath2c ratios .
we calculated the column densities of the normal species and the five @xmath2c isotopologues using @xmath52 of 5.9 k and 7.1 k , which are the lowest and highest values in the 3@xmath6 error range of the excitation temperature .
we derived the column densities of the normal species to be ( @xmath73)@xmath62 @xmath63 and ( @xmath74)@xmath62 @xmath63 for 5.9 k and 7.1 k , respectively .
the column densities of the @xmath2c isotopologues were not significantly affected by the change of the excitation temperature .
the @xmath10c/@xmath2c ratios averaged for the five isotopologues are @xmath75 ( 1@xmath6 ) and @xmath76 ( 1@xmath6 ) for 5.9 k and 7.1 k , respectively .
although the @xmath10c/@xmath2c ratios slightly depend on the excitation temperature , the @xmath10c/@xmath2c ratio of hc@xmath3n is still higher than that of the elemental ratio in the local interstellar medium .
we also evaluated the ratio by using the excitation temperature derived for each @xmath2c isotopologue .
we derived @xmath52 to be @xmath77 , @xmath77 , @xmath78 , @xmath66 , and @xmath78 k ( 1@xmath6 ) for h@xmath2cccccn , hc@xmath2ccccn , hcc@xmath2cccn , hccc@xmath2ccn , and hcccc@xmath2cn , respectively , from the intensities of the observed two rotational transitions .
these results are consistent with the value obtained from the normal species ( 6.5 k ) within the errors , which implies that we can neglect the effect of self - trapping for the normal species .
we then calculated the column densities using @xmath52 obtained for each @xmath2c isotopologue .
the @xmath2c isotopic fractionation is not affected by the changes in @xmath52 , and the @xmath10c/@xmath2c ratios are also not affected within the errors .
the results are also shown in table [ tab2 ] .
we here consider possible reactions leading to cyanopolyynes , using the umist database for astrochemistry 2012 @xcite . by focusing on the rate constants , activation energies , abundances and formation pathways of precursors
, we can sketch formation pathways of cyanopolyynes , as shown in figure [ fig3 ] .
three types of formation pathways are possible as follows : + pathway 1 : the reactions of c@xmath21h@xmath15 + cn , + pathway 2 : the growth of the cyanopolyyne carbon chains via c@xmath15h@xmath15@xmath79 + hc@xmath23n , and + pathway 3 : the reactions between hydrocarbon ions and nitrogen atoms followed by electron recombination reactions . for hc@xmath12n , pathway 1 is suggested to play a dominant role in its production : @xcite reported that the main formation pathway of hc@xmath12n is the reaction of c@xmath15h@xmath15 + cn on the basis of their observational results of @xmath2c isotopic fractionation .
fractionation of @xmath2c in cn proceeds through the following reaction @xcite : @xmath80 this reaction is exothermic with energy of 34 k due to decreasing the zero - point vibrational energy .
hence , its backward reaction is not effective in cold dark clouds , where the typical temperature is approximately 10 k ( e.g. , * ? ? ?
as a result , the @xmath10c/@xmath2c ratio for cn decreases in comparison with those for hydrocarbons such as c@xmath15h@xmath15 and c@xmath13h@xmath15 , because similar fractionation processes of @xmath2c do not occur in hydrocarbons @xcite . if hc@xmath3n is mainly formed by pathways 1 and/or 2 , at least one carbon atom in hc@xmath3n , possibly that adjacent to the nitrogen atom , is not equivalent to the others . assuming that hc@xmath3n is mainly formed by pathway 2 ,
the @xmath2c isotopic fractionation of hc@xmath3n should be @xmath81 , where @xmath82 and @xmath83 are arbitrary values , for @xmath7h@xmath2cccccn@xmath8 : [ $ ] hc@xmath2ccccn@xmath8 : [ $ ] hcc@xmath2cccn@xmath8 : [ $ ] hccc@xmath2ccn@xmath8 : [ $ ] hcccc@xmath2cn@xmath8 $ ] , unless scrambling of the carbon atoms is efficient .
however , our observed abundances are roughly equal for all the @xmath2c isotopologues .
it is not likely that pathway 1 and 2 are the main formation pathways of hc@xmath3n . on the basis of the above consideration
, we propose that pathway 3 should significantly contribute to formation of hc@xmath3n . in pathway 3 , all carbon atoms in hc@xmath3n originate from the hydrocarbon ions . if all the carbon atoms in the mother hydrocarbon ion have the similar @xmath10c/@xmath2c ratio , the observational result can be explained .
this seems indeed possible for a large hydrocarbon ion , because it is generally produced through various processes .
this pathway is similar to the chemical model calculation by @xcite , where hc@xmath3n can be produced through the reaction between a hydrocarbon ion ( c@xmath3h@xmath12@xmath79 ) and a nitrogen atom .
thus , we propose that pathway 3 overwhelms pathways 1 and 2 , resulting in almost equivalent abundances of the @xmath2c isotopologues of hc@xmath3n . based on the above discussion on hc@xmath3n
, we also propose that the longer cyanopolyynes may be formed by a mechanism similar to that of hc@xmath3n , namely the reactions between hydrocarbon ions and nitrogen atoms .
however , it is difficult to confirm our suggestion by observations of @xmath2c isotopic fractionation at the present . @xcite
tentatively detected the @xmath2c isotopologues of hc@xmath84n using the _ j _ = @xmath85 and @xmath86 transitions for the first time in the interstellar medium .
they calculated the averaged @xmath10c/@xmath2c ratio of hc@xmath84n , but they could not study its @xmath2c isotopic fractionation as they only derived upper limits for each of the seven @xmath2c isotopologues . hence , more sensitive observations are needed to confirm the flat @xmath2c abundance over the seven @xmath2c isotopologues of hc@xmath84n . in the theoretical model by @xcite , the pathways leading to hc@xmath84n and hc@xmath87n via reactions between hydrocarbon ions ( c@xmath84h@xmath12@xmath79 and c@xmath87h@xmath12@xmath79 , respectively ) and nitrogen atoms followed by the electron recombination reactions are presented . the most interesting point to emerge from our analysis
is that cyanopolyynes longer than hc@xmath3n are mainly produced by the reactions between hydrocarbon ions and nitrogen atoms , whereas hc@xmath12n is mainly produced by the neutral - neutral reaction between c@xmath15h@xmath15 and cn .
one possible reason is high abundance of c@xmath15h@xmath15 @xcite .
we compile @xmath10c/@xmath2c ratios of some carbon - chain molecules at tmc-1 cp in table [ tab3 ]
. the @xmath10c/@xmath2c ratios of carbon - chain molecules show more or less higher than that of the local interstellar elemental ratio ( @xmath11 ) ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this can be explained by the following reaction , which should be effective under low - temperature conditions ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , @xmath88 the reaction leads to a reduction in @xmath2c@xmath79 .
consequently , @xmath2c isotopologues of carbon - chain molecules are also reduced , because c@xmath79 and c are important raw materials of carbon - chain molecules .
we derived the five @xmath10c/@xmath2c ratios of hc@xmath3n , and they are slightly higher than those of hc@xmath12n ( see table [ tab3 ] ) .
@xcite were able to detect the @xmath2c isotopologues only by combining the observations of all seven isotopologues of hc@xmath84n into one spectrum .
they derived the averaged @xmath10c/@xmath2c ratio of hc@xmath84n by combination of multiple lines of the seven @xmath2c isotopologues of hc@xmath84n .
these ratios indicate that the @xmath2c species of the cyanopolyynes are less diluted than other carbon - chain molecules .
for example , for hc@xmath12n and hc@xmath3n , the @xmath10c/@xmath2c ratios coincide among their @xmath2c isotopologues within about 20% and 10% , respectively . on the other hand , for ccs
, @xmath10c/@xmath2c of @xmath2ccs is about 4 times larger than that of c@xmath2cs . in other words ,
an interesting characteristic of the @xmath10c/@xmath2c ratios of the two cyanopolyynes is that their differences appear to be small among the @xmath2c isotopologues of each cyanopolyyne , compared to other carbon - chain molecules .
the reason why the @xmath2c species of the cyanopolyynes are not significantly diluted is unclear
. mechanisms of the dilution of the @xmath2c species are still controversial and are left for future studies .
we carried out observations of the _ j _ = @xmath0 and @xmath1 rotational transitions of the normal species and the five @xmath2c isotopologues of hc@xmath3n toward tmc-1 cp with the nro 45-m telescope .
the abundance ratios of the five @xmath2c isotopologues are found to be @xmath4 ( @xmath5 ) ( 1@xmath6 ) for @xmath7h@xmath2cccccn@xmath8 : [ $ ] hc@xmath2ccccn@xmath8 : [ $ ] hcc@xmath2cccn@xmath8 : [ $ ] hccc@xmath2ccn@xmath8 : [ $ ] hcccc@xmath2cn@xmath8 $ ] .
based on the result , we suggest that reactions between hydrocarbon ions and nitrogen atoms play an important role in the formation of hc@xmath3n .
we derived the @xmath10c/@xmath2c ratios from the five @xmath2c isotopologues of hc@xmath3n .
the averaged value is @xmath9 ( 1@xmath6 ) , and the @xmath2c species of cyanopolyynes are less diluted than other carbon - chain molecules .
we are grateful to the staff of nobeyama radio observatory .
nobeyama radio observatory is a branch of the national astronomical observatory of japan , national institutes of natural sciences . in particular , we deeply appreciate mr . jun maekawa for developing a software for smoothed bandpass calibration analysis , and dr .
izumi mizuno for discussing the method of smoothed bandpass calibration analysis .
we express our thanks to prof .
kazuhito dobashi and dr .
tomomi shimoikura ( tokyo gakugei university ) for discussing analysis of the data obtained by using the z45 receiver , and to all the z45 receiver group members for their kind support .
the z45 receiver is supported in part by a grant - in - aid for science research of japan ( 24244017 ) and the national science foundation under grant ( nsf phy11 - 25915 ) .
alexander , a. j. , kroto , h. w. , & walton , d. r. m. 1976 , jmosp , 62 , 175 benson , p.j . , & myers , p. c. 1989 , , 71 , 89 dickens , j. e. , langer , w. d. , & velusamy , t. 2001 , , 558 , 693 fukuzawa , k. , osamura , y. , & schaefer , h. f. 1998 , , 505 , 278 hirahara , y. , suzuki , h. , yamamoto , s. , et al .
1992 , , 394 , 539 kaifu , n. , ohishi , m. , kawaguchi , k. , et al .
2004 , , 56 , 69 kaiser , m. e. , hawkins , i. , & wright , e. l. 1991 , , 379 , 267 langer , w. d. , graedel , t. e. , frerking , m. a. , & armentrout , p. b. 1984 , , 277 , 581 langer , w. d. , & penzias , a. a. 1990 , , 357 , 477 langer , w. d. , & penzias , a. a. 1993 , , 408 , 539 langer , w. d. , velusamy , t. , kuiper , t. b. h. , et al .
1995 , , 453 , 293 langston , g. , & turner , b. 2007 , , 658 , 455 markwick , a. j. , millar , t. j. , & charnley , s. b. 2000 , , 535 , 256 mcelroy , d. , walsh , c. , markwick , a. j. , et al .
2013 , , 550 , a36 milam , s. n. , savage , c. , brewster , m. a. , ziurys , l. m. , & wyckoff , s. 2005 , , 634 , 1126 m@xmath89ller , h. s. p. , schl@xmath90der , f. , stutzki , j. , & winnewisser , g. 2005 , jmost , 742 , 215 nakamura , f. , ogawa , h. , yonekura , y. , et al .
2015 , , in press ( arxiv:1509.02631 ) sakai , n. , ikeda , m. , morita , m. , et al .
2007 , , 663 , 1174 sakai , n. , saruwatari , o. , sakai , t. , takano , s. , & yamamoto , s. 2010 , , 512 , a31 sakai , n. , takano , s. , sakai , t. , et al .
2013 , jpca , 117 , 9831 sanz , m. e. , mccarthy , m. c. , & thaddeus , p. 2005
, , 122 , 194319 savage , c. , apponi , a. j. , ziurys , l. m. , & wyckoff , s. 2002 , , 578 , 211 seki , k. , yagi , m. , he , m. , halpern , j. b. , & okabe , h. 1996 , cpl , 258 , 657 sims , i. r. , queffelec , j. -l . , travers , d. , et al . 1993 , cpl , 211 , 461 suzuki , h. , yamamoto , s. , ohishi , m. , et al .
1992 , , 392 , 551 takano , s. , suzuki , h. , ohishi , m. , et al .
1990 , , 361 , l15 takano , s. , masuda , a. , hirahara , y. , et al .
1998 , , 329 , 1156 wakelam , v. , loison , j. -c . ,
herbst , e. , et al .
2015 , , 217 , 20 watson , w. d. , anicich , v. g. , & huntress , jr . , w. t. 1976 , , 205 , l165 woon , d. e. , & herbst , e. 1997 , , 477 , 204 yamaki , h. , kameno , s. , beppu , h. , mizuno , i. , & imai , h. 2012 , , 64 , 118 yoshida , k. , sakai , n. , tokudome , t. , et al .
2015 , , 807 , 66 lccccccc hc@xmath3n & @xmath0 & 23.9639007 ( 1 ) & 1600 ( 20 ) & 0.88 ( 1 ) & 5.90 ( 9 ) & 1.50 ( 3 ) & 2.6 + & @xmath1 & 42.6021529 ( 2 ) & 1874 ( 57 ) & 0.46 ( 2 ) & 5.9 ( 2 ) & 0.93 ( 4 ) & 3.3 + h@xmath2cccccn & @xmath0 & 23.3400887 ( 5 ) & 44 ( 2 ) & 0.57 ( 3 ) & 6.0 ( 3 ) & 0.027 ( 2 ) & 2.2 + & @xmath1 & 41.4931720 ( 9 ) & 31 ( 3 ) & 0.44 ( 4 ) & 5.8 ( 5 ) & 0.0147 ( 19 ) & 3.2 + hc@xmath2ccccn & @xmath0 & 23.7183241 ( 5 ) & 35 ( 2 ) & 0.67 ( 4 ) & 5.8 ( 3 ) & 0.025 ( 2 ) & 2.4 + & @xmath1 & 42.1655778 ( 9 ) & 31 ( 3 ) & 0.44 ( 4 ) & 5.8 ( 4 ) & 0.0148 ( 16 ) & 3.0 + hcc@xmath2cccn & @xmath0 & 23.9390819 ( 6 ) & 25 ( 1 ) & 0.97 ( 6 ) & 5.8 ( 3 ) & 0.026 ( 2 ) & 2.1 + & @xmath1 & 42.5580316 ( 9 ) & 31 ( 3 ) & 0.49 ( 4 ) & 5.8 ( 4 ) & 0.0165 ( 16 ) & 3.1 + hccc@xmath2ccn & @xmath0 & 23.9419974 ( 5 ) & 31 ( 2 ) & 0.81 ( 5 ) & 5.8 ( 3 ) & 0.027 ( 2 ) & 2.1 + & @xmath1 & 42.5632150 ( 9 ) & 35 ( 3 ) & 0.45 ( 4 ) & 5.9 ( 4 ) & 0.0166 ( 19 ) & 3.1 + hcccc@xmath2cn & @xmath0 & 23.7271659 ( 6 ) & 43 ( 2 ) & 0.69 ( 4 ) & 5.9 ( 3 ) & 0.031 ( 2 ) & 2.4 + & @xmath1 & 42.1812962 ( 10 ) & 35 ( 3 ) & 0.47 ( 4 ) & 5.8 ( 5 ) & 0.0172 ( 18 ) & 3.0 + lcccc hc@xmath3n & ( 6.2 @xmath91 0.3 ) @xmath92 & & & + h@xmath2cccccn & @xmath67 & @xmath93 & @xmath94 & @xmath95 + hc@xmath2ccccn & @xmath69 & @xmath96 & @xmath97 & @xmath98 + hcc@xmath2cccn & @xmath70 & @xmath99 & @xmath97 & @xmath98 + hccc@xmath2ccn & @xmath71 & @xmath100 & @xmath101 & @xmath102 + hcccc@xmath2cn & @xmath72 & @xmath103 & @xmath104 & @xmath105 + lcc hc@xmath12n / h@xmath2cccn & 79 @xmath91 11 ( 1@xmath6 ) & 1 + hc@xmath12n / hc@xmath2ccn & 75 @xmath91 10 ( 1@xmath6 ) & 1 + hc@xmath12n / hcc@xmath2cn & 55 @xmath91 7 ( 1@xmath6 ) & 1 + hc@xmath3n / h@xmath2cccccn & 98 @xmath91 14 ( 1@xmath6 ) & 2 + hc@xmath3n / hc@xmath2ccccn & 101 @xmath91 14 ( 1@xmath6 ) & 2 + hc@xmath3n / hcc@xmath2cccn & 95 @xmath91 12 ( 1@xmath6 ) & 2 + hc@xmath3n / hccc@xmath2ccn & 93 @xmath91 13 ( 1@xmath6 ) & 2 + hc@xmath3n / hcccc@xmath2cn & 85 @xmath91 11 ( 1@xmath6 ) & 2 + hc@xmath84n / average @xmath2c isotopologues & 87 @xmath106 ( 1@xmath6 ) & 3 + cch/@xmath2cch & > 250 & 4 + cch / c@xmath2ch & > 170 & 4 + c@xmath13h/@xmath2cccch & 141 @xmath91 44 ( 3@xmath6 ) & 5 + c@xmath13h / c@xmath2ccch & 97 @xmath91 27 ( 3@xmath6 ) & 5 + c@xmath13h / cc@xmath2cch & 82 @xmath91 15 ( 3@xmath6 ) & 5 + c@xmath13h / ccc@xmath2ch & 118 @xmath91 23 ( 3@xmath6 ) & 5 + ccs/@xmath2ccs & 230 @xmath91 130 ( 3@xmath6 ) & 6 + ccs / c@xmath2cs & 54 @xmath91 5 ( 3@xmath6 ) & 6 + c@xmath12s/@xmath2cccs & > 206 ( 3@xmath6 ) & 5 + c@xmath12s / c@xmath2ccs & 48 @xmath91 15 ( 3@xmath6 ) & 5 + c@xmath12s / cc@xmath2cs & 30 - 206 & 5 +
|
we observed the _ j _ = @xmath0 and @xmath1 rotational transitions of the normal species and five @xmath2c isotopologues of hc@xmath3n to study its formation mechanisms toward the cyanopolyyne peak in taurus molecular cloud-1 , with the 45-m radio telescope of nobeyama radio observatory .
we detected the five @xmath2c isotopologues with high signal - to - noise ratios between 12 and 20 , as well as the normal species .
the abundance ratios of the five @xmath2c isotopologues of hc@xmath3n are found to be @xmath4 ( @xmath5 ) ( 1@xmath6 ) for @xmath7h@xmath2cccccn@xmath8 : [ $ ] hc@xmath2ccccn@xmath8 : [ $ ] hcc@xmath2cccn@xmath8 : [ $ ] hccc@xmath2ccn@xmath8 : [ $ ] hcccc@xmath2cn@xmath8 $ ] .
we do not find any significant differences among the five @xmath2c isotopologues .
the averaged [ hc@xmath3n]/[@xmath2c isotopologues ] abundance ratio is determined to be @xmath9 ( 1@xmath6 ) , which is slightly higher than the local interstellar elemental @xmath10c/@xmath2c ratio of @xmath11 .
possible formation pathways are discussed on the basis of these results .
|
compact neutron star binaries are among the most promising sources of gravitational waves . at low frequencies ,
the continuous inspiral signal may be detectable by the space antenna lisa , while ground based interferometers such as virgo @xcite , ligo @xcite , geo @xcite or tama @xcite , are expected to detect the last few minutes prior coalescence , at frequencies up to 1.4 - 1.6 khz . in a first paper @xcite ( hereafter paper i ) ,
we have investigated the high frequency signal and its detection with the first generations of ground based detectors . using new estimates of the mean merging rate in the local universe , that account for the galactic star formation history derived directly from observations and
include the contribution of elliptical galaxies , we predict a detection every 148 and 125 years in the volume probed by initial virgo and ligo , and up to 6 detections per year in their advanced configuration . in a second paper @xcite ( hereafter paper ii ) , we used numerical simulations to estimate the gravitational wave emission produced by the superposition of unresolved extra - galactic sources .
as in paper i , we were interested in the few sixteen minutes before the last stable orbit is reached , when more than 96% of the gravitational wave energy is released and when the frequency evolves in the range 10 - 1500 hz , covered by ground based interferometers .
we find that above the critical redshift @xmath4 , the sources produce a truly continuous background , with a maximal gravitational density parameter of @xmath5 around 670 hz .
the signal is below the sensitivity of the first generation of ground based detectors but could be detectable by the next generation .
correlating two coincident advanced - ligo detectors or two third generation ego interferometers , we derived a s / n ratio of 0.5 and 10 respectively . in this article , we extend our previous simulations to investigate the stochastic background from the early low frequency inspiral phase
. estimates of the emission from the various population of compact binaries ( see for instance @xcite for the galactic contribution and @xcite for the extra - galactic contribution ) or from captures by supermassive black holes ( @xcite ) , which represent the main source of confusion noise for lisa , are of crucial interest for the development of data analysis strategies . the paper will be organized as follow . in ii . , we describe the simulations of the dns population , in iii . the contribution of inspiralling dns to the stochastic background is calculated and compared to the lisa instrumental noise and to the wd - wd galactic foreground . finally , in iv .
the main conclusions are summarized .
to generate a population of dnss ( each one characterized by its redshift of formation and coalescence timescale ) we follow the same procedure as in paper ii , to which the reader is referred for a more detailed description .
the redshift of formation is randomly selected from the probability distribution @xcite : @xmath6 constructed by normalizing the differential dns formation rate , @xmath7 the normalization factor in the denominator corresponds to the rate at which massive binaries are formed in the considered redshift interval , e.g. , @xmath8 which numerically gives @xmath9 s@xmath10 in the two cosmic star formation rate adopted in this paper . the formation rate of massive binaries per redshift interval eq .
[ eq - frate1 ] writes : @xmath11 where @xmath12 is the cosmic star formation rate ( sfr ) expressed in m@xmath13 mpc@xmath14yr@xmath10 and @xmath15 is the mass fraction converted into dns progenitors .
hereafter , rates per comoving volume will always be indicated by the superscript * , while rates with indexes z@xmath16 " or
z@xmath17 " refer to differential rates per redshift interval , including all cosmological factors . the ( 1+z ) term in the denominator of eq .
[ eq - frate ] corrects the star formation rate by time dilatation due to the cosmic expansion .
the element of comoving volume is given by @xmath18 with @xmath19^{1/2}\ ] ] where @xmath20 and @xmath21 are respectively the present values of the density parameters due to matter ( baryonic and non - baryonic ) and vacuum , corresponding to a non - zero cosmological constant . throughout this paper , the 737 flat cosmology @xcite , with @xmath20 = 0.30 , @xmath21 = 0.70 and hubble parameter h@xmath22
= @xmath23 @xcite is assumed .
recent measurements of galaxy luminosity function in the uv ( sdss , galex , combo17 ) and fir wavelengths ( spitzer space telescope ) , after dust corrections and normalization , allowed to refine the previous models of star formation history , up to redshift @xmath24 , with tight constraints at redshifts @xmath25 . in our computations
, we consider the recent parametric fits of the form of @xcite , provided by @xcite , constrained by the super kamiokande limit on the electron antineutrino flux from past core - collapse supernovas .
it is worth mentioning that the final results are not significantly different if we adopt , as in paper ii , the sfr given in @xcite . throughout this paper , we assume that the parameter @xmath26 does not change significantly with the redshift and can be considered as a constant .
in fact , this term is the product of three other parameters , namely , @xmath27 where @xmath28 is the fraction of binaries which remains bounded after the second supernova event , @xmath29 is the fraction of massive binaries formed among all stars and @xmath30 is the mass fraction of neutron star progenitors . from paper
i we take @xmath28 = 0.024 and @xmath29 = 0.136 .
@xcite investigated the effect of the initial mass function ( imf ) assumption on the normalization of the sfr .
they showed that top heavy imfs are preferred to the traditional salpeter imf @xcite and their fits are optimized for imfs of the form : @xmath31 with a turnover below @xmath32 m@xmath13 , and normalized within the mass interval 0.1 - 100 m@xmath13 such as @xmath33 = 1 . in this paper , we assume @xmath34 ( a modified salpeter ) , but taking @xmath35 ( @xcite ) would not change the final results significantly . to be consistent with the adopted models of the sfr
, we follow @xcite and assume a minimal initial mass of 8 m@xmath13 for ns progenitors , the upper mass being 40 m@xmath13 .
it results finally , @xmath36 m@xmath37 . the next step consists to estimate the redshift @xmath38 at which the progenitors have already evolved and the system is now constituted by two neutron stars .
this moment fixes also the beginning of the inspiral phase .
if @xmath39 ( @xmath40 yr ) is the mean lifetime of the progenitors ( average weighted by the imf in the interval 8 - 40 m@xmath13 ) then @xmath41 once the beginning of the inspiral phase is established , the redshift at which the coalescence occurs should now be estimated .
the duration of the inspiral depends on the orbital parameters and the neutron star masses .
the probability for a given dns system to coalesce in a timescale @xmath42 was initially derived by @xcite and confirmed by subsequent simulations @xcite and is given by @xmath43 simulations indicate a minimum coalescence timescale @xmath44 yr and a considerable number of systems having coalescence timescales higher than the hubble time . the normalized probability in the range @xmath45 yr up to 20 gyr implies @xmath46 .
therefore , the redshift @xmath47 at which the coalescence occurs is derived by solving the equation @xmath48 where @xmath49
compared with the previous study , slight complications arise when calculating the spectral properties of the signal . in paper ii , we were interested in the signal emitted between 10 - 1500 hz , which duration in the source frame is short enough , less than 1000 s , to be considered as a burst located at the redshift of coalescence .
now , as we are looking at inspiral phase in the lisa frequency window ( @xmath50 hz ) , the evolution of the redshift of emission with frequency must be taken into account .
for the same reason , the critical redshift at which the population of dnss constitute a truly continuous background ca nt be determined by simply solving the condition @xmath51 , where d is the duty cycle defined as the ratio of the typical duration of a single burst @xmath52 to the average time interval between successive events ( see paper ii ) .
instead , we remove the brightest sources , lying in the close universe , where the distribution of galaxies is expected to be highly anisotropic , in particular due to the strong concentration of galaxies in the direction of the virgo cluster and the great attractor ( see paper i ) , and where the density is well below the average value derived from the sfr ( @xcite ) . in a second step
, we verify that the resulting background is continuous , by using both jaque bera and lilliefors gaussianity tests with significant level @xmath53 over a set of 100 realizations . for each realization
, the number of simulated dnss corresponds to the expected number of extragalactic dnss observed today and is derived as , @xmath54 the dnss present at redshift z were formed at @xmath55 , with a coalescence time larger than the cosmic time between @xmath56 and z. thus , @xmath57 where @xmath58 with @xmath59= 20 gyr and @xmath60=max(@xmath45 yr;@xmath61 ) , which gives @xmath62 . we obtain that after @xmath0 , corresponding to the distance beyond the virgo cluster ( @xmath63 mpc ) , when the relative density fluctuations are @xmath64 ( @xcite ) , the background is continuous in the range @xmath50 hz .
we consider also a more conservative threshold of @xmath65 , corresponding to the distance beyond the great attractor ( @xmath66 mpc ) , when the relative density fluctuations are @xmath67 and when the universe is expected to become homogeneous(@xcite ) .
the gravitational flux in the observer frame produced by a given dns is : @xmath68 where @xmath69 is the redshift of emission , @xmath70 the distance luminosity , @xmath71 the proper distance , which depends on the adopted cosmology , @xmath72 the gravitational spectral energy and @xmath73 the frequency of emission in the source frame . in the quadripolar approximation and for a binary system with masses @xmath74 and @xmath75 in a circular orbit : @xmath76 where the fact that the gravitational wave frequency is twice the orbital frequency was taken into account .
then @xmath77 taking @xmath78 m@xmath13 , one obtains @xmath79 erg hz@xmath80 .
the redshift of emission is derived by solving the relation @xcite : @xmath81 where @xmath82 ( or numerically @xmath83 yr@xmath10 ) , and where the time spent between redshifts @xmath69 and @xmath38 is given by : @xmath84 the spectral properties of the stochastic background are characterized by the dimensionless parameter @xcite : @xmath85 where @xmath86 is the wave frequency in the observer frame and @xmath87 the critical mass density needed to close the universe , related to the hubble parameter @xmath88 by , @xmath89 @xmath90 is the gravitational wave flux at the observer frequency ( given here in erg @xmath91hz@xmath10s@xmath10 ) @xmath92 , integrated over all sources above the critical redshift @xmath93 , namely @xmath94 in our simulations , the integrated gravitational flux is calculated by summing individual fluences ( eq .
[ eq - fluence ] ) , scaled by the ratio between the total formation rate of progenitors ( eq .
[ eq - revents ] ) and the number of simulated massive binaries , as : @xmath95 figure [ fig - omega ] shows the density parameter @xmath96 as a function of the observed frequency averaged over 10 simulations .
the continuous line corresponds to sources located beyond the critical redshift to have a continuous background , @xmath0 ( or @xmath97 mpc ) , while the dashed line corresponds to the more conservative estimate , with a cutoff at @xmath98 ( or @xmath99 mpc ) .
after a fast increase below @xmath100 hz , @xmath96 increases as @xmath101 to reach @xmath102 at 0.1 hz for @xmath0 ( @xmath103 for @xmath98 ) .
this is one order of magnitude larger than the previous predictions by @xcite .
, corresponding to ns - ns inspirals occurring beyond @xmath104 mpc ( continuous curve ) and @xmath99 mpc ( dashed curve ) .
[ fig - omega ] ] astrophysical backgrounds represent a confusion noise for lisa , which spectral density is given by @xcite : @xmath105 figure [ fig - sh ] shows @xmath106 compared to the lisa sensitivity @xcite and to the contribution of unresolved galactic wd - wd @xcite .
the contribution from dnss may dominate the lisa instrumental noise between @xmath107 - @xmath108 hz for @xmath0 ( @xmath109 - @xmath110 hz for @xmath98 ) and the galactic wd - wd confusion noise after @xmath111 .
however , the resulting reduction in the sensitivity should be less than a factor 4 and thus should nt affect significantly signal detection . , corresponding to ns - ns inspirals occurring beyond @xmath104 mpc ( continuous curve ) and @xmath99 mpc ( dashed curve ) , along with the lisa instrumental noise ( black ) and the galactic wd - wd foreground ( light gray ) .
[ fig - sh ] ] the critical redshift at which a single dns can be resolved and detected by lisa ( figure [ fig - zmax ] ) is at least one order of magnitude smaller ( @xmath112 at frequencies @xmath113 ) than our threshold to have a gaussian background @xmath0 ( or @xmath98 ) .
sources lying between the two regimes are responsible for a non gaussian and non isotropic cosmic popcorn " noise ( see @xcite for a recent review ) that will be investigated in a further work .
this contribution , which statistical properties differ from those of both the instrumental noise and the cosmological background , could be detected and removed by new data analysis techniques currently under investigation , such as the search for anisotropies @xcite that can be used to create a map of the gw background @xcite , the maximum likelihood statistic @xcite , or methods based on the probability event horizon concept @xcite , which describes the evolution , as a function of the observation time , of the cumulated signal throughout the universe . .
the galactic wd - wd foreground is taken into account in the lisa sensitivity .
[ fig - zmax ] ]
in this paper , we have modelled the contribution of extra galactic inspiralling double neutron stars , to the to the astrophysical gravitational wave foreground , in the range of sensitivity of the space detector lisa . using a recent fit of the star formation rate @xcite , optimized for a top heavy a modified salpeter imf ,
we show that sources beyond @xmath0 ( or @xmath98 for a more conservative estimate ) constitute a truly continuous background .
we find that the density parameter @xmath96 , after a fast increase below @xmath100 hz , increases as @xmath101 to reach @xmath114 ( @xmath103 ) at 0.1 hz , which is one order of magnitude above the previous estimate of @xcite . as a result ,
the signal may dominate the lisa instrumental noise in the range @xmath107 - @xmath108 hz ( @xmath109 - @xmath110 ) and overwhelm the galactic wd - wd confusion noise at frequencies larger than @xmath3 .
sources located closer than @xmath93 , but still too far to be detected by lisa , constitute a non gaussian and non isotropic cosmic popcorn " noise , which will be investigated in a further work .
k. kuroda et al . , in _ proceedings of the international conference on gravitational waves : sources and detectors , 1997 _ edited by i. ciufolini and f. fidecaro ( singapore : world scientific , 1997 ) , p. 1007 .
|
we review the contribution of extra galactic inspiralling double neutron stars , to the lisa astrophysical gravitational wave foreground . using recent fits of the star formation rate ,
we show that sources beyond @xmath0 contribute to a truly continuous background , which may dominate the lisa instrumental noise in the range @xmath1 - @xmath2 hz and overwhelm the galactic wd - wd confusion noise at frequencies larger than @xmath3 .
|
quasars have been used as background sources to study the gaseous phases of a variety of objects that are located along our sight - lines to them .
these objects include not only _ intervening _ absorbers such as intervening galaxies , the intergalactic medium ( igm ) , clouds in the halo of the milky way , and the host galaxies of the quasars themselves , but also _ intrinsic _ absorbers that are physically associated with the quasar central engines .
one of the most promising candidates for the intrinsic absorbers are outflowing winds from the quasars that could be accelerated by radiation pressure from the accretion disk ( murray et al .
1995 ; arav et al . 1995 ; proga et al .
2000 ) or by magnetocentrifugal forces ( e.g. , everett 2005 ) .
outflowing winds are important components of quasar central engines because they carry away angular momentum from the accretion disk and allow the remaining gas in the disk to accrete onto the central black hole .
quasar outflows are also important for cosmology since they deliver energy , momentum , and metals to the interstellar and intergalactic media , thus significantly affecting star formation and galaxy assembly ( e.g. , granato et al .
2004 ; scannapieco & oh 2004 ; springel , di matteo & hernquist 2005 ) .
intrinsic narrow absorption lines ( nals ; fwhm @xmath3 ) present a powerful way to determine physical parameters of the accretion disk winds . unlike the _ broad _ absorption lines ( bals ; fwhm @xmath4 ; weymann et al .
1991 ) that are also associated with quasars outflows , nals do not suffer from self - blending or from saturation , so that line parameters are more easily evaluated .
bals are thought to be associated with radiatively - driven outflows from accretion disks ( e.g. , weymann et al .
1991 ; becker et al .
1997 ) , while it is difficult to separate intrinsic nals from intervening nals .
mini - broad absorption lines ( mini - bals ) are an intermediate subclass between nals and bals , which are typically wider than those of nals , but narrower than bals .
mini - bals have the advantages of both bals ( i.e. , high probability of being intrinsic lines ) and nals ( i.e. , line profiles can be resolved into individual components ) , which make them useful targets ( e.g. , churchill et al .
1999 ; hamann et al .
1997a ; narayanan et al .
bals could probe the low - latitude , dense , fast portion of the wind , while the mini - bals and nals may probe the lower - density portion of the wind at high latitudes above the disk .
thus , the study of mini - bals and nals complements the study of bals because the corresponding absorbers reside in different regions of the outflow and allow us to sample different sets of physical conditions . to isolate intrinsic nals ,
two tests are commonly used : ( i ) partial coverage , i.e. , trough dilution by un - occulted light ( e.g. , hamann et al . 1997b ; barlow & sargent 1997 ; ganguly et al .
1999 ) , and ( ii ) variability of the absorption profiles within a few years in the quasar rest frames ( e.g. , hamann et al . 1997a ; narayanan et al . 2004 ; wise et al .
these effects occur for intrinsic absorbers that are very compact and very dense compared to intervening absorbers . the fraction of quasars that host intrinsic associated nals ( nals within 5,000 of the quasar emission redshifts ) has been estimated to be @xmath527% at @xmath6 , using the time variability technique ( narayanan et al . 2004 ; wise et al .
2004 ) , and @xmath723% at @xmath8 , based on partial coverage analysis ( misawa et al .
these signatures are a sufficient , but not a necessary , condition to demonstrate the intrinsic nature of an absorber .
it is very likely that some nals without time variability or partial coverage are also intrinsic .
both of the above methods have been applied to bal and mini - bal systems .
for example , partial coverage analysis has been carried out by petitjean & srianand ( 1999 ) , srianand & petitjean ( 2000 , ) yuan et al .
( 2002 ) , and ganguly et al .
( 2003 ) , while time - variability has been studied by foltz et al .
( 1987 ) , turnshek et al .
( 1988 ) , smith & penston ( 1988 ) , barlow et al .
( 1992 ) , barlow ( 1993 ) , vilkoviskij & irwin ( 2001 ) , ma ( 2002 ) , narayanan et al .
( 2004 ) , and lundgren et al .
it is quite difficult ( or almost impossible ) to deblend kinematic components in those heavily blending absorption features using low / intermediate resolution spectra .
however , some authors successfully deblended narrower absorption components from each other in such systems using high resolution spectra of @xmath9 ( e.g. , yuan et al .
2002 ; ganguly et al .
systematic monitoring of individual mini - bals using high - resolution spectra taken on several epochs can potentially provide very important constraints on the properties of the absorbers .
however , no such campaign has been attempted so far , to our knowledge .
the optically bright quasar hs 1603 + 3820 ( @xmath10=2.542 , b=15.9 ) , first discovered in the hamburg / cfa bright quasar survey ( hagen et al .
1995 ; dobrzycki et al .
1996 ) , is known to have a large number ( 13 ) of doublets at @xmath11 ( dobrzycki , engels , & hagen 1999 ) . using high - resolution spectra ( @xmath12 ) taken with the subaru telescope , misawa et al .
( 2003 , 2005 ) classified all doublets into 8 systems , and found that only a mini - bal system at @xmath1 ( shift velocity , @xmath1310,600 ) shows both partial coverage and time variability in an interval of 0.36 years in the quasar rest frame .
based on these results , misawa et al .
( 2005 ) placed constraints ( i ) on the electron density ( @xmath14 ) and the absorber s distance from the quasar ( @xmath15 kpc ) , if a change in the ionization state causes the variability , or ( ii ) on the time scale for the absorber to cross the continuum source and the distance from the continuum source , if gas motion across the background uv source causes the variability . in the latter case , the crossing velocity is constrained to be @xmath16 , and the distance from the continuum source , @xmath17 pc .
this is larger than the size of the continuum source , @xmath18 pc , but smaller than that of the blr , @xmath19@xmath73 pc , estimated for this quasar . on the other hand
, the other nals found towards the quasar show no sign of being intrinsic .
further monitoring observations make it possible to ( i ) discriminate between changes in the ionization state and motion of the absorbers across the continuum source as the cause of the time variability seen in the mini - bal , and ( ii ) to confirm whether the other nals really show neither partial coverage nor variability . in this paper , we present the results of monitoring hs 1603 + 3820 over 4.2 years ( 1.2 years in the quasar rest frame ) based on eight spectra .
six of these spectra were taken with the high dispersion spectrograph ( hds ; noguchi et al .
2002 ) on the subaru telescope , and two with the medium - resolution spectrograph ( mrs ; horner , engel , & ramsey , 1998 ) on the hobby - eberly telescope ( het ) .
the first two spectra in this time series were presented and discussed by misawa et al ( 2003 , 2005 ) . with the new spectra we are able to sample the variations of the absorption lines more densely and
probe the internal structure of the absorber .
in 2 , we describe the observations and data reduction .
the methods used for model fitting are outlined in 3 .
the properties of the mini - bal and of the other nal systems are examined in 4 .
the possible origins of the time variability seen in the mini - bal system are discussed in 5 , and a summary is given in 6 .
we adopt @xmath20 as the systemic redshift of the quasar , which was estimated from its narrow emission lines ( misawa et al .
time intervals between observations are given in the quasar rest frame throughout the paper , unless otherwise noted .
we use a cosmology with @xmath21=75 mpc@xmath22 , @xmath23=0.3 , and @xmath24=0.7 .
we observed hs 1603 + 3820 eight times over a period of 4.2 years in the observed frame , from march 2002 to may 2006 .
we will call the dates of these observations epoch 1 through epoch 8 .
six of the observations were obtained with subaru+hds , using a 08 or 10 slit width ( @xmath25 or 36,000 ) , and adopting 2@xmath261 pixel binning along the slit .
the red grating , with a central wavelength of 4900 , was used to cover as many systems as possible , except for the first and the last observations that used central wavelengths of 6450 and 5700 .
the other two spectra were taken with the het+mrs .
the mrs features a pair of fibers : one for the target and another for the sky , which enables us to perform sky subtraction effectively .
we used a 15 fiber to get @xmath27 spectra .
a setup with a central wavelength of 7,000 is required for the queue observations with mrs+het .
we reduced both the subaru and het data in a standard manner with the iraf software .
we used a th - ar spectrum for wavelength calibration .
we directly fitted the continuum , which also includes substantial contributions from broad emission lines 1 year ( e.g , giveon et al .
1999 , hawkins 2001 , kaspi et al .
2007 ) . ] , with a third - order cubic spline function .
around heavily absorbed regions , in which direct continuum fitting is difficult , we used different techniques for the subaru and het spectra . for the subaru spectra
, we adopted the interpolation technique introduced in misawa et al .
since the echelle blaze profile in hds shows time variability during a night , we can not directly use the instrumental blaze function as a continuum profile . in this scheme
, we obtain the continuum shape by interpolating between the two echelle orders adjacent to the order of interest , we use orders @xmath28 and @xmath29 as reference orders .
however , in our data those adjacent orders are also somewhat affected by broad absorption features .
therefore , we used orders @xmath30 and @xmath31 as reference orders . ] , after weighting flux levels based on the flat - frame spectrum [ see eqn .
( 1 ) of misawa et al .
we then divide the quasar spectrum by this interpolated continuum shape to obtain the normalized spectrum .
we have verified the validity of this technique by applying it to a stellar spectrum ; the continuum error is always less than 3.3% . for het spectra , we used a blaze profile constructed from the flat field spectrum as a continuum model after multiplying it by a scaling factor to adjust its count level to that of the quasar spectrum ( the scaling factor is a ratio of counts in the quasar spectrum to those in a flat spectrum for the same _ unabsorbed _ region )
. an observation log is given in table 1 , in which we list the observation epoch , date of observation , relative time in the quasar rest frame , instrument , wavelength coverage , total exposure time , spectral resolving power , and signal - to - noise ratio ( s / n ) per pixel ( after rebinning to pixel scales of 0.03 and 0.15 for subaru and het data , respectively ) .
the s / n is evaluated around 5370 , close to the mini - bal . in figure 1
, we show a normalized spectrum , after combining 6 spectra taken with subaru+hds in the region redward of the ly@xmath32 emission line of the quasar ( @xmath33 ) .
we used the line - fitting software package minfit ( churchill 1997 ; churchill et al . 2003 ) , with which we can fit absorption profiles with four free parameters : redshift ( @xmath34 ) , column density ( ) , doppler parameter ( @xmath35 ) , and coverage fraction ( ) . here , the coverage fraction is the fraction of the background continuum source and the broad - emission line region that is covered by the intrinsic absorber . the coverage fraction can be systematically evaluated in an unbiased manner by considering the optical depth ratio of resonant , rest - frame uv doublets of lithium - like species ( e.g. , , , and ) , namely , @xmath36 where @xmath37 and @xmath38 are the residual fluxes of the blue and red members of the doublets in the continuum normalized spectrum ( hamann et al . 1997b ; barlow & sargent 1997 ; crenshaw et al .
a @xmath39 value less than unity signifies that a portion of the background source is not occulted by the absorber .
this , in turn , means the doublet is probably produced by an intrinsic absorber ( e.g. , wampler et al . 1995 ; barlow & sargent 1997 ) .
we found 9 systems at @xmath402.55 , as identified in dobrzycki et al .
( 1999 ) , but we did not detect a system at @xmath41 .
this system was probably a false detection ( dobrzycki 2005 ; private communication ) . among the 9 systems ,
only the system at @xmath422.45 is classified as a mini - bal system , because its line width is larger than 500 .
we carried out model fits to all systems in the same manner as in our previous paper ( misawa et al . 2005 ) , except for the mini - bal .
misawa et al .
( 2005 ) performed a manual voigt profile fit to the mini - bal , because self - blending ( blending of the blue and red members of the doublet ) occurs in this system . in its original form ,
the minfit code was not able to handle such self - blended regions .
the code was originally written for narrow absorption lines , thus it fitted the models to the blue and red members of doublets _
separately_. this procedure overproduced the residual intensity at the self - blending regions because it added contributions from two components . to avoid this problem , we revised the code so as to fit models to the blue and red members of doublets _ simultaneously _ ,
i.e. , by multiplying the contributions from the two doublet members . in this paper
, we fit the observed mini - bals automatically using this improved minfit . as in misawa et al .
( 2005 ) , we fitted only the first ( bluest ) two kinematic components in the system , at @xmath435318 , because the other components are so heavily blended with each other that we can not separate them , even with the improved minfit ( i.e. , minfit gives multiple solutions ) . at first
, we remove the contamination from @xmath441527 in system b , at @xmath455318 .
the @xmath441527 line in the system did not show any detectable variability between the first two observations .
the doublets for system b show neither partial coverage nor time variability , and are consistent with it being an intervening system .
therefore , the @xmath441527 line is also not likely to vary .
thus , we removed the contamination from @xmath441527 by dividing the observed spectra by a model profile synthesized with the line parameters of @xmath441527 , determined from an average of the first two observations presented in misawa et al .
( 2005 ) . after removing the narrow @xmath441527 components
, we applied the improved minfit to the spectrum using two components , one narrow and one broad , as in misawa et al .
( 2005 ) . during the fitting process , we initially forced the same coverage fraction for the narrow and broad components , as in misawa et al .
( 2005 ) . in that paper we showed that , during the first two epochs of observation , this solution produced an equally good fit as solutions that allowed these coverage fractions to differ .
we also considered the possibility of differing coverage fractions for the narrow and broad components in later epochs of observation .
figure 2 shows 6 spectra taken with subaru+hds , for which we find good model fits .
the fits themselves are shown in figure 3 .
we can not fit spectra obtained with het+mrs in the same manner , because of their lower resolution and s / n .
therefore , on the het+mrs spectra , we overplotted model profiles synthesized with parameters that are linearly interpolated between the two adjacent epochs observed with subaru .
these models provide reasonably good fits to the het+mrs profiles .
the fitting parameters for the narrow and broad components , at different epochs , are summarized in table 2 .
columns ( 1)(3 ) give the epoch number , observation date and the relative time in the quasar rest frame .
columns ( 4)(8 ) list , for the narrow component , the shift velocity , column density , doppler parameter , coverage fraction , and observed - frame equivalent width evaluated from the synthetic model spectrum . columns ( 9)(13 ) are the same as columns ( 4)(8 ) , but for the broad component .
we confirmed that the fit parameters for the first two epochs are consistent within 1@xmath46 errors with those presented in misawa et al .
( 2005 ) , based on a manual fitting procedure .
we also applied minfit to all metal lines in the other nal systems .
since , as described later , we did not see any remarkable variability in these nal systems , we combined all 6 subaru spectra to produce a final spectrum with higher s / n ( @xmath796 per pixel ) before applying minfit to the nals .
the best fit parameters are summarized in table 3 .
column ( 1 ) is an identification number for the line .
columns ( 2 ) and ( 3 ) give the observed wavelength and redshift . column ( 4 ) gives the shift velocity from the quasar systemic redshift .
columns ( 5 ) and ( 6 ) give the column density and the doppler parameter with 1@xmath46 errors .
column ( 7 ) reports the coverage fraction with its 1@xmath46 error ( which includes the error due to continuum level uncertainty as described below ) . during a fitting trial
, minfit sometimes gives unphysical coverage fractions such as @xmath47 or @xmath48 .
when this happens , the other fit parameters such as the column density and doppler parameter have no physical meaning .
these unphysical fit results could be caused by errors in the continuum fit .
misawa et al .
( 2005 ) found that the derived coverage fractions can be significantly affected by continuum level errors , especially for very weak components whose values are close to 1 .
therefore , if minfit gives unphysical values for some components , we carry out the fit again for the entire line , assuming @xmath49 for those components , following misawa et al .
( 2007 ) . in this case
, we do not evaluate the error in .
the best fit models are overlaid with the observed spectra in figure 4(a i ) .
there are also other error sources for the fit parameters , such as blending with other lines , and convolution of the true spectrum with the instrumental line spread function ( lsf ) .
both of these are negligible if ( 1 ) the doublet is strong enough , i.e. , @xmath50 in the case of doublets , ( 2 ) the normalized residual flux is smaller than @xmath70.5 , and ( 3 ) the line profile is much wider than the instrumental lsf ( misawa et al . 2005 ) .
most components in our sample are broad enough compared with the lsf ( corresponding to @xmath51 ) , and not severely blended with each other , as described in section 4.2 .
we previously reported that the mini - bal of system a shows trough dilution ( misawa et al . 2002 ) , and variability within 0.36 years ( misawa et al . 2005 ) . with the additional spectra of this paper
, we follow the variability of this system for 1.2 years in the quasar rest frame , sampling it at eight epochs . and
ly@xmath32 could not be identified because of heavy contamination from the ly@xmath32forest .
was not detected , even in our final , high - s / n spectrum .
figure 5 presents graphically the evolution of the fit parameters of the mini - bal , including ( from top to bottom ) the total rest frame equivalent width at @xmath525346 , the column density , the doppler parameter , the coverage fraction , and the shift velocity .
we discuss each in detail below . the total rest frame equivalent width increased by about a factor of two over a rest frame time interval of 0.36 years , and then decreased over an interval of 0.6 years . during this entire period ,
the general appearance of the line profile did not change significantly . based on the results of just the first two epochs , misawa et al .
( 2005 ) speculated that this mini - bal may soon evolve into a bal , with more than 10% of the flux absorbed over at least a 2,000 range .
this idea is not confirmed by the subsequent epochs , since the mini - bal was found to weaken .
we also monitored the equivalent widths of the narrow and broad components , separately . at first , both components have almost the same equivalent width .
but after epoch 3 , the equivalent width of the broad component was much higher than that of narrow component .
the variability of the column density in the mini - bal is complex .
this is especially true for the broad component , which showed a large increase in column density within 0.13 years , from epoch 3 to epoch 7 , as shown in figure 5 .
this increase in column density was concurrent with the increase of equivalent width , as shown in the top window of the figure .
one possible explanation for this dramatic increase is a change of the ionization state of the gas ( hamann et al .
1997b ; narayanan et al . 2004 ) , which we discuss further in 5 .
an interesting trend is seen in doppler parameter variability . from epoch 3 to epoch 5 ( an interval of 0.09 years in the quasar rest frame ) , the doppler parameter of the broad component increased suddenly from @xmath53 to @xmath54 , and then it decreased gradually .
this abrupt jump is correlated with the sudden increase of the column density in the broad component , as described above . on the other hand ,
the narrow component did not show a significant change in doppler parameter .
the coverage fraction ( assuming it is the same for the narrow and broad components ) shows a variability similar to that of the total equivalent width . after increasing to 0.45
, the coverage fraction decreased to 0.24 .
inspection of the separate contributions of the broad and narrow components to the total equivalent width shows that the coverage fraction evolves similarly to the equivalent width of the narrow component .
if the change of coverage fraction is caused by the absorber s motion across the background flux source , the absorber size is constrained to be smaller than the background source ( about half of the background source because of the maximum value of @xmath70.5 ) . at the same time
, the column density increased , suggesting an inhomogeneous absorber .
a more detailed discussion will be presented in 5 .
the column density and equivalent width of the broad component increased over the same period ( epoch 3 to epoch 8) as its coverage fraction decreased .
however , since we assumed that the broad and narrow components had the same coverage fractions , we must assess whether this assumption could lead to the observed trends .
specifically , we consider the effect of assuming a constant for the broad component ( adopting @xmath55 for epoch 3 through epoch 8 , and allowing the narrow component to change as before ) . this model , which also produced an adequate fit to the data ,
is illustrated with the dotted lines in the 2nd , 3rd , and 4th panels from the top of figure 5 .
for this case , the column density of the broad component does not increase as rapidly , but it still does increase .
the same applies for the equivalent width .
thus our conclusions regarding the evolution of the column density and equivalent width are not qualitatively changed by our assumption about the value of the broad component .
we have looked for correlated changes in all of the above parameters of the mini - bal profile by plotting them against each other . in figures 6 and 7 , respectively , we show the relevant plots for the narrow and broad components of the mini - bal .
other than a tentative trend between @xmath56 and @xmath35 in figure 6 and between @xmath56 and @xmath57 in figure 7 , we see no convincing correlation . this lack of correlated variability between model parameters is an important clue , which we take into consideration in our later discussion .
the observed shift velocity represents the ejection velocity of the absorber from the central engine projected along the line of sight . while the line center of the narrow component remains nearly constant , the broad component shifts by about 300 ( @xmath58 ) from epoch 1 to epoch 7 .
if this velocity shift is really due to the radiative pressure from the continuum source , the acceleration of the gas can be calculated to be @xmath7 0.01 m s@xmath59 . however , the measurement of the center of the broad component is highly uncertain , because the observed velocity shift is much smaller than the total line width because of self - blending of the blue and red members of the doublet .
moreover , the center of the broad component moved in both directions ( i.e. , blueward and redward ) during our observations .
although it is still not clear if there is a truly systematic increase in this parameter , we could regard the above value as an upper limit on the acceleration of the absorber .
in addition to the mini - bal system , we detected 8 nal systems in the hs 1603 + 3820 spectra .
for these nals , we do not detect significant variability .
model fit results to the combined spectrum of the 6 subaru observations ( with a total exposure time of 8.9 hours and a s / n per 0.03 pixel of about 91 at @xmath60 ) are consistent with the results based on only the epoch 2 spectrum ( a 1.7 hour exposure with @xmath61 per pixel ; misawa et al .
2005 ) , with only a few small exceptions as described below . in this section
we discuss individual systems , other than system a ( the mini - bal ) .
the profiles of the lines of these systems are shown on a common velocity scale in figure 4(a i ) .
system b : : we fit this strong nal , shown in figure 4a , b , with 15 components , although misawa et al .
( 2005 ) used 19 components .
the coverage fractions of two components , 11 and 14 , deviate from unity by more than 1@xmath46 , but both are heavily blended with other components and their values are not very reliable .
all components in the nal are consistent with full coverage . only the epoch 8 spectrum covered the doublet of this system . for the doublet
, only component 7 suggests partial coverage , but the model fit is not good around the left wing of this component .
we conclude that this system does not show strong evidence for partial coverage , and is probably an intervening system , most likely a galaxy because of the strength of the low - ionization absorption lines .
system c : : the nal in this system with a simple absorption profile ( see figure 4c ) is fitted with only one component , as in misawa et al .
both fits are consistent with full coverage .
we also did not find any evidence for time variability . except for its small shift velocity , @xmath62 ,
there is no evidence to suggest that this system is intrinsic .
system d : : misawa et al .
( 2005 ) fitted this system s and profiles with 4 and 2 components , respectively , and did not find any partial coverage .
however , we found an additional component at @xmath63 from the system center that shows partial coverage , and we have no reason to suspect systematic error .
the line profiles are shown in figure 4d .
this component was not studied by misawa et al .
( 2005 ) because of a data defect .
we also now find that component 1 in the nal possibly shows partial coverage , although it was consistent with full coverage in the shorter exposure presented by misawa et al .
we also note that it is surprising that the @xmath441548 is black at the position of this component if the partial coverage is true .
however , the fact that partial coverage is apparent for both a and a component suggests that this system may be an intrinsic system .
the fact that this system is redshifted from the quasar systemic redshift also supports this idea .
although this system is located near the top of the emission line , there is no residual flux at the position of the doublet for components 4 and 5 .
these components must therefore cover both the continuum source and broad emission line region .
we do not see any variability over the course of our observations .
system e : : misawa et al .
( 2005 ) fitted the nal with only two components , and one of them showed evidence of partial coverage .
however , our higher - s / n spectrum ( shown in figure 4e ) requires 5 components to fit the nal , and three of them are consistent with full coverage .
although components 3 and 5 deviate from full coverage , they are heavily blended with the other components , so this result is quite uncertain .
the shift velocity of the system is also extremely large , @xmath760,000 .
there is no compelling reason to believe that this system is intrinsic to the quasar .
system f : : our best - fitting model is almost consistent with the previous one , having the same number of components at almost the same positions ( see figure 4f ) .
all components are consistent with full coverage except for component 1 that is blended with the strong component 2 .
this system also has a very large shift velocity , @xmath64 .
we consider it probable that this is an intervening system .
system g : : because the nal in this system was positioned at the edge of the blue ccd of the hds ( around which bad pixels are common ) in some of our observing setups , the absorption profiles are slightly or severely different in some spectra .
therefore , we combined only spectra from epochs 3 , 7 , and 8 ( these are not affected by the data defect ) to produce a final spectrum for this system ( figure 4 g ) .
we fit the nal with 6 components , and all of them are consistent with full coverage .
component 5 was not fitted by misawa et al .
( 2005 ) because of the unphysical line ratio of the blue and red component , due to line blending .
this system is probably an intervening system .
system h : : two components are necessary to fit this nal ( see figure 4h ) .
the shallower absorption feature , centered on component 2 , has an unphysical ratio of its blue and red members .
probably the blue member is contaminated by other lines . only component 1
is useful , and it yields @xmath65 . in this system , we again do not see any evidence of intrinsic properties .
system i : : this system was detected only in the epoch 8 spectrum , because it was not covered in the other epochs .
we used 7 components to fit the nal ( see figure 4i ) .
component 4 may show partial coverage , but its value differs from unity by only slightly more than 1@xmath46 .
the case for partial coverage for component 6 is somewhat more compelling , since the minfit fitting procedure yielded a value @xmath66 .
component 4 is not blended with other components , and component 6 is only weakly blended . although the s / n for this spectrum is not very high in the relevant region , this system may be another candidate for an intrinsic system .
in view of the observational constraints derived above , we evaluate here different scenarios for the origin of the variability of the mini - bal . a fundamental assumption underlying
our discussion is that the kinematic components making up system a represent parcels of absorbing gas at different positions along the line of sight . in effect
, we assume that the absorber is embedded in an accelerating outflow from the quasar , thus the blueshift of a kinematic component increases with its distance from the quasar continuum source .
a very important observational clue from the data presented above is the fact that all the kinematic components of the mini - bal system vary in concert ( i.e. , the depths of all troughs increase or decrease together ) .
these changes appear to be driven largely by changes in the coverage fraction , as shown in figure 5 .
other physical parameters ( @xmath67 and @xmath35 ) are varying as well , but there does not seem to be an obvious correlation between changes of different profile parameters , at least in the bluest troughs of the mini - bal ( see figures 6 and 7 ) . in misawa et al .
( 2005 ) we favored an interpretation in which motion of the ( homogeneous ) absorber across the line ( or cylinder ) of sight causes the observed variability ( see also hamann et al .
the assumed scenario is depicted in the cartoon of figure 2a of hamann & sabra ( 2004 ) . with time , an increasing fraction of the cylinder of sight was covered by the absorber , leading to changes in the coverage fraction . in this context and based on two snapshots of the spectrum , we derived a transverse velocity of 8,000 and a constraint of @xmath68 pc on the distance of the absorber from the continuum source ( see details in misawa et al .
2005 ) .
that model was motivated largely by the observation that the coverage fraction was the only quantity that changed substantially between the first two observations .
the subsequent observations show that scenario to be too simple , namely the column density of the absorber changes with time as well , suggesting that the absorber is inhomogeneous and its internal structure plays an important role in the variability of the mini - bal profile .
thus , a more plausible scenario is the one depicted in figure 2b of hamann & sabra ( 2004 ) and in figure 5 of de kool , korista , & arav ( 2002 ) , where the absorbing medium is inhomogeneous , with clumps spanning a range of column densities scattered over the cylinder of sight .
there is a remaining difficulty with this scenario , however .
the fact that all the troughs in the mini - bal vary in concert would require a coincidence between the motions of the individual clumps . more specifically , the projected areas of all the clumps within the cylinder of sight should vary in the same way ( i.e. , rise and fall with time in the same manner for all clumps ) .
we consider such a coordinated sequence of variations rather unlikely , therefore , we regard this interpretation as implausible .
we note that if future observations show this pattern to be common , then this interpretation can be conclusively rejected . at first glance ,
compression of the absorber might be able to produce a trend similar to that of figure 7b , where the coverage fraction decreases as the column density increases in one of the kinematic components of the mini - bal .
however , on closer examination , such a compression would require highly supersonic motions within the absorber .
assuming the background continuum source size is 0.02 pc ( this is the uv - emitting region of the accretion disk ; see misawa et al .
2005 ) , a change in the coverage fraction from 0.36 to 0.27 within 0.13 years ( epochs 5 and 7 ) implies a speed of order @xmath69 . in comparison ,
the speed of sound in a @xmath70 k gas is only @xmath71 .
thus we consider such a scenario untenable .
the observed partial coverage signature can , in principle , be explained if continuum photons are re - directed by a scattering medium towards the observer , when they were initially traveling in a different direction .
thus , the absorber can be homogeneous and its projected size need not be smaller than the cylinder of sight towards the continuum source .
the dilution of the troughs by photons scattered around the absorber can change with time as the absorber moves relative to the scattering medium or as the conditions in the scattering medium change . in the context of this scenario ,
the _ apparent _ coverage fraction can vary according to the behavior of the scattering medium .
moreover , the variations of the coverage fraction need not be tied in any way to variations of other parameters of the absorber , such as the column density , the @xmath35-parameter , or the shift velocity .
this idea can be tested observationally through spectropolarimetry .
in particular , we would expect the fractional polarization in the absorption mini - bal troughs to be higher than in the continuum ( see the analogous test in bal quasars by brotherton et al .
1997 , ogle et al . 1999 , and lamy & hutsemkers 2000 , for example ) .
furthermore , we would also expect this fractional polarization to increase as the coverage fraction decreases .
changes in the level of the ionizing continuum can also bring about absorption - line variability since they change the relative ionic populations in the absorber ( in this particular case , the relative populations of , and ) . in principle , a monotonic change in the continuum intensity can produce a non - monotonic change in the absorption line strength .
for example , an absorber initially at a low ionization state that sees a rising ionizing flux will first respond with an increase in the strength of its absorption lines ( as @xmath72 ) and then with a decrease ( as @xmath72 ) .
the variability time scale of the absorption lines provides an upper limit on the ionization or recombination time of the absorbing gas , hence its electron density ( see the detailed discussion in hamann et al .
1997b and narayanan et al .
the density limit can be obtained from @xmath73 , where @xmath74 is a recombination coefficient and @xmath75 is a variability time scale .
this is a simplified picture in which the gas is in ionization equilibrium and is the dominant ionic species of c. since the ionic column densities have fluctuated up and down _ more than once _ over the course of our monitoring observations , the ionizing continuum seen by the absorber must be fluctuating too .
moreover , the continuum intensity must be fluctuating by at least a factor of 3 on rest - frame time scales of 6 months or less to produce the observed changes in the ionic column density ( see the calculations of hamann 1997 ) . taking this interpretation at face value ( but see discussion below ) , we can use the shortest variability time scale that we have observed ( 16 days in the quasar rest frame , epochs 5@xmath727 ; see fig .
5 ) to constrain the density of the gas . we consider both an increase and a decrease in the ionization state of the gas as a cause for the observed change in the column density ( i.e. , @xmath72 or @xmath72 ) , with respective recombination coefficients of @xmath76 and @xmath77 ( from arnaud & rothenflugh 1986 ; assuming a nominal gas temperature of 20,000 k , following hamann 1997 ) .
thus we obtain a limit on the electron density of @xmath78 , using the methodology of narayanan et al .
( 2004 ) and a bolometric luminosity of @xmath79 ( misawa et al . 2005 ) . to convert this limit to a limit on the distance of the absorber from the continuum source , @xmath80
, we must assume a value of the ionization parameter .
if the column density fluctuations reflect conversions of @xmath81 in a lower - ionization absorber , then the calculations of hamann ( 1997 ) indicate an ionization parameter of @xmath82 and lead to @xmath83 kpc .
if , on the other hand , the column density fluctuations reflect conversions of @xmath81 in a higher - ionization absorber , then the calculations of hamann ( 1997 ) indicate an ionization parameter of @xmath84 and lead to @xmath85 kpc . in this context
the simultaneous variability of all the mini - bal troughs implies that these troughs should represent gas parcels of similar densities for the following reason .
since all absorbing gas parcels must be within the cylinder of sight to the continuum source , the time lag between changes in the continuum intensity at the source and our observation of a change in the mini - bal profile is the sum of the light travel times from the continuum source to the absorber and from the absorber to the observer , plus the recombination time of the gas .
the sum of light travel times is the same regardless of the location of a gas parcel _
along _ the cylinder of sight , therefore , a difference in time lag can only result from different recombination times , i.e. , different gas densities .
we can then conclude that the density does not vary significantly across the line profile , which implies in turn that the shape of the mini - bal profile is determined primarily by the combination of column density ( i.e. , physical thickness ) and coverage fraction as a function of velocity .
fluctuations in the ionization state of the absorber are an attractive explanation of the observed variations in the strength of the mini - bal in hs 1603 + 3820 .
however , the mini - bal appears to vary much faster than any expected variations in the uv continuum of such a luminous quasar ( see for example , giveon et al . 1999 , hawkins 2001 , and kaspi et al .
moreover , fleming & kennefick ( 2006 ) found no variability of this particular quasar in 10 days in the observed frame .
therefore we propose that the variations of the ionizing continuum _ seen by the absorber _
are caused by a screen of variable optical depth between it and the ( relatively steady ) continuum source .
this screen could be the ionized , clumpy , inner part of the outflow , i.e. , the shielding or hitchhiking gas invoked in the outflow models of murray et al .
( 1995 ) , and arising naturally in the more detailed calculations of proga et al .
the two - dimensional , axisymmetric models of proga et al .
( 2000 ) also show that disk winds can be accelerated up to 15,000 , and that they are unsteady and generate dense knots periodically . the time variability seen in the mini - bal system of hs 1603 + 3820
may be related to this mechanism .
the conditions in the screen may be similar to those of `` warm '' absorbers , observed in the x - ray spectra of many seyfert galaxies , where it often co - exists with a uv absorber ( e.g. , crenshaw et al .
we note , by way of example , that models for the warm absorbers in ngc 5548 , iras 1339 + 2438 , and ngc 3783 require multiple ionization phases for an adequate fit to the x - ray spectrum ( kaastra et al . 2000 , sako et al .
2001 , blustin et al .
2002 ) , one of which may fulfill the requirements for the inner screen
. if the optical depth of the screen at wavelengths around the edge is close to unity , then small fluctuations in the column density can cause large changes in the shape and intensity of the transmitted ionizing continuum ( e.g. , fig .
6 of murray et al .
1995 , or fig . 1 of hamann 1995 ) ,
leading to the observed changes in the mini - bal of hs 1603 + 3820 .
this may be analogous to the best - studied warm absorber , in the seyfert galaxy ngc 3783 , where fluctuations in the ionization structure of the absorbing gas have been observed over the course of 6 months ( e.g. , netzer et al .
there is also evidence that the opacity of the warm absorber in ngc 3783 fluctuates over time intervals as short as a month , in response to variations in the x - ray continuum by a factor of 2 ( krongold et al . 2005 ) . within this context
we can also understand why the variations in the coverage fraction of the mini - bal gas are unrelated to variations of its column density .
because both media can be clumpy and since differential rotation of the wind causes the inner screen to move faster than the mini - bal gas across the cylinder of sight , the intensity of the continuum transmitted through the screen and received by the mini - bal gas would vary with time .
moreover , the coverage fraction need not be correlated with the intensity of the transmitted continuum .
we have considered the possibility that one of the observed nal systems at smaller blueshifts relative to the quasar could represent the inner screen . however , none of the three candidate systems , b , c , and d , can satisfy the requirements for the screen .
the very fact that they have absorption lines suggests that their ionization state is too low , and moreover , systems b and d have even lower - ionization lines .
in addition , none of the three systems are variable , which is necessary in order for the screen to perform the desired function .
more promising ways of finding the screen signature are x - ray observations in search of the classic warm absorber feature , and edges .
thus , we will be able to determine which of the ly@xmath32 absorption lines seen at shorter wavelengths might be associated with the screen material .
we have been monitoring the absorption lines in the spectrum of the quasar hs 1603 + 3820 since 2002 , using the subaru and hobby - eberly telescopes .
we have obtained six subaru+hds and two het+mrs spectra spanning an interval of approximately 1.2 years in the quasar rest frame and probing rest - frame time scales as short as 16 days .
we have determined the physical parameters of the 9 absorption - line systems by fitting voigt models to the line profiles .
our main results are as follows : 1 .
we detect one mini - bal and eight nal systems in the quasar spectrum . the partial coverage signature and time variability of the mini - bal show it to be intrinsic to the quasar .
two of the nal systems ( d and i ) may also be intrinsic , based upon our partial coverage analysis .
we are able to fit models only to the bluest portion of the mini - bal profile where self - blending is not severe .
all fit parameters ( i.e. , column density , doppler parameter , coverage fraction , and shift velocity ) as well as the total equivalent width of the system vary significantly with time , even on short time scales
. however , the profile parameters do not appear to change in concert with each other , with one exception : the equivalent widths of all the troughs in this system vary together and approximately follow the variations of the coverage fraction determined from the model fits .
we have examined a number of ways of explaining the above variations of the mini - bal and we have found two viable possibilities .
the first possibility is that the observed partial coverage signature is the result of continuum photons scattering around the absorber and into our cylinder of sight .
the observed changes in mini - bal equivalent widths are thus produced by variations in the scattered continuum that dilutes the absorption troughs . in the second possibility
, the illumination of the uv absorber fluctuates on short time scales by a factor of up to 3 .
we suggest that these fluctuations are caused by a screen of variable optical depth between the mini - bal gas and the continuum source .
this screen might be identified with the shielding gas invoked or predicted in some outflow models .
moreover , it could be analogous to the `` warm '' absorbers observed in the x - ray spectra of seyfert galaxies and some quasars .
this picture can also explain the variations in the coverage fraction , which appear to be unrelated to the ionic column density changes .
this work was supported by nasa grant nag5 - 10817 .
we are grateful to the staff of the subaru telescope , which is operated by the national astronomical observatory of japan .
we would also like to thank christopher churchill for providing us with the minfit software package .
me acknowledges partial support from the theoretical astrophysics visitors fund at northwestern university and thanks the members of the theoretical astrophysics group for their warm hospitality .
the hobby - eberly telescope ( het ) is a joint project of the university of texas at austin , the pennsylvania state university , stanford university , ludwig - maximillians - universitt mnchen , and georg - august - universitt gttingen .
the het is named in honor of its principal benefactors , william p. hobby and robert e. eberly .
we also thank the anonymous referee for helpful comments and suggestions .
arav , n. , korista , k.t . ,
barlow , t.a . , & begelman , m.c .
, 1995 , , 376 , 576 arnaud , m. , & rothenflugh , r. , 1985 , , 60 , 425 bachev , r. , strigachev , a. , & semkov , e. , 2005 , , 358 , 774 barlow , t.a . , & sargent , w.l.w . , 1997 , , 113 , 136 barlow , t.a . , 1993 , ph.d . thesis , university of california , san diego , ( 1994 , pasp , 106 , 548 ) barlow , t.a . ,
junkkarinen , v.t .
, burbidge , e.m . ,
weymann , r.j . ,
morris , s.l . , & korista , k.t . , 1992 , , 397 , 81 becker , r.h . ,
gregg , m.d . ,
hook , i.m . ,
mcmahon , r.g . ,
white , r.l .
, & helfand , d.j . , 1997 , , 479 , l93 blustin , a. j. , branduardi - raymont , g. , behar , e. , kaastra , j. s. , kahn , s. m. , page , m. j. sako , m. , & steenbrugge , k. c. 2002 , , 392 , 453 brotherton , m. s. , tran , h. d. , van breugel , w. , dey , a. , & antonucci , r. 1997 , , 487 , l113 chiang , j. & murray , n. , 1996 , , 466 , 704 churchill , c.w . , vogt , s.s . , & charlton , j.c . , 2003 , , 125 , 98 churchill , c.w . , schneider , d.p . , schmidt , m. , & gunn , j.e . , 1999 , , 117 , 2573 churchill , c.w . , 1997 , ph.d .
thesis , univ .
california , santa cruz crenshaw , d.m . , kraemer , s.b .
, boggess , a. , maran , s.p . ,
mushotzky , r.f . , &
wu , c .- c . , 1999 , , 516 , 750 dobrzycki , a. , engels , d. , & hagen , h .- j .
, 1999 , , 349 , l29 ( d99 ) dobrzycki , a. , engels , d. , hagen , h .- j . , elvis , m. , huchra , j. , & reimers , d. , 1996 , , 188.0602 elvis , m. , 2000 , , 545 , 63 everett , j.e . , , 2005 , 631 , 689 fleming , b.t .
, & kennefick , j. , 2006 , , 141 , 01 foltz , c.b . , weymann , r.j . ,
morris , s.l . , & turnshek , d.a . , 1987 , , 317 , 450 ganguly , r. , eracleous , m. , charlton , j.c . , & churchill , c.w . , 1999 , ,
117 , 2594 ( g99 ) ganguly , r. , bond , n.a . ,
charlton , j.c . ,
eracleous , m. , brandt , w.n .
, & churchill , c.w . , 2001 , , 549 , 133 ( g01 ) ganguly , r. , masiero , j. , charlton , j.c . ,
& sembach , k.r . , 2003 , , 598 , 922 giveon , u. , maoz , d. , kaspi , s. , netzer , h. , & smith , p. s. 1999 , , 306 , 673 granato , g.l . , de zotti , g. silva , l. , bressan , a. , & danese , l. , 2004 , , 600 , 580 hagen , h .- j . , groote , d. , engels , d. , & reimers , d. , 1995 , , 111 , 195 hamann , f. 1997 , apjs , 109 , 279 hamann , f. , barlow , t.a . , & junkkarinen , v. , 1997a , , 478 , 87 hamann , f. , barlow , t.a . ,
junkkarinen , v. , & burbidge , e.m .
, 1997b , , 478 , 80 hamann , f. , & sabra , b. , 2004 , 2004 , aspc , 311 , 203 hawkins , m.r.s . , 2001 , , 553 , 97 horner , s.d . ,
engel , l. & ramsey , l.w . , 1998 ,
spie conf .
3355 , 375 kaspi , s. , brandt , w.n .
, maoz , d. , netzer , h. , schneider , d.p . , and shemmer , o. , 2007 , , in press ( astro - ph/0612722 ) kaastra , j. s. , mewe , r. , liedahl , d. a. , komossa , s. , & brinkman , a. c. 2000 , , 354 , l83 krongold , y. , nicastro , f. , brickhouse , n. , elvis , m. , & mathur , s. 2005 , , 622 , 842 de kool , m. korista , k.t . , & arav , n. 2002 , , 580 , 54 lamy , h. & hutsemkers , d. 2000 , , 356 , l9 lundgren , b.f . , wilhite , b.c , brunner , r.j , hall , p.b . , schneider , d.p . ,
york , d.g .
, vanden berk , d.e . , & brinkmann , j. , 2006 , astro - ph/0610656 ma , f. , 2002 , , 335 , 99l misawa , t. , charlton , j.c . , eracleous , m. , ganguly , r. , tytler , d. , kirkman , d. , suzuki , n. , & lubin ,
d. , 2007 , , submitted misawa , t. , eracleous , m. , charlton , j.c . , & tajitsu , a. , 2005 , , 629 , 115 misawa , t. , yamada , t. , takada - hidai , m. , wang , y. , kashikawa , n. , iye , m. , & tanaka , i. , 2003 , , 125 , 1336 misawa , t. , tytler , d. , iye , m. , storrie - lombardi , l.j . ,
suzuki , n. , & wolfe , a.m. , 2002 , , 123 , 1863 murray , n. , chiang , j. , grossman , s.a . , & voit , g.m . , 1995 , , 451 , 498 narayanan , d. , hamann , f. , barlow , t. , burbidge , e.m . , cohen , r.d . ,
junkkaribeb , v. , & lyons , r. , 2004 , , 601 , 715 netzer , h. et al . , 2003 , , 599 , 933 noguchi , k. , aoki , w. , kawanomoto , s. , ando , h. , honda , s. , izumiura , h. , kambe , e. , okita , k. , sadakane , k. , sato , b. , tajitsu , a. , takada - hidai , m. , tanaka , w. , watanabe , e. , and yoshida , m. , 2002 , , 54 , 6 ogle , p. m. , cohen , m. h. , miller , j. s. , tran , h. d. , goodrich , r. w. , martel , a. r. 1999 , , 125 , 1 petitjean , p. & srianand , r. , 1999 , , 345 , 73 proga , d. , stone , j.m .
, & kallman , t.r . , 2000 , , 543 , 686 richards , g.t . , 2001 , , 133 , 53 sako , m. et al . 2001 , , 365 , l168 sargent , w.l.w . , boksenberg , a. , & steidel , c.c . , 1988 , , 68 , 539 scannapieco , e. & oh , s. p. 2004
, , 608 , 62 smith , l.j . & penston , m.v . , 1988 , , 235 , 551 springel , v. , di matteo , t. , & hernquist , l. 2005 , , 620 , l79 srianand , r. & petitjean , p. , 2000
, , 357 , 414 steidel , c.c .
, 1990 , , 72 , 1 turnshek , d.a . ,
grillmair , c.j . ,
foltz , c.b . , & weymann , r.j . , 1988 , , 325 , 651 vestergaard , m. , 2002 , , 571 , 733 vilkoviskij , e.y .
& irwin , m.j . , 2001 , , 321 , 4 weymann , r.j . ,
morris , s.l . ,
foltz , c.b . , & hewett , p.c . , 1991 , , 373 , 23 wise , j.h . ,
eracleous , m. , charlton , j.c . ,
& ganguly , r. , 2004 , , 613 , 20 yuan , q. , green , r.f . , brotherton , m. , tripp , t.m . ,
kaiser , m.e . , & kriss , g.a .
, 2002 , , 575 , 687 clcllcrc 1 & 2002 mar 23 & 0.00 & subaru+hds & 50806420 & 2700 & 45,000 & 46.5 + 2 & 2003 jul 7 & 0.36 & subaru+hds & 35204850 , 49306260 & 6000 & 45,000 & 48.7 + 3 & 2005 feb 26 & 0.83 & subaru+hds & 35204855 , 49256275 & 7100 & 36,000 & 43.0 + 4 & 2005 may 10 & 0.89 & het+mrs & 45006200 & 1800 & 9,200 & 17.4 + 5 & 2005 jun 29 & 0.92 & subaru+hds & 35304855 , 49256280 & 3600 & 45,000 & 31.1 + 6 & 2005 aug 3,8 & 0.95 & het+mrs & 45006170 & 4000 & 9,200 & 28.5 + 7 & 2005 aug 19 & 0.96 & subaru+hds & 35204850 , 49206235 & 3600 & 36,000 & 43.6 + 8 & 2006 may 31jun 1 & 1.18 & subaru+hds & 43205630 , 57407050 & 9000 & 45,000 & 52.5 + clccccccccccccc 1 & 2002 mar 23 & 0.00 & 10648 & [email protected] & 85@xmath863 & [email protected] & 1.49 & & 10118 & [email protected] & 233@xmath86 6 & [email protected] & 1.88 + 2 & 2003 jul 7 & 0.36 & 10637 & [email protected] & 98@xmath865 & [email protected] & 2.79 & & 10273 & [email protected] & 274@xmath8615 & [email protected] & 3.32 + 3 & 2005 feb 26 & 0.83 & 10644 & [email protected] & 84@xmath862 & [email protected] & 2.15 & & 10309 & [email protected] & 348@xmath8615 & [email protected] & 2.22 + 4 & 2005 may 10 @xmath87 & 0.89 & 10643 & 14.70 & 81 & 0.33 & 1.83 & & 10259 & 14.67 & 692 & 0.33 & 2.86 + 5 & 2005 jun 29 & 0.92 & 10642 & [email protected] & 80@xmath863 & [email protected] & 1.72 & & 10235 & [email protected] & 863@xmath8699 & [email protected] & 3.20 + 6 & 2005 aug 38 @xmath88 & 0.95 & 10641 & 14.72 & 71 & 0.28 & 1.49 & & 10345 & 14.82 & 793 & 0.28 & 3.29 + 7 & 2005 aug 19 & 0.96 & 10641 & [email protected] & 69@xmath862 & [email protected] & 1.43 & & 10381 & [email protected] & 770@xmath8631 & [email protected] & 3.36 + 8 & 2006 may 31jun 1 & 1.18 & 10637 & [email protected] & 71@xmath862 & [email protected] & 1.32 & & 10291 & [email protected] & 587@xmath8618 & [email protected] & 3.27 + lcccccc + + 1 ............ & 5381.8 & 2.47618 & 5627 & 13.36 @xmath86 0.01 & 8.2 @xmath86 0.2 & 1.00 + 2 ............ & 5382.2 & 2.47642 & 5606 & 14.16 @xmath86 0.03 & 6.3 @xmath86 0.2 & 1.00 + 3 ............ & 5382.8 & 2.47685 & 5569 & 13.52 @xmath86 0.01 & 22.4 @xmath86 0.4 & 1.00 + 4 ............ & 5383.8 & 2.47749 & 5514 & 13.83 @xmath86 0.00 & 15.6 @xmath86 0.1 & 1.00 + 5 ............ & 5383.9 & 2.47752 & 5511 & 14.69 @xmath86 0.46 & 5.5 @xmath86 1.1 & 1.00@xmath89 + 6 ............ & 5384.3 & 2.47779 & 5488 & 12.87 @xmath86 0.29 & 4.1 @xmath86 2.1 & 0.90@xmath90 + 7 ............ & 5384.5 & 2.47793 & 5476 & 13.58 @xmath86 0.01 & 6.7 @xmath86 0.2 & 1.00 + 8 ............ & 5385.0 & 2.47826 & 5447 & 14.06 @xmath86 0.00 & 14.8 @xmath86 0.2 & 1.00 + 9 ............ & 5385.6 & 2.47862 & 5416 & 13.96 @xmath86 0.10 & 14.7 @xmath86 2.2 & 0.97@xmath89 + 10 ........... & 5386.2 & 2.47899 & 5384 & 14.13 @xmath86 0.00 & 27.1 @xmath86 0.2 & 1.00 + 11 ........... & 5386.2 & 2.47899 & 5384 & 13.62 @xmath86 0.21 & 5.6 @xmath86 1.2 & 0.76@xmath91 + 12 ........... & 5387.1 & 2.47959 & 5333 & 12.96 @xmath86 0.21 & 7.3 @xmath86 1.0 & 1.00@xmath92 + 13 ........... & 5387.4 & 2.47977 & 5317 & 13.74 @xmath86 0.01 & 7.8 @xmath86 0.1 & 1.00 + 14 ........... & 5388.2 & 2.48028 & 5273 & 13.56 @xmath86 0.18 & 28.1 @xmath86 2.2 & 0.45@xmath93 + 15 ........... & 5388.6 & 2.48055 & 5250 & 14.03 @xmath86 0.03 & 4.7 @xmath86 0.1 & 1.00 + + 1 ............ & 4308.0 & 2.47748 & 5514 & 13.26 @xmath86 0.01 & 7.7 @xmath86 0.3 & 1.00 + 2 ............ & 4308.4 & 2.47784 & 5483 & 12.73 @xmath86 0.06 & 9.4 @xmath86 1.5 & 1.00 + 3 ............ & 4308.9 & 2.47824 & 5449 & 13.13 @xmath86 0.06 & 19.7 @xmath86 2.7 & 1.00 + 4 ............ & 4309.6 & 2.47882 & 5399 & 13.22 @xmath86 0.05 & 28.2 @xmath86 4.0 & 1.00 + + 1 ............ & 4845.2 & 2.47635 & 5612 & 12.44 @xmath86 0.03 & 15.5 @xmath86 0.8 & 1.00 + 2 ............ & 4845.3 & 2.47642 & 5606 & 13.20 @xmath86 0.01 & 4.0 @xmath86 0.1 & 1.00 + 3 ............ & 4846.8 & 2.47751 & 5512 & 12.49 @xmath86 0.01 & 7.5 @xmath86 0.2 & 1.00 + 4 ............ & 4847.4 & 2.47793 & 5476 & 12.72 @xmath86 0.01 & 2.9 @xmath86 0.1 & 1.00 + 5 ............ & 4847.8 & 2.47824 & 5449 & 13.05 @xmath86 0.10 & 21.3 @xmath86 0.9 & 0.94@xmath94 + 6 ............ & 4848.3 & 2.47860 & 5418 & 13.12 @xmath86 0.00 & 9.3 @xmath86 0.1 & 1.00 + 7 ............ & 4848.9 & 2.47899 & 5384 & 13.05 @xmath86 0.07 & 11.9 @xmath86 0.3 & 0.69@xmath95 + 8 ............ & 4849.7 & 2.47960 & 5332 & 12.98 @xmath86 0.01 & 27.7 @xmath86 0.4 & 1.00 + 9 ............ & 4849.9 & 2.47976 & 5318 & 12.86 @xmath86 0.01 & 4.6 @xmath86 0.1 & 1.00 + 10 ........... & 4851.0 & 2.48054 & 5251 & 13.39 @xmath86 0.01 & 4.9 @xmath86 0.1 & 1.00 + + 1 ............ & 4194.8 & 2.47687 & 5567 & 15.14 @xmath86 0.30 & 20.1 @xmath86 1.4 & 1.00 + 2 ............ & 4195.8 & 2.47770 & 5496 & 13.97 @xmath86 0.01 & 84.7 @xmath86 1.6 & 1.00 + 3 ............ & 4196.9 & 2.47857 & 5420 & 14.31 @xmath86 2.23 & 6.9 @xmath86 4.5 & 1.00 + 4 ............ & 4197.4 & 2.47903 & 5381 & 12.57 @xmath86 0.04 & 8.4 @xmath86 0.8 & 1.00 + 5 ............ & 4198.0 & 2.47949 & 5341 & 13.71 @xmath86 1.57 & 4.6 @xmath86 2.7 & 1.00 + 6 ............ & 4198.3 & 2.47976 & 5318 & 14.43 @xmath86 0.57 & 3.1 @xmath86 0.7 & 1.00 + 7 ............ & 4198.3 & 2.47971 & 5322 & 13.06 @xmath86 0.03 & 32.7 @xmath86 1.4 & 1.00 + 8 ............ & 4199.2 & 2.48052 & 5252 & 13.99 @xmath86 0.38 & 5.6 @xmath86 0.7 & 1.00 + 9 ............ & 4200.1 & 2.48122 & 5192 & 13.10 @xmath86 2.41 & 2.2 @xmath86 3.2 & 1.00 + 10 ........... & 4200.3 & 2.48136 & 5180 & 12.51 @xmath86 0.04 & 4.8 @xmath86 0.7 & 1.00 + + 1 ............ & 4639.4 & 2.47642 & 5606 & 13.52 @xmath86 0.07 & 2.6 @xmath86 0.3 & 1.00 + 2 ............ & 4641.7 & 2.47811 & 5460 & 14.06 @xmath86 0.01 & 11.9 @xmath86 0.2 & 1.00 + 3 ............ & 4642.2 & 2.47852 & 5425 & 14.41 @xmath86 0.03 & 13.9 @xmath86 0.4 & 1.00 + 4 ............ & 4642.5 & 2.47871 & 5408 & 14.15 @xmath86 2.21 & 2.6 @xmath86 3.4 & 1.00 + 5 ............ & 4642.9 & 2.47904 & 5380 & 12.98 @xmath86 0.02 & 9.9 @xmath86 0.9 & 1.00 + 6 ............ & 4643.4 & 2.47944 & 5345 & 13.08 @xmath86 0.02 & 7.0 @xmath86 0.6 & 1.00 + 7 ............ & 4643.8 & 2.47970 & 5323 & 13.97 @xmath86 0.08 & 4.3 @xmath86 0.9 & 1.00 + 8 ............ & 4643.9 & 2.47981 & 5314 & 12.72 @xmath86 1.02 & 10.6 @xmath8613.0 & 1.00 + 9 ............ & 4644.7 & 2.48037 & 5265 & 12.26 @xmath86 0.70 & 1.1 @xmath86 5.1 & 1.00 + 10 ........... & 4644.9 & 2.48052 & 5252 & 14.92 @xmath86 0.80 & 3.6 @xmath86 0.9 & 1.00 + 11 ........... & 4645.8 & 2.48120 & 5194 & 15.33 @xmath86 1.70 & 2.7 @xmath86 1.2 & 1.00 + 12 ........... & 4646.0 & 2.48134 & 5182 & 13.89 @xmath86 0.08 & 3.7 @xmath86 0.6 & 1.00 + 13 ........... & 4646.3 & 2.48162 & 5158 & 12.72 @xmath86 0.03 & 5.0 @xmath86 0.8 & 1.00 + + 1 ............ & 5811.2 & 2.47814 & 5458 & 11.95 @xmath86 0.01 & 10.6 @xmath86 0.4 & 1.00 + 2 ............ & 5811.9 & 2.47856 & 5421 & 12.92 @xmath86 0.02 & 8.8 @xmath86 0.1 & 1.00 + 3 ............ & 5813.4 & 2.47944 & 5345 & 11.28 @xmath86 0.03 & 4.2 @xmath86 0.9 & 1.00 + 4 ............ & 5813.9 & 2.47972 & 5321 & 11.70 @xmath86 0.01 & 4.0 @xmath86 0.4 & 1.00 + 5 ............ & 5815.2 & 2.48053 & 5252 & 12.57 @xmath86 3.50 & 1.5 @xmath86 3.2 & 1.00 + 6 ............ & 5816.4 & 2.48121 & 5193 & 11.84 @xmath86 0.18 & 1.7 @xmath86 1.1 & 1.00 + 7 ............ & 5816.6 & 2.48135 & 5181 & 11.50 @xmath86 0.04 & 3.5 @xmath86 1.2 & 1.00 + + 1 ............ & 4381.3 & 2.47607 & 5636 & 12.08 @xmath86 0.02 & 4.1 @xmath86 0.6 & 1.00 + 2 ............ & 4381.7 & 2.47641 & 5607 & 13.26 @xmath86 1.78 & 1.8 @xmath86 1.5 & 1.00 + 3 ............ & 4383.6 & 2.47791 & 5477 & 12.55 @xmath86 2.19 & 1.5 @xmath86 3.3 & 1.00 + 4 ............ & 4383.9 & 2.47813 & 5458 & 13.11 @xmath86 0.01 & 9.0 @xmath86 0.4 & 1.00 + 5 ............ & 4384.3 & 2.47845 & 5431 & 15.74 @xmath86 0.39 & 1.9 @xmath86 0.1 & 1.00 + 6 ............ & 4384.6 & 2.47865 & 5414 & 13.48 @xmath86 0.37 & 5.8 @xmath86 1.9 & 1.00 + 7 ............ & 4384.8 & 2.47887 & 5395 & 12.07 @xmath86 0.04 & 3.2 @xmath86 1.5 & 1.00 + 8 ............ & 4385.1 & 2.47905 & 5379 & 11.97 @xmath86 0.03 & 3.7 @xmath86 0.9 & 1.00 + 9 ............ & 4385.6 & 2.47944 & 5345 & 12.22 @xmath86 0.02 & 5.9 @xmath86 0.5 & 1.00 + 10 ........... & 4385.9 & 2.47971 & 5322 & 13.03 @xmath86 0.07 & 4.1 @xmath86 0.3 & 1.00 + 11 ........... & 4386.9 & 2.48052 & 5252 & 14.13 @xmath86 0.28 & 3.1 @xmath86 0.3 & 1.00 + 12 ........... & 4387.8 & 2.48122 & 5192 & 13.63 @xmath86 1.30 & 3.2 @xmath86 2.0 & 1.00 + 13 ........... & 4388.0 & 2.48135 & 5181 & 12.93 @xmath86 4.77 & 2.0 @xmath86 6.3 & 1.00 + + 1 ............ & 5595.1 & 2.47857 & 5420 & 13.65 @xmath86 0.01 & 4.6 @xmath86 0.1 & 1.00 + + 1 ............ & 4529.7 & 2.47855 & 5422 & 14.58 @xmath86 0.03 & 6.9 @xmath86 0.2 & 1.00 + 2 ............ & 4533.2 & 2.48125 & 5190 & 13.41 @xmath86 0.02 & 10.7 @xmath86 0.8 & 1.00 + + 1 ............ & 5303.3 & 2.47369 & 5841 & 13.05 & 21.0 & 1.00 + 2 ............ & 5304.0 & 2.47416 & 5801 & 13.12 & 22.8 & 1.00 + 3 ............ & 5304.9 & 2.47476 & 5749 & 12.95 & 21.4 & 1.00 + 4 ............ & 5306.4 & 2.47570 & 5668 & 13.62 & 11.1 & 1.00 + 5 ............ & 5307.6 & 2.47647 & 5602 & 13.04 & 13.1 & 1.00 + 6 ............ & 5310.1 & 2.47815 & 5457 & 13.24 & 14.2 & 1.00 + 7 ............ & 5310.8 & 2.47857 & 5420 & 14.07 & 11.2 & 1.00 + 8 ............ & 5312.6 & 2.47976 & 5318 & 13.17 & 12.3 & 1.00 + 9 ............ & 5313.8 & 2.48055 & 5250 & 13.43 & 5.2 & 1.00 + 10 ........... & 5314.8 & 2.48124 & 5190 & 13.22 & 7.5 & 1.00 + 11 ........... & 5315.2 & 2.48147 & 5171 & 13.46 & 11.6 & 1.00 + + + + 1 ............ & 5475.9 & 2.53693 & 430 & 13.41 @xmath960.05 & 10.9 @xmath86 0.2 & 0.94@xmath95 + + + + 1 ............ & 5498.0 & 2.55125 & 782 & 13.18 @xmath86 0.22 & 3.0 @xmath86 0.7 & 0.30@xmath97 + 2 ............ & 5500.1 & 2.55261 & 897 & 13.05 @xmath86 0.14 & 4.5 @xmath86 0.6 & 0.81@xmath98 + 3 ............ & 5500.4 & 2.55280 & 913 & 13.54 @xmath86 0.15 & 8.2 @xmath86 3.3 & 0.23@xmath99 + 4 ............ & 5501.0 & 2.55314 & 942 & 14.26 @xmath86 0.02 & 11.7 @xmath86 0.5 & 0.99@xmath100 + 5 ............ & 5501.5 & 2.55351 & 973 & 16.66 @xmath86 0.06 & 5.6 @xmath86 0.1 & 0.99@xmath100 + + 1 ............ & 4952.2 & 2.55313 & 941 & 13.05 @xmath86 0.06 & 9.4 @xmath86 0.3 & 0.66@xmath101 + 2 ............ & 4952.7 & 2.55350 & 972 & 13.41 @xmath86 0.02 & 8.4 @xmath86 0.1 & 0.96@xmath89 + + + + 1 ............ & 4469.8 & 1.88708 & 60493 & 13.07 @xmath86 0.30 & 7.5 @xmath86 1.3 & 0.73@xmath102 + 2 ............ & 4470.1 & 1.88730 & 60471 & 13.61 @xmath86 0.08 & 14.2 @xmath86 2.3 & 0.93@xmath103 + 3 ............ & 4470.5 & 1.88753 & 60448 & 13.35 @xmath86 0.16 & 8.4 @xmath86 2.7 & 0.64@xmath104 + 4 ............ & 4470.7 & 1.88771 & 60430 & 13.06 @xmath86 0.33 & 7.7 @xmath86 3.1 & 0.78@xmath105 + 5 ............ & 4471.0 & 1.88788 & 60414 & 13.34 @xmath86 0.17 & 10.8 @xmath86 1.6 & 0.67@xmath106 + + + + 1 ............ & 4587.4 & 1.96305 & 52980 & 13.49 @xmath86 0.12 & 11.2 @xmath86 0.9 & 0.52@xmath107 + 2 ............ & 4587.7 & 1.96327 & 52959 & 13.86 @xmath86 0.02 & 5.9 @xmath86 0.2 & 1.00 + 3 ............ & 4588.0 & 1.96345 & 52941 & 13.78 @xmath86 0.01 & 7.5 @xmath86 0.3 & 1.00 + 4 ............ & 4588.6 & 1.96381 & 52906 & 13.04 @xmath86 0.44 & 5.1 @xmath86 0.9 & 0.37@xmath108 + 5 ............ & 4589.4 & 1.96436 & 52852 & 14.00 @xmath86 0.03 & 8.3 @xmath86 0.2 & 0.99@xmath89 + 6 ............ & 4589.8 & 1.96459 & 52829 & 13.26 @xmath86 0.17 & 9.4 @xmath86 1.2 & 0.71@xmath109 + 7 ............ & 4590.2 & 1.96486 & 52803 & 13.40 @xmath86 0.08 & 8.3 @xmath86 0.5 & 0.89@xmath110 + 8 ............ & 4590.5 & 1.96506 & 52783 & 12.85 @xmath86 0.02 & 5.9 @xmath86 0.4 & 1.00 + + + + 1 ............ & 4752.1 & 2.06947 & 42665 & 12.70 @xmath86 0.03 & 11.6 @xmath86 1.1 & 1.00 + 2 ............ & 4752.6 & 2.06974 & 42639 & 13.75 @xmath86 0.00 & 10.0 @xmath86 0.1 & 1.00 + 3 ............ & 4752.9 & 2.06997 & 42617 & 14.04 @xmath86 0.04 & 14.8 @xmath86 1.4 & 1.00@xmath100 + 4 ............ & 4753.2 & 2.07016 & 42599 & 13.47 @xmath86 0.01 & 12.0 @xmath86 0.2 & 1.00 + 5 ............ & 4753.7 & 2.07045 & 42571 & 13.01 @xmath86 0.02 & 9.1 @xmath86 0.6 & 1.00 + 6 ............ & 4754.0 & 2.07070 & 42547 & 12.56 @xmath86 0.05 & 11.8 @xmath86 1.8 & 1.00 + + 1 ............ & 4306.3 & 2.06984 & 42629 & 13.16 @xmath86 0.02 & 14.6 @xmath86 0.7 & 1.00 + 2 ............ & 4306.6 & 2.07010 & 42604 & 12.83 @xmath86 0.03 & 7.3 @xmath86 0.6 & 1.00 + + + + 1 ............ & 5055.2 & 2.26521 & 24357 & 13.05 @xmath86 0.22 & 5.9 @xmath86 0.5 & 0.75@xmath111 + 2 ............ & 5055.5 & 2.26542 & 24337 & 13.85 @xmath86 0.16 & 13.8 @xmath86 2.8 & 0.14@xmath112 + + 1 ............ & 3939.5 & 2.26520 & 24357 & 12.76 @xmath86 0.13 & 4.2 @xmath86 0.8 & 1.00 + 2 ............ & 3939.6 & 2.26534 & 24345 & 11.82 @xmath86 0.21 & 1.7 @xmath86 3.6 & 1.00 + + + + 1 ............ & 4914.9 & 2.17460 & 32722 & 12.83 @xmath86 0.01 & 9.9 @xmath86 0.5 & 1.00 + 2 ............ & 4915.2 & 2.17481 & 32703 & 12.97 @xmath86 0.22 & 8.7 @xmath86 1.4 & 0.90@xmath113 + 3 ............ & 4915.4 & 2.17494 & 32691 & 12.25 @xmath86 0.03 & 2.0 @xmath86 1.3 & 1.00 + 4 ............ & 4915.8 & 2.17520 & 32666 & 13.33 @xmath86 0.11 & 14.8 @xmath86 0.7 & 0.70@xmath114 + 5 ............ & 4916.2 & 2.17542 & 32646 & 13.17 @xmath86 0.46 & 0.8 @xmath86 0.6 & 0.37@xmath115 + 6 ............ & 4916.4 & 2.17558 & 32631 & 13.65 @xmath86 0.02 & 8.6 @xmath86 0.3 & 0.87@xmath100 + 7 ............ & 4916.7 & 2.17578 & 32612 & 12.76 @xmath86 0.03 & 27.4 @xmath86 2.4 & 1.00 + + 1 ............ & 4424.6 & 2.17462 & 32720 & 11.96 @xmath86 0.05 & 8.2 @xmath86 1.7 & 1.00 + 2 ............ & 4424.9 & 2.17480 & 32704 & 11.67 @xmath86 0.08 & 1.2 @xmath86 4.9 & 1.00 + 3 ............ & 4425.1 & 2.17495 & 32690 & 11.74 @xmath86 0.08 & 4.7 @xmath86 2.1 & 1.00 + 4 ............ & 4425.4 & 2.17519 & 32667 & 12.04 @xmath86 0.04 & 9.3 @xmath86 1.2 & 1.00 + 5 ............ & 4426.0 & 2.17557 & 32632 & 12.51 @xmath86 0.01 & 7.2 @xmath86 0.3 & 1.00 + + 1 ............ & 3830.9 & 2.17524 & 32663 & 12.35 @xmath86 0.03 & 7.7 @xmath86 0.8 & 1.00 + 2 ............ & 3831.4 & 2.17560 & 32629 & 12.61 @xmath86 0.03 & 6.8 @xmath86 0.5 & 1.00 + + 1 ............ & & + , , and are marked with downward arrows .
the region blueward of the ly@xmath32 emission line is not shown because the spectrum normalization by continuum fitting is not reliable in the ly@xmath32 forest .
the lower line is the 1@xmath46 error spectrum .
the region of the mini - bal in system a ( 5288 5348 ) is shaded.,title="fig:",width=642 ] , , and are marked with downward arrows .
the region blueward of the ly@xmath32 emission line is not shown because the spectrum normalization by continuum fitting is not reliable in the ly@xmath32 forest .
the lower line is the 1@xmath46 error spectrum .
the region of the mini - bal in system a ( 5288 5348 ) is shaded.,title="fig:",width=642 ] , , and are marked with downward arrows .
the region blueward of the ly@xmath32 emission line is not shown because the spectrum normalization by continuum fitting is not reliable in the ly@xmath32 forest .
the lower line is the 1@xmath46 error spectrum .
the region of the mini - bal in system a ( 5288 5348 ) is shaded.,title="fig:",width=642 ] , for 8 observing epochs from march 23 , 2002 to may 31 , 2006 .
filled stars with 1@xmath46 errors denote the coverage fractions of narrow ( left ) and broad ( right ) components .
the stars are placed at the wavelengths of the blue members of the doublet ( i.e. , 1548 ) .
open stars plotted in the het spectra are coverage fractions that are interpolated from the subaru spectra . in the top left panel , we also present models of the narrow and broad components as well as the @xmath441527 components in system b from the first observation.,width=453 ] error bars are coverage fractions determined by minfit .
if a star is not displayed , the value of the coverage fraction is 1.0 .
small dots with error bars denote the coverage fractions evaluated by the pixel - by - pixel method .
the error values contain both poisson noise and error from continuum level uncertainty.,width=359 ] ) parameters over our monitoring period . from top to bottom , the frames show ( 1 ) rest frame equivalent width , ( 2 ) column densities , ( 3 ) doppler parameters , ( 4 ) coverage fractions , and ( 5 ) shift velocities .
the solid line and triangles in the top panel refer to the equivalent width of the entire mini - bal profile ( see figures 1 and 2 ) , while the model parameters refer to the bluest troughs ( see figure 3 ) .
the open and filled triangles in the top panel are measurements of the rest frame equivalent width , directly from the subaru and het data respectively .
these include all and @xmath441527 components ( the components make only a small contribution ) . in all panels ,
the blue line and open stars refer to the narrow model component while the red line and open circles refer to the broad model component .
the horizontal axis gives the observation time , both as the observed year ( top label ) and as the time in the quasar rest - frame , relative to the first observation ( bottom label ) .
only the results from subaru spectra are shown , except for the equivalent width panel , because the low - resolution and low - s / n ratio of the het+mrs spectra prevented us from fitting models .
the dotted lines in the 2nd , 3rd , and 4th panels show the best fit parameters for epoch 3 through epoch 8 , assuming a constant ( = 0.36 ) for the broad component.,width=283 ]
|
we present six new and two previously published high - resolution spectra of the quasar hs 1603 + 3820 ( @xmath0 ) taken over an interval of 4.2 years ( 1.2 years in the quasar rest frame ) .
the observations were made with the high - dispersion spectrograph on the subaru telescope and medium - resolution spectrograph on the hobby - eberly telescope .
the purpose was to study the narrow absorption lines ( nals ) .
we use time variability as well as coverage fraction analysis to separate intrinsic absorption lines , which are physically related to the quasar , from intervening absorption lines . by fitting models to the line profiles , we derive the parameters of the respective absorbers as a function of time .
only the mini - bal system at @xmath1 ( @xmath2 ) shows both partial coverage and time variability , although two nal systems possibly show evidence of partial coverage .
we find that all the troughs of the mini - bal system vary in concert and its total equivalent width variations resemble those of the coverage fraction .
however , no other correlations are seen between the variations of different model parameters .
thus , the observed variations can not be reproduced by a simple change of ionization state nor by motion of a homogeneous parcel of gas across the cylinder of sight .
we propose that the observed variations are a result of rapid continuum fluctuations , coupled with coverage fraction fluctuations caused by a clumpy screen of variable optical depth located between the continuum source and the mini - bal gas .
an alternative explanation is that the observed partial coverage signature is the result of scattering of continuum photons around the absorber , thus the equivalent width of the mini - bal can vary as the intensity of the scattered continuum changes .
|
x - ray timing has historically led to fundamental understanding of many phenomena , such as accretion - powered pulsars , low - mass x - ray binaries , and strongly magnetized neutron stars .
x - ray timing has unique power because timing signals are often clean diagnostics of systems , and hence lead directly to physical understanding . here
we discuss recent results of x - ray timing .
in particular , we focus on aspects of fundamental physics that can be addressed with timing observations .
physics in ultrastrong magnetic fields , high densities , and strong gravity can all be tested with such observations , which reveal truly small - scale structure , from the @xmath0 cm sizes of neutron stars and black holes to the @xmath1 cm radii of nucleons . in
2 we consider strongly magnetized neutron stars , and the evidence for them apparent in timing observations . in 3
we discuss thermonuclear propagation in x - ray bursts , and what it might tell us about such propagation in other astrophysical settings , such as type ia supernovae . in 4
we examine what has been learned about strong gravity and dense matter through x - ray timing observations . finally , in 5 we consider future prospects , and the exciting qualitative advances in these fundamental subjects that are expected from observations with a large - area x - ray timing mission
it has long been recognized ( e.g. , erber 1966 ) that in magnetic fields @xmath2 g , many otherwise insignificant microphysical processes become important or even dominant .
such processes include single photon pair production , photon splitting , and the existence of a vacuum resonance frequency . terrestrial laboratories
are unable to reproduce the combination of high magnetic field and high photon energy or high electron lorentz factor necessary to test the predictions of quantum electrodynamics in these extreme conditions .
the study of high - field neutron stars is therefore important for fundamental physics as well as for the astrophysical understanding it brings .
observations of soft gamma - ray repeaters provide the best current evidence for supercritical magnetic fields @xmath3 .
the first detection of soft gamma - ray repeaters was made in 1979 , with the detection of a giant burst on 5 march 1979 from the source sgr 052966 in the large magellanic cloud ( mazets et al .
1979 ) it was only much later that this source and two similar ones ( sgr 1900 + 14 and sgr 180620 ) were recognized as a class potentially distinct from classical gamma - ray bursts .
these three sources , a more recently detected fourth member of the class ( sgr 162741 ; woods et al .
1999 ) , and possibly even a fifth ( sgr 180123 ; cline et al .
2000 ) have long quiescent periods of relatively steady @xmath4 erg s@xmath5 x - ray emission punctuated by episodes of bursting behavior .
the bursts are typically short , from @xmath6 s to @xmath7 s , and have peak estimated isotropic luminosities of @xmath8 erg s @xmath5 .
the spectra are quasithermal , with a peak energy at @xmath9 kev .
the first suggestion that soft gamma - ray repeaters might be magnetars " with supercritical magnetic field strengths was made by thompson & duncan ( 1995 , 1996 ) .
the cleanest evidence in favor of the magnetar hypothesis has come from timing observations .
modulation of the persistent x - ray emission from sgr 180620 was detected during the november 1996 active period ( kouveliotou et al .
1998 ) , at a period of 7.5 s and with a period derivative @xmath10 s yr@xmath5 .
interpreted using the standard dipole braking model , these values of @xmath11 and @xmath12 imply a surface dipole field of @xmath13 g , nearly twenty times the quantum critical field . similarly , the @xmath14 s yr@xmath5 spindown of sgr 1900 + 14 ( kouveliotou et al . 1999 ) plus its 5.2 s period also implies a surface dipole field of @xmath15 g in this source .
the discovery of the high magnetic field inferred from dipole spindown models of sgr 180620 and sgr 1900 + 14 is a dramatic confirmation of the magnetar model .
alternative models must of course be considered .
for example , marsden , rothschild , & ligenfelter ( 1999 ) have noted that the environments of sgrs are unusually dense , and have suggested that accretion must therefore be reconsidered as the driving mechanism for sgr behavior .
however , the nature of the spindown supports strongly the interpretation that sgrs have supercritical magnetic fields , and hence are unique laboratories for the predictions of quantum electrodynamics in ultrastrong magnetic fields .
prior to the launch of the rossi x - ray timing explorer ( rxte ) it was expected theoretically ( e.g. , joss 1978 ; fryxell & woosley 1982 ; nozakura , ikeuchi , & fujimoto 1984 ; bildsten 1995 ) that ignition would happen at some point in the fuel layer and then spread , so that burning would not be uniform over the entire star .
these expectations were confirmed by the discovery of brightness oscillations during thermonuclear x - ray bursts from several neutron stars in low - mass x - ray binaries ( for reviews see strohmayer , zhang , & swank 1997 ; van der klis 2000 ) .
these brightness oscillations are highly coherent ( @xmath16fwhm@xmath17 ; see , e.g. , strohmayer & markwardt 1999 ) and are extremely stable in their frequency from burst to burst in a given source ( strohmayer et al .
1998 ) . their coherence and stability , plus the recent report of a burst oscillation in the x - ray millisecond pulsar sax j1808 - 3658 with a frequency equal to the spin frequency ( in t zand et al . 2000 ) , establishes that the oscillations occur at the spin frequency or its first overtone .
the physical picture is that ignition happens at some spot on the surface , and the burning then propagates away from this spot . at a given moment
, there is therefore a hot spot on the surface .
rotational modulation as seen by a distant observer then produces the observed oscillations .
the linear rotational velocity , @xmath18 , is much greater than the @xmath19 cm s@xmath5 velocity expected for thermonuclear propagation , and hence the rotational modulation provides a series of snapshots of the burning as it propagates .
the study of thermonuclear propagation in x - ray bursts has broad implications for other propagation phenomena , such as classical novae and type ia supernovae .
sne ia are especially important to understand because of their unique value as probes of the expansion of the universe and of the possible existence of a cosmological constant . however , these sources are difficult to model numerically , due to the large range of length scales involved ( from the @xmath20 cm thickness of flame fronts to the @xmath21 cm radius of the pre - sn white dwarf ) .
observations are also challenging , because the propagation of burning happens primarily deep within the white dwarf , and a given source only has one episode of such propagation in its lifetime .
this means that , even on such issues as whether the propagation happens as a deflagration ( slower than the sound speed @xmath22 in the unburned material ) or a detonation ( faster than @xmath22 ) , and the expected transition between those propagation regimes , there is considerable uncertainty .
observation of thermonuclear propagation during x - ray bursts offers a different path by which these processes can be understood .
the burning happens close to the surface ( @xmath23meters ) , where it is relatively easily observed , and happens frequently for a given source , so many observations can be made .
careful measurement of burst brightness oscillations , especially with a future high - area timing instrument , therefore holds promise for the elucidation of thermonuclear propagation in other circumstances such as type ia supernovae .
one recent result about propagation emerged from study of the burst source 4u 1636536 , in which the dominant 580 hz brightness oscillation is at twice the spin frequency ( miller 1999 ) .
this is interpreted as the existence of two nearly antipodal hot spots on the surface , perhaps the result of two pools of fuel ( e.g. , at the magnetic poles ) . in general ,
ignition is expected to happen at one pole first , then propagate to the other , so the rapidity with which the 580 hz oscillation appears indicates the speed of propagation . in the case of 4u 1636536 , the 580 hz oscillation appears within 0.03 s of the burst onset .
combined with the @xmath24 cm between antipodes , this implies a propagation velocity @xmath25 cm s@xmath5 , much larger than the @xmath0 cm s@xmath5 expected in some models of turbulent deflagration waves ( see , e.g. , fryxell & woosley 1982 ; nozakura et al . 1984 ) .
these numbers point to one of many qualitative advances that would be possible with an x - ray timing mission with 10x the collecting area of rxte .
the speed of sound in the unburned medium is expected to be @xmath26 cm s@xmath5 . with the observed signal strength , an instrument with ten times
rxte s area would be able to resolve propagation velocities up to @xmath27 cm s@xmath5 , and would therefore be able to distinguish between deflagration and a detonation and possibly see a transition between the two burning regimes .
this would be a major advance in the understanding of thermonuclear propagation in general .
many of the most dramatic advances afforded by x - ray timing have occurred in the last few years , in the subjects of strong gravity and dense matter .
this is a direct result of the high area , fast timing , and high telemetry rate of the rossi x - ray timing explorer . in this section
we summarize what has been learned from high - frequency brightness oscillations from neutron stars and black holes , and in the next section we discuss the dramatic qualitative advances in our understanding that would be made possible by a future timing mission with ten times the area of rxte .
in addition to brightness oscillations during x - ray bursts ( discussed in the previous section ) , the major discovery with rxte was the existence of khz quasiperiodic brightness oscillations ( qpos ) in neutron - star low - mass x - ray binary systems and the existence of a number of qpos in accreting black holes .
the qpos in neutron stars are seen from @xmath2320 systems , and a number of trends have been established ( see van der klis 2000 for a review ) .
these qpos ( 1 ) often appear as two simultaneous peaks in a power density spectrum , ( 2 ) are relatively sharp and strong , with @xmath28fwhm up to @xmath29 in some cases and fractional rms amplitudes up to 15% in the 2 - 60 kev range of the proportional counter array on rxte , ( 3 ) tend to increase in frequency as the inferred mass accretion rate increases , ( 4 ) have a separation frequency between the two peaks that often decreases with increasing peak frequency , ( 5 ) have a maximum separation frequency that is close to the spin frequency inferred from burst brightness oscillations .
a number of models have been proposed for this phenomenon ( see van der klis 2000 and references therein ) . for the reasons discussed in miller ( 2000 ) , we favor beat - frequency mechanisms , in which the upper peak frequency is close to the frequency of a circular orbit at some special radius , and the lower peak is generated by a beat of that frequency with radiation modulated at the stellar spin frequency ( miller , lamb , & psaltis 1998 ) . however , many of the most important inferences about strong gravity and dense matter from these oscillations can be derived more generally , from the widely accepted idea that the upper peak frequency @xmath30 is an orbital frequency .
there are two simple considerations that yield a constraint on the mass @xmath31 and the radius @xmath32 of a neutron star , from the observation of an orbital frequency @xmath33 .
first , the orbit is obviously outside the star .
second , the orbit must also be outside the innermost stable circular orbit ( isco ) , because otherwise there would be rapid inspiral and any resulting oscillation would last at most a few cycles , leading to a broad qpo , in conflict with observations . as shown by miller et al .
( 1998 ) , these limits constrain the mass of the star to be @xmath34 and the circumferential radius to be @xmath35 km , to first order in the dimensionless spin parameter @xmath36 , where @xmath37 is the stellar angular momentum ( @xmath38 for a neutron star with a spin frequency of 300 hz , and the spacetime is unique to first order in @xmath39 ) .
together , the observation of a given orbital frequency restricts the mass and radius to be inside a wedge in a @xmath40 diagram ; the larger the frequency , the more restrictive the constraint .
the highest frequency yet observed , 1330 hz ( van straaten et al . 2000 ) , rules out the hard equation of state l ( labeled as in miller et al .
this is the first time that an astrophysical observation has been able to rule out a hard equation of state .
this is therefore a major advance in our understanding of the dense matter in the cores of neutron stars .
as first discussed in 1996 , an even more major advance would occur if strong evidence were found for the existence of the isco .
this is because the existence of unstable circular orbits is a qualitatively new feature of general relativity compared to newtonian gravity , and is an essential component of models of accreting black holes on all scales as well as of inspiraling compact objects and other phenomena . given that observed frequencies increase with increasing @xmath41 , but that @xmath42 is an upper limit to the frequency of the upper peak , it was expected that there would be a rollover in the slope of the @xmath43 curve for both the upper and lower peaks ( miller et al .
moreover , the frequency at which the curve rolled over has to be a constant , for many observations of a given source .
this may have been seen in rxte data from 4u 182030 by zhang et al .
if so , this is a tremendous discovery ; not only is this strong confirmation of the prediction of the isco , but it implies a neutron star gravitational mass of @xmath44 , which limits the possible high - density equation of state severely . given these dramatic implications , care must be taken in the interpretation of the data .
the main question is whether the rollover really happens as a function of @xmath41 , or whether the countrate change is indicative of a spectral state change .
further observations and additional sources with this signature would help immensely .
this is one of the many potential benefits of a future large - area timing instrument .
there are many different types of quasiperiodic oscillations from black hole sources discovered with rxte .
one that has been seen in the microquasars grs 1915 + 105 and gro j165540 is particularly intriguing .
the frequency appears not to change with luminosity in either source ( fixed at 67 hz for grs 1915 + 105 and at @xmath45300 hz for gro j165540 ) . in grs 1915 + 105
the feature can be fairly sharp , with a maximum @xmath46 .
these qpos are much weaker than many qpos seen in neutron star sources , with typical rms amplitudes @xmath231% .
they are therefore more difficult to study , and less is known about them .
however , their fixed frequency suggests that they are related to some fundamental quantity or radius in the system ; possibilities include frame - dragging frequencies due to lense - thirring precession or periastron advance ( e.g. , stella , vietri , & morsink 1999 ) , and disk oscillation modes ( e.g. , nowak et al .
regardless of the physical origin of these qpos , the matter generating the oscillations is close enough to the black hole that it must in some sense reflect the strongly curved spacetime .
it is therefore expected that , especially with the well - characterized waveforms available with a large - area timing mission , the mass and spin of the black hole could be inferred , and potentially the spacetime itself might be probed .
an x - ray timing satellite with ten times the collecting area of rxte would be the first instrument able to characterize accurately the waveforms and energy - dependent phase lags of the high - frequency qpos observed from compact objects in low - mass x - ray binaries . as described below
, these waveforms and phase lags encode otherwise inaccessible information about strong gravity and the properties of the compact objects . with this capability ,
such an instrument would test quantitatively the predictions of general relativity in the highly curved spacetime near neutron stars and black holes .
it would also make fundamentally new contributions to our understanding of the properties of matter at supranuclear densities .
the waveform from a rotating hot spot during a burst oscillation or persistent emission is affected by gravitational light deflection and doppler shifts . by fitting the waveform
one can therefore constrain simultaneously the mass and radius of the star , and hence the equation of state of matter beyond nuclear saturation density , which is an issue of central importance in nuclear physics research . fig .
1a shows that observations from a large - area timing instrument would bring an unprecedented level of precision to our understanding of high - density matter . in this simulation
we assume a neutron star of gravitational mass 1.8 @xmath47 with high - density equation of state uu , and assume that the hot spot is a small spot on the rotational equator , as seen by a distant observer in the rotational plane .
we simulate five bursts which had 5% rms oscillations at 364 hz for five seconds each .
modern equations of state have mass - radius curves that span approximately the range between the fps and uu equations of state ( akmal , pandharipande , & ravenhall 1998 ) .
it is clear that measurements of this type for several different neutron stars would exclude most of these eos .
these results would therefore have direct feedback on nuclear physics .
similarly , accurate measurement of the waveforms of luminous clumps around a black hole would yield a precise mass for the hole and strong constraints on the kerr spin parameter @xmath48 .
it may even be possible to detect the frequency chirp produced as a clump moves inside the last stable orbit ( sunyaev 1973 , stoeger 1980 ) .
the high velocity of surface rotation during bursts or persistent coherent emission also doppler - boosts the energies of photons from the hot spot , causing high - energy photons to lead low - energy photons by a phase that depends on the precise value of the surface velocity and therefore on the stellar radius , for fixed frequency .
this effect was first pointed out by ford ( 1999 ) in an analysis of rxte data on bursts from aql x-1 , and has also been analyzed for the coherent oscillations from the millisecond x - ray pulsar sax j18083658 by weinberg , miller , & lamb ( 2000 ) and ford ( 2000 ) . in fig .
1b we simulate a fit to such phase lag data , with an assumed gravitational mass of @xmath49 and equation of state uu , and phase uncertainties taken from the sax j18083658 data .
this figure shows that although rxte is not sufficiently sensitive to provide useful constraints , an observation with a future large - area timing instrument would yield tight limits on the mass and the radius of this neutron star , and hence on the equation of state of high - density matter .
the most exciting prospect of a high - area timing mission is that data from it could test quantitatively the predictions of general relativity in strong gravity . in the previous sections we have discussed several independent methods of estimating the mass , radius , or angular momentum of an individual neutron star or black hole .
these include waveform and phase lag fits and the detection of an isco and the associated orbital frequency .
the combination of all these independent measurements overdetermines the problem so that the underlying theories get tested .
for example , the various estimates of the mass of the central object will agree with each other only if general relativity is the correct theory of gravity , because general relativity is assumed in making the estimates .
this would provide a powerful test of its predictions in a regime that is otherwise not accessible , and would continue the legacy of x - ray timing as a clean measuring tool for fundamental physics and astrophysics .
|
x - ray timing observations of neutron stars and black holes are among the few available probes of ultrastrong magnetic fields , strong gravity , high densities , and the propagation of thermonuclear burning . here
we review the evidence for these effects revealed with data from the rossi explorer in the last five years .
we also discuss the exciting prospects for making the first quantitative tests of strong - gravity general relativistic predictions with a large - area x - ray timing mission .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in
|
cluster - tilted algebras are finite dimensional associative algebras which were introduced in @xcite and , independently , in @xcite for the type @xmath12 .
one motivation for introducing these algebras came from fomin and zelevinsky s cluster algebras @xcite . to every cluster in an acyclic cluster algebra one can associate a cluster - tilted algebra , and the indecomposable rigid modules over the cluster - tilted algebra correspond bijectively to the cluster variables outside the chosen cluster .
generalizations of cluster - tilted algebras , the jacobian algebras of quivers with potentials , were introduced in @xcite , extending this correspondence to the non - acyclic types .
many people have studied cluster - tilted algebras in this context , see for example @xcite .
the second motivation came from classical tilting theory .
tilted algebras are the endomorphism algebras of tilting modules over hereditary algebras , whereas cluster - tilted algebras are the endomorphism algebras of cluster - tilting objects over cluster categories of hereditary algebras .
this similarity in the two definitions lead to the following precise relation between tilted and cluster - tilted algebras , which was established in @xcite .
there is a surjective map @xmath13 & \{\textup{cluster - tilted algebras}\ } , & c\ar@{|->}[r]&b = c\ltimes e , } \ ] ] where @xmath14 denotes the @xmath0-@xmath0-bimodule @xmath15 and @xmath16 is the trivial extension .
this result allows one to define cluster - tilted algebras without using the cluster category .
it is natural to ask how the module categories of @xmath0 and @xmath7 are related , and several results in this direction have been obtained , see for example @xcite .
the hochschild cohomology of the algebras @xmath0 and @xmath7 has been compared in @xcite . in this paper
, we use a new approach to study the relation between the module categories of a tilted algebra @xmath0 and its cluster - tilted algebra @xmath1 , namely _ induction _ and _
coinduction_. the induction functor @xmath2 and the coinduction functor @xmath17 from @xmath18 to @xmath19 are defined whenever @xmath0 is a subring of @xmath7 which has the same identity .
if we are dealing with algebras over a field @xmath20 , we can , and usually do , write the coinduction functor as @xmath21 , where @xmath22 is the standard duality . induction and coinduction are important tools in classical representation theory of finite groups . in this case , @xmath7 would be the group algebra of a finite group @xmath23 and @xmath0 the group algebra of a subgroup of @xmath23 ( over a field whose characteristic is not dividing the group orders ) .
in this situation , the algebras are semi - simple , induction and coinduction are the same functor , and this functor is exact . for arbitrary rings , and even for finite dimensional algebras , the situation is not that simple .
in general , induction and coinduction are not the same functor and , since the @xmath0-module @xmath7 is not projective ( and not flat ) , induction and coinduction are not exact functors
. however , the connection between tilted algebras and cluster - tilted algebras is close enough so that induction and coinduction are interesting tools for the study of the relation between the module categories .
induction and coinduction have been studied for split extension algebras in @xcite , and we apply some of their results to our situation .
our first main result is on the @xmath0-@xmath0-bimodule @xmath24 considering its right @xmath0-module structure @xmath25 as well as its left @xmath0-module structure @xmath26 but as a right @xmath0-module @xmath27 .
we show the following .
[ thm 1 ] if @xmath0 is a tilted algebra and @xmath7 is the corresponding cluster - tilted algebra , then + @xmath4 is a ( classical ) partial tilting and @xmath28-rigid @xmath0-module , and its corresponding induced module @xmath29 is a ( classical ) partial tilting and @xmath30-rigid @xmath7-module .
+ @xmath14 is a partial cotilting and @xmath28-corigid @xmath0-module , and its corresponding coinduced module @xmath31 is a partial cotilting and @xmath30-corigid @xmath7-module .
we think it would be an interesting problem to study the possible completions of these partial ( co)-tilting modules and their endomorphism algebras .
our second main result compares induction and coinduction with the well - known equivalence of categories @xmath32 of buan , marsh and reiten @xcite . here
@xmath33 denotes the cluster category . recall that mod@xmath34 naturally embeds in @xmath33
we show that induction commutes with this equivalence on the subcategory @xmath35 given by the torsion subcategory determined by the tilting module @xmath9 , @xmath36 we also show a similar statement for the coinduction functor using the torsion free subcategory @xmath37 determined by @xmath9 , @xmath38 we show the following .
[ thm 3 ] let @xmath0 be a tilted algebra and @xmath7 the corresponding cluster - tilted algebra .
then @xmath39 let us point out that these formulas implicitly use the tilting theorem of brenner and butler @xcite .
finally , we address the question which @xmath7-modules are actually induced or coinduced from a tilted algebra . we obtain the following results .
[ thm 4 ] let @xmath7 be a cluster - tilted algebra .
* if @xmath7 is of finite representation type , then _ every _
@xmath7-module is induced and coinduced from some tilted algebra . *
if @xmath7 is of arbitrary representation type , then every _ transjective _ @xmath7-module is induced or coinduced from some tilted algebra . * if @xmath7 is cluster concealed , then _ every _
@xmath7-module is induced or coinduced from some tilted algebra . *
if @xmath7 is of tame representation type and such that there are no morphisms between indecomposable projective non - transjective modules , then _ every _
indecomposable @xmath7-module is induced or coinduced from some tilted algebra .
we remark that for general cluster - tilted algebras , there are indecomposable modules that are not induced and not coinduced .
example [ ex 7.7 ] depicts a cluster - tilted algebra with infinitely many such modules , all lying in the same tube .
we think it would be interesting to study the structure of these modules . using induction and coinduction functors we also constructed explicit injective presentations for the induced modules in cluster - tilted algebras in @xcite .
the paper is organized as follows .
we start by recalling some background on tilting theory in section [ sect : basics ] and reviewing general results on induction and coinduction in section [ sect 3 ] . in section [ sect 4 ] , we study induction and coinduction from tilted to cluster - tilted algebras and prove theorem [ thm 1 ] .
the relation to cluster categories and the proof of theorem [ thm 3 ] are contained in section [ sect 6 ] . in section [
sect 7 ] , we study the question which @xmath7-modules are induced or coinduced from a tilted algebra and prove theorem [ thm 4 ] .
in particular , part ( a ) follows from theorem [ 7.0 ] , part ( b ) follows from theorem [ 7.1 ] , part ( c ) follows from theorem [ 7.2 ] , and pard ( d ) follows from part ( b ) and theorem [ 7.3 ] .
throughout this paper all algebras are assumed to be basic , finite dimensional over an algebraically closed field @xmath20 .
suppose @xmath40 is a connected quiver without oriented cycles .
by @xmath41 we denote the path algebra of @xmath42 .
if @xmath43 is a @xmath20-algebra then denote by mod@xmath44 the category of finitely generated right @xmath43-modules and by ind@xmath44 a set of representatives of each isoclass of indecomposable right @xmath43-modules .
given @xmath45 mod@xmath44 , the projective dimension of @xmath46 in mod@xmath44 is denoted by pd@xmath47 and its injective dimension by id@xmath47 .
let @xmath5 be the auslander - reiten translation . also define @xmath48 to be the first syzygy and @xmath49 the first cosyzygy of @xmath46 .
by @xmath50 we understand the standard duality functor hom@xmath51 .
finally , let @xmath52 be the nakayama functor and @xmath53 be the inverse nakayama functor .
for further details on representation theory we refer to @xcite .
we recall the classical definition of a tilting module .
[ def 1 ] let @xmath43 be an algebra .
a @xmath43-module @xmath9 is called a _ partial tilting module _ if the following two conditions are satisfied : * pd@xmath54 , * @xmath55 . a partial tilting module @xmath9
is called a _ tilting module _ if it also satisfied the following additional condition : * there exists a short exact sequence @xmath56 with @xmath57 .
let @xmath10 be a hereditary algebra and let @xmath9 be a basic tilting @xmath10-module .
it is well known that in this case , @xmath9 is tilting if ext@xmath58 and the number of indecomposable direct summands of @xmath9 equals the number of isomorphism classes of simple @xmath10-modules .
the corresponding algebra @xmath8 is called a _ tilted algebra_. consider the following full subcategories of @xmath59 .
@xmath60 @xmath61 then @xmath62 is a torsion pair in mod@xmath34 that determines another torsion pair @xmath63 in @xmath18 , where @xmath64 @xmath65 for more details refer to ( * ? ?
* , and viii ) . with this notation
consider the following theorem due to brenner and butler , see @xcite for details .
[ 2.1 ] let @xmath10 be a hereditary algebra , @xmath9 a tilting @xmath10-module , and . then + the functors @xmath66 and @xmath67 induce quasi - inverse equivalences between @xmath35 and @xmath68 ; + the functors @xmath69 and @xmath70 induce quasi - inverse equivalences between @xmath37 and @xmath71 .
+ moreover , for any @xmath72 we have @xmath73 and @xmath74 , and for any @xmath75 we have @xmath76 and @xmath77 . let @xmath43 be a @xmath20-algebra , then let @xmath78 denote the auslander - reiten quiver of mod@xmath44 . a connected component of @xmath78 is called _ regular _ if it contains neither projective nor injective modules .
an indecomposable @xmath43-module is called _ regular _ if it belongs to a regular component of @xmath78 .
there is a close relationship between the auslander - reiten quivers of a hereditary algebra and its tilted algebras .
we shall need the following result .
[ 2.3]@xcite let @xmath10 be a representation - infinite hereditary algebra , @xmath9 be a preprojective tilting @xmath10-module , and @xmath8 .
then + @xmath35 contains all but finitely many nonisomorphic indecomposable @xmath10-modules , and any indecomposable @xmath10-module not in @xmath35 is preprojective ; + the image under the functor @xmath79 of regular @xmath10-modules yields all regular @xmath0-modules ; + the injective and projective dimension of all regular @xmath0-modules is at most one .
the following proposition describes several facts about tilted algebras , which we will use throughout the paper .
we recall some terminology first .
let @xmath43 be an algebra .
a _ path _ in mod@xmath44 with source @xmath80 and target @xmath81 is a sequence of non - zero morphisms @xmath82 where @xmath83 for all @xmath84 , and @xmath85 . in this case , @xmath80 is called a _ predecessor _ of @xmath81 in mod@xmath44 and @xmath81
is called a _ successor _ of @xmath80 in mod@xmath44 . a _ path _ in @xmath78 with source @xmath80 and target @xmath81 is a sequence of arrows @xmath82 in the auslander - reiten quiver . in addition , such path is called _ sectional _ if for each @xmath84 with @xmath86 , we have @xmath87 . [ 2.2 ] let @xmath10 be a hereditary algebra , @xmath9 a tilting @xmath10-module , and the corresponding tilted algebra . then * c @xmath88 . * for all @xmath89 or @xmath89 . * for all @xmath90 @xmath89 . * for all @xmath91 @xmath89 .
* @xmath63 is splitting , which means that every indecomposable c - module belongs either to @xmath71 or @xmath68 .
* @xmath68 is closed under predecessors and @xmath71 is closed under successors .
there is also a precise description of injective modules in a tilted algebra in terms of the corresponding hereditary algebra .
* proposition vi 5.8)[2.4 ] let @xmath10 be a hereditary algebra , @xmath9 a tilting @xmath10-module , and @xmath8 .
let @xmath92 be a complete set of pairwise nonisomorphic indecomposable direct summands of @xmath9 .
assume that the modules @xmath93 are projective , the remaining modules @xmath94 are not projective , and @xmath95 are indecomposable injective @xmath10-modules with @xmath96 , for @xmath97 . then the @xmath0-modules @xmath98 form a complete set of pairwise nonisomorphic indecomposable injective modules .
let @xmath99 and let @xmath100 denote the derived category of bounded complexes of @xmath10-modules .
the _ cluster category _
@xmath33 is defined as the orbit category of the derived category with respect to the functor @xmath101 $ ] , where @xmath102 is the auslander - reiten translation in the derived category and @xmath103 $ ] is the shift .
cluster categories were introduced in @xcite , and in @xcite for type @xmath12 , and were further studied in @xcite .
they are triangulated categories @xcite , that have serre duality and are 2-calabi yau @xcite .
an object @xmath9 in @xmath33 is called _ cluster - tilting _
if @xmath104 and @xmath9 has @xmath105 non - isomorphic indecomposable direct summands .
the endomorphism algebra @xmath106 of a cluster - tilting object is called a _ cluster - tilted algebra _ @xcite .
the following theorem will be useful later . [
2.5 ] @xcite if @xmath9 is a cluster - tilting object in @xmath33 , then @xmath107 induces an equivalence of categories @xmath108 .
let @xmath0 be an algebra of global dimension at most two and let @xmath14 be the @xmath0-@xmath0-bimodule @xmath109 .
relation extension _ of @xmath0 is the trivial extension algebra @xmath1 , whose underlying @xmath0-module is @xmath110 , and multiplication is given by @xmath111 .
relation extensions where introduced in @xcite . in the special case where @xmath0 is a tilted algebra , we have the following result .
@xcite let @xmath0 be a tilted algebra .
then @xmath112 is a cluster - tilted algebra .
moreover all cluster - tilted algebras are of this form .
this shows that the tilted algebra @xmath0 is a subalgebra and a quotient of the cluster - tilted algebra @xmath7 .
let @xmath113 be the corresponding surjective algebra homomorphism .
we will often consider a @xmath0-module @xmath46 as a @xmath7-module with action @xmath114 .
in particular , @xmath0 and @xmath14 are right @xmath7-modules , and there exists a short exact sequence in mod@xmath115 @xmath116&e\ar[r]^i&b\ar[r]^{\pi}&c\ar[r]&0}.\ ] ] let @xmath43 be a @xmath20-algebra . a _ slice _ @xmath117 in @xmath78 is a set of indecomposable @xmath43-modules such that a. @xmath117 is sincere , meaning @xmath118 for any projective @xmath43-module @xmath119 . b. any path in mod@xmath44 with source and target in @xmath117 consists entirely of modules in @xmath117 .
c. if @xmath46 is an indecomposable non - projective @xmath43-module then at most one of @xmath46 , @xmath120 belongs to @xmath117 .
d. if @xmath121 is an irreducible morphism with @xmath122 and @xmath123 , then either @xmath46 belongs to @xmath117 or @xmath46 is non - injective and @xmath124 belongs to @xmath117 .
given an indecomposable @xmath43-module @xmath81 in a connected component @xmath125 of @xmath78 let @xmath126 @xmath127 be two subquivers of @xmath125 . consider the following result due to ringel .
@xcite [ 2.6 ] let @xmath81 be an indecomposable sincere module in a preprojective or preinjective component .
then both @xmath128 and @xmath129 are slices .
the existence of slices is used to characterize tilted algebras in the following way . [ 2.7 ] @xcite let @xmath8 be a tilted algebra . then the @xmath0-module @xmath130 forms a slice in @xmath131 .
conversely , any slice in any module category is obtained in this way .
the following notion of local slices has been introduced in @xcite in the context of cluster - tilted algebras . a _ local slice _
@xmath117 in @xmath78 is a set of indecomposable @xmath43-modules inducing a connected full subquiver of @xmath78 such that a. if @xmath132 and @xmath133 is an arrow in @xmath78 then either @xmath134 or @xmath135 . b. if @xmath134 and @xmath133 is an arrow in @xmath78 then either @xmath136 or @xmath137 .
c. for every sectional path @xmath82 in @xmath78 with @xmath138 we have @xmath139 , for @xmath140 .
d. the number of indecomposables in @xmath117 equals the number of nonisomorphic summands of @xmath9 , where @xmath9 is a tilting @xmath43-module .
the definition of a local slice makes sense if we replace the algebra @xmath43 by a cluster category @xmath33 , and then consider objects of @xmath33 instead of @xmath43-modules .
there is a relationship between tilted and cluster - tilted algebras given in terms of slices and local slices . given a @xmath43-module @xmath46
let @xmath141 be its _
right annihilator_. [ 2.8]@xcite let @xmath0 be a tilted algebra and
@xmath7 be the corresponding cluster - tilted algebra .
then any slice in @xmath131 embeds as a local slice in @xmath142 and any local slice @xmath117 in @xmath142 arises in this way .
moreover , @xmath143 .
the existence of local slices in a cluster - tilted algebra gives rise to the following definition .
the unique connected component of @xmath144 that contains local slices is called the _ transjective _ component . in particular
if @xmath7 is of finite representation type then the transjective component is the entire auslander - reiten quiver of mod@xmath115 . in this case
there is the following statement .
[ 2.9]@xcite let @xmath7 be a representation - finite cluster - tilted algebra .
then any indecomposable @xmath7-module lies on a local slice .
we end this section with a lemma that we will need later .
[ lemls ] let @xmath145 be a local slice in the cluster category @xmath33 .
then @xmath146 is a local slice in @xmath19 if and only if @xmath145 contains no summand of @xmath147 .
@xmath148 is a bijection if and only if @xmath145 contains no summand of @xmath147 .
now the statement follows from ( * ? ? ?
* lemma 17 ) .
in this section we define two functors called induction and coinduction and describe some general results about them .
suppose there are two @xmath20-algebras @xmath0 and @xmath7 with the property that @xmath0 is a subalgebra of @xmath7 and they share the same identity .
then there is a general construction via the tensor product , also known as _ extension of scalars _ , that sends a @xmath0-module to a particular @xmath7-module .
we give a precise definition below .
let @xmath0 be a subalgebra of @xmath7 , such that @xmath149 , then @xmath150 is called the _ induction functor _ , and dually @xmath151 is called the _
coinduction functor_. moreover , given @xmath152 mod@xmath153 the corresponding _ induced module _ is defined to be @xmath154 , and the _ coinduced module _ is defined to be @xmath155 .
first observe that both functors are covariant .
the induction functor is right exact , while the coinduction functor is left exact .
now consider the following lemma . [ * ] let @xmath0 and @xmath7 be two @xmath20-algebras and @xmath156 a @xmath0-@xmath7-bimodule , then as @xmath7-modules for all .
@xmath157 .
the next proposition describes an alternative definition of these functors , and we will use these two descriptions interchangeably . [ 3.3 ] let @xmath0 be a subalgebra of @xmath7 such that @xmath149 , then for every @xmath152 @xmath153 + both parts follow from lemma [ * ] .
next we show some basic properties of these functors . [
3.4 ] let @xmath0 be a subalgebra of @xmath7 such that @xmath149 . if @xmath158 is an idempotent then + @xmath159 .
+ @xmath160 .
+ in particular , if @xmath161 and @xmath162 are indecomposable projective and injective @xmath0-modules at vertex @xmath84 , then @xmath163 and @xmath164 are respectively indecomposable projective and injective @xmath7-modules at vertex @xmath84 .
observe that @xmath165 , and similarly we have @xmath166 .
the rest of the proposition follows from above if we let @xmath167 be the primitive idempotent given by the constant path at vertex @xmath84 .
we can say more in the situation when @xmath7 is a split extension of @xmath0 .
let @xmath7 and @xmath0 be two algebras .
@xmath7 is a split extension of @xmath0 by a nilpotent bimodule @xmath14 _ if there exist algebra homomorphisms @xmath113 and @xmath168 such that @xmath169 is the identity on @xmath0 and @xmath170 is a nilpotent ( two - sided ) ideal of @xmath7 .
in particular , there exists the following short exact sequence of @xmath7-modules . @xmath171&e\ar[r]^{i}&b \ar@<.5ex>[r]^{\pi}&c \ar@<.5ex>[l]^{\sigma}\ar[r]&0}\ ] ] for example , relation extensions are split extensions .
if @xmath7 is a split extension of @xmath0 then @xmath172 is injective , which means @xmath0 is a subalgebra of @xmath7 .
also , @xmath14 is a @xmath0-@xmath0-bimodule , and we require @xmath14 to be nilpotent so that @xmath173 . observe that @xmath174 as @xmath0-modules , and there is an isomorphism of @xmath0-modules @xmath175 .
similarly , @xmath176 as @xmath0-modules .
this shows that induction and coinduction of a module @xmath46 yields the same module @xmath46 plus possibly something else .
the next proposition shows a precise relationship between a given @xmath0-module and its image under the induction and coinduction functors . [ 3.6 ] suppose @xmath7 is a split extension of @xmath0 by a nilpotent bimodule @xmath14 , then for every @xmath152 @xmath153 there exist two short exact sequences of @xmath7-modules : + .
+ . to show part ( a ) we apply @xmath177 to the short exact sequence and
obtain the following long exact sequence @xmath178&m\otimes_c e\ar[r ] & m\otimes _ c b \ar[r]&m\otimes _ c c\ar[r]&0}.\ ] ] however , tor@xmath179 since @xmath0 is a projective @xmath0-module , so part ( a ) follows .
similarly , to show part ( b ) we apply @xmath180 to sequence .
this yields a long exact sequence @xmath181&d(c\otimes_c dm)\ar[r]&d(b\otimes_c dm)\ar[r]&d(e\otimes_c dm)\ar[r]&d\text{tor}^c_1 ( c , dm)}.\ ] ] the last term in the sequence is again zero , which shows part ( b ) .
thus , in this situation each module is a quotient of its induced module and a submodule of its coinduced module . [ 3.7 ] suppose @xmath7 is a split extension of @xmath0 by a nilpotent bimodule @xmath14 , then for every @xmath182 @xmath153 + if @xmath183 then @xmath184 and @xmath155 lie in @xmath185 .
+ @xmath186 if and only if @xmath187 .
+ @xmath188 if and only if @xmath187 .
part ( a ) .
suppose @xmath189 , but @xmath190 where neither @xmath191 nor @xmath192 is the zero module . because @xmath7 is a split extension of @xmath0
, it follows that @xmath0 has a left @xmath7-module structure .
thus , @xmath193 , and we conclude that @xmath194 . by assumption @xmath46
is indecomposable , so without loss of generality let @xmath195 . but @xmath196 , which is a contradiction . therefore , @xmath184 is indecomposable .
a similar argument implies that if @xmath197 then @xmath198 .
part ( b ) .
suppose @xmath199 .
again , since @xmath0 has a left @xmath7-module structure , we have that @xmath200 , which means @xmath201 .
the proof of part ( c ) is similar to that of part ( b ) and we omit it .
the next lemma describes a relationship between the auslander - reiten translations in @xmath18 and @xmath19 , and induction and coinduction functors .
this lemma together with the following theorem were shown in @xcite .
[ 3.8 ] suppose @xmath7 is a split extension of @xmath0 by a nilpotent bimodule , then for every @xmath45 @xmath153 + @xmath202 .
+ @xmath203 .
[ 3.9 ] suppose @xmath7 is a split extension of @xmath0 by a nilpotent bimodule @xmath14 and @xmath204 @xmath153 , then + @xmath205 is a ( partial ) tilting @xmath7-module if and only if @xmath9 is a ( partial ) tilting @xmath0-module , @xmath206 and @xmath207 . + @xmath208 is a ( partial ) cotilting @xmath7-module if and only if @xmath9 is a ( partial ) cotilting @xmath0-module , @xmath209 and @xmath210 .
in this section we develop properties of the induction and coinduction functors particularly when @xmath0 is an algebra of global dimension at most two and @xmath1 is the trivial extension of @xmath0 by the @xmath0-@xmath0-bimodule @xmath211 . in the specific case when @xmath0 is also a tilted algebra , @xmath7 is the corresponding cluster - tilted algebra .
some of the results in this section only hold when @xmath0 is tilted , but many hold in a more general situation when gl.dim@xmath212 , and we make that distinction clear in the assumptions of each statement . however , throughout this section we always assume @xmath213 and a tensor product @xmath214 is a tensor product over @xmath0 .
the main result of this section holds when @xmath0 is a tilted algebra .
it says that @xmath4 is a partial tilting and @xmath215-rigid @xmath0-module , and the corresponding induced module @xmath29 is a partial tilting and @xmath216-rigid @xmath7-module .
we begin by establishing some properties of @xmath14 and @xmath4 when the corresponding algebra @xmath0 has global dimension at most two . [ 4.1 ] let @xmath0 be an algebra of global dimension at most 2
. then + since the global dimension of @xmath0 is at most 2 , we have id@xmath217 ( resp .
pd@xmath218 ) for all @xmath0-modules @xmath46 . therefore ,
when applying the auslander - reiten formula below , we obtain the full hom - space and not its quotient by the space of morphisms factoring through projectives ( respectively injectives ) .
+ part ( a ) .
@xmath219 . + part ( b ) .
@xmath220 . + part ( c ) . using lemma [ * ]
we have @xmath221 , which in turn by part ( b ) is isomorphic to @xmath222 .
then , we have the following chain of isomorphisms @xmath223 .
part ( d ) can be shown in a similar manner as above . [ 4.3 ] let @xmath0 be an algebra of global dimension at most 2 , and let @xmath1 .
suppose @xmath152 @xmath153 , then + @xmath224 if and only if @xmath225 .
+ @xmath224 if and only if @xmath226 .
part ( a ) .
recall that @xmath227 as @xmath0-modules .
therefore , it suffices to show that @xmath228 if and only if id@xmath224 . proposition [ 4.1](c ) implies that @xmath229 , which is zero if and only if id@xmath224 .
+ part ( b ) .
similarly it suffices to show that @xmath230 if and only if pd@xmath224 .
however , proposition [ 4.1](d ) implies that @xmath231 , which again is zero if and only if pd@xmath224 .
[ 4.2 ] let @xmath0 be an algebra of global dimension 2 . then for all + @xmath232 for all nonzero .
+ @xmath232 for all nonzero .
part ( a ) . by proposition
[ 4.1 ] ( c ) , @xmath233 , which is nonzero if and only if id@xmath234 .
now , consider a minimal injective resolution of @xmath46 and @xmath235 @xmath236&&&&\\ 0\ar[r]&m\ar[r]&i^0\ar[rr]\ar[ur]^{\pi}&&i^1\ar[r]&i^2\ar[r]&0.}\ ] ] apply @xmath237 to find a minimal projective resolution of @xmath238 .
@xmath239&\nu^{-1}\omega^{-1}m\ar[r]&\nu^{-1}i^1\ar[r]&\nu^{-1}i^2\ar[r]&\tau^{-1}\omega^{-1}m\ar[r]&0.}\ ] ] since @xmath0 has global dimension two , @xmath240 is a projective @xmath0-module .
it remains to show that it is nonzero . by definition @xmath241 , which is nonzero since we have a nonzero map @xmath242 .
finally , observe that the argument above holds if we replace @xmath238 by a nonzero direct summand of @xmath238 .
this completes the proof .
part ( b ) can be shown in a similar manner as above . [
4.6 ] let @xmath0 be an algebra of global dimension at most 2 , then + we will show parts ( a ) and ( c ) , and the rest of the lemma can be proven similarly .
+ part ( a ) . by proposition [ 4.1](a )
, we see that ext@xmath243 , which in turn by the auslander - reiten formula is isomorphic to @xmath244 .
let @xmath245 be an injective envelope of @xmath0 , thus we have the following short exact sequence @xmath246&c\ar[r]^i & i\ar[r]^-{\pi}&\omega^{-1}c\ar[r]&0}.\ ] ] applying hom@xmath247 to this sequence we obtain an exact sequence @xmath181 & \text{hom}_c(c , c)\ar[r]&\text{hom}_c ( c , i)\ar[r]^-{\pi_*}&\text{hom}_c(c,\omega^{-1}c)\ar[r]&\text{ext}^1_c(c , c)}.\ ] ] however , ext@xmath248 shows that @xmath249 is surjective .
this implies that every morphism from @xmath0 to @xmath250 factors through the injective @xmath251 .
thus , @xmath252 , and this shows part ( a ) .
+ part ( c ) . as above
observe that ext@xmath253 . applying hom@xmath254 to the sequence we get an exact sequence @xmath181 & \text{hom}_c(e , c)\ar[r]&\text{hom}_c ( e , i)\ar[r]^-{\pi_{**}}&\text{hom}_c(e,\omega^{-1}c)\ar[r]&\text{ext}^1_c(e , c)}.\ ] ] but then by part ( a ) we have ext@xmath255 , which shows that @xmath256 is surjective .
thus @xmath257 , and this completes the proof of part ( c ) . [ 4.7 ] if the global dimension of @xmath0 is at most two , then both @xmath258 and @xmath259 are rigid modules .
observe that @xmath260 .
the first summand is zero because of lemma [ 4.6 ] , and the second is zero because @xmath0 is projective .
the proof of the rigidity of @xmath259 is similar .
observe that , unless @xmath0 is hereditary , @xmath258 is rigid but not tilting , because pd@xmath261 , by lemma [ 4.2 ] .
next , we consider the case when @xmath0 is tilted .
[ 4.4 ] let @xmath0 be a tilted algebra . then for all + @xmath262 .
+ @xmath263 .
part ( a ) follows from lemma [ 4.2](a ) and proposition [ 2.2](b ) .
similarly , part ( b ) follows from lemma [ 4.2](b ) and proposition [ 2.2](b ) .
part ( a ) of the following proposition is well - known , see @xcite .
let @xmath0 be a tilted algebra .
then + part ( a ) .
proposition [ 4.1](c ) implies that @xmath264 , but this is zero since id@xmath265 , by lemma [ 4.4](a ) with @xmath266 .
part ( b ) follows directly from part ( a )
. the above proposition does not hold if @xmath0 has global dimension 2 , but is not tilted .
for example , consider the algebra @xmath0 given by the following quiver with relations .
@xmath267^{\alpha}&2\ar[r]^{\beta}&3\ar[r]^{\gamma}&4\ar[r]^{\delta}&5 & & \alpha\beta=\gamma\delta=0.}\ ] ] the auslander - reiten quiver of @xmath0 is the following .
@xmath268 & & & & & & { \begin{smallmatrix}1\\2\end{smallmatrix}}\ar[dr ] \\ { \begin{smallmatrix}5\end{smallmatrix}}\ar[ur ] & & { \begin{smallmatrix}4\end{smallmatrix}}\ar[dr ] & & { \begin{smallmatrix}3\end{smallmatrix}}\ar[dr ] & & { \begin{smallmatrix}2\end{smallmatrix}}\ar[ur ] & & { \begin{smallmatrix}1\end{smallmatrix}}\\ & & & { \begin{smallmatrix}3\\4\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}2\\3\end{smallmatrix}}\ar[ur ] & & \\ & & & & { \begin{smallmatrix}2\\3\\4\end{smallmatrix}}\ar[ur ] & & } \ ] ] here @xmath269 , where @xmath270 is the simple module at vertex @xmath84 .
hence , @xmath271 .
this particular algebra @xmath0 also appears in ( * ? ? ?
* remark 4.21 ) , where the authors stress that @xmath16 is not a cluster - tilted algebra . on the other hand ,
the tensor algebra @xmath272 , which in this case equals @xmath273 , is a cluster - tilted algebra . as a consequence of proposition [ 4.3 ] and lemma [ 3.8 ]
we obtain the following corollary , which says that a slice in a tilted algebra together with its @xmath5 and @xmath274-translates fully embeds in the cluster - tilted algebra .
this result was already shown in [ abs4 ] relying on the main theorem of [ az ] . here
we present a new proof using induction and coinduction functors .
[ 4.5 ] let @xmath0 be a tilted algebra and @xmath7 the corresponding cluster - tilted algebra .
let @xmath117 be a slice in @xmath18 and @xmath46 a module in @xmath117 .
then + @xmath275 , + @xmath276 .
we show part ( a ) and the proof of part ( b ) is similar . since the module @xmath46 lies on a slice in mod@xmath153 then pd@xmath224 and id@xmath224 . by lemma [ 3.8](a )
we have the following isomorphism @xmath277 it follows from proposition [ 4.3](a ) that the left hand side is isomorphic to @xmath278 .
it remains to show that the right hand side is isomorphic to @xmath279 .
we may suppose without loss of generality that @xmath8 and @xmath280 lies in @xmath68 . hence , in particular @xmath281 .
proposition [ 2.2](f ) implies that @xmath282 , so part ( d ) of the same proposition shows that pd@xmath283 .
therefore , @xmath284 by proposition [ 4.3](b ) , and this finishes the proof . the following theorem is the main result of this section .
following @xcite we say that a @xmath43-module @xmath46 is @xmath285__-rigid _ _ if @xmath286 , and similarly that @xmath46 is @xmath285__-corigid _ _ if @xmath287 . also , @xmath46 is called _ partial cotilting _ if id@xmath288 and @xmath289 .
[ 4.8 ] if @xmath0 is a tilted algebra and @xmath7 is the corresponding cluster - tilted algebra , then + @xmath4 is a partial tilting and @xmath28-rigid @xmath0-module , and its corresponding induced module @xmath29 is a partial tilting and @xmath30-rigid @xmath7-module .
+ @xmath14 is a partial cotilting and @xmath28-corigid @xmath0-module , and its corresponding coinduced module @xmath31 is a partial cotilting and @xmath30-corigid @xmath7-module .
we show part ( a ) , and the proof of part ( b ) is similar .
first let us show that @xmath4 is a partial tilting and @xmath28-rigid @xmath0-module . because @xmath0 is tilted , lemma [ 4.4](b ) implies that pd@xmath290 , but also @xmath4 is rigid by lemma [ 4.6](d ) .
this shows that @xmath4 is a partial tilting @xmath0-module . on the other hand , by lemma [ 4.2](b ) , all nonzero indecomposable summands of @xmath4 have injective dimension 2 .
thus , pd@xmath291 , by proposition [ 2.2](b ) .
so , applying the auslander - reiten formula we have @xmath292 , where the last step follows from lemma [ 4.6](d ) .
this shows that @xmath4 is @xmath28-rigid .
now , it remains to shows that @xmath29 is a partial tilting and @xmath30-rigid @xmath7-module .
first we show that @xmath29 is partial tilting . from above we know that @xmath4 is a partial tilting @xmath0-module , thus by theorem [ 3.9 ] it suffices to show that hom@xmath293 and hom@xmath294 .
the second identity follows from the work above , so we need to show the first identity .
observe that by lemma [ 4.2](a ) , pd@xmath295 .
then proposition [ 2.2 ] implies that @xmath296 .
similarly , by lemma [ 4.2](b ) , id@xmath297 , so proposition [ 2.2 ] implies that @xmath298 .
however , by proposition [ 2.2](f ) , @xmath68 is closed under predecessors , which means @xmath299 . by definition of a torsion pair
there are no nonzero morphisms from @xmath71 to @xmath68 , so hom@xmath300 .
this shows that @xmath29 is a partial tilting @xmath7-module .
now we show that @xmath29 is @xmath30-rigid , that is hom@xmath301 .
first observe that lemma [ 3.8](a ) yields @xmath302 , which in turn by proposition [ 4.3](b ) is isomorphic to @xmath303 .
let @xmath304 .
then we have the following diagram , whose top row is the short exact sequence of proposition [ 3.6](a ) .
@xmath305&de\otimes e \ar[r]^{i_*}&de\otimes b \ar[r]^-{\pi_*}\ar[d]^{f}&de\ar[r]\ar@{.>}[dl]^g&0\\ & & \tau_c de&&}}.\ ] ] observe that @xmath306 , which is zero by our calculation above .
next , the universal property of coker@xmath307 implies that there exists @xmath308 such that @xmath309 .
however , we know that @xmath4 is @xmath215-rigid , which implies that @xmath310 .
this means @xmath311 .
thus , we conclude that @xmath29 is @xmath30-rigid . unlike @xmath4 ,
the @xmath0-module @xmath14 is not @xmath28-rigid , that is hom@xmath312 .
consider , for example , the tilted algebra @xmath0 of type @xmath313 given by the following quiver with relations .
@xmath314^{\alpha}\ar[dr]_{\beta}&&4 \ar@<-.5ex>[ll]_{\delta } \ar@<.5ex>[ll]^{\gamma } & & \delta \alpha = \gamma\beta = 0.\\&3&}}\ ] ] here @xmath315 and @xmath316 .
also , in this case @xmath317 , a direct sum of six simple projective @xmath0-modules .
it is easy to see that @xmath4 and @xmath29 are partial tilting and @xmath5-rigid in mod@xmath153 and mod@xmath115 respectively , because @xmath4 is a projective @xmath0-module .
let @xmath0 be the tilted algebra of type @xmath318 given by the following quiver with relations .
@xmath314^{\alpha}\ar[dr]_{\beta}&&4 \ar[ll]_{\delta } & & \delta \alpha = \delta\beta = 0.\\&3&}}\ ] ] observe that @xmath4 equals the indecomposable module @xmath319 . on the other hand
, @xmath14 equals @xmath320 , a direct sum of three isomorphic simple modules .
in this section we study the relationship between induction and coinduction functors and the cluster category . here
, we assume that @xmath10 is a hereditary algebra and @xmath204 mod@xmath34 is a basic tilting module .
let @xmath8 be the corresponding tilted algebra , and @xmath7 be the associated cluster - tilted algebra .
finally , let @xmath33 denote the cluster category of @xmath10 .
we know that mod@xmath34 naturally embeds in @xmath33 , which in turn maps surjectively onto mod@xmath115 via the functor @xmath11 .
recall that the induction functor @xmath2 and the coinduction functor @xmath3 both map mod@xmath153 to mod@xmath115 , while the module categories of @xmath10 and @xmath0 are closely related via the tilting theorem [ 2.1 ] .
now , we want to study how the induction and the coinduction functors fit into this larger picture .
@xmath321^{\text{tilting theorem}}\ar[d]&\text{mod}\,c\ar[d]^{\begin{smallmatrix } -\otimes _ c
b\\\\d(b\otimes_c d-)\end{smallmatrix}}\\ \mathcal{c}_a\ar[r]^{\text{hom}_{\mathcal{c}_a}(t,-)}&\text{mod}\,b}\ ] ] the following theorem describes the relationship between the induction functor and the cluster category . [ 9.1 ] let @xmath10 be a hereditary algebra and @xmath322 a basic tilting module .
let @xmath33 be the cluster category of @xmath10 , @xmath8 be the corresponding tilted algebra , and @xmath1 the corresponding cluster - tilted algebra .
recall the definitions of @xmath35 and @xmath68 , the associated torsion class of mod@xmath34 and the torsion free class of mod@xmath153 below .
@xmath323then the following diagram commutes .
@xmath324^{\text{hom}_a(t,-)}\ar[d]&\mathcal{y}(t)\ar[d]^{-\otimes _ c b}\\ \mathcal{c}_a\ar[r]^{\text{hom}_{\mathcal{c}_a}(t,-)}&\text{mod}\,b}\]]that is , @xmath325 for every @xmath326 .
let @xmath46 be an indecomposable module belonging to @xmath35 .
if @xmath327 , then @xmath328 is a projective @xmath0-module , and @xmath329 is the corresponding projective @xmath7-module . on the other hand
, @xmath330 is also the same projective @xmath7-module .
hence , in this case the theorem above holds . if @xmath331 , then proposition [ 2.2](d ) implies that the projective dimension of @xmath332 is one .
let @xmath333&p_c^1\ar[r]^{f}&p_c^0\ar[r]&\text{hom}_a(t , m)\ar[r]&0}\ ] ] be its minimal projective resolution in mod@xmath153 . by theorem [ 2.1](a ) , there exist @xmath334 , and @xmath335 such that @xmath336 moreover , since @xmath326 , we have that @xmath337 . applying @xmath67 to the projective resolution (
[ 12 ] ) and using theorem [ 2.1](a ) , we obtain an exact sequence @xmath338&t^1\ar[r]^{g}&t^0\ar[r]&m\ar[r]&0.}\ ] ] furthermore , @xmath73 by tilting theorem [ 2.1 ] , which means that the sequence above is a short exact sequence in mod@xmath34 .
observe that @xmath339 is also a morphism in the cluster category , and the short exact sequence above corresponds to a triangle @xmath340^g&t^0\ar[r]&m\ar[r]&t^1[1]}\ ] ] in @xmath33 .
applying @xmath11 to this triangle yields an exact sequence in mod@xmath115 @xmath341^{g_*}&\text{hom}_{\mathcal{c}_a}(t , t^0)\ar[r]&\text{hom}_{\mathcal{c}_a}(t , m)\ar[r]&0.}\ ] ] note that there is zero on the right of the sequence as @xmath342)=0 $ ] , because @xmath9 is a tilting object in @xmath33 .
also , @xmath343 and @xmath344 , thus @xmath345 on the other hand , applying the induction functor @xmath346 to the sequence ( [ 12 ] ) we obtain @xmath347^{f\otimes 1}&p^0_b\ar[r]&\text{hom}_a(t , m)\otimes_c b\ar[r]&0.}\ ] ] hence , we see that @xmath348 but since @xmath349 , we have @xmath350 because @xmath351 is projective both maps @xmath352 and @xmath353 are determined by their restrictions to @xmath354 . therefore , we construct a commutative diagram @xmath355^{\text{hom}_{\mathcal{c}_a}(t , g)}\ar@{=}[d]&p^0_b\ar[r]\ar[d]^{\cong}&\text{hom}_{\mathcal{c}_a}(t , m)\ar[r]\ar[d]^{h}&0\\ p^1_b\ar[r]^{\text{hom}_a(t , g)\otimes 1}&p^0_b\ar[r]&\text{hom}_a(t , m)\otimes_c b\ar[r]&0}\ ] ] where @xmath356 is an isomorphism by the five lemma .
this shows that for every @xmath357 we have @xmath358 , and finishes the proof of the theorem .
this result shows that the induction functor can be reformulated in terms of other well - studied functors .
in particular consider the following corollary .
recall the definitions of @xmath37 and @xmath71 , the associated torsion free class of mod@xmath34 and the torsion class of mod@xmath153 below .
@xmath359 [ 9.5 ] let @xmath10 be a hereditary algebra and @xmath322 a basic tilting module .
let @xmath33 be the cluster category of @xmath10 , @xmath8 be the corresponding tilted algebra , and @xmath1 the corresponding cluster - tilted algebra .
then @xmath360 for every @xmath197 .
first , note that every @xmath197 belongs to either @xmath71 or @xmath68 by proposition [ 2.2](e ) .
if @xmath361 , proposition [ 2.2](c ) implies that id@xmath89 , and then proposition [ 4.3](a ) yields @xmath362 .
if @xmath281 , then theorem [ 2.1](a ) implies that @xmath363 for @xmath364 .
hence @xmath365 , by theorem [ 9.1 ] .
similarly , there is a dual statement that provides an alternative description of the coinduction functor . [ 9.2 ] let @xmath10 be a hereditary algebra and @xmath322 a basic tilting module .
let @xmath33 be the cluster category of @xmath10 , @xmath8 be the corresponding tilted algebra , and @xmath1 the corresponding cluster - tilted algebra . then the following diagram commutes .
@xmath366^{\text{ext}^1_a(t,-)}\ar[d]&\mathcal{x}(t)\ar[d]^{d(b\otimes _ c d-)}\\ \mathcal{c}_a\ar[r]^{\text{ext}^1_{\mathcal{c}_a}(t,-)}&\text{mod}\,b}\]]that is , @xmath367 , for every @xmath368 .
let @xmath46 be an indecomposable module belonging to @xmath37 .
consider @xmath369 , which belongs to @xmath71 by theorem [ 2.1](b ) .
if @xmath370 is an injective @xmath0-module , then @xmath371 is the corresponding injective in mod@xmath115 .
on the other hand , by proposition [ 2.4 ] we have @xmath372 for some @xmath373 .
therefore , by serre duality @xmath374 this shows that the theorem holds if @xmath369 is injective . if @xmath375 is not injective then according to proposition [ 2.2](c ) it has injective dimension one . consider a minimal injective resolution of this module in mod@xmath153 below .
@xmath376&\text{ext}_a^1(t , m)\ar[r]&i^0_c\ar[r]^f&i^1_c\ar[r]&0}\ ] ] because @xmath377 are successors of @xmath369 , proposition [ 2.2](f ) implies that these injectives belong to @xmath71 . by proposition
[ 2.4 ] and theorem [ 2.1](b ) , there exist @xmath378 , and @xmath379 such that @xmath380 applying @xmath381 to sequence ( [ 13 ] ) and using theorem [ 2.1 ] , we obtain a long exact sequence @xmath382&m\ar[r]&\tau t^0\ar[r]^g&\tau t^1\ar[r ] & \text{ext}_a^1(t , m)\otimes_c t.}\ ] ] observe that by the same theorem the last term in the sequence above is zero , likewise @xmath383 , because @xmath9 as a left @xmath0-module has projective dimension at most one . thus we obtain a short exact sequence in mod@xmath34 @xmath181&m\ar[r]&\tau t^0\ar[r]^g&\tau t^1 \ar[r]&0}\ ] ] which induces a triangle @xmath384\ar[r]&m\ar[r]&\tau t^0 \ar[r]^g&\tau t^1}\ ] ] in the cluster category .
applying @xmath385 to this triangle we obtain an exact sequence in mod@xmath115 @xmath386)\ar[r]&\text{ext}^1_{\mathcal{c}_a}(t , m)\ar[r]&\text{ext}^1_{\mathcal{c}_a}(t,\tau t^0)\ar[r]^{g_*}&\text{ext}^1_{\mathcal{c}_a}(t,\tau t^1)}.\ ] ] observe that because @xmath9 is a tilting object in @xmath33 , we have @xmath387)=0 $ ] .
moreover , equation ( [ 14 ] ) yields @xmath388 and @xmath389 .
thus , @xmath390 on the other hand , applying the coinduction functor to sequence ( [ 13 ] ) we obtain @xmath391&d(b\otimes _ c d \text{ext}^1_a(t , m))\ar[r]&i^0_b\ar[r]^{d(1\otimes d g)}&i^1_b.}\ ] ] hence , we see that @xmath392 because @xmath393 is injective , both maps @xmath394 and @xmath395 are determined by their preimages of @xmath396 .
but since @xmath397 , we have @xmath398 thus we have a commutative diagram @xmath399&\text{ext}^1_{\mathcal{c}_a}(t , m)\ar[r]\ar[d]^h&i^0_b\ar[d]^{\cong}\ar[r]^{\text{ext}^1_{\mathcal{c}_a}(t , g)}&i^1_b\ar@{=}[d]\\ 0\ar[r]&d(b\otimes_c d\text{ext}^1_a(t , m))\ar[r]&i^0_b\ar[r]^{d(b\otimes_c d\text{ext}^1_a(t , g))}&i^1_b}\ ] ] and @xmath356 is an isomorphism by the five lemma .
this shows that for every @xmath368 we have @xmath400 , and finishes the proof of the theorem .
let @xmath10 be a hereditary algebra and @xmath322 a basic tilting module .
let @xmath33 be the cluster category of @xmath10 , @xmath8 be the corresponding tilted algebra , and @xmath1 the corresponding cluster - tilted algebra .
then @xmath401 for every @xmath402 .
the proof is similar to that of corollary 6.2 and we omit it .
let @xmath10 be the path algebra of type @xmath403 given by the following quiver .
@xmath404\\1&2\ar[l]&3\ar[l]&4\ar[l]\ar[dr]\\&&&&6}\ ] ] let @xmath405 be the tilting @xmath10-module .
the auslander - reiten quiver of mod@xmath34 is shown at the top of figure [ figure 1 ] .
the modules belonging to @xmath35 are in the dark shaded regions enclosed by a solid line , while modules belonging to @xmath37 are in the light shaded regions enclosed by a dotted line . then the corresponding tilted algebra @xmath8 is of type @xmath403 , and it is given by the following quiver with relations .
@xmath406_{\beta}&4\ar[l]_{\alpha}&5\ar[l]_{\gamma}&6\ar[l]_{\delta}\ar[dl]^{\epsilon}&&\alpha\beta=0\\&&&3\ar[ull]^{\sigma}&&&\delta\gamma\alpha=\epsilon\sigma.}\ ] ] the auslander - reiten quiver of mod@xmath153 is shown in the middle of figure [ figure 1 ] .
the modules belonging to @xmath68 are in the dark shaded regions enclosed by a solid line , while the modules belonging to @xmath71 are in the light shaded regions enclosed by a dotted line .
the cluster - tilted algebra @xmath1 is represented by the following quiver with relations .
@xmath407^{\omega}&2\ar[l]^{\beta}\ar@/^-15pt/[rrr]_{\;\;\;\rho}&4\ar[l]_{\alpha}&5\ar[l]_{\gamma}&6\ar[l]_{\delta}\ar[dl]^{\epsilon}&&{\begin{matrix}\alpha\beta=\omega\alpha=0 & \gamma\alpha\rho=\alpha\rho\delta=0&\sigma\rho=\rho\epsilon=0\end{matrix}}\\&&&3\ar[ull]^{\sigma}&&&{\begin{matrix}\beta\omega=\rho\delta\gamma & \delta\gamma\alpha=\epsilon\sigma.\end{matrix}}}\ ] ] the auslander - reiten quiver of mod@xmath115 is shown at the bottom of figure [ figure 1 ] , where we identify the modules which have the same labels .
the modules in the dark shaded regions enclosed by a solid line correspond to the set @xmath408 , while those in the light shaded regions enclosed by a dotted line correspond to the set @xmath409 .
note that there are four modules that lie in both sets .
as stated in theorems [ 9.1 ] and [ 9.2 ] the shape and relative position of @xmath35 and @xmath37 in mod@xmath34 correspond exactly to the set of induced @xmath68-modules and the set of coinduced @xmath71-modules respectively . moreover ,
if we apply @xmath410 to the set of modules @xmath409 , then their position relative to the induced @xmath68 modules will correspond exactly to the position of @xmath37 relative to @xmath35 in mod@xmath34 .
if we want to study the module category of a cluster - tilted algebra via the induction and coinduction functors , then it is natural to ask which modules in a cluster - tilted algebra are actually induced or coinduced from the modules over some tilted algebra .
recall from theorem [ 2.8 ] that , given a cluster - tilted algebra @xmath7 , every local slice @xmath117 gives rise to a tilted algebra @xmath411 , whose relation extension is @xmath7 . 1 .
@xmath46 is said to be _ induced from some tilted algebra _ if there exists a tilted algebra @xmath0 and a @xmath0-module @xmath80 such that @xmath7 is the relation extension of @xmath0 and @xmath412 .
@xmath46 is said to be _ coinduced from some tilted algebra _ if there exists a tilted algebra @xmath0 and a @xmath0-module @xmath80 such that @xmath7 is the relation extension of @xmath0 and @xmath413 .
[ 7.0 ] if @xmath7 is a cluster - tilted algebra of finite representation type , then for every indecomposable @xmath7-module @xmath46 there exists a tilted algebra @xmath0 and a @xmath0-module @xmath156 such that @xmath46 is induced and coinduced from @xmath156 .
if @xmath7 is of finite representation type and @xmath414 then by theorem [ 2.9 ] there exists a local slice @xmath415 containing @xmath46 .
theorem [ 2.8 ] implies that there is a tilted algebra @xmath0 such that @xmath117 is a slice in @xmath416 and @xmath7 is the relation extension of @xmath0 .
hence , @xmath46 is an indecomposable @xmath0-module lying on a slice , which means id@xmath89 and pd@xmath224 .
it follows from proposition [ 4.3 ] that @xmath417 .
therefore , @xmath414 is both induced and coinduced from the same module @xmath197 , for some tilted algebra @xmath0 .
suppose @xmath7 is a cluster - tilted algebra and @xmath46 is an indecomposable @xmath7-module .
if @xmath46 is also an indecomposable @xmath0-module , for some tilted algebra @xmath0 , whose relation extension is @xmath7 , then @xmath46 is induced or coinduced .
observe that if @xmath46 is a @xmath0-module , then proposition [ 2.2](b ) implies that id@xmath418 or pd@xmath89 .
in the first case @xmath419 and in the second one @xmath420 by proposition [ 4.3 ] .
hence , in either case @xmath46 is induced or coinduced from a @xmath0-module .
[ 7.1 ] let @xmath7 be a cluster - tilted algebra . then for every transjective indecomposable @xmath7-module @xmath46 , there exists a tilted algebra @xmath0 , such that @xmath7 is the relation extension of @xmath0 , and @xmath46 is an indecomposable @xmath0-module . in particular , every transjective @xmath7-module is induced or coinduced from @xmath0 .
let @xmath10 be a hereditary algebra and @xmath422 a cluster - tilting object such that @xmath423 .
let @xmath46 be an indecomposable @xmath7-module lying in the transjective component @xmath424 of @xmath144 , and let @xmath425 be an indecomposable object such that @xmath426 .
finally , let @xmath427 be the full subquiver of the auslander - reiten of @xmath33 defined in section [ sect 2.4 ] .
since @xmath428 , for all @xmath429 , we may assume without loss of generality that @xmath430 lies in the preprojective component of @xmath59 .
furthermore , we may assume that every preprojective successor of @xmath430 in @xmath59 is sincere .
indeed , this follows from the fact that there are only finitely many ( isoclasses of ) indecomposable preprojective @xmath10-modules that are not sincere . for tame algebras this holds , because non - sincere modules are supported on a dynkin quiver , and for wild algebras see @xcite .
now since @xmath431 is a sincere @xmath10-module , proposition [ 2.6 ] implies that @xmath145 is a slice in @xmath59 , and therefore @xmath432 is a local slice in @xmath33 .
let @xmath433 .
then @xmath434 , and thus , if @xmath435 is a local slice in @xmath19 , we are done .
suppose to the contrary that @xmath435 is not a local slice .
then lemma [ lemls ] implies that @xmath430 contains an indecomposable summand @xmath436 of @xmath147 .
let @xmath437 in @xmath33 .
then @xmath438 contains @xmath439 .
since @xmath440 is sincere , it again follows from proposition [ 2.6 ] that @xmath438 is a local slice in @xmath33
. moreover @xmath438 can not contain any summands of @xmath147 , because if it did , there would be a sectional path from @xmath440 to a summand of @xmath441 , hence @xmath442 , which is impossible , since @xmath9 is a cluster - tilting object . therefore lemma [ lemls ] implies that @xmath443 is a local slice in @xmath19 containing @xmath444 .
following @xcite , we say that a cluster - tilted algebra @xmath7 is _ cluster - concealed _ if @xmath445 where @xmath9 is obtained from a preprojective @xmath10-module .
this means that all projective @xmath7-modules lie in the transjective component of @xmath144 . in this case
, we show that the theorem above holds not only for the transjective modules but for all @xmath7-modules .
[ 7.2 ] let @xmath7 be a cluster - concealed algebra . then for every indecomposable @xmath7-module @xmath46 there exists a tilted algebra @xmath0 whose relation extension is @xmath7 , such that @xmath46 is an indecomposable @xmath0-module . moreover , for all non - transjective modules one can take the same @xmath0 . in particular , every @xmath7-module is induced or coinduced from some tilted algebra . because @xmath7 is cluster - concealed there exists a hereditary algebra @xmath10 , and a preprojective tilting @xmath10-module @xmath9 such that @xmath446 .
observe that @xmath10 is of infinite representation type .
let @xmath447 be the corresponding tilted algebra .
then @xmath7 is the relation extension of @xmath0 , thus @xmath448 , where @xmath15 .
let @xmath449 be the set of all regular @xmath10-modules .
it is nonempty because @xmath10 is of infinite representation type .
it follows from theorem [ 2.3](a ) that @xmath450 . according to theorem [ 2.3](b )
the set of regular @xmath0-modules @xmath451 is obtained from @xmath449 by applying the functor @xmath79 .
also , because of theorem [ 2.5 ] , the set of regular @xmath7-modules @xmath452 is obtained from @xmath449 by applying @xmath11 .
hence , given @xmath453 there exists @xmath454 such that @xmath455 . because @xmath450 , theorem [ 9.1 ] implies that @xmath456
hence , @xmath46 is induced from the indecomposable @xmath0-module @xmath457 , and now it remains to show that @xmath46 is actually an indecomposable @xmath0-module itself , that is @xmath458 .
it follows from theorem [ 2.3](c ) that id@xmath459 , so proposition [ 4.3](a ) implies that @xmath460 this shows that @xmath46 is an indecomposable @xmath0-module .
[ 7.3 ] let @xmath7 be a tame cluster - tilted algebra .
let @xmath461 be a tube in mod@xmath115 and let @xmath462 denote all distinct indecomposable projective @xmath7-modules belonging to @xmath461 . if @xmath463then every module in @xmath461 is induced or coinduced from the same tilted algebra @xmath0 .
let @xmath10 be a hereditary algebra and @xmath9 a tilting @xmath10-module such that @xmath464 .
thus @xmath10 is tame , and we can suppose without loss of generality that @xmath9 has no preinjective summands .
the regular components of @xmath59 form a family of pairwise orthogonal tubes .
let @xmath465 be the tube in @xmath59 whose image under @xmath11 is the tube @xmath461 in the statement of the theorem .
let @xmath466 denote the rank of the tube @xmath465 .
let @xmath467 denote the connected component in @xmath18 given as the image of @xmath465 under the functor @xmath79 .
thus @xmath467 lies inside @xmath68 , which implies that each module in @xmath467 has projective dimension at most one , by proposition [ 2.2 ] , and therefore the coinduction functor is the identity on @xmath467 , by proposition [ 4.3 ] .
let @xmath469 be the indecomposable summands of @xmath9 that lie inside @xmath465 , such that @xmath470 , for @xmath471 , are the indecomposable projective modules in @xmath461 . because of our assumption @xmath472 , if @xmath473 , we have that each @xmath440 lies on the mouth of @xmath465 , see for example ( * ? ? ?
* proposition xii.2.1 ) .
@xmath474&&\tau_a^2 t_i\ar[dr]&&\tau_a t_i\ar[rd ] & & t_i\ar[dr]&&\cdots\\ & \cdots \ar[ur]\ar[dr]&&\bullet\ar[ru ] \ar[dr]&&\bullet\ar[ur]\ar[dr]&&\cdots\ar[ru]\\ & & \cdots \ar[ur]&&\bullet \ar[ur]&&\cdots\ar[ru]}\ ] ] @xmath475&&i_b(i)\ar[dr]&&&&p_b(i)\ar[dr]&&\cdots\\ & \cdots \ar[ur]\ar[dr]&&\tau_b\,\text{rad}\,p_b(i ) \ar[dr]&&\text{rad}\,p_b(i)\ar[ur]\ar[dr]&&\cdots\ar[ru]\\ & & \cdots \ar[ur]&&\bullet \ar[ur]&&\cdots\ar[ru]}\ ] ] since @xmath476 , and @xmath477 lies on the mouth of @xmath465 , we see that @xmath478 consists of the @xmath479 corays not ending in one of the @xmath477 , @xmath480 .
therefore @xmath481 has @xmath482 corays , because of our commutative diagram .
so these @xmath479 corays are in the image of the induction functor @xmath2 .
moreover the @xmath483 corays in @xmath461 that are not in the image of the induction functor are precisely those ending in @xmath484 , @xmath485 we will show that these corays are equal to the corays in @xmath467 ending in @xmath486 , which implies that these corays are coinduced . since @xmath487 we have @xmath488 by our assumption on @xmath9 , we have @xmath489 and thus @xmath490 is simple .
moreover , by proposition [ 3.6](b ) , there is a short exact sequence in @xmath19 of the form @xmath181&i_c(i)\ar[r]&i_b(i)\ar[r]&\tau_c\omega_ci_c(i)\ar[r]&0,}\ ] ] and thus finally , let @xmath46 be any indecomposable @xmath0-module on the coray in @xmath467 ending in @xmath495 , but @xmath496 .
then @xmath184 lies on the coray in @xmath461 ending in @xmath497 .
lemma [ 3.8 ] implies that @xmath498 where the last identity holds because the coinduction functor is the identity on the tube @xmath467 .
this shows that the @xmath30-translate of the coray in @xmath461 ending in @xmath497 is equal to the coray in @xmath467 ending in @xmath499 .
in particular the modules on this coray are coinduced from @xmath0 .
[ ex 7.7 ] let @xmath7 be a cluster - tilted algebra of type @xmath500 given by the following quiver with relations .
@xmath501\ar@<-.5ex>[rr]_{\beta}&&5\ar[dl]^{\epsilon}&&\alpha\beta=\beta\epsilon=\epsilon\alpha=0\\&3\ar[ul]^{\alpha}\ar[dr]^{\delta}&&&\gamma\delta=\delta\sigma=\sigma\gamma=0.\\2\ar[ur]^{\gamma}&&4\ar[ll]_{\sigma}}\ ] ] then there are exactly two tilted algebras @xmath0 and @xmath502 , both of infinite representation type , whose relation extension is @xmath7 .
@xmath0 and @xmath502 can be represented by the following quivers with relations .
@xmath503\ar@<-.5ex>[rr]_{\beta}&&5&\alpha\beta=0 & & & 1\ar@<.5ex>[rr]\ar@<-.5ex>[rr]_{\beta}&&5\ar[dl]^{\epsilon}&\beta\epsilon=0\\ c : & & 3\ar[dr]^{\delta}\ar[ul]^{\alpha}&&\gamma\delta=0.&&c ' : & & 3\ar[dr]^{\delta}&&\gamma\delta=0.\\ & 2\ar[ur]^{\gamma}&&4 & & & & 2\ar[ur]^{\gamma}&&4}\ ] ] there is a tube in @xmath144 of rank four containing three distinct projective modules .
we describe it below and identify the modules that have the same label .
we will show that the four modules emphasized in bold are not induced and not coinduced .
@xmath504 & & & & { \begin{smallmatrix}2\\3\\1\\5\\3\\4\end{smallmatrix}}\ar[dr ] & & & & { \begin{smallmatrix}4\\2\end{smallmatrix}}\\ & { \begin{smallmatrix}4\end{smallmatrix}}\ar[dr ] & & { \begin{smallmatrix}3\\1\\5\\3\\4\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}2\\3\\1\\5\\3\end{smallmatrix}}\ar[dr ] & & { \begin{smallmatrix}2\end{smallmatrix}}\ar[ur]&&\\ & & { \begin{smallmatrix}3\\4\;1\\\;\;\;5\\\;\;\;3\\\;\;\;4\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}\bf{3}\\\bf{1}\\\bf{5}\\\bf{3}\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}2\;\;\;\\3\;\;\;\\1\;\;\;\\5\;2\\3\end{smallmatrix}}\ar[dr]\ar[ur]\\ & { \begin{smallmatrix}1\\5\\3\\4\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}\bf{3}\\\bf{4}\;\bf{1}\\\;\;\;\bf{5}\\\;\;\;\bf{3}\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}\bf{3}\;\;\;\\\bf{1}\;\;\;\\\bf{5}\;\bf{2}\\\bf{3}\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}2\\3\\1\\5\end{smallmatrix}}\ar[dr]&&\\ { \begin{smallmatrix}2\;\;\;\\3\;\;\;\\1\;1\\\;5\;5\\\;\;\;\;3\\\;\;\;\;4\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}1\\5\\3\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}\bf{3}\;\;\;\;\\\bf{4}\;\bf{1}\;\;\\\;\;\;\;\bf{5}\;\bf{2}\\\;\;\;\;\;\bf{3}\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}3\\1\\5\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}2\;\;\;\\3\;\;\;\\1\;1\\\;5\;5\\\;\;\;\;3\\\;\;\;\;4\end{smallmatrix}}\\ & { \begin{smallmatrix}2\;\;\;\\3\;\;\;\\1\;1\\\;5\;5\\\;\;\;\;3\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}1\;\;\;\\5\;2\\3\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}3\\4\;1\\\;\;\;5\end{smallmatrix}}\ar[dr]\ar[ur ] & & { \begin{smallmatrix}3\;\;\;\\1\;1\\\;5\;5\\\;\;\;3\\\;\;\;4\end{smallmatrix}}\ar[dr]\ar[ur]\\ \dots\ar[ur]&&\dots\ar[ur]&&\dots\ar[ur ] & & \dots\ar[ur ] & & \dots}\ ] ] the corresponding skew tube in @xmath416 is given below .
observe that it consists of one coray ending in the projective at vertex 2 , labeled @xmath505 , and four rays starting in @xmath506,@xmath507,@xmath508 and @xmath509 .
let @xmath510 and @xmath511 , then we make the following computation .
@xmath512 @xmath513&&\dots \\ & & & & { \begin{smallmatrix}2\\3\\1\\5\end{smallmatrix}}\ar[dr ] & & { \begin{smallmatrix}1\\5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}\;3\;\;\;\;\\4\;1\;1\\\;\;\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr]\\ & { \begin{smallmatrix}4\end{smallmatrix}}\ar[dr ] & & { \begin{smallmatrix}3\\1\\5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}2\;\;\;\;\\3\;\;\;\;\\\;1\;1\\\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}1\;1\\\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & \dots\\ & & { \begin{smallmatrix}3\\4\;1\\\;\;\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}3\;\;\;\\1\;1\\\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}2\;\;\;\;\;\;\\3\;\;\;\;\;\;\\\;1\;1\;1\\\;\;\;5\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & \dots\ar[ur]\\ & { \begin{smallmatrix}1\\5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}\;3\;\;\;\;\\4\;1\;1\\\;\;\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}3\;\;\;\;\;\;\\1\;1\;1\\\;\;5\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & \dots\ar[ur]\\ { \begin{smallmatrix}2\;\;\;\;\\3\;\;\;\;\\\;1\;1\\\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}1\;1\\\;\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & { \begin{smallmatrix}\;\;3\;\;\;\;\;\;\\4\;1\;1\;1\\\;\;\;\;\;5\;5\;5\end{smallmatrix}}\ar[ur]\ar[dr ] & & \dots\ar[ur]\\ & \dots\ar[ur ] & & \dots\ar[ur ] & & \dots\ar[ur ] } \ ] ] let @xmath514 denote the coray in @xmath416 ending in the projective at vertex 2 . observe that if we induce the modules in @xmath515 , that is apply @xmath516 , then we obtain a coray in the tube of @xmath144 that ends in the projective @xmath7-module at vertex 2 .
recall that there is an isomorphism of @xmath0-modules @xmath517 for every @xmath518 . on the other hand coinduction @xmath17 of modules in @xmath515 will act as the identity map , because there is a @xmath0-module isomorphism @xmath519 and the last summand is zero for all @xmath520 .
therefore , we see the coray @xmath514 appearing in @xmath144 . finally , the only module in the tube above that does not belong to @xmath514 is the simple at 4 .
observe , that this module is also a projective @xmath0-module , so inducing it we obtain the corresponding projective @xmath7-module @xmath521 .
similarly , in @xmath522 there is a skew tube consisting of four corays and one ray starting in @xmath523 , the injective @xmath502-module at vertex 4 .
denote this ray by @xmath524 .
analogous calculations yield that the ray in @xmath144 starting in @xmath525 the injective @xmath7-module , is coinduced from @xmath524 . but the induction of @xmath524 acts as the identity map , so we see this exact same ray appearing in @xmath144 . observe that every module on this ray is indeed an indecomposable @xmath502-module .
finally , the simple injective @xmath502-module at vertex 2 , belongs to this skew tube , and its induction is again the simple module supported at vertex 2 . in particular we observe that the @xmath7-modules which make up the mesh emphasized in bold are not induced and not coinduced from any tilted algebra .
moreover , as one goes down the tube in @xmath144 there will appear infinitely many such meshes , where no module is induced or coinduced .
simsko3 t. adachi , o. iyama and i. reiten , @xmath5-tilting theory , _ compos .
_ * 150 * ( 2014 ) , no .
3 , 415452 . c. amiot ,
cluster categories for algebras of global dimension 2 and quivers with potential , _ ann .
. fourier _ * 59 * no 6 , ( 2009 ) , 25252590 . , cluster - tilted algebras as trivial extensions , _ bull .
_ * 40 * ( 2008 ) , 151162 .
, cluster - tilted algebras and slices , _ j. of algebra _ * 319 * ( 2008 ) , 34643479 .
, on the galois covering of a cluster - tilted algebra , _ j. pure appl .
* 213 * ( 7 ) ( 2009 ) 14501463 .
, cluster - tilted algebras without clusters , _ j. algebra _ * 324 * , ( 2010 ) , 24752502
. i. assem , m. j. redondo and r. schiffler , on the first hochschild cohomology group of a cluster - tilted algebra , _ algebr .
theory _ * 18 * ( 2015 ) , no .
6 , 15471576 .
i. assem , d. simson and a. skowronski , _ elements of the representation theory of associative algebras , 1 : techniques of representation theory _ , london mathematical society student texts 65 , cambridge university press , 2006 .
i. assem and d. zacharia , full embeddings of almost split sequences over split - by - nilpotent extensions , _ coll .
_ * 81 * , ( 1 ) ( 1999 ) 2131 . , from iterated tilted to cluster - tilted algebras , _ adv .
_ * 223 * ( 2010 ) , no .
4 , 14681494 .
l. beaudet , t. brstle and g. todorov , projective dimension of modules over cluster - tilted algebras , _ algebr .
represent . theory _
* 17 * ( 2014 ) , no . 6 , 17971807 .
m. a. bertani - kland , s. oppermann and a wrlsen , constructing tilted algebras from cluster - tilted algebras , _
j. algebra _ * 323 * ( 2010 ) , no . 9 , 24082428 .
s. brenner and m. butler , generalizations of the bernstein - gelfand - ponomarev reflection functors .
representation theory , ii ( proc .
second internat .
carleton univ . ,
ottawa , ont . , 1979 ) , pp .
103169 , lecture notes in math . ,
832 , springer , berlin - new york , 1980 .
a. b. buan , r. marsh , m. reineke , i. reiten and g. todorov , _ tilting theory and cluster combinatorics _ , adv . math . * 204 * ( 2006 ) , no . 2 , 572 - 618 . , cluster - tilted algebras , _ trans .
_ * 359 * ( 2007 ) , no .
1 , 323332 ( electronic ) . , quivers with relations and cluster tilted algebras , _ algebr . and represent . theory _ * 9 * , ( 2006 ) , no .
4 , 359376 . , from triangulated categories to cluster algebras , _ invent .
_ * 172 * ( 2008 ) , 169211 .
l. david - roesler and r. schiffler , algebras from surfaces without punctures . _ j. algebra _ * 350 * ( 2012 ) , 218244 .
h. derksen , j. weyman and a. zelevinsky , quivers with potentials and their representations . i. mutations .
_ selecta math .
_ ( n.s . ) * 14 * ( 2008 ) , no .
1 , 59119 . s. fomin and a. zelevinsky , cluster algebras i : foundations , _
* 15 ( 2002 ) , 497529 . *
, on triangulated orbit categories , _ documenta math . _ * 10 * ( 2005 ) , 551581 .
b. keller and i. reiten , cluster - tilted algebras are gorenstein and stably calabi - yau , _ adv .
_ * 211 * ( 2007 ) , no .
1 , 123151 . o. kerner , more representations of wild quivers .
igusa , kiyoshi , a. martsinkovsky , and g. todorov , _ expository lectures on representation theory : _ maurice auslander distinguished lectures and international conference , april 25 - 30 , 2012 , woods hole oceanographic institute , quisset campus , falmouth , ma . providence : american mathematical society , 2014 . 3556 .
s. ladkani , hochschild cohomology of the cluster - tilted algebras of finite representation type , preprint , arxiv:1205.0799 .
plamondon , cluster algebras via cluster categories with infinite - dimensional morphism spaces .
math . _ * 147 * ( 2011 ) , no . 6 , 19211954 .
ringel , _ tame algebras and integral quadratic forms _ , lecture notes in math .
1099 , springer - verlag , 1984 .
ringel , cluster - concealed algebras . _
_ * 226 * ( 2011 ) , no .
2 , 15131537 .
r. schiffler , _ quiver representations _ , cms books in mathematics , springer international publishing , 2014 .
r. schiffler and k. serhiyenko , injective presentations of induced modules over cluster - tilted algebras , preprint .
d. simson and a. skowroski , _ elements of the representation theory of associative algebras 3 : representation - infinite tilted algebras _ london mathematical society student text 72 , cambridge university press , 2007 .
|
we propose a new approach to study the relation between the module categories of a tilted algebra @xmath0 and the corresponding cluster - tilted algebra @xmath1 .
this new approach consists of using the induction functor @xmath2 as well as the coinduction functor @xmath3 .
we show that @xmath4 is a partial tilting and a @xmath5-rigid @xmath0-module and that the induced module @xmath6 is a partial tilting and a @xmath5-rigid @xmath7-module . furthermore ,
if @xmath8 for a tilting module @xmath9 over a hereditary algebra @xmath10 , we compare the induction and coinduction functors to the buan - marsh - reiten functor @xmath11 from the cluster - category of @xmath10 to the module category of @xmath7 .
we also study the question which @xmath7-modules are actually induced or coinduced from a module over a tilted algebra .
|
many countries throughout the developed world are currently in the long - term process of changing their electricity supply system from one based on fossil and nuclear fuels to one based on renewable energy sources ( res ) .
multiple studies have shown that an electricity supply system based entirely on res is possible not only in large - interconnected systems like europe ( e.g. @xcite ) and the united states ( e.g. @xcite ) , but also in isolated national systems ( e.g. @xcite ) . in many studies dealing with a prospective european power supply systems based on res , a large share of the electricity
will be provided by variable renewable energy sources ( vres ) , which are in particular solar and wind resources @xcite .
a power supply system based mainly on vres requires measures to cope with the natural variability of the power production from these resources in order to ensure our current high security of supply .
an option widely considered is the utilisation of storage devices . here
, a combination of small , highly efficient storage capacities for balancing fluctuations in the order of a day or less and large , less efficient seasonal storage capacities for balancing long - term fluctuations appears suitable @xcite .
it is still an open question though , which storage dimensions storage capacity and power requirements are most suitable for integrating a very high share of vres into a power supply system .
two large groups of approaches exist to address this question : many studies rely on cost estimates to come up with possible good combinations of storage and generation capacities ( e.g. @xcite ) . due to the often relatively high number of parameters and
generally well - founded
assumptions , the sensitivity with respect to individual parameters and even more to combinations thereof is often difficult to determine .
a different approach is to be as general as possible by considering an abstract view on the system and focusing on the meteorological aspect of future power supply systems , since this aspect will play an important role in future power supply systems based on res @xcite . using mainly long - term res power production time - series and corresponding load data ,
these works have investigated e.g. the large - scale need for the installation of storage capacities in a 100%-res scenario for europe and how these needs can be reduced using excess generation as well as transmission capacities @xcite .
besides the energy capacity of the storage devices , their power requirements are also an important dimension to be considered for future power systems based on ( v)res .
weiss et al . investigated the storage requirements for the german power supply system in which 80% of the total electricity consumption would be provided by a combination of solar , wind and hydropower resources ; and it was found that in this case 31 gw of discharging power and slightly less for charging power would be required for the long - term storage devices when combined with the current pumped - hydro capacity of 8 gw as short - term storage @xcite . like the need for the energy capacity of the storage devices rises with increasing share of res penetration , so does the need for storage power .
using a holistic model with a hierarchical management approach , bussar et al .
recently found that a renewable power supply system for the eu - mena region ( europe , middle east and north africa ) would be most economic with a combination of storage devices having a total discharging power on the same order as the peak load , whereupon long - term storage was also found to need higher charging power than discharging power @xcite .
furthermore , the discharge power of the long - term storage was found to be higher than the combined discharge power of both short - term and medium - term storage @xcite , which stresses the importance of long - term storage in future power systems based on ( v)res . in this work
, we present an optimisation approach to systematically study power systems with multiple storage systems , well defined by their energy content , charging and discharging power as well as their efficiency . here
, a special focus is set on the power requirements of the storage systems .
the modelling approach is an advancement of our previous modelling approaches ( cf .
@xcite ) and is intended to provide deeper insights into urgent questions in the design of prospective power systems based on energy generated from fluctuating renewable sources .
to investigate combinations of energy storages for the integration of high shares of vres into power systems , we suggest a novel modelling approach using an economic optimisation procedure . for our approach
we build on previous contributions on the integration of vres in germany and europe @xcite .
while these contributions were restricted to meteorological aspects of the integration and were by construction limited to the investigation of the effects of one single storage class , the optimisation approach developed here will allow to study economic aspects and power systems with multiple storages in a yet rather straightforward and instructive manner .
we start from time series data of the residual load for a certain area of interest , @xmath0 , sampled at time lag @xmath1 , i.e. @xmath2 with @xmath3 .
@xmath0 can be obtained by subtracting the power generation of both must - run units and variable , non - dispatchable renewable energy sources from the load , details will be described later in section [ ssec : resources ] . if @xmath4 there is a surplus of energy which could be curtailed , whereas at times @xmath5 with @xmath6 there is a lack of energy which needs to be compensated . for balancing of the loads in particular in the second case
it is aimed to use either backup power plants or storages , with the latter being charged in times of surplus of energy .
a storage @xmath7 is characterised by its charging ( @xmath8 ) and discharging ( @xmath9 ) power , its round - trip efficiency is split equally between charging and discharging in this work .
individual values for charging efficiency ( @xmath10 ) and discharging efficiency ( @xmath11 ) can easily be implemented but would only influence the interpretation of the value of the storage capacity @xmath12 .
the storage capacity @xmath12 corresponds to the energy stored , e.g. in the form of chemical energy , while @xmath10 and @xmath11 determine the flow of electrical energy from / to the overall system ( cf .
( [ eq : storagelevel ] ) ) . ]
@xmath13 , and its maximum capacity in terms of energy @xmath12 .
we assume that each of these parameters implies costs , which are characterised by their respective equivalent annual installation and operating costs per unit , @xmath14 , @xmath15 , and @xmath16 . if we for the time being neglect grid limitations and grid losses , storage units with the same technology ( i.e. in particular with the same efficiencies for charging and discharging ) can be combined to a single storage , which significantly reduces the amount of different storages which need to be considered .
the approach we would like to propose can be characterised as an economic optimisation approach to identify which combination of storages and backup power would be most beneficial from an economic perspective ( i.e. cheapest ) to guarantee a non - positive residual load at each moment in time .
the costs for installing and operating the generation capacities are explicitly not included in the optimisation approach , since these costs depend on the scenario considered and be can calculated and added separately ( c.f . section [ ssec : resources ] ) .
other costs of prospective power systems which are not included in our approach ( e.g. transmission costs , social costs ) need to be investigated and compared in more detailed analyses in future work .
the linear optimisation problem with the corresponding cost function @xmath17 reflecting the annual costs for storage and backup is defined as : @xmath18 here , the generally not unique
time series @xmath19 , @xmath20 and @xmath21 determine the charge and discharge of the respective storage and the backup power required ( i.e. the operation of storages and backup ) . for the backup
, @xmath21 describes the actual time series of backup usage , @xmath22 is the length of the time series in multiples of years , and @xmath23 are the costs for backup per energy unit .
the backup energy is modelled in this rather simple way to be able to include a wide variety of backup options , ranging from biogas power plants with high usage rates to peak power plants with low number of full - load hours as well as load shedding options , in a rather straightforward manner .
a sufficiently high value for the backup costs @xmath23 implies that backup energy is not primarily used in comparison to wind and solar energy .
the real costs for some of the technologies considered as backup power might be cheaper if they achieve a high number of full - load hours .
however , the latter is not known until the optimisation process is finished . for most of the investigated scenarios in this work ,
the maximum backup power @xmath24 is not limited so as to include the wide variety of backup options mentioned above
. a closer look at the backup power plants is taken in section [ ssec : minimal_backup ] .
the constraints guarantee that the parameters and the storage level @xmath25 remain within the boundaries associated with the minimum required installation costs , namely ( 1 ) meeting the residual load at all times by ( net ) discharging the storages and utilisation of backup power , the ( 2 ) charging and ( 3 ) discharging power must not exceed the charging and discharging power limit of the respective storage , ( 4 ) the storage level for each storage is always between zero ( empty ) and the size of the storage @xmath12 , ( 5 ) the storage levels at the end of the simulation must not be lower than at the beginning , and ( 6 ) backup power plants can only provide positive backup power .
the structure of is related to a linear minimisation problem also referred to as linear program . the main difference to a deterministic linear optimisation problem is the fact that the upper boundaries of the constraints are not fixed during the optimisation process . to solve with flexible constraints , standard solver for linear programs ( e.g. simplex or interior point methods )
can be applied on piecewise equivalent subproblems @xcite .
additionally , the approach allows integration of other power plants by integrating them as a ( virtual ) storage .
the storage size @xmath12 corresponds to the highest possible energy production in the period of investigation , while the charging power @xmath8 of the corresponding storage / power plant would be set to zero , that is @xmath26 this procedure is not applied in this work .
overall , the optimisation approach can be seen as an investment decision approach with a perfect and unlimited prediction horizon , while the actual dispatch of the individual units would have to be decided upon in a more detailed simulation with a much smaller prediction horizon . for prospective power systems
the residual load depends strongly on the amount of non - dispatchable solar and wind energy in the system .
provided that time series for the load ( @xmath27 ) and for the availability of solar ( @xmath28 ) and wind ( @xmath29 ) resources are available , it is straightforward to investigate the implications of their installation on the residual loads . for this purpose
we follow a procedure outlined in @xcite . with @xmath30 representing the long - term average of a time series
, the residual load @xmath31 is modelled by @xmath32 here , the respective shares of wind and solar power generation of the gross electricity demand are given by @xmath33 and @xmath34 .
the parameter @xmath35 is termed the average renewable energy power generation factor and determines the total electricity which can be produced from vres ( cf .
a value of @xmath36 corresponds to a scenario , in which the long - term averaged total generation from solar and wind resources is equal to the average demand . in this work
, we concentrate on the technical aspect of a future power system .
hence , we focus in our model on the costs for the installation and the operation of combinations of generation , storage and backup capacities . for every investigated scenario ,
the total equivalent annual costs @xmath37 are given by the costs for storage and backup @xmath17 as reflected by the cost function of and the costs for the generation capacities @xmath38
@xmath39 the generation costs @xmath38 are given by @xmath40 here , the parameters @xmath41 and @xmath42 describe the respective installed capacities of wind and solar as given by the wind share @xmath43 and the generation factory @xmath35 for each scenario .
the parameters @xmath44 and @xmath45 describe the equivalent annual installation and operating costs per power unit , which are fixed for each scenario ( cf .
section [ ssec : data ] ) .
the parameter @xmath46 reflects the costs for must - run units ; the latter are not considered in this work .
overall , the generation costs are fixed for each investigated scenario and not relevant for the optimisation process described in the section [ ssec : optimisation ] .
the modelling approach described in section [ sec : modelling ] is now applied to data for germany .
germany is chosen in this work as it has significantly increased its res share over the last years with more than 32% of electricity originating from res in 2015 @xcite .
furthermore , it has also been the subject of our previous publication dealing with a more meteorologically focused perspective on the storage demand in future renewable power systems @xcite .
the modelling approach requires long - term power generation for renewable energy sources as well as corresponding load data . in this work ,
we use wind and solar power generation data with hourly resolution ( @xmath1=1h ) .
the division between wind and solar power generation is determined by the wind share @xmath43 ( cf .
section [ ssec : resources ] ) , while for the wind data itself holds that 52% of the installed capacity is onshore .
the data sets span the year 2006 - 2012 ( @xmath47=7 ) and thus contain n=61368 data points each .
this way , the variability of the power production from vres over multiple years is included in the data .
the load data originate from the data from the transmission system operators ( tsos ) in germany for the same period and are available from entso - e .
the average load in the period of investigation is @xmath48 , subsequent energy values are normalised where appropriate to average load hours , abbreviated av.l.h .
, with 1 av.l.h .
@xmath49 55.1 gwh .
the data sets were also used in previous publications @xcite .
more details regarding the wind and solar power generation data sets can be found in @xcite .
three different storage technologies are used in this work to balance fluctuations .
first are pumped hydro storages , which are nowadays mainly used in germany to balance fluctuations on the scale of multiple hours ( phs , @xmath50 @xcite ) .
the current installed capacity is around 39 gwh according to @xcite with a small potential for additional capacities @xcite .
given germany s proximity to the alps region and the large hydro facilities there , we assume a maximum pumped - hydro storage size available to the german power system as @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh . as
this value is still below the required storage capacities found in other works investigating fully renewable power supply systems(@xcite ) , further storage technologies are considered .
batteries based on lithium - ion ( lib , @xmath52
@xcite ) are the second storage technology considered in our work . finally , as a rather long - term storage we use synthetic hydrogen stored in caverns ( h2s , @xmath53 @xcite ) .
by not including any restrictions on the operation of the storage devices , all storages can in principle be used as short - term and long - term storage .
the equivalent annual costs @xmath54 for the generation and storage units is composed of installation costs as well as annual operating and maintenance ( o&m ) costs . for consistency with respect to the underlying assumptions , the data are to a large extend based on a single source and
are assumed costs for germany by mid - century .
more precisely , the costs for storages and generation capacities used in this work are based on @xcite ( in turn based on @xcite ) as well as the updated version @xcite and shown in table [ tab : parameters ] .
the prices @xmath54 are achieved by converting these costs to equivalent annual costs using a fixed interest rate of r=6% throughout the work .
details regarding this conversion to equivalent annual costs can be found in [ sec : app_eac ] . the costs for backup energy is generally set to @xmath23=0.15 eur / kwh .
as mentioned already above in section [ ssec : optimisation ] , the backup energy is modelled this way to include a wide variety of backup options .
.parameters for base case of economic evaluation : @xmath55 are installation costs , @xmath56 describe o&m costs in relation to installation costs , n is the lifetime ( cf .
[ sec : app_eac ] ) , data are based on @xcite . [ cols="<,>,>,^,^ " , ] let us start our investigation of the cost - optimal design of the prospective german power supply system by studying the influence of the wind share @xmath43 and the power generation factor @xmath35 on the total system costs @xmath37 .
for this , we use the cost function as defined in equation and consider the three storage technologies mentioned in section [ ssec : data ] .
figure [ fig : heatmap_costs ] shows the total equivalent annual costs @xmath37 for scenarios with different combinations of @xmath43 and @xmath35 considering all three storage technologies ( @xmath57 ) . in these scenarios , the size of the pumped - hydro storages was limited to @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh ] in order to reflect the limited resources for pumped hydro ( see also section [ ssec : data ] ) and the backup power was in practical terms unlimited ( @xmath58 ) .
as we can see from the figure , the sensitivity with respect to the wind share @xmath43 is higher than the sensitivity with respect to the capacity factor @xmath35 . as an example , for a fixed generation factor @xmath36 , the total system costs @xmath37 vary from @xmath59 65.1 bn eur as equivalent annual costs for a solar - only scenario ( @xmath43=0.0 ) to @xmath59 41.1 bn eur as equivalent annual costs for a scenario with @xmath60 . with respect to the generation factor @xmath35 ,
the costs variation is significantly less , e.g. for @xmath60 from @xmath59 40.8 bn eur as equivalent annual costs for @xmath61 to @xmath59 46.9 bn eur as equivalent annual costs for @xmath62 . for scenarios with different values of wind share @xmath43 and generation factor @xmath35 .
for all scenarios , the pumped - hydro storage size was limited to @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh and the backup power was in practical terms unlimited ( @xmath58).,scaledwidth=95.0% ] overall , this figure illustrates that the total equivalent annual system costs @xmath37 as defined in this work are strongly influenced by the costs for different technology ( cf .
table [ tab : parameters ] ) .
based on the previous results , we will now investigate scenarios with two different wind shares as well as different capacity factors and draw comparisons between them .
for this , we will focus on scenarios with a wind share of @xmath63 ( cheapest option and also share in related works ( e.g. @xcite ) as well as scenarios with a high solar share , that is @xmath64 , in the upcoming investigations .
the total systems costs illustrated in figure [ fig : heatmap_costs ] are made up of different components .
figure [ fig : histogram_cost_basecase ] shows the total equivalent annual costs @xmath37 and its breakdown to system components for different values of the generation factor @xmath35 as well as for a wind share @xmath64 ( fig . [ fig : histogram_cost_basecase ] , left panel ) and a wind share @xmath63 ( fig . [
fig : histogram_cost_basecase ] , right panel ) .
it can be seen in this figure that the share of costs related to storage capacities in all investigated scenarios make up to only 11% of the total costs , while in all scenarios the highest share of costs is related to wind and solar generation capacities as well as backup costs .
the share of costs related to generation and backup is higher than the results found by bussar et al . , where those costs were responsible for about 74% of the total costs ( excluding grid costs ) .
this is likely due to differences in the particular costs assumptions .
the overall trend that the major shore of costs is related to the production costs is consistent in both works though .
furthermore , the figure illustrates that increasing the generation factor @xmath35 significantly decreases the need for energy originating from backup power plants , while the costs for storage and hence its capacities are increased .
as already seen in figure [ fig : heatmap_costs ] , the overall costs @xmath37 increase slightly only when the generation factor @xmath35 is changed , implying that the increased costs for generation capacities are approximately equivalent to the combined costs of decreased costs for backup and increased costs for storage capacities .
besides the overall difference in costs related to the higher generation costs in the high - solar scenario ( @xmath64 ) , a major difference between the scenarios with different wind shares @xmath43 is the presence of lithium - ion batteries in scenarios with a wind share of @xmath43=0.50 , while no lithium - ion batteries at all are present in the results for scenarios with a wind share of @xmath63 .
this can be explained by considering that lithium - ion batteries have rather low charging and discharging power costs and high capacity costs , and these are favourable for scenarios with a relatively high solar - share and hence strong daily fluctuations . on the contrary ,
the cost - minimised setup in wind - dominated scenarios requires the utilisation of hydrogen storage which is characterised by comparably high charging and discharging power costs and low capacity costs . and its breakdown on system components for different values of generation factor @xmath35 and a wind share @xmath64 ( left panel ) and a wind share @xmath63 ( right panel ) .
the lower panels show a detail of the upper panels for a more straightforward discussion of the costs originating from storages .
for all scenarios , the pumped - hydro storage size was limited to @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh and the backup power was in practical terms unlimited ( @xmath58).,scaledwidth=95.0% ] the aforementioned characteristics of the different storage technologies can also be identified when further splitting the storage costs into costs for charging and discharging power on one side and capacity on the other side .
for all the scenarios shown in figure [ fig : histogram_cost_basecase ] the equivalent annual costs for charging and discharging power are significantly higher than the respective equivalent annual costs for capacity when pumped - hydro storage or synthetic hydrogen storage is considered . on the contrary , the equivalent annual costs for storage capacity of the lithium - ion batteries are found to be up to ten times the equivalent annual costs for the charging and discharging power . with the actual power dimensioning of the storage being a focus of this work , figure [
fig : pcpd_basecase ] examines the previous results in more detail and illustrates the corresponding @xmath8 and discharging power @xmath9 for pumped - hydro storage , lithium - ion storage and hydrogen storage .
a wind share of @xmath64 was used for the left panel and a wind share of @xmath63 for the right panel . and discharging power @xmath9 for pumped - hydro , lithium - ion and hydrogen storage for different values of generation factor @xmath35 and a wind share @xmath64 ( left panel ) and a wind share @xmath63 ( right panel ) .
in each block , the generation factor @xmath35 increases from @xmath35=0.90 to @xmath35=1.30 in step sizes of 0.05 .
for all scenarios , the pumped - hydro storage size was limited to @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh and the backup power was in practical terms unlimited ( @xmath58).,scaledwidth=95.0% ] for scenarios with a wind share @xmath64 , the charging @xmath65 and discharging power @xmath66 of the pumped - hydro storage is almost irrespective of the generation factor @xmath35 , whereas the discharging power @xmath66 is always lower than the corresponding charging power @xmath65 .
this can be understood by considering that there are clear daily patterns in the residual load of these high - solar scenarios with comparably high surpluses during the day .
the charging @xmath67 and discharging power @xmath68 for the lithium - ion batteries reach their respective maximum value for a generation factor of @xmath69 and decreases for larger values of @xmath35 . for the latter , the deficit hours in the residual load are reduced , furthermore the peaks of the surplus generation at midday can be curtailed , leading to a reduced need for the rather expensive lithium - ion batteries . for scenarios with a generation factor @xmath70 , the charging @xmath71 and
discharging power @xmath72 of the hydrogen storage exceeds the corresponding values of the lithium - ion batteries .
this implies that the hydrogen storage replaces the lithium - ion storage in some respects .
furthermore , it is worth mentioning that the combined discharging power of all three storages increases with increasing generation factor @xmath35 , reaching a combined sum almost as high as the average load @xmath73 .
comparing the high - solar scenario ( @xmath74 , figure [ fig : pcpd_basecase ] , left panel ) with the wind - dominated scenario ( @xmath60 , right panel ) shows that the charging @xmath65 and discharging power @xmath66 of the pumped - hydro storage is significantly lower in the scenario with @xmath60 . furthermore , the synthetic hydrogen is found to be the storage technology with the highest charging @xmath71 and discharging power @xmath72 .
the results presented in this section are meant to represent a base case of our investigations .
based on these results , we will continue by studying three different aspects of the scenarios .
first , the backup energy and its power requirements ( section [ ssec : minimal_backup ] ) .
second , the need of seasonal storage ( section [ ssec : noptg ] ) as well as , third , the influence of the yearly variations on the results ( section [ ssec : yearlyvariations ] ) . in the previous investigations in sections [ ssec : generalalphagamma ] and
[ ssec : basecase ] , the backup power @xmath24 was not limited so as to include a variety of backup technologies .
that was also the reason why we chose an energy - only price for backup usage and did not include any power price . in this section ,
we aim to limit the backup power and investigate the resulting effects on the average equivalent full - load hours of the backup power plants ( cf .
condition [ eq : limitbackup ] in section [ sec : modelling ] ) . for this , the average full - load hours of the backup power plants are defined as @xmath75 figure [ fig : backup_fullload ] shows the total equivalent annual costs @xmath37 ( top ) and corresponding average full - load hours flh ( bottom ) as a function of the backup power limit @xmath24 for different values of the generation factor @xmath35 , where a wind share @xmath74 was used for the left and a wind share @xmath60 was used for the right panel .
( top ) and corresponding average full - load hours per year flh ( bottom ) as function of backup power limit @xmath24 for different values of generation factor @xmath35 : wind share is @xmath64 ( left panel ) and @xmath63 ( right panel ) .
for all scenarios , the pumped - hydro storage size was limited to @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh . ] starting with the upper end of the investigated backup power limits ( right side of each half , @xmath76 50 gw ) , the total equivalent annual costs @xmath37 is almost equal to the corresponding costs without backup power limits . in this case
, the average full - load hours reach values of up to about 1890 equivalent full - load hours per year for @xmath77 and @xmath60 , slightly less for @xmath74 , and as low as about 335 equivalent full - load hours for @xmath78 and @xmath60 . decreasing the backup power limit down to
@xmath76 30 gw only increases the total equivalent annual costs @xmath37 for all investigated scenarios .
however , even then the backup power plants reach about 3000 equivalent full - load hours only .
reduction of the backup power limit @xmath24 below this threshold leads to a significant increase in total equivalent annual costs @xmath37 .
the most extreme values of @xmath37 should be ignored , as the backup power plants are almost used permanently in these scenarios , which would presumably result in production costs much below the fixed costs used in the approach and therefore does seem to be inconsistent with the assumptions . in summary , the investigations in this section show that a compromise between the installation of excess generation capacities and backup power plant capacities is required . for our base case scenarios , we assumed that the synthetic hydrogen storage technology will make large advancements in comparison to today and that cheap storage capacities in underground caverns will become available , leading to the costs assumptions in table [ tab : parameters ] . using these assumptions ,
the hydrogen storage was found to be installed in large capacities , in particular in the scenarios with a wind share @xmath60 ( cf .
section [ ssec : basecase ] ) . to further study the role of hydrogen storage
, we also investigated scenarios where the proposed hydrogen technology does not exist , that is only pumped - hydro storage and lithium - ion batteries exist as storage technologies . using this restriction
, we find that the total equivalent annual costs @xmath37 increase significantly in most of the investigated scenarios . as an example , for scenarios with a wind share of @xmath74 the total equivalent annual costs @xmath37 increase from @xmath59 45.2 bn eur to @xmath59 46.8 bn eur for @xmath35=1.0 , and from @xmath59 48.5 bn eur to @xmath59 51.4 bn eur for @xmath35=1.3 .
the increase in costs is largely due to an increase in costs for lithium - ion batteries as well as an increased backup usage . reducing the backup costs to @xmath23=0.12 eur / kwh to take into account the latter effect
was found to reduce the total equivalent annual costs @xmath37 , e.g. for a wind share @xmath64 and a generation factor @xmath35=1.0 to @xmath59 44.4 bn eur . for scenarios with a wind share @xmath60 ,
the increase in costs for scenarios without hydrogen storage is even higher .
as an example , the total equivalent annual costs @xmath37 increase from @xmath59 41.1 bn eur to @xmath59 43.9 bn eur for @xmath35=1.0 , and from @xmath59 42.4 bn eur to @xmath59 47.2 bn eur for @xmath35=1.3 . in these cases ,
the lack of hydrogen storage is compensated to a large extent by increasing usage of backup power , while also the capacities for pumped - hydro storage are increased and small capacities of lithium - ion batteries are installed . in this case , the reduction of the backup costs to @xmath23=0.12 eur / kwh led to a decrease in total system costs for e.g. a generation factor @xmath35=1.0 to @xmath59 40.1 bn eur .
an iterative process could be used to adjust the backup costs to the resulting full - load hours for different backup technologies .
however this would contradict the simple design of our approach and is beyond the scope of this work . as a final variation of our base case scenarios , we investigate an aspect which is related to the meteorological perspective of the investigated scenarios .
in contrast to the previous parts of this work , we treat each year of the scenarios individually in this section , so as to reduce the statistical variation of the original data set . for this purpose ,
all years are treated individually , implying that the variations in average load are eliminated by normalisation and the installed capacities for wind and solar generation capacities are adjusted to the respective generation pattern in the corresponding year .
figure [ fig : heatmap_yearlyvariations ] shows the standard variation of the total equivalent annual system costs @xmath37 between different years for different values of the wind share @xmath43 and the generation factor @xmath35 .
one can see that the variations are lowest with a low wind share @xmath43=0.5 and low generation factor @xmath35=0.9 . in this case , a significant share of the energy is provided by backup power ( cf .
figure [ fig : histogram_cost_basecase ] ) which allows to overcome variations in the generation and load patterns to a large extent . increasing the wind share @xmath43 and the generation factor
@xmath35 increases the variations in the total equivalent annual costs @xmath37 , reaching a standard variation of up to 4% . in absolute values and considering the variations in the original data sets
, this corresponds to e.g. total equivalent annual system costs varying between @xmath59 38.9 bn eur for the year 2009 to @xmath59 43.8 bn eur for the year 2010 in scenarios with a wind share @xmath43=0.8 and a generation factor @xmath35=1.0 . in comparison to the seven - years annual system costs of @xmath59 41.1 bn eur , these variations correspond to a deviation of up to 7% . in terms of the installed storage capacities , the differences between individual years
lead to e.g. a discharging power @xmath72 for the hydrogen storage of @xmath79 10.0 gw for the year 2009 and @xmath79 17.0 gw for the year 2010 , while a value of @xmath79 12.0 gw was found for the seven - years data in the optimal case .
the system component which was found to vary most between different years is the size of the hydrogen storages @xmath80 with @xmath81 119 av.l.h . for the year 2010 and @xmath81
364 av.l.h . for the year 2006 .
it is worth noting that even the smallest value was found to be higher than the value for the seven - years data @xmath81 113 av.l.h .
indicating that particularly the required size of the hydrogen storage is overestimated when individual years are considered in comparison to long - term data .
overall , these results illustrate the large variations in the underlying generation patterns for wind and solar energy and consequently the need to consider multi - year data when studying prospective power supply systems .
furthermore , this result implies that the previously found low average full - load hours might be even lower in some years , so that some backup power capacities are not used for a year or longer while always being available on short - term basis . between different years for different values of the wind share @xmath43 and the generation factor @xmath35 . for all scenarios ,
the pumped - hydro storage size was limited to @xmath51 = 4 av.l.h .
@xmath49 220.6 gwh and the backup power was in practical terms unlimited ( @xmath58).,scaledwidth=95.0% ]
future scenarios based on res will have high shares of electricity originating from variable resources like wind and solar energy . to guarantee a stable system
, measures have to be taken to balance load and generation at all times . in this work
, we looked at prospective power supply systems with very high shares of vres , in which the fluctuations are balanced using a combination of storage devices as well as fully - flexible backup power plants and curtailment .
we presented a modelling approach to systematically study which combination of these balancing options leads to minimum costs . in accordance with results from related works
, we found that the highest shares in costs are related to the generation capacities , while only a minor share is related to the storage costs .
for the storage , the costs for capacities were found to be generally lower compared to the costs for charging and discharging .
furthermore , our results indicate that storage based on synthetic hydrogen as well as pumped - hydro storages might be favourable to be installed in scenarios in which a significant share of the vres generation originates from wind .
lithium - ion storage with its high efficiencies but high capacity costs was only found to be present in case of scenarios with high shares of pv . in this vein
, we can confirm that the daily fluctuations in the solar scenarios lead to a need for efficient , small daily storage , whereas seasonal storage are required for scenarios dominated by wind energy .
this implies that the cheap capacity costs of hydrogen storage can partly overcome its low efficiency .
investigating scenarios without hydrogen storage being available significantly increased the overall system costs , which underlines the importance of this technology .
we also took a closer look at backup power need and found that backup would be needed with high backup power up to about 30 gw . these power plants would only run on average about 2000 - 3000 equivalent full - load hours .
furthermore , we found that the yearly variations reach values up to 4% , which corresponds to a difference up to five billion eur equivalent annual costs between cheapest and most expensive year . in practical terms , these results imply that the system ( operator and/or society ) has to come up with solutions how to finance those large capacities which are used at rare intervals .
solutions such as demand - side management might help to shift consumption by a few hours , yet concepts need to be found how to deal with long periods of dead calm which occur only once every years and how many backup capacities should be kept as final reserve for this purpose .
we conclude this work with recommendations regarding future work : as we have seen in this work , the results vary significantly depending on the weather situation in the investigated scenario .
hence , future work should also be done with multi - year data so as to include the statistical variations of these long - term weather patterns .
furthermore , as also done in this work , we recommend to not explicitly include the generation costs in the optimisation part but instead study scenarios with different wind - solar - mixes and draw comparisons between them . regarding the modelling approach presented in this work , a further development could trigger the modelling of the backup process by using e.g. an iterative approach to match the price of backup with resulting full - load hours . including large capacities of controllable power plants like biomass or hydro power as virtual storage
might lead to lower storage needs .
this way , a stable and affordable power supply system based mainly on renewable energy sources will likely be found .
the authors thank thomas vogt and lukas wienholt for helpful discussions . financial support of the first author through a phd scholarship by the lower saxony ministry for science and culture ( mwk ) is acknowledged .
the authors thank lueder von bremen and alexander kies ( forwind centre for wind energy research , university of oldenburg , germany ) for providing the wind and solar power generation data set used in this work .
this data set was developed in the research project restore 2050 ( no . 03sf0439 ) funded by the german federal ministry of education and research in the energy storage funding initiative .
the parameters @xmath54 corresponding to the equivalent annual costs for technology @xmath82 are composed of @xmath83 , an annual share of the installation costs @xmath84 , and the annual operating & maintenance costs @xmath85 : @xmath86 with @xmath87 and @xmath88 being the lifetime in years and @xmath89 being the interest rate .
an interest rate of @xmath90 is used throughout this work .
the annual operating & maintenance costs @xmath85 are directly related to the overall installation costs through the factor @xmath56 : @xmath91 the values for @xmath84 , @xmath56 and @xmath88 are given in table [ tab : parameters ] .
10 dominik heide , lueder von bremen , martin greiner , clemens hoffmann , markus speckmann , and stefan bofinger .
seasonal optimal mix of wind and solar power in a future , highly renewable europe .
, 35(11):24832489 , 2010 . elaine k hart and mark z jacobson
. a monte carlo approach to generator portfolio planning and carbon emissions assessments of systems with large penetrations of variable renewables .
, 36(8):22782286 , 2011 .
cory budischak , deanna sewell , heather thomson , leon mach , dana e veron , and willett kempton .
cost - minimized combinations of wind power , solar power and electrochemical storage , powering the grid up to 99.9% of the time . , 2012 .
sarah becker , bethany a. frew , gorm b. andresen , mark z. jacobson , stefan schramm , and martin greiner .
renewable build - up pathways for the us : generation costs are not system costs . , 81:437445 , 2015 . bert jm de vries , detlef p van vuuren , and monique m hoogwijk .
renewable energy sources : their global potential for the first - half of the 21st century at a global level : an integrated approach . , 35(4):25902610 , 2007 .
mark z jacobson and mark a delucchi . providing all global energy with wind , water , and solar power , part i : technologies , energy resources , quantities and areas of infrastructure , and materials .
, 39(3):11541169 , 2011 .
danel frstenwerth , lars waldmann , michael sterner , martin thema , fabian eckert , albert moser , andreas schfer , tim drees , christian rehtanz , ulf hger , jan kays , andre seack , dirk uwe sauer , matthias leuthold , and philipp stcker .
stromspeicher in der energiewende . , 2012 .
lueder von bremen , k knorr , b lange , and s bofinger . a fully renewable power supply scenario for europe : the weather determines storage and transport . in _
8th international workshop on large scale integration of wind power into power systems as well as on transmission networks for offshore wind farms , bremen _
, 2009 .
dominik heide , martin greiner , lueder von bremen , and clemens hoffmann .
reduced storage and balancing needs in a fully renewable european power system with excess wind and solar power generation .
, 36(9):25152523 , 2011 .
thomas weiss and detlef schulz .
development of fluctuating renewable energy sources and its influence on the future energy storage needs of selected european countries . in _ proceeding of the 4th international youth conference on energy
( iyce ) , 2013 _ , pages 15 , june 2013 .
christian bussar , philipp stcker , zhuang cai , luiz moraes jr . , dirk magnor , pablo wiernes , niklas van bracht , albert moser , and dirk uwe sauer .
large - scale integration of renewable energies and impact on storage demand in a european renewable power system of 2050 - sensitivity study .
, 6:110 , 2016 .
alexander kies , kabitri nag , lueder von bremen , elke lorenz , and deltev heinemann .
investigation of balancing effects in long term renewable energy feed - in with respect to the transmission grid .
, 12(1):9195 , 2015 .
franziska adamek , t aundrup , w glaunsinger , m kleimaier , h landinger , m leuthold , b lunz , a moser , c pape , h pluntke , n rotering , d u sauer , m sterner , and w wellow .
: energiespeicher fr die energiewende gesamttext . ,
2012 .
joachim nitsch , thomas pregger , yvonne scholz , michael sterner , norman gerhardt , amany von oehsen , carsten pape , yves - marie saint - drenan , and bernd wenzel .
langfristszenarien und strategien fr den ausbau der erneuerbaren energien in deutschland bei bercksichtigung der entwicklung in europa und global - leitstudie 2010 .
arbeitsgemeinschaft dlr , iwes , ifne , 2010 .
joachim nitsch , thomas pregger , yvonne scholz , michael sterner , norman gerhardt , amany von oehsen , carsten pape , yves - marie saint - drenan , and bernd wenzel .
langfristszenarien und strategien fr den ausbau der erneuerbaren energien in deutschland bei bercksichtigung der entwicklung in europa und global - datenanhang ii zum schlussbericht .
arbeitsgemeinschaft dlr , iwes , ifne , 2012 .
christoph kost , johannes n. mayer , jessica thomsen , niklas hartmann , charlotte senkpiel , simon philipps , sebastian nold , simon lude , and thomas schlegl .
stromgestehungskosten erneuerbare energien .
fraunhofer - institut fr solare energiesysteme ise , 2013 .
|
prospective power supply systems based on renewable energy sources require measures to balance power generation and load at all times .
the utilisation of storage devices and backup power plants is widely suggested for this purpose , whereas the best combination is still to be found . in this work
, we present a modelling approach to systematically study scenarios of future power supply systems with a high share of electricity originating from wind and solar resources . by considering backup as a subordinate source of electricity with energy - only costs ,
the approach is independent of the actual full - load hours of the backup power plants . applying the approach to multi - year meteorological data for germany
, cost - optimised combinations of storage devices and backup power are identified .
we find that even in scenarios with significant excess generation capacities there is a need for storage devices or backup power plants with discharging power on the same order as the average load to balance the system at all times .
furthermore , these capacities seem to be required in some years of the multi - year period only .
our results imply that the societal need of having electricity available at all times can likely be satisfied by installing over - capacities only , whereas a balance has to be found between installing additional backup or storage or generation capacities .
energy storage , power system , optimisation
|
cygnus x-1 , identified with the bright ( v@xmath18 ) star hde226868 ( bolton 1972 ; webster & murdin 1972 ) , has long been regarded as the best black hole candidate among the high - mass x - ray binaries . as such , it has been an object of extensive observation over the past two and a half decades . perhaps surprisingly , there remain some important uncertainties and discrepancies in the derived properties of the system . in particular , ninkov , walker , & yang ( 1987 , hereafter nwy ) report evidence of possible period variation , discussed below , and additional periodicities on timescales ranging from 39 days to 4.5 years have been suggested by kemp , herman , & barbour ( 1978 ) , wilson and fox ( 1981 ) , priedhorsky , terrell , & holt ( 1983 ) , and walker and quintanilla ( 1978 ) .
+ hde226868 is a single - lined spectroscopic binary .
the optical spectrum has been classified o9.7iab by walborn ( 1973 ) , with variable emission at heii @xmath24686 and , less prominently , at the balmer lines .
the radial velocity period of 5.6 days has been established for over two decades , and the definitive orbital elements and ephemeris were published by gies and bolton ( 1982 , hereafter gb ) .
they give a period of 5.59974 @xmath3 0.00008 days , a precision which translates to an uncertainty of only @xmath30.02 cycle in phase at the present time .
+ however , nwy , in a subsequent radial - velocity study , reported a somewhat longer period of 5.6017 @xmath3 0.0001 days , and suggest the possibility of a period increase with time .
these results have the effect that the orbital phase of the cyg x-1 system is then highly uncertain at the present epoch .
the phase calculated using the ephemeris of nwy differs from the gb phase by 0.5 cycle , and if the period is in fact varying , the phase at the present epoch is completely indeterminate .
+ if correct , a time - varying period would have important implications for the mass transfer rate and hence evolution of the binary . furthermore , with this indeterminacy in phase , it is impossible to ascribe and interpret accurately features in the x - ray light curve such as the dipping behaviour , in particular the distribution of dipping with orbital phase ( e.g. remillard and canizares 1984 ) . to resolve these ambiguities we therefore undertook a program of observations to re - establish the orbital ephemeris of hde226868 .
these data also allowed us to undertake a more detailed investigation of the long - term stability of the orbit , as high quality data now exist over a @xmath130 year baseline .
our new ephemeris has thereby allowed us to complete a survey of the distribution of x - ray dipping with orbital phase ( balucinska - church et al .
through the la palma service programme , and with the assistance of regularly scheduled observers , we obtained ccd spectra of hde226868 using the intermediate dispersion spectrograph on the isaac newton telescope on 17 nights in may and june 1996 .
these included 11 consecutive nights covering two complete orbital cycles . in all we obtained 37 spectra of the object , each covering a spectral range of 4100 to 4900 , with a dispersion of approximately 0.8 / pixel .
exposures varied from 100 to 200 seconds , and yielded a s / n in excess of 100 .
in addition , we observed 19 cep , a radial velocity standard of similar spectral type , on two nights , for reference and calibration purposes .
+ the spectra were extracted and reduced using standard iraf routines . the extracted spectra were wavelength calibrated using a cu - ar arc recorded immediately before and after the program spectra , and then normalized .
our wavelength calibration accuracy had rms residuals of @xmath10.10 .
typical hde226868 and 19 cep spectra are shown in figure 1 .
we determined radial velocities by cross - correlation of the hde226868 spectra with the 19 cep spectra , using iraf routines .
separate determinations were made for the hydrogen and helium lines .
we used the hei absorption lines at 4388 , 4472 , 4713 and 4921 .
the resulting velocities are given in table 1 , and were reduced to heliocentric velocities using the catalogue value ( hoffleit 1982 ) for the radial velocity of 19 cep .
we performed period searches on both sets of velocities , as well as on our velocities in combination with previously determined velocities from the literature .
we used several different period - search techniques , including scargle periodograms , fourier searches , and chi - square minimization to a circular orbit fit .
all methods yielded similar results , as follows .
+ the best fit period to the hydrogen line velocities for our data alone is 5.566 @xmath3 0.012 days , while the best fit to our helium line data is 5.629 @xmath3 0.015 days .
the apparent discrepancy here is startling , as is the amount of apparent change in period from the canonical value in either case . both of these issues can be readily resolved . + it is well established that the hydrogen lines are contaminated by emission whose velocity is in approximate antiphase with the absorption lines ( gb ; hutchings , crampton , & bolton 1979 ) .
this produces variable and irregular line profiles which yield spurious velocities and periods when cross - correlated with standard spectra .
so any period determined from the hydrogen line velocities is highly suspect .
for this reason we reject all velocity determinations based on the hydrogen lines , and in working with historical data below , we use only published velocities determined for the helium lines . + the fact that the best - fit period to the helium lines seems to be @xmath4 larger than previously determined values should also be viewed with some scepticism .
in fact , with a dense grid of observations over a relatively short period of time , the quality of fit is good over a much broader period range than suggested by the quoted uncertainty .
fitting our data to the nwy period , or to the gb period , produces @xmath5 residuals which are not significantly worse than those for our best fit .
in fact , the formal uncertainty calculated by the fitting routines is undoubtedly too small ; this is because the scatter of the data is such that even for the best orbital models the resulting @xmath5 is too large for the formal uncertainty calculations to be valid .
we shall see below that this is not unique to our results but in fact has been a problem for most , and perhaps all , previous published orbital solutions for cyg x-1 ; uncertainties have been systematically underestimated , resulting in apparent discrepancies where none really exist . + to test for the possibility of period variation , we combined our observations with all previously published helium - line velocity data ( brucato and zappala 1974 ; gb ; nwy ) and performed a @xmath5 minimization weighted fit to a circular orbit both with a fixed period and with a period which varies linearly with time , using the minuit package of function - fitting routines ( james 1994 ) .
the results are summarized in table 2 and plotted in figure 2 .
the best fixed period is 5.5998 @xmath3 0.0001 days , with a @xmath5 of 4284 ; the best variable period is 5.5997 @xmath3 0.0001 days , with @xmath6 of 3.8 @xmath7 10@xmath8 and @xmath5 of 4231 .
but with only 216 data points , this means that the @xmath9 is @xmath120 in both cases .
our new period agrees well with the photometric period derived recently by voloshina , lyutyi & tarasov ( 1997 ) and a solution from hei @xmath26678 measurements by sowers et al ( 1998 ) .
our values for @xmath10 and k are within 2@xmath11 of previous results ( nwy and gb ) .
+ the variable - period model produces a slightly better fit to the data , but , again , with a @xmath5 difference of less than 1.5% , the improvement in quality of fit between that and the fixed - period model is _ not _ statistically significant .
we conclude that there is no compelling evidence for period variation in cygnus x-1 .
we have therefore adopted the orbital elements of the best - fit constant period model in all further discussion .
+ we computed o c residuals for all the helium - line velocities and performed period searches on these .
no significant periodicity was found in these residuals . in particular
, we find no evidence of the suggested periodicities at 39d or 78d ( kemp , herman , & barbour 1978 ) , or 4.5 years ( wilson and fox 1981 ) or of the 294 day x - ray modulation ( priedhorsky , terrell , & holt 1983 ) or the 91d photometric period ( walker and quintanilla 1978 ) .
the absence of a 294 d signal was also noted by gies & bolton ( 1984 ) .
our best - fit period for the combined datasets , 5.5998 @xmath3 0.0001 d is identical to well within the standard errors with the gb period of 5.59974 @xmath3 0.00008 d , and our t@xmath12 , or epoch of inferior conjunction , agrees with the prediction of the gb ephemeris to within 0.014 phase , or 0.078 d. since the uncertainty in the periods yields an uncertainty in the gb prediction of t@xmath12 of @xmath3 0.12 days at this epoch , this is excellent agreement .
thus our results confirm the ephemeris of gb and refute suggestions of period variation by nwy .
+ why , then , did nwy s result show such a discrepancy and suggest a change in period ?
they reported that the period determined from their data alone was 5.60172 @xmath3 0.00003 d , differing by 20 @xmath11 from the period determined from all previous observations .
we suggest several sources for this error .
first of all , we performed our own period searches on the data published by nwy ; the best fit we find is at 5.6002 d , or only 2 @xmath11 greater than the gb value , so we may be looking at much ado about a typographical error .
in addition , there is excellent phase agreement between the nwy data and the gb ephemeris , again suggesting no real period discrepancy .
+ finally , it seems that the error estimates for both the nwy period and their period determined from historical data are too small by as much as an order of magnitude .
this is primarily because the formal error calculations used to determine these error estimates are based on the assumption of a very good fit of the model to the data , i.e. that @xmath131 .
this condition is _ not _ met by any of the data sets and fits used in this work , presumably because the estimates of the uncertainties in the velocities were too small , although perhaps because of variability in the source .
we recalculated the fits , trying larger estimates for the velocity uncertainties until the reduced chi - square criterion was satisfied .
we found that the period uncertainties then were about an order of magnitude larger than reported by the original authors .
for example , treating nwy s data in this way yields a best - fit period of 5.6002 @xmath3 0.0003 d , and the discrepancy with gb disappears .
+ an additional source of underestimated error uncertainty is the inclusion by both nwy and bolton ( 1975 ) of two velocities obtained by seyfert and popper ( 1941 ) , ostensibly to improve the precision of the period determinations . for example , bolton ( 1975 ) uses these points to decrease his uncertainty estimate by a factor of 10 ; nwy do not discuss the effect of including these points on their uncertainty .
popper ( 1996 , personal communication ) suggests that these velocities are `` very weak reeds on which to hang significant conclusions '' , the velocities having considerable uncertainties and being based on averaging velocities from lines of several different species . because of the problems with hydrogen - line velocities discussed above , inclusion of these points is therefore unlikely to improve the period uncertainty .
+ gb also introduced `` velocity corrections '' to many of their velocities , shifting spectra to fit a mean interstellar k - line velocity due to instabilities in their spectrograph .
it is not unlikely that this introduced uncertainties larger than the mean errors they report .
we have carried out a programme of radial velocity determinations for the black hole binary cyg x-1 , and have combined these data with data used previously to determine the period , to provide a new orbital ephemeris for the source .
a critical consideration of the errors associated with previous work has shown these to be underestimated . based on this
, our main conclusion is that there is no evidence for a change in orbital period as suggested by ninkov et al .
finally , our new ephemeris allows the orbital phase calculation for cyg x-1 with an error that is much reduced compared with the error that would be attached to extrapolating the gies and bolton ephemeris with its quoted accuracy to the present epoch .
we are very grateful to the la palma support astronomers , and particularly don pollacco , who operate the service programme , and to those astronomers who allowed the cyg x-1 observations to be taken during their own time .
the isaac newton group of telescopes is operated on the island of la palma by the royal greenwich observatory in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias .
jls wishes gratefully to acknowledge the following sources of partial support for his participation in this work : the margaret cullinan wray charitable lead annuity trust ( through the aas small grants program ) ; the nasa jove program ; a theodore dunham , jr . , grant of the fund for astrophysical research ; the international astronomical union exchange of astronomers program ; and the department of astrophysics of the university of oxford .
heliocentric&phase & & & + julian date & & @xmath14 & @xmath15 & @xmath14 & @xmath15 & 0 - c + 2450000 + & & & & km s@xmath16 + + 228.739 & 0.830 & -71.6 & 6.8 & -54.4 & 5.2 & 17.1 + 228.742 & 0.831 & -74.8 & 7.2 & -59.0 & 5.3 & 12.3 + 228.744 & 0.831 & -75.5 & 7.5 & -56.2 & 5.8 & 15.1 + 229.715 & 0.004 & -5.6 & 6.6 & 12.4 & 7.0 & 17.8 + 229.724 & 0.006 & -7.6 & 6.6 & 10.3 & 6.1 & 12.8 + 230.731 & 0.185 & 73.7 & 7.1 & 86.2 & 5.6 & 23.1 + 230.733 & 0.186 & 69.5 & 6.9 & 88.0 & 5.5 & 24.0 + 231.736 & 0.365 & 55.6 & 5.0 & 64.9 & 4.4 & 13.7 + 231.738 & 0.366 & 52.8 & 4.8 & 61.8 & 4.4 & 10.9 + 232.736 & 0.544 & -33.8 & 6.9 & -34.2 & 5.8 & -8.2 + 232.738 & 0.544 & -35.9 & 6.4 & -36.9 & 5.8 & -10.9 + 233.723 & 0.720 & -75.2 & 7.0 & -64.6 & 3.9 & 14.9 + 233.726 & 0.721 & -80.7 & 7.2 & -68.8 & 4.5 & 10.8 + 234.730 & 0.900 & -53.2 & 6.0 & -43.7 & 5.0 & 6.2 + 234.733 & 0.901 & -52.4 & 6.0 & -39.9 & 5.3 & 9.4 + 235.726 & 0.078 & 25.1 & 7.2 & 47.4 & 4.1 & 17.3 + 235.730 & 0.079 & 26.6 & 7.1 & 46.5 & 5.0 & 15.9 + 236.737 & 0.258 & 69.3 & 7.1 & 80.6 & 6.9 & 10.7 + 236.740 & 0.259 & 68.7 & 6.8 & 77.0 & 7.3 & 6.7 + 237.738 & 0.437 & 12.0 & 7.9 & 7.4 & 5.6 & -16.3 + 238.745 & 0.617 & -64.5 & 5.5 & -60.6 & 3.4 & -4.6 + 238.746 & 0.617 & -69.0 & 7.0 & -65.2 & 4.2 & 0.0 + 238.748 & 0.618 & -71.4 & 6.9 & -75.4 & 5.2 & -19.1 + 241.722 & 0.149 & 51.4 & 5.5 & 60.8 & 4.0 & 5.4 + 242.704 & 0.324 & 54.8 & 5.3 & 60.0 & 4.0 & -2.1 + 242.705 & 0.324 & 56.7 & 5.9 & 61.6 & 3.4 & -0.5 + 243.721 & 0.506 & -2.8 & 5.4 & 2.9 & 4.9 & 11.1 + 243.722 & 0.506 & -4.9 & 5.0 & -7.8 & 2.8 & 0.4 + 253.726 & 0.292 & 72.0 & 9.8 & 49.5 & 5.9 & -17.9 + 253.727 & 0.292 & 74.2 & 10.2 & 44.1 & 5.4 & -23.4 + 254.721 & 0.470 & 8.4 & 9.9 & -7.3 & 6.3 & -16.0 + 254.721 & 0.470 & 12.0 & 10.3 & -7.9 & 6.3 & -16.6 + 255.673 & 0.640 & -64.6 & 6.3 & -70.6 & 5.0 & -7.1 + 255.674 & 0.640 & -63.2 & 6.3 & -67.0 & 6.5 & -3.5 + 255.675 & 0.640 & -68.0 & 5.9 & -71.1 & 4.9 & -7.5 + & variable period model & fixed period model + + @xmath17 ( days ) & 5.5997(1 ) & 5.5998(1 ) + p & 3.8(4 ) x 10@xmath8 & 0 ( fixed ) + @xmath18 ( jd 2,440,000 + ) & 1869.10(5 ) & 10235.29(1 ) + @xmath10 ( km s@xmath16 ) & -5.4(1 ) &
-5.4(1 ) + k ( km s@xmath16 ) & 75.53(15 ) & 75.48(15 ) + @xmath5 & 4231 & 4283 + * figure 1 : * typical rectified spectra of hde226868 ( top ) and the radial velocity standard 19 cep ( bottom ) obtained with the la palma 2.5 m int .
the spectral types are very similar , but note the heii @xmath24686 emission in hde226868 .
|
we present epoch 1996 , high - quality radial velocity data for hde 226868 , the optical counterpart of cygnus x-1 . combining our results with all published historical data , we have derived a new ephemeris for the system of hjd2450235.29 @xmath0 , which allows accurate orbital phase calculations to be made for any x - ray observations over the last 30 years .
we find no evidence for any period change as has been suggested by ninkov , walker & yang ( 1987 ) .
we discuss the shortcomings of previous work in establishing the period and orbital elements .
# 1*#1 * x - rays : stars binaries : close stars : individual : cyg x-1 accretion , accretion discs
|
choptuik s discovery of critical phenomena in the gravitational collapse of a scalar field @xcite sparked a surge of interest in gravitational collapse just at the threshold of black hole formation .
the discovery of critical behavior in several other matter models quickly followed .
despite the fact that the evolution equations are very complex and highly non - linear , the dynamics of the near - critical field evolution is relatively simple and , in some important aspects , universal .
the critical solution , which depends on the matter model only , serves as an intermediate attractor in the phase space of solutions , and often has an additional peculiar symmetry called self - similarity .
the mass of the black hole produced in supercritical evolution scales as a power law @xmath0 with parameter @xmath1 describing initial data , and mass - scaling exponent @xmath2 is dependent only on the matter model , but not on the initial data family . an interesting consequence of mass scaling which has direct bearing on the cosmic censorship conjecture is the fact that arbitrarily small black holes can be produced in near - critical collapse , with the critical solution exhibiting a curvature singularity and no event horizon .
the explanation of the universality of the critical behavior lies in perturbation analysis and renormalization group ideas .
it turns out that critical solutions generally have only one unstable perturbation mode , making them the most important solutions for understanding the dynamics of field evolution , after the stable ones ( flat space and schwarzschild or kerr - newman black hole ) .
as the near - critical field configuration evolves , all its perturbation modes decay , losing information about the initial data and bringing the solution closer to critical , except the one growing mode which will eventually drive the solution to black hole formation or dispersal , depending on its content in the initial data .
thus the critical solution acts as an intermediate attractor ( of codimension one ) in the phase space of field configurations .
finding the eigenvalue of the growing perturbation mode allows one to calculate important parameters of the critical evolution , the mass - scaling exponent in particular .
an important question is how generic the critical behavior is with respect to initial data , or , in phase space language , how big is the basin of attraction of the critical solution .
so far most of the work on critical gravitational collapse , numerical or analytic , has been restricted to the case of spherical symmetry , simply because of the enormous difficulties in treating fully general non - symmetric solutions of einstein equations .
a natural concern is whether the critical phenomena observed so far are limited to spherical symmetry , and whether the evolution of non - spherical data will lead to the same results .
the numerical study of abrahams and evans on axisymmetric gravitational wave collapse and recent numerical perturbation calculations by gundlach @xcite give numerical evidence for the claim that critical phenomena are not restricted to spherical symmetry , and that the critical solutions are indeed attractors in the full phase space .
in this paper we search for analytical evidence to support that claim .
one of the few known closed form solutions related to critical phenomena is the roberts solution , originally constructed as a counterexample to the cosmic censorship conjecture @xcite , and later rediscovered in the context of critical gravitational collapse .
it is a continuously self - similar solution of a spherically symmetric gravitational collapse of a minimally coupled massless scalar field .
while it is not a proper critical solution , as it has more than one growing mode @xcite , it is still a good ( and simple ) toy model of the critical collapse of the scalar field .
this paper considers fully general perturbations of the roberts solution in a gauge - invariant formalism . due to the symmetries of the background , the linear perturbation equations decouple and the variables separate , so an exact analytical treatment is possible .
we find that there are no growing perturbation modes apart from spherically symmetric ones described earlier @xcite .
so all the non - sphericity of the initial data decays in the collapse of the scalar field , and only the spherically symmetric part will play a role in the critical behavior . to our knowledge , this is the first paper to obtain analytical results on non - spherical critical collapse .
the spacetime we will use as a background in our calculations is a continuously self - similar spherically symmetric solution of the gravitational collapse of a massless scalar field ( the roberts solution ) .
the einstein - scalar field equations @xmath3 can be solved analytically in spherical symmetry by imposing continuous self - similarity on the solution , i.e. by assuming that there exists a vector field @xmath4 such that @xmath5 where @xmath6 denotes lie derivative .
self - similar solutions form a one - parameter family , which is most easily derived in null coordinates .
the critical solution is given by the metric @xmath7 where @xmath8.\ ] ] the global structure of the critical spacetime is shown in fig .
[ fig : roberts ] .
the influx of the scalar field is turned on at the advanced time @xmath9 , so that the spacetime is minkowskian to the past of this surface .
the initial conditions for the field equations ( [ eq : r ] ) and ( [ eq : box ] ) are specified there by the continuity of the solution .
it is instructive to rewrite roberts solution in new coordinates so that the self - similarity becomes apparent . for this purpose
we introduce scaling coordinates @xmath10 , \hspace{1em } s = - \ln(-u),\ ] ] with the inverse transformation @xmath11 the signs are chosen to make the arguments of the logarithm positive in the region of interest ( @xmath12 , @xmath13 ) , where the field evolution occurs . in these coordinates the metric ( [ eq : metric ] ) becomes @xmath14 + r^2\ , d\omega^2,\ ] ] and the critical solution ( [ eq : crit ] ) is simply @xmath15 observe that the scalar field @xmath16 does not depend on the scale variable @xmath17 at all , and the only dependence of the metric coefficients on the scale is through the conformal factor @xmath18 .
this is a direct expression of the geometric requirement ( [ eq : ss ] ) in scaling coordinates ; the homothetic killing vector @xmath4 is simply @xmath19 .
to avoid complicated gauge issues of fully general perturbations , we will use the gauge - invariant formalism developed by gerlach and sengupta .
this formalism describes perturbations around a general spherically symmetric background @xmath20 which in our case we take to be the roberts solution ( [ eq : metric ] ) . here and
later capital latin indices take values @xmath21 , and lower - case latin indices run over angular coordinates . @xmath22 and @xmath23 are defined on a spacetime two - manifold , while @xmath24 is the metric of the unit two - sphere .
because the background spacetime is spherically symmetric , perturbations around it can be decomposed in spherical harmonics .
scalar spherical harmonics @xmath25 have even parity under spatial inversion , while vector spherical harmonics @xmath26 have odd parity .
we will only concern ourselves with even - parity perturbations here , since odd - parity perturbations can not couple to scalar field perturbations .
we will focus on non - spherical perturbation modes ( @xmath27 ) , as the spherically symmetric case ( @xmath28 ) was studied earlier @xcite . for clarity , angular indices @xmath29 and the summation over all harmonics will be suppressed from now on .
the most general even - parity metric perturbation is @xmath30 dx^a dx^b,\end{aligned}\ ] ] and the scalar field perturbation is @xmath31 as you can see , metric perturbations are described by a two - tensor @xmath32 , a two - vector @xmath33 , and two two - scalars @xmath34 and @xmath35 ; the scalar field perturbation is described by a two - scalar @xmath36 . however , these perturbation amplitudes do not have direct physical meaning , as they change under the ( even - parity ) gauge transformation induced by the infinitesimal vector field @xmath37 one can construct two gauge - invariant quantities from the metric perturbations @xmath38 and one from the scalar field perturbation @xmath39 where @xmath40
only gauge - invariant quantities have physical meaning in the perturbation problem .
all physics of the problem , including the equations of motion and boundary conditions , should be written in terms of these gauge - invariant quantities . once gauge - invariant quantities have been identified ,
one is free to convert between gauge - invariant perturbation amplitudes and their values in whatever gauge choice one desires .
we will work in longitudinal gauge ( @xmath41 ) , which is particularly convenient since perturbation amplitudes in it are just equal to the corresponding gauge - invariant quantities .
the above condition fixes the gauge uniquely for non - spherical modes .
( there is some gauge freedom left over for the @xmath28 mode , but remember that we are only concerned with higher @xmath42 modes . )
expressions for the components of the linear perturbation equations @xmath43 for a fully general perturbation in longitudinal gauge are collected in appendix [ sec : long ] . by inspection of the @xmath44 component of the equations ,
it is clear that the equations of motion require that @xmath45 for @xmath27 . with the change of notation @xmath46 and @xmath47 , the remaining equations of motion for non - spherical modes
are [ eq : pert ] @xmath48 @xmath49 @xmath50 @xmath51 @xmath52 @xmath53 @xmath54 equation ( [ eq : pert : box ] ) comes from the scalar wave equation , and equations ( [ eq : pert : r : uu][eq : pert : r : vt ] ) are the @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath59 , and @xmath60 components of the einstein equations , correspondingly .
as usual with a scalar field , the system ( [ eq : pert ] ) has one redundant equation , so equation ( [ eq : pert : r : uv ] ) is satisfied automatically by virtue of other equations
. equations ( [ eq : pert : r : ut ] ) and ( [ eq : pert : r : vt ] ) are constraints , and the remaining four equations are dynamic equations for four perturbation amplitudes @xmath61 , @xmath62 , @xmath34 , and @xmath36 .
boundary conditions for the system ( [ eq : pert ] ) are specified at @xmath9 and the spatial infinity .
continuity of matching with flat spacetime at the hypersurface @xmath9 requires the vanishing of the perturbations there .
we also require well - behavedness of the perturbations at @xmath63 and @xmath64 , so that the perturbation expansion holds .
thus , the boundary conditions are @xmath65 equations ( [ eq : pert ] ) together with boundary conditions ( [ eq : pert : bc ] ) constitute our eigenvalue problem .
it is possible to decouple the dynamic equations ( [ eq : pert : box][eq : pert : r : tt ] ) by combining them with the constraints ( [ eq : pert : r : ut ] ) and ( [ eq : pert : r : vt ] ) , and their first derivatives .
after somewhat cumbersome algebraic manipulations , which we will not show here , the system of linear perturbation equations ( [ eq : pert ] ) can be rewritten as [ eq : uv ] @xmath66 @xmath67 @xmath68 @xmath69 @xmath70 @xmath71 this decoupled system of partial differential equations can be further simplified by exploiting continuous self - similarity of the background to separate spatial and scale variables . with this intent , we rewrite equations ( [ eq : uv ] ) in terms of the scaling coordinates ( [ eq : coord : xs ] ) [ eq : xs ] @xmath72 @xmath73 @xmath74 @xmath75
@xmath76 @xmath77 we decompose the perturbation amplitudes into modes that grow exponentially with the scale @xmath17 ( which amounts to doing laplace transform on them ) @xmath78 the summation runs over the perturbation mode eigenvalues @xmath79 , which could , in general , be complex .
modes with @xmath80 grow and are relevant for critical behavior , while modes with @xmath81 decay and are irrelevant .
the growing perturbation mode amplitudes vanish at @xmath82 , so the boundary condition at @xmath63 is satisfied automatically . for clarity , the perturbation mode subscript @xmath79 and the explicit summation over all modes
will be suppressed from now on , so henceforth @xmath36 , @xmath61 , @xmath62 , and @xmath34 will mean @xmath83 , @xmath84 , @xmath85 , and @xmath86 for the mode with eigenvalue @xmath79 .
the decomposition ( [ eq : laplace ] ) converts the system of partial differential equations ( [ eq : xs ] ) into a system of ordinary differential equations , which is much easier to analyze : [ eq : x ] @xmath87
@xmath88 @xmath89 @xmath90 @xmath91 @xmath92 the prime denotes a derivative with respect to spatial variable @xmath93 .
these equations can be converted into standard algebraic form by the change of variable @xmath94 so that the system ( [ eq : x ] ) becomes [ eq : y ] @xmath95 \dot{\phi } - [ 3\kappa/2 + l(l+1)/2 ] \phi = 0,\ ] ] @xmath96 \dot{u } - [ -\kappa/2 + l(l+1)/2 ] u = 0,\ ] ] @xmath97 \dot{v } - [ 3\kappa/2 + l(l+1)/2 ] v = 0,\ ] ] @xmath98 \dot{k } - [ \kappa/2 - 1 + l(l+1)/2 ] k = 2 \phi + v,\ ] ] @xmath99 @xmath100 the dot denotes a derivative with respect to @xmath101 , and we redefined the scalar field perturbation amplitude as @xmath102 to cast the equations into standard table form .
the boundary conditions ( [ eq : pert : bc ] ) are @xmath103 imposed on the decoupled system of ordinary differential equations ( [ eq : y ] ) , these boundary conditions give an eigenvalue problem for the perturbation spectrum @xmath79 .
in the previous section we formulated an eigenvalue problem for the spectrum of non - spherical perturbations of the critical roberts solution . we now proceed to solve it .
observe that equations ( [ eq : y : f][eq : y : k ] ) governing the dynamics of the perturbations are hypergeometric equations of the form @xmath104 \dot{x } - ab x = 0.\ ] ] equation ( [ eq : y : k ] ) is not homogeneous , but we will deal with that shortly .
the hypergeometric equation coefficients are different for equations describing the perturbations @xmath105 , @xmath61 , @xmath62 , and @xmath34 ; they are summarized in the table below .
@xmath106 hypergeometric equations have been extensively studied ; for complete description of their properties see , for example , .
hypergeometric equation ( [ eq : hypergeom ] ) has three singular points at @xmath107 , and its general solution is a linear combination of any two different solutions from the set @xmath108 where @xmath109 is the hypergeometric function , which is regular at @xmath110 and has @xmath111 .
any three of the functions ( [ eq : solns ] ) are linearly dependent with constant coefficients . in particular , @xmath112 the functions @xmath113 , @xmath114 are appropriate for discussing the behavior of solution near @xmath115 , while @xmath116 , @xmath117 give the behavior at infinity @xmath118 as we said before , imposing the boundary conditions ( [ eq : y : bc ] ) on solutions of equations ( [ eq : y ] ) leads to a perturbation spectrum .
we will now investigate what restrictions the boundary conditions place on the hypergeometric equation coefficients .
the vanishing of perturbation amplitudes at @xmath115 rules out @xmath113 as a component of the solution and requires that @xmath119 to make @xmath114 go to zero .
the solution @xmath114 has non - zero content of both @xmath116 and @xmath117 by virtue of ( [ eq : connection ] ) , hence for it to be bounded at infinity , both @xmath120 and @xmath121 must be positive to guarantee convergence of @xmath116 and @xmath117 .
so , unless there is degeneracy , the boundary conditions translate to the following conditions on the hypergeometric equation coefficients : [ eq : bc : coeff ] @xmath122 we are now ready to take on system ( [ eq : y ] ) .
take a look at equation ( [ eq : y : v ] ) for @xmath62 .
condition ( [ eq : bc : coeff:1 ] ) for it is @xmath123 , i.e. there are no growing @xmath62 modes ! with the amplitude of relevant @xmath62 perturbation modes being zero , the constraints ( [ eq : y : ut ] ) and ( [ eq : y : vt ] ) become @xmath124 and right hand side of equation ( [ eq : y : k ] ) can be absorbed by the left hand side , making the equation for @xmath34 homogeneous ( with @xmath125 ) .
indeed equations ( [ eq : y : k ] ) and ( [ eq : y : f ] ) for @xmath34 and @xmath105 are just derivatives of equation ( [ eq : y : u ] ) for @xmath61 @xmath126 \dot{u } - [ -\kappa/2 + l(l+1)/2 ]
u = 0,\ ] ] which is the homogeneous hypergeometric equation with coefficients @xmath127 imposing the boundary condition at @xmath115 for the solution of the above equation and its derivatives , which behave like @xmath128 produces restriction on the non - spherical mode eigenvalue @xmath129 which is the strongest of restrictions ( [ eq : bc : coeff:1 ] ) for equations for @xmath61 , @xmath34 , and @xmath105 .
but then @xmath130 and hence @xmath120 and @xmath121 can not be both positive , and so the boundary condition at infinity can not be satisfied . a more careful investigation of degenerate cases of relation ( [ eq : connection ] )
shows that the contradiction between boundary conditions at @xmath115 and infinity still persists if @xmath131 .
it can only be resolved by the trivial solution @xmath132 .
thus we have shown that there are no growing non - spherical perturbation modes around the critical roberts solution .
in fact , an even stronger statement is true .
the contradiction between boundary conditions at @xmath115 and infinity can not be resolved by a non - trivial solution so long as @xmath131 , i.e. so long as @xmath133 .
hence non - trivial non - spherical perturbation modes of critical roberts solution must decay faster than @xmath134 .
in this paper we used the gauge - invariant perturbation formalism to explore the critical behavior in the gravitational collapse of a massless scalar field .
perturbing around a continuously self - similar critical solution ( the roberts solution ) , we obtained an eigenvalue problem for the spectrum of perturbations .
the remarkable feature of this model of critical scalar field collapse is that it allows an exact analytical treatment of the perturbations as well as of the critical solution , due to the highly symmetric background .
an exact analysis of the perturbation eigenvalue problem reveals that there are no growing non - spherical perturbation modes . however , there are growing spherical perturbation modes .
their spectrum is continuous and occupies a big chunk of the complex plane @xcite . in view of these findings ,
the following picture of dynamics of scalar field evolution near self - similarity emerges : as we evolve generic initial data which is sufficiently close to the critical roberts solution , non - spherical modes decay and the solution approaches the spherically symmetric one .
asymmetry of the initial data does not play a role in the collapse .
the growing spherical modes , on the other hand , drive the solution farther away from the continuously self - similar one . in this sense ,
the critical roberts solution is an intermediate attractor for non - spherical initial data .
an interesting question , which is not answered by perturbative calculations , is the further fate of the scalar field evolution as it gets away from the roberts solution . in all likelihood
, it evolves towards the discretely self - similar choptuik solution , which is a local attractor of lower codimension ( one ) , as the continuous self - similarity is broken by oscillatory growing modes . after staying near the choptuik solution for a while ,
the scalar field will eventually disperse or settle into a black hole , with these final states being global attractors in the phase space of field configurations .
this evolution from attractor to attractor in phase space is somewhat analogues to a ball rolling down the stairs , going from a step to a lower step , until it reaches the bottom .
the results of this paper shed some light at the complicated problem of critical collapse of generic initial data from the analytical viewpoint , confirming the hypothesis that critical phenomena are not restricted to spherical symmetry .
investigation of the fate of a scalar field solution as it breaks away from continuous self - similarity , as outlined above , will further our understanding of the dynamics of scalar field collapse , and presents an interesting ( and challenging ) analytical problem .
numerical simulations might also help to establish a clearer picture of near - critical scalar field evolution .
this research was supported by the natural sciences and engineering research council of canada and by the killam trust .
in this appendix we collect expressions for components of the perturbed einstein - scalar equations calculated in longitudinal gauge ( @xmath135 ) .
the perturbed metric in longitudinal gauge is @xmath136 and the perturbed scalar field is @xmath137 + f y.\ ] ] the einstein equations for scalar field are equivalent to the vanishing of the tensor @xmath138 .
its non - trivial components , calculated to the first order in the perturbation amplitude using the above metric and scalar field , are
|
this paper studies the non - spherical perturbations of the continuously self - similar critical solution of the gravitational collapse of a massless scalar field ( the roberts solution ) .
the exact analysis of the perturbation equations reveals that there are no growing non - spherical perturbation modes .
|
superfluid states in the nuclear environment have been extensively studied as they have a significative role in several physical processes , such as the structure of nuclei out the stability valley and the cooling dynamics of proto - neutron stars . + a variety of models and approximations
have been used for this purpose @xcite , mainly non - relativistic potentials or effective forces , such as the density - dependent skyrme or gogny ones .
it is not unusual a mixed treatment that combines both schemes in order to simplify involved calculations @xcite .
+ approximately two decades ago a covariant model of the field theory of hadrons , generally known as quantum hadro - dynamics ( qhd ) @xcite , was used for the first time to study nuclear matter superfluidity @xcite .
there are several reasons to use this theoretical framework to deal with nuclear superfluidity , in first place there are practical reasons , some self - consistent calculations are more simply stated and easily solved .
this property gave rise to a version of qhd , known as density dependent hadron field theory @xcite which casts dirac - brueckner outputs in the qhd language .
in second place one must consider field theory as a more adequate tool to make contact with the fundamental theory of strong interactions and the fact that covariant formulations are desirable for astrophysical applications , among others conceptual reasons .
furthermore , the formalism has the versatility to include vacuum effects and finite renormalizations in a coherent way @xcite . +
since qhd models are formulated as a many - body theory , one of its basic premises is the reproduction of the nuclear matter saturation properties .
this can be fulfilled with a few adjustable parameters and even at the lowest order of approximation . once the free - parameters have been fixed , the qhd model has a noticeable prediction power .
+ although a wide spreading of numerical results for the superfluid phase in infinite nuclear matter can be found in the literature , there is a qualitative agreement that the superfluid gap should not exceed 3 mev , and it should vanish for densities around the saturation density @xcite . unfortunately , the findings of @xcite for superfluid nuclear matter in the @xmath0 phase do not agree with these expectations .
a maximum @xmath2 mev was found there , and only an unphysical reduction of 15% in the omega - meson mass yield results comparable with currently accepted values .
+ this situation has not been changed substantially after the evaluation of different corrections into the original scheme .
however , the good properties of this treatment have motivated mixed descriptions combining qhd models and conventional potentials @xcite . + more recently it was claimed that a coherent inclusion of meson proper self - energy @xcite , or adjustable quenching factor @xcite , could bring numerical calculations to the likely values .
further studies about the effect of corrections of the meson propagators and the influence of low density instabilities on the superfluid gap can be found for instance in @xcite .
+ it was stressed in reference @xcite that the high momenta behavior of the repulsive potential has a crucial role in the exceedingly large values obtained for @xmath3 . it must be noticed that the relative strength of repulsive and attractive contributions have been calibrated for momenta below the fermi surface in order to produce the saturation mechanism
. therefore it would be desirable a pairing potential which preserves the relative strength of its components in the fermi sphere , but having a repulsive component decreasing as faster as the attractive one in the high momenta domain .
+ taking these facts into account we try in the present work , to obtain an effective and concise model able to deal with the superfluid phase of nuclear matter . for this purpose
we use and compare several models of relativistic nuclear fields interacting through scalar and vector mesons . as a first approach
, we reduce to its minimal expression the complexity of the nuclear interaction , but further refinements can be considered .
we introduce a characteristic length scale in the effective interaction , which could eventually be traced back to the confining mechanism of the fundamental theory of strong interactions .
a similar approach was applied in the past to describe heavy ion collisions @xcite as well as nuclear matter properties @xcite .
it has been known as finite volume correction since it takes into account the spread of the nucleon localization .
the relevance of this effect upon the evaluation of some bulk properties of the nuclear matter has been stressed long time ago @xcite .
+ in the next section we present the theoretical deduction of the general gap equation at finite temperature in a context of landau - fermi liquid .
the last part of this section is devoted to the theoretical deduction of the self - consistent expression defining the superfluid gap for the @xmath0 phase within the qhd framework .
in section 3 we show the numerical results and discuss them .
finally , conclusions are drawn in section 4 .
the different superfluid phases in a fermionic system can be described in a general and compact way within the formalism of the fermi liquid , using for instance the formalism of reference @xcite .
it is assumed that the low - lying excitations of the system are represented in terms of quasi - particles and , circumstantially , collective modes .
+ we use the notation @xmath4 for the equilibrium distribution function of a quasi - particle state , where the label @xmath5 comprise spin , isospin , and momentum quantum numbers . the fermionic contribution to every conserved quantity , such as particle number and energy ,
can be expressed in terms of a summation over @xmath4 .
+ in the following we will be interested in nucleon pairs coupled to singlet spin and triplet isospin , so that an anomalous distribution function @xmath6 and a energy gap @xmath7 are introduced . according to @xcite
, we make the decomposition @xmath8_{a\ , b } \sigma^{(2)}_{s\,s ' } , \label{gdist}\\ \delta_\beta&=&\delta_k(p)\left [ \tau^{(k ) } \tau^{(2)}\right]_{a\ , b } \sigma^{(2)}_{s\,s ' } , \label{gapdecomp}\end{aligned}\ ] ] where momentum , isospin @xmath9 , and spin @xmath10 dependencies has been clearly distinguished . here
@xmath11 stands for the pauli matrices for isospin and spin , respectively .
+ on the other hand , the normal phase is filled with quasi - particle states in a isospin duplet described similarly by @xmath12 , with @xmath13 accordingly , the quasi - particle spectra is assumed in matrix form @xmath14 .
+ in the landau theory of fermi liquids , the energy of the system @xmath15 is considered as a functional of the distribution functions @xmath16 .
first variations respect to them give the quasi - particle and gap spectra matrix components @xmath17 since @xmath18 and @xmath19 are itself functions of @xmath20 , the equations above are self - consistent relations .
+ within the block diagonalization procedure of @xcite the distribution functions are written @xmath21^{-1 } , \label{ndist}\\ k&=&\left(1 + x x^\dag \right)^{-1},\end{aligned}\ ] ] here @xmath22 indicates matrix transposition , @xmath23 , @xmath24 , and the diagonal matrix @xmath25diag@xmath26 collects the proton ( 1 ) and nucleon ( 2 ) chemical potentials .
the unknown matrix @xmath27 satisfies the condition @xmath28 we have solved this system of equations for symmetric nuclear matter coupled to @xmath29 .
we obtained @xmath30 , \nonumber\\ & & f_3=0,\ , e_\delta=\sqrt{\delta_3 ^ 2+(\varepsilon_0-\mu_0)^2 } ; \nonumber\end{aligned}\ ] ] in the first line @xmath31 stands for either the proton or the neutron chemical potential .
+ the particle density of protons ( k=1 ) or neutrons ( k=2 ) is given by + @xmath32\nonumber\ ] ] these equations are used to relate the chemical potentials to the conserved isospin and baryonic number density .
+ in the next subsection we show the model which provides the quasi - particle interaction and spectra .
models of the nuclear interaction , formulated in the covariant field theory , have been widely used in the study of the dynamics and structure of infinite matter as well as finite nuclei . since the pioneering work of references @xcite
, the simple @xmath33 model has grown in different directions and it was completed in order to cover a multitude of manifestations of the nuclear force .
+ in particular the subject of the nuclear superfluidity was first treated within this context in reference @xcite , by using the original @xmath33 model plus a pseudo - scalar pion interaction .
the scheme of approximation used there consisted in a mean field treatment of the meson and nucleon fields , a gorkov factorization of the pairing interaction and a instantaneous assumption which allows a time - independent resolution of the gap equation .
subsequently , this procedure was extended to consider the effect of vacuum , and the variation of the in - medium meson properties @xcite .
+ most of these studies agree in a excessively large value for the gap in infinite nuclear matter .
the realization of a correlated state of two nucleons is a consequence of the equilibrium between a repulsive and an attractive component of the pairing potential , originating in the exchange of virtual @xmath34-mesons and @xmath35-mesons respectively .
this mechanism is also found in the binding energy of nuclear matter .
+ the slow decrease of the repulsive potential as a function of the transferred momenta , has been pointed out as the main cause of the mismatch .
this situation can not be modified without a substantial redefinition of the couplings , which should lead to a destruction of the saturation mechanism .
+ in this work we adopt the simplest version of nucleons interacting through @xmath35 and @xmath34 mesons , whose lagrangian density is + @xmath36 where summation over the repeated isospin index @xmath37 is assumed , @xmath38 and @xmath39 are adimensional coupling constants .
the equations of motion of the classical fields are @xmath40 denoting by @xmath41 the propagators of scalar and vector - meson fields , the eqs .
( [ sigma ] ) and ( [ omega ] ) can be formally solved as @xmath42 the hamiltonian density @xmath43 is given by the canonical procedure , and the energy density of the system @xmath44 is evaluated by taking its expectation value @xmath45 under the hypothesis of static meson fields the second and third terms vanish . + inserting the meson equations of motion ( [ sigma ] ) , ( [ omega ] ) together with the eq .
( [ mesons ] ) into ( [ e1 ] ) , we obtain @xmath46 at this point we introduce an expansion of the nucleon fields in terms of quasi - particle creation and annihilation operators , similar to that of a free field @xmath47 , \label{psiquant } \\ \bar{\psi}^a_\alpha(x)&=&\int { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3p \frac{m^\ast_a}{e_a(p)}\left[b^{\dag \ , a } _ s(p ) \bar{u}^s_\alpha(p ) e^{i p_a x}+d^a_s(p ) \bar{v}^s_\alpha(p ) e^{-i p_a x}\right ] , \label{psibarquant } \\ u_a^s(p)&=&\frac{\not \ ! p_a+m^\ast_a}{\sqrt{2
m^\ast_a(m^\ast_a + e_a ( p))}}\;u^s(0),\nonumber \\ v_a^s(p)&=&\frac{\not \ ! p_a+m^\ast_a}{\sqrt{2 m^\ast_a ( m^\ast_a + e_a ( p))}}\;u^s(0 ) , \nonumber\end{aligned}\ ] ] but creation and annihilation operators are referred to the lowest energy state of correlated nucleons .
we have used the quasi - particle properties @xmath48 , @xmath49 . due to the isospin invariance of the interaction proton and
neutron properties are actually independent .
furthermore , as we are interested in isospin symmetric matter proton and neutron are indistinguishable and the isospin index @xmath37 becomes superfluous .
from now on we will omit it , and a degeneracy factor 2 will be included when necessary .
+ the quantities @xmath50 stand for the mean values of the @xmath35 and @xmath34 meson fields in homogeneous , isotropic matter .
as usual , they can be deduced from eqs .
( [ sigma ] ) , ( [ omega ] ) by neglecting derivatives and taking expectation values of the fermionic bilinears @xmath51 , @xmath52 . + as a part of the approximation we neglect in eq .
( [ dense ] ) the contribution of particle - antiparticle or antiparticle - antiparticle terms .
we define the equilibrium distribution functions for the normal and superfluid phases @xmath53 it must been taken into account that @xmath54 , whereas pairs of nucleons are assumed to couple to zero momentum , so that @xmath55 .
+ within the mean field approach , corrections to the meson propagation in the nuclear environment are dismissed , although a random phase approximation could be considered , as in @xcite . in the first case ,
we obtain @xmath56 the term proportional to @xmath57 in ( [ omegapro ] ) , produce zero contribution in integrals combining @xmath58 with nucleon fields , because of the conservation of the baryonic current .
+ we are interested in static homogeneous matter , therefore @xmath59 in eq .
( [ fouriert ] ) .
if we neglect time retardation in the meson propagation , _ i. e. _ @xmath60 , then terms containing @xmath61 are absent in eqs .
( [ sigmapro ] ) , ( [ omegapro ] ) as well as in the exponential of the fourier transform shown lines above . + all these elements together in eq .
( [ dense ] ) produce @xmath62 ^ 2-\mid \mathbf{p}-\mathbf{q } \mid^2-m_s^2+i \varepsilon}\nonumber\\ & & - 2\ , g_w^2\int { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3p\ , { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3q \frac{f(p ) f(q ) } { e(p)e(q ) } \frac{2 m^{\ast\,2}-e(p ) e(q)+\mathbf{p } \cdot \mathbf{q } } { \left[\varepsilon(p)-\varepsilon(q)\right]^2-\mid \mathbf{p}-\mathbf{q } \mid^2-m_w^2+i \varepsilon } \label{efock}\\ \mathcal{e}_s&=&-g_s^2\int { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3p\ , { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3q \,\frac{g^{\dag}(p ) g(q ) \lambda_s(p , q ) } { \left[\varepsilon(p)-\varepsilon(q)\right]^2-\mid \mathbf{p}-\mathbf{q } \mid^2-m_s^2+i \varepsilon}\;\eta(p , q)\nonumber \\ & & + g_w^2\int { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3p\ , { \stackrel{^-}{}\!\!\!\!\!\!d\,}^3q \,\frac{g^{\dag } ( p ) g(q ) \lambda_w(p , q ) } { \left[\varepsilon(p)-\varepsilon(q)\right]^2-\mid \mathbf{p}-\mathbf{q } \mid^2-m_s^2+i \varepsilon}\;\eta(p , q ) .
\label{esup}\end{aligned}\ ] ] here we have separated mean field ( @xmath63 ) , fock ( @xmath64 ) and superfluid ( @xmath65 ) contributions .
certain integrals appearing in ( [ dense ] ) vanish because of the isotropy of infinite matter .
the following notation is used @xmath66\right),\nonumber\end{aligned}\ ] ] with @xmath67 .
+ the factor @xmath68 entering in these expressions can be used for integration respect to @xmath69 . for this purpose
it is useful the distribution identity + @xmath70 denominators in the integrands of eqs .
( [ esup ] ) , ( [ efock ] ) come from the meson propagators , in particular the combinations @xmath71 come from its @xmath61 dependence . in the next step
we apply the instantaneous approximation @xcite , which result in the elimination of all these combinations .
+ within this approach eq .
( [ esup ] ) can be re - written + @xmath72 from now on , we use only the mfa and neglect the fock term .
+ the superfluid gap can be determined by using eqs .
( [ gapdecomp ] ) and ( [ spectra ] ) @xmath73 finally it must be stressed that in applying eq.([spectra ] ) for evaluating the quasi - particle spectra , the effective mass @xmath74 must be considered a functional of the distribution function @xmath18 .
standard field theory considers physical particles as structureless , point - like objects .
this could be a serious shortcoming when composed states in a dense medium are described .
it is well known that , for instance , the energy and charge density of a soliton spreads over a finite range of space @xcite .
therefore it is legitimate to assign a intrinsical length scale to nucleons immersed in a dense environment .
+ this was the argument supporting many phenomenological studies of the nuclear interaction @xcite . from a practical point of view , one can consider @xmath75 fermions of class @xmath37 distributed over a finite volume @xmath76 , then the available space for quantization is @xmath77 , where @xmath78 is the spatial extension of this state .
but the canonical procedure uses the full volume @xmath76 , this situation is corrected by introducing a correction factor @xmath79 in the second quantization of fields .
this , in turn , modifies fermion bilinears like particle number @xmath80 and normal energy density by a factor @xmath81 .
the superfluid energy density requires a correction @xmath82 , which is transferred to the formulae of the gap function .
+ the effective volume inaccessible for other particles due to the presence of a spherical object of radius @xmath83 is @xmath84 the parameter @xmath83 can be understood in a simple minded model as the geometrical size of a particle .
actually it introduces into the model a characteristic scale of the strong interaction , i. e. the spatial spreading of a bounded state of quarks and gluons . the value selected for
this radius must be compatible with similar lengths adopted in hybrid models of the nuclear interaction , see for instance @xcite .
returning to the schematic picture , the factor @xmath5 takes account of the fact that the inaccessible volume exceeds the actual size of each particle and it depends on the spatial arrangement adopted by the collection of objects .
the minimal volume configuration for identical particles corresponds to a face centered cubic arrangement . in such a case
is @xmath85 , which is the value adopted in the present calculations . + it is worthwhile to mention that the normalization of the nucleon field with an excluded volume coefficient is not equivalent to the introduction of a hard - core potential .
the normalized nucleon field interacts dynamically with the meson fields , both scalar and vector .
the in - medium properties of protons and neutrons , as well as the meson fields configuration arise simultaneously from this interaction .
the sigma meson gives rise to the attractive channel of the nuclear force , whereas the omega meson is responsible for the repulsive component .
therefore , the proposed normalization affects both attractive and repulsive channels .
this fact is evident from eq .
( [ sigma ] ) and the discussion given above .
it should be clear that the mean - field value of the sigma meson is strongly affected by the normalization of the nucleons .
moreover , the relation is highly non - linear .
furthermore , the treatment of the mesons is not symmetrical since the omega meson is coupled to a conserved charge . in consequence the omega meson mean field value
is completely determined by the conserved baryonic density .
the sigma meson mean field value , instead , come forth the hadronic dynamics .
+ since @xmath86 introduces an explicit dependence upon the baryonic densities , the quasi - particle energy @xmath87 , see eq .
( [ spectra ] ) , gets an extra term + @xmath88 it must be noticed that the additional term depends on density and temperature , but not on the momentum .
the model of nucleons and mesons with finite volume corrections has several parameters , we used @xmath89 mev for the mass of the degenerate nucleons , and @xmath90 mev for the omega - meson mass .
the sigma - meson mass has been fixed at @xmath91 mev , in agreement with @xcite .
a discussion about variation of @xmath92 can be found in @xcite .
+ the length scale @xmath83 is not determined by the model , therefore we consider it as a constant value ranging between 0.5 fm and 0.9 fm . +
the coupling constants @xmath93 are fixed in order to reproduce the saturation properties of symmetric nuclear matter , the binding energy @xmath94 mev and a saturation density corresponding to the fermi momentum @xmath95 @xcite . to have a look of the performance of qhd models in describing the nuclear superfluidity , we have examined two different models of the nucleon - meson interaction . in first place we study the density dependent coupling model of @xcite , which translate dirac - bruckner calculations with bonn a potential into the covariant field theory .
we have used the rational function parametrization given there for the couplings , obtaining a maximum value @xmath96 mev at @xmath97 . as another qhd example
we take the non - linear meson - nucleon model of @xcite , it predicts a lower compressibility and higher effective mass in the mfa than the @xmath33 model @xcite does . in our calculations
we get @xmath98 mev at @xmath99
. + [ table ] .the gap function on the fermi surface and the quotient of coupling constants for several values of the characteristic nucleon size @xmath100 . [ cols="^,^,^",options="header " , ] as the next step we consider the effects of the finite extension of nucleons , taking @xmath100 as an adjustable parameter .
the results are summarized in table 1 , where the treatments regarding the finite extension of nucleons ( fec ) or neglecting it ( nfec ) have been distinguished .
there is a sensible reduction of @xmath101 , between 30% and 40 % , respect to the previous calculations .
a non - monotonous dependence on @xmath100 is obtained , the lowest value @xmath102 mev is reached for @xmath103 fm at @xmath104 @xmath105 . + choosing @xmath106 for subsequent calculations ,
we compare the momentum dependence of the gap function for a fixed density @xmath97 .
results are shown in fig.1 .
the square points show the magnitude of the gap at the fermi surface .
a comparison of the two cases nfec and fec , shows that the latter yields the lowest absolute value for @xmath107 gev .
a reduction of almost @xmath108 is registered at the fermi surface in the fec case . in fig .
[ fig2 ] we display @xmath109 as a function of the density , here one can see that the fec treatment not only lowers @xmath101 , but also reduces more than 20% the range of densities where the gap is effective . in this case
@xmath110 @xmath105 is the upmost value for non - zero pairing gap .
+ to improve understanding of this outcome , we investigate separately contributions of scalar and vector character to the gap function .
the attractive ( @xmath111 ) and repulsive ( @xmath112 ) potentials evaluated at the fermi surface @xmath113\nonumber \\ v_w(q)&= & -\frac{g_w^2}{8}\frac{2 e_f e(q)-m^2}{p_f \,q\ , e_f\ , e(q)}\ ; \ln\left(\frac{(p_f - q)^2+m_w^2}{(p_f+q)^2+m_w^2 } \right)\nonumber\end{aligned}\ ] ] where @xmath114 , have been defined in order that eq .
( [ snmgap ] ) can be simplified to @xmath115 .
\nonumber\ ] ] as can be seen in figure [ fig3 ] , both @xmath111 and @xmath112 appears diminished in the fec case , with a stronger suppression of @xmath111 as compared to @xmath112 in the high momenta regime .
therefore @xmath116 asymptotically goes to zero faster in fec than in nefc , minimizing contributions from higher momenta .
+ on the other hand , for @xmath117 @xmath105 the attractive potential is stronger than the repulsive one in fec than in nfec approaches .
+ there is a surprising numerical similitude between the values described above for @xmath3 and some findings in @xcite , where a sophisticated evaluation of in medium effects over the @xmath0 pairing is presented , within the brueckner theory of nuclear matter .
the bcs gap function is found to have a maximum value @xmath118 mev in symmetric nuclear matter at @xmath119 .
the upper limit for the existence of the gap is given by @xmath120 @xmath105 .
+ another interesting comparison can be made with the results found by matsuzaki et al .
@xcite . that work , as the present approach ,
evaluate the possibility of giving a unified description of both particle - particle and particle - hole channels . in that case the sigma - omega model is used in the mean field approximation , and the results are rendered physically acceptable by a direct intervention over the integrals in momentum space .
this is achieved by the introduction of a form factor depending on a single parameter , which is adjusted to obtain the best fit to microscopic calculations .
it must be pointed out that the correction factor @xmath86 used in our procedure can not be rigorously considered as a form factor , since for a given fermi momentum it reduces to a constant value and therefore it does not modify the integrals . despite the procedural differences , our results are comparable to those of @xcite . for instance , if we consider the momentum dependence of the single - particle potential @xmath121 , evaluated at @xmath122 such that a maximum of @xmath3 is obtained , it takes values -4 mev @xmath123 3 mev in @xcite . in our calculations , instead , we found -5 mev @xmath124 1 mev , see fig .
3 . furthermore
, the asymptotic behavior is similar in both calculations .
more appreciable differences are found for the momentum dependence of the gap corresponding to the same @xmath122 . at very low momentum
we have @xmath125 mev , then as the transfer momentum @xmath126 is increased , the gap decreases smoothly , passes through zero at @xmath127 @xmath105 , reaches a minimum value of -1 mev at @xmath128 @xmath105 , and finally tends asymptotically to zero from negative values , see fig .
1 . as can be seen in fig .
3b of @xcite , the low momentum gap is sensibly lower @xmath129 mev , the gap passes trough zero at a lower value @xmath130 @xmath105 and reaches the same minimum value but at @xmath127 @xmath105 . from this observations
, we can conclude that in our approach , the contributions coming from momenta @xmath131 mev are overestimated as compared with the treatment of @xcite .
consequently , the behavior of @xmath3 has a maximum value that exceeds by 2.5 mev the results shown in fig .
2a of @xcite .
however , this maximum value is reached for @xmath132 @xmath105 in both cases .
the spreading is also similar , a drop of about 75@xmath133 is verified at @xmath134 @xmath105 in both calculations . + as a last application we study the temperature behavior of the gap function evaluated at the fermi momentum in the fec approach . in fig .
[ fig4 ] we select some definite values of the particle density such that the gap has magnitude higher than 0.1 mev at zero temperature . for the lower densities a steep fall
is registered around @xmath135 mev .
phase in terms of the temperature for several densities . ]
in this work we have studied the superfluid phase in a nuclear environment , within the relativistic field theory of hadrons .
the fermi - landau liquid framework , as stated by @xcite , has been used in order to obtain expressions for the @xmath0 superfluid gap energy and the distribution functions for both normal and superfluid phases .
this is equivalent to a hartree approximation , which provides a quasi - particle picture of nucleons dressed by the mesonic interaction , solved in a self - consistent way with a bcs scheme for the superfluid phase .
+ it is well known that bcs approaches in terms of qhd models , produce excessively large values for the gap in isospin symmetric nuclear matter . in this work we explored the possibility to state a easy to handle formalism , capable to retain the good properties of the qhd theory but modifying the above mentioned failure . with this purpose
we took the simplest sigma - omega interaction and we introduced short range correlations inspired by the finite volume extension of nucleons . in our approach an additional model parameter is required , which is related to the characteristic length of the finite size of nucleons .
+ we have found a significative reduction , between @xmath136 , of the @xmath0 gap in symmetric matter .
so , the results obtained are closer to non - relativistic estimates .
the modifications proposed modify the high momenta performance producing a pairing potential which goes to zero faster than in @xcite .
further refinements , such as fock term , additional mesons , non - linear sigma terms , etc . , could improve the agreement with expected values .
+ finally , the temperature behavior of the superfluid gap obtained is comparable with previous estimates @xcite , major differences correspond to the lowest densities .
this work was partially supported by the conicet , argentina .
|
properties of the @xmath0 superfluid phase are studied for symmetric nuclear matter at finite temperature .
it is described within a covariant hadronic field model , of the @xmath1 type , with addition of density dependent correlations simulating effects due to finite extension of nucleons .
the model is solved in a selfconsistent hartree - bogoliubov approach , assuming instantaneous interactions in the superfluid phase .
a comparison with the results obtained from several hadronic field models is made .
main characteristics of our description of the superfluid gap are in qualitative agreement with some studies using microscopic potentials , although further refinements could improve its performance .
|
the orion kl nebula , that lies at a distance of 414@xmath27 pc @xcite is a focal source for studies of the physics and chemistry of high mass star forming regions .
its rich molecular diversity and complex physical structure have been revealed by a number of spectral line surveys and interferometric studies of the region ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
recently , a broadband spectral line survey of orion kl carried out with the herschel / hifi spectrometer @xcite has been obtained as part of the hexos key program .
the unprecedented 1.2 thz wide frequency coverage of this survey has allowed for strong constraints on physical parameters and molecular abundances due to the detection of many transitions for each molecule , with a wide range of excitation conditions , revealing the presence of thermal gradients ( e.g. * ? ? ? * ) . among the molecular components associated with orion kl lie the so - called hot core and compact ridge .
although the precise nature of these regions remains enigmatic , evidence has been recently provided that they may have originated from the interaction of the remnants of a recent explosion with ice grains mantles ( @xmath3720 years ago , e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , leading to the evaporation and the heating of organic material @xcite .
deuterium fractionation is intrinsically a low - temperature process .
therefore , deuterated molecules in warm ( t@xmath4100k ) region , such as orion - kl , offer a window into its physical conditions at the earlier , colder stage ( t@xmath51030 k ) when this material likely formed @xcite .
numerous deuterated species have been observed toward orion kl : deuterated water @xcite , deuterated ammonia @xcite , deuterated formaldehyde @xcite and deuterated methanol @xcite .
recently , @xcite used the hifi survey , and tens of transitions , to set strong constraints in the d / h ratio of organics associated with orion kl .
their finding strongly suggests that d / h fractionation is inhomogenuous in this source . in this study
, we extend the hifi study of deuterated species performed by @xcite ; especially since the herschel observations are spatially unresolved , so even with well determined column densities within the beam , there can be large deviations in the deuterium fractionation as a function of spatial position . in section [ sec : observations ] , we present the sma observations to complement the hifi analysis . in particular , we focus on formaldehyde ( h@xmath1co and hdco )
. the d / h ratios of this molecule is key to obtaining a full picture of the oxygen chemistry in this region , as formaldehyde is believed to be an intermediate in the grain surface synthesis of methanol through sequential reactions of h or d atoms with co ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
alternatively , gas phase chemistry might independently lead to the formation of formaldehyde via reactions involving ch@xmath6d@xmath7 products under warmer conditions ( typically t@xmath8 50k , see * ? ? ?
* ; * ? ? ?
incidentally , the tentative detection of the ch@xmath6d@xmath7 ion toward orion - kl by @xcite suggests that gas phase chemistry might be efficient for the hdco formation . in this analysis
we present the first high angular resolution image of the spatial distribution of ( deuterated ) formaldehyde in orion kl and discuss the different chemical pathways that can be involved in its formation .
more specifically , results and analysis are presented in section [ sec : results ] and discussed in section [ sec : discussion ] , with conclusions set out in section [ sec : conc ] .
observations of orion - kl were carried out with the sma @xcite in its compact configuration on december 12 , 2013 for about 6.75h .
the phase - tracking centre was @xmath9 = 05@xmath1035@xmath1114@xmath1220 , @xmath13 = -05@xmath1422@xmath1533@xmath1600 .
the new broader - bandwidth sma capability of the 230 ghz band was used to cover the following frequency ranges : 192.6194.6 ghz and 197.6199.6 ghz in lsb band and , 207.6209.6 ghz and 212.6214.6 ghz in the usb band .
an enormous advantage of this broadband capability is the possibility to observe transitions of both hdco and h@xmath1@xmath17co in a single tuning ( see table [ tab1 ] ) .
the correlator was configured with a uniform spectral resolution over 4 ghz bandwidth in each sideband .
each 4 ghz bandwidth is divided into 48 ` chunks ' , each is further divided into 128 channels with a channel width of 0.812 mhz .
the weather conditions were good and stable with an atmospheric opacity of about 0.14 at 225 ghz .
the sma data were calibrated using the mir / idl packagecqi / mircook.html ] @xcite .
the nearby quasars 0423@xmath18013 and 0510@xmath19180 were used as complex gain ( phase and amplitude ) calibrators .
continuum subtraction and data imaging were performed using the miriad software package .
the final continuum emission maps obtained at 198 ghz and 213 ghz are shown in figure [ fg1 ] .
they were restored using a robust weighting of 0.0 , resulting in a synthesized beam size of 3.5@xmath0 3.01(p.a .
= 32.7@xmath14 ) at 198 ghz and of 3.2@xmath0 2.8(p.a .
= 26.7@xmath14 ) at 213 ghz .
positions of the four main continuum sources derived from gaussian fits in the @xmath20 lsb dataset image plane are given in figure [ fg1 ] .
positions of these continuum sources are commensurate between the lsb and usb data sets and , in agreement , within the synthesized beam size , with previous pdbi observations @xcite . in this analysis
, we only focus on 2 emission lines of hdco and 2 of para
h@xmath1@xmath17co ( see spectroscopic parameters listed in table [ tab1 ] ) .
the spectral resolution is 1.14 km s@xmath21 for h@xmath22co at 212/213 ghz and 1.26 km s@xmath21 for the hdco lines at 193 ghz .
the resulting synthesized beam sizes ( @xmath5@xmath23@xmath0@xmath23 , p.a .
= 30 @xmath14 ) are given in table [ tab1 ] at the different frequencies .
the hdco and h@xmath1@xmath17co emission maps integrated over the line profile are shown in figure [ fg2 ] .
four main molecular emission peaks , labeled df1 to df4 , are associated with the hdco emission .
more specifically , they are located toward the left of hot core component ( df1 ) , the north of the compact ridge ( df2 ) and in the quiescent ridge ( df3 and df4 ) located south of the compact ridge .
their coordinates are given in table [ tab1 ] .
the regions df1 , df2 and df3 are also associated with the h@xmath1@xmath17co emission but not the df4 one .
incidentally , it is important to note that the integrated emission map of the h@xmath1@xmath17co transition at 213293.57 mhz is strongly contaminated by the emission from an unidentified species toward the hot core region .
this is shown in fig .
[ fg3 ] , which displays the integrated emission map along with the spectrum of this h@xmath1@xmath17co line ; the latter being taken at the position of the contaminated emission peak .
spectra of the formaldehyde transitions observed towards the regions df1 to df4 are shown in figure [ fg4 ] and the observed line parameters are summarized in table [ tab1 ] . the bulk of the formaldehyde emission towards those peaks appears to lie between @xmath24=7.5 km s@xmath21 and @xmath24=9.8 km s@xmath21 .
this result is consistent with the hdco lsr velocity derived by @xcite toward the quiescent ridge ( @xmath58.5 km s@xmath21 , single - dish millimeter wave observatory observations ) and by @xcite toward the region ( @xmath58 and 10.4 km s@xmath21 , herschel / hifi observations ) .
it is important to note that @xcite have associated the hdco emission with two velocity components in compact ridge .
this difference results from the fact that _
i ) _ some complex molecules , such as methyl formate or dimethyl ether , are seen with two distinct velocity components emitting at @xmath57.5 and 9.2 km s@xmath21 in the compact ridge ( e.g. * ? ? ?
* ; * ? ? ?
* ) and that , _
ii ) _ the herschel observations are spatially unresolved .
we infer that the two hdco velocity components seen by @xcite , actually correspond to our hdco emission observed toward df3 and df1 , respectively . assuming optically thin emission , we use the integrated line intensity measurements of the observed h@xmath1@xmath17co transitions to derive the formaldehyde excitation temperature ( t@xmath25 ) within orion kl following eq . [ eq1 ] : @xmath26 where e , w , a , g and @xmath27 refer to the state energies ( k ) , the integrated line intensities ( k km s@xmath21 ) , the einstein a coefficients ( s@xmath21 ) , the statistical weights and the line frequencies ( hz ) of the 3@xmath282@xmath29 ( u ) and 3@xmath302@xmath31 ( l ) transitions .
it is important to note that equation [ eq1 ] can not be directly applied to hdco since both of the detected transitions have the same upper energy level ( see table [ tab1 ] ) .
thus , the below analysis hinges upon the assumption that t@xmath25 is the same for both hdco and h@xmath1@xmath17co .
figure [ fg5]a presents the excitation temperature map for h@xmath1@xmath17co within orion - kl .
the formaldehyde excitation temperature derived toward the df2 , df3 and df4 peaks is low ( 5257 k , see table [ tab1 ] ) contrary to df1 , for which it is higher ( @xmath590 k ) .
our derived temperatures in the quiescent ridge ( i.e regions df3 and df4 ) are commensurate within the uncertainties , with the one derived by ( * ? ? ?
* k ) using ch@xmath32c@xmath1h , a known tracer of ambient ridge gas . regarding the df2 emission peak ,
our derived temperature is in agreement with one of the kl w methanol clump that lies in its vicinity ( i.e @xmath52@xmath33 east , see * ? ? ?
as for the df1 emission peak , there is a slight discrepancy with the temperature derived by @xcite .
in particular , these authors report a temperature that is lower than the typical temperature derived in the vicinity of this region ( @xmath5100300 k , e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
however , looking at their figure b.4 .
, a 130 k temperature seems to be consistent with their data , suggesting that their error bars might have been underestimated .
overall the temperatures derived here are lower than the typical temperatures derived for organic species towards the hot core and compact ridge components associated with the inner part of orion kl ( e.g. * ? ? ?
assuming that _ i ) _ lte is reached , which implies that the excitation temperature is equal to the rotational temperature and , _ ii ) _ the rotational temperature is the same for both hdco and h@xmath6@xmath17co , we derive the total column density , n , by using the following equation @xcite : @xmath34 where , @xmath35 the column densities derived toward the regions df1 to df4 for both hdco and para
h@xmath6@xmath17co are listed in table [ tab1 ] .
in addition , table [ tab2 ] gives the total column densities for hdco , h@xmath6@xmath17co and h@xmath6co toward the df1 to df4 peaks .
the total column densities have been calculated as follow : _
i ) _ the column densities derived for the two hdco transitions , that are listed in table [ tab1 ] , have been averaged .
the same applies to the two para h@xmath6@xmath17co transitions . _
ii ) _ regarding h@xmath6@xmath17co and hdco , we assume an ortho : para ratio of 3:1 , that is commensurate with measurements performed towards orion kl by @xcite and @xcite , and a @xmath36c/@xmath17c isotopic ratio of 70 ( see , * ? ? ?
* and reference therein ) , that is in agreement within the error bars with previous studies ( e.g. * ? ? ?
* ; * ? ? ?
using the total column densities for hdco and for h@xmath6@xmath17co ( see section [ sec : coldens ] and table [ tab2 ] ) , we derive the d / h ratio for the regions df1 to df4 .
the resulting d / h ratios are given in table [ tab2 ] .
figure [ fg5]b shows the distribution map of the d / h ratio for formaldehyde in orion kl .
the hdco / h@xmath6co ratio is found to be [email protected] toward df4 , [email protected] toward df1 and [email protected] and [email protected] toward df2 and df3 , respectively .
our results are commensurate within the uncertainties with previous d / h ratio measurements in the region ( e.g. * ? ? ?
* ; * ? ? ?
* ) and suggest a possible gradient in the hdco / h@xmath6@xmath17co ratio at a 3@xmath37 level .
a recent study by @xcite using the plateau de bure interferometer derived the d / h ratios of methanol ( ch@xmath1doh and ch@xmath32od ) at comparable resolution to our sma observations .
a notable result is that the spatial distribution of formaldehyde emission ( hdco and h@xmath1@xmath17co , see fig .
[ fg2 ] ) clearly differs from that of ( deuterated ) methanol ( see fig . 3 of * ? ? ?
the methanol distribution follows a v - shaped structure linking the hot core component to the compact ridge @xcite while formaldehyde exhibits a spatial structure which appears to surround this molecular v - shaped structure on four sides .
this difference is illustrated in figure [ fg2 ] ( bottom right panel ) in which the deuterated methanol emission peaks identified by @xcite are indicated on the hdco integrated emission map at at 193391.6 mhz . furthermore , the d / h ratio for methanol derived by @xcite lies in the range 0.81.3@xmath010@xmath38 , which is up to an order of magnitude ( within the uncertainties ) lower than the d / h ratio for formaldehyde derived in this study ( see table [ tab2 ] ) .
all the nearby well characterized sources ( i.e. hot core , compact ridge ) are not coincident with formaldehyde emission . in fact , it is possible and even likely that we are seeing quiescent material surrounding the hotter sources , such as the well known quiescent ridge in orion kl .
incidentally , the spatial difference between the formaldehyde and methanol emission has already been observed toward others sources , such as the horsehead photon - dominated region ( pdr ) and its associated dense core @xcite and the orion bar pdr @xcite . in their respective analyses
, the authors have shown that the uv radiation field affects the gas phase abundance of formaldehyde and methanol .
more specifically , the h@xmath1co / ch@xmath32oh abundance ratio is expected to be higher in high fuv illuminated environments in comparison to denser and less illuminated regions . in this light , the entire orion
kl quiescent ridge gas is a face on pdr ( see * ? ? ?
* ; * ? ? ?
* ) , therefore photodissociation of gas - phase methanol into formaldehyde might occur ( e.g. see , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . in this context
we determine the abundance ratio of h@xmath1co / ch@xmath32oh in the formaldehyde clumps .
comparison to ratios measured towards other pdrs would then illuminate whether the orion pdr is influencing the local chemistry of dense material surrounding the hot embedded cores . the determination of the methanol column density ( see eq .
2 ) relies on the assumptions that the ch@xmath32oh e transitions listed in table [ tab3 ] are optically thin and , that the rotational temperature is the same as that of h@xmath1@xmath17co .
it is important to note that toward the df1 and df2 emission peaks the ch@xmath32oh e lines are strongly blended ( see figure [ fg6 ] ) , preventing us from deriving the column density of methanol at those positions .
regarding the df3 and df4 regions , the resulting h@xmath1co / ch@xmath32oh abundance ratio lies in the range 3.54.5 , that might suggest a slight decrease in the methanol abundance in the quiescent ridge .
however , our derived ratio is lower by at least a factor of 3400 in comparison to that estimated by @xcite toward the orion bar , suggesting that the production of formaldehyde in orion kl through photodissociation of gas - phase methanol is not the main formation route .
finally , it is noticeable that our estimated h@xmath1co / ch@xmath32oh abundance ratio is more consistent with that found in hot corino regions ( see @xcite and @xcite for a review ) .
this finding leads us to argue that formaldehyde emission in orion
kl is more likely probing the edge of the hot gas .
formaldehyde and methanol are believed to be formed on ice grain mantles during the pre - stellar cold phase ( t@xmath3950 k ) from successive hydrogenation reactions of h or d atoms with co ( see e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) : @xmath40 and @xmath41 laboratory experiments have shown that the sequential co hydrogenation process ( i.e. reaction [ eq3 ] ) efficiently form both formaldehyde and methanol at low temperatures ( @xmath320 k , see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
thus , if methanol and formaldehyde both purely originate from ice grain mantles , the measured d / h ratio should be similar ( within a factor 2 ) for both molecules and , in particular , these molecules should trace similar environments while desorbed as they share a common origin .
this is inconsistent with our observations ( see section [ sec : anti ] ) and thus strongly argues in favor of differences in the chemical formation pathways of these deuterated molecules .
thus ch@xmath32oh and h@xmath1co emission is probing different spatial regions within the central region of orion kl that may not be coeval .
we infer that a likely scenario is a hdco production through hot gas - phase deuterium chemistry in the colder gas associated with orion kl , as suggested by @xcite . more specifically , at * relatively * high temperature ( t@xmath570k100k , * ? ? ?
* ; * ? ? ?
* ) , such as that derived in this study ( see figure [ fg5 ] ) , reaction [ eq5 ] is active due to its high exothermicity ( @xmath42e of 654 k , see * ? ? ?
* ) and favors deuterium fractionation through reactions involving ch@xmath1d@xmath7 as a parent molecule @xcite .
@xmath43 in that context , reaction [ eq5 ] can then lead to the formation of hdco and h@xmath1co ( see further details in * ? ? ?
incidentally , the temperature of this colder region associated with orion - kl is not sufficient enough to desorb ch@xmath32oh from the surface of icy grain mantles ( t@xmath4120160 k see , * ? ? ?
* ; * ? ? ?
however , in the inner hottest part of orion kl where methanol emits , most of the formaldehyde might have been converted into methanol by grain surface reactions @xcite during a colder phase prior to the explosive event .
then , upon heating methanol is released from grain surfaces into gas phase along with little formaldehyde .
we stress that the release of methanol into gas phase in this region is mainly due to internal radiation heating ( i.e outflow , shock , ir sources ) but not heat from the pdr ( see section 4.2 ) . in this instance
, the d / h ratio and the spatial distribution of methanol should differ from the one of formaldehyde .
we have characterized , for the first time , the spatial distribution of deuterated formaldehyde ( hdco ) toward orion kl with an angular resolution of 3.4@xmath0 3.0using the sma . the new sma broadband capability of the 230 ghz band allowed simultaneous observation of both hdco and h@xmath1@xmath17co within relatively sparse spectra .
four main emission peaks , labelled in this study df1 to df4 , are identified in the formaldehyde emission image .
a salient result is that the spatial distribution the emission of the hdco and h@xmath1@xmath17co emission differ from that of ch@xmath32oh and its deuterated isotopologues .
the emission maps show that methanol emits in the inner part of orion kl , while formaldehyde is found in the colder gas enveloping the ch@xmath32oh emission on four sides : one located left of the hot core ( df1 ) , one toward the north of the compact ridge ( df2 ) and two in the quiescent ridge ( df3 and df4 ) .
in addition , we derive a hdco / h@xmath1co ratio which lies in the range of 0.003 to 0.009 within the region ; this value is an order of magnitude higher than the d / h measured for methanol .
we suggest that the high measured d / h ratio for formaldehyde and its spatial distribution in orion kl are a combination of gas phase chemistry in the cooler gas associated with orion kl and , a conversion of formaldehyde into methanol on ice grain mantles in the earlier history of the inner part .
llllcc|ccccc|ccccc| 192893.26 & 3@xmath302@xmath31 & 18.5 & 1.94& & & & & & & & & & & & + 193907.46 & 3@xmath282@xmath44 & 50.4 & 1.10 & 3.44 @xmath0 3.01&30 & 8.50(1.54)&9.84(0.14)&1.80(0.46 ) & 89(16 ) & 4.65(0.86 ) & 7.38(0.43)&7.75(0.10 ) & 2.20(0.18 ) & 56(7 ) & 2.92(0.22 ) + 193391.61 & 3@xmath452@xmath28 & 50.4 & 1.09 & 3.47 @xmath0 3.02 & 30&9.05(0.63 ) & 9.65(0.10 ) & 1.69(0.16 ) & 89(16 ) & 4.96(0.37)&7.07(0.35)&7.57(0.07 ) & 2.40(0.15 ) & 56(7 ) & 2.80(0.18 ) + & & & & & & & + & & & & & & @xmath46dv & @xmath47 & @xmath48 & t@xmath25 & n&@xmath46dv & @xmath47 & @xmath48&t@xmath25 & n + & & & & & & ( k km s@xmath21 ) & ( km s@xmath21 ) & ( km s@xmath21 ) & ( k ) & ( 10@xmath49@xmath50 ) & ( k km s@xmath21 ) & ( km s@xmath21 ) & ( km s@xmath21 ) & ( k ) & ( 10@xmath49@xmath50 ) + 192893.26 & 3@xmath302@xmath31 & 18.5 & 1.94 & & & & & & & & & & & & + 193907.46 & 3@xmath282@xmath44 & 50.4 & 1.10&3.44 @xmath0 3.01&30 & 7.13(0.98 ) & 7.80(0.10 ) & 1.71(0.32 ) & 57(8 ) & 2.77(0.42 ) & 11.25(1.45 ) & 8.61(0.05 ) & 1.13(0.14 ) & 52(8 ) &
4.26(0.62 ) + 193391.61 & 3@xmath452@xmath28 & 50.4 & 1.09 & 3.47 @xmath0 3.02 & 30 & 8.00(2.29)&8.00(0.03)&1.39(0.44 ) & 57(8 ) & 3.12(0.93 ) & 9.79(1.02 ) & 8.30(0.10)&1.66(0.24 ) & 52(8 ) & 3.71(0.45 ) + + + + & & & & & & + & & & @xmath010@xmath51 & & @xmath46dv & @xmath47 & @xmath48 & t@xmath25 & n&@xmath46dv & @xmath47 & @xmath48&t@xmath25 & n + ( mhz ) & & ( k ) & ( s@xmath21 ) & ( @xmath33 @xmath0 @xmath33)&pa ( @xmath14 ) & ( k km s@xmath21 ) & ( km s@xmath21 ) & ( km s@xmath21 ) & ( k ) & ( 10@xmath49@xmath50 ) & ( k km s@xmath21 ) & ( km s@xmath21 ) & ( km s@xmath21 ) & ( k ) & ( 10@xmath49@xmath50 ) + 212811.18 & 3@xmath302@xmath31 & 20.4 & 2.64 & 3.17 @xmath0 2.75 & 30 & 17.99(1.73 ) & 9.88(0.09 ) & 1.83(0.24 ) & 89(16 ) & 5.44(0.53 ) & 16.04(1.37 ) & 7.44(0.11 ) & 2.54(0.29 ) & 56(7 ) & 2.83(0.26 ) + 213037.34 & 3@xmath452@xmath28 & 67.7 & 1.46 & & & & & & & & & & & & + 213293.57 & 3@xmath282@xmath29 & 67.7 & 1.46 & 3.15 @xmath0 2.75 & 30 & 5.87(0.85 ) & 9.89(0.19 ) & 2.21(0.42 ) & 89(16)&5.43(0.82)&3.84(0.40 ) & 7.43(0.13 ) & 2.16(0.30 ) & 56(7 ) & 2.83(0.35 ) + & & & & & & & + & & & & & & @xmath46dv & @xmath47 & @xmath48 & t@xmath25 & n&@xmath46dv & @xmath47 & @xmath48&t@xmath25 & n + & & & & & & ( k km s@xmath21 ) & ( km s@xmath21 ) & ( km s@xmath21 ) & ( k ) & ( 10@xmath49@xmath50 ) & ( k km s@xmath21 ) & ( km s@xmath21 ) & ( km s@xmath21 ) & ( k ) & ( 10@xmath49@xmath50 ) + 212811.18 & 3@xmath302@xmath31 & 20.4 & 2.64 & 3.17 @xmath0 2.75 & 30 & 12.93(0.28 ) & 7.94(0.03 ) & 1.55(0.04 ) & 57(8 ) & 2.27(0.64)&9.70(1.16 ) & 8.44(0.03 ) & 1.58(0.25 ) & 52(8 ) & 1.58(0.20 ) + 213037.34 & 3@xmath452@xmath28 & 67.7 & 1.46 & & & & & & & & & & & & + 213293.57 & 3@xmath282@xmath29 & 67.7 & 1.46 & 3.15 @xmath0 2.75 & 30 & 3.14(0.55 ) & 8.03(0.08 ) & 1.56(0.37 ) & 57(8 ) & 2.27(0.44)&2.16(0.26 ) & 8.63(0.17 ) & 1.30(0.23 ) & 52(8)&1.58(0.23 ) + rrrrr df1 & 4.81(0.62 ) & 21.70(2.72 ) & 15.20(1.90 ) & 0.0032(0.0008 ) + df2 & 2.86(0.20 ) & 11.30(1.22 ) & 7.92(0.86 ) & 0.0036(0.0006 ) + df3 & 2.94(0.67 ) & 9.06(1.00 ) & 6.34(0.70 ) & 0.0046(0.0016 ) + df4 & 3.99(0.54 ) & 6.34(0.85 ) & 4.44(0.60 ) & 0.0090(0.0024 ) + rlrrl 193415.37 & 4@xmath52 3@xmath53 @xmath54=0 & 36.3 & 3.03 + 193441.61 & 4@xmath55
3@xmath56 @xmath54=0 & 28.8 & 2.84 + 193474.33 & 4@xmath57
3@xmath58 @xmath54=0 & 70.9 & 1.33 + 193488.99 & 4@xmath59
3@xmath57 @xmath54=0 & 85.9 & 1.33 + 193506.60 & 4@xmath56
3@xmath60 @xmath54=0 & 44.3 & 2.91 +
|
we report the first high angular resolution imaging ( 3.4@xmath0 3.0 ) of deuterated formaldehyde ( hdco ) toward orion
kl , carried out with the submillimeter array ( sma ) .
we find that the spatial distribution of the formaldehyde emission systematically differs from that of methanol : while methanol is found towards the inner part of the region , hdco is found in colder gas that wraps around the methanol emission on four sides .
the hdco / h@xmath1co ratios are determined to be 0.0030.009 within the region , up to an order of magnitude higher than the d / h measured for methanol .
these findings strengthen the previously suggested hypothesis that there are differences in the chemical pathways leading to hdco ( via deuterated gas phase chemistry ) and deuterated methanol ( through conversion of formaldehyde into methanol on the surface of icy grain mantles ) .
|
single transverse spin asymmetries ( stsas ) generated in semi - inclusive deep inelastic scattering ( sidis ) and in hadronic collisions are one of the hot topics of research in quantum chromodynamics ( qcd ) , promising unparalleled insight in the physics of chiral symmetry breaking and quark confinement . in the factorization framework involving transverse momentum - dependent distribution functions ( tmds ) @xcite the origin of stsas
is chiefly ascribed to either the quark tmds ( sivers effect @xcite ) , to multiple partonic rescattering @xcite , or to the quark fragmentation functions ( collins effect @xcite ) . while both the quark tmd and the fragmentation function are non - perturbative , and , according to the conventional wisdom , can not be calculated form first principles , it is desirable to understand the detailed physical mechanism leading to generation of stsas in qcd . to that end
a significant progress has been achieved by brodsky , hwang and schmidt ( bhs ) in @xcite ( see also @xcite ) , where , in a quark di - quark proton model calculation , it has been shown that the stsa in sidis can be generated through an interference of the final - state parton rescattering diagram with the born - level amplitude .
in essence , it was shown in @xcite that multiple partonic rescatterings are key to generating the asymmetry .
the multiple rescatterings are often referred to as the `` lensing '' interaction , since , in sidis , the associated color - lorentz force tries to attract the knocked - out quarks back into the hadron @xcite , thus `` focusing '' them .
the effects of such multiple rescatterings can be absorbed into the sivers distribution function of a polarized hadron in sidis @xcite .
a consequence of this understanding of the origin of stsa in sidis , is that the sivers function ( and , hence , the asymmetry itself ) has to change sign between sidis and the drell - yan process ( dy ) . at the level of the operator matrix element
this conclusion has been reached in @xcite , while an illustration of this result in the bhs model was completed only recently @xcite ( see also @xcite for the outline of the calculation ) .
it is our understanding that in the `` lensing '' interpretation of stsas this sign change corresponds to the color - lorentz force changing sign from attractive to repulsive between a knocked - out quark in sidis and the incoming anti - quark in dy .
the goal of the present work is to extend our understanding of the physical mechanism behind the stsa beyond the quark di - quark model of the proton used in @xcite ( see @xcite for other efforts in a similar direction ) . in particular ,
multiple partonic rescatterings in high energy scattering can be particularly simply accounted for in the framework of the quasi - classical approximation to qcd employed in the glauber mueller ( gm ) @xcite and , equivalently , mclerran venugopalan ( mv ) @xcite models . in these approaches
the hadron is modeled by a large nucleus , with a large number @xmath0 of nucleons in it .
the large number of nucleons leads to high density of small-@xmath1 gluons in the nuclear wave function , which , in turn , generates a hard scale @xmath2 known as the parton saturation scale , justifying the use of perturbative qcd calculations .
( for reviews of the saturation / color glass condensate ( cgc ) physics see @xcite . )
the fact that the quasi - classical approximation generates a hard scale justifying the approach indicates that it is not simply a `` model '' of qcd , but , in fact , it represents a limiting behavior of strong interactions at high energy .
multiple rescatterings can be resummed in the gm / mv model as an expansion in powers of the parameter @xmath3 @xcite : the presence of a resummation parameter allows for a controlled approximation to the problem at hand . in the past
there was a number of efforts to include spin effects in the saturation / cgc framework @xcite . to alleviate the worry about whether a large nucleus can adequately represent a proton ( or any other hadron ) in sidis and dy experiments , let us point out that in unpolarized scattering the proton may have a significant number of non - perturbatively generated large-@xmath1 ( @xmath4 ) partons , which are modeled by `` nucleons '' in this large - nucleus approximation .
the large-@xmath1 partons/nucleons , in turn , give rise to small-@xmath1 gluons .
the resulting expressions for the deep inelastic scattering ( dis ) structure functions have been quite successful in describing hera low-@xmath1 data @xcite , also indicating relevance of the large-nucleus approximation to the proton wave function at small-@xmath1 . in
what follows we would have to slightly modify the original mv model of the nucleus by giving the `` nucleus '' both a non - zero spin and a non - zero orbital angular momentum ( oam ) . here
this would mean that free nucleons in an approximately spherical bag , as considered originally in @xcite , would now be polarized and would be orbiting the nuclear spin axis . in a realistic polarized nucleus
the nucleons tend to form pairs with zero net oam , such that the net spin of the nucleus is carried by the few unpaired nucleons and does not get very large ( does not grow directly with @xmath0 ) . since it is not clear whether such effect ( at the level of quarks and gluons ) takes place in the proton we are trying to model , we will not make any particular assumptions about the polarizations and oams of the nucleons in our `` nucleus '' .
the main physical mechanism for generating stsa in the quasi - classical framework is as follows .
imagine a large spinning nucleus .
the nucleus is so large that it is almost completely opaque to a colored probe .
this strong nuclear shadowing is due to multiple rescatterings in the nucleus generating a short mean free path for the quark , anti - quark , or a gluon .
let us first consider the drell - yan process on such a rotating nucleus with shadowing , as shown in in the nuclear rest frame with the rotation axis of the nucleus perpendicular to the collision axis .
the incoming anti - quark ( generated in the wave function of the other hadron ) scatters on the `` front '' surface of the polarized nucleus due to the strong shadowing . since the anti - quark interacts with the nucleons which
, at the `` front '' of the nucleus preferentially rotate with the nucleus out of the plane of the page in , the produced time - like virtual photons are produced preferentially out of the page , generating left - of - polarized - beam single spin asymmetry .
the same mechanism can be applied to generate stsa in sidis , as illustrated in , also in the rest frame of the nucleus .
now the incoming virtual photon interacts with the transversely polarized nucleus , producing a quark . for the quark to escape out of the nucleus and be produced the interaction has to take place at the `` back '' of the nucleus , to minimize the path the quark needs to travel through the nucleus , maximizing its chances to escape .
the nucleons in the `` back '' of the nucleus rotate preferentially into the page of : scattering of a virtual photon on such nucleons results in the right - of - beam single spin asymmetry for the outgoing quarks ( quarks produced preferentially with transverse momentum pointing into the page ) .
spin asymmetries in dy and sidis shown in figs .
[ dy_fig ] and [ sidis_fig ] are generated through a combination of oam effects and nuclear shadowing .
the two asymmetries are opposite - sign ( left- and right - of - beam ) , and , assuming that scattering in the two processes happens equal distance from the nuclear edge , are likely to be equal in magnitude , in agreement with the prediction of @xcite . as we will see below in the actual calculations , the stsas in figs .
[ dy_fig ] and [ sidis_fig ] do require multiple rescatterings , but they are needed solely to generate nuclear shadowing .
thus the physical mechanism of figs .
[ dy_fig ] and [ sidis_fig ] is quite different from the `` lensing '' interaction @xcite , in which the knocked - out quark in sidis `` feels '' the net color charge of the remainder of the proton , and is attracted back by this charge @xcite . in the presence of shadowing , it would be much harder for the quark in sidis to `` see '' the whole remainder of the polarized proton ( nucleus ) coherently : thus one expects the `` lensing '' effect to weaken with increasing shadowing ( if we could increase shadowing without modifying the degree of the polarization of the nucleus ) .
this is qualitatively different from the mechanism in figs .
[ dy_fig ] and [ sidis_fig ] , in which the asymmetry actually increases with shadowing .
clearly , the more opaque the nucleus is , the more likely the interactions to happen at its `` front '' in dy and at its `` back '' in sidis , making the asymmetry larger . in the paper below we will outline the calculations leading to the physical picture presented in figs .
[ dy_fig ] and [ sidis_fig ] . after some generalities in sec .
[ sec : general ] we proceed in sec .
[ sec : sidis ] with the quasi - classical analysis of stsa in sidis . as mentioned above , to model the oam of a polarized nucleus we have to assume that the nucleus is rotating .
this implies a generalization of the original mv and gm models , in which nucleons are static , to include rotational motion of the nucleons .
hence the nucleons need to have both the well - defined positions and momenta : this is only possible in the classical limit .
the classical mv model limit is achieved in sec .
[ sec : sidis ] using the wigner functions approach , which allows to specify both the positions and momenta of the nucleons in the polarized nucleus .
we then proceed to the calculation of stsa in sidis , identifying two mechanisms for stsa generation : one is due to the coupling of the produced quarks transverse momentum to the oam of the nucleus , while another one is due to the stsa generated in the scattering of the virtual photon on an individual nucleon along the lines of the bhs mechanism @xcite ( sivers function density ) .
the former mechanism is leading in the saturation framework , being dominant in the saturation power counting ( for non - zero oam ) : it is order - one for @xmath5 .
the latter mechanism is order-@xmath6 for @xmath5 , and is thus subleading . at large values of the produced quark transverse momentum @xmath7
the oam mechanism gives the contribution to the sivers function of the order @xmath8 with @xmath9 the typical transverse momentum of the valence quarks in the polarized nucleus due to orbital motion and @xmath10 the nucleon mass ( with @xmath11 roughly the constituent quark mass ) , while the sivers function density gives a contribution proportional to @xmath12 .
assuming that @xmath13 , we see that the sivers function density mechanism dominates for @xmath14 ; conversely , the oam mechanism is dominant for @xmath15 , the domain including everything inside of the saturation region and a phase - space sector outside of that region .
a similar quasi - classical stsa calculation is carried out for the drell - yan process in sec .
[ sec : dy ] , where we also explicitly show the mechanism for the sign reversal of the sivers function outlined in this introduction .
we conclude in sec . [
sec : discussion ] by summarizing our results and outlining possible improvements of our results left for the future work .
the single transverse spin asymmetry is defined as @xmath16 for producing a hadron with transverse momentum @xmath17 in sidis on a transversely polarized target and in polarized proton
proton collisions or a di - lepton pair with transverse momentum @xmath17 in dy process on a polarized proton . the asymmetry @xmath18 singles out a part of the production cross section proportional to @xmath19 , where @xmath20 is the 3-momentum of the polarized hadron pointing along the collision axis . throughout this paper we will use light - cone
coordinates @xmath21 with the corresponding metric @xmath22 . accordingly , we denote four - vectors as @xmath23 , with the transverse momentum @xmath24 and @xmath25 where @xmath26 is the quark field with dirac index @xmath27 , while the quark is taken with transverse momentum @xmath28 and the longitudinal momentum fraction @xmath1 .
the proton ( or polarized nucleus ) spin four - vector is @xmath29 , while @xmath30 is the gauge link necessary to make the object on the right of gauge - invariant .
below , when considering sidis and dy , we will work in the light - cone gauge of the projectile .
choosing the polarized proton ( nucleus ) to move along the light - cone @xmath31-direction , such that @xmath32 is large , we will work in the @xmath33 gauge . in the quasi - classical approximation
the @xmath33 gluon field of a large ultrarelativistic nucleus moving along the @xmath31-direction has zero transverse component , @xmath34 , such that the only non - zero component is @xmath35 . defining the wilson line @xmath36 \equiv \mathcal{p } \ , \exp \left [ \frac{i g}{2 } \int\limits_{a^-}^{b^- } dx^- a^+ ( x^+ = 0 , x^- , { \underline{x } } ) \right]\end{aligned}\ ] ] we write for the case of sidis @xcite @xmath37 \ , v_{{\underline{x } } } [ + \infty \ , , \ , x^-],\end{aligned}\ ] ] while for dy we have @xmath38 \ , v_{{\underline{x}}}^\dagger [ x^- \ , , \ , -\infty ] .\end{aligned}\ ] ] in both cases we neglected the transverse gauge link at @xmath39 since @xmath34 in the gauge we chose .
as will become apparent below , the direction of the wilson lines in the @xmath40 s is given by the direction of motion of the outgoing quark in sidis and the incoming anti - quark in dy .
this results in different definitions of the correlator @xmath41 for the two processes , which is usually referred to as the controlled process - dependence of the tmds @xcite .
the correlation function @xmath41 is decomposed as @xcite @xmath42_{ij},\end{aligned}\ ] ] where @xmath43 is the mass of the polarized proton or nucleus . in
the following we will be using the sivers function @xmath44 and the unpolarized quark tmd @xmath45 .
these functions can be obtained from the correlator @xmath41 using the following expressions [ tmds ] @xmath46 @xmath47
we first consider the process of quark production in semi - inclusive deep inelastic lepton scattering on a transversely polarized heavy nucleus : @xmath48 .
the leptonic tensor can be factorized out in the usual way , so we represent the process as the scattering of a virtual photon : @xmath49 .
this photon carries a large spacelike virtuality @xmath50 and knocks out a quark from one of the nucleons , which may then rescatter on the nuclear remnants .
the nucleus is taken in the classical gm / mv approximation , which we augment by requiring that the nucleons are polarized and the nucleus rotates around the transverse spin axis , which leads to a non - zero oam .
power - counting .
a quark is ejected from a nucleon in the nucleus by the high - virtuality photon , which escapes without rescattering .
different solid horizontal lines represent valence quarks from different nucleons in the nuclear wave function , with the latter denoted by the vertical shaded oval.,height=151 ] consider first the lowest - order process shown in fig . [ fig : dis1 ] , in which a quark is ejected without rescattering . physics quark production is dominated by a higher - order in @xmath6 process , where the virtual photon splits into a @xmath51 pair before hitting the target : since in this work @xmath52 , the dipole process is not dominant , constituting an order-@xmath6 correction to the channel shown in .
] we work in a frame ( such as the photon - nucleus center - of - mass frame ) in which the virtual photon moves along the @xmath53-axis with a large momentum @xmath54 and the nucleus moves along the @xmath31-axis with a large momentum @xmath32 . in this frame , the kinematics are @xmath55 where @xmath56 is the mass of the nucleus and the on - mass - shell valence quark with momentum @xmath57 is a part of the light - cone wave function of the nucleus . in what follows
we will model nucleons as made out of single valence quarks : in the end of the calculation , to go back to the nucleons one simply would need to replace distribution functions in a valence quark by the distribution functions in the nucleons .
let us denote the photon - nucleus center - of - mass energy squared by @xmath58 and the photon - nucleon ( valence quark ) center - of - mass energy squared by @xmath59 .
we consider the kinematic limit @xmath60 and work to leading order in the small kinematic quantities @xmath61 , which we denote collectively as @xmath62 . since we are operating in the limit in which @xmath63 , the formalism of tmd factorization applies , justifying the use of the correlator and decomposition .
additionally , to a good accuracy one can assume that a typical scale for the momentum fraction @xmath64 is @xmath65 , where @xmath0 is the mass number of the nucleus ( in fact , @xmath66 for the single - valence quark `` nucleons '' at hand ) .
in this limit , @xmath67 where @xmath68 is the bjorken scaling variable per nucleon .
the corresponding scaling variable for the entire nucleus is @xmath69 . the kinematic limit at hand
, @xmath70 corresponds to @xmath71 .
the on - shell condition for the outgoing gluon is @xmath72 which fixes the struck quark to be ejected along the @xmath53-direction , so that its light - cone plus momentum @xmath73 is small since @xmath74 .
this also fixes the momentum fraction of the active quark just before interaction with the photon to be @xmath75 in the usual way .
( note that @xmath76 . ) in our frame , the @xmath53-extent of the lorentz - contracted nucleus is @xmath77 , where @xmath78 is the radius of the nucleus in its rest frame .
the incoming virtual photon and outgoing quark interact with the nucleus based on their corresponding coherence lengths : @xmath79 and @xmath80 , respectively . comparing these to the size of the nucleus
, @xmath81 we see that the photon s coherence length is short , but the coherence length of the ejected quark is parametrically large for @xmath82 .
thus , for our calculation in which @xmath83 , the virtual photon interacts incoherently ( locally ) on a single nucleon , but the ejected quark interacts coherently with all of the remaining nucleons it encounters before escaping the nucleus .
this limit thus combines the local `` knockout '' picture of the deep inelastic scattering process with the coherent rescattering that usually characterizes the small-@xmath1 limit . in the formal limit of a large nucleus in which @xmath84 and @xmath85 such that @xmath86 , these coherent interactions with subsequent nucleons must be re - summed according to this saturation - based power counting .
in general it is rather straightforward to write an answer for the quasi - classical quark production in sidis . as we mentioned in the introduction , here
the problem is a little more subtle than usual since we are interested in also including transverse and longitudinal motion of the nucleons in the nucleus in order to model its oam .
thus our quasi - classical description of the nucleus has to provide us both with the positions and momenta of the nucleons .
this can be done using wigner distributions .
let us illustrate the method with a simple single - rescattering process from .
just like in the parton model , the time scale of inter - nucleon interactions is lorentz - dilated in the infinite momentum frame of the nucleus that we are working in
. we can , therefore , write the scattering amplitude for the process in as a product of the light - cone wave function @xmath87 of the valence quarks in the nucleus ( defined according to light - front perturbation theory rules @xcite in the boost - invariant convention of @xcite ) with the quark virtual photon scattering amplitude @xmath88 : @xmath89 here @xmath90 is the boost - invariant light - cone wave function of a valence quark ( in one of the nucleons ) in the nucleus , while @xmath88 is the scattering amplitude for the `` knock - out '' process @xmath91 .
the sum over valence quark spin and color is implied in . in calculating the quark production process
we need to square this amplitude , integrate it over the momentum of the final state gluon and sum over all nucleons in the nucleus .
since momenta @xmath92 and @xmath93 are fixed , this amounts to integrating over @xmath94 .
one gets @xmath95 first let us introduce a fourier transform of the valence quark wave function , @xmath96 with @xmath97 .
next we define the wigner distribution for the valence quarks ( which is closely related to the wigner distribution of the nucleons in the quasi - classical mv model employed here ) with the help of the fourier transform : @xmath98 note that the wave function is normalized such that @xmath99 giving @xmath100 since @xmath101 we can recast as @xmath102 finally , in the following , as usual in the saturation framework , it would be convenient to calculate the scattering amplitude in ( partial ) transverse coordinate space .
writing @xmath103 ( with @xmath104 and @xmath105 fixed by eqs . and )
we rewrite as @xmath106 note that the fourier transform appears to imply that @xmath107 is the transverse position of the outgoing gluon in , whereas in the wigner distribution @xmath107 is the position of the valence quark @xmath94 .
as we will shortly see such interpretation is not inconsistent : in the classical limit of a large nucleus the wigner distribution is a slowly varying function of @xmath108 , with changes in @xmath109 becoming significant over the variations of @xmath108 over distances of the order of nucleon size 1 fm or larger .
the valence quark and outgoing gluon in are perturbatively close to each other ( being the part of the same feynman diagram ) , and hence the difference in their positions is outside the precision of @xmath110 and can be taken to be the same in the wigner distribution . in appendix [ a ]
we show that the formula holds not only at the lowest order , but when multiple rescatterings are included as well , such that in the kinematics outlined above @xmath111 where we define the energy - independent ( at the lowest non - trivial order ) @xmath112 scattering amplitudes by ( see also eqs . and )
@xcite @xmath113 and @xmath114 in denotes the sum over rescatterings of the virtual photon on any number of nucleons in the nucleus . of the nucleon
, we only keep one convolution with the wigner function explicitly . ]
( note that for a `` nucleus '' made out of a single nucleon @xmath115 , which allows one to reduce to by neglecting the `` spectator '' nucleons . )
we therefore conclude that the quark production cross section for the @xmath116 process can be written as @xmath117 where the cross section for producing a quark in @xmath118 scattering on the nucleons is @xmath119\end{aligned}\ ] ] with the semi - infinite fundamental dipole scattering amplitude given by ( cf . )
@xmath120 = \left\langle \frac{1}{n_c } \ , \mbox{tr } \left [ v_{{\underline{x } } } [ + \infty , b^- ] \ , v^\dagger_{{\underline{y } } } [ + \infty , b^- ] \right ] \right\rangle\end{aligned}\ ] ] and with some @xmath121 and @xmath122-dependent prefactor @xmath123 . here
@xmath124 is the rapidity of the produced quark and a factor of @xmath0 in accounts for the fact that the first scattering can take place on any of the @xmath0 nucleons .
we fixed the normalization of by requiring it to be valid for a nucleus made out of a single nucleon , which would be described by a trivial wigner distribution fixing the momentum and position of the nucleon by simple delta - functions .
( alternatively one could require the formula to be valid for the case of cross section @xmath125 independent of @xmath94 and @xmath126 . )
as already mentioned before , with the accuracy of the large-@xmath0 classical approximation , the argument @xmath108 in the wigner distribution can be replaced by any other transverse coordinate involved in the scattering process .
hence one can replace @xmath108 in @xmath127 from by either @xmath128 or @xmath129 from , or by any linear combination of those variables .
replacing @xmath108 in @xmath127 from by @xmath130 and employing we write @xmath131,\end{aligned}\ ] ] where @xmath132 substituting into yields @xmath133.\end{aligned}\ ] ] is our starting point for exploring the stsa in sidis : it gives the quark production cross section in the quasi - classical approximation .
the expression is illustrated in : the first interaction between the incident virtual photon and a nucleon in the transversely polarized nucleus happens at the longitudinal coordinate @xmath134 .
a quark is knocked out , which proceeds to interact with the rest of the nucleons in the nucleus .
this latter interaction is recoilless and is encoded in a wilson line .
the wigner distribution in allows to take the quasi - classical gm / mv limit of a large nucleus in a controlled way . for a large `` classical '' nucleus
we usually can replace @xmath135 by the following classical expression for it ( neglecting longitudinal orbital motion of the nucleons ) @xmath136 where @xmath137 is the nucleon number density normalized such that @xmath138 the function @xmath139 in is responsible for the transverse momentum distribution of the nucleons and , to satisfy , is normalized such that @xmath140 as originally formulated @xcite , the mv model contained no dependence on the spin or transverse momentum of the valence quarks .
this result is recovered by using @xmath141 .
substituting the classical wigner distribution into yields @xmath142,\end{aligned}\ ] ] which is a simplified version of .
imagine a large nucleus with the total spin @xmath143 such that @xmath144 where @xmath145 is the oam of all the nucleons in the nucleus and @xmath146 is the net spin of all the nucleons .
in the quasi - classical approximation at hand the oam is generated by rotation of the nucleons around a preferred axis .
the nucleus is transversely polarized to the beam : we assume that both @xmath145 and @xmath146 point along the ( positive or negative ) @xmath147-axis .
the result for the quark production cross section in sidis can be utilized to write down an expression for sidis sivers function of the large nucleus with the help of .
we first note that the quark production cross section in sidis is proportional to the correlator with the future - pointing wilson line given by ( cf .
eqs . and ) .
the gauge link in begins and ends at the same @xmath134 , while the more general gauge link in has different endpoints at @xmath148 and @xmath53 .
the reason is that the nuclear wave function is composed of color - neutral `` nucleons '' localized in @xmath134 ; hence there is only a contribution to the correlator when the gauge link both begins and ends at the same @xmath134 .
the dirac @xmath149-matrix of is also present in the quark production cross section since the dirac structure of the large-@xmath104 outgoing quark line is given by @xmath150 . to obtain
the sivers function one only needs to eliminate the gamma
matrices stemming from the quark
photon vertices in the amplitude and in the complex conjugate amplitude ; this can be done by simply contracting the lorentz indices of these gamma
matrices @xcite . while such contraction is not allowed in a calculation of the sidis cross section due to non - trivial structure of the lepton tensor , it is a legitimate method of extracting the sivers function @xcite , since @xmath151 .
we thus see that and equation like would still hold for tr@xmath152 $ ] instead of sidis cross section , since to obtain the former one simply needs to repeat all the steps of the cross section derivation that led to without inserting the photon polarizations ( implicit in ) , and adding a contraction over lorentz indices of the gamma
matrices from the quark photon vertices in the end . by analogy with we can express the quark correlation function @xmath153 of the nucleus in terms of the quasi - classical distribution @xmath154 of nucleons , the quark correlators @xmath155 of individual nucleons , and the semi - infinite wilson line trace @xmath156 $ ] : @xmath157 = a \int \frac{d p^+ \ , d^2 p \ , d b^-}{2(2\pi)^3 } & \ , d^2x \
, d^2 y \ , \sum_\sigma w_n^\sigma \left ( p , b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \int \frac{d^2 k'}{(2\pi)^2 } \ , e^ { - i \ , ( { \underline{k } } - { \underline{k ' } } ) \cdot({\underline{x}}-{\underline{y } } ) } \notag \\ & \times { \mathrm{tr}}[\phi_n(x , { \underline{k } } ' - x \ , { { \underline{p } } } ; p , \sigma ) \ , \gamma^+ ] \ , d_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^-].\end{aligned}\ ] ] is illustrated in . in we
explicitly show the sum over the polarizations @xmath158 of the nucleons along the @xmath1-axis .
note that @xmath159 and it varies with @xmath160 inside the integral ; at the same time the `` averaged '' value of bjorken-@xmath1 per nucleon is @xmath161 .
the quark correlator of the nucleus @xmath153 is defined by , @xmath162 along with the corresponding nucleonic correlator is @xmath163 these definitions are done in a frame in which the parent particle s transverse momentum is zero .
the @xmath164 in the argument of @xmath155 in is obtained by making a transverse boost from the frame in which the nucleon has transverse momentum @xmath165 into a frame in which @xmath166 ( and the definition applies ) .
note that our lab frame corresponds to the photon - nucleus center - of - mass frame in which @xmath167 .
the polarization - dependent wigner functions are normalized as ( cf . ) @xmath168 as in @xcite , the correlation functions in can be parametrized in terms of the tmd distribution functions , of which the most relevant to the problem at hand are the unpolarized distribution @xmath169 and the sivers function @xmath170 .
using eqs .
we write [ tmd4 ] @xmath171 & = 2 \ , f_1^a ( { \bar x } , k_t ) + \ , \frac{2}{m_a } \ , { \hat z } \cdot ( { \underline{j } } \times { \underline{k}})\ , f_{1t}^{\perp a } ( { \bar x } , k_t ) \label{tmd4a } \\ { \mathrm{tr}}[\phi_n ( x , { \underline{k}}'- x \ , { \underline{p } } ; p , \sigma ) \ , \gamma^+ ] & = 2 \ , f_1^n ( x , |{\underline{k}}'- x \ , { \underline{p}}| ) + \notag \\ & + \ , \frac{2}{m_n } \ , { \hat z } \cdot \left ( { \underline{\sigma } } \times ( { \underline{k}}'- x \ , { \underline{p } } ) \right ) \
f_{1t}^{\perp n } ( x , |{\underline{k}}'- x \ , { \underline{p}}| ) , \label{tmd4b}\end{aligned}\ ] ] where we introduced the unpolarized quark tmds ( @xmath172 and @xmath173 ) and sivers functions ( @xmath174 and @xmath175 ) for the nucleus and nucleons respectively , along with the masses @xmath56 and @xmath10 of the nucleus and nucleons .
we may extract the sivers function of the nucleus @xmath174 by antisymmetrizing with respect to either the nuclear spin or the momentum @xmath17 of the produced quark : , which is indeed the case due to its @xmath176-symmetry properties . ] @xmath177 - ( { \underline{k } } \rightarrow -{\underline{k}}).\end{aligned}\ ] ] using in we write @xmath178 \ , d_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] - ( { \underline{k } } \rightarrow - { \underline{k } } ) .
\end{aligned}\end{aligned}\ ] ] we can decompose the quark correlator in a nucleon @xmath155 into the nucleon s unpolarized quark distribution @xmath173 and sivers function @xmath179 using . substituting this into yields @xmath180 \\ & \times \ , d_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] - ( { \underline{k } } \rightarrow - { \underline{k } } ) .
\end{aligned}\end{aligned}\ ] ] we can understand the sources of the @xmath176-odd nuclear sivers function @xmath174 by explicitly ( anti)symmetrizing the various terms on the right of . to start with , perform the nucleon spin sum @xmath181 in a basis parallel or antiparallel to the nuclear spin @xmath182 .
this can be done using the definitions @xmath183 where we will refer to @xmath184 as the distribution of unpolarized nucleons and to @xmath185 as the nucleon transversity distribution .
note that @xmath186 as follows from the definition and from . becomes @xmath187 \notag \\ & \times \ , d_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] - ( { \underline{k } } \rightarrow - { \underline{k}}).\end{aligned}\ ] ] now , in the terms with @xmath188 being subtracted , we also redefine the dummy integration variables @xmath189 , @xmath190 , and @xmath191 .
this leaves the fourier factors and the distribution functions @xmath192 , @xmath193 unchanged , giving @xmath194 - w_{unp } \left ( p^+ , -{\underline{p}},b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ , d_{{{\underline{y } } } { { \underline{x}}}}[+\infty , b^- ] \right ] \notag \\ & + \frac{1}{m_n } \ , { \hat z } \cdot \left ( { \underline{s } } \times ( { \underline{k}}'- x \ , { \underline{p } } ) \right ) \ , f_{1t}^{\perp n } ( x , |{\underline{k}}'- x \ , { \underline{p}}| ) \\ & \times \ , \left [ w_{trans } \left ( p^+ , { \underline{p } } , b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ , d_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] + w_{trans } \left ( p^+ , - { \underline{p } } , b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ , d_{{{\underline{y } } } { { \underline{x}}}}[+\infty , b^- ] \right ] \bigg\}. \notag\end{aligned}\ ] ] at this point it is convenient to explicitly ( anti)symmetrize the distribution functions with respect to @xmath195 and the wilson lines with respect to @xmath196 .
define @xmath197 and @xmath198 , \end{aligned}\end{aligned}\ ] ] where we have used the `` oam '' label to indicate that the preferred direction of transverse momentum in the antisymmetric case reflects the presence of net orbital angular momentum .
we can decompose @xmath109 into symmetric and oam parts for both the unpolarized distribution @xmath184 and the transversity distribution @xmath185 . using
the ( anti)symmetrized quantities in we can evaluate the factors in the square brackets of as @xmath199 & - w_{unp}(-{\underline{p}},b ) \
, d_{{{\underline{y } } } { { \underline{x}}}}[+\infty , b^- ] \bigg ] = \\ & = 2 \bigg ( w_{unp}^{oam}(p , b ) \ , s_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] + w_{unp}^{symm}(p , b ) \ ,
i \ , o_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] \bigg ) \\ \bigg[w_{trans}(p ,
b ) \ , d_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] & + w_{trans}(-{\underline{p}},b ) \ , d_{{{\underline{y } } } { { \underline{x}}}}[+\infty , b^- ] \bigg ] = \\ & = 2 \bigg ( w_{trans}^{symm}(p , b ) \ , s_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] + w_{trans}^{oam}(p , b ) \ ,
i \ , \ , o_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] \bigg ) \end{aligned}\end{aligned}\ ] ] giving @xmath200 + w_{unp}^{symm } \left ( p^+ , { \underline{p}},b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ ,
i \ , o_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] \right ] \notag \\ & + \frac{1}{m_n } \ , { \hat z } \cdot \left ( { \underline{s } } \times ( { \underline{k}}'- x \ , { \underline{p } } ) \right ) \ , f_{1t}^{\perp n } ( x , |{\underline{k}}'- x \ , { \underline{p}}| ) \\ & \times \ , \left [ w_{trans}^{symm } \left ( p^+ , { \underline{p } } , b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ , s_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] + w_{trans}^{oam } \left ( p^+ , { \underline{p } } , b^- , \frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ , i \ , o_{{{\underline{x } } } { { \underline{y}}}}[+\infty , b^- ] \right ] \bigg\}. \notag\end{aligned}\ ] ] altogether , the symmetry arguments presented above allow us to decompose the nuclear sivers function @xmath174 into four distinct channels with the right quantum numbers to generate the @xmath176-odd asymmetry .
these four channels correspond to the negative @xmath176-parity occurring in the nucleon distribution @xmath201 , in the quark sivers function of the nucleon @xmath175 , in the antisymmetric `` odderon '' rescattering @xmath202 , or in all three simultaneously .
we now will neglect the odderon contributions in . the way to understand this approximation is as follows . as shown in @xcite , the preferred direction generated by
odderon - type rescattering couples to transverse gradients of the nuclear profile function , @xmath203 .
the length scale over which these gradients become important is on the order of the nuclear radius ; these gradients are therefore @xmath204 suppressed ( in addition to an extra power of @xmath6 entering the lowest - order odderon amplitude corresponding to the triple - gluon exchange @xcite ) and are beyond the precision of the quasi - classical formula . neglecting the odderon channels in we arrive at @xmath205 .
\end{aligned}\end{aligned}\ ] ] shifting the integration variable @xmath206 we write @xmath207 .
\end{aligned}\end{aligned}\ ] ] to further simplify the obtained expression we need to impose a constraint on the transverse momentum of the nucleons .
consider the nucleus in its rest frame , as shown in .
the net oam @xmath208 of the transversely - polarized nucleus corresponds to the rotation of the nucleus around the spin axis ( the @xmath1-axis in ) .
the rotational invariance around the @xmath1-axis implies that the average magnitude of the rotational transverse momentum is constant for a given distance from the @xmath1-axis and for fixed @xmath1-coordinate .
( in appendix [ b ] we show that , as a consequence of @xmath209-symmetry , only rotational motion of the nucleons in the nucleus rest frame is allowed . )
consider a nucleon at the point @xmath210 in the @xmath211 coordinate system , as illustrated by the black circle in .
its 3-momentum is @xmath212 , where @xmath94 denotes some typical rotational momentum of a nucleon . after a longitudinal boost along the @xmath213-axis to the infinite - momentum frame of we
get the large light - cone component of the momentum to be @xmath214 the corresponding bjorken-@xmath1 is ( see ) @xmath215 where we have used @xmath216 . the @xmath217 constraint in ( cf .
) gives @xmath218 since @xmath219 is not a small number , in fact @xmath220 , we conclude that @xmath221 .
therefore , the magnitude of the rotational momentum in the nuclear rest frame is bounded by @xmath222 from above .
the typical transverse momentum @xmath9 in , being boost - invariant , is also bounded by the nucleon mass from above , @xmath223 .
since we assume that @xmath7 is perturbatively large , @xmath224 , we do not consistently resum all powers of @xmath225 .
( saturation approach resums mainly @xmath226-enhanced power corrections , that is , powers of @xmath227 , but not powers of @xmath228 . )
the bound provides us with the condition on when the sidis process on the nucleon highlighted in can take place .
violation of this bound would imply that sidis on that nucleon is kinematically prohibited , and consequently sidis may take place only on some of the other nucleons in the nucleus . while such situation where the nucleus is spinning so fast that sidis is only possible on a subset of its nucleons
is highly unlikely in the real physical experiments , this presents a theoretical example where the sivers function would , in fact , depend on the direction of @xmath165 and , hence , of spin @xmath229 , presumably through even powers of @xmath230 .
while such dependence is impossible for spin-@xmath231 particles such as protons @xcite , it has been considered for targets with different spin @xcite ; in our case it arises due to the classical model at hand with the value of spin @xmath232 not restricted to @xmath231 . to avoid potential formal complications and unrealistic effects associated with large rotational momentum , below we will assume that @xmath233 such that the bound is satisfied . without such assumption
, would be our final result for the sivers function in the quasi - classical approximation .
we see that we have to limit the calculation to the lowest non - trivial power of @xmath234 contributing to the sivers function . expanding in the powers of @xmath165 to the lowest non - trivial order , and remembering that @xmath201 is an odd function of @xmath165 we obtain @xmath235 .
\end{aligned}\end{aligned}\ ] ] is our main formal result .
it relates the sivers function of a nucleus to the quark tmd and quark sivers function in a nucleon .
it shows that within the quasi - classical approximation , there are two leading channels capable of generating the sivers function of the composite nucleus : 1 . : an unpolarized nucleon in a transversely polarized nucleus with a preferred direction of transverse momentum generated by the oam of the nucleus has a quark knocked out of its symmetric @xmath173 transverse momentum distribution which rescatters coherently on spectator nucleons .
the multiple rescatterings bias the initial knockout process to happen near the `` back '' of the nucleus , where , due to oam motion of the nucleons , the outgoing quark gets an asymmetric distribution of its transverse momentum , generating stsa .
( see or left panel of below . ) 2 . :
a polarized nucleon with its preferred transverse spin direction inherited from the nucleus has a quark knocked out of its sivers @xmath175 distribution which rescatters coherently on spectator nucleons .
the single spin asymmetry is generated at the level of the `` first '' nucleon , and , unlike the oam channel , the presence of other nucleons is not essential for this channel ( see ) .
the oam and transversity channels are depicted in in terms of their space - time structure and feynman diagrams .
the diagrams resummed in arriving at are the square of the graph shown in the left panel of ( oam channel ) and the diagram looking like the interference between the two panels in ( transversity channel ) .
the difference between the two channels outlined above is in the first `` knockout '' interaction : the oam channel couples to quark tmd , while the transversity channel couples to the nucleon sivers function . at the lowest order in perturbation theory
the two functions are illustrated in : indeed sivers function shown in the panel b of requires at least one more rescattering as compared to the quark tmd in panel a , according to the conventional wisdom @xcite .
note that , in the oam channel , the unpolarized quark distribution @xmath173 enters parametrically at @xmath236 if calculated at the lowest - order in the perturbation theory ( see panel a in ) , which is @xmath237 in the saturation power counting ( where @xmath5 ) . in the transversity channel ,
the nucleonic sivers function @xmath175 enters at @xmath238 at the lowest order in perturbation theory , because it requires an extra @xmath239 gluon to be exchanged with the same nucleon to obtain the necessary lensing effect @xcite ( see panel b in ) .
the transversity channel is therefore @xmath240 in the saturation power counting and is subleading by @xmath239 to the oam channel in this sense . in @xmath173
runs with some non - perturbative momentum scale , and is large , @xmath241 ; however , a simple application of the blm @xcite prescription to the calculation of @xcite can show that in @xmath242 the two powers of the coupling run as @xmath243 . while one of the couplings is also non - perturbatively large , the other one is perturbatively small for @xmath244 , indicating suppression .
] indeed the non - trivial transverse motion of nucleons due to oam should be present for the oam channel to be non - zero : this channel is leading only if there is an oam . in our estimate here
we have assumed that the net spin of our `` nucleons '' scales linearly with the atomic number , @xmath245 ; perhaps a more realistic ( both for protons and nuclei ) slower growth of @xmath246 with @xmath0 would introduce extra @xmath0-suppression for the transversity channel . despite the transversity channel being subleading ,
it is more dominant than the @xmath247 corrections we neglected when arriving at the quasi - classical formula ( again , for @xmath248 ) .
order @xmath249 quantum corrections to the oam channel also enter at the same order as the nucleonic sivers function and are also within the precision of the formalism . an essential role is played by the rescattering factor @xmath250 $ ] in the oam channel .
the wigner function due to orbital motion of nucleons around the axis of the transverse spin is an odd function of the longitudinal coordinate @xmath251 in the rest frame of the nucleus , @xmath252 which follows simply from the fact that in the left panel of we have as many nucleons moving outside the page to the left of the nuclear center as there are nucleons moving into the page to the right of the nuclear center . in appendix [ b ]
we show how the result can be obtained by requiring that our `` nucleus '' is @xmath209-symmetric .
the @xmath134-integral of the first term in the curly brackets of would have been zero , if it was not for the @xmath134-dependent factor of @xmath253 $ ] .
this multiple - rescattering factor approaches @xmath254 for @xmath134 values near the `` back '' of the nucleus ( right end of the nucleus in ) and is a monotonically decreasing function of @xmath134 . due to this factor
, different @xmath134 regions contribute differently to the integral , making it non - zero .
the region near the `` back '' of the nucleus dominates , which has a physical interpretation that it is easier for the quark to escape the nucleus if it is produced near the edge .
hence we arrive at the interpretation of the sidis in the oam channel outlined in the introduction : the quarks are produced predominantly toward the `` back '' of the nucleus , where the nucleons rotate preferentially into the page ( see left panel of or ) .
therefore , the quark has more transverse momentum into the page than out of the page , which leads to stsa for the produced quarks . to complete
we need to construct an expression for the nuclear spin @xmath255 .
the oam of the nucleons in the nucleus from in the nuclear rest frame is @xmath256 where @xmath257 , @xmath258 , and @xmath259 is the wigner distribution in the rest frame of the nucleus expressed in terms of 3-vectors @xmath260 and @xmath261 . to boost this into the infinite momentum frame of we define the pauli - lubanski vector of the nuclear spin @xmath262 where @xmath263 with @xmath264 and @xmath265 the expectation values of the oam and spin generators of the lorentz group in the nuclear state .
the oam generator is @xmath266 as usual , with the hat denoting operators .
the nuclear oam four - vector is then defined by @xmath267 note that @xmath268 in are the momentum operators of the nucleons , while @xmath269 in eqs . and is the net momentum of the whole nucleus . in the rest frame of the nucleus
gives @xmath270 as expected ( for @xmath271 ) .
the nuclear oam four - vector can then be written as @xmath272 in the infinite momentum frame of the nucleus .
since boosts preserve transverse components of four - vectors , the boost along the @xmath273-axis of the nucleus in would preserve its oam three - vector @xmath208 . hence gives us the transverse components of oam in the infinite momentum frame as well .
we thus write @xmath274,\end{aligned}\ ] ] where the integration over @xmath94 and @xmath126 needs to be carried out in the nucleus rest frame . combining eqs . with allows one to extract the sivers function @xmath174 of the nucleus .
we will now illustrate the properties of the sivers function by studying a specific simplified example .
consider the model of the nucleus as a non - relativistic rigid rotator , with the rotational momentum in its rest frame being much smaller than the nucleon mass , @xmath275 .
the corresponding classical wigner distribution is ( cf . ) @xmath276 where @xmath277 is the extent of the nucleus in the @xmath134-direction at @xmath278 ( with @xmath279 ) , and @xmath280 is the maximum value of the rotational momentum at a given @xmath281 . in writing down the distribution we have neglected possible longitudinal rotational motion of the nucleons , which is justified in the @xmath275 limit .
we also assume that a fraction @xmath282 of the nucleons in the nucleus are polarized in the @xmath283-direction , such that their net spin is @xmath284 and ( see ) @xmath285 substituting eqs . and into and integrating over @xmath160 and @xmath165 yields @xmath286,\end{aligned}\ ] ] where we also replaced @xmath287 by @xmath288 and @xmath146 by @xmath289 .
to further simplify we need to make specific assumptions about the form of @xmath173 and @xmath175 .
inspired by the lowest - order expressions for both quantities @xcite we write @xmath290 where @xmath291 and @xmath292 are some @xmath1-dependent functions and @xmath293 is an infrared cutoff .
inserting into and integrating over @xmath294 yields @xmath295.\end{aligned}\ ] ] in the classical mv / gm approximation the ( symmetric part of the ) dipole scattering matrix is @xcite @xmath296 = \exp \left[-\frac{1}{4 } |{\underline{x}}- { \underline{y}}|^2 \ , q_s^2 \left(\frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \
, \left(\frac{r^- ( { \underline{b } } ) - b^-}{2 r^- ( { \underline{b}})}\right ) \ , \ln\frac{1}{|{\underline{x } } -{\underline{y}}| \lambda } \right],\end{aligned}\ ] ] where @xmath297 and the quark saturation scale is @xmath298 with the nuclear profile function @xmath299 as usual @xmath300 is the number of colors and @xmath301 is the casimir operator of su@xmath302 in the fundamental representation .
in arriving at we assumed that the nuclear density is constant within the nucleus , such that @xmath303 employing along with eqs . and , and neglecting all logarithms @xmath304 ( which is justified as long as @xmath7 is not too much larger than @xmath305 @xcite ) we can integrate over @xmath134 and @xmath306 obtaining @xmath307 \notag \\ & + { \alpha_s}\ , \beta \ , m_n \ , c_2 \ , e^{-k_t^2/q_s^2 ( { { \underline{b } } } ) } \bigg\},\end{aligned}\ ] ] where now @xmath308 and @xmath309 . the @xmath108-integral appears to be rather hard to perform for a realistic spherical nucleus : we leave expression in its present form . to obtain a final expression for the sivers function we need to determine the spin of the nucleus @xmath232 .
for a rigid rotator spinning around the @xmath147-axis with the maximum nucleon momentum @xmath310 we readily get @xmath311 in the nuclear rest frame . using this in along with @xmath312 we obtain @xmath313 inserting into gives @xmath314 + { \alpha_s}\beta \ , m_n \ , c_2 \ , e^{-k_t^2/q_s^2 ( { { \underline{b } } } ) } \bigg\}.\end{aligned}\ ] ] is our final expression for the sivers function of a nucleus in the quasi - classical approximation with the rigid rotator model for the nucleus and @xmath7 not too much larger than @xmath305 . analyzing this expression
we see that the oam term ( the first term in the curly brackets ) does change sign as a function of @xmath7 , while the sivers density term ( the second term in the curly brackets of ) is positive - definite .
still the first term in the curly brackets is positive for most of the @xmath7 domain , in agreement with the qualitative argument in the introduction corresponding to quarks being produced preferentially into the page in . to study the @xmath315 case we have to return to : this time we do not neglect the logarithms .
large @xmath7 limit implies that @xmath316 is small , and we need to expand the exponential in to the lowest non - trivial ( contributing ) order in each term in . for the sivers density term
this corresponds to replacing the exponent by @xmath254 .
the remaining evaluation is easier to carry out in , which yields an intuitively clear formula @xmath317 in the first term in the curly brackets of we need to expand the exponential in @xmath253 $ ] one step further , obtaining after some straightforward algebra for the whole sidis sivers function @xmath318 \notag \\ = & \frac{\beta}{\beta+ \tfrac{8}{5 } \
, p_{max } \ , r } \left [ - \frac{4 \ , { \alpha_s}\ , m_n \ , { \bar x } \ , c_1}{3 \beta \ , k_t^6 } \ln \frac{k_t^2}{\lambda^2 } \int d^2 b \
t({\underline{b } } ) \ , p_{max } ( { \underline{b } } ) \ , q_s^2 ( { \underline{b } } ) + \frac{a \ , { \alpha_s}^2 \ , m_n^2 \ , c_2}{k_t^4 } \ , \ln \frac{k_t^2}{\lambda^2 } \right].\end{aligned}\ ] ] since @xmath319 we see that the oam channel contribution in ( the first term ) is proportional to @xmath320 , while the transversity channel contribution ( the second term ) scales as @xmath321 . ( note that @xmath322 , such that powers of @xmath1 do not generate suppression . ) assuming that @xmath13 ( see the discussion following ) we observe that the ratio of the oam to transversity channel contributions is @xmath323 .
( note that for @xmath13 the prefactor of gives a factor @xmath324 multiplying both terms , but not affecting their ratio . )
we conclude that the oam channel dominates for @xmath325 that is both inside the saturation region , and in a sector of phase space outside that region . for @xmath14
the transversity channel dominates , mapping onto the expected perturbative qcd result .
while the main aim of this calculation is to model a nucleon at high energies , a few comments are in order about the application of this rigid rotator toy model to a realistic nucleus .
certainly a classical rigid rotator is a poor model for a real nucleus ; a better approach would be to use our general result with the wigner distribution @xmath135 given by the realistic single - particle wave functions taken from nuclear structure calculations .
in such realistic cases , the total angular momentum @xmath232 of the nucleus is typically small , and the fraction @xmath282 that comes from the nucleons spins is also small due to nucleon spin pairing . if one were to approximate a real nucleus with this rigid rotator toy model , appropriately small @xmath232 and @xmath282 would need to be used in and
. the smallness of the total oam @xmath232 does not affect the sivers function @xmath326 because the magnitude is contained in the prefactor @xmath327 and cancels in the @xmath328 ratio .
the smallness of the spin contribution @xmath329 , however , would suppress the transversity channel and ensure the dominance of the oam term .
but regardless of its applicability to a real nucleus , the rigid rotator toy model illustrates the ability of this formalism to capture the interplay of spin and angular momentum in a dense system at high energy .
we now wish to perform a similar analysis for the drell - yan process @xmath330 , where the antiquark from the unpolarized hadron scatters on the transversely polarized hadron / nucleus , producing a space - like photon which later decays into a di - lepton pair .
the annihilation sub - process @xmath331 is related to the sidis process by time reversal , which leads to the famous prediction @xcite that the sivers functions entering observables in the two processes should have equal magnitudes and opposite signs .
the lowest - order drell - yan annihilation process is shown in fig .
[ fig : dy1 ] , without including initial - state rescattering of the antiquark on nuclear spectators .
labeling the momenta as in fig .
[ fig : dy1 ] and following along the same lines as for sidis , we can write the kinematics in the @xmath332 center - of - mass frame as @xmath333 where @xmath334 as with sidis , we are working in the kinematic limit @xmath335 , with @xmath336 . again
, we can compare the coherence lengths @xmath337 of the antiquark and @xmath338 of @xmath339 analogous to sidis , this shows that the coherence length of the incoming antiquark is large ; in fact it would be infinite if we dropped the quark mass @xmath340 as we have elsewhere in the calculation .
we conclude that the long - lived antiquark is able to rescatter off of many nucleons before it finally annihilates a quark .
the annihilation occurs locally , as indicated by the short coherence length of the virtual photon , and thereafter the produced photon / dilepton system does not rescatter hadronically .
this again motivates the resummation of these initial state rescatterings into a wilson line dipole trace .
the entire drell - yan ( dy ) process in the quasi - classical approximation is shown in at the level of the scattering amplitude : the incoming antiquark coherently scatters on the nucleon in the transversely polarized nucleons , until the last interaction in which the virtual photon is produced , which later generated the di - lepton pair .
physics the dy process is dominated by the @xmath341 scattering @xcite , which is @xmath342-subleading compared to the diagram in : since in our calculation @xmath52 , we will neglect the @xmath343 process here as an @xmath342 correction . ] by analogy with in sidis we write the following relation for the quark correlators in dy , @xmath344 = a \int & \frac{d p^+ \ , d^2 p \ , d b^-}{2(2\pi)^3 } \int \frac{d^2 k ' \ , d^2 x \ , d^2 y}{(2\pi)^2 } \ , e^{i { \underline{k ' } } \cdot ( { \underline{x } } - { \underline{y } } ) } \ , \sum_\sigma w_n^\sigma \left(p , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \\ & \times { \mathrm{tr}}[\phi_n ( x , { \underline{q } } -{\underline{k}}'-x \ , { \underline{p } } ) ; p , \sigma ) \gamma^+ ] \ , d_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] , \end{aligned}\end{aligned}\ ] ] where @xmath345 = \left\langle \frac{1}{n_c } \ , \mbox{tr } \left [ v_{{\underline{y } } } [ b^- , -\infty ] \ , v^\dagger_{{\underline{x } } } [ b^- , -\infty ] \right ] \right\rangle\end{aligned}\ ] ] and the quark correlators are defined by equations similar to and , but now using a different gauge link : @xmath346 @xmath347 here @xmath348 .
is illustrated in .
the main difference compared to is that now @xmath349 and @xmath350 . projecting out the dy sivers function of the nucleus @xmath351 using gives @xmath352 \ , d_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] - ( { \underline{q } } \rightarrow -{\underline{q}}).\end{aligned}\ ] ] with the help of we write @xmath353 \notag \\ & \times \ , d_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] - ( { \underline{q } } \rightarrow -{\underline{q}}).\end{aligned}\ ] ] performing the spin sums gives @xmath354 \ , d_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] - ( { \underline{q } } \rightarrow -{\underline{q}}).\end{aligned}\ ] ] in the terms being subtracted in with @xmath355 , we can also reverse the dummy integration variables @xmath356 , @xmath357 , and @xmath358 .
this leaves the fourier factor and the distributions @xmath173 , @xmath175 invariant , giving @xmath359 - w_{unp } \left(p^+ , - { \underline{p } } , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \ , d_{{{\underline{x } } \ , { \underline{y } } } } [ b^- , -\infty ] \bigg ] \notag \\ & + \frac{1}{m_n } { \hat z } \cdot \left({\underline{s } } \times ( { \underline{q } } -{\underline{k}}'-x \ , { \underline{p}})\right ) \ , f_{1t}^{\perp n } ( x , |{\underline{q } } -{\underline{k}}'- x { \underline{p}}|_t ) \\ & \times \bigg [ w_{trans } \left(p^+ , { \underline{p } } , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \ , d_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] + w_{trans } \left(p^+ , - { \underline{p } } , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \ , d_{{{\underline{x } } \ , { \underline{y } } } } [ b^- , -\infty ] \bigg ] \bigg\}. \notag\end{aligned}\ ] ] we recognize the factors in brackets from the sidis case , rewriting as @xmath360 + w_{unp}^{symm } \left(p^+ , - { \underline{p } } , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \ ,
, o_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] \bigg ] \notag \\ & + \frac{1}{m_n } { \hat z } \cdot \left({\underline{s } } \times ( { \underline{q } } -{\underline{k}}'-x \ , { \underline{p}})\right ) \ , f_{1t}^{\perp n } ( x , |{\underline{q } } -{\underline{k}}'- x { \underline{p}}|_t ) \\ & \times \bigg [ w_{trans}^{symm } \left(p^+ , { \underline{p } } , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \ , s_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] + w_{trans}^{oam } \left(p^+ , - { \underline{p } } , b^- , \frac{{\underline{x } } + { \underline{y}}}{2 } \right ) \
, i \ , o_{{{\underline{y } } \ , { \underline{x } } } } [ b^- , -\infty ] \bigg ] \bigg\}. \notag\end{aligned}\ ] ] as before , we drop contributions from the odderon @xmath361 as being outside the precision of the quasi - classical formula to get @xmath362 . \notag\end{aligned}\ ] ] since the rotational momentum of the nucleons @xmath9 is assumed to be small , we have to expand in it to the lowest non - trivial order . shifting the integration variable @xmath363 in and expanding the exponential to the lowest non - trivial order in @xmath9
we obtain ( cf . ) @xmath364 , \notag\end{aligned}\ ] ] where we have also interchanged @xmath358 and @xmath365 .
is our main formal result for the dy sivers function .
we again see that the sivers function in dy can arise through two distinct channels in this quasi - classical approach : the oam channel that contains its preferred direction in the distribution @xmath366 and the transversity / sivers density channel that generates is preferred direction through a local lensing mechanism @xmath175 . to demonstrate the importance of the wilson lines for the sivers function , for the moment , let us ignore the contribution of the wilson lines associated with initial - state rescattering in . without any such initial - state interactions , the nucleonic sivers function is zero , @xmath367 @xcite , leaving @xmath368 which vanishes after @xmath134 integration because of the rotational and @xmath209-symmetry conditions .
now that the dy sivers function is expressed in the same form as the sivers function for sidis , we can compare both expression to see how the nuclear sivers functions have changed between sidis and dy and understand the origin of the sidis / dy sign - flip relation @xcite @xmath369 first , we notice that the transversity / sivers density channel ( the second term in the curly brackets ) has changed signs as required between and .
mathematically , this occurs because of the @xmath365 interchange , simply because the momentum going into the wilson line in sidis corresponds to the momentum coming from the wilson line in dy ( cf . figs .
[ fig : tmd1 ] and [ fig : dy2 ] ) .
the transversity / sivers density channel contribution thus automatically satisfies the sign - flip relation .
the oam channel contribution to is more subtle ; although the prefactor has not changed as compared to , the longitudinal coordinate @xmath134 integral entering for dy can be modified using @xmath370 substitution along with to give @xmath371 = - \int db^- \
, w_{unp}^{oam } ( p , b ) \ , s_{{{\underline{x } }
\ , { \underline{y } } } } [ - b^- , -\infty ] .
\end{aligned}\ ] ] when evaluating the dipole @xmath246-matrix we neglect the polarization effects as being energy suppressed .
therefore , for the purpose of this @xmath246-matrix , the nucleus has a rotational symmetry around the @xmath213-axis ( see for axes labels ) .
we thus write @xmath372 \overset{pt}{= } s_{-{{\underline{x } } , \
, -{\underline{y } } } } [ + \infty , b^- ] \overset{z - rotation}{= } s_{{{\underline{x } } \ , { \underline{y } } } } [ + \infty , b^-],\end{aligned}\ ] ] where @xmath213-rotation denotes a half - revolution around the @xmath213-axis . using in we arrive at @xmath373}^{\mathrm{dy } } & = - \overbrace{\int db^-
\ , w_{unp}^{oam } ( p , b ) \ , s_{{{\underline{x } } \ , { \underline{y } } } } [ + \infty , b^- ] } ^{\mathrm{sidis}}. \end{aligned}\end{aligned}\ ] ] one can also simply see that is true by using the quasi - classical gm / mv dipole @xmath246-matrix from on its right - hand - side , along with @xmath374 = \exp \left[-\frac{1}{4 } |{\underline{x}}- { \underline{y}}|^2 \ , q_s^2 \left(\frac{{{\underline{x } } } + { { \underline{y}}}}{2 } \right ) \ , \left(\frac{b^- + r^-}{2 r^-}\right ) \ , \ln\frac{1}{|{\underline{x } } -{\underline{y}}| \lambda } \right]\end{aligned}\ ] ] on its left - hand - side .
we conclude that the oam channel contributions to the sidis sivers function and the dy sivers function also satisfy the sign - flip relation .
therefore , for any wigner distribution @xmath135 , the sivers functions at the quasi - classical level for sidis and for dy are equal in magnitude and opposite in sign , .
this statement is a direct consequence of the invariance of @xmath135 under rotations and @xmath209-reversal , , and it mirrors in this context the original derivation by collins @xcite .
the advantage of our approach here , apart from providing the explicit formal results and , is in the new physical interpretation of the transverse spin asymmetry in the oam channel .
as described in the introduction following , the incoming antiquark is more likely to interact with the `` front '' of the nucleus , due to shadowing effects , thus scattering on the nucleon moving out of the page in .
this is justified by the @xmath375 $ ] function in ( see also ) , which is largest for @xmath376 .
thus the virtual photon is produced preferentially out of the page ; this leads to a non - zero stsa in dy .
the sign reversal relation follows from comparing figs .
[ dy_fig ] and [ sidis_fig ] : in dy the particles are produced preferentially left - of - beam , while in sidis the produced hadrons come out mainly right - of - beam . the rigid - rotator toy model of sec . [ toy_model ]
can also be constructed for dy sivers function . however , due to the sign - reversal relation we can read off the answer for the dy sivers function in the rigid - rotator model as being negative of that in for moderate @xmath7 and negative of for @xmath315 .
all the conclusions about the relative importance of the two contributing channels remain the same .
the main goal of this work was to construct sidis and dy sivers functions in the quasi - classical gm / mv approximation , which models a proton as a large nucleus , and which we modified by giving the nucleus a non - zero oam .
the main formal results are given in eqs .
( sidis ) and ( dy ) .
we showed that there are two main mechanisms generating the quasi - classical sivers function : the oam channel and the transversity channel .
the former is leading in saturation power counting ; it also dominates for @xmath377 , that is both inside and , for @xmath378 , outside of the saturation region . at higher @xmath7
the transversity channel dominates . in the future
our quasi - classical calculation can be augmented by including evolution corrections to the sivers function , making the whole formalism ready for phenomenological applications , similar to the successful use of nonlinear small-@xmath1 evolution equations @xcite to description ( and prediction ) of high energy scattering data @xcite .
perhaps just as important , we constructed a novel physical mechanism of the stsa generation .
this is the oam channel .
the oam mechanism , while diagrammatically very similar to the original bhs mechanism @xcite , provides a different interpretation from the lensing effect @xcite or the color - lorentz force of @xcite .
the oam mechanism is described in the introduction , in the discussion around figs .
[ dy_fig ] and [ sidis_fig ] .
it is based on interpreting the extra rescattering proposed by bhs as a shadowing - type correction .
the stsa is then generated by the combination of the oam and shadowing .
the shadowing makes sure the projectile interacts differently with the front and the back of the target , generating the asymmetry of the produced particles . while shadowing is a high - energy phenomenon , and our calculation
was done in the high - energy approximation @xmath379 ( though for @xmath71 ) , it may be that the oam mechanism for generating stsa is still valid for lower - energy scattering , though of course the formulas derived above would not apply in such regime . at lower energies
the difference between the interactions of the projectile with the front / back of the target may result from , say , energy loss of the projectile as it traverses the target .
again , combined with the target rotation this would generate stsa , and , hence , the sivers function . the formalism needed to describe such a low - energy process would be quite different from the one presented above ; moreover , the correct degrees of freedom may not be quarks and gluons anymore .
however , the main physics principle of combining oam with the difference in interaction probabilities between the projectile and front / back of the target to generate stsa may be valid at all energies .
returning to higher energies and the derived formulas and , let us point out that these results , when applied to experimental data , may allow one to determine the distribution of intrinsic transverse momentum @xmath380 of partons in the hadronic or nuclear target , along with the transversity / sivers function density in the target .
this would complement the existing methods of spatial imaging of quarks and gluons inside the hadrons and nuclei @xcite , providing a new independent cross - check for those methods .
we would like to thank daniel boer , dick furnstahl , leonard gamberg and feng yuan for useful correspondence . this research is sponsored in part by the u.s .
department of energy under grant no .
de - sc0004286 .
the aim of this appendix is to justify the result given in . to study the interplay between the local `` knockout '' channel of deep inelastic scattering and the coherent multiple rescattering on the nuclear remnants , it is illustrative to consider a minimal case with both features .
this process , shown in fig .
[ fig : dis2 ] , consists of the knockout sub - process followed by a single rescattering on a different quark from a second nucleon in the nucleus .
rescattering on a second nucleon receives a combinatoric enhancement of order @xmath381 compared to rescattering on the same nucleon ; the former is @xmath240 in the saturation power counting , while the latter is @xmath239 . the minimal sidis process containing both the `` knockout '' of a quark from the nuclear wave function and rescattering on a different quark from a second nucleon .
the short thick vertical line indicates that the pole of the intermediate quark propagator is picked up in the calculation . , height=151 ] the total sidis amplitude @xmath382 depicted in fig .
[ fig : dis2 ] consists of a loop integral connecting the mean - field single - particle wave functions @xmath383 of the nucleus to a scattering amplitude @xmath384 denoting both the `` knockout '' and rescattering processes : @xmath385 where a sum over spins and colors of the participating quarks is implied . squaring both sides of and integrating out the final - state momenta @xmath386 and @xmath387 gives @xmath388 ^ 2 \ , ( p_1^+ + q^+ ) \ ,
p_2^+ } |m_{tot}|^2 \\ = & \int \frac{d p_1^+ \ , d^2 p_1 \ ,
dp_2^+ \ , d^2p_2}{[2 \ , ( 2 \pi)^3]^2 \ , ( p_1^+ + q^+ ) \
, p_2^+ } \frac{d r^+ \ , d^2 r}{2 \ , ( 2 \pi)^3 } \frac{d r^{\prime + } \ , d^2 r'}{2 \ , ( 2 \pi)^3 } \frac{a \ , ( a -1 ) \ , \left ( p^+ \right)^2}{\sqrt{(p_1^+ - r^+ ) \ , ( p_2^+ + r^+ ) \ , ( p_1^+ - r^{\prime + } ) \ , ( p_2^+ + r^{\prime + } ) } } \notag \\ & \times \int d b_1 ^ - \ , d^2 b_1 \ , d b_2 ^ - \ , d^2 b_2 \ , e^{- i \ , ( r - r')\cdot(b_1 - b_2 ) } \ , w \left(p_1 - \frac{r + r'}{2},b_1 \right ) w \left(p_2 + \frac{r + r'}{2 } , b_2 \right ) \notag \\ & \times
m_{k+r}(p_1 - r , p_2 + r , q , k , r ) \ , m^*_{k+r}(p_1 - r',p_2 + r',q , k , r ' ) , \notag \ ] ] where we have employed the wigner distributions defined in above and summed over all pairs of nucleons .
is still far from because in we do not have the amplitude squared : instead we have the product of @xmath384 and @xmath389 with different arguments .
it is easier to further analyze the expression separately for the transverse and longitudinal degrees of freedom .
we proceed by taking the classical limits , in which case the wigner distributions give us the position and momentum distributions of nucleons simultaneously . moreover , for the large nucleus at hand the wigner distributions depend on @xmath390 and @xmath391 weakly over the perturbatively short distances associated with the feynman diagrams .
we thus define @xmath392 and @xmath393 and write @xmath394 now the difference in the arguments of @xmath384 and @xmath389 is only in the longitudinal momenta @xmath395 and @xmath396 . to integrate over these momenta
we notice that , as follows from , in the high energy kinematics at hand the leading contribution to the amplitude @xmath384 comes from the region where @xmath397 . in this regime
we combine eqs . and to write @xmath398 ^ 2 \ , ( p_1^+ + q^+ ) \
, p_2^+ } \frac{d r^+ \ , d r^{\prime + } \ , d^2 r}{4 \ , ( 2 \pi)^4 } \frac{\left ( p^+ \right)^2}{p_1^+ \ ,
p_2^+ } d b_1 ^ - \ , d b_2 ^ - d^2 b \ ,
e^{- i \tfrac{1}{2 } \ , ( r^+ - r^{\prime + } ) \ , ( b_1 ^ - - b_2 ^ - ) } a \ , ( a -1 ) \notag \\ & \times \ , w \left(p_1^+ , { { \underline{p}}}_1 - { { \underline{r } } } , b_1 ^ - , { { \underline{b } } } \right ) w \left(p_2^+ , { { \underline{p}}}_2 + { { \underline{r } } } , b_2 ^ - , { { \underline{b } } } \right ) \ , m_{k+r}(p_1^+ , { { \underline{p}}}_1 - { { \underline{r } } } , p_2^+ , { { \underline{p}}}_2 + { { \underline{r } } } , q , k , r^+ , { { \underline{r } } } ) \notag \\ & \times \ , m^*_{k+r}(p_1^+ , { { \underline{p}}}_1 - { { \underline{r } } } , p_2^+ , { { \underline{p}}}_2 + { { \underline{r } } } , q , k , r^{\prime + } , { { \underline{r}}}).\end{aligned}\ ] ] in the @xmath397 kinematics the amplitude @xmath384 contains only one pole in @xmath395 resulting from the denominator of the @xmath399 quark propagator ( cf .
@xcite ) . we can thus write @xmath400 where @xmath401 denotes the rest of the diagram which does not contain singularities in @xmath395 in the @xmath402 approximation .
( note that @xmath401 also contains the numerator of the @xmath399 quark propagator . ) since @xmath403 we can use to integrate over @xmath395 , @xmath404 here we assumed that @xmath405/k^- \approx 0 $ ] in our kinematics .
after putting the @xmath399 quark propagator on mass shell the amplitude @xmath401 factorizes into a product of separate amplitudes for knockout @xmath406 and rescattering @xmath407 @xcite , as employed in , where the sum over quark polarizations and colors is implicit . with the help of ( and a similar one for the @xmath396-integration of @xmath389 )
we write @xmath408 ^ 2 \ , ( p_1^+ + q^+ ) \ ,
p_2^+ } \frac{d^2 r}{4 \ , ( 2 \pi)^2 } \frac{a \ , ( a-1 ) \ , \left ( p^+ \right)^2}{p_1^+ \ , p_2^+ \ , ( k^-)^2 } \ , d b_1 ^ - \ , d b_2 ^ - \ , d^2 b \ , \theta ( b_1 ^ - - b_2 ^ - ) \notag \\ \times \ , w \left(p_1^+ , { { \underline{p}}}_1 - { { \underline{r } } } , b_1 ^ - , { { \underline{b } } } \right ) \ , w \left(p_2^+ , { { \underline{p}}}_2 + { { \underline{r } } } , b_2 ^ - , { { \underline{b } } } \right ) \notag \\ \times \ , |m_k(p_1 -r , q , k - r)|^2 \ , |m_r(p_2+r , k - r , k , r)|^2.\end{aligned}\ ] ] defining the energy - independent ( at the quasi - classical level ) rescattering amplitude by @xcite @xmath409 and denoting the average of this object in the wigner distribution by the angle brackets @xmath410 we rewrite as @xmath411 in arriving at we have shifted the momentum @xmath412 .
we now define the `` energy - independent '' total and `` knockout '' amplitudes @xcite @xmath413 employing the fourier transform we reduce to @xmath414 with @xmath415 comparing to we see that , just like in all high energy qcd scattering calculations @xcite the rescattering can be factored out into a multiplicative factor in the transverse coordinate space .
similar to the above one can show that all further rescatterings would only introduce more multiplicative factors . defining a somewhat abbreviated notation @xmath416
we see that reduces to , as desired .
the above discussion also demonstrates how multiple rescatterings factorize in the transverse coordinate space : in the high energy kinematics they are included through the wilson lines of eqs . and .
the wilson line correlator @xmath417 $ ] from contains a @xmath134-ordered product of multiple rescattering factors @xmath418 from all the interacting nucleons @xcite .
the decompositions and essentially break the wilson line operator @xmath419 in the definition into two parts : the coherent rescattering @xmath253 $ ] on other spectator nucleons which is a leading - order contribution in the saturation power counting , and the subleading lensing interaction with the same nucleon which generates @xmath175 . if we neglect the wilson line operator @xmath419 entirely , then we know that the sivers function of the nucleus @xmath174 must vanish , as first proved by collins @xcite .
but if we drop @xmath175 and @xmath253 $ ] from , we do not obviously get zero : the right - hand side of this equation must vanish for wave functions described by @xmath366 that are @xmath209 eigenstates @xcite ; we can see this explicitly by considering the constraints on @xmath420 due to rotational invariance and @xmath209 symmetry .
it is most convenient to enumerate the rotational symmetry properties of the nucleon distribution @xmath420 in the rest frame of the nucleus , using a cylindrical vector basis coaxial to the transverse spin vector @xmath182 .
this basis @xmath422 is shown in fig .
[ fig : w_rot ] and is defined by @xmath423 where @xmath424 and @xmath425 .
first , the distribution must be symmetric under rotations about the transverse spin @xmath426 , which are easy to express in this cylindrical basis : @xmath427 second , if the nucleus is in a @xmath209-symmetric eigenstate of the qcd hamiltonian , then @xmath420 should be invariant under @xmath209 transformations .
these transformations reverse the coordinates @xmath428 and pseudovectors like the spin @xmath429 , but leave the momentum vector @xmath94 unchanged . using this transformation , together with rotational invariance as shown in fig .
[ fig : pt_rot ] we obtain @xmath430 where the rotation @xmath431 is a half - revolution in the @xmath432-plane .
illustration of the @xmath209 transformation and rotational symmetry in the rest frame used in . left panel : illustration of the momentum flow represented by @xmath420 .
center panel : under a @xmath209 transformation , the spins @xmath433 and coordinate @xmath126 are reversed , but the momentum @xmath94 is invariant .
right panel : rotation of the center panel by @xmath434 about the vector @xmath435 returns the distribution to its original position @xmath126 , with @xmath436 and @xmath437 having been reversed . ]
this means that in a @xmath209 eigenstate with transverse spin @xmath426 , the only allowed direction of net momentum flow corresponds to the azimuthal orbital momentum @xmath438 and explains the naming convention @xmath201 in .
the distributions that enter and , however , are the ( anti)symmetrized distributions under reversal of the transverse momenta @xmath439 . for these purposes , it is more convenient to write the distribution @xmath440 in terms of the cartesian basis @xmath441 using the symmetry properties and , we can write the @xmath165-reversed distribution in terms of the distribution at a point @xmath442 on the opposite side of the nucleus : @xmath443 thus a nucleon on the back side of the nucleus has an opposite transverse momentum to a corresponding nucleon in the front of the nucleus .
therefore , the ( anti)symmetrized distributions have definite parity under @xmath444 : @xmath445 = + w_\sigma^{symm}(p,\overline{b } ) \\ w_\sigma^{oam}(p , b ) & \equiv \frac{1}{2 } \left [ w_\sigma(p , b ) - ( { \underline{p } } \rightarrow - { \underline{p } } ) \right ] = - w_\sigma^{oam}(p,\overline{b } ) . \end{aligned}\end{aligned}\ ] ] eq . tells us that a consequence of @xmath209 invariance in the nucleus is that the orbital angular momentum encountered at any point in the front of the nucleus is compensated by an equal and opposite orbital angular momentum from a corresponding point on the back of the nucleus .
this is the resolution to the apparent paradox : when we neglect all wilson line contributions ( both @xmath253 $ ] and @xmath175 ) , the net asymmetry in the quark distribution is zero since @xmath446 . hence neglecting all wilson line contributions yields zero sivers function , consistent with @xcite .
an essential role is played in , then , by the rescattering factor @xmath253 $ ] . in the oam channel , even though the rescattering @xmath447 $ ] is not the source of a preferred transverse direction , without it the net contribution to the sivers function from oam would vanish after integration over @xmath134 , as can be gleaned from the left panel in .
the rescattering factor @xmath253 $ ] is essential because it introduces shadowing that breaks this front - back symmetry by screening quarks ejected from the front of the nucleus more than those ejected the back .
the sivers function relevant for sidis is therefore more sensitive to oam from the back of the nucleus than from the front , which prevents the complete cancellation of the oam contribution as in .
this analysis is strikingly similar to the arguments that historically established the existence of the sivers function . as collins argued in @xcite , @xmath209-invariance of any hadronic eigenstate prohibits a preferred direction that can generate the sivers function .
this is directly reflected in the vanishing of without the effects of multiple rescattering . and
as brodsky , hwang , and schmidt demonstrated in @xcite , the rescattering represented by the semi - infinite wilson lines breaks this symmetry and permits a preferred direction for the asymmetry . unlike that calculation , however , here the rescattering does not occur as color - correlated `` lensing '' due to rescattering on the remnants of the active quark . here
the interaction is explicitly color - decorrelated because the rescattering occurs on many nucleons whose colors are not correlated . despite this difference ,
the rescattering effects are still sufficient to break the front - back symmetry and give rise to a net preferred direction for the asymmetry .
s. j. brodsky , d. s. hwang , and i. schmidt , _ final state interactions and single spin asymmetries in semiinclusive deep inelastic scattering _ , _
phys.lett . _
* b530 * ( 2002 ) 99107 , [ http://xxx.lanl.gov/abs/hep-ph/0201296[hep-ph/0201296 ] ] .
y. koike and s. yoshida , _ probing the three - gluon correlation functions by the single spin asymmetry in @xmath448 _ , _ phys.rev .
_ * d84 * ( 2011 ) 014026 , [ http://xxx.lanl.gov/abs/1104.3943 [ arxiv:1104.3943 ] ] .
y. kanazawa and y. koike , _ chiral odd contribution to single transverse spin asymmetry in hadronic pion production _ , _ phys.lett . _ * b478 * ( 2000 ) 121126 , [ http://xxx.lanl.gov/abs/hep-ph/0001021 [ hep - ph/0001021 ] ] .
y. kanazawa and y. koike , _ estimate of a chiral odd contribution to single transverse spin asymmetry in hadronic pion production _ , _ phys.lett . _
* b490 * ( 2000 ) 99105 , [ http://xxx.lanl.gov/abs/hep-ph/0007272[hep-ph/0007272 ] ] .
j. c. collins , _ fragmentation of transversely polarized quarks probed in transverse momentum distributions _ , _
_ * b396 * ( 1993 ) 161182 , [ http://xxx.lanl.gov/abs/hep-ph/9208213 [ hep - ph/9208213 ] ] .
s. j. brodsky , d. s. hwang , y. v. kovchegov , i. schmidt , and m. d. sievert , _ single - spin asymmetries in semi - inclusive deep inelastic scattering and drell - yan processes _ , _ phys.rev .
_ * d88 * ( 2013 ) 014032 , [ http://xxx.lanl.gov/abs/1304.5237[arxiv:1304.5237 ] ] .
s. j. brodsky , d. s. hwang , and i. schmidt , _ initial state interactions and single spin asymmetries in drell - yan processes _ , _ nucl.phys .
_ * b642 * ( 2002 ) 344356 , [ http://xxx.lanl.gov/abs/hep-ph/0206259 [ hep - ph/0206259 ] ] .
l. gamberg and m. schlegel , _ final state interactions and the transverse structure of the pion using non - perturbative eikonal methods _ , _ phys.lett . _ * b685 * ( 2010 ) 95103 , [ http://xxx.lanl.gov/abs/0911.1964[arxiv:0911.1964 ] ] . l. d. mclerran and r. venugopalan , _ gluon distribution functions for very large nuclei at small transverse momentum _ , _ phys . rev . _
* d49 * ( 1994 ) 33523355 , [ http://xxx.lanl.gov/abs/hep-ph/9311205 [ hep - ph/9311205 ] ] . l. d. mclerran and r. venugopalan , _ computing quark and gluon distribution functions for very large nuclei _ , _ phys .
rev . _ * d49 * ( 1994 ) 22332241 , [ http://xxx.lanl.gov/abs/hep-ph/9309289 [ hep - ph/9309289 ] ] .
j. jalilian - marian and y. v. kovchegov , _ saturation physics and deuteron gold collisions at rhic _ , _ prog .
* 56 * ( 2006 ) 104231 , [ http://xxx.lanl.gov/abs/hep-ph/0505052 [ hep - ph/0505052 ] ] .
y. v. kovchegov , _ quantum structure of the non - abelian weizscker - williams field for a very large nucleus _ , _ phys .
* d55 * ( 1997 ) 54455455 , [ http://xxx.lanl.gov/abs/hep-ph/9701229 [ hep - ph/9701229 ] ] .
d. boer , a. dumitru , and a. hayashigaki , _ single transverse - spin asymmetries in forward pion production at high energy : incorporating small - x effects in the target _ , _ phys.rev . _ * d74 * ( 2006 ) 074018 , [ http://xxx.lanl.gov/abs/hep-ph/0609083[hep-ph/0609083 ] ] .
d. boer , a. utermann , and e. wessels , _ the saturation scale and its x - dependence from lambda polarization studies _ , _ phys.lett .
_ * b671 * ( 2009 ) 9198 , [ http://xxx.lanl.gov/abs/0811.0998 [ arxiv:0811.0998 ] ] .
f. dominguez , j .- w .
qiu , b .- w .
xiao , and f. yuan , _ on the linearly polarized gluon distributions in the color dipole model _ , _ phys.rev .
_ * d85 * ( 2012 ) 045003 , [ http://xxx.lanl.gov/abs/1109.6293 [ arxiv:1109.6293 ] ] .
k. golec - biernat and m. wsthoff , _ saturation effects in deep inelastic scattering at low @xmath122 and its implications on diffraction _ , _ phys .
* d59 * ( 1998 ) 014017 , [ http://xxx.lanl.gov/abs/hep-ph/9807513[hep-ph/9807513 ] ] .
p. tribedy and r. venugopalan , _ saturation models of hera dis data and inclusive hadron distributions in p+p collisions at the lhc _ , _ nucl.phys .
_ * a850 * ( 2011 ) 136156 , [ http://xxx.lanl.gov/abs/1011.1895[arxiv:1011.1895 ] ] .
d. boer , l. gamberg , b. musch , and a. prokudin , _ bessel - weighted asymmetries in semi inclusive deep inelastic scattering _ , _ jhep _ * 1110 * ( 2011 ) 021 , [ http://xxx.lanl.gov/abs/1107.5294 [ arxiv:1107.5294 ] ] .
y. v. kovchegov , l. szymanowski , and s. wallon , _ perturbative odderon in the dipole model _ , _
phys.lett . _
* b586 * ( 2004 ) 267281 , [ http://xxx.lanl.gov/abs/hep-ph/0309281[hep-ph/0309281 ] ] .
dedicated to the memory of jan kwiecinski .
a. bacchetta , m. boglione , a. henneman , and p. mulders , _ bounds on transverse momentum dependent distribution and fragmentation functions _ , _ phys.rev.lett .
_ * 85 * ( 2000 ) 712715 , [ http://xxx.lanl.gov/abs/hep-ph/9912490[hep-ph/9912490 ] ] .
k. itakura , y. v. kovchegov , l. mclerran , and d. teaney , _ baryon stopping and valence quark distribution at small x _ , _ nucl .
phys . _ * a730 * ( 2004 ) 160190 , [ http://xxx.lanl.gov/abs/hep-ph/0305332 [ hep - ph/0305332 ] ] .
y. v. kovchegov and a. h. mueller , _ gluon production in current nucleus and nucleon nucleus collisions in a quasi - classical approximation _ , _ nucl .
phys . _ * b529 * ( 1998 ) 451479 , [ http://xxx.lanl.gov/abs/hep-ph/9802440[hep-ph/9802440 ] ] .
b. kopeliovich , j. raufeisen , a. tarasov , and m. johnson , _ nuclear effects in the drell - yan process at very high - energies _ , _ phys.rev . _ * c67 * ( 2003 ) 014903 , [ http://xxx.lanl.gov/abs/hep-ph/0110221 [ hep - ph/0110221 ] ] .
j. jalilian - marian , a. kovner , and h. weigert , _ the wilson renormalization group for low x physics : gluon evolution at finite parton density _ , _ phys .
* d59 * ( 1998 ) 014015 , [ http://xxx.lanl.gov/abs/hep-ph/9709432[hep-ph/9709432 ] ] .
e. gardi , j. kuokkanen , k. rummukainen , and h. weigert , _ running coupling and power corrections in nonlinear evolution at the high - energy limit _ , _ nucl .
_ * a784 * ( 2007 ) 282340 , [ http://xxx.lanl.gov/abs/hep-ph/0609087[hep-ph/0609087 ] ] .
j. l. albacete , n. armesto , j. g. milhano , p. quiroga - arias , and c. a. salgado , _ aamqs : a non - linear qcd analysis of new hera data at small - x including heavy quarks _ , _ eur .
phys . j. _ * c71 * ( 2011 ) 1705 , [ http://xxx.lanl.gov/abs/1012.4408[arxiv:1012.4408 ] ] .
a. accardi , j. albacete , m. anselmino , n. armesto , e. aschenauer , _ et .
_ , _ electron ion collider : the next qcd frontier - understanding the glue that binds us all _ , http://xxx.lanl.gov/abs/1212.1701 [ arxiv:1212.1701 ] .
j. jalilian - marian , a. kovner , l. d. mclerran , and h. weigert , _ the intrinsic glue distribution at very small x _ , _ phys .
rev . _ * d55 * ( 1997 ) 54145428 , [ http://xxx.lanl.gov/abs/hep-ph/9606337 [ hep - ph/9606337 ] ] .
|
we calculate the sivers function in semi - inclusive deep inelastic scattering ( sidis ) and in the drell - yan process ( dy ) by employing the quasi - classical glauber mueller/ mclerran venugopalan approximation . modeling the hadron as a large `` nucleus '' with non - zero orbital angular momentum ( oam )
, we find that its sivers function receives two dominant contributions : one contribution is due to the oam , while another one is due to the local sivers function density in the nucleus .
while the latter mechanism , being due to the `` lensing '' interactions , dominates at large transverse momentum of the produced hadron in sidis or of the di - lepton pair in dy , the former ( oam ) mechanism is leading in saturation power counting and dominates when the above transverse momenta become of the order of the saturation scale .
we show that the oam channel allows for a particularly simple and intuitive interpretation of the celebrated sign flip between the sivers functions in sidis and dy .
+ pacs numbers : 12.38.bx , 13.88.+e , 12.38.cy , 24.85.+p
|
a hyperuniform many - particle system is one in which the structure factor approaches zero in the infinite - wavelength limit.@xcite in such systems , density fluctuations ( measured by the variance of number of particles inside a spherical window ) are anomalously suppressed at very large lengths scales , a `` hidden '' order that imposes strong global structural constraints.@xcite all structurally perfect crystals and quasicrystals are hyperuniform,@xcite but typical disordered many - particle systems , including gases , liquids , and glasses , are not .
disordered hyperuniform many - particle systems are exotic states of amorphous matter that have attracted considerable recent attention.@xcite materials that are simultaneously disordered and hyperuniform can be regarded to be exotic states of matter that lie between a crystal and a liquid ; they behave more like crystals in the manner in which they suppress large - scale density fluctuations , and yet they also resemble typical statistically isotropic liquids and glasses with no bragg peaks.@xcite an important class of disordered hyperuniform many - particle systems is comprised of the classical ground states of `` stealthy potentials,''@xcite which are bounded , long - range , pairwise additive potentials designed in fourier space .
these classical ground states are of particular fundamental interest because they can be degenerate and noncrystalline .
a nonnegative parameter inversely proportional to the number density , @xmath0 , controls the degree of order of such ground states . for @xmath4 ,
the ground states are overwhelmingly highly degenerate and disordered .
as @xmath0 increases above 0.5 , long - range translational and rotational order begins to emerge and eventually the system crystallizes .
we have previously studied these disordered ground states , and computed their pair correlation functions,@xcite structure factors,@xcite voronoi cell volume distribution,@xcite and particle - exclusion probabilities.@xcite some initial studies have demonstrated that stealthy hyperuniform systems are endowed with novel thermodynamic and physical properties .
for example , their low - temperature excited states are characterized by negative thermal expansion behavior.@xcite it has also been shown that dielectric networks derived from stealthy disordered hyperuniform point configurations possess complete photonic band gaps comparable in size to those of a photonic crystal , while at the same time maintain statistical isotropy , enabling waveguide geometries not possible with photonic crystals as well as high - density disordered transparent materials.@xcite however , the determination of physical / chemical properties of stealthy disordered hyperuniform materials is generally an unexplored area of research . in this paper , we investigate steady - state and time - dependent diffusion properties of certain decorations of stealthy disordered hyperuniform ground - state point configurations in two and three dimensions .
in particular , we derive two - phase heterogeneous media from point configurations by decorating the point configurations with spheres ( circles ) ; specifically , all points are circumscribed by spheres of radius @xmath2 that generally may overlap with one another . by varying the radius ,
the fraction of space occupied by the spheres will vary .
we study the effective transport properties of these disordered two - phase systems , including the effective diffusion coefficient,@xcite and static and time - dependent characteristics of diffusion - controlled reactions at the interfaces between the two continuous phases , as well as the trapping rate ( or its inverse , the mean survival time ) as well as the principal ( largest ) relaxation time.@xcite quantifying the effective diffusion coefficient is of importance not only because it has direct applications ( e.g. , diffusion of fuel and oxygen in a fuel cell @xcite , diffusion tensor magnetic resonance imaging , @xcite regulation and metabolism of normal organs , @xcite and drug release from porous matrices @xcite ) , but also because its determination translates immediately into equivalent results for the effective thermal and electric conductivity , the effective dielectric constant , and the effective magnetic permeability for reasons of mathematical analogy,@xcite and is therefore related to a host of applications .
diffusion - controlled reactions arise in widely different processes , such as heterogeneous catalysis,@xcite gas sensor operation,@xcite cell metabolism,@xcite crystal growth,@xcite and nuclear magnetic resonance ( nmr).@xcite these transport properties are related to several statistical geometrical and topological characteristics , which we therefore also study .
these include the pore - size functions ( the distribution of the distance from a randomly chosen location in the void phase to the closest phase boundary),@xcite the quantizer error ( a moment of the pore - size function , which is related to the principal relaxation time),@xcite the order metric @xmath5 ( a measure of the translational order of point configurations),@xcite and the percolation threshold or the critical radius ( the radius of the spheres at which a specific phase becomes connected ) of each phase.@xcite we compare the aforementioned physical and geometrical properties of our two - phase system derived from decorated stealthy ground states , as a function of the tuning parameter @xmath0 , with those of two other two - phase media : ( 1 ) equilibrium disordered ( fluid ) hard - sphere systems and ( 2 ) decorated poisson point processes ( ideal - gas configurations ) .
the former has short - range order that is tunable by its volume fraction but no long - range order .
the latter has neither short - range order nor long - range order . through comparison
, we find that some of these quantities are dramatically affected by the degree of long - range order , while other quantities are much more sensitive to the degree of short - range order . because many of these quantities depend on the density , we re - scale all systems to unit number density to ensure a fair comparison . among our major findings ,
we show that these transport , geometrical and topological properties of our two - phase media are generally distinctly different from those of equilibrium hard - sphere systems and spatially uncorrelated overlapping spheres . at high @xmath0 values
, the stealthy disordered two - phase media can attain nearly maximal effective diffusion coefficient , while also maintaining isotropy .
this novel property could have practical implications , e.g. , optimal and isotropic drug release from designed nanoparticles .
stealthy ground states tend to ideal gases configurationally in the @xmath6 limit .
@xcite nevertheless , we find that even in the low-@xmath0 regime , our two - phase media have much lower principal relaxation time than that of equilibrium hard - sphere systems of similar degrees of order , indicating that the formation of large holes in the stealthy systems is strongly suppressed .
lastly , we also find that the aforementioned geometrical and topological quantities are strongly correlated with each other .
the rest of the paper is organized as follows : in sec .
[ definition ] , we give precise definitions of the stealthy potential and the aforementioned transport , geometrical and topological quantities . in sec .
[ simulation ] , we present our numerical method to calculate them . we present our results in sec .
[ result ] and conclusions in sec .
[ conclusion ] .
this paper studies properties of point - particle systems as well as two - phase heterogeneous media derived from certain decorations of these point configurations .
a point - particle system consists of @xmath7 point particles with a certain probability density function @xmath8 , where @xmath9 , @xmath10 , ... , @xmath11 is the particle positions , in a simulation box of volume @xmath12 under periodic boundary conditions in @xmath13-dimensional euclidean space @xmath14 , where @xmath13 is 2 or 3 .
the number density is defined as @xmath15 .
the `` poisson point process ''
( also called `` ideal gas '' ) is produced by the probability density function @xmath16 that does not depend on particle positions @xmath17 .
the equilibrium hard - sphere point process of radius @xmath2 is another point process with @xmath8 equal to a positive constant if the distance between every pair of points is larger than @xmath18 and zero otherwise .
a realization of a two - phase medium can be mathematically described as a partition of a domain of space @xmath19 with volume @xmath20 into two separate regions , @xmath21 and @xmath22 .
it is characterized by an indicator function , @xmath23 , where @xmath24 is any position in the two - phase medium .
the indicator function @xmath23 is one if @xmath25 and zero if @xmath26 .
the volume fraction of phase 1 is given by @xmath27 , where @xmath28 denotes an ensemble average .
that of the other phase is given by @xmath29 .
let @xmath30 be the interface between @xmath21 and @xmath22 , the specific surface , i.e. , the total area of @xmath30 divided by @xmath20 , is given by : @xmath31 the two - phase media that we consider here are derived from point configurations by decorating the point configurations with spheres ( circles ) ; specifically , each point is circumscribed by a sphere of radius @xmath2 that generally may overlap with one another .
therefore , it is composed of a void region ( phase 1 ) and a particle region ( phase 2 ) .
when such a mapping is applied to a poisson point process , the decorated system is also called `` fully penetrable spheres '' @xcite or `` spatially uncorrelated spheres . ''
@xcite consider point processes that are obtained from the canonical ensemble probability distribution function defined by @xmath32/z , \label{canonical}\ ] ] where @xmath33 is an interaction potential , @xmath34 is the inverse temperature , and @xmath35 d \mathbf r^n$ ] is the partition function .
of particular interest in this paper is the `` stealthy '' interaction potential : @xmath36 where the sum is over all reciprocal lattice vector @xmath37 s of the simulation box such that @xmath38 , @xmath39 , @xmath40/2v_f \label{temp1}\ ] ] is a constant independent of the particle positions @xmath17 , and the second equal sign in eq .
( [ pot_f ] ) can be proved by parseval s theorem .
such potential is interesting not only because it is a pairwise additive potential [ as the right side of eq .
( [ pot_f ] ) shows ] , but also because it allows one to directly tune the structure factor @xmath41 the ground state ( i.e. , @xmath42 or zero - temperature state ) of this potential is obtained by constraining @xmath43 for all @xmath38.@xcite let @xmath44 be half the number of @xmath37 points in the summation of eq . ; , @xmath44 is the number of independent constraints . ]
the parameter @xmath45 determines the degree to which the ground states are constrained and therefore the degeneracy and disorder of the ground states.@xcite for @xmath4 , the ground states are typically disordered and uncountably infinitely degenerate.@xcite therefore , there are multiple ways to assign different weights ( i.e. , probabilities ) to different sets of ground states .
one way of particular interest is the zero - temperature ( @xmath42 ) limit of eq .
( [ canonical ] ) .
ground states drawn from such distribution are called `` entropically favored ground states''.@xcite it is interesting to note that in the @xmath6 and @xmath46 limit , both entropically favored ground states of stealthy potentials and equilibrium hard - sphere point processes tend to poisson point process geometrically . in the rest of the paper
this fact will be frequently used to test our simulation results since many properties of the poisson point process have been studied previously .
this paper studies the following steady - state and time - dependent diffusion properties in phase 1 ( the void phase ) of decorated entropically favored ground states of stealthy potentials , and compare them with that of decorated poisson point process and equilibrium disordered ( fluid ) hard - sphere system at unit number density .
consider the steady - state diffusion problem of some species with concentration field @xmath47 in a two - phase medium in which phase 1 is the space in which diffusion occurs and phase 2 are `` obstacles '' that the diffusing species can not enter . in phase 1 ,
the flux of the species , @xmath48 , is predicted by fick s first law : @xmath49 where @xmath50 is a diffusion coefficient which we set to unity for simplicity .
however , eq . ( [ eq_diffusion ] ) is valid only in phase 1 and has to be paired with the following neumann boundary condition : @xmath51 where @xmath52 is the normal vector of the surface .
we see that the inclusion of such obstacles adds a complicated boundary condition and makes the overall diffusion problem difficult . nevertheless , on a length scale
much larger than the characteristic length of the obstacles , the system can be homogenized @xcite and characterized by an `` effective '' diffusion coefficient , @xmath53 , defined by the average fick s first law : @xmath54 where angular brackets denote ensemble averages .
the effective diffusion coefficient of an isotropic two - phase medium must satisfy the hashin - shtrikman ( hs ) upper bound.@xcite for our case where phase 1 has unit diffusion coefficient and phase 2 can not be entered , this bound in @xmath13 dimensions is given by : @xmath55 this bound is optimal because it is realizable by certain model microstructures , including the `` coated - sphere model '' described in ref . , and is therefore the best possible bound for isotropic systems given volume - fraction information only .
consider the problem of diffusion and reaction among absorbing `` traps '' in the random medium
. let phase 1 be the region in which diffusion occurs and phase 2 be the trap region , the diffusion process in phase 1 is governed by the same fick s first law but with time dependency : @xmath56 this equation , combined with the conservation of the diffusing species inside phase 1 , @xmath57 , yields fick s second law : @xmath58 if phase 2 are absorbing `` traps '' ( rather than impenetrable obstacles as in the aforementioned effective diffusion problem ) , the boundary condition has to be changed . in the diffusion - controlled limit ,
i.e. , when the reaction rate at the interface is infinite , we have the following boundary condition:@xcite @xmath59 if we also set the initial concentration to be uniform outside of traps : @xmath60 then we have the survival problem . the `` survival probability , '' @xmath61 is equal to the fraction of reactant not yet absorbed at time @xmath62:@xcite @xmath63 the mean survival time of the reactant is the zeroth moment of @xmath61 : @xcite . here
, we use @xmath64 to avoid confusion with the order metric @xmath5 . ] @xmath65 the survival probability can be decomposed as a sum of exponential functions : @xmath66 where @xmath67 are coefficients and @xmath68 are relaxation times .
the largest relaxation time is called `` principal relaxation time '' and by convention denoted @xmath69 .
these quantities can be measured directly by nmr experiments since in nmr experiment of fluid - saturated porous media , proton magnetization decays mainly on the phase boundary.@xcite it is worth noting that although the above problems involve differential equations , @xmath53 , @xmath61 , and @xmath64 can actually be calculated much more efficiently by simulating brownian motions using the so - called `` first - passage time '' technique .
( see the sec .
[ simulation ] for details . ) the effective diffusion coefficient can be found from the ratio of the mean square displacement of such brownian particles and the time spent .
the survival probability @xmath61 is equal to the probability that a brownian particle have never reached any trap at time @xmath62 .
the mean survival time , @xmath64 , can be calculated by integrating @xmath61 but can also be calculated , more easily , by finding the average time needed for a particle to reach a trap the first time .
it is also worth noting that while the effective diffusion coefficient is identically zero as long as the void phase is not percolating , @xmath64 and @xmath69 are both positive until the spheres cover the entire space.@xcite this paper also studies the following geometrical and topological properties that are intimately related to the aforementioned diffusion characteristics .
as we have explained earlier , a hyperuniform many - particle system is one in which the structure factor , eq .
( [ structurefactor ] ) , approaches zero in the @xmath70 limit .
the name `` hyperuniform '' refers to an anomalous suppression of density fluctuations : consider random placements of a spherical observation window of radius @xmath71 in a @xmath13-dimensional many - particle system .
the number of points contained in such window , @xmath72 , is a random variable .
for a uniform but not hyperuniform many - particle system ( e.g. , ideal gas without a gravity field ) , @xmath73 for large @xmath71 scales as @xmath74 . for a hyperuniform system , @xmath73 for large
@xmath71 grows more slowly than @xmath74 .
it has been proved that the above - mentioned two conditions of hyperuniformity , @xmath75 and @xmath73 for large @xmath71 grows more slowly than @xmath74 , are mathematically equivalent.@xcite a similar definition exists for two - phase media.@xcite one can compute the volume fraction of either phase inside a spherical observation window of radius @xmath71 and find its variance . for large @xmath71
, this variance scales as @xmath76 for typical ( non - hyperuniform ) random two - phase media and decreases faster than @xmath76 for hyperuniform two - phase media .
an equivalent condition for hyperuniformity is that @xmath77 , where @xmath78 is called the `` spectral density '' and @xmath79 is the fourier transform of @xmath80.@xcite stealthy hyperuniform many - particle systems or two - phase media are subsets of hyperuniform many - particle systems or two - phase media in which @xmath81 or @xmath82 is zero for a range of @xmath37 vectors around the origin , i.e. , @xmath83 where @xmath84 is some positive number . for the many - particle systems mentioned in this paper , the ground state of `` stealthy '' potentials are stealthy and hyperuniform while equilibrium hard - sphere systems and poisson point process are neither stealthy nor hyperuniform .
when we decorate a point - particle configuration by replacing points with spheres of radius @xmath2 , the whole collection of spheres is considered a `` sphere packing '' if each pair of point particles is separated by a distance of at least @xmath18 ( i.e. , if the spheres do not overlap ) .
the fraction of space occupied by the union of spheres , @xmath85 , is called the packing fraction @xmath86 . of particular interest in this paper
is the maximum packing radius @xmath87 , which is half the minimum separation distance between two particles , and maximum packing fraction @xmath88 , which is the volume fraction of phase 2 when @xmath89 .
why should we study the maximum packing fraction ?
one important reason is that when we decorate a point configuration and map it into a two - phase medium , if spheres do not overlap , then the spectral density @xmath90 of the two - phase medium is proportional to the structure factor @xmath81 of the underlying point configuration:@xcite @xmath91 where @xmath92 is the volume of a @xmath13-dimensional sphere of radius @xmath2 and @xmath93 is the bessel function of order @xmath94 .
therefore , a decorated stealthy point configuration is a stealthy two - phase medium if @xmath95 . when @xmath96 , however , eq . ( [ chi_s ] ) no longer holds and we will see in sec .
[ packing_stealthy ] that decorated systems are generally no longer stealthy or hyperuniform . given a point - particle system ,
the void - exclusion probability @xmath97 is the probability that a spherical cavity of radius @xmath98 , centered at a random location , is empty of particles .
a related quantity is @xmath99/(\partial r)$ ] , the probability density function of the distance to the nearest particle from a randomly chosen location .
a different interpretation of @xmath100 is that if each point particle is replaced with a sphere of radius @xmath2 , then @xmath101 is the volume fraction of the space outside of the spheres , i.e. , @xmath102 since @xmath103 is the negative derivative of @xmath100 , @xmath104 is the specific surface @xmath105.@xcite another quantity related to @xmath97 is the scaled dimensionless quantizer error @xmath106 . for a point configuration with positions @xmath107 ,
a quantizer is a device that takes as an input a position @xmath24 in @xmath108 and outputs the nearest point @xmath109 of the configuration to @xmath24 .
assuming @xmath24 is uniformly distributed , one can define a mean square error , which can be obtained from @xmath97 via the relation:@xcite @xmath110 finally , two more related quantities can be defined for two - phase media .
the pore - size cumulative distribution function , @xmath111 , is defined as the fraction of pore space ( i.e. , space covered by phase 1 ) which has a pore radius larger than @xmath112 .
the function @xmath111 of our decorated system is trivially related to @xmath97 of the underlying point - particle system : @xmath113 moreover , the associated pore - size probability density function is given by @xmath114/(\partial \delta)$ ] .
this pore - size function at the origin is related to the specific surface , @xmath105 , by @xmath115 it is interesting to note that the moments of @xmath111 are related to the mean survival time and principle relaxation time via the following rigorous lower bounds @xcite : @xmath116 and @xmath117 we see that @xmath106 is proportional to the first moment of @xmath111 in the @xmath46 limit and is therefore related to the principal relaxation time .
we will be studying the above properties for systems of varying degrees of order .
therefore , it is desirable to have a way to quantify such order .
moreover , since the underlying point - configurations we study include both stealthy ground states , which have long - range order , and equilibrium liquid hard - sphere systems , which have short - range order , we desire an order metric that reflects short - range order and long - range order equally well .
a suitable choice is the order metric @xmath5 , introduced in ref . and
defined as : @xmath118 ^ 2 d\mathbf r = \frac{1}{(2\pi)^dd^d } \int_0^\infty [ s(k)-1]^2 d\mathbf k , \label{tau}\ ] ] where @xmath50 is some characteristic length scale , @xmath119 is the pair correlation function,@xcite @xmath120 is the angular average of @xmath81 , and the second equal sign can be proved by parseval s theorem . in this paper , we simply let @xmath121 because we always rescale the configuration to make the number density unity .
since the effective diffusion coefficient is trivially zero when the void phase is topologically disconnected , it is important to quantify when the phases are connected . to do this
, we will be considering the percolation properties of the systems .
as we specified earlier , we map point configurations into two - phase media by replacing each point with a sphere of radius @xmath2 . for phase 2 , the critical or percolation radius , @xmath122 , is the minimum @xmath2 such that a connected part of phase 2 becomes infinite in size .
the percolation volume fraction , @xmath123 , is the fraction of space occupied by the union of spheres of radius @xmath122 .
we can define similar percolation characteristics of the void phase.@xcite the percolation radius of the void phase , @xmath124 , is defined as the maximum @xmath2 such that there is still an infinite - sized connected part of phase 1 .
the percolation volume fraction , @xmath125 , is the volume fraction of phase 1 at radius @xmath124 . in two dimensions ,
it is very rare to have both phases percolating simultaneously ( see ref .
for such a rare example ) . in our case , @xmath126 and @xmath127 . in three dimensions
, however , both phases can simultaneously percolate , i.e. , the two - phase system is bicontinuous .
indeed , this is the case for our 3d systems and hence we must compute @xmath124 and @xmath122 separately .
we generate entropically favored ground states of stealthy potentials using the same protocol as our previous work.@xcite this protocol involves performing molecular dynamics ( md ) simulations at a very low temperature ( @xmath128 in 2d and @xmath129 in 3d in dimensionless units ) , taking snapshots periodically , and performing a local energy minimization starting from each snapshot . because the md temperature is sufficiently low ,
the snapshots before energy minimization are already very close to ground states .
therefore , the ground states produced by the subsequent energy minimization closely follow the canonical distribution in the zero - temperature limit .
we generate 20,000 configurations per @xmath0 value , same as ref . .
the only two differences between this work and our previous work @xcite are ( 1 ) system sizes are different ( see appendix [ number ] for our choice of system sizes and the justification ) , and ( 2 ) each configuration is rescaled to unit number density ( in order to ensure a fair comparison ) . we also generate equilibrium disordered hard - sphere systems via standard monte - carlo techniques in order to compare their statistics with entropically favored stealthy ground states statistics .
depending on the packing fraction @xmath130 , an equilibrium hard - sphere system can be disordered ( liquid - like ) or crystalline .
disordered equilibrium hard - sphere system exists for @xmath131 in 2d and @xmath132 in 3d.@xcite therefore , the packing fraction we used include @xmath133 , 0.1 , 0.15 , 0.2 , 0.25 , 0.3 , 0.35 , 0.4 , 0.45 , 0.5 , 0.55 , 0.6 , 0.65 , and 0.68 in 2d and @xmath133 , 0.1 , 0.15 , 0.2 , 0.25 , 0.3 , 0.35 , 0.4 , 0.45 , and 0.48 in 3d . for each @xmath130 in each dimension , we generate equilibrium hard - sphere systems with @xmath134 , 300 , and 500 particles . in each case , the system was first equilibrated with @xmath135 trial moves .
after that , we sample a configuration every @xmath136 trial moves until we obtain 20,000 configurations .
similar to the stealthy ground states , we keep the number density @xmath137 .
therefore , we adjust sphere radius to attain a desired packing fraction . because the method we used to calculate the effective diffusion coefficient is an extension of the method to calculate survival probability and mean survival time , we will explain the latter method first .
the survival probability @xmath61 and mean survival time @xmath64 can be calculated by simulating particles undergoing brownian motions .
the brownian motion can be simulated very efficiently using the first - passage - time technique.@xcite the key idea of this technique is that for a brownian particle at a particular location , let @xmath138 be the distance between it and the closest phase boundary . construct a sphere centered at the particle with radius @xmath138 ( which is called a first - passage - time sphere ) .
let @xmath139 be the time needed for the particle to reach the surface of such sphere for the first time , the distribution of @xmath139 can be calculated analytically . in 3d , the cumulative distribution function ( cdf ) of @xmath139 is@xcite @xmath140 in 2d , ref .
did not provide the distribution of @xmath141 . here
we find the following explicit 2d expression for @xmath141 : @xmath142 where @xmath143 is the bessel function of order @xmath144 , and @xmath145 is the @xmath144th root of @xmath146 .
the mean of @xmath139 , in any dimension , is simply @xmath147 .
therefore , the brownian motion inside the first - passage - time sphere does not need to be simulated in detail .
one simply moves the particle to a random location on the surface of such sphere , and increase the time by a certain amount , as detailed below . when calculating the mean survival time @xmath64 ,
the time increment can simply be @xmath147 , the mean of @xmath139 .
when calculating @xmath61 , however , the time increment has to be a random number drawn from the distributions given in eq .
( [ cdf_3d ] ) or ( [ cdf_2d ] ) .
the process of finding @xmath138 , moving the particle , and increasing the time is repeated until the brownian particle gets very close ( @xmath148 ) to a trap , at which time the brownian particle is deemed trapped . in our implementation , eqs .
( [ cdf_3d])-([cdf_2d ] ) are pre - computed and tabulated to accelerate the simulation . for each configuration , we simulate 10 brownian trajectories to calculate @xmath64 and 1000 trajectories to calculate @xmath61 . when calculating @xmath61 , each trajectory is additionally sampled 100 times , with different random time increments drawn from distributions ( [ cdf_3d])-([cdf_2d ] ) . after calculating @xmath61 ,
we calculate the principal relaxation time @xmath69 by fitting @xmath61 in the range @xmath149 to the asymptotic equation @xmath150 \approx c + t / t_1,\ ] ] where @xmath151 and @xmath69 are fitting parameters .
the effective diffusion coefficient @xmath152 can also be calculated using first - passage - time techniques .
@xcite in this case , however , the brownian particle can not be deemed trapped when it is sufficiently close to the phase boundary because phase 2 is now non - absorbing obstacles rather than absorbing traps . instead , we construct a first - passage - time sphere of radius @xmath153 , find a random place on the surface of the first - passage - time sphere that is outside of the obstacle phase , and move the brownian particle to that random place .
although this first - passage - time sphere contains two phases , the mean time taken for the brownian particle to reach such surface could still be computed analytically and was given in ref .
: @xmath154 where @xmath155 is the volume of the obstacle phase divided by the volume of the conducting phase inside the first - passage - time sphere and can be found analytically .
the process of constructing a first - passage - time sphere and moving the point particle is repeated to form a brownian trajectory . in the infinite - time limit
, the effective diffusion coefficient is given by:@xcite @xmath156 where @xmath157 is the mean - squared displacement of a brownian particle at time @xmath62 .
in practice , in a finite - time simulation , one should only consider the time regime in which the mean square displacement is strongly linear in time , since for sufficiently early times the mean square displacement is either ballistic or grows faster than linear in time.@xcite we find @xmath152 by fitting @xmath157 versus @xmath62 and extracting the slope of the line after some sufficiently large dimensionless time .
we define the unit of time to be @xmath158 and set both @xmath159 and @xmath50 to be unity . the point in time
in which @xmath157 first becomes a strongly linear function occurs when the brownian particle sufficiently samples the two - phase system such that it can be viewed effectively as brownian motion in a homogeneous medium .
for the microstructures that we considered here , we find that the linear regime occurs in the dimensionless time interval @xmath160 , i.e. , we determine @xmath152 from the linear relationship @xmath161 to get a sense of the possible behaviors of the mean square displacements as a function of time , we show examples in fig .
[ r2vt ] at several values of @xmath2 for a three - dimensional system at @xmath162 and indicate the linear fit in each case .
this fitting procedure works especially well near percolation , which is the most difficult regime to simulate . for this particular system ,
the void phase stops percolating at obstacle radius @xmath163 .
figure [ r2vt ] shows that for @xmath164 , @xmath157 is linear with @xmath62 . for @xmath165 , only a fraction of configurations
still have a percolating void phase , and our fitting procedure was able to distinguish the initial uprise in @xmath157 ( contributions mainly from brownian particles moving inside a `` cage '' , i.e. , a disconnected part of the void phase ) from the steady increase in @xmath157 ( contributions from brownian particles that are in a percolating part of the void phase ) . for @xmath166 ,
all brownian particles are caged , and the fit has a virtually zero slope ( and therefore produces a virtually zero @xmath152 ) . ,
versus time , @xmath62 , for a three - dimensional system at @xmath162 with obstacle radii @xmath167 ( top ) , 0.8 ( middle ) , and 0.9 ( bottom ) . for this particular system ,
the percolation threshold of the void phase is @xmath163 , and hence @xmath152 must vanish for larger values of @xmath2.,title="fig:",scaledwidth=30.0% ] , versus time , @xmath62 , for a three - dimensional system at @xmath162 with obstacle radii @xmath167 ( top ) , 0.8 ( middle ) , and 0.9 ( bottom ) . for this particular system ,
the percolation threshold of the void phase is @xmath163 , and hence @xmath152 must vanish for larger values of @xmath2.,title="fig:",scaledwidth=30.0% ] , versus time , @xmath62 , for a three - dimensional system at @xmath162 with obstacle radii @xmath167 ( top ) , 0.8 ( middle ) , and 0.9 ( bottom ) . for this particular system ,
the percolation threshold of the void phase is @xmath163 , and hence @xmath152 must vanish for larger values of @xmath2.,title="fig:",scaledwidth=30.0% ] we simulate 1 brownian trajectory per configuration to calculate @xmath152 . in fig .
[ de_withperco ] , we compare the computed @xmath152 with the distribution of the void - phase percolation threshold and find that @xmath152 becomes zero right after all configurations stop percolating .
the fact that our measured @xmath152 diminishes to zero at the percolation threshold indicates that our choice of the fitting range in eq .
( [ eq_de_fit ] ) is appropriate . , with the probability density function ( pdf ) of the void - phase percolation threshold , @xmath168 , for three dimensions , @xmath169 .
the calculated @xmath152 becomes zero when the void phase stops percolating.,scaledwidth=50.0% ] generally speaking , the precise calculation of the percolation threshold of disordered systems require very large system sizes .
for example , to accurately determine the percolation threshold of 3d fully penetrable spheres , ref .
employed systems of up to @xmath170 particles .
the whole system is divided into smaller cubes and the content particles in each cube is generated only when such cube is being probed .
unfortunately , our protocol of low - temperature md and a subsequent energy minimization does not allow us to save time by only generating required parts of the configuration .
moreover , in order to accurately follow the canonical - ensemble distribution at zero - temperature limit , the md temperature has to be so low such that many ( @xmath171 ) time steps are required to produce a sufficiently long trajectory .
the requirement of a very large number of time steps forces us to further sacrifice system size . as a result ,
our system is limited to several hundred particles .
therefore , accurate determination of the percolation threshold is extremely challenging .
thus we experimented with two advanced algorithms to minimize finite - size effect in order to obtain relatively accurate results .
we will first explain how to use these two methods to determine the percolation threshold for the particle phase , and then describe the generalizations to the void phase .
one of them , which we call `` @xmath172 maximum method , '' is described in ref . .
starting from a random particle in a configuration , one randomly chooses two of its periodic images in two different directions . the quantity @xmath172 ( denoted as @xmath173 in ref . )
is defined as the probability that this particle is connected to one of the chosen periodic images but not the other . at the percolation threshold
, @xmath172 attains its maximum .
therefore , one can numerically find @xmath172 as a function of sphere radius @xmath2 and find its maximum in order to find the percolation threshold . in our implementation
, we calculate @xmath174 for various @xmath2 s starting from @xmath175 , with increment @xmath176 , until @xmath174 develops a peak and then returns to zero .
we then select all data points such that @xmath177 , where @xmath178 is the maximum of @xmath174 , and perform a quadratic fit of the selected data points .
the maximum of the fitted function gives the percolation radius @xmath122 .
measures the percolation radii @xmath122 using several different system sizes and then extrapolates to the infinite - system - size limit .
however , when we perform the same fitting procedure using different system sizes , we did not find a clear trend : in each dimension , for some @xmath0 values larger systems produces larger @xmath122 while for other @xmath0 values larger systems produces smaller @xmath122 .
moreover , the extrapolated @xmath122 , as a function of @xmath0 , is not as smooth as the un - extrapolated one .
we therefore conclude that random noise is probably more important than finite - size effect in this case and extrapolation is not proper .
thus , we will simply use @xmath122 of our largest system as an estimate of the infinite - system - size @xmath122 . after finding the percolation radius @xmath122
, we determine the percolation volume fraction @xmath123 .
one could have simply read this quantity from a plot of the quantity @xmath97 , since @xmath179 .
however , we decide to use a somewhat more accurate method : we divide the whole simulation box into 12000@xmath18012000 pixels ( in 2d ) or 1200@xmath1801200@xmath1801200 voxels ( in 3d ) and find out if the center of each pixel or voxel is inside any sphere of radius @xmath181 .
we then count the number of pixels or voxels that are centered inside spheres to find out the volume fraction . from our experience
, this procedure gives us a four - significant - figures precision in @xmath182 .
the other method we employed , which we call `` @xmath183 intersection method , '' is introduced in ref . .
at a given radius @xmath2 , define @xmath184 to be the size of the largest cluster , @xmath183 ( denoted as @xmath185 in ref . )
is defined as : @xmath186 where @xmath187 denotes an ensemble average . as ref .
shows , at the percolation threshold , @xmath183 is the same for different system sizes .
therefore , one can compute @xmath183 as a function of @xmath2 , and find the intersection of @xmath188 for different system sizes to find @xmath122 . following ref . , we use three different @xmath7 s for each @xmath0 value , and perform an extrapolation to find @xmath122 in the infinite-@xmath7 limit .
after that , we use the same procedure discussed in the previous paragraph to calculate @xmath123 from @xmath122 .
we have used both methods to calculate the percolation volume fraction @xmath123 of decorated stealthy ground states at various @xmath0 s in 2d and 3d .
they are presented in fig .
[ perco ] .
figure [ perco ] also presents @xmath123 for decorated poisson point processes obtained from refs . and , which are @xmath189 in 2d and @xmath190 in 3d .
these results can be used as benchmarks since poisson point processes are geometrically equivalent to entropically favored stealthy ground states at @xmath191 .
we see that in 2d , while both methods give results that approaches the poisson value very well in the @xmath192 limit , the @xmath172 maximum method produces much smoother results . in 3d , however , although both methods produce relatively smooth results , only results from the @xmath183 intersection method approaches the poisson value very well in the @xmath192 limit .
therefore , we decide to choose the @xmath172 maximum method in 2d and the @xmath183 intersection method in 3d for the rest of the paper .
it is interesting to note that to our knowledge , the @xmath172 maximum method has been demonstrated to work well in 2d @xcite but not in 3d , while the @xmath183 intersection method has been demonstrated to work well in 3d @xcite but not in 2d .
it is possible that these two methods are just more suited to their respective dimensions .
of entropically favored stealthy ground states at different @xmath0 s in 2d ( top ) and 3d ( bottom).,title="fig:",scaledwidth=45.0% ] of entropically favored stealthy ground states at different @xmath0 s in 2d ( top ) and 3d ( bottom).,title="fig:",scaledwidth=45.0% ] of entropically favored stealthy ground states at different @xmath0 s in 3d.,scaledwidth=45.0% ] besides the percolation threshold of the spheres , we also study the percolation threshold of the void phase . in two dimensions ,
the percolation radius of the void phase , @xmath124 , is equal to the percolation radius of the spheres , @xmath122 . in three dimensions , however , @xmath124 has to be calculated separately .
we compute @xmath124 in three dimensions by performing a voronoi tessellation of each configuration , and then computing @xmath183 of the voronoi vertices . as in the particle - phase case ,
the intersection of @xmath188 at different system sizes gives @xmath124 .
similar to the particle - phase case , @xmath124 can then be converted to @xmath125 by digitization , the result of which is presented in fig .
[ perco1 ] .
similar to the particle - phase case , we compare @xmath125 for our systems with that for the decorated poisson point processes obtained from ref .
, @xmath193 . combining the @xmath123 and @xmath125 results
, we see that as @xmath0 increases from 0 to 0.46 , the @xmath194 range for bicontinuity moves upwards , from @xmath195 to @xmath196 , respectively .
the quantities @xmath97 and @xmath106 are calculated by first computing @xmath197 . for each configuration of @xmath7 point particles , @xmath198 random locations in 2d or @xmath199 random locations in 3d
are generated in the simulation box .
for each location , the distance from it to its nearest particle is found .
these distances are then binned to yield @xmath197 .
we then integrate @xmath197 using trapezoidal rule to find @xmath97 .
the quantizer error @xmath106 is obtained by another integration of @xmath200 , using trapezoidal rule , as eq .
( [ quantizerdefinition ] ) shows .
the numerically obtained @xmath197 always have compact support , and thus the above - mentioned integrations does not need to be truncated .
the order metric @xmath5 can be computed from either @xmath119 or @xmath120 , as eq .
( [ tau ] ) shows .
we have tried both approaches .
the real - space integration in eq .
( [ tau ] ) is truncated at half the simulation box side length and the reciprocal space integration in eq .
( [ tau ] ) is truncated at @xmath201 , where @xmath84 is the cutoff of the stealthy potential ( as detailed in sec .
[ subsection_stealthypotential ] ) .
we present visualizations of our two - phase systems derived from decorated stealthy ground states in figs .
[ configurations_2 ] - [ configurations_32 ] . in three dimensions , we present separate figures for the particle phase and the void phase for clarity . in the rest of the section ,
we present the above - mentioned properties of our two - phase systems , and compare them with decorated poisson point process and hard - sphere point process .
( top ) and @xmath202 ( bottom ) , at @xmath203 .
the void phase is marked green .
, title="fig:",scaledwidth=48.0% ] ( top ) and @xmath202 ( bottom ) , at @xmath203 .
the void phase is marked green .
, title="fig:",scaledwidth=48.0% ] ( top ) and @xmath204 ( bottom ) , at @xmath203 .
each sphere is randomly assigned to one of four colors in order to improve visual clarity.,title="fig:",scaledwidth=48.0% ] ( top ) and @xmath204 ( bottom ) , at @xmath203 .
each sphere is randomly assigned to one of four colors in order to improve visual clarity.,title="fig:",scaledwidth=48.0% ] ( top ) and @xmath204 ( bottom ) , at the void - phase percolation threshold @xmath205 ( top ) or 0.6992 ( bottom).,title="fig:",scaledwidth=48.0% ] ( top ) and @xmath204 ( bottom ) , at the void - phase percolation threshold @xmath205 ( top ) or 0.6992 ( bottom).,title="fig:",scaledwidth=48.0% ] we present the maximum packing fraction of decorated stealthy ground states , @xmath206 , in fig .
[ packingfraction ] . in each dimension , as @xmath0 increases , @xmath206 remains to be zero for @xmath0 up to about 0.3 and then start to increase .
this indicates that for @xmath207 , particles in entropically favored stealthy ground states can become arbitrarily close to each other .
as @xmath0 becomes higher , particles develop an effective hard core that are impenetrable .
the development of such hard core was also observed in ref . .
, averaged over all configurations , of decorated stealthy ground states in two and three dimensions as a function of @xmath0.,scaledwidth=48.0% ] when we decorate a stealthy ground state and map it into a two - phase medium , if @xmath208 , then eq . ( [ chi_s ] ) ensures that the resulting two - phase medium is also stealthy .
however , if @xmath209 , will the resulting two - phase medium also be stealthy or hyperuniform ? to answer this question , we decorated a two - dimensional stealthy ground state of @xmath210 particles at @xmath211 with several different sphere radii @xmath2 , digitized the resulting two - phase medium into @xmath212 pixels , and calculated the spectral density @xmath213 using eq .
( [ chi_v ] ) .
the result is presented in fig .
[ stealthy_spectraldensity ] . for this particular system ,
the maximum packing radius is @xmath214 .
we see that for @xmath215 , @xmath213 is zero for @xmath216 . for @xmath217 , however , @xmath213 is positive and does not tend to zero as @xmath218 .
therefore , a decorated stealthy ground state is generally neither stealthy nor hyperuniform if @xmath209 . of a two - phase medium obtained from decorating a two - dimensional stealthy ground state with @xmath210 particles at @xmath211 with several different sphere
radii @xmath2 . , scaledwidth=48.0% ] , for our two - phase systems derived from decorated stealthy ground states in two ( top ) and three ( bottom ) dimensions as a function of sphere radius @xmath2 .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=48.0% ] , for our two - phase systems derived from decorated stealthy ground states in two ( top ) and three ( bottom ) dimensions as a function of sphere radius @xmath2 .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=48.0% ] we present the calculated effective diffusion coefficient for our two - phase systems derived from decorated stealthy ground states in fig . [ diffusioncoefficient ] .
it is interesting to note that in two dimensions , the curves of @xmath219 cross over each other for different values of @xmath0 : while for smaller @xmath2 higher @xmath0 produces a higher @xmath152 , for larger @xmath2 higher @xmath0 produces a smaller @xmath152 .
an explanation for such phenomenon will be presented in the next paragraph .
it is also useful to plot @xmath152 versus the particle - phase volume fraction , @xmath85 , by mapping @xmath2 to @xmath85 using eq .
( [ ev_phi ] ) .
we present such plots in fig .
[ diffusioncoefficient2 ] .
these plots show that higher @xmath0 values ( more ordered arrangements of the obstacle phase ) always produce higher @xmath152 at the same volume fraction , which is consistent with our intuition : a more ordered arrangement of the obstacles leaves more space between them , and produces a higher @xmath152 .
so why did we see the opposite relationship between @xmath0 and @xmath152 in fig .
[ diffusioncoefficient ] , except for smaller @xmath2 in 2d ?
it turns out that a lower @xmath0 induces more overlap between the spherical obstacles and thus results in a lower @xmath85 .
this in turn produces a higher @xmath152 . with @xmath152
plotted versus @xmath85 , it is interesting to compare our result with the hs upper bound given in eq .
( [ hsbound ] ) .
we make such comparison in fig .
[ diffusioncoefficient2 ] .
our result is consistent with the upper bound for any @xmath0 and @xmath85 except for small fluctuations , but the bound is sharp only for smaller @xmath85 . if it is desired to find structures that maximizes @xmath152 , then any one of the degenerate structures that achieves the hs bound is optimal .
we see that our two - phase systems derived from decorated stealthy ground states at very high @xmath0 s are very close to being optimal for @xmath85 up to 0.4 - 0.5 . in two dimensions ,
this @xmath85 range coincides with @xmath206 at high @xmath0 s .
since decorated stealthy ground states loses stealthiness as @xmath85 increases beyond @xmath206 , our results suggest that a loss of stealthiness causes @xmath152 to stop being optimal . in three dimensions , however , @xmath152 is less sensitive to structures . thus , although decorated stealthy ground states ( at high @xmath0 s ) stops being a packing at around @xmath220 , @xmath152 does not deviate from the optimal value until about @xmath221 . our observation that @xmath152 is less sensitive to structures in 3d than in 2d is consistent with the trends indicated in ref . , which found that in the infinite-@xmath13 limit , @xmath152 is given exactly by the arithmetic average of the diffusion coefficients of the two phases ( weighted by their volume fraction ) , independent of the structure .
in fig . [ diffusioncoefficient2 ]
we also present @xmath152 of decorated lattice structures ( i.e. , periodic arrays of spherical inclusions ) . since these lattice structures are stealthy with even higher @xmath0 values , unsurprisingly , their @xmath152 sticks with the hs upper bound for an even larger @xmath85 range .
lastly , we would like to mention a difference between the support of @xmath152 as a function of @xmath85 in 2d versus 3d . while @xmath152 for for our two - phase systems in 2d diminishes to zero at @xmath222 , in 3d @xmath152 does not vanish until @xmath223 .
this difference emerges from the difference in the topological ( connectedness ) characteristics of the void phase between these dimensions .
, for our two - phase systems derived from decorated stealthy ground states in two ( top ) and three ( bottom ) dimensions as a function of particle - phase volume fraction @xmath85 .
the number density @xmath159 is fixed to be unity .
the optimal hashin - shtrikman ( hs ) upper bound , @xmath152 for triangular lattice and face - centered cubic ( fcc ) lattice , and @xmath152 for equilibrium hard disks and spheres are also plotted.,title="fig:",scaledwidth=48.0% ] , for our two - phase systems derived from decorated stealthy ground states in two ( top ) and three ( bottom ) dimensions as a function of particle - phase volume fraction @xmath85 .
the number density @xmath159 is fixed to be unity .
the optimal hashin - shtrikman ( hs ) upper bound , @xmath152 for triangular lattice and face - centered cubic ( fcc ) lattice , and @xmath152 for equilibrium hard disks and spheres are also plotted.,title="fig:",scaledwidth=48.0% ] we have computed the mean survival time , @xmath64 , of a diffusing reactant with unit diffusion coefficient , in our two - phase systems derived from decorated stealthy ground states .
these results are summarized in fig .
[ meansurvivaltime ] . for comparison ,
the same quantity for equilibrium disordered hard - sphere systems are also included . clearly , increasing order ( increasing @xmath0 for stealthy ground states or increasing @xmath85 for hard spheres ) suppresses @xmath64 .
however , there is a crossover between the curves for stealthy ground states and that for equilibrium disordered hard spheres .
this crossover is expected because as @xmath85 increases , an equilibrium hard - sphere system becomes more ordered , and therefore comparable to a stealthy two - phase medium with a higher @xmath0 . in fig .
[ meansurvivaltime2 ] , we plot @xmath64 versus @xmath0 for @xmath220 and 0.5 .
we see that in 2d , @xmath64 is somewhat more sensitive to @xmath0 than in 3d .
, as a function of particle - phase volume fraction @xmath85 , for our two - phase systems derived from decorated stealthy ground states in 2d ( top ) and 3d ( bottom ) . the same quantity for equilibrium disordered hard - sphere system is also included for comparison .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=40.0% ] , as a function of particle - phase volume fraction @xmath85 , for our two - phase systems derived from decorated stealthy ground states in 2d ( top ) and 3d ( bottom ) . the same quantity for equilibrium disordered hard - sphere system
is also included for comparison .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=40.0% ] for our two - phase systems derived from decorated stealthy ground states in two and three dimensions at phase 2 volume fraction @xmath220 and 0.5 .
the number density @xmath159 is fixed to be unity.,scaledwidth=40.0% ] in fig .
[ survivalprobability ] we present the survival probability , @xmath61 , at @xmath224 , for our two - phase systems derived from decorated stealthy ground states and equilibrium disordered hard spheres .
the same crossover phenomenon also appears here , suggesting that the long - range order possessed by stealthy ground states suppresses @xmath61 at large @xmath62 more efficiently , while the short - range order possessed by equilibrium disordered hard spheres suppresses @xmath61 at small @xmath62 more efficiently . for our two - phase systems derived from decorated stealthy ground states in 2d ( top ) and
3d ( bottom ) at phase 2 volume fraction @xmath220 . the same quantity for equilibrium disordered hard - sphere system
is also included for comparison .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=48.0% ] for our two - phase systems derived from decorated stealthy ground states in 2d ( top ) and 3d ( bottom ) at phase 2 volume fraction @xmath220 . the same quantity for equilibrium disordered hard - sphere system
is also included for comparison .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=48.0% ] in fig .
[ relaxationtime ] we present the principal relaxation time @xmath69 for our two - phase systems derived from decorated stealthy ground states .
it turns out that @xmath69 is much more sensitive to @xmath0 in 2d than in 3d .
more interestingly , one can compare @xmath69 of stealthy ground states and equilibrium disordered hard disks at the same order metric @xmath5 .
we present such comparison in fig .
[ relaxationtime_tau ] . in two dimensions
, one can see that at @xmath220 , @xmath69 of equilibrium disordered hard disks is much higher than that of our two - phase systems derived from decorated stealthy ground states with similar @xmath5 s .
as we explained earlier , @xmath69 is related to the pore - size distribution .
therefore , our results suggest that hyperuniformity suppresses the formation of large holes , even in the very disordered regime
. as @xmath85 increases to 0.5 , however , the difference between the two systems diminishes .
our finite - sized simulation results suggest that at this value of @xmath85 , even equilibrium hard - sphere systems suppress the formation of large holes very well . of equilibrium disordered hard - sphere systems
is actually infinite because of a non - zero probability of forming arbitrarily large holes ( i.e. , @xmath225 is non - zero for arbitrarily large @xmath112 ) . @xcite for finite - sized systems , however , @xmath69 is finite .
for decorated stealthy ground states , it is unclear whether or not @xmath69 would be infinite in the infinite - system - size limit . ] for our two - phase systems derived from decorated stealthy ground states in two and three dimensions at phase 2 volume fraction @xmath220 and 0.5 .
the number density @xmath159 is fixed to be unity.,scaledwidth=48.0% ] for our two - phase systems derived from decorated stealthy ground states and equilibrium disordered hard spheres in 2d ( top ) and 3d ( bottom ) at volume fraction @xmath220 and 0.5 .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=48.0% ] for our two - phase systems derived from decorated stealthy ground states and equilibrium disordered hard spheres in 2d ( top ) and 3d ( bottom ) at volume fraction @xmath220 and 0.5 .
the number density @xmath159 is fixed to be unity.,title="fig:",scaledwidth=48.0% ] the percolation volume fraction for both phases in 2d and 3d was already presented in fig .
[ perco ] and [ perco1 ] .
the void - exclusion probability @xmath97 , quantizer error @xmath106 , and order metric @xmath5 are presented in figs .
[ ev]-[ordermetric ] .
we see that in each dimension , as @xmath0 increases , @xmath182 increases , @xmath97 at any @xmath98 decreases , @xmath106 decreases , and @xmath5 increases .
the order metric @xmath5 can be computed from either @xmath119 or @xmath120 , as shown in fig .
[ ordermetric ] . in 2d ,
the results from these two approaches have good consistency .
however , in 3d , @xmath5 computed from @xmath119 is often slightly lower than @xmath5 computed from @xmath120 .
we discovered that this is because @xmath119 is still oscillating around 1 at half the simulation box side length , where the integration in eq .
( [ tau ] ) has to be cut off .
therefore , such a cutoff should make @xmath5 computed from @xmath119 too low .
we thus use @xmath5 computed from @xmath120 in the rest of the paper .
it is seen that @xmath5 is very sensitive at detecting the rise in short - range and long - range order as @xmath0 increases .
of entropically favored stealthy ground states in two ( top ) and three ( bottom ) dimensions at unit number density.,title="fig:",scaledwidth=50.0% ] of entropically favored stealthy ground states in two ( top ) and three ( bottom ) dimensions at unit number density.,title="fig:",scaledwidth=50.0% ] of entropically favored stealthy ground states in two and three dimensions at unit number density.,scaledwidth=50.0% ] of entropically favored stealthy ground states in two and three dimensions at unit number density , calculated from pair correlation function @xmath119 and structure factor @xmath120 .
we also include an analytical approximation for @xmath5 , given in ref .
18 , which is @xmath226.,scaledwidth=50.0% ] , order metric @xmath5 , and percolation volume fraction @xmath227 at unit number density.,title="fig:",scaledwidth=48.0% ] , order metric @xmath5 , and percolation volume fraction @xmath227 at unit number density.,title="fig:",scaledwidth=48.0% ] , order metric @xmath5 , and percolation volume fraction @xmath227 at unit number density.,title="fig:",scaledwidth=48.0% ] in fig .
[ correlations ] , we explore the correlation between @xmath106 , @xmath5 , and @xmath227 , and compare that for entropically favored stealthy ground states with that for equilibrium disordered hard - sphere systems .
it is interesting to note that these two systems behave very differently . at the same @xmath5 ,
stealthy systems give lower values of @xmath228 , indicating @xmath228 is more sensitive to long - range order than to short - range order .
at the same @xmath5 or @xmath228 , stealthy systems give higher values of @xmath123 , indicating that @xmath123 is even more sensitive to long - range order than to short - range order .
in this work , we decorated stealthy disordered hyperuniform point configurations of different degrees of order with spheres of various radii , and computed several transport and structural properties of these decorated systems .
the transport properties that we studied include effective diffusion coefficient @xmath152 , mean survival time @xmath64 , survival probability @xmath61 , and principal relaxation time @xmath69 . the structural properties examined include hyperuniformity and stealthiness , maximum packing fraction @xmath206 , the void - exclusion probability @xmath100 , the order metric @xmath5 , and the percolation thresholds @xmath125 and @xmath123 .
we showed that the order metric @xmath5 is an exquisite detector of both short- and long - range translational order . while all geometrical and topological quantities are strongly correlated ( positive correlation between @xmath123 and @xmath5 , and negative correlation between @xmath106 and the former two quantities ) , the relation between the physical quantities are more complex : while @xmath152 increases as @xmath0 increases or as @xmath85 decreases , @xmath64 and @xmath69 increases as @xmath0 decreases or as @xmath85 decreases .
therefore , there is no simple relationship between @xmath152 and @xmath64 or @xmath69 .
another reason why there is no such relationship is because if phase 1 ceases to percolate , then @xmath152 becomes zero but @xmath64 and @xmath69 are still positive @xcite . besides finding correlations between geometrical and topological properties , we find that in the highly disordered ( @xmath229 ) regime , @xmath69 of our two - phase systems derived from decorated stealthy ground states is much lower than that of equilibrium hard - sphere system .
together with the void - exclusion probability , low @xmath69 suggests that the formation of large holes is strongly suppressed , even though the configuration appears completely disordered . in the higher order ( @xmath230 ) regime
, @xmath152 of our disordered isotropic two - phase systems derived from decorated stealthy ground states is very close to the hashin - shtrikman upper bound for @xmath231 , where @xmath206 is the maximum packing fraction . since such decorated systems maintain stealthiness if and only if @xmath232 , our results suggest a connection between stealthiness and the ability to have a nearly optimal ( maximal ) @xmath152 .
the fact that stealthy disordered two - phase media have nearly optimal @xmath152 could have practical implications , e.g. , optimal and isotropic drug release from designed nanoparticles . although nearly optimal @xmath152 can also be achieved by lattice structures ( i.e. , periodic arrays of inclusions ) , the latter are always anisotropic .
thus , if one desires isotropic two - phase media with highest possible @xmath152 at a specific volume fraction , disordered stealthy two - phase media could be the best choice .
disordered stealthy ground states are uncountably infinitely degenerate.@xcite the maximum packing fraction @xmath206 varies among configurations . in the future , it would be interesting to design algorithms that sample stealthy ground states with a bias toward configurations with higher @xmath206 values .
with such an algorithm , one would be able to design isotropic two - phase media with nearly optimal @xmath152 with very high @xmath85 .
lastly , we would like to mention that although here we only study the diffusion problem of point brownian particles , the diffusion problem of finite - sized brownian particles has also been of interest.@xcite it is noteworthy that our results can be trivially extended to the latter case .
the diffusion of brownian particles of radius @xmath233 among obstacles of radius @xmath2 is equivalent to the diffusion of point brownian particles among obstacles of radius @xmath234 .
this mapping was previously exploited to quantify diffusion of finite - sized spheres in various models of porous media.@xcite interestingly , one can relate the transport properties computed here ( @xmath152 , @xmath64 , and @xmath69 ) to different physical properties of the same systems via cross - property relations , including those that relate them to the elastic moduli , @xcite as well as fluid permeability .
@xcite in future work , we will carry out such analyses .
the authors thank duyu chen and jaeuk kim for their careful reading of the manuscript .
as discussed in sec . [ generatingstealthy ] , it is nontrivial to choose the system size @xmath7 and parameter @xmath0 , especially because one of our protocol to calculate the percolation threshold requires three different @xmath7 s for each @xmath0 .
we enumerated all possible choices of @xmath7 s and @xmath0 s for @xmath235 and picked up some @xmath0 values that allow at least three different choices of @xmath7 s .
our choice of @xmath7 and @xmath0 in 2d and 3d are listed in tables [ n_2d ] and [ n_3d ] .
it is desirable to consider values of @xmath0 higher than 0.4133 ... in 3d , but our enumeration did not find such a @xmath0 value that satisfies the above condition .
therefore , we chose three more @xmath7 s that allow @xmath0 to be very close to @xmath236 but makes @xmath0 differ in the fifth decimal place .
see the caption for table [ n_3d ] for details .
except for the percolation threshold calculation , we only use the largest @xmath7 for each @xmath0 and @xmath13 .
71ifxundefined [ 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 * * , ( ) s .
atkinson , g. zhang , a. b. hopkins , and s. torquato , phys .
e * 94 * , 012902 ( 2016 ) .
@noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ ( , ) pp .
@noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) https://books.google.com/books?id=utfobwaaqbaj[__ ] , interdisciplinary applied mathematics ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) https://books.google.com/books?id=yvglmuank8ic[__ ] , no .
( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1016/0022 - 5096(63)90060 - 7 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) i. c. kim , and s. torquato , j. appl . phys . * 68 * , 3892 ( 1990 ) .
|
disordered hyperuniform many - particle systems have attracted considerable recent attention , since they behave like crystals in the manner in which they suppress large - scale density fluctuations , and yet also resemble statistically isotropic liquids and glasses with no bragg peaks .
one important class of such systems is the classical ground states of `` stealthy potentials . ''
the degree of order of such ground states depends on a tuning parameter @xmath0 .
previous studies have shown that these ground - state point configurations can be counterintuitively disordered , infinitely degenerate , and endowed with novel physical properties ( e.g. , negative thermal expansion behavior ) . in this paper , we focus on the disordered regime ( @xmath1 ) in which there is no long - range order , and control the degree of short - range order .
we map these stealthy disordered hyperuniform point configurations to two - phase media by circumscribing each point with a possibly overlapping sphere of a common radius @xmath2 : the `` particle '' and `` void '' phases are taken to be the space interior and exterior to the spheres , respectively .
the hyperuniformity of such two - phase media depend on the sphere sizes : while it was previously analytically proven that the resulting two - phase media maintain hyperuniformity if spheres do not overlap , here we show numerically that they lose hyperuniformity whenever the spheres overlap .
we study certain transport properties of these systems , including the effective diffusion coefficient of point particles diffusing in the void phase as well as static and time - dependent characteristics associated with diffusion - controlled reactions . besides these effective transport properties
, we also investigate several related structural properties , including pore - size functions , quantizer error , an order metric , and percolation thresholds .
we show that these transport , geometrical and topological properties of our two - phase media derived from decorated stealthy ground states are distinctly different from those of equilibrium hard - sphere systems and spatially uncorrelated overlapping spheres .
as the extent of short - range order increases , stealthy disordered two - phase media can attain nearly maximal effective diffusion coefficients over a broad range of volume fractions while also maintaining isotropy , and therefore may have practical applications in situations where ease of transport is desirable .
we also show that the percolation threshold and the order metric are positively correlated with each other , while both of them are negatively correlated with the quantizer error . in the highly disordered regime ( @xmath3 ) , stealthy point - particle configurations are weakly - perturbed ideal gases .
nevertheless , reactants of diffusion - controlled reactions decay much faster in our two - phase media than in equilibrium hard - sphere systems of similar degrees of order , and hence indicate that the formation of large holes is strongly suppressed in the former systems .
|
quantum technologies are often developed in the qubit paradigm , where the basic carrier of quantum information is a two - level quantum system .
qubits are a natural choice because binary is the language of classical technologies .
however , even here , despite the prevalence of binary , its supremacy is questionable .
indeed , donald knuth has advocated the use of balanced ternary , a 3-state classical logic @xcite . in the quantum domain ,
qudits offer a state space with a richer structure than their two - level counterparts , and the merits of this for quantum information have been explored in many contexts @xcite .
recent experiments have shown even large @xmath0 quantum systems can be precisely controlled @xcite . here
we present a qudit generalisation of a powerful class of quantum error correcting codes , the color codes @xcite .
along with surface codes @xcite , they constitute the most successful topological codes . using topology ,
quantum codes have achieved high error thresholds , whilst proving more practical than concatenated codes .
thresholds of qudit surface codes indicate improvements with qudit dimension @xcite .
qubit color codes have several advantages over qubit surface codes , and we show these features can be transferred over into the qudit setting . there has been some prior investigation into qudit color codes for prime - power @xmath0 @xcite and color codes based on more general groups @xcite , although these works were restricted to 2d topologies . in this paper
, we generalize color codes to any qudit dimension @xmath0 and spatial dimensions @xmath1 up to the point where @xmath2 factorial is a multiple of @xmath0 . specifically , given any lattice suitable for constructing qubit color codes
, we show how to use the same lattice to construct a qudit color code . for qubits ,
a non - clifford gate can be implemented in color codes in 3 and higher spatial dimensions transversally , i.e. by a tensor product of local unitary gates , an inherently fault - tolerant procedure @xcite . to avoid confusion between the spatial dimension of the lattice and the hilbert space dimension of the qudit
, we shall always denote the former by the letter @xmath1 and the latter by the letter @xmath0 .
recently , it has been shown by bravyi and knig @xcite that a quantum error correcting code in @xmath1 spatial dimensions can support a gate with constant depth from at most the @xmath1th level of the clifford hierarchy .
the fact that color codes can be shown to saturate this bound with transversal gates is a very promising feature , and when combined with gauge fixing techniques @xcite enables universal quantum computation without the need of magic state distillation @xcite .
the structure of the qudit clifford group is very different from its qubit counterpart @xcite .
nevertheless , we find that 3d color codes also provide transversal non - clifford gates in the qudit case .
recently the color codes were generalized to gauge color codes @xcite subsystem codes with many advantageous features including low weight error detection measurements , universal transversal gates via gauge fixing for @xmath3 , fault - tolerant conversion @xcite between codes of different spatial dimension @xcite , and for the @xmath4 case , single shot error correction @xcite a robustness to measurement errors without the need for repeated measurements .
we show that the qudit color codes introduced here can also be generalized to gauge color codes .
the main technique which we employ is a bipartition of the vertices in the graph that defines the code into _ starred _ and _ unstarred _ vertices .
we call this the _ star - bipartition _ , to distinguish it from the other important colorings which define the color codes . the commutation properties of the stabilizer and logical operators of the color codes ( and gauge color codes ) in the qubit setting can be reduced to the fact that the pairs of operators @xmath5 and @xmath6 commute .
the star - bipartition we introduce replaces some operators with their complex conjugate with respect to the computational basis of @xmath7 eigenstates .
for example , replacing the above operators with @xmath8 and @xmath9 , respectively .
crucially , this latter pair of operators does commute for qudits of any dimension , and this becomes the starting point for a generalisation of the color codes from qubits to qudits of any hilbert space dimension @xmath0 and any spatial dimension @xmath1 .
the sharp - eyed reader will note that @xmath10 .
we write the star explicitly on a real matrix here , and throughout , to emphasize symmetry and to simplify notation .
thus we see that as long as a pair of @xmath11- and @xmath7-type stabilizers have in common an equal number of starred qudits , they will commute .
furthermore , the star - bipartition provides a general framework for constructing transversal gates from higher levels of the clifford hierarchy .
while elements of this technique can be seen in earlier work @xcite , this is the first time that it has been exploited systematically .
the second key technical component of our work is a generalisation of the triorthogonal matrix technique by bravyi and haah @xcite .
this paper is structured as follows .
we start in sec .
[ sec : stabilizer ] by reviewing the stabilizer formalism for @xmath0-level systems . in sec .
[ sec : colour_codes ] we describe how to generalize a qubit color code in arbitrary spatial dimensions to a qudit color code by employing the notion of star - conjugation . in sec .
[ sec : m_matrices ] we review and generalize the triorthogonal matrix construction . in sec .
[ sec : transversal ] we derive conditions on the lattice that must hold for a transversal non - clifford gate to be implementable on the code .
[ sec : gauge_fixing ] explains how the hadamard can be implemented on the same lattice , although not in the color code , using the technique of gauge fixing . in sec .
[ sec : gauge_color_codes ] we show gauge color codes can naturally be defined for all the codes we have introduced , inheriting the favorable features of the qubit gauge color codes .
we conclude in sec .
[ sec : conclusion ] .
we will consider @xmath0-level quantum systems ( qudits ) as the building blocks for the constructions of quantum codes @xcite . unless stated otherwise , the qudit dimension @xmath0 is assumed to be any integer greater than two .
the conventional basis states for the @xmath0-level system are taken to be @xmath12 , for @xmath13 .
the single qubit pauli matrices @xmath11 and @xmath7 have natural extensions in higher dimensions @xcite .
the qudit analogues are @xmath14 where @xmath15 . with a slight abuse of terminology
, we shall say that two operators @xmath16 and @xmath17 `` @xmath18-commute '' if @xmath19 and note that @xmath18-commutation holds for @xmath11 and @xmath7 .
these generalized operators simplify to the familiar pauli operators for @xmath20 .
the single qudit pauli group is generated ( up to global phases ) by @xmath21 and @xmath7 and the @xmath22-qudit pauli group @xmath23 is the @xmath22-fold tensor product of the single qudit pauli group .
consider an abelian subgroup @xmath24 , such that @xmath25 for nonzero @xmath26 , then we say that @xmath27 is a stabilizer group , and we refer to its elements as the _
stabilizers_. the stabilizer group defines an error correcting code with codewords corresponding to states @xmath28 that are stabilized by the stabilizers , i.e. @xmath29 for all @xmath30 .
the logical operators correspond to the set of operators that commute with @xmath31 but are not contained in it . a pair of logical operators @xmath32 and @xmath33 @xmath18-commute with each other and hence encode one qudit .
gottesman and chuang @xcite introduced a classification of quantum gates known as the _ clifford hierarchy _ ( ch ) , which can be defined recursively @xcite as @xmath34 where @xmath35 is the pauli group .
for example , @xmath36 is the group of operators that leave the pauli group invariant under conjugation and is called the clifford group .
we shall describe some important clifford group gates @xmath37 , @xmath38 and @xmath39 below . in prime qudit dimensions , these are known to generate the whole clifford group @xcite .
the gate @xmath40 is the qudit version of the hadamard gate ( also known as the discrete fourier transform ) , @xmath41 the @xmath42 gate is the generalisation of the qubit @xmath43-phase gate , @xmath44 and @xmath45 is the controlled-@xmath46 gate ( also known as the sum gate ) , @xmath47 where @xmath48 and @xmath49 are the control and target qudits , respectively .
the set of gates @xmath50 in the hierarchy contains all gates from lower levels of the hierarchy . to refer to gates in level @xmath51 of the hierarchy but not level @xmath52
we shall say that the level @xmath51 of the hierarchy is the lowest level of the hierarchy for which the gate is a member . in prime dimensions
, it is known that it suffices to supplement the clifford group with just one non - clifford gate from the third level of the ch in order to obtain a universal set of gates @xcite .
such a gate is not unique , and in the qubit case , the @xmath53 gate @xmath54 is usually chosen for this purpose . for the qudit case , we choose the following particularly convenient definition for the @xmath53 gate , which is valid in all dimensions except when @xmath55 @xcite , @xmath56 it is a consequence of lemmas [ rlemma1 ] and [ rlemma2 ] ( below ) that for @xmath57 this gate is non - clifford and inhabits the third level of the ch . in @xmath58 and @xmath59 ,
the gate defined in equation is not non - clifford , as it reduces to the pauli @xmath7 gate since @xmath60 mod @xmath61 and mod @xmath62 .
however , the following definition provides a suitable alternative @xmath53 gate for these dimensions .
the gate is non - clifford and in the third level of the ch : @xmath63 where @xmath64 and where the function in the exponent @xmath65 is evaluated in regular arithmetic ( or equivalently modulo @xmath66 ) .
when we refer to `` the @xmath53 gate '' in this paper we will always mean a gate of the form of equations or , depending on the qudit dimension under consideration .
for notational convenience we suppress the dependence of @xmath53 on @xmath67 , since it will always be clear by the context which @xmath53 gate we require .
notice how the @xmath68 gate has a cubic power in the exponent of @xmath18 , in contrast to the quadratic power in the case of the @xmath38 gate . for prime dimensions ,
the first investigation characterising all the phase gates from the third level of the ch was performed by howard and vala @xcite . in general
, there is a close correspondence between the order of the polynomial in the exponent of @xmath18 and the lowest level of the ch the phase gate belongs to .
let us define the following family of phase gates in terms of a polynomial function @xmath69 of degree @xmath70 , such that @xmath71 with coefficients @xmath72 , so @xmath73 : @xmath74 one can then prove the following useful lemmas . [ rlemma1 ] for all @xmath0 , all @xmath75 and all functions
@xmath76 , the gate @xmath77 is in the @xmath70th level of the clifford hierarchy .
the proof of this is simple and concise .
we begin by calculating @xmath78 where @xmath79 is a new function @xmath80 .
expanding out , the degree @xmath70 terms cancel , so the leading term is @xmath81 and extra terms are all degree @xmath82 or smaller .
we now observe that if @xmath83 , @xmath77 is a pauli operator . using the definition of the clifford hierarchy in equation , the lemma follows by induction .
[ rlemma2 ] for all @xmath0 , all @xmath75 and all functions @xmath76 satisfying @xmath84 , the gate @xmath77 is not in the @xmath85th level of the clifford hierarchy .
if the gate is in the @xmath85th level of the hierarchy , then applying the inductive transformation in the previous proof @xmath86 times must return a pauli operator , and thus applying it @xmath70 times results in an operator proportional to the identity . by chaining together transformations of the form of equation @xmath87 times
, we obtain @xmath88 .
if the corresponding operator is proportional to the identity then this function must be constant , and thus @xmath89 .
these two lemmas provide us with a simple way to generate gates at all levels of the clifford hierarchy as required .
they fail when @xmath89 , which was the case above for @xmath90 and @xmath91 . in those exceptional cases
, gates can be discovered by moving to higher roots of unity as illustrated by equation .
color codes @xcite are a class of topological qubit stabilizer codes that be defined on a topological space of any spatial dimension @xcite @xmath92 .
the @xmath1-dimensional manifold is celluated into a lattice of objects called @xmath93-_cells _ for all spatial dimensions @xmath94 .
for example , vertices are @xmath95-cells , edges are @xmath96-cells and a @xmath97-cell is a cell defined on the faces , or plaquettes , of the lattice .
we define [ mucolex ] a lattice @xmath98 is called a @xmath1-colex ( colex is short for `` color complex '' ) whenever 1 .
it is a celluation of an orientable @xmath1-dimensional manifold without a boundary ; and 2 .
every vertex has @xmath99 neighbors ( @xmath99-valency ) ; and 3 .
the @xmath1-cells are @xmath99-colorable .
any @xmath1-colex defines a qubit color code .
for example , the smallest 3-dimensional color code is a 15-qubit code defined on the lattice illustrated in figure [ fig : tetra](a ) .
the stabilizer group is generated by face operators ( with a @xmath7 operators assigned to the vertices around the face or 2-cell ) and cell operators ( with an @xmath11 operators assigned to the vertices around each 3-cell ) , see figure [ fig : tetra](b ) .
our results show a @xmath1-colex also defines a qudit color code .
an alternative description of a @xmath1-colex is that its dual @xmath100 is a simplical lattice .
this is lattice where each @xmath93-cell is a simplex , an object with @xmath101 vertices . in the dual
, we have conditions : ( @xmath102 ) every @xmath1-cell has @xmath99 neighbors ; ( @xmath103 ) the vertices are @xmath99 colorable .
such codes can be constructed @xcite by starting with a closed hyperspherical lattice and then removing a vertex to `` puncture '' the surface . alternatively , as we show in sec .
[ sec : code_distance ] , one can construct them directly by defining a suitable boundary on a regular lattice structure .
such a code encodes a single qubit and has the attractive feature of a transversal non - clifford @xmath53 gate .
the qudit color codes have many useful properties . for more details ,
we refer the reader to @xcite .
the stabilizer generators are guaranteed to commute in this construction .
those of the same type ( @xmath11 or @xmath7 ) trivially commute , while those of different types commute since they always meet in pairs .
as already remarked , @xmath5 and @xmath104 commute only in @xmath20 .
however , by replacing one operator of each pair with its complex conjugate we produce a pair of operators that commute for all @xmath0 .
these are @xmath8 and @xmath105 .
to define color codes in all qudit dimensions , we need , therefore , to find a construction which allows us to take advantage of the commutation of @xmath8 and @xmath105 .
this can be achieved by identifying and exploiting a bipartition or bicoloring of the graph defining the lattice . to avoid confusion with other colorings of the lattice important for color codes we call this the star - bipartition .
we note that a similar construction was defined ( for prime power qudit dimension only ) in @xcite . here
we go further and show that the star - bipartition is the starting point for an identification of a broad family of transversal gates on these codes , including non - clifford gates saturating bravyi and knig s bound .
the star - bipartition is a bipartition of the color code lattice @xmath98 .
it thus divides the set of lattice vertices into starred and unstarred vertices where neighboring vertices always belong to different sets .
here we prove the following useful bipartition lemmas [ starlem ] let @xmath98 be a @xmath1-colex , then its vertices can be 2-colored into starred and unstarred sets , @xmath106 and @xmath107 , respectively .
recall that the @xmath1-colexes have several properties listed in def .
[ mucolex ] that are essential to proving the above lemma . using @xmath108 to denote the number of elements in a set , we also have [ starcelllem ] let @xmath98 be a @xmath1-colex , so that its vertices are partitioned into @xmath107 and @xmath106 according to lemma [ starlem ] .
it follows that 1 .
@xmath109 ; and 2 .
any @xmath93-cell @xmath110 with @xmath111 contains a equal number of vertices from each partition , so @xmath112 .
lemmas [ starlem ] and [ starcelllem ] , or variants thereof , were already proved in prior research @xcite .
these results are especially important to qudit color codes and play a fundamental role , and so , for completeness , we present our own proofs .
our presentation of lemma [ starlem ] has the merit of being more concise than that of ref .
@xcite . before presenting our proof
, we review the concept of orientation from topology theory @xcite .
an orientation on a @xmath113-simplex is an ordered list of its vertices , with orientations considered equivalent whenever they differ by a permutation generated from an even number of transpositions ( a transposition is swap of two elements ) .
all permutations can be generated by either an even or odd number of transpositions , and so there are two possible orientations for a @xmath113-simplex .
the @xmath113-simplex contains subsimplices that can be obtained by removing a single vertex .
this vertex removal provides an induced orientation on the subsimplex , such that removing @xmath114 from a simplex with orientation @xmath115 induces an orientation @xmath116 .
consider two oriented @xmath113-simplices with a common subsimplex ( that is , they are are neighbors ) .
we say their orientations are consistent if they induce opposite orientations to their common subsimplex .
an entire lattice of simplices is said to be orientable if there exists a choice of orientations such that all pairs of simplices are consistently orientated .
an example of a consistently orientated lattice is shown in fig .
[ fig : lemma3 ] .
orientablity of a lattice is a topological feature that depends on the underlying manifold and from this definition one can show that many familiar manifolds are orientable , including euclidean , spherical and toroidal manifolds .
we now employ these concepts in our proof . moving clockwise round the simplex . for neighboring simplices @xmath16 and @xmath17
, we also show the induced orientation on their edges . for an edge orientation @xmath117
, we illustrate this with an arrow going from @xmath118 to @xmath119 .
the simplices @xmath16 and @xmath17 induce opposite orientations to their common edge , and so are consistently orientated .
if a simplex has an orientation @xmath120 with colors @xmath121 then we label it starred , and otherwise unstarred .
the star - bipartition lemma further formalises this notion of starring in terms of color and orientation . ] to prove the star - bipartition lemma , we switch to the dual lattice @xmath100 . in the dual picture , we need to show that the @xmath1-dimensional cells can be 2-colored .
property 2 of def .
[ mucolex ] ensures that the dual lattice is simplical .
property 3 of def .
[ mucolex ] ensures the dual vertices are @xmath99 colored .
this coloring can be described by a map @xmath122 that assigns color label @xmath123 to vertex @xmath114 . given a simplex @xmath124 with an orientation , say @xmath125 , we can apply the color map to the orientation to obtain a colored - orientation @xmath126 .
this defines a bipartition ( the star - bipartition ) with two simplicies belonging to the same partition if their colored - orientation differs by an even permutation .
the remainder of the proof establishes that neighbouring simplicies belong to different partitions .
two neighbouring oriented simplices @xmath16 and @xmath17 always share a common @xmath127 subsimplex .
let us call the vertices they share in common @xmath128 .
simplex @xmath16 also contains a vertex @xmath129 and we label its orientation @xmath130 .
similarly , simplex @xmath17 contains additional vertex @xmath131 and has orientation @xmath132 where @xmath133 is some permutation of vertex labels .
no generality is lost by placing @xmath129 and @xmath131 last in their respective orientations as this can always be achieved with an even number of transpositions ( assuming @xmath134 ) . since the manifold is orientable , we can assume a consistent orientation on their common subsimplex .
we see that @xmath16 induces the orientation @xmath135 and @xmath17 induces the orientation @xmath136 , and so consistency demands that @xmath133 is an odd permutation of the vertex labels .
next , we show this lifts to odd permutation in the color - orientation of @xmath16 and @xmath17 .
simplex @xmath16 has color - orientation @xmath137 and simplex @xmath17 has color - orientation @xmath138 .
since the lattice is @xmath139 colored , we know @xmath129 and @xmath131 must possess the same color , which is whatever color is absent from their common subsimplex .
permutation of vertex labels results in a corresponding permutation of colors .
therefore , the color - orientations of @xmath16 and @xmath17 differ by an odd permutation of their first @xmath113 elements , which corresponds to an overall odd permutation . since all neighboring @xmath1-simplicies must have opposite color - orientations , they belong to separate partitions of our star bipartition , proving the lemma .
let us now turn to lemma [ starcelllem ] .
we shall begin with part _ 1 _ of the lemma .
consider a bipartite lattice that is @xmath70-valent .
we count the number of edges @xmath140 .
every edge is incident to one , and only one , starred vertex .
furthermore , each vertex is contained in @xmath70 edges , and so the total count is @xmath141 .
the same argument applies to unstarred vertex , so @xmath142 . equating @xmath143
, we see that @xmath144 entails @xmath145 .
since a @xmath1-colex is @xmath99-valent , we have proven part _ 1 _ of lemma [ starcelllem ] .
consider any @xmath93-cell @xmath110 .
it defines a sublattice @xmath146 with same vertex set as @xmath110 .
if @xmath146 is a @xmath93-valent with @xmath147 , then the above argument applies and @xmath110 contains equal number of starred and unstarred vertices .
the following reasoning parallels that in ref .
we show explicitly that @xmath146 is a @xmath93-colex for the @xmath148 case , but any @xmath93 is reached by iteratively applying the argument down to the desired @xmath93 .
again , we must switch to the dual lattice where @xmath149 is obtained by removing a single vertex @xmath150 from @xmath100 .
the sublattice @xmath149 retains all simplices containing @xmath150 , but with their dimension reduced by the removal of @xmath150 .
therefore @xmath149 is a simplical lattice of dimension @xmath151 and so @xmath98 is @xmath1-valent .
our proof only requires the correct valency , which we have shown , but note that @xmath146 also has the correct coloring for a @xmath151-colex .
specifically , if @xmath152 and @xmath110 intersect on some @xmath151-cell of @xmath146 , then we assign it the color of @xmath152 . the above results concern a lattice without a boundary
however , if one punctures the code removing a starred vertex then we have [ cor_punc ] let @xmath98 be puncturing of a @xmath1-colex , with the inherited bipartition into @xmath106 and @xmath107
. then 1 .
@xmath153 ; and 2 .
any @xmath93-cell @xmath110 with @xmath111 contains a equal number of vertices from each partition , so @xmath112 .
property _ 1 _ is immediately follows from point _ 1 _ of lemma [ starcelllem ] as a single starred vertex has been removed .
2 _ is identical to its partner in lemma [ starcelllem ] .
the punctured lattice only keeps cells that did not contain the punctured vertex , and so the property is directly inherited . the dual lattice @xmath100 has been an essential proof tool for these lemmas , but in the rest of the paper we shall only consider the primal lattice .
-type stabilizers colored red , green , blue and yellow . the 1-cells ( edges ) of the code
are also colored .
the vertices are also colored so the set @xmath107 is represented by black circles and the set @xmath106 is represented by white stars .
( b ) a single @xmath7 stabilizer of the tetrahedral 3d color code . the plaquette can take the color yellow if considered as a face of the green 3-cell , and vice versa .
( c ) a single @xmath11 stabilizer of the tetrahedral 3d color code . ]
before we write down the stabilizer generators for these codes , we shall use the star - bipartition to define notation for an important family of transversal operators which we call star - conjugate transversal . throughout this paper , when discussing color codes we identify qudits with vertices in the lattice defining the code .
let @xmath154 denote a set of vertices ( and thus the corresponding qudits ) .
let us introduce the notation @xmath155=u ^{\otimes|v|}\ ] ] to denote the tensor product of unitary @xmath156 acting on each qudit identified with the vertices in @xmath154 .
we call an operator _ star - conjugate transversal _ when it has the following form @xmath157 \otimes u^{*}[v_{\star } ]
. \label{eqn : u_trans } \ ] ] in other words , @xmath158 consists of @xmath156 applied to all unstarred vertices and @xmath159 applied to all starred vertices . to define qudit color codes we need to introduce two types of stabilizer generators which are defined with respect to cells of different dimensions within the lattice .
as in the qubit case , we associate @xmath7-type stabilizer generators with the @xmath160-cells of the lattice and @xmath11-type stabilizers with the @xmath161-cells @xcite . in this section , for simplicity of presentation , we will take the example of @xmath162 so the @xmath7-type stabilizers are associated with plaquettes ( @xmath97-cells ) and @xmath11-type stabilizers are associated with @xmath1-cells , although the construction holds in the more general case , which we shall consider later .
for example in 3d , the @xmath11 stabilizers act on the vertices contained in a 3-cell of the lattice , and the @xmath7 stabilizers act on the vertices contained in a 2-cell ( plaquette ) .
this can be seen in figure [ fig : tetra ] .
setting @xmath162 the stabilizer generators therefore take the form @xmath163 \otimes x^{*}[v_{\star } \cap c ] , \label{eqn : x_stab}\\ s_{z , p } & = & z[v_{\bullet } \cap p ] \otimes z^{*}[v_{\star } \cap p ] .
\label{eqn : z_stab}\end{aligned}\ ] ] for all @xmath97-cells @xmath164 , and @xmath1-cells @xmath110 of the lattice .
recall that we write the conjugate operator @xmath165 explicitly to emphasis the ubiquity of starring , even though @xmath11 is a real operator in the computational basis and @xmath166 .
it is not just the @xmath167 operators where starring is non - trivial , but many logical operators also require starring .
let us denote the group generated by @xmath168 operators @xmath169 and @xmath167 operators by @xmath170 .
all elements of @xmath171 commute as do all elements of @xmath172 .
it remains to show that all elements of @xmath171 commute with all elements of @xmath172 .
this follows from the fact that @xmath173 commutes with @xmath9 .
the above construction ensures that whenever cell @xmath110 and cell @xmath164 overlap they overlap on an equal number of starred and unstarred vertices , which is a point discussed further in sec .
[ moitysec ] .
hence these stabilizer generators commute as required .
note that the qubit @xmath20 case is included in our definition .
we noted above that , in the constructions we consider , the code will contain one more unstarred qudit than there are starred qudits .
we can thus define the logical encoded pauli operators for the code star - conjugate transversally .
@xmath174 \otimes x^{*}[v_{\star } ] , \\
\bar{z } = \tilde{z } = z[v_{\bullet } ] \otimes z^{*}[v_{\star } ] .
\label{eqn : log_ops}\end{aligned}\ ] ] the fact that @xmath153 ensures that the logical operators satisfy the same commutation properties as @xmath11 and @xmath7 .
one can verify that these operators commute with the stabilizer operators defined above by recalling that each 2-cell @xmath164 and and @xmath1-cell @xmath110 contains an equal number of starred and unstarred vertices . in topological stabilizer codes
we expect to be able to increase the number of physical systems encoding the quantum information in order to protect the information more effectively a property characterized by the code distance .
we have remarked already that suitable codes can be constructed by puncturing a hypersphere . in this section
we provide , however , a constructive alternative method to show that in @xmath1 spatial dimensions the color code distance can be made arbitrarily large .
the code distance is the weight of the smallest logical operator .
it was convenient above , in equations , to define the logical operator basis star - conjugate transversally .
however , multiplying by a subset of the stabilizer group we can localize @xmath175 to any edge of the polytope formed by the lattice , where such an object is a 1-dimensional sub - manifold of the lattice between two of the vertices of the polytope .
each such vertex is contained in only a single @xmath1-cell ( as opposed to those contained in two or more cells of the lattice ) , and hence there are exactly ( @xmath176 ) vertices of the polytope . in @xmath1 spatial dimensions ,
an edge of the polytope is comprised of 1-cells of two different colors .
similarly , 2-cells contain 1-cells of two colors .
beginning from the transversal definition of the operators in equations and multiplying @xmath175 by the subset of @xmath7 stabilizers defined on 2-cells containing 1-cells of the same pair of colors , the logical operator is localised to the edge of the polytope uniquely defined as containing 1-cells matching this pair of colors .
the constraints placed on the subset of stabilizers chosen , and the initial choice of transversal operator basis ensures that the minimum length of a string - like logical operator defined in this way is the length of the edge of the polytope colored as indicated .
the minimum length of a string - like @xmath175 in such a definition , and hence the code distance , is then length of the shortest edge of the polytope .
thus by showing that the length of an edge of the polytope can be made arbitrarily large we demonstrate the ability to create codes of any distance .
we begin with an example in 3d and then generalize the argument to higher spatial dimensions .
the particular choice we make for this example is a 4-colorable tiling of _ truncated octahedra _ , part of which is illustrated in figure [ fig : truncated_octa ] .
this is not a unique choice for an alternative see @xcite where the 3d color code is defined on a lattice constructed from cubes and truncated octahedra .
in addition to the two lattices already mentioned , we have identified two further possible choices of regular tilings in 3d euclidean space on which a color code may be defined .
one is the cantitruncated cubic honeycomb , comprising truncated cuboctahedra ( polyhedra with square , hexagonal and octahedral faces ) , truncated octahedra , and cubes .
the second is the omnitruncated cubic honeycomb comprising truncated cuboctahedra and octahedral prisms .
this list is not exhaustive , for instance many more tilings may exist in curved space .
the method proceeds by carefully cutting a block of the lattice out , in order to form a tetrahedron , similar to that illustrated in figure [ fig : tetra](a ) .
the main difference is that the new tetrahedron contains more qudits and can have an arbitrarily large ( though always odd , in order to ensure the @xmath18-commutation of the logical operators ) number of qudits contained in its edges .
the shape cut from the lattice may not look tetrahedral at first but may be deformed to a regular tetrahedron , with vertices belonging only to a single cell forming the four vertices of the tetrahedron . in @xmath1 spatial dimensions a suitable tiling of the space as outlined at the beginning of section [ sec : colour_codes ] must be found . besides these requirements placed on the lattice construction , the `` block '' ( polytope ) cut out of the lattice to form the higher - dimensional analogue of the tetrahedron must adhere to certain rules . 1
. the polytope formed must have @xmath99 boundaries , each of a different color .
the color of the boundary corresponds to the color of the string - like operators that can end on that boundary .
2 . every edge of the polytope must contain an odd number of qudits . as stated above , this is to ensure the @xmath18-commutation of the logical operators is preserved .
3 . there should be exactly @xmath1 vertices that belong only to a single cell of the lattice .
these cells should be different colors and these vertices will form the vertices of the polytope .
errors are detected by measuring all the stabilizer generators .
we shall not present a detailed analysis here , since errors and syndrome are related for these codes in a similar way to the qubit case .
for example , in 3 spatial dimensions a single @xmath7-type error on an unstarred qudit will result in a measured eigenvalue of @xmath18 in the @xmath11 stabilizers corresponding to the four cells containing that vertex ( see figure [ fig : tetra ] ) .
similarly , a single @xmath11-type error on an unstarred qudit will result in a measured eigenvalue of @xmath18 in the six faces that contain the vertex ( see figure [ fig : tetra ] ) .
in general , a @xmath177 or @xmath178 error will lead to a measurement outcome of @xmath179 .
syndromes arising from sets of multiple errors can be calculated using standard techniques .
a classical algorithm called a _ decoder _ may be used to interpret the syndrome and infer the correction operator that must be applied to return the code to its original state .
there are many examples of decoding algorithms for 2d qubit color codes that have been developed recently @xcite , but very little is known for the case of higher qudit and spatial dimensions .
however , in the case of surface codes , it has been shown that some proposed renormalization group ( rg ) algorithms can be generalized to qudit codes in a straightforward manner @xcite .
moreover , recent work for the 3d fault - tolerant implementation of the qudit surface code using a hard - decision rg decoder @xcite suggests that adaptations of such algorithms for higher spatial dimensions should be possible , however , the error correction thresholds would likely be degraded as the spatial dimension increases due to the larger stabilizer generators .
we shall not present a decoder in this paper , but remark that , unlike in qubit codes where neighboring errors lead to cancellation of syndromes in between , such cancellation becomes unlikely with increasing qudit dimension , and thus in general more information is available to the decoder .
in fact , it has been shown that it is possible to design decoders to exploit this additional information to obtain higher thresholds with respect to qubit codes @xcite .
for the case of gauge color codes , very recently a hard - decision rg decoder @xcite was implemented for the qubit 3d construction .
as with the color code , we expect a generalization of such an algorithm to gauge color codes in arbitrary spatial and qudit dimensions to be straightforward .
brayvi and haah introduced a powerful framework for defining codes with a transversal non - clifford gate ; codes defined in terms of triorthogonal matrices . here
we introduce a significant generalisation to bravyi and haah s approach , extending to qudits codes with transversal gates from higher up the clifford hierarchy .
we begin by reviewing how bravyi and haah represent css codes with matrices @xcite .
they use a matrix @xmath181 over @xmath182 that has linearly independent rows under modulo 2 arithmetic .
the matrix @xmath181 is broken up into two blocks @xmath183 ( with @xmath93 rows ) and @xmath184 ( with @xmath185 rows ) .
this defines a quantum code with @xmath93 logical qubits .
the elementary logical basis state @xmath186 is written @xmath187 } \ket{f}\nonumber\\ & = \frac{1}{\sqrt{2^s } } \sum_{y \in \mathbb{z}_{2}^{s } } \ket { y^{t } \cdot g_0 } .
\label{eqn : log_0_state}\end{aligned}\ ] ] where `` @xmath188 '' denotes matrix multiplication modulo @xmath0 , and we take advantage of the fact that the rows of @xmath189 are linearly independent to represent the terms in the superposition by a row vector @xmath119 transposed and multiplied with @xmath184 . for the other logical computational basis states @xmath190 where @xmath191
, we have @xmath192 where addition of row vectors is elementwise modulo 2 as acknowledged by the symbol @xmath193 .
note that in the special case that @xmath194 , @xmath118 is a scalar ( or a @xmath195 row vector ) and the transpose operation is trivial . for qudits ,
quantum codes are defined by a matrix @xmath181 taking elements from @xmath196 , so that @xmath197 the key differences are that we now sum over all @xmath198 and that arithmetic is modulo @xmath67 . in the css formalism
, a quantum code has a stabilizer group which is generated by a product of two distinct generating sets , the @xmath11-stabilizer generators , which are tensor products of @xmath11 and @xmath199 alone and the @xmath7-stabilizer generators , which are tensor products of @xmath7 and @xmath199 alone .
in addition , one must define logical operators , and in the css formalism , logical encoded @xmath200 operators consist of a tensor product of @xmath11 and ( optionally ) @xmath199 alone , and the logical encoded @xmath201 consists of a tensor product of @xmath7 and ( optionally ) @xmath199 alone . once the @xmath11-stabilizer generators and logical @xmath202 operator ( or operators if there are multiple encoded qubits ) are defined , the remaining @xmath7-stabilizer generators are fixed by conjugation relations which can be elegantly captured via the use of dual codes from classical coding theory . in the qudit
setting , @xmath184 will again define the @xmath11-stabilizer generators for the code , via the mapping @xmath93 to @xmath178 , e.g. the row vector @xmath203 defines the operator @xmath204 , and @xmath183 will similarly define the logical @xmath11 operators for the code .
bravyi and haah @xcite prove that a so - called triorthogonal code supports a transversal non - clifford gate in the 3rd level of the clifford hierarchy .
we now generalize bravyi and haah s construction .
first , we focus on generalising from qubits to qudits .
we defined qudit color codes using a star - bipartition , and it is useful to incorporate that into the definition .
we introduce the star sign - flip matrix @xmath205 .
each column of matrix @xmath181 corresponds to a qudit , and let us order these columns such that the first @xmath206 columns correspond to unstarred qudits , and the latter @xmath207 columns to starred qudits .
we then define the @xmath208 matrix @xmath205 : @xmath209 where the first @xmath210 elements on the diagonal are @xmath96 and the remaining elements on the diagonal are @xmath211 .
we note that the ordering of the columns here is arbitrary , so we do not lose any generality by ordering the columns in this manner
. the purpose of @xmath205 will be to flip the sign of the entries of row vectors corresponding to the starred qudits , e.g. if @xmath212_1 , [ g]_2 , \ldots , [ g]_n \}$ ] is a row vector , we define @xmath213 as the row vector where @xmath214_j=[g]_j$ ] for the first @xmath206 elements ( corresponding to unstarred qudits ) and @xmath214_j =- g_j$ ] for the remaining elements ( corresponding to starred vertices ) .
we also define the weight @xmath215 of a row vector as @xmath216_j$ ] .
we now will define a significant generalisation of triorthogonal matrices , which we call @xmath180-orthogonality ( pronounced @xmath217 star orthogonality " ) .
[ mdefuniv ] an @xmath208 matrix @xmath181 over @xmath218 with a a bipartition of columns into @xmath219 is @xmath180-orthogonal if both of the following conditions hold . 1 .
the weight of every elementwise product of any @xmath217 rows of @xmath220 ( including repeated rows ) is equal to @xmath221 , except for the following case : 2 .
for each row of submatrix @xmath222 , the weight of the row vector raised to the @xmath217th power elementwise is @xmath223 . where @xmath205 is defined as in equation and where the matrices columns are ordered to respect to the star - bipartition @xmath219 .
notice the above makes no mention of modular arithmetic .
if we had instead defined the weights modulo @xmath67 , we would have a weaker notion of @xmath180-orthogonality .
color codes turn out to satisfy this stronger notion , and so this definition suffices for this paper .
we note that the results below rely on the stronger form to deal with the exceptional cases , e.g. @xmath224 for the @xmath53 gate ( see app .
[ sec : m - ortho ] )
. we also remark on the weaker notion to clarify that for @xmath225 it exactly corresponds to the brayvi - haah definition of triorthogonality .
other applications of weak @xmath180-orthogonality include quantum reed - muller codes @xcite .
a more symbolic statement of @xmath180-orthogonality will prove useful in subsequent sections .
we make use of of @xmath226 symbol to denote element - wise products of vectors @xmath227 and @xmath154 , such that it has elements @xmath228_{j } = [ g_{a}]_j [ g_{b}]_j , \label{eqn : circle_prod}\ ] ] which generalizes for an arbitrary number of vectors , e.g. for three vectors @xmath229_{j } = [ g_{a}]_j [ g_{b}]_j [ g_{c}]_j .\ ] ] let us now consider a reformulation of points _ 1 _ and _ 2 _ of def .
[ mdefuniv ] . for any set of @xmath217 rows
, we have a list of @xmath217 row indices @xmath230 .
@xmath180-orthogonality demands that for all the @xmath231_{j } = 0\ ] ] unless all rows are identical and come from @xmath183 , in which case the weight is unity .
the concept of @xmath180-orthogonality is a powerful tool for studying the properties of qudit color codes due to the following lemma [ colorcode_morth_lemma ] a qudit color code for any @xmath232 defined on a lattice in spatial dimension @xmath1 where @xmath11-stabilizer generators are defined by @xmath160-cells with @xmath233 , and with a star - bipartition @xmath219 defined by the bipartition of the lattice , is a @xmath180-orthogonal code for all @xmath234 . to prove this ,
let us first prove a convenient lemma . [ colorcode_morth_sublemma ]
any @xmath180-orthogonal matrix @xmath181 which includes , as one of its rows , the all - ones vector is also @xmath235-orthogonal for all @xmath236 .
this follows since an elementwise product of @xmath217 vectors including the all - ones vector is equal to the elementwise product of @xmath237 vectors excluding the all - ones vector .
since the logical @xmath11 for all color codes is the transversal @xmath11 acting on all qudits , the matrix @xmath238 for these codes contains the all - ones row .
hence to prove lemma [ colorcode_morth_lemma ] we now only need to prove that a color code in @xmath217 spatial dimensions is @xmath180-orthogonal .
recall that the @xmath11-generators of the code are defined by @xmath160-cells .
taking as an example the 3d lattice in figure [ fig : tetra ] , the @xmath11-stabilizer generators are defined by 3-cells .
consider the following geometric properties of these cells .
when @xmath239 distinct @xmath160-cells intersect non - trivially , where they meet defines a cell of smaller dimension . for a general lattice , there is no further restriction on the dimension of this cell .
however , for a @xmath1-colex it is well known @xcite that the intersection of @xmath239 objects of dimension @xmath160 yields either an empty set or a cell of dimension @xmath240 .
for instance , the intersect of two neighboring @xmath160-cells defines a @xmath241-cell , where three such @xmath160-cells meet defines a @xmath242-cell , and so on .
where @xmath160 @xmath160-cells meet defines a 1-cell , or lattice edge .
for example , in 3d , two adjacent 3-cells meet at a face , and three adjacent 3-cells meet at an edge ( four adjacent 3-cells meet at a point ) .
we use this geometric fact to prove the theorem .
any product of @xmath217 row vectors in @xmath184 has a geometric representation as the intersection of the vertices of these cells .
this corresponds to either an empty cell or cell of dimension no less than @xmath243 .
each cell in the lattice , of any dimension greater than zero , has an equal number of starred and unstarred vertices ( see cor .
[ cor_punc ] ) .
when we multiply the product row vector by @xmath205 we invert the sign of the columns corresponding to the starred vertices , but this corresponds to precisely half of the non - zero elements in the vector .
hence the weight of the resultant vector is zero provided @xmath244 .
therefore , the rows of @xmath184 satisfy the conditions for @xmath180-orthogonality whenever @xmath245 .
the other case to consider is the product where all @xmath217 vectors in the product are the all - ones vector .
the element - wise product results trivially in the all - ones vector .
the number of unstarred qudits is one more than the starred qudits ( see cor . [ cor_punc ] ) , hence the weight of this vector after multiplying by @xmath205 is 1 .
together with lemma [ colorcode_morth_sublemma ] this proves lemma [ colorcode_morth_lemma ] .
note that the bounds in this lemma are tight .
a code whose @xmath11-stabilizer generators are defined by @xmath160-cells is not @xmath246-orthogonal because @xmath99 @xmath160-cells meet at a point , and a single vertex represents a vector which does not have zero weight .
our attention now turns to transversal gates .
the qubit color codes are the only family of topological codes known to support transversal gates satisfying the bravyi - knig bound .
we prove in this paper that ( apart from a minority of special cases which need to be treated individually ) the qudit codes we have defined also saturate this bound .
[ thm : mainresult ] the qudit color codes on a @xmath1-colex , where @xmath11-stabilizer generators are defined by @xmath247 cells and where @xmath248 , support transversal gates in the @xmath160th level of the clifford hierarchy ( ch ) ( which are not in the @xmath151th level of the ch ) , saturating the bravyi - knig bound .
the qualification @xmath248 is due to lemma [ rlemma2 ] , namely that when @xmath249 , gates of the form of equation are in the @xmath241th level of the ch ( and therefore does not saturate bravyi - knig ) as discussed in sec .
[ sec : stabilizer ] . for the most important @xmath4 case , however , we additionally prove transversality for the exceptional @xmath58 and @xmath59 cases , and can thus state the unqualified result : [ thm : mainresulttgates ] the qudit color codes on a @xmath1-colex , where @xmath11-stabilizer generators are defined by @xmath247 cells support transversal non - clifford gates in the @xmath61rd level of the clifford hierarchy ( ch ) for all @xmath250 .
we believe that the qualification in theorem [ thm : mainresult ] can be removed , following a similar approach to proving theorem [ thm : mainresulttgates ] , but we leave this for future work . to prove theorem [ thm : mainresult ] , we show that gates of the form of equation are transversal in all @xmath180-orthogonal codes for polynomial degree @xmath251 . to prove theorem [ thm : mainresulttgates ] we also prove that gates of the form in equation are transversal in @xmath252-orthogonal codes .
the theorems then follow by virtue of lemmas [ rlemma1 ] , [ rlemma2 ] and [ colorcode_morth_lemma ] , together with the known results for qubit color codes @xcite .
in particular , the transversal gates we consider are of star - conjugate transversal form , see equation [ eqn : u_trans ] .
[ mtranslemma ] a @xmath180-orthogonal code has a transversal implementation of the unitary gate @xmath253 such that : @xmath254 \otimes r_{f_{{m}}}^{*}[v_{\star } ] \label{eqn : r_trans}\ ] ] implements @xmath253 on the code space .
we prove this lemma in app .
[ sec : m - ortho ] . as a simplified presentation for the main text of the paper
, we prove the special and important case of @xmath90 .
the general proof follows the same approach , but is a little notationally unwieldy .
[ tgatelemma ] a @xmath252-orthogonal code has a transversal implementation of the unitary gate @xmath255 such that : @xmath256 \otimes t^{*}[v_{\star } ] \label{eqn : t_trans}\ ] ] implements @xmath53 on the code space .
we now examine the conditions that the star - conjugate transversal @xmath257 implements a logical @xmath258 . for this to be true
, the phases for each computational basis component of each logical codeword in equation , must agree with the phases defining the gate in equation .
noting that the same phase applies to all terms in the sum in equation , we can write @xmath259 where @xmath260 and @xmath261 , where @xmath185 is the number of stabilizer generators . in other words
we fix the number of logical qubits @xmath194 . applying equations and to the state on the right hand side of equation
, we recover the following relationship between the phases on both sides of the equation : @xmath262 where we have right multiplied the @xmath205 matrix to vector @xmath263 .
it is useful to define @xmath264 so @xmath265 now define @xmath266 where @xmath267 is the value of the @xmath268th position of the vector @xmath269 , and where @xmath270 are the basis vectors over @xmath271 .
for example @xmath272 , @xmath273 and so on .
re - writing the left hand side of equation we obtain @xmath274 note that @xmath275 is the @xmath268th row of the matrix g , which we denote @xmath276 . therefore @xmath277 we can re - write this as @xmath278_{j } z_{a } z_{b } z_{c } \pmod{{d}}. \label{eqn : ref_from_appendix}\end{aligned}\ ] ] summing over @xmath279 gives @xmath280 we break the sum into two parts , so that @xmath281 now we use the definition of @xmath180-orthogonal matrices .
this ensures that the first term is equal to @xmath282 and the second term is equal to zero .
hence equation is satisfied , which completes the proof of lemma [ tgatelemma ] . for larger @xmath217
, the proof follows the same lines but is with more cumbersome algebra .
the proof of lemma [ mtranslemma ] which we present in app .
[ sec : m - ortho ] completes the proof of theorem [ thm : mainresult ] . to prove theorem [ thm : mainresulttgates ] we require one further lemma : [ tgatelemma36 ] a @xmath252-orthogonal qudit code where @xmath283 or @xmath284 has a transversal implementation of the unitary gate @xmath285 defined in equation such that : @xmath286 \otimes t_{3,6}^{*}[v_{\star}]\ ] ] the proof is presented in app .
[ sec : triortho_3_and_6 ] , completing the proof of theorem [ thm : mainresulttgates ] . finally , we remark , that as css codes , all color codes admit a transversal @xmath39 . we conclude this section with a table of star - conjugate transversal gates in qudit color codes of different spatial dimension , see table [ gatetable ]
. | c |c | y | c| logical operator & defined in equation & star - conjugate transversal implementation & spatial dimension + @xmath37 & equation & @xmath287 \otimes h^{*}[v_{\star}]$ ] & @xmath97 + @xmath39 & equation & applied blockwise transversally & @xmath288 + @xmath38 & equation & @xmath289 \otimes s^{*}[v_{\star}]$ ] & @xmath288 + @xmath53 & equation & @xmath290 \otimes t^{*}[v_{\star}]$ ] & @xmath291 + @xmath77 & equation & @xmath292 \otimes r_{f_{r}}^{*}[v_{\star}]$ ] & @xmath293 +
in this section and sec . [ sec : gauge_color_codes ] we show how a universal gate set can be achieved via gauge fixing , or by defining a gauge color code .
although these results follow directly from the qubit cases presented in @xcite and @xcite without the need for any further technical innovations , we include a discussion and explanation of these techniques for completeness .
first we consider gauge fixing .
we shall define a _ subsystem code _ occupying the same lattice as the color code .
this code is capable of realising the logical hadamard gate on the same encoded qudit as the color code , while a set of gauge qudits become corrupted .
fortunately there is a simple protocol to fix the corrupted gauge qudits to re - initialize the color code . in a @xmath294 $ ] subsystem code @xcite
there are three sets of operators that define the code .
first , a set of gauge generators is defined , a set of pauli operators defined on the lattice that do not necessarily mutually commute .
these generate the gauge group @xmath295 .
the center of the gauge group ( the elements of the group which commute with all other elements ) represents the stabilizer of the code @xmath31 ( up to a free choice of plus or minus phases which must be chosen to define the zero - error code space ) .
the final set of operators to define is the set of logical pauli operators on the code space , the pauli operators that commute with @xmath295 .
subsystem codes can be understood as stabilizer codes in which some of the logical qudits have been demoted to gauge qubits .
these gauge qubits are not used to carry information and can be corrupted or measured during operations on the code .
the logical operators @xmath202 and @xmath175 are products of @xmath11 and @xmath7 ( and @xmath165 and @xmath296 ) operators acting on every qudit .
therefore , a natural candidate for transversal @xmath297 is a global star - conjugate hadamard .
however , for color codes of @xmath3 it can be seen from the structure of the stabilizer generators that such a unitary will not leave the code space invariant .
in particular , the @xmath11 stabilizers act on the @xmath160-cells of the lattice whereas the @xmath7 stabilizers act on the @xmath161-cells of the lattice . therefore unless @xmath298 , the @xmath7 stabilizers can not be transformed into valid @xmath11 ( cell ) stabilizers , by a global hadamard . the solution to this is to switch between the stabilizer color codes and a different code in which @xmath11 stabilizer generators and @xmath7 stabilizer generators have identical support . in @xmath1 spatial dimensions
the subsystem code stabilizers are defined as the star - conjugate @xmath11 and @xmath7 operators acting on the qudits contained in a @xmath1-cell of the lattice .
the logical operators matching the color code logical operators are the transversal @xmath202 and @xmath175 defined in equations .
in addition there are @xmath11 and @xmath7 gauge operators defined on the @xmath151-cells , @xmath299-cells , and so on to the @xmath97-cells of the lattice , such that they commute with the stabilizer group .
they can be identified in @xmath18-commuting pairs , with each pair protecting a gauge qudit . in higher spatial dimensions
there are a larger number of gauge generators and hence more gauge qudits than in lower spatial dimensions .
nevertheless , the basic protocol for realising the logical hadamard gate in the subsystem code and fixing the gauge to return to the stabilizer code is independent of the spatial dimensions . the star - conjugated transversal hadamard gate
, @xmath300 \otimes h^{*}[v_{\star } ] , \label{eqn : trans_h}\end{aligned}\ ] ] transforms the transversal logical operators and the stabilizers of the code correctly .
however the gauge generators are corrupted under the action of @xmath297 , meaning that @xmath301 ( the @xmath7 logical operator for qudit @xmath279 ) is transformed not to @xmath302 as desired but instead to the operator for a different qudit , @xmath303 , causing the gauge qudits to become entangled .
these corrupted gauge operators must be _ fixed _ , to return the code to the original stabilizer code in which the other transversal operators can be realized .
this fixing is enacted by measuring the stabilizer generators .
the measured eigenvalues will indicate the pauli correction operator that should be applied this is an operator that restores the corrupted gauge qudit to its initial state while commuting with the other logical operators .
once the corrections are applied , the gauge qudits are re - initialized to their original state so that any entanglement between the gauge qudits is destroyed .
this is equivalent to the stabilizer color code with a logical hadamard applied to the encoded qudit .
another approach to the implementation of the transversal hadamard is to treat the subsystem code as the base code , and to achieve the transversal non - clifford gates by gauge fixing to the ( stabilizer ) color code .
this switch in perspective leads to the _ gauge color codes _ , recently introduced by bombin @xcite .
far more than a change in perspective , however , gauge color codes gain many useful new properties , which the original color codes do not possess .
the first advantage of the gauge color code construction is that the outcome of stabilizer measurements can be inferred from measurements of the gauge operators .
this offers a practical advantage , since the weight of the gauge generators can be significantly smaller than stabilizer generators .
for example , in 3d the @xmath11 stabilizer generators would include 24-body operators associated with truncated octahedrons , discussed in sec .
[ sec : code_distance ] , requiring coherent 24-body measurement . in the equivalent gauge color code , however , the @xmath11 stabilizers can be decomposed into 4- or 6-body @xmath11 gauge operator measurements , a much more practically feasible proposition . a second advantage ( in the @xmath4 case )
is that the outcome of stabilizer measurements is more robustly encoded in the gauge generator measurement .
the extra redundancy leads to to `` single - shot fault - tolerant error correction '' @xcite meaning that measurement errors can be accounted for without the need for repeated measurements .
a third advantage is that one can use gauge fixing to fault tolerantly convert @xcite between codes of different spatial dimension @xcite , allowing one to combine the advantageous features of both .
for example , one could perform all clifford computations in a 2-d code while only resorting to a 3-d code to achieve a non - clifford gate .
we can define qudit gauge color codes in any spatial dimension analogously to bombin s qubit codes @xcite .
the only difference is that operators are star - conjugated according to the star - bipartition of the lattice .
the most studied gauge color code exists on a 3d lattice and we shall describe the qudit analogue as an example .
gauge generators consist of star - conjugated face operators of both @xmath11 and @xmath7-type .
the stabilizer is generated by 3-cell operators whose values can be obtained from the products of the face operators .
the star - conjugate transversality of the hadamard gate for this code is trivial via its definition . to achieve a transversal @xmath53 gate
one must gauge fix back to the stabilizer color code by promoting @xmath7 face operators to the stabilizer .
although the basic properties of color codes transfer immediately to the general qudit case via this definition and by the results presented in previous sections in this paper , the performance of single - shot error correction is less straightforward to analyze , and we leave a detailed study to further work .
single - shot error correction works , in the qubit case , because the relationship between gauge and stabilizer measurement outcomes in the code is such that the values of the stabilizer are mapped to delocalized structures ( branching points in string - like extended structures ) in the gauge outcomes .
this arises because the parity of the outcomes of the face operators around the cell must be equal to the eigenvalue of the stabilizer representing the cell , even parity in for a trivial syndrome , and odd - parity for the non - trivial case . in qudit codes
, the stabilizer measurement outcome may take any of @xmath304 non - trivial values .
the outcomes of gauge measurements must sum to these values .
thus the relationship between gauge measurements and stabilizer measurements is more complicated than the simple `` branching point or no branching point '' behavior seen in the qubit code .
this extra structure , however , should represent additional information about the location and charge of stabilizer outcomes which a decoder could exploit .
thus , with the development of a suitable decoder , we expect single - shot error correction robust to measurement error to be achieved in qudit gauge color codes , but we leave its the construction of such a decoder , and analysis of its performance , to future work .
the outcomes of the gauge measurements for each of the red faces , or each of the blue faces , or each of the green faces.,scaledwidth=30.0% ]
we have shown how any existing qubit color code can be generalized to support a qudit color code in arbitrary spatial dimensions , and with an arbitrarily large distance . to do this
we introduced the notion of a star - conjugate lattice and showed how this construction allows the clifford phase , controlled-@xmath11 and non - clifford phase gates to be implemented transversally .
the set of clifford group generators is completed by the hadamard , which can be implemented in a subsystem code . techniques for switching between the two codes have been outlined .
an additional advantage of considering the color codes as gauge codes is reduction of the weight of measurements required to obtain the error syndrome .
there are still many open questions remaining here we shall outline some of the ones we find most interesting .
although a fault - tolerant decoder has been implemented for a 2d code , no decoder has , to our knowledge , been proposed for color codes in higher spatial dimensions . in ref .
@xcite a decoder for a 3d gauge color code is discussed , unfortunately however , a classical algorithm capable of performing the decoding is not specified .
one would expect that in a fully fault - tolerant simulation the gate error threshold for the @xmath1-dimensional color codes would be prohibitively low , since the stabilizers defined on the @xmath1-cells would require circuits of many time steps to measure .
there is hope however that by switching to the corresponding gauge color code the measurements require fewer gates and therefore the threshold may recover @xcite .
the development of efficient decoders for qudit gauge color codes is therefore of significant interest .
the construction of the transversal gates in the color codes and gauge color codes presented in this work holds for qudits of any dimension , @xmath0 .
however , for non - prime @xmath0 it is not clear that @xmath38 , @xmath37 and @xmath39 are sufficient to generate the clifford group .
we desire a better understanding of universality and the clifford hierarchy when using systems of non - prime dimension .
we have identified some other lattices that can support a color code in 3d .
it would be interesting to see a comprehensive list of possible lattices .
furthermore , our star - conjugate construction is valid for all of these lattices , but there may be one that offers advantages in terms of the necessary qudit overhead to achieve a desired distance . finally , there are many reed - muller codes known for qudits for which no equivalent color code is known @xcite .
is it possible to use one of these to construct a topological code similar to a color code and study its properties ?
here , we found codes where the distance is purely topological in nature , and so depends on the lattice not on whether we use qudits .
in contrast , qudit reed - muller codes provide an improved distance that is algebraic rather than topological in origin .
the potential exists for qudit quantum codes which abstract topological ideas to a more general setting to generate novel codes of arbitrary code distance with a rich family of transversal logical gates .
deb would like to thank hector bombin and ben brown for helpful and elucidatory discussions on color codes and gauge color codes .
fhew and ha acknowledge the financial support of the epsrc ( grant numbers ep / g037043/1 and ep / k022512/1 , respectively ) . 80 d.e .
knuth , art of computer programming vol .
2 : seminumerical algorithms 3rd edition , _ addison - wesley professional _ ( 1997 ) . d. gottesman , chaos , solitons & fractals * 10 * , 1749 - 1758 ( 1999 ) .
d. l. zhou , b. zeng , z. xu and c. p. sun , phys .
a * 68 * , 062303 ( 2003 ) .
m. howard and j. vala , phys .
a * 86 * , 022316 ( 2012 ) .
h. anwar , e.t .
campbell and d.e .
browne , new j. phys .
* 14 * , 063006 ( 2012 ) .
campbell , h. anwar and d.e .
browne , phys .
x * 2 * 041021 , ( 2012 ) .
g. duclos - cianci and d. poulin , phys .
a * 87 * 062338 ( 2013 ) .
h. anwar , b.j .
brown , e.t .
campbell and d.e .
browne , new j. phys .
* 16 * , 063038 ( 2014 ) .
f.h.e . watson , h. anwar and d.e .
browne , arxiv:1411.3028 .
a. hutter , d. loss and j.r .
wootton , arxiv:1410.4478 .
a. smith , b.e .
anderson , h. sosa - martinez , c.a .
riofro , i.h .
deutsch and p.s .
jessen , phys .
lett . * 111 * , 170502 ( 2013 ) .
anderson , h. sosa - martinez , c.a .
riofro , i.h .
deutsch and p.s .
jessen , arxiv:1410.3891 .
h. bombin and m.a .
martin - delgado , phys .
* 97 * , 180501 ( 2006 ) . h. bombin and m.a .
martin - delgado , phys .
lett . * 98 * , 160502 ( 2007 ) h. bombin , arxiv:1311.0879 . h. bombin and r.w . chhajlany and m. horodecki and m.a .
martin - delgado , new j. phys .
* 15 * , 055023 ( 2013 ) .
a. kubica and m.e .
beverland , arxiv:1410.0069 .
kitaev , ann .
phys . * 303 * , 2 ( 2003 ) .
s.b . bravyi and a.y .
kitaev , arxiv : quant - ph/9811052 .
p. sarvepalli , _ ieee itw _ 15 ( 2010 ) .
brell , arxiv:1408.6238 .
e. knill , r. laflamme and w.h .
zurek , arxiv : quant - ph/9610011 .
e. knill , r. laflamme and w.h .
zurek , science * 279 * , 342 - 345 ( 1998 ) .
r. raussendorf , j. harrington , and k. goyal , ann . phys . * 321 * , 2242 - 2270 ( 2006 ) .
s. bravyi and r. knig , phys .
lett . * 110 * , 170503 ( 2013 ) .
a. paetznick and b.w reichardt , phys .
lett . * 111 * , 090505 ( 2013 ) .
anderson , g. duclos - cianci and d. poulin , phys .
rev . lett . *
113 * , 080501 ( 2014 ) .
t. jochym - oconnor and r. laflamme , phys .
lett . * 112 * , 010505 ( 2014 ) .
s. bravyi and a. kitaev , phys . rev .
a * 71 * , 022316 ( 2005 ) .
reichardt , quant .
inf . and comp . * 11&12 * , 1030 ( 2009 ) .
s. bravyi and j. haah , phys .
a * 86 * 052329 ( 2012 ) . c. jones , phys .
a * 87 * , 042305 ( 2013 ) .
a.m. meier , b. eastin and e. knill , quant . inf . and comp . * 3&4 * , 195 ( 2013 ) .
campbell , phys .
lett . * 113 * , 230501 ( 2014 ) .
appleby , j. math . phys .
* 46 * , 052107 ( 2005 ) .
h. bombin , arxiv:1412.5079 .
h. bombin , arxiv:1404.5504 .
d. gottesman and i.l .
chuang , nature * 402 * , 390393 ( 1999 ) .
f. pastawski and b. yoshida , phys .
rev . a * 91 * 012305 ( 2015 ) .
s. clark , j. phys .
39 * , 2701 ( 2006 ) .
h. zhu , j phys . a - math
* 30 * , 305305 ( 2010 ) .
h. bombin and m.a martin - delgado , phys .
b * 75 * , 075103 ( 2007 ) .
j. lee , introduction to topological manifolds , springer ( 2010 ) d.s .
wang , a.g .
fowler , c.d .
hill and l.c.l .
hollenberg , arxiv:0907.1708 . h.g .
katzgraber , h. bombin and m.a .
martin - delgado , phys .
lett . * 103 * , 090501 ( 2009 ) .
m. ohzeki , phys .
e * 80 * , 011141 ( 2009 ) .
andrist , h.g .
katzgraber , h. bombin and m.a .
martin - delgado , new j. phys .
* 13 * , 083006 ( 2011 ) .
landahl , j.t .
anderson and p.r .
rice , arxiv:1108.5738 .
n. delfosse , phys .
rev . a * 89 * , 012317 ( 2014 ) .
p. sarvepalli and r. raussendorf , phys .
a * 85 * , 022317 ( 2012 ) .
brown , n.h .
nickerson and d.e .
browne , arxiv:1503.08217 .
g. duclos - cianci and d. poulin , quant .
inf . comp . * 9&10 * , 0721 ( 2014 ) .
d. bacon , phys .
a * 73 * , 012340 ( 2006 ) .
d. kribs , r. laflamme and d. poulin , phys .
* 94 * , 180501 ( 2005 ) .
d. poulin , phys .
lett . * 95 * , 230504 ( 2005 ) .
d. kribs , r. laflamme , d. poulin and m. lesosky , arxiv:0504189 .
a @xmath180-orthogonal code has a transversal implementation of the unitary gate @xmath253 such that : @xmath254 \otimes r_{f_{{m}}}^{*}[v_{\star } ] \label{appendixlemma7}\ ] ] implements @xmath253 on the code space .
for equation to be true , for all @xmath264 we have @xmath305 applying equations and to the state on the right hand side of equation , the phases on both sides of the equation are equal if @xmath306 where the notation @xmath307 means a circle - product ( see equation ) taken between @xmath217 identical elements .
proceeding using the same notation as in sec .
[ sec : transversal ] we find @xmath308^{{m } } } z_{\vec{a } } \ : g_{\vec{a}}.f \right)_{j } \pmod{{d}},\end{aligned}\ ] ] where @xmath309^{c}$ ] is a vector of length @xmath48 with elements that take values between @xmath268 and @xmath310 , and where @xmath311 is the bitwise product of @xmath217 rows of @xmath181 , so @xmath312 we can re - write this as @xmath313^{{m } } } \left [ g_{\vec{a } } \cdot f \right]_{j } z_{\vec{a } } \pmod{{d}}.\end{aligned}\ ] ] summing over @xmath279 gives @xmath314^{{m } } } \left| g_{\vec{a } } \cdot f \right| z_{\vec{a } } \pmod{{d}}. \label{eqn : long_m - ortho}\end{aligned}\ ] ] we now use the definition of a @xmath180-orthogonal code .
this means that all terms satisfying @xmath315 in the summation on the right hand side vanish .
this implies that the only non - zero contribution is represented by @xmath316 , a @xmath217-element vector where every element is @xmath268 . recall also
that such terms vanish unless @xmath317 , in other words it is the @xmath217-fold elementwise product of the vector in @xmath238 .
we note that @xmath318 , leaving only @xmath319 which proves the desired equation is satisfied .
a @xmath252-orthogonal qudit code where @xmath283 or @xmath284 has a transversal implementation of the unitary gate @xmath285 defined in equation such that : @xmath320 \otimes t_{3,6}^{*}[v_{\star}]\ ] ] where @xmath64 and where the function in the exponent @xmath65 is evaluated in regular arithmetic ( or equivalently modulo @xmath66 ) .
we shall prove lemma [ tgatelemma36 ] for @xmath58 and @xmath59 separately , since each requires calculation in arithmetic with a different modularity .
first we consider the case of qutrits , where @xmath283 . in this case ,
@xmath322 . hence
, our calculation will make use of addition modulo 9 .
proceeding as before we start with : @xmath323 the phases on both sides of this equation match if @xmath324 this expression is a mix of modulo 3 and modulo 9 arithmetic so we must proceed with care .
the matrix multiplication is mod 3 , whereas all other arithmetic is mod 9 .
re - writing the left hand side as @xmath325_{j } z_{a } \pmod{3 } \right\ } \circ \nonumber \\ & \quad \left\ { [ g_{b } \cdot f]_{j } z_{b } \pmod{3 } \right\ } \circ \nonumber \\ & \qquad \left\ { [ g_{c } \cdot f]_{j } z_{c } \pmod{3 } \right\},\end{aligned}\ ] ] we can use the following identity to express the modulo 3 reduction in terms of standard addition @xmath326 setting aside the modulo 9 arithmetic for the moment and evaluating this expression first in standard arithmetic , the left hand side now expands to the following form : @xmath327_{j } z_{a } - 3\lfloor \frac{[g_{a } \cdot f]_{j } z_{a}}{3 } \rfloor \right ) \nonumber \\ & \qquad \circ \left ( [ g_{b } \cdot f]_{j } z_{b } - 3\lfloor \frac{[g_{b } \cdot f]_{j } z_{b}}{3 } \rfloor \right ) \circ \left ( [ g_{c } \cdot f]_{j } z_{c } - 3\lfloor \frac{[g_{c } \cdot f]_{j } z_{c}}{3 } \rfloor \right ) \\ & = \sum_{j } \sum_{a , b , c=1}^{s+1 } [ ( g_{a } \circ g_{b } \circ g_{c } ) \cdot f]_{j } z_{a } z_{b } z_{c } - 9 [ ( g_{a } \circ g_{b})\cdot f]_{j } z_{a } z_{b } \lfloor \frac{[g_{c } \cdot f]_{j } z_{c}}{3 } \rfloor \nonumber\\ & \qquad + 27 [ g_{a } \cdot f]_{j } z_{a } \lfloor \frac{[g_{b } \cdot f]_{j } z_{b}}{3}\rfloor \lfloor \frac{[g_{c } \cdot f]_{j } z_{c}}{3}\rfloor \nonumber\\ & \qquad\qquad - 27 \lfloor \frac{[g_{a}\cdot f]_{j } z_{a}}{3}\rfloor\lfloor \frac{[g_{b } \cdot f]_{j } z_{b}}{3}\rfloor \lfloor \frac{[g_{c}\cdot f]_{j } z_{c}}{3}\rfloor .
\end{split}\ ] ] the last three terms in this summation are integer multiples of 9 and are thus equal @xmath328 . thus if we reimpose modulo 9 arithmetic we are left with : @xmath329_{j } z_{a } z_{b } z_{c } \nonumber \\ & \hspace{4 cm } \pmod{9}\end{aligned}\ ] ] when @xmath284 the phase in @xmath285 is @xmath330 and thus proceed with modulo 18 arithmetic .
@xmath323 we expand the right hand side and find that @xmath331_{j } z_{a } z_{b } z_{c } \pmod{18}.\end{aligned}\ ] ] @xmath332_{j } z_{a } z_{b } z_{c } - 18 [ ( g_{a } \circ g_{b})\cdot f]_{j } z_{a } z_{b } \lfloor \frac{[g_{c } \cdot f]_{j } z_{c}}{6 } \rfloor \nonumber\\ & \qquad + 108 [ g_{a } \cdot f]_{j } z_{a } \lfloor \frac{[g_{b } \cdot f]_{j } z_{b}}{6}\rfloor \lfloor \frac{[g_{c } \cdot f]_{j } z_{c}}{6}\rfloor \nonumber\\ & \qquad\qquad - 216 \lfloor \frac{[g_{a}\cdot f]_{j } z_{a}}{6}\rfloor\lfloor \frac{[g_{b } \cdot f]_{j } z_{b}}{6}\rfloor \lfloor \frac{[g_{c}\cdot f]_{j } z_{c}}{6}\rfloor .
\end{split}\end{aligned}\ ] ]
|
two - level quantum systems , qubits , are not the only basis for quantum computation .
advantages exist in using qudits , @xmath0-level quantum systems , as the basic carrier of quantum information .
we show that color codes a class of topological quantum codes with remarkable transversality properties can be generalized to the qudit paradigm . in recent developments
it was found that in three spatial dimensions a qubit color code can support a transversal non - clifford gate , and that in higher spatial dimensions additional non - clifford gates can be found , saturating bravyi and knig s bound [ phys .
rev .
lett . * 110 * , 170503 ( 2013 ) ] .
furthermore , by using gauge fixing techniques , an effective set of clifford gates can be achieved , removing the need for state distillation .
we show that the qudit color code can support the qudit analogues of these gates , and show that in higher spatial dimensions a color code can support a phase gate from higher levels of the clifford hierarchy which can be proven to saturate bravyi and knig s bound in all but a finite number of special cases .
the methodology used is a generalisation of bravyi and haah s method of triorthogonal matrices [ phys .
rev .
a * 86 * 052329 ( 2012 ) ] , which may be of independent interest . for completeness
, we show explicitly that the qudit color codes generalize to gauge color codes , and share the many of the favorable properties of their qubit counterparts .
|
chaos synchronization has attracted considerably increasing attention and become an active area of research for both theoretical interests and practical applications , since seminal work of pecora and carroll @xcite . over the last decade , following the complete synchronization @xcite , several new types of synchronization have been found in interacting chaotic systems , such as generalized synchronization @xcite , phase synchronization @xcite , anti - phase synchronization @xcite , projective synchronization @xcite , lag synchronization @xcite and anticipating synchronization @xcite .
complete synchronization is characterized by the convergence of the two chaotic trajectories , @xmath3 .
it appears only when interacting systems are identical .
generalized synchronization means the amplitude of the slave s state variable correlated with that of master s by a generic function , @xmath4 .
phase synchronization is defined as the entrainment of phases of chaotic oscillators , @xmath5 ( @xmath6 and @xmath7 are integers ) , whereas their amplitude remains chaotic and uncorrelated .
projective synchronization is the dynamical behavior in which the amplitude of the master s state variable and that of the slave s synchronizes up to a constant scaling factor @xmath0 ( a proportional relation ) .
complete synchronization and anti - phase synchronization are the special cases of the projective synchronization in cases of @xmath1 and @xmath2 , respectively .
lag synchronization means a coincidence of shifted - in - time states of two coupled systems , the state variable of the drive system is delayed by positive @xmath8 in comparison with that of the driven : @xmath9 , @xmath10 . whereas for anticipating synchronization , the driven system anticipates the driver : @xmath11 , @xmath10 . among those synchronization ,
projective synchronization is one of the most interesting problems , because of its proportion between the synchronized dynamical states . in application to secure communications , this feature can be used to extend binary digital to m - nary digital communication @xcite for achieving fast communication .
the early study of projective synchronization reported that the projective synchronization was usually observable only in the coupled partially linear systems @xcite .
following work has extended that to a general class of chaotic systems without the limitation of partial - linearity @xcite .
recently , hoang _ et al_. investigated a new synchronization in time - delay chaotic system , which they called projective - anticipating synchronization @xcite .
projective - anticipating synchronization is a combination of the well - known schemes of projective and anticipating synchronization .
that is , the driver synchronizes with the driven under the anticipating synchronization scheme and the amplitude is correlated by a scaling factor @xmath0 : @xmath12 . in the case of projective - lag synchronization @xcite , the amplitude of the master s and slave s state variables
is correlated by a scale factor @xmath0 : @xmath13 . however , most existing work about projective - anticipating and projective - lag synchronization focused only on two coupled chaotic systems .
systems consisting of many interconnected subsystems are ubiquitous in nature and social science which can be described by complex networks @xcite .
complex networks are usually composed by a large number of interconnected components ( nodes ) representing individuals or organizations and edges mimicking the interaction among them . with the development of research on complex systems , more and more researchers carried out the study about complex dynamical behaviors on networks .
while most studies focused on complete @xcite and phase synchronization @xcite in various networks , little attention has been paid to projective synchronization .
very recently , hu et al .
@xcite studied the projective synchronization on drive - response dynamical networks by considering coupled lorenz chaotic systems .
but due to finite signal transmission times , switching speeds and memory effects with both single and multiple delays are ubiquitous in nature , technology and society @xcite .
time - delayed systems are also interesting because the dimension of their chaotic dynamics can be increased by increasing the delay time sufficiently @xcite . from this point of view
, these systems are especially appealing for secure communication schemes .
in addition , time - delayed system can be considered as a special case of spatiotemporal systems @xcite .
so it is natural to consider an interesting topic whether we can achieve projective - anticipating , projective , and projective - lag synchronization of time - delayed chaotic systems on complex dynamical networks ? in this paper , inspired by the above discussions , we extends the work on projective - anticipating , projective , and projective - lag synchronization of two coupled chaotic systems to complex dynamical networks .
we attempt to achieve projective - anticipating , projective and projective - lag synchronization in a general class of time - delayed chaotic systems related to optical bistable or hybrid optical bistable device on complex dynamical networks .
the layout of this paper is as follows . in sec .
ii , according to the lyapunov stability theory @xcite , we theoretically analyze both the existence and sufficient stability conditions of the projective - anticipating , projective , and projective - lag synchronization of time - delayed chaotic systems on complex dynamical networks . in sec .
iii , the well - known ikeda and mackey - glass systems are considered as the dynamics of single node to prove the validity of the proposed theoretical approach in sec .
ii , respectively .
finally we end this paper by a short conclusion in last section .
for convenience , we study different types of projective synchronization of infinite - dimensional chaotic systems on erds - rnyi ( er ) random networks @xcite .
consider an er network consisting of @xmath14 nodes which constitute the drive - response dynamical networks @xcite . in response dynamical networks , we connect each pair of nodes with probability @xmath15 , and denote the degree of node @xmath16 with @xmath17 .
each individual node is a delay - differential system related to the optical bistable or hybrid optical device @xcite @xmath18 where @xmath19 is the dimensionless output of the system at time @xmath20 , @xmath21 is the delay time of the feedback loop , @xmath22 is the response time of the nonlinear medium , @xmath23 is proportional to the intensity of the incident light and @xmath24 is the parameter . in eq.([single1 ] ) , @xmath25 is a nonlinear function of @xmath26 , characterizing the system , e.g. , @xmath27 $ ] for ikeda model @xcite , @xmath28 $ ] for valle model @xcite , @xmath29 for the sine - square model @xcite , and @xmath30 for mackey - glass model @xcite . in order to observe the different types of projective synchronization , the drive - response dynamical networks are described by the following equations : @xmath31 @xmath32 where @xmath26 denotes the drive system , @xmath33 denotes the response dynamical network systems , @xmath34 is the coupling strength , @xmath0 is a desired scaling factor present for the projective - anticipating , projective , and projective - lag synchronization , @xmath35 is the feedback delay time in the coupled systems and @xmath36 is the coupling delay time between systems @xmath26 and @xmath33 . the graph topology is encoded in the laplacian g , a symmetric matrix with zero row - sum , @xmath37 the eigenvalues @xmath38 of @xmath39 are real and nonpositive @xcite . throughout this study , for simplification
we use the notation @xmath40 . in the following
, we will investigate different kinds of projective synchronization .
one can find that under conditions @xmath41 eqs .
( [ driver1 ] ) and ( [ responser1 ] ) allow for synchronization manifold @xmath42 we define the error systems as the difference between the eqs .
( [ driver1 ] ) and ( [ responser1 ] ) : @xmath43 , one obtains error dynamics @xmath44,\\ \label{error } \end{split}\ ] ] according to eq.([manifold ] ) , @xmath45 contained in eq.([error ] ) can be expressed as @xmath46 .
therefore , combining eq.([existence ] ) , eq.([error ] ) can be reduced as @xmath47 .
\label{error2}\ ] ] consider a lyapunov function in the form @xmath48 according to the lyapunov stability theory @xcite , if the function ( [ lyapunov ] ) satisfies the following conditions @xmath49 and @xmath50 then @xmath51 will asymptotically converge to zero as time tends to infinity leading to @xmath52 , i.e. , projective - anticipating , projective , and projective - lag synchronization of time - delayed chaotic systems can be realized on er networks .
surely , the lyapunov function ( [ lyapunov ] ) satisfies the condition ( [ stability1 ] ) . for the condition ( [ stability2 ] ) , from eqs .
( [ error2 ] ) and ( [ lyapunov ] ) , we get @xmath53 where @xmath54 , @xmath55 . because @xmath39 have zero row - sum matrix , according to the stability theory and the gerschgorin s disk theorem , we can obtain that the sufficient stability condition for the synchronized manifold ( [ manifold ] ) can be written as @xmath56 and @xmath57 . under this sufficient stability and existence conditions ( [ existence ] ) , with the change of magnitude relation between feedback delay time @xmath35 and coupling delay time @xmath36 , we can observe different types of projective synchronization .
let @xmath58 .
if @xmath59 , the synchronization manifold @xmath60 ( [ manifold ] ) can be written as @xmath61 , i.e. , we can observe projective - anticipating synchronization .
if @xmath62 , the synchronization manifold @xmath60 ( [ manifold ] ) can be written as @xmath63 , i.e. , we can observe exact projective synchronization .
when @xmath64 , we can derive @xmath65 , that is to say , projective - lag synchronization can be observed . * remark 1 * the value of scaling factor @xmath0 has no effect on the error dynamics of the system ( eq .
( [ error2 ] ) ) because the values of @xmath66 are independent of the scaling factor @xmath0 .
so we can arbitrarily direct the scaling factor @xmath0 onto any desired value .
in this section , we give two illustrative examples .
* remark 2 * because the whole response dynamical networks realize complete synchronization , we can get the projective - anticipating , projective , and projective - lag synchronization between the drive system and the response dynamical networks with the same scaling factor @xmath0 .
* remark 3 * in the first example and the second one , the er networks consist of @xmath67 nodes . in response dynamical networks systems ,
connectivity probability is @xmath68 .
because the size of er networks is large , it is difficult to get the numerical results between the drive system and the whole response dynamical networks . in simulations ,
we give the numerical results with three arbitrarily selected response systems .
if they can achieve projective - anticipating , projective , and projective - lag synchronization with the drive system @xmath26 , the other response systems can do it with the same scaling factor @xmath0 .
* example 1 . *
consider the nodes of the drive - response dynamical networks are ikeda systems as follows : @xmath69 @xmath70 with @xmath71 and @xmath72 .
here , @xmath24 is the relation coefficient for the drive system @xmath26 and response dynamical networks @xmath33 , @xmath73 is proportional to the power of the incident light , the delay @xmath35 is the time required for light to make a round trip in the cavity .
ikeda model was introduced to investigate the dynamics of an optical bistable resonator , playing an important role in electronical and physiological study and is well known for delay - induced chaotic behavior @xcite . in the following
, we will confirm that the numerical simulations fully support the analytical results presented above .
the coupling strength in response dynamical networks is set as @xmath74 .
if @xmath59 , one can observe the projective - anticipating synchronization for parameters @xmath75 , with the anticipating time @xmath76 . in fig .
[ fig1](a ) , we can observe that the driven systems anticipates the driver , at the same time , the amplitude of @xmath77 and @xmath78 correlates with each other by @xmath79 .
figure [ fig1](b ) shows the time - shifted plot of @xmath77 and @xmath78 .
it is clear from the scales of the coordinate axes that the slope of the line is @xmath80 . in fig .
[ fig1](c ) , we give numerical results of the time series of the error systems @xmath81 and @xmath82 , where @xmath83 indicates the average of nodes of response dynamical networks . from it
we can see that the whole response dynamical networks already realize complete synchronization when projective - anticipating synchronization among drive - response complex dynamical networks can be observed . with @xmath84
, we can observe the exact projective synchronization between eqs .
( [ driver3 ] ) and ( [ responser3 ] ) for parameters @xmath85 . from fig .
[ fig2](a ) , one can find that the phase angle between the synchronized trajectories is zero .
it reduces to the complete synchronization if @xmath86 . from fig .
[ fig2](b ) , we can see the synchronization between @xmath19 and @xmath78 for @xmath87 . for coupling delay @xmath36 being larger than feedback delay @xmath35
, we can observe the projective - lag synchronization for parameters @xmath88 . in fig .
[ fig3](a ) , @xmath89 instead of @xmath19 is convenient for our study .
the response systems lag the state of the drive system with constant lag time @xmath90 and the amplitude of the driver s and driven s state variables is correlated by @xmath91 .
the time - shifted plot of @xmath92 and @xmath93 is shown in fig . [
fig3](b ) .
it is also clear from the scales of the coordinate axes that the slope of the line is @xmath94 .
* example 2 . *
consider the nodes of the drive - response dynamical networks are mackey - glass systems as follows : @xmath95 initially , it was introduced as a model of blood generation for patients with leukemia .
@xmath19 represents the density of circulating cells at time @xmath20 , when it is produced , and @xmath96 is the density when the `` request '' for more blood cells is made .
later , this system became popular in chaos theory as a model for producing high - dimensional chaos to test various methods of chaotic time - series analysis , controlling chaos , etc .
the electronic analog of this system has been proposed by pyragas and his collaborators @xcite .
the response dynamical network systems are written as @xmath97 in our simulation , the coupling strength in response dynamical networks is set as @xmath98 . with @xmath36 being less than @xmath35 , one can observe the projective - anticipating synchronization for parameters @xmath99 , with the anticipating time @xmath100 . in fig .
[ fig4](a ) , we observe that the driven systems anticipates the driver , at the same time , the amplitude of @xmath101 and @xmath33 correlates with each other by @xmath102 .
figure [ fig4](b ) shows the time - shifted plot of @xmath101 and @xmath33 .
it is clear from the scales of the coordinate axes that the slope of the line is @xmath80 . with @xmath62
, one can observe the exact projective synchronization for parameters @xmath103 .
when @xmath104 , we observe the anti - phase synchronization shown in fig .
[ fig5](a ) , synchronized chaotic attractors with an anti - phase pattern where the phase angle between the synchronized trajectories is @xmath105 . from fig .
[ fig5](b ) , we see the anti - phase synchronization between @xmath19 and @xmath78 . with @xmath36 being larger than @xmath35 , one can observe the projective - lag synchronization for parameters @xmath106 . in fig .
[ fig6](a ) , we can observe that in this case the response systems lag the state of the drive system with constant lag time @xmath107 and the amplitude of the driver s and driven s state variables is correlated by @xmath108 .
figure [ fig6](b ) shows the time - shifted plot of @xmath109 and @xmath78 .
to summarize , we have analytically estimated and numerically simulated projective - anticipating , projective , and projective - lag synchronization properties of time - delayed chaotic systems related to optical bistable or hybrid optical bistable device on random dynamical networks .
our work may lead to several advantages over existing work:(1 ) it is capable of realizing projective - anticipating and projective - lag synchronization of a general class of time - delayed chaotic system related to optical bistable or hybrid optical bistable device on random networks , while previous work @xcite only mentioned projective - anticipating and projective - lag synchronization on only two coupled chaotic systems .
( 2 ) although previous work @xcite has achieved projective synchronization on complex dynamical networks , the dynamics of the nodes are coupled partially linear chaotic systems ; we realize projective synchronization of time - delayed chaotic systems on random dynamical networks , without the limitation of the partial - linearity ; so it can be considered as an extension of the dynamics of each individual node from partially linear chaotic systems to non - partially - linear chaotic systems , or an extension from finite - dimension to infinite - dimensional chaotic systems . according to the lyapunov stability theory ,
we have achieved the sufficient stability and necessary conditions for the projective - anticipating , projective , and projective - lag synchronization manifolds .
we derive that the transition among projective - anticipating , projective , and projective - lag synchronizations can be achieved by adjusting the magnitude relation between the feedback delay time and the coupling delay time .
the validity and feasibility of our method have been verified by computer simulations of ikeda and mackey - glass systems .
|
we study projective - anticipating , projective , and projective - lag synchronization of time - delayed chaotic systems on random networks .
we relax some limitations of previous work , where projective - anticipating and projective - lag synchronization can be achieved only on two coupled chaotic systems . in this paper
, we can realize projective - anticipating and projective - lag synchronization on complex dynamical networks composed by a large number of interconnected components . at the same time , although previous work studied projective synchronization on complex dynamical networks , the dynamics of the nodes are coupled partially linear chaotic systems . in this paper , the dynamics of the nodes of the complex networks are time - delayed chaotic systems without the limitation of the partial - linearity . based on the lyapunov stability theory ,
we suggest a generic method to achieve the projective - anticipating , projective , and projective - lag synchronization of time - delayed chaotic systems on random dynamical networks and find both the existence and sufficient stability conditions . the validity of the proposed method is demonstrated and verified by examining specific examples using ikeda and mackey - glass systems on erds - rnyi networks .
+ + _ pacs _ : 05.45.xt , 05.45.jn , 89.75.hc + + _ keywords _ : projective - anticipating synchronization ; projective synchronization ; projective - lag synchronization ; time - delayed chaotic systems ; complex networks * * in 1999 , mainieri and rehacek observed projective synchronization in coupled partially linear chaotic systems where the drive and response vectors synchronize up to a constant ratio @xmath0 ( scaling factor ) .
complete synchronization and anti - phase synchronization are proved to be the special cases of projective synchronization in cases of @xmath1 and @xmath2 , respectively .
this proportional feature can be used to extend binary digital to m - nary digital for achieving fast communication . with the development of research on complex systems , more and more researchers carried out the study about complex dynamical behaviors on networks .
while most studies focused on complete and phase synchronization in various networks , little attention has been paid to projective synchronization . in this paper , based on the lyapunov stability theory , we theoretically analyze both the existence and sufficient stability conditions of the projective - anticipating , projective , and projective - lag synchronization of time - delayed chaotic systems on random networks .
numerical experiments for the ikeda system and the mackey - glass system show that the control method works .
|
an approach to spin - sensitive devices for electronic applications , that may exhibit improved spin injection , is the use of iii - v diluted magnetic semiconductors ( dmss).@xcite these alloys , which incorporate a small percentage of magnetic atoms into the semiconductor host , have been shown to exhibit ferromagnetic behavior up to 110 k.@xcite however , for these materials to find widespread applications , the ferromagnetism should be stable at room temperature .
consequently , much of the current experimental work is directed towards increasing their ferromagnetic transition temperature , @xmath8 . since theory predicts that the transition temperature increases with the magnetic - ion concentration ,
efforts have centered on increasing such concentration.@xcite the solubility of magnetic ions , however , is often quite low .
iii - v dms have been grown using low - temperature molecular beam epitaxy ( lt - mbe ) at temperatures lower than 300 @xmath4c to prevent phase separation .
nevertheless , recently we have demonstrated the growth of single - phase , ferromagnetic in@xmath0mn@xmath1as , with @xmath9 , using metalorganic vapor phase epitaxy ( movpe ) at temperatures as high as 520 @xmath4c.@xcite in this report , we present the results of our investigation of the magnetic properties of these films .
we have observed room - temperature ferromagnetic behavior in a nominally single - phase @xmath10-type in@xmath11mn@xmath12as film .
in contrast , a comparable high @xmath8 has not been observed in single - phase ( in , mn)as grown using lt - mbe .
a possible mechanism for the observed high - temperature ferromagnetism , based on transition metal clustering at the atomic level , will be discussed .
( in , mn)as films were prepared using atmospheric pressure movpe , as described earlier.@xcite films were grown at temperatures between 475520 @xmath4c on semi - insulating gaas(001 ) substrates .
the precursors used were trimethylindium ( tmin ) , tricarbonyl(methylcyclopentadienyl)manganese ( tcmn ) and 0.3% arsine ( ash@xmath13 ) in hydrogen .
phase composition was determined using double - crystal x - ray diffraction ( xrd ) using cu k@xmath14 radiation .
energy dispersive x - ray spectroscopy ( eds ) was used to determine the mn concentration in the films .
figure [ fig1 ] shows the @xmath15 - 2@xmath15 x - ray diffraction pattern with the zinc - blende ( 004 ) and ( 002 ) reflections for a 300 nm thick in@xmath11mn@xmath12as film grown at 520 @xmath4c .
@xmath15 - 2@xmath15 x - ray diffraction scan of in@xmath11mn@xmath12as / gaas(001 ) showing the respective ( 002 ) and ( 004 ) reflections .
no evidence of mnas second phase was observed . ] no other peaks were observed , indicating that the alloy was phase pure to within 0.1 volume percent .
the rocking - curve fwhm for this film was 0.25@xmath4 .
azimuthal ( @xmath16 ) scans for the \{202 } reflections of the film and substrate were recorded to confirm epitaxy .
figure [ fig2 ] shows the four - fold - symmetric diffraction pattern for the \{202 } reflections of the ( in , mn)as film compared with that of the gaas substrate .
azimuthal , @xmath16 , x - ray diffraction scans for the \{202 } reflections of in@xmath11mn@xmath12as and gaas indicating the degree of epitaxy with the gaas substrate .
the angle of the \{202 } diffracting planes with respect to the sample surface is @xmath17@xmath4 . ]
excellent epitaxial alignment is observed despite the @xmath18 lattice mismatch between the film and the substrate .
mnas was observed in films grown at temperatures lower than 475 @xmath4c and for mn concentrations greater than @xmath19 .
growth at temperatures higher than @xmath20 @xmath4c resulted in very little deposition presumably due to increased indium and manganese desorption from the surface .
the magnetic properties of the ( in , mn)as thin films were measured using a squid magnetometer .
the magnetic field was applied perpendicular to the plane of the film along the easy axis of magnetization for ( in , mn)as.@xcite the substrate diamagnetic contribution was subtracted from the total magnetization signal .
the remaining data consisted of the paramagnetic and ferromagnetic contributions from the film .
the total magnetization is given by : @xmath21h\ ] ] where @xmath22 is the ferromagnetic component of the film , @xmath23 is the paramagnetic susceptibility of the film , @xmath24 is the substrate diamagnetic susceptibility , @xmath25 is the temperature , and @xmath26 is the applied magnetic field . the magnetic response of a nominally single - phase in@xmath11mn@xmath12as film is shown in figure [ fig3 ] for applied magnetic fields up to 20 koe and temperatures of 5 , 150 and 300 k. magnetization as a function of applied magnetic field for a in@xmath11mn@xmath12as sample ( ajb070 ) measured at 5 , 150 , and 300 k. magnetic field was applied perpendicular to the plane of the film .
inset is the complete hysteresis loop at 5 k. ] the inset of figure [ fig3 ] shows the complete hysteresis loop for this sample at 5 k. at this temperature the measured saturation magnetization normalized to the film volume was 62 emu/@xmath7 with a remanence of 10 emu/@xmath7 and a coercive field of 400 oe .
when the temperature was increased to 300 k , the saturation magnetization decreased only to 49 emu/@xmath7 .
the magnetization as a function of temperature was measured for the single - phase in@xmath11mn@xmath12as sample discussed above .
the sample was zero - field cooled and then subjected to a 10 koe applied magnetic field perpendicular to the plane of the film .
after subtraction of the diamagnetic substrate contribution , the resulting magnetization is shown in figure [ fig4 ] .
a curie temperature of 333 k was measured for this sample .
such a high curie temperature for a iii - v dilute magnetic semiconductor has been previously attributed to the presence of mnas@xcite which has a curie temperature of 318 k. however , the presence of mnas was not observed by x - ray diffraction in this sample as shown in figure [ fig1 ] .
if a second phase were present , the measured magnetization of 62 emu/@xmath7 would correspond to a mnas volume fraction of approximately 10% .
this should certainly be observable by x - ray diffraction .
if on the other hand the mnas was present as nanoprecipitates , the magnetization would be considerably smaller as observed in ( ga , mn)as films.@xcite thus the magnetization data is consistent with the alloy film being a single - phase .
currently , we are investigating precipitate formation further through transmission electron microscopic analysis .
the value of the saturation magnetization of ( in , mn)as is given by @xmath27 , where @xmath28 is the nominal concentration of mn ions , @xmath29 is the lande factor and is equal to 2 for mn , @xmath30 is the bohr magneton , and @xmath31 is the spin of mn .
the measured @xmath32 using @xmath6 , corresponding to @xmath28=1.8@xmath3310@xmath34 @xmath35 , gives a value of @xmath36 .
this corresponds to a value of @xmath31 between 4/2 and 3/2 .
this is interpreted as an indication of the presence of valence states other than mn@xmath37.@xcite the temperature dependence of the magnetization was compared to the brillouin function for different values of the total angular momentum , @xmath38 .
the lack of agreement between the brillouin function and the measured data above 150 k indicates that the transition to the ferromagnetic state may not be second order , in which the change in magnetization value is continuous with increasing temperature .
it may be first order .
magnetic disordering as a first order phase transformation has previously been reported and a model to describe this behavior was developed by bean and rodbell.@xcite the model predicts that if the exchange interaction is a strong function of the inter - atomic spacing and the lattice is compressible , then the ferromagnetic to paramagnetic transition will be first - order . by introducing a strain energy term into the free energy formulation and minimizing with respect to the reduced magnetization , it was shown@xcite that the temperature dependence of the magnetization for @xmath39 is given by @xmath40 where @xmath41
, @xmath42 is the reduced magnetization ( normalized to the saturation magnetization ) , @xmath43 is the number of magnetic dipoles per unit volume , @xmath44 is the boltzmann constant , @xmath45 is the compressibility , @xmath46 is the slope of the dependence of @xmath8 on volume , and @xmath47 is the curie temperature if the lattice were not compressible . @xmath48 and @xmath25 are the pressure and temperature respectively . for @xmath49 ,
the transition is second order , and equation [ eq2 ] reduces to the brillouin function for @xmath50 . for @xmath51 ,
the transition is first order and the magnetization exhibits a discontinuity at the transition temperature .
the measured magnetization was compared with the model of bean and rodbell for different values of @xmath52 as shown in figure [ fig4 ] .
temperature dependence of the magnetization for the in@xmath11mn@xmath12as sample ( ajb070 ) compared with the model of bean and rodbell for @xmath39 and @xmath53.@xcite magnetic field strength was 10 koe an was applied perpendicular to the plane of the film .
a @xmath54 k and @xmath55 emu/@xmath7 were used for the fitting . ] as can be seen , the best fit occurred for @xmath56 .
this indicates that the phase transformation for the alloy has an intermediate character between first and second order .
a deviation between theory and experiment at temperatures above the curie temperature was also observed .
this was addressed in the initial work by bean and rodbell and is a consequence of the long - range interaction implicit in the model.@xcite molecular field theory which was used in the model , indicates no order above the curie temperature for a normal ferromagnet .
experimentally , however , significant short - range order can be observed . for comparison ,
the ferromagnetic to paramagnetic phase transition of mnas has been shown to be first - order@xcite with a value of @xmath57.@xcite we have measured the temperature dependence of the magnetization for an ( in , mn)as sample which exhibited mnas phase formation .
the magnetization is shown in figure [ fig5 ] as a function of temperature .
temperature dependence of the magnetization for a two phase sample ( in@xmath58mn@xmath59as + mnas ) ( ajb076 ) compared with the model of bean and rodbell for @xmath39 and @xmath53 .
magnetic field strength was 1 koe and was applied perpendicular to the plane of the film .
a @xmath60 k and @xmath61 emu/@xmath7 were used for fitting . ] in this case the temperature dependence appears to be clearly first - order .
fitting of the magnetization data using the first - order model resulted in a value of @xmath62 .
thus it can be seen that the temperature dependence of the magnetization for ( in , mn)as / mnas two phase alloys differs from that of ( in , mn)as single - phase alloys .
the temperature dependent magnetization of in@xmath11mn@xmath12as sample shown in figure [ fig4 ] is closer to second - order and can not simply be explained by the presence of mnas nanoprecipitates . the ferromagnetism observed in iii - v dms films has been previously explained within the framework of sp - d exchange interactions between the fermi sea and the localized magnetic moments.@xcite similar to rkky theory , this hole - mediated ferromagnetism theory predicts that the curie temperature is dependent upon both magnetic ion concentration and hole concentration . using this model dietl et al . predicted
that the transition temperature for ( in , mn)as with a mn concentration of 5% should be 35 k , which is much lower than that observed in the present study .
one possible explanation for the difference is that mn in the alloy is not randomly distributed on in sites but is present as atomic clusters .
this is supported by the recent calculations of van schilfgaarde and mryasov based on local density functional theory.@xcite these calculations predict that due to the strong attractive coupling between the mn ions and the semiconductor cation nuclei , there is a large driving force for the formation of mn clusters consisting of two or more mn atoms located at nearest - neighbor cation sites .
these atomic clusters in turn can stabilize the ferromagnetic state ; according to this model , the presence of free carriers is not required .
although the model does not consider kinetic effects , which may determine whether mn clustering occurs , the growth temperature used for our films is likely sufficient to overcome any kinetic barrier to cluster formation .
indeed clustering of the transition - metal ion has been observed in other iii - v dmss grown at high temperatures@xcite and may also occur in other materials that have been shown to exhibit ferromagnetic behavior while having @xmath63-type conductivity.@xcite presently , we are pursuing extended x - ray absorption fine structure ( exafs ) measurements of these films to determine the local environment of mn .
in conclusion , epitaxial , single - phase in@xmath11mn@xmath12as films have been grown using metalorganic vapor phase epitaxy at temperatures as high as 520 @xmath4c .
temperature- and field - dependent magnetization measurements indicated a single - phase film ( as measured by x - ray diffraction ) to be ferromagnetic with a curie temperature of 333 k. modelling of the magnetization data indicated that the ferromagnetic - to - paramagnetic phase transition was intermediate between first and second order .
the origin of the observed ferromagnetism was discussed in the light of a new theory based on the formation of transition - metal clusters .
extensive use was made of the materials research center facilities at northwestern university .
this work was supported by the national science foundation through the mrsec program under grant number dmr-0076097 .
|
in@xmath0mn@xmath1as diluted magnetic semiconductor ( dms ) thin films have been grown using metalorganic vapor phase epitaxy ( movpe ) .
tricarbonyl(methylcyclopentadienyl)manganese was used as the mn source . nominally single - phase , epitaxial films were achieved with mn content as high as @xmath2 using growth temperatures @xmath3 @xmath4c . for lower growth temperatures and higher mn concentrations , nanometer scale mnas precipitates were detected within the in@xmath0mn@xmath5as matrix .
magnetic properties of the films were investigated using a superconducting quantum interference device ( squid ) magnetometer .
room - temperature ferromagnetic order was observed in a sample with @xmath6 .
magnetization measurements indicated a curie temperature of 333 k and a room - temperature saturation magnetization of 49 emu/@xmath7 .
the remnant magnetization and the coercive field were small , with values of 10 emu/@xmath7 and 400 oe , respectively .
a mechanism for this high - temperature ferromagnetism is discussed in light of the recent theory based on the formation of small clusters of a few magnetic atoms .
|
the high - proper motion star nltt 16249 is a double degenerate system showing hydrogen lines and molecular carbon and cyanogen bands in its optical spectrum @xcite .
although large radial velocity variations were noted by @xcite , the orbital parameters , period and separation , are yet to be determined .
the detection of photospheric nitrogen in the carbon - rich ( dq ) component of this system was a first occurrence for this class of objects and , according to the dredge - up scenario commonly applied to dq white dwarfs @xcite , it implied the presence of nitrogen in the white dwarf core .
the other component is a hydrogen - rich ( da ) white dwarf with a mass above average but with a luminosity similar to that of the dq white dwarf .
the two characteristics of the system , a nitrogen - enriched dq component and a likely close orbit , may or may not be related .
@xcite established a clear evolutionary link between born - again stars and a nitrogen / oxygen enrichment in pg1159 stars and their dq descendents .
members of the pg1159 class have a helium - rich surface with notable enrichment in carbon , nitrogen and oxygen @xcite .
@xcite argued that not all dq stars followed the born - again path because the carbon abundance is lower in some of these objects than would be expected following this evolutionary path .
however , the @xmath4 versus trend established by @xcite and @xcite does appear to follow the model predictions of @xcite .
moreover , @xcite noted two discernible tracks with one at a higher carbon abundance that was attributed to a thinning of the helium layer .
therefore , data and born - again models may be compatible for the bulk of dq white dwarfs . however , nitrogen is not detected in any dq stars other than in the dq in nltt 16249 . in this letter , we present convincing evidence that the components of the binary nltt 16249 are in a close orbit . our new radial velocity measurements
( section 2 ) and our original model atmosphere analysis @xcite help constrain the component properties and offer clues to the origin and evolutionary prospect of the system ( section 3 ) .
we summarize and discuss these new results in section 4 .
[ cols="^,^,^,^,^,^,^,^,^ " , ] + the cooling age of the da white dwarf is estimated at @xmath5 gyr .
a high mass for both white dwarfs implies relatively massive , @xmath6 ( see a discussion in * ? ? ?
* ) , hence short - lived progenitors ( @xmath7 gyr , * ? ? ?
* ) , constraining the total age of the system to less than 3 gyr . using our measured period ( 1.17 d ) and mass ratio @xmath8 , and constraining the da mass to @xmath9 , the merger time
is estimated at @xmath10 yr following @xcite .
therefore , the system is not a candidate type 1a supernova progenitor .
the systemic velocity was obtained by subtracting the calculated gravitational redshift ( @xmath11 ) from the mean velocity of the da white dwarf ( @xmath12 ) , @xmath13 . adopting a distance of @xmath14 pc , the systemic velocity and proper - motion measurements ( @xmath15 mas yr@xmath16 , * ? ? ?
* ) correspond to the galactic velocity components @xmath17 following @xcite .
the system belongs to the thin - disk population @xcite in agreement with its young age ( @xmath18 gyr ) .
we have shown that the peculiar dq white dwarf in nltt 16249 is in a close 1.17 d orbit with a da white dwarf companion . close
double - degenerate stars are common and @xcite reported a fraction of 5 to 19% of close pairs in their radial velocity survey .
white dwarfs with composite spectra indicative of da plus db ( hei ) or da plus dc ( cool he - rich ) pairs are also relatively common in large surveys ( see , e.g. , * ? ? ?
* ; * ? ? ?
however , only two da+dq pairs are known : nltt 16249 and sdss j153210.04 + 135615.0 ( nltt 40489 ) which was recently identified by @xcite .
the velocity amplitudes in nltt 16249 imply that the dq is slightly more massive than its companion in which case it is more likely that the dq formed first , both from @xmath19 progenitors , unless stable mass transfer reversed the initial mass ratio ( for detailed scenarios see * ? ? ?
* ) . the relatively long orbital period reported here rules out the prospect of a merger within a hubble time .
the dq white dwarf is peculiar and the nitrogen concentration in its atmosphere is , so far , unique .
@xcite followed the evolution of a 2.7@xmath1 main - sequence star past the pg1159 spectroscopic stage and down the cooling track up to the dq stage ( 0.6@xmath1 ) .
the star follows a `` born - again '' he - flash loop that resets its evolution onto the post - asymptotic giant - branch ( post - agb ) , considerably modifying the chemical structure of the outer layers that an otherwise normal post - agb would possess . in particular , @xcite predicted measurable concentrations of nitrogen and oxygen in the atmosphere of dq descendents of born - again stars .
the nitrogen abundance measured in the dq component of nltt 16249 is qualitatively similar to the abundance predicted for the 10,500 k model , the lowest temperature considered by @xcite .
however , the @xmath20 ratio for that particular model is close to 300 , i.e , much larger than measured in nltt 16249 ( @xmath21 ) . in the born - again context
, the difference could be attributed to a different progenitor mass which could affect the outcome of the chemical evolution , although such models are not available to us .
@xcite also predicted a concentration of oxygen between that of carbon and nitrogen ( @xmath22 ) .
it is possible to constrain the @xmath23 abundance ratio in dq white dwarfs using the near and far ultraviolet co fourth - positive " ( 4p ) bands along with cn violet and c@xmath3 swan bands . finally , the @xmath24 isotopic ratio expected for the same evolutionary track is well below detection limits reported by @xcite in the case of the dq in nltt 16249
. what would be the role , if any , played by binary interactions in shaping the chemical structure of the dq progenitor in nltt 16249 ?
the dq white dwarfs are not known to show nitrogen in their spectra ( see * ? ? ?
* ) . therefore , it may not be a coincidence that the only known detection of nitrogen in a dq white dwarf is also that of a white dwarf in a close double - degenerate system . in that context
as well , it remains to be shown that detectable traces of nitrogen would be the expected outcome of mass - transfer or early envelope ejection .
s.v . and a.k .
acknowledge support from the grant agency of the czech republic ( ga r p209/10/0967 , ga r p209/12/0217 ) .
this work was also supported by the project rvo:67985815 in the czech republic .
we thank donna burton for invaluable help during the december observations at siding spring observatory .
j.r.t . gratefully acknowledges support from nsf grants ast-0708810 and ast-1008217 .
althaus , l. g. , serenelli , a. m. , panei , j. a. , crsico , a. h. , garca - berro , e. , scccola , c. g. , 2005 a&a , 435 , 631 benvenuto , o. g. , althaus , l. g. 1999 , mnras , 303 , 30 dopita , m. , hart , j , mcgregor , p. , oates , p. , jones , d. 2007 , ap&ss , 310 , 255 dreizler , s. , heber , u. 1998 , a&a , 334 , 618 dufour , p. , bergeron , p. , fontaine , g. 2005 , apj , 627 , 404 ferrario , l. , wickramasinghe , d. , liebert , j. , williams , k. a. 2005 , mnras , 361 , 1131 giammichele , n. , bergeron , p. , dufour , p. 2012 , apjs , 199 , 29 johnson , d. r. h. , soderblom , d. r. 1987 , aj , 93 , 864 koester , d. , knist , s. 2006 , a&a , 454 , 951 limoges , m .- m . , bergeron , p. 2010 , apj , 714 , 1037 macdonald , j. , hernanz , m. , jose , j. 1998 , mnras , 296 , 523 maxted , p. f. l. , marsh , t. r. 1999 , mnras , 307 , 122 nelemans , g. , yungelson , l. r. , portegies zwart , s. f. , verbunt , f. 2001 , a&a , 365 , 491 pauli , e .- m . ,
napiwotzki , r. , heber , u. , altmann , m. , odenkirchen , m. 2006 , a&a , 447 , 173 pelletier , c. , fontaine , g. , wesemael , f. , michaud , g. , wegner , g. 1986 , apj , 307 , 242 ritter , h. 1986 , a&a , 169 , 139 salim , s. , gould , a. 2003 , apj , 582 , 1011 schaller , g. , schaerer , d. , meynet , g. , maeder , a. 1992 , a&as , 96 , 269 tremblay , p .- e . , bergeron , p. , gianninas , a. 2011 , apj , 730 , 128 vennes , s. , & kawka , a. 2012 , apjl , 745 , l12 werner , k. , heber , u. , hunger , k. 1991 , a&a , 244 , 437 werner , k. , & herwig , f. 2006 , , 118 , 183
|
new spectroscopic observations show that the double degenerate system nltt 16249 is in a close orbit ( @xmath0 ) with a period of 1.17 d. the total mass of the system is estimated between 1.47 and 2.04@xmath1 but it is not expected to merge within a hubble time - scale ( @xmath2 yr ) .
vennes & kawka ( 2012 , apj , 745 , l12 ) originally identified the system because of the peculiar composite hydrogen ( da class ) and molecular ( c@xmath3dq class and cn ) spectra and the new observations establish this system as the first da plus dq close double degenerate . also , the dq component was the first of its class to show nitrogen dredged - up from the core in its atmosphere
. the star may be viewed as the first known dq descendant of the born - again pg1159 stars .
alternatively , the presence of nitrogen may be the result of past interactions and truncated evolution in a close binary system .
|
experimental observations mainly of dynamics of spiral galaxies and galaxy clusters indicate the existence of dark matter ( dm ) .
cosmological observations confirm the existence of dm and in addition show that the bulk of it must be non - baryonic @xcite . in this paper
we consider weakly interacting massive particles ( wimps ) as candidates for dm , but we adopt a model independent and phenomenological approach .
the mass and cross section of the dm ( which is expected to have a local density of about 0.3 gev/@xmath3 and velocity w.r.t the earth of about 200 km / sec @xcite ) is probed by direct detection experiments like xenon @xcite , cdms @xcite , dama @xcite and cogent @xcite .
these experiments detect dm scattering off nuclei by measuring the recoil energy of the nuclei .
the energy threshold of such detectors is typically of the order of a few kev . of these experiments ,
the dama experiment observed an annual modulation in its signal which could have been due to dm scattering .
however none of the other experiments conducting direct dm searches have seen evidence of such an event .
theoretically if one considers spin - independent interaction @xcite of these wimps with the nuclei then it is found that only in the low mass range from about 5 to 10 gev one can reconcile dama signal with the null results from other experiments .
there have been proposals of inelastic scattering of dm @xcite by nuclei which results in heavier dm mass being allowed for explaining all the existing data .
although dm has zero electric charge it may couple to photons through loops in the form of electric and magnetic dipole moments .
here we study wimps which are endowed with such dipole moments and thus can interact feebly via electromagnetic interaction @xcite .
we first will consider the case that dm is a dirac fermion .
the effective lagrangian for coupling of a dirac fermion @xmath4 having an electric dipole moment @xmath5 and a magnetic dipole moment @xmath6 to a electromagnetic field @xmath7 is @xmath8 in this case , we have elastic scattering of dm by nuclei through a photon exchange ( fig.[feyn ] ) .
this is studied in detail for the case of electric dipole moment in section [ edm ] and of magnetic dipole moment in section [ mdm ] .
next we consider the case of dm being majorana fermions . these have only non - zero transition moments between different mass eigenstates .
their interaction with photons is described by @xmath9 where @xmath10 is the transition magnetic moment and @xmath11 is the transition electric moment .
majorana dm would have inelastic scattering off nuclei through a photon exchange .
we study the case of transition electric dipole moment in section [ inel ] and of transition magnetic dipole moment in section [ mdm ] . in section [ relic ]
we calculate the bounds on the dipole moments which would give the desired relic density abundance of dm .
in this paper we present two main types of results : 1 .
we find bounds on dipolar moments , direct and transition , coming from wimp search experiments , which update the results in @xcite .
in addition we show that there are bounds on dipole moments from single photon search at lep , an effect not discussed in @xcite .
we find regions in parameter space that are consistent with the positive signal from dama and with the null results from other experiments .
all these results are presented in section [ results ] . in appendix
a we give some details of our dm scattering calculations , and in appendix b we show the dm annihilation cross section calculations which are relevant for the relic density results . in appendix c
we review the contact scalar interaction case , which is useful for comparison with our results .
the differential cross section for a dm - proton elastic scattering via interaction of the electric dipole moment of dm with the proton charge is given by @xmath12 where @xmath13 is defined in ( [ l_elastic ] ) . here
@xmath14 is the recoil energy of proton and @xmath15 is the speed of dm relative to the nucleus .
the differential rate for nuclear scattering of dm through electric dipole moment interaction for a nucleus of @xmath16 protons and dm of mass @xmath17 is given by @xmath18 where @xmath19 is the number of target nuclei in the detector and @xmath20 is the local dm density .
the velocity distribution @xmath21 and minimum speed of dm @xmath22 , for a given energy threshold @xmath23 are given by @xcite @xmath24 where @xmath25 is the mass of the target nucleus , @xmath26 is the reduced mass of the dm - nucleus system and @xmath27 km / sec .
the factor @xmath28 appearing in eqn.([dr ] ) shows the fact that this is a coherent scattering of the @xmath16 protons in the target nucleus . to determine the expected event rate @xmath29 we integrate eqn.([dr ] ) over the nuclear recoil energy : @xmath30 here @xmath31 is the efficiency of the detector which depends on the recoil energy and @xmath32 is the quenching factor that relates the observed energy with the actual recoil energy i.e @xmath33 .
@xmath32 depends on the target nucleus and the nature of the detector . @xmath34 and @xmath35 are limits of the observed energy interval in a direct search experiment
. therefore @xmath36 and @xmath37 give the corresponding limits for the recoil energy @xmath14 .
observed energies are usually quoted in units of electron equivalent energies ( kevee ) .
the detectors usually detect scintillation caused by electrons and thereby measure energy transferred to electrons by the nucleus .
@xmath32 then represents the efficiency with which the recoil energy of nucleus is transferred to electrons which are detected by scintillation .
the quenching factors for the different experiments analyzed here are given in table [ t ] .
c c c + experiment & target nucleus & quenching + & & factor(@xmath32 ) + + dama & na & 0.3 + & i & 0.09 + cdms & ge & 1 + xenon & xe & 1 + cogent&ge & 0.2 + in fig.[edm - plot ] we show our results .
the experiments give an upper limit which is a function of the dm mass @xmath17 , except for dama whose result is represented as an allowed band .
we would like to point out that one can get a bound from the @xmath38 process whose signal is a single photon detection with missing energy . from the analysis of the collaboration
l3 @xcite we get @xmath39 e - cm , valid for @xmath40 gev ; this limit is also shown in fig.[edm - plot ] .
we shall see the implications of this figure in section [ results ] ; there we shall also discuss the issue of the dm relic density .
in this section we consider the case that the wimp is a majorana particle , with a a transition electric dipole moment @xmath41 as defined in ( [ l_inelastic ] ) . in this case
we can have inelastic scattering @xmath42 where @xmath43 and @xmath44 are two different mass eigenstates , and in general there is a mass difference between @xmath43 and @xmath44 , @xmath45 . due to this mass difference ,
the minimum dm kinetic energy needed for the nucleon scattering becomes higher .
dama has lower detection threshold compared to most experiments and therefore can be more sensitive to this scattering mode than the other experiments .
this was first proposed in @xcite . for inelastic scattering for different experiments .
shaded region shows the allowed parameter space for dama which is consistent with all other experiments and in case of @xmath46 gev it is also consistent with the relic density.,title="fig:",width=283,height=283 ] for inelastic scattering for different experiments .
shaded region shows the allowed parameter space for dama which is consistent with all other experiments and in case of @xmath46 gev it is also consistent with the relic density.,title="fig:",width=283,height=283 ] for inelastic scattering for different experiments .
shaded region shows the allowed parameter space for dama which is consistent with all other experiments and in case of @xmath46 gev it is also consistent with the relic density.,title="fig:",width=340,height=340 ] for a fixed @xmath5 which gives the correct relic density of dm.,width=340,height=340 ] if we drop terms of higher order in @xmath47 ( compared to @xmath17 ) we end up with eqn.([dromega ] ) for the differential cross section .
that is , @xmath48 however the total cross section in the elastic and inelastic case are related to each other by a factor which enforces the condition of the minimum kinetic energy required for inelastic scattering .
this relation has been derived at the end of appendix a , and is given by , @xmath49 but we do not use the total cross section in our analysis , only the differential cross section and therefore the square root factor does not figure in our calculations .
we integrate the differential cross section per unit recoil energy over the energy interval in which the experimental data have been observed .
the only change from the analysis in the elastic case is that the minimum speed for scattering is now given by @xmath50 where notation has the same meaning as in the previous section . in fig.[edm - inel ]
we plot our results .
we show a line for each direct search experiment and the allowed range is on the right of the line , except for dama that again is an allowed band . in the figure we also plot the limit coming from lep
mentioned before , @xmath51 e - cm , which is the same than in the direct transition case if @xmath52 .
the collider limit applies provided @xmath40 gev . since in the inelastic case we have three parameters , namely , @xmath53 , and @xmath47 , it is instructive to present fig.[edm - inel - fix ] where we fix @xmath41 and show the constraint in the ( @xmath54 ) plane . here
the allowed regions is on the right of the curves of the different experiments except for dama which is again an allowed band .
we shall discuss the implications of all these limits in section [ results ] and especially for the relic density in section [ relic ] .
the differential cross section per unit energy transfer for elastic scattering by a magnetic dipole moment interaction is given by @xmath55 in the case of inelastic scattering we get the same formula with @xmath6 substituted by the transition magnetic moment @xmath10 , provided we drop terms of higher order in the mass difference @xmath47 .
again , the only difference between elastic and inelastic is in the kinematics , because the minimum speed for inelastic scattering is given by eqn.([vmin ] ) .
fig.[mu ] and fig.[mu - inel ] show the results for magnetic dipole moment which are essentially the same as those for electric dipole moment in terms of the allowed parameter space from various experiments . in fig.[mu - inel - fix ] we plot the allowed regions in the space ( @xmath17 , @xmath10 ) . in the next section
we discuss the implications of dipolar dm vis a vis the dm relic density . for inelastic scattering for different experiments .
shaded region shows the allowed parameter space for dama which is consistent with all other experiments and in case of @xmath46 gev it is also consistent with the relic density.,title="fig:",width=283,height=283 ] for inelastic scattering for different experiments .
shaded region shows the allowed parameter space for dama which is consistent with all other experiments and in case of @xmath46 gev it is also consistent with the relic density.,title="fig:",width=283,height=283 ] for inelastic scattering for different experiments .
shaded region shows the allowed parameter space for dama which is consistent with all other experiments and in case of @xmath46 gev it is also consistent with the relic density.,title="fig:",width=340,height=340 ] for a fixed @xmath6 which gives the correct relic density of dm.,width=340,height=340 ]
the relic abundance of dipolar dm is determined by the annihilation cross section @xmath56 through a dipole vertex . in the case of majorana fermions
only non - identical fermions can annihilate ( @xmath57 ) through the dipole channel as only the transition dipole moments ( @xmath58 ) are non - zero .
when the mass difference @xmath47 between @xmath43 and @xmath44 is small the cross section for majorana annihilation process is identical to that of the dirac fermions ( with @xmath59 replaced with @xmath60 ) .
the annihilation cross section for electric dipole annihilation of dm for @xmath56 is given by ( see appendix b ) @xmath61\\ \nonumber\ ] ] and the expression for the magnetic dipole annhilation case is @xmath62\\ \nonumber \label{ann - mag}\end{aligned}\ ] ] where @xmath63 is the relative velocity of the two annihilating wimps .
the thermal averaged cross section , @xmath64 , can be parameterized in terms of the temperature , @xmath65 , where @xmath66 , such that @xmath67 where @xmath68 and the parameter @xmath69 for magnetic and electric dipole interactions take the following values @xmath70 considering cold dark matter ( cdm ) we want to estimate the relic density in the context of dipolar interactions of such cdm candidates .
the expression for @xmath71 , where the subscript @xmath72 indicates the freeze out condition , is given up to a reasonable approximation by @xcite @xmath73\\ & -\left(n+\frac{1}{2}\right)\ln\lbrace\ln[0.038(n+1 ) ( g / g_{*}^{1/2})m_{pl}m_\chi \sigma_0]\rbrace \end{split}\ ] ] where @xmath74 is planck mass , while @xmath75 and @xmath76 are the effective number of relativistic particles at the time of decoupling .
the dm relic density is then given by @xcite @xmath77 the limit on cold dark matter density from wmap @xcite is @xmath78 . using ( [ omegah2 ] ) , this gives us the bound on @xmath79 , which gives the acceptable relic density of dark matter , to be @xmath80 e - cm .
for the magnetic dipole case again using ( [ omegah2 ] ) we find that the dipole moment must be @xmath81 . for the case of non - identical majorana annihilation
the annihilation cross sections are identical to the case of dirac fermions and the limits on the dirac electric and magnetic moments from relic density abundance are identical to the bounds on transition magnetic moments given above .
we have explored the parameter space of electric and magnetic dipole moments of dm from the results of nuclear recoil experiments .
we find that the limit on direct electric dipole moment is @xmath82 e - cm if the dm mass @xmath83 gev , see fig.[edm - plot ] . in this mass range
there is no allowed parameter space which is consistent with the positive signal from dama and with the null results of other experiments . for a dm mass @xmath84 gev and a moment @xmath85 e - cm
there is a region in parameter space where the dama signal is not ruled out by other experiments , see fig.[edm - plot ] . for smaller masses , @xmath86 gev
, the figure shows that cogent rules out dama . in appendix c
we review the scalar interaction mediated scattering ( see fig.[sigma ] ) where we see that dama is allowed by cogent for any dm mass below about 15 gev .
this can be attributed to the fact that the dm interaction via electric dipole moment has a factor of recoil energy ( @xmath14 in the denominator of its cross section in eqn.([dr ] ) .
this means that for low threshold this cross section is large and in the case of cogent this threshold happens to be smaller than that for dama .
but as the dm mass increases this advantage in favour of cogent is nullified by the dm mass factor which again appears in the denominator of the cross section formulae .
the limit on direct magnetic dipole moment is @xmath87 if the dm mass @xmath88 gev , see fig.[mu ] . in this mass range
there is no allowed parameter space which is consistent with a positive signal from dama with null results of other experiments . for dm masses
@xmath89 gev and @xmath90 there is a region in parameter space where the dama signal is not ruled out by other experiments , see fig.[mu ] .
actually , in the low dm mass range the collider bound @xcite is more restrictive than the direct wimp search limits , as can be see in fig.[mu ] .
if the only interaction of dm with standard model particles had been via electromagnetic dipole interaction then the relic density of the dm particles would be determined uniquely from the annihilation cross section of the em - dipole mediated process @xmath91 and @xmath17 .
such a calculation shows @xcite that to get relic cold dm density consistent with wmap measurement the dipole moments have to be in the range of @xmath92 e - cm or @xmath93 when @xmath94 gev . the limits on dm dipole moments that we obtain on the basis of nuclear scattering experiments are much lower .
this means that the dm particles must have some other more dominant interaction which decouples at a temperature @xmath95 to give the correct relic abundance while the dipole interactions decouple much earlier and have no bearing on the present dm density .
these other dominant interactions however must not dominate over dipole interactions in nuclear scattering in order for our bounds on dipole moments to be meaningful .
a good example of such a situation is seen in certain classes of wimp models @xcite , where the interaction with standard model particles can be by higgs exchange which gives rise to spin - independent ( si ) interactions with nucleons or by @xmath16 exchange which gives rise to spin - dependent ( sd ) interactions @xcite .
when the nuclear scattering is by sd interaction then there is no coherent enhancement of the cross section by the atomic number .
the bounds on the cross section for si interactions is therefore more stringent than the bounds on sd interactions .
the bounds for si interactions from nuclear scattering experiments is @xmath96 cm@xmath2 and for sd interactions it is @xmath98 cm@xmath2 ( for @xmath99 gev ) .
our bounds are relevant for wimps which can have a large spin - dependent cross section @xmath100 cm@xmath2 to give the correct relic density , but this cross section is too small to be observed in nuclear scattering experiments . in figs.[edm - inel ] and [ mu - inel ] we plot the constraints on the transition dipole moments @xmath101 and @xmath102 , respectively , versus the dm mass @xmath17 .
we have fixed three values for the mass @xmath103 and 100 gev , in order to illustrate our findings . we see that a common feature of the plots is that the region where dama is consistent with other experiments is quite large and extends to higher and higher inelasticities .
however for @xmath104 gev the region is further constrained by the collider bound from l3 , and the allowed region is reduced .
the allowed area is more and more reduced as we lower @xmath17 and for @xmath105 gev there is no region at all .
for @xmath106 gev there is no collider bound and the region is large .
the most interesting aspect of the experimentally allowed parameter space for transition dipole moments is that the cosmologically preferred value ( for getting the correct relic abundance ) @xmath107 e - cm or @xmath108 @xcite is consistent with the allowed values dama and other experiments when @xmath109 gev and the mass split @xmath110 kev .
thus , we find that transition dipole scattering has a large parameter space where the results of dama , the null results from other experiments and cosmological relic abundance are all consistent .
the event rates depend upon @xmath111 and @xmath79 or @xmath6 as @xmath112 or @xmath113 respectively .
the dipole moments can increase by four orders of magnitude from the @xmath114 e - cm to @xmath115 e - cm ( and similarly for the case of magnetic moments ) and the same event rates would be obtained if mass is increased by eight orders of magnitude from @xmath116 gev to @xmath117 gev .
the range of dm mass for which dama is consistent with cosmology and l3 is therefore quite large , @xmath118 .
this parameter space of @xmath17 and electric /magnetic moments can give rise to a missing energy signal via the process @xmath119hadronic jets , and the dipolar model of dark matter may be testable at the lhc .
e.m . would like to thank alessio provenza for useful comments .
e.m . acknowledges support by the cicyt research project fpa 2008 - 01430 and the departament duniversitats , recerca i societat de la informaci ( dursi ) , project 2005sgr00916 and partly by the european union through the marie curie research and training network `` universenet '' ( mrtn - ct-2006 - 035863 ) .
here we consider the elastic scattering process @xmath120 . we calculate the cross section in the lab frame where the proton is at rest initially .
we assume that even after the scattering the proton remains at rest since the momentum transferred is very small in this process .
the entire calculation is done in the non - relativistic limit .
* electric dipole moment * the initial momenta of dm and proton are denoted by @xmath121 and @xmath122 respectively , while the final momenta are denoted by @xmath123 and @xmath124 respectively .
the amplitude squared for this process is given by @xmath125\,\times tr[(\slashi{p_f}+m_p)\gamma_\mu(\slashi{p_i}+m_p)\gamma_\nu]\ ] ] @xmath126\ ] ] where @xmath127 is the momentum transferred , while @xmath17 and @xmath128 are masses of dm and proton respectively . in writing the above equation we have made use of @xmath129 and @xmath130 . here , @xmath131 and @xmath132 .
hence we can drop terms containing @xmath133 and its higher orders , keeping just the last term in eqn.([ampel ] ) .
this gives us @xmath134 using @xmath135 we finally arrive at the following expression for the amplitude squared , @xmath136 now , the differential cross section is given by @xmath137 where @xmath34 and @xmath35 are energies of the initial and final dm states .
also the recoil energy can be written as follows : @xmath138 therefore , @xmath139 and we know that @xmath140 , so we can write @xmath141 using this in eqn.([ds ] ) we get @xmath142 we use the approximations , @xmath143 and @xmath144 , which gives us @xmath145 substituting for @xmath146 from eqn.([ampel1 ] ) we get @xmath147 which is same as eqn.([dromega ] ) .
* magnetic dipole moment * the trace in this case is same as in the case of edm but without any @xmath148 in it .
the amplitude squared for this process is given by @xmath149\,\times tr[(\slashi{p_f}+m_p)\gamma_\mu(\slashi{p_i}+m_p)\gamma_\nu]\ ] ] @xmath150\\ & - \frac{4 e^2 \mu^2}{q^4 } \left[4(q\cdot p_i)^2 m_\chi^2\right ] .
\end{split}\ ] ] here we drop terms which are higher order in @xmath151 except the last term which is enhanced by a factor of @xmath152 in the denominator .
thus we get @xmath153\ ] ] where @xmath154 using these results in eqn.([m2 ] ) we get @xmath155.\ ] ] again using @xmath156 gives @xmath157\\ & = \frac{8 e^2 \mu^2}{e_r } \left[m_\chi^2 m_p v^2 + \frac{m_\chi^2 e_r}{2 } - m_\chi m_p e_r \right].\end{aligned}\ ] ] the differential cross section per unit recoil energy is given by eqn.([ds1 ] ) . substituting from the above equation in eqn.([ds1 ] )
we get @xmath158\\ & = \frac{e^2\mu^2}{4\pi e_r } \left[1+\frac{e_r}{2\mu_p v^2}-\frac{3e_r}{2m_\chi v^2}\right]\end{aligned}\ ] ] where @xmath159 is the reduced mass of the dm - proton system .
the last term in the above equation can be dropped for large values of dm mass .
and so we get the following expression for the differential cross section per unit energy , @xmath160.\ ] ] * total cross section for inelastic scattering * in the case of inelastic scattering we consider the wimps as majorana particles . as mentioned before in the main section that the differential scattering cross section in both elastic and inelastic scattering is the same . but
the total cross sections in the two cases are related to each other as in eqn.([el - inel ] ) .
now for inelastic scattering we have , @xmath161 the magnitude of the momentum transferred ( working in cm frame ) is then , @xmath162 where we have used @xmath163 . and so , @xmath164 but @xmath165 , therefore @xmath166 the above equation in the elastic case becomes @xmath167 now the total cross section for elastic scattering is given by , @xmath168 for the inelastic case we have , @xmath169
the process involved here is @xmath170 , @xmath72 being a fermion of charge @xmath171 with negligible mass .
now we work in the cm frame where , @xmath172 the amplitude squared for this process is given by @xmath173 [ u(p_1)\bar u(p_1)\gamma^\mu v(p_2)\bar v(p_2)\gamma^\nu]\ ] ] where @xmath174 and @xmath175 for magnetic dipole interaction , whereas for electric dipole interaction @xmath176 and @xmath177 . for majorana particles , @xmath178\\ \ ] ] in the majorana representation of @xmath179-matrices , @xmath180 and the majorana condition implies that @xmath181 . imposing @xmath182 on ( [ majorana ] )
we get the relation @xmath183 . using this relation
we can rewrite eqn.([m2i ] ) as @xmath184 [ u(p_1)\bar u(p_1)\gamma^\mu v(p_2)\bar v(p_2)\gamma^\nu].\end{aligned}\ ] ] now the spin sum rules for majorana fermions are given by @xcite @xmath185 where @xmath186 .
we then take the sum over spins of eqn.([m2maj ] ) and use eqns.([srule1 ] ) and ( [ srule2 ] ) to obtain @xmath187\ , tr[\slashi{p_1}\gamma^\mu\slashi{p_2}\gamma^\nu]\ ] ] where @xmath188 and @xmath189 are masses of @xmath43 and @xmath44 respectively .
now we discuss the two cases , beginning with the electric dipole case first . * electric dipole interaction * eqn.([majav ] ) with @xmath190 and @xmath191 yields the following expression for the averaged amplitude squared , @xmath192\ ] ] where @xmath193 . now in cm frame , @xmath194 using these in eqn.([mjed ] ) we get @xmath195\\\label{mjed2 } & = 4e^2\mathcal d^2 e_1 e_2\left[\left(1-\frac{m_1 m_2}{e_1 e_2}\right ) + \left(1-\frac{(s - m_1 ^ 2-m_1 ^ 2)}{2e_1 e_2}\right)\cos^2\theta\right]\end{aligned}\ ] ] where we use @xmath196 with @xmath197 .
the differential cross section in the cm frame is given by @xmath198\end{aligned}\ ] ] where @xmath199 .
integrating above equation over @xmath200 gives the total cross section as @xmath201.\ ] ] eqn.([ann - elec ] ) can be obtained by taking @xmath202 and @xmath203 in the above equation .
* magnetic dipole interaction * for the magnetic dipole case , with @xmath204 and @xmath205 in eqn.([majav ] ) , the averaged amplitude squared reads @xmath206\ ] ] and this equation in the cm frame looks like @xmath207.\ ] ] the total cross section is then given by @xmath208.\ ] ] one can obtain eqn.([ann - mag ] ) by taking @xmath202 and @xmath203 in the above formula .
here we consider scattering of dm and nucleus through scalar mediated spin - independent interaction .
the details of this calculation are given in @xcite . here
we shall give only the results of the calculation .
the differential recoil rate per unit detector mass for dm - proton scattering can be written as @xmath209 where @xmath210 is the nucleus recoil momentum , @xmath211 is the dm - proton cross - section , @xmath212 is the number of nucleons and @xmath213 is the reduced mass of the dm - proton system . here
@xmath214 is the woods - saxon form factor given by @xmath215 where @xmath216 is the momentum transferred , @xmath25 being the nucleus mass , @xmath217 fm and @xmath218 with @xmath219 fm .
the spin - independent differential cross section for contact interaction scattering is given by @xmath220 ^ 2\\\label{sig0 } & \sigma_0=\int_0^{4\mu_n^2 v^2}\quad{\frac{d\sigma(q=0)}{d|\mathbf{q}|^2}d|\mathbf{q}|^2 } = \frac{4\mu_n^2}{\pi}\left[zf_p+(a - z)f_n\right]^2\end{aligned}\ ] ] here @xmath221 is the magnitude square of the momentum tranferred , @xmath222 is the reduced mass of the dm - nucleus system , @xmath16 is the number of protons and @xmath212 is the number of nucleons . taking @xmath223 we can write the dm - nucleus cross section in terms of the dm - proton cross section@xcite using eqn.([sig0 ] ) as @xmath224 where @xmath213 is the reduced mass of the proton - dm system . substituting from eqn.([sigp ] ) in eqn.([dsig1 ] ) we have @xmath225
the differential rate per unit detector mass can be written as @xmath226 where the quantity @xmath72 is as defined in the main section . substituting from eqn.([dsig2 ] ) in the above equation we get @xmath227 d. tucker - smith and n. weiner , phys .
d * 64 * , 043502 ( 2001 ) [ arxiv : hep - ph/0101138 ] ; d. tucker - smith and n. weiner , phys .
d * 72 * , 063509 ( 2005 ) [ arxiv : hep - ph/0402065 ] ; s. chang , g. d. kribs , d. tucker - smith and n. weiner , arxiv:0807.2250 [ hep - ph ] .
e. komatsu _ et al .
_ [ wmap collaboration ] , astrophys .
j.suppl . * 180 * , 330 ( 2009 ) [ arxiv:0803.0547 [ astro - ph ] ] .
v. barger , w. y. keung and g. shaughnessy , phys .
d * 78 * , 056007 ( 2008 ) [ arxiv:0806.1962 [ hep - ph ] ] .
e. w. kolb and m. s. turner , front .
phys . * 69 * , 1 ( 1990 ) .
|
if dark matter ( dm ) has non - zero direct or transition , electric or magnetic dipole moment then it can scatter nucleons electromagnetically in direct detection experiments . using the results from experiments like xenon , cdms , dama and cogent we put bounds on the electric and magnetic dipole moments of dm .
if dm consists of dirac fermions with direct dipole moments , then dm of mass less than 10 gev is consistent with the dama signal and with null results of other experiments .
if on the other hand dm consists of majorana fermions then they can have only non - zero transition moments between different mass eigenstates .
we find that majorana fermions with mass @xmath0 gev and mass splitting of the order of ( 50 - 200 ) kev can explain the dama signal and the null observations from other experiments and in addition give the observed relic density of dm by dipole - mediated annihilation .
this parameter space for the mass and for dipole moments is allowed by limits from l3 but may have observable signals at lhc .
2.5 cm + 0.5 cm @xmath1_grup de fsica terica and institut de fsica daltes energies + universitat autnoma de barcelona , 08193 bellaterra , spain _ + @xmath2_physical research laboratory ,
ahmedabad 380009 , india _ +
|
a powerful classical double radio source , also known as an frii source ( fanaroff & riley 1974 ) , is powered by large scale outflows from a supermassive black hole system that resides at the center of a galaxy ( e.g. blandford & rees 1974 ; scheuer 1974 ) .
the energy per unit time , or beam power , carried from the vicinity of the supermassive black hole to the large scale radio source is a fundamental physical parameter that , in part , describes the physical state of the black hole system .
the beam power can be used to study many aspects of the source , the source population , interactions of the source with its environment , and the role of feedback between the black hole system , its environment , and the host galaxy ( e.g. silk & rees 1998 ; willott et al .
1999 ; eilek & owen 2002 ; birzan et al .
2004 ; croston et al . 2005 ; dunn et al . 2005 ; dunn & fabian 2006 ; rafferty et al .
2006 ; belsole et al . 2007 ;
jetha et al .
2007 ; birzan et al . 2008 ; cavagnolo et al . 2008
; daly 2009 , 2011 ; cavagnolo et al .
2010 ; martnez - sansigre & rawlings 2011 ) .
it is possible to determine the beam power of a classical double ( frii ) radio source using multifrequency radio data that covers most of the radio emitting region and resolves several different regions of each radio lobe ( e.g. rawlings & saunders 1991 ; daly 1994 ; kharb et al .
2008 ; odea et al . 2009 ; daly et al .
2010 ) ; see odea et al .
( 2009 ) for a detailed discussion of the method .
obtaining these data requires a considerable amount of observing time .
if a relationship between beam power and radio power can be established , it can be applied to other frii sources so that their beam power can be estimated from their radio power .
the relationship between beam power and radio power has been studied for samples of radio sources that reside in gas rich environments .
these sources are primarily fri sources ( fanaroff & riley 1974 ) at relatively low redshift ( cavagnolo et al . 2010 ; birzan et al .
2008 ; rafferty et al . 2006 ; dunn et al . 2005 ; birzan et al .
it has also been studied for a small sample of 14 frii galaxies in the context of an open cosmological model ( wan & daly 1998 ) . here , the relationship between beam power and radio power is determined for a sample of 31 frii radio galaxies with a broad range of redshift .
the sample is described in section 2 .
the results are presented in section 3 , discussed and compared with other results and model predictions in section 4 , and summarized in section 5 . a standard spatially flat cosmological model with normalized mean mass density @xmath4 , cosmological constant @xmath5 , and hubble constant @xmath6 is assumed throughout .
the sample studied consists of 31 very powerful frii radio galaxies with redshifts ranging from about 0.056 to 1.79 and core - hot spot sizes ranging from about 30 to 400 kpc drawn from the sample studied by odea et al .
the sources have both beam powers and radio powers determined , and are 3crr sources ( laing , riley , & longair 1983 ) .
the sources and source properties are listed in table 1 .
the intrinsic 178 mhz energy per unit time per unit frequency , @xmath7 , is listed in column 3 , and is determined from the 178 mhz flux density , @xmath8 , and the radio spectral index , @xmath9 , defined as @xmath10 .
radio flux densities and spectral indices are obtained from jackson & rawlings ( 1997 ) and willott et al .
the 178 mhz radio power , @xmath11 , is obtained by multiplying @xmath7 by 178 mhz , and the integrated radio power , @xmath12 is obtained by integrating @xmath13 over the frequency range from 200 to 400 mhz using the radio spectral index listed in column 9 .
this frequency range is selected to match that studied by cavagnolo et al .
( 2010 ) so that the results can be easily compared .
values of @xmath11 and @xmath12 are listed in columns 4 and 5 , respectively , of table 1 .
beam powers , @xmath0 , are the weighted sum of the beam power from each side of a given source obtained from odea et al .
( 2009 ) and are listed in column 6 .
the ratio of the beam power to the radio power is listed in columns 7 and 8 for @xmath11 and @xmath12 , respectively .
the classical double radio galaxies studied here are among the most powerful extended radio sources known in the cosmos .
these powerful radio sources are often much larger than their parent galaxies and have cigar - like radio lobe structure indicating that the forward region of the radio emitting lobe is moving into the ambient gas supersonically ( e.g. alexander & leahy 1987 ; leahy , muxlow , & stephens 1989 ; liu , pooley , & riley 1992 ) .
since the structure of the sources indicates that each is growing at a supersonic rate , the equations of strong shock physics may be applied to the sources ( leahy , muxlow , & stephens 1989 ) .
as discussed in detail by leahy ( 1990 ) , rawlings & saunders ( 1991 ) , daly ( 1990 , 1994 ) , wan & daly ( 1998 ) , wan et al .
( 2000 ) , and odea et al .
( 2009 ) , for example , this means that the beam power @xmath14 , where @xmath15 is a constant , @xmath16 is the rate of growth of the source , @xmath17 is the radius of the cross - sectional area of the forward region of the shock , and @xmath18 is the postshock pressure . each of these parameters has been determined empirically .
odea et al . ( 2009 ) provide a detailed description of the method and related uncertainties and a comprehensive list of beam powers obtained using this method ; these are the sources studied here .
remarkably , odea et al .
( 2009 ) found that the beam power is independent of offsets from minimum energy conditions due to a cancellation of the way the offset enters into the determinations of @xmath18 and @xmath16 , so the beam power is insensitive to assumptions regarding minimum energy conditions .
the beam power studied here has been obtained using the equation @xmath14 , where @xmath19 ( e.g. leahy 1990 ; wan et al .
the rate of growth of the source @xmath16 is obtained using the equation @xmath20 .
the change in the radio spectrum across a region @xmath21 along the symmetry axis of the source is used to obtain the time @xmath22 that has elapsed for the source length to change by an amount @xmath21 .
the spectral aging method of meyers & spangler ( 1985 ) was applied to the kardashev - pacholczyk and jaffe - perola models ( jaffe & perola 1973 ) to obtain @xmath22 , and similar results were obtained with both models .
details of the method used to obtain the quantities described above and the computations of their uncertainties are given in sections 2.1 , 2.2 , 3.1 , 3.2 , and 3.3 of odea et al .
( 2009 ) . in brief ,
uncertainties in the injection spectral index , and the flux density at each frequency studied are included in the uncertainty of @xmath22 obtained for each value of @xmath21 .
several different values of @xmath21 were considered for each source , and the value of @xmath23 obtained was found to be independent of the value of @xmath21 . in the computation of the rate of growth of the source , offsets of the true magnetic field strength @xmath24 from minimum energy conditions were parameterized by @xmath25 where @xmath26 is the minimum energy magnetic field , and the geometric mean of the minimum energy fields 10 and 25 kpc from the hot spot along the symmetry axis of the source were used to obtain @xmath26 over the spectral aging region . the postshock pressure @xmath18
was taken to be the pressure in the region 10 kpc behind the hot spot moving toward the core of the host galaxy along the symmetry axis of the radio source ; the pressure can be written @xmath27 ( e.g. wan et al .
2000 ) , where @xmath26 here is obtained 10 kpc from the hot spot .
it is shown in section 3.3 of odea et al .
( 2009 ) that the offset from minimum energy conditions entering through @xmath16 cancels that entering through @xmath18 so that , for the sources studied here , @xmath0 is independent of offsets from minimum energy conditions .
the uncertainty of the beam power is dominated by the uncertainties of the source rate of growth and pressure .
the half - width of the source @xmath17 is taken to be the deconvolved half - width of the source , which is equal to @xmath28 times the deconvolved fwhm of the source ( see , for example , section 5.1 of wellman et al .
widths were extracted at 1.4 ghz a distance 10 kpc behind the hot spot .
detailed studies of these sources show that the source width is independent of the frequency of observation for the range of frequencies studied ( daly et al .
for the new sources studied by odea et al .
( 2009 ) , the widths were obtained using a cross - sectional slice taken perpendicular to the symmetry axis of the source , as described by daly et al .
the deconvolved fwhm , @xmath29 , is given by @xmath30 , where @xmath31 is the fwhm of the best fit gaussian to the slice and @xmath32 is the observing beamwidth .
the widths obtained were found to be in good agreement with those obtained earlier by wellman et al .
the uncertainty of the deconvolved fwhm , @xmath33 , was obtained using the expression @xmath34 , where @xmath35 is the uncertainty of the fwhm of the best fit gaussian profile and @xmath36 is the uncertainty of the observing beam . in practice
, @xmath33 was always very close to @xmath35 ; only widths very close to the hot spot had a larger uncertainty , and these were at most about 35 % larger than @xmath35 .
the beam power was studied as a function of radio power for the full sample of 31 sources and for a sample of 30 galaxies obtained by excluding the lone low redshift source in the sample , cygnus a. very similar results are obtained using the radio power defined by @xmath11 and @xmath12 , and considering the samples of 30 and 31 sources .
the results are illustrated in figure 1 , and best fit parameters are listed in table 2 . the uncertainties on the best fit parameters have been adjusted to bring the reduced chi - square of the fit to unity .
overall , the data are described by the equation @xmath37 where @xmath38 , and @xmath39 .
for powerful frii sources , most of the radio power is produced by the outer half of the radio lobe ( near the radio hot spots ) ( e.g. leahy , muxlow , & stephens 1989 ) .
the result obtained here , @xmath40 , is consistent with @xmath41 at about one sigma .
these results are consistent with those obtained with a sample of 14 radio galaxies ( wan & daly 1998 ) .
the efficiency with which beam power is converted to radio power is @xmath42 , and the constant of proportionality suggests the beam power may be converted to radio power with an overall efficiency of about 0.7 % . the radio power and the beam power each increase with redshift , so there is a concern that the intrinsic relationship between beam power and radio power may be masked because each is correlated with redshift .
this effect can be circumvented by studying the ratio @xmath43 rather than studying each quantity separately . to explore this
, the ratio of the beam power to the radio power , @xmath43 , is studied as a function of redshift with and without cygnus a. the results are illustrated in figure 2 , and best fit parameters are listed in table 2 . without cygnus
a , the ratio @xmath43 is independent of redshift , indicating that @xmath41 is intrinsic to the sources .
when cygnus a is included , there is a tendency for the ratio to increase with redshift though the significance of the result is only about one sigma .
this suggests that the relationship between beam power and radio power obtained for powerful classical double radio sources is intrinsic to the sources and is independent of redshift for redshifts ranging from about 0.4 to 2 , and possibly for redshifts from about zero to 2 .
the expression used to obtain the beam power can be compared with that used to obtain the radio power of a given source .
the radio power is obtained by integrating the radio surface brightness of the the source to obtain the flux density , multiplying this by the luminosity distance squared and appropriate redshift factors , and multiplying by the rest frame radio frequency or integrating over radio frequency to obtain @xmath11 or @xmath12 , respectively .
the beam power depends on @xmath44 , where @xmath45 and @xmath46 are the widths and @xmath47 and @xmath48 are the mean surface brightnesses of cross - sectional slices of the radio lobe 10 and 25 kpc , respectively , from the hot spot along the symmetry axis of the source and @xmath49 is the spectral aging break frequency .
this expression assumes that synchrotron cooling dominates over inverse compton cooling and that @xmath50 , both of which apply to the sources studied here ( odea et al .
as noted above , the length @xmath21 is measured from the hot spot to the location along the symmetry axis of the lobe where the break frequency @xmath49 is measured , and the product @xmath51 was found to be independent of @xmath21 , thus , the beam power is also independent of @xmath21 ( odea et al .
the beam power depends upon several factors that do not enter into the calculation of the radio power , and the radio power depends upon factors that do not enter the beam power , so the correlation obtained here between these two parameters is likely to characterize an intrinsic physical relationship between beam power and radio power . in addition , the results obtained here are consistent with those obtained by cavagnolo et al .
( 2010 ) who used a completely different and independent method of determining the beam power , as discussed in section 4 .
the results obtained here can be compared with those obtained by cavagnolo et al .
( 2010 ) who studied radio sources in gas rich clusters of galaxies and other gas rich environments .
cavagnolo et al .
( 2010 ) expanded upon the sample studied by birzan et al .
( 2008 ) by adding very lower power radio sources to the sample .
the sources studied by these groups are primarily fri sources .
the method used to obtain the beam power for these sources is significantly different from that described in section 3 . for the radio sources in gas rich environments studied by birzan et al .
( 2008 ) and cavagnolo et al .
( 2010 ) , the p dv work done by the relativistic plasma on the intracluster medium is obtained and combined with the buoyancy timescale of the radio source to obtain the beam power ( e.g. fabian et al .
2006 ; rafferty et al .
cavagnolo et al .
( 2010 ) obtain @xmath52 , or @xmath53 , and the coefficient may be approximated as @xmath54 . as noted by these authors ,
the buoyancy timescale is likely to be an overestimate of the true source age and some of the sources may be affected by multiple outflows ; these uncertainties would increase the uncertainties on parameters stated above .
the slope of the relationship obtained here , @xmath55 , is within one sigma of that obtained by cavagnolo et al .
( 2010 ) , @xmath56 . the normalization of the relationship obtained by cavagnolo et al .
( 2010 ) , @xmath54 , differs from that obtained here , @xmath57 by less than 2 sigma .
thus , the results obtained are remarkably similar even though different types of radio sources at different redshifts are studied . the small differences that do exist
may be accounted for by uncertainties in fitted parameters , or by intrinsic differences between the sources .
more data will be needed to compare results obtained with different methods in more detail .
the agreement between the results obtained here and those of cavagnolo et al .
( 2010 ) is particularly remarkable because the interaction of the sources with their environments is quite different : the forward region of frii sources is moving supersonically with respect to the ambient gas resulting in a strong shock whereas fri sources are not moving supersonically and in any case the method used by cavagnolo et al .
( 2010 ) does not include energy that has gone into shocks .
the results of cavagnolo et al .
( 2010 ) described above imply an efficiency of conversion of beam power to radio power of @xmath58 ; the slope is consistent with zero at about 2.5 sigma and the coefficient suggests a conversion efficiency of about 0.3% .
if the slope is taken to be non - zero , it would suggest that the conversion efficiency increases with radio power and is about 0.3% at a power of @xmath59 .
thus , both types of source appear to convert beam power to radio power with similar efficiencies : about 0.7% for frii sources and about 0.3% for fri sources .
this provides a hint that the physics of the conversion of beam power to relativistic electrons and magnetic fields may be similar for fri and frii sources , perhaps with the efficiency of conversion increasing with radio power for fri sources .
alternatively , it is possible that different physical processes for the conversion of beam power to radio power have similar efficiencies .
the results obtained here can be compared with the theoretical model described by willott et al .
( 1999 ) , who found that @xmath60 , where the term @xmath61 is expected to lie between 1 and 20 and is a combination of factors including the filling factor of the lobes , departure from minimum energy conditions , and the fraction of energy in non - radiating particles .
this model prediction is obtained in the context of a detailed model based on the work of falle ( 1991 ) and kaiser & alexander ( 1997 ) for extended radio sources .
the dependence of beam power on radio power predicted by the model , which has an exponent of @xmath62 , is remarkably close to the value of @xmath63 obtained here .
the normalization of the relationship obtained here , @xmath64 , indicates a value of @xmath65 . the fact that both the results obtained here and those obtained by cavagnolo et al . (
2010 ) are in fairly good agreement with the model predictions suggests that the fundamental model provides a good description of the sources .
a sample of 31 powerful frii radio galaxies is studied to determine the relationship between beam power and radio power for sources of this type .
the data indicated that @xmath66 , where @xmath38 , and @xmath39 .
the results listed in table 2 may be applied to samples of frii sources with known radio power to obtain estimates of the beam power . to determine whether the relationship between beam power and radio power evolves with redshift
, the ratio of the beam power to radio power was studied as a function of redshift with and without the one low redshift source , cygnus a. excluding cygnus a , there is no redshift evolution of the ratio of beam power to radio power . including cygnus a
, there is a tendency for the ratio to increase with redshift , but the significance of the result is marginal .
thus , there is no conclusive evidence at this time that the ratio evolves with redshift .
the results obtained here are broadly consistent with those obtained by cavagnolo et al .
( 2010 ) for samples of radio sources in gas rich environments , which consist primarily of low redshift fri sources .
interestingly , the results obtained here are consistent with the radio power being proportional to the beam power , with an efficiency of conversion of beam power to radio power of about 0.7% that is independent of or very weakly dependent upon radio power .
this is comparable to the efficiency of about 0.3% indicated by fri sources , which may be weakly dependent upon radio power in the sense that the efficiency of conversion of beam power to radio power may increase weakly with radio power .
the relationship between beam power and radio power obtained here may be compared with theoretical predictions obtained in the context of a specific model for the sources .
the agreement between the results obtained here and the detailed model described by willott et al .
( 1999 ) is quite impressive .
we would like to thank the referee for very helpful comments and suggestions .
this work is supported in part by penn state university ( r.a.d . and t.b.s . ) and the radcliffe institute for advanced study at harvard university ( s.a.b . ) .
|
beam power is a fundamental parameter that describes , in part , the state of a supermassive black hole system .
determining the beam powers of powerful classical double radio sources requires substantial observing time , so it would be useful to determine the relationship between beam power and radio power so that radio power could be used as a proxy for beam power .
a sample of 31 powerful classical double radio sources with previously determined beam and radio powers are studied ; the sources have redshifts between about 0.056 and 1.8 .
it is found that the relationship between beam power , @xmath0 , and radio power , @xmath1 , is well described by @xmath2 , where both @xmath0 and @xmath1 are in units of @xmath3 .
this indicates that beam power is converted to radio power with an efficiency of about 0.7% .
the ratio of beam power to radio power is studied as a function of redshift ; there is no significant evidence for redshift evolution of this ratio over the redshift range studied .
the relationship is consistent with empirical results obtained by cavagnolo et al .
( 2010 ) for radio sources in gas rich environments , which are primarily fri sources , and with the theoretical predictions of willott et al .
( 1999 ) .
[ firstpage ] black hole physics galaxies : active
|
using multi - band light curve of the rxte - asm , we had reported a significant change in the x - ray spectrum of cen x-3 in the period december 2000 to april 2004 .
this was interpreted as a change in the accretion mode in this period ( paul , raichur & mukherjee , 2005 ) .
recently , mller et al . ( 2011 ) reported non detection of this feature in the rxte - asm light curve of the same time span and also in some additional data from rxte - asm , maxi , integral and swift .
we fully agree with the findings of mller et al .
( 2011 ) here we report that the earlier finding was not due to any error in analysis of the light curves used in paul , raichur & mukherjee ( 2005 ) but due to a discrepancy in the asm light curve for the same period that was made available in 2005 and what is available now .
we would like to note here that the files used by us for the report of 2005 were all for observations from 4th or 5th january 1996 to 7th november 2004 and all the fits files were created on 9th november 2004 . below
we show a plot of the light curves in the three energy bands that was available in 2005 ( with no further analysis done on the original lightcurves ) and the light curves now available from the nasa s high energy astrophysics science archive center ( heasarc ) .
all these light curves are the dwell - by - dwell data as used in the mller et al ( 2011 ) research note .
the time bin used to plot the light curves is 2.08702 d. the red points are from the older light curves and the black points are from the current light curves .
only data points upto tjd
13600 are ploted for clarity . while small differences are seen between the two sets for the 1.5 - 3 kev and 3 - 5 kev bands , the two light curves in the 5 - 12 kev band are significantly different in the second half of the plot .
we have also verified that this difference is present only for the light curves of cen x-3 and not in a significant way for the other sources namely vela x-1 , her x-1 and smc x-1 for which we have the older version of light curves to compare .
the older version of the light curves for cen x-3 , her x-1 , vela x-1 and smc x-1 are available with the authors and can be shared on request .
the multiband rxte - asm light curves of cen x-3 that were available at heasarc in 2005 are different from the light curves that are now available at the same site .
we would also like to note here that before publishing the results in 2005 , we had discussed with the rxte - asm team members about any possible artifacts .
the difference in the light curves was noticed in 2010 and the same was communicated to the rxte - asm team . as
yet , no clarification is available .
this information about possible artifacts will be useful for users of asm light curves .
|
in paul , raichur and mukherjee ( 2005 ) we had reported a spectral mode change in the x - ray binary pulsar cen x-3 using the then available asm light curves of cen x-3 .
being mindful of the fact that instrument calibration or other issues like gain changes , software issues etc . might give rise to some of the artifacts , we had used the asm light curves of three other x - ray binary pulsars namely her x-1 , smc x-1 and vela x-1 for comparison for the same observation times . in light of the recent
finding of mller et al .
( 2011 ) ; which reports the non - detection of such features in the current asm light curves available on the heasarc website and with other instruments like maxi , swift and integral , we would like to point out the inconsistency between the asm ligthcurves available in 2005 and now .
this information will be very useful for other users of asm light curves .
|
financial derivatives ( also known as contingent claims ) are now a ubiquitous tool in risk management with approximately 600 trillion dollars worth of such contracts currently in the market .
the pricing of such derivatives is therefore an active area of research in both mathematics and finance ( see for example @xcite and the references therein ) . in this paper
, we will apply the perturbative ( asymptotic ) method introduced in @xcite for numerically solving parabolic equations and then use this method to price european options .
one of the earliest models used in pricing derivatives is the black - scholes - merton model @xcite , for which the movement in the price @xmath0 of the underlying asset on which the claim is based is modeled by geometric brownian motion . for the black - scholes - merton as well as for other models given by stochastic differential equations
, the pricing of european options can be reduced to the calculation of certain solutions of parabolic equations , obtained through ito s lemma ( and the change of variables @xmath1 ) in the backward kolmogorov equation .
the resulting equation is a fokker - planck equation , which is an equation of parabolic type .
fokker - planck equations more generally have important applications in statistical mechanics and in probability ( see for example the monographs @xcite ) .
given that the asset price is always assumed positive , the fokker - planck equation is solved on the positive half - line .
one difficulty in treating this type of the equation is that the coefficients of the fokker
planck operator typically vanish at the boundary , making the equation degenerate . for example , for the black - scholes - merton model
, the resulting fokker - planck equation is given by @xmath2 where @xmath3 is the _ black - scholes operator _
, a degenerate elliptic operator , @xmath4 is the time to expiry , and @xmath5 is the so - called _ pay - off _ function . for a _
european call option with strike @xmath6 _ and expiry ( or exercice ) date @xmath7 , the pay - off function @xmath5 is given by the formula @xmath8 , where @xmath9 is the price of the underlying asset at time @xmath7 .
above , @xmath10 and @xmath11 are constant parameters , representing respectively the _ volatility _ of the underlying asset , and the current _
interest rate_. since the operator is degenerate at the boundary @xmath12 , it can be shown that the solution automatically vanishes there and no explicit boundary condition need to be imposed .
a popular model related to the black - scholes - merton model is the cev model @xcite . in the cev model ,
the operator @xmath13 is the form @xmath14 where @xmath10 , @xmath15 , @xmath11 are constant .
yet another popular model is dupire s local volatility model , for which we allow the volatility to change with time : @xmath16 except in special cases , such as the black - scholes - merton equation above and when @xmath13 has constant coefficients , very few exact solution formulas to the problem are available .
it is therefore important to devise fast , accurate approximate solution methods .
the focus of this paper is on obtaining approximate solution methods that are fast and accurate by combining standard numerical methods with the asymptotic techniques developped in @xcite .
fast solution methods are crucial when calibrating unknown parameters , especially in the baeysian inference framework .
we hope to address this question in a forthcoming paper . in view of the above discussion , it is justified to study the forward initial - value problem for the general case when @xmath13 is an operator of the form : @xmath17 we therefore allow for _ variable _ coefficients _ in both space and time_. we assume throughout that @xmath18 , for @xmath19 and that the coefficients @xmath20 , @xmath21 , @xmath22 are smooth functions .
the perturbative method introduced in @xcite for the study of parabolic equations in arbitrary dimensions was fully justified in the case when @xmath20 , @xmath21 , and @xmath22 and all their derivatives are bounded , and are bounded away from zero : @xmath23 . in this paper
we complete the results of @xcite with explicit formulas for the 1d case .
then we numerically test our formulas for the black - scholes - merton and cev models , obtaining an excellent agreement between our theoretical results and the numerical tests .
both the black - scholes - merton model and and the cev model are more general than the models considered in @xcite in that their coefficients do not satisfy the assumptions of the paper , yet the numerical tests indicates that the results of that paper are still valid for the more general models considered here .
this observation suggests that the theoretical framework of @xcite is applicable in greater generality .
we plan to study this point in a forthcoming paper .
to explain our method , let us recall that , under certain conditions on the operator @xmath13 and initial value @xmath5 , described in details in the next section , there exists a _
smooth function _
@xmath24 such that the solution to has the representation @xmath25 the kernel function @xmath24 in is the fundamental solution or the so - called _ green function _ for the problem .
[ rem.kernel ] given that @xmath24 arises in several different contexts , we will call the function @xmath24 the _ transition density kernel , pricing kernel , heat kernel _ , or _
green function _ interchangeably , depending on the context in which the object arises . as mentioned above , except for some very special cases no explicit formulas for @xmath24 or @xmath26 are available . for the black - scholes - merton model ,
a change of variables reduces the pde to a heat equation that can then be solved explicitly .
therefore , exact formulas for the kernel @xmath27 and the solution @xmath28 exist , which we recall now for further reference : @xmath29 where @xmath30 is the cumulative normal distribution function ( cumulative gaussian distribution function ) and @xmath31 however , for the time - dependent black - scholes - merton model , where @xmath10 and @xmath11 are time - dependent , or local volatility models in general , closed form solutions are generally given by series expansions and difficult to use in practice or are not known ( see , for instance , @xcite ) .
the method that we use in this paper is to give an approximate closed - form solution for the equation by giving an approximate closed - form expansion for the green s function @xmath24 .
since our approximation of the green s function is in terms of gaussian - type integrals , it gives a closed - form for the approximate price of a european call option for _ any _ one - dimensional model where the operator @xmath13 is given by .
in fact , as an application , we give the prices and greeks ( that is , suitable derivatives ) of a european call option and perform an error analysis in section [ sec.error ] .
there exists a vast literature on obtaining asymptotic expansions of the green s function @xmath24 when @xmath4 small and @xmath32 is close to @xmath33 , especially in the case that @xmath13 is independent of time @xcite .
( see also @xcite ) .
many of these methods are based on a geometric interpretation of the operator @xmath13 ( or at least its principal part ) as a laplace operator on curved space , and require computing the geodesics in this space , which very often must be done numerically .
other approaches are based on pseudo - differential calculus .
in particular , corielli , foschi , and pascucci @xcite use a _ parametrix _ construction for the problem to obtain a closed - form approximate solution .
we recently developed in @xcite a complementary approach to computing short - time asymptotics for @xmath34 , based on parabolic rescaling , taylor s expansions of the coefficients , duhamel s and dyson s formulas , and exact commutator expansion .
we called this method the _ dyson - taylor commutator method_. our method is more elementary and appears very stable in practical implementations .
let us fix a function @xmath35 with the properties that @xmath36 and all its derivatives are bounded .
the function @xmath37 will represent the _ basepoint _ for a parabolic rescaling of the green s function .
then our short - time asymptotics give an expansion for the kernel in the form : @xmath38}(x , y;z ) + t^{(n+1)/2 } { { \mathcal e}}_t^{[n]}(x , y ; z ) \\ { { \mathcal g}}_t^{[n]}(x , y;z ) & = g_t^{[0]}(x , y ; z ) + t^{1/2 } g_t^{[1]}(x , y ; z ) + \cdots + t^{n/2 } g^{[n]}_t(x , y ; z ) , \\ \end{split}\end{aligned}\ ] ] where @xmath39}$ ] is the sum on the first @xmath40 terms of the expansion and represents the @xmath40-th order approximate kernel , while @xmath41}$ ] is the remainder .
the first term , @xmath42}$ ] is given by a dilated gaussian function @xmath43}(x , y ) = g_{t}^{[0]}(x - y ) = \frac{1}{\sqrt{2 t \pi a(0,z)^2 } } \exp ( - \frac{|x - y|^2}{2 t a(0,z)^2}).\ ] ] the dyson - taylor commutator method yelds an explicit algorithm to compute the terms @xmath44}$ ] for _ any _
@xmath40 , if @xmath13 is an operator of the form and corresponding analogs in higher dimension .
more precisely , our main result in @xcite is that for the local volatility operator , the @xmath40-th order approximate kernel has the form @xmath45}(x , y ) : = t^{-1/2 } \mathfrak{p}^k(\,z , \,z+\ , \frac{x - z}{t^{1/2}}\ , , \,\frac{x - y}{t^{1/2}}\ , ) g^{[0]}_t ( \,\frac{x - y}{t^{1/2}}\,),\ ] ] where the functions @xmath46 are algorithmically computable ( recall that @xmath47 ) . in this paper
we shall compute the functions @xmath48 , for @xmath49 at an _ arbitrary _ basepoint @xmath37 .
the details , based on the dyson - taylor commutator method method , can be found in section [ sec.framework ] and [ sec.basepoints ] .
we therefore obtain _ new _ closed form asymptotic expansions of the green function for local volatility models .
in particular , the first order asymptotic expansion at arbitrary @xmath47 is given by @xmath50}(x , y;z ) = \frac{1}{\sqrt{2 \pi\ , t\ , a(0,z)^2 } } \left [ 1 + \frac{3 a(0,z ) a ' ( 0,z ) - 2 b(0,z ) } { 2 a(0,z)^2 } ( x - y ) \right .
\\ \left .
- \frac{a'(0,z)}{2 t\ , a(0,z)^3}(x - y)^3 + ( x - z ) \frac{(x - y)^2 - t\,a(0,z)^2}{t\ , a(0,z)^3 } \right ] e^ { - \frac{|x - y|^2}{2 t\ , a(0,z)^2}. } \end{gathered}\ ] ] we provide an explicit formula for the second order expansion of the green function at the end of section [ sec.framework ] .
this algorithm can be implemented very efficiently at least in dimension 1 and for @xmath40 small , @xmath51 , @xmath52 .
the numerical tests in section [ sec.error ] show that already the second - order approximation is adequate for the black - scholes and cev models . for each term @xmath53}$ ] in the expansion of the green function , let @xmath54}$ ] denote the corresponding term in the expansion of the solution , @xmath55}(t , x ) = \int_0^\infty g_t^{[k]}(x , y ) h(y;k)dy.\end{aligned}\ ] ] then using and , we arrive at the expansion of the value of the contingent claim , @xmath56}(t , x ) + t^{(n+1)/2}{\mathfrak e}^{[n]}(t , x ) \\ & u^{[n]}(t , x ) = { { \mathcal u}}^{[0]}(t , x ) + t^{1/2 } { { \mathcal u}}^{[1]}(t , x ) + t { { \mathcal u}}^{[2]}(t , x ) + \cdots t^{n/2 } { { \mathcal u}}^{[n]}(t , x ) \end{split}\end{aligned}\ ] ] where @xmath57}(t , x ) : = t^{(n+1)/2}\int_0^\infty { { \mathcal e}}_t ^{[n]}h(y;k ) = u(t , x ) - u^{[n]}(t , x)\end{aligned}\ ] ] is the remainder term ( or error ) in the expansion of the solution . in @xcite
we have shown that the remainder can be controlled in exponentially weighted sobolev norms , when the operator @xmath13 is uniformly strongly elliptic .
these bounds on the remainder imply that , in this case , the error made by replacing @xmath34 with @xmath58}_t$ ] in is of order @xmath59 _ globally _ in space , the expected optimal rate . in @xcite , we consider degenerate operators , the symbol of which is strongly elliptic with respect to some complete metric of bounded geometry .
for example , the black - scholes and the sabr models fit into this framework .
by contrast , the cev model with @xmath60 does not fit into this framework .
our numerical tests indicate nevertheless that the error term has the same order in @xmath4 even for the cev model with @xmath61 . for pedagogical purposes and error analysis we will list all the details for the time - dependent black - scholes and cev models , although our results are more general . in section [ sec.error ] we perform a numerical error analysis by computing both the numerical solution @xmath62 and expansion @xmath63}$ ] and estimating the error @xmath64}(t , x ) |\end{aligned}\ ] ] pointwise for the basepoint @xmath65 , when @xmath66 .
the error analysis is in good agreement with the theoretical results , even though the local volatility operators considered in this paper do not necessarily satisfy the assumptions on the coefficients of @xmath13 needed to establish the analytic error estimates performed in @xcite . in section [ sec.error ]
we then perform an error analysis . for the black - scholes - merton model ,
for which an exact solution formula is readily available , we compare the expansions at the basepoint @xmath65 with the exact solution .
( note however , that numerical errors arise also in the calculation of exact solutions , due to round - off errors and other approximations . ) for the cev model , we compare the expansions with benchmark formulas in the literature , in particular the hagan - woodward implied volatility approximation @xcite .
given that the kernel approximation is asymptotic in time , it guarantees good error control _ a piori _ only for sufficiently small @xmath4 . in section [ bootstrap ] , we shall introduce a bootstrap scheme to extend our method to arbitrary large time .
this strategy is based on the evolutionary property of the solution operator to . by doing so ,
we show that the error is remarkedly reduced . as an application in portfolio management
, we also compute the greeks ( or hedging parameters ) of a european call option and compare our approximations with the true black - scholes greeks in section [ greeks ] and section [ bootstrap ] .
these applications again underline the accuracy of our methods .
the authors would like to thank marco avellaneda for valuable suggestions and comments on the manuscript , and jim gatheral for useful discussions .
victor nistor also gladly acknowledges support from the max planck institute for mathematics , where part of this work has been performed .
we begin by recalling the dyson - taylor commutator method , which we introduced in @xcite , to obtain small - time asymptotic expansions for the solution of the initial - value problem : @xmath67 throughout the paper , the operator @xmath13 will be given by , and we will omit the explicit dependence of @xmath13 and of its coefficients on @xmath32 .
in addition , we tacitly assume that all the coefficients of @xmath13 are regular enough to carry our the manipulations described next . for
a rigorous justification in the case @xmath13 is not degenerate , we refer to @xcite .
if there is a unique solution to the initial - value problem , then the linear operator that maps the initial data @xmath5 to the solution @xmath62 is well defined .
we refer to such operator as the _ solution operator_. for constant - coefficient second - order operators , @xmath68 , the solution operator forms a _ semigroup _ , denoted by @xmath69 , @xmath70 ; that is , the solution operator has the following properties : 1 .
2 . @xmath72 , @xmath73 .
[ i.semigroup2 ] the same conclusion hold for variable - coefficient , but time - independet operators @xmath13 , under some conditions , for instance if @xmath13 is strongly elliptic @xcite ( that is , @xmath74 for all @xmath32 ) . when @xmath13 is a time - dependent operator , @xmath75 , the solution operator is no more a semigroup , but under some additional mild conditions , forms an _ evolution system _
@xmath76 @xcite . for an evolution system , property ( [ i.semigroup2 ] ) is replaced by @xmath77 , if @xmath78 . following the notation set forth in the introduction
, we denote the kernel or green s function of the solution operator to the problem by @xmath79 .
our method relies heavily on the study of distribution kernels of the evolution operators defined by our fokker - planck operator , so a brief discussion of distribution kernels and of our conventions is in order .
[ rem.genkerneldef ] given a linear operator @xmath7 mapping smooth functions with compact support into distributions , there exist a _
distribution kernel _
@xmath80 such that @xmath81 the integral above is interpreted as the pairing between test functions and distributions . in this paper , we will be interested in the integral representation in the case that @xmath7 is a _ smoothing _ operator , that is , an operator that maps compactly supported distributions into smooth functions .
then , the kernel @xmath80 is a _
smooth function _ , and the notation @xmath82 is justified pointwise .
( for a more detailed dicussion , see for example @xcite . ) in this case , we will write @xmath83 to denote the kernel @xmath82 , and in general , _ we shall identify an operator with its distribution kernel_. let @xmath84 be a smooth function , then we denote the operators of multiplication by @xmath84 also with @xmath84 .
additonally , we notice that there is no confusion when writing @xmath85 or @xmath86 since the distribution kernels of these operators are given by @xmath87 or @xmath88 .
similarly , there is no confusion when writing @xmath89 , since the distribution kernel of @xmath90 is @xmath91 .
however @xmath92 .
we now introduce parabolic rescaling , which is a basic tool used in this paper .
let @xmath37 be a fixed , but arbitrary point in @xmath93 and @xmath94 a parameter . given a function @xmath95 we denote by @xmath96 the _ parabolic rescaling by @xmath97 of the function @xmath84 about @xmath98_. thus @xmath99 for a function that does not depend on @xmath4 .
we will refer to @xmath37 as the _ basepoint _ for the rescaling .
similarly , we define a rescaled operator @xmath100 by @xmath101 if @xmath62 solves the initial - value problem , then @xmath102 solves the rescaled problem @xmath103 consequently , the green functions of the operator @xmath104 and of the rescaled operator @xmath105 are related by @xmath106 we now proceed to compute the green s function @xmath107 of the rescaled problem when @xmath108 . in order to do so
, we shall consider the taylor expansion in @xmath97 at @xmath109 of the rescaled operator @xmath100 , given in equation , up to order @xmath40 . by `` taylor expansion '' we mean that we taylor expand the coefficients of @xmath100 and group all terms of the same order in @xmath97 .
the operator @xmath100 can then be written as follows @xmath110 where @xmath111 contains all the remainder terms from the taylor expansion of the coefficients . in this paper , we concentrate on calculating explicitly the second - order approximation of the green function of @xmath13 .
hence , _ we fix @xmath52 from now on_. for notational convenience , we denote @xmath112 and @xmath113 . then the second - order taylor expansion in @xmath97 of @xmath114 at @xmath109 is given by @xmath115 with @xmath116 the remainder . below @xmath117 and
all the other functions are to be evaluated at @xmath98 , unless stated otherwise .
we then readily have the second order taylor expansion of @xmath100 in @xmath97 at @xmath109 : @xmath118 and , @xmath119 , where @xmath120 hence @xmath121 where @xmath122 is the remainder term .
[ remark.1 ] each @xmath123 in has polynomial coefficients of order @xmath124 in @xmath125 and of order @xmath126 in @xmath4 .
in particular , @xmath127 is a constant coefficient operator , for which the green s function is computed explicitly in .
thus , in order for the expansion to capture the time dependence of the coefficients , the coefficient must be expanded at least to second order in @xmath97 .
time - dependent corrections will therefore appear only at order @xmath128 in the expansion of @xmath129 .
let @xmath130 be the green function of the parabolic problem , that is , the solution is given by @xmath131 .
we begin the approximation scheme for @xmath79 by decomposing @xmath13 into a constant - coefficient , second - order operator @xmath68 , for which we can explicitly compute the solution operator , and a remainder : @xmath132 where @xmath133 is a time - dependent , variable coefficient , second order operator . by duhamel
s principle we then have @xmath134 repeated applications of duhamel s formula leads to a recursive representation of @xmath130 as a time - ordered expansion : @xmath135 where , for notational convenience , we have set @xmath136
. this expansion can be rigorously justified , at least in the case when @xmath13 uniformly strongly elliptic and all the coefficients of @xmath13 and their derivatives are bounded .
see @xcite for details . in the limit @xmath137
, it yields an asymptotic time - ordered series , also called a _
dyson series _
, for the green s function .
the integer @xmath138 stands for the _ iteration level _ in the time - ordered expansion , which at this point is distinct from the order @xmath40 of the taylor expansion of the operator @xmath13 . for consistency
we need @xmath139 @xcite . _
we set from now on @xmath140 .
_ a similar formula holds for the green s function @xmath141 of the solution operator for the rescaled problem .
we recall that it is enough to compute an approximate green s function at @xmath142 for the rescaled problem by .
we now choose the operator @xmath68 to be exactly the zeroth - order taylor expansion of @xmath100 , given in .
then : @xmath143 and using with @xmath140 and @xmath108 yields @xmath144 where @xmath145 even though we set @xmath142 , we still keep the @xmath4 dependence explicit in @xmath146 to emphasize this term comes from taylor expansion in @xmath4 .
the term @xmath147 in contains all the higher order terms and will be included in the remainder .
the approximation for the green s function @xmath79 of the original problem is now obtained as follows .
let @xmath148 be the distribution kernel of the operator @xmath149 .
the desired second order approximation is then given by @xmath150}(x ,
y ) = t^{-1/2 } t^{\sqrt{t},z}(z + ( x - z)/\sqrt{t } , z + ( y - z)/\sqrt{t}),\ ] ] where @xmath47 is an admissible function . in particular , the kernels @xmath151}_t$ ] appearing in are given by @xmath152}_t ( x , y ) : = t^{-1/2 } g_{n}(z + ( x - z)/\sqrt{t } , z + ( y - z)/\sqrt{t } ; z(x , y)).\ ] ] we thus need to compute the distribution kernels of the operators @xmath153 , @xmath154 , @xmath155 , @xmath156 , @xmath146 . in order to do so , we exploit the semigroup property of @xmath157 to carry out explicitly the time integration in . before we proceed
, we introduce some useful notation . by @xmath158:=t_1 t_2 - t_2 t_1 = - [ t_2 , t_1]$ ]
we shall denote the commutator of two operators @xmath159 and @xmath160 .
two operators @xmath161 commute if @xmath158 = 0.$ ] for any two operators @xmath159 and @xmath160 , we define @xmath162 by @xmath163,$ ] and , for any integer @xmath164 , we define @xmath165 recursively by @xmath166 we next recall a baker - campbell - hausdorff - type identity proved and used in this setting in @xcite ( note that the operators @xmath167 are unbounded ) . namely , for any @xmath168 and differential operator @xmath169 with _ polynomials coefficients _ in @xmath32 , we have @xmath170 where @xmath171 is a differential operator with polynomial coefficients in @xmath32 given by @xmath172 in proving this formula , we use the fact that the series is actually a finite sum , as we show below . in particular , @xmath173 can be _ explicitly _ computed . a simple calculation gives the following lemma .
[ lemma.f.comm ] let @xmath174 be a second - order differential operator with polynomial coefficients of degree at most @xmath175 .
then @xmath176 for @xmath177 .
in particular , we have @xmath178 = 0 $ ] , @xmath179 , and @xmath180 .
the proof is a simple calculation .
we now proceed to compute the integrals in @xmath181)e^{l_0^{z}}d{\tau}_1\\ & = & ( l_1^{z}+\frac{1}{2}[l_0^{z},l_1^{z}])e^{l_0^{z}},\\\end{aligned}\ ] ] @xmath182)(l_1^{z}+ ( 1-{\tau}_2 ) [ l_0^{z},l_1^{z}])e^{l_0^{z}}d\tau_2 d { \tau}_1 , \\ & = & ( \frac{1}{2 } ( l_1^{z})^2 + \frac{1}{3 } l_1^{z}[l_0^{z},l_1^{z } ] + \frac{1}{6}[l_0^{z},l_1^{z}]l_1^{z } + \frac{1}{8}[l_0^{z},l_1^{z}]^2)e^{l_0^{z}},\\\end{aligned}\ ] ] @xmath183+\frac{(1-{\tau}_1)^2}{2}[l_0^{z},[l_0^{z},l_{2,x}^{z } ] ] ) e^{l_0^{z}}d{\tau}\\ & = & \big ( l_{2,x}^{z } + \frac{1}{2}[l_0^{z},l_{2,x}^{z } ] + \frac{1}{6}[l_0^{z},[l_0^{z},l_{2,x}^{z } ] ] \big ) e^{l_0^{z } } , \\\end{aligned}\ ] ] @xmath184 hence becomes @xmath185 where @xmath186 , \\ q_2 = \frac{1}{2 } ( l_1^{z})^2 + \frac{1}{3 } l_1^{z}[l_0^{z } , l_1^{z } ] + \frac{1}{6 } [ l_0^{z } , l_1^{z } ] l_1^{z } + \frac{1}{8 } [ l_0^{z } , l_1^{z}]^2 , \\
+ l_{2,x}^{z } + \frac{1}{2 } [ l_0^{z } , l_{2 , x}^{z } ] + \frac{1}{6 } [ l_0^{z } , [ l_0^{z } , l_{2 , x}^{z } ] ] + \frac{1}{2 } l_{2,t}^{z } , \end{gathered}\ ] ] and @xmath147 is again the error term as in .
therefore , we only need to compute the commutators in the above formula to get the second - order approximation of @xmath187 .
we recall that we agreed that all functions in the commutator formulas below are evaluated at @xmath98 .
hence @xmath117 , @xmath188 , and so on .
we have @xmath189 = a(0,z)^3 a ' ( 0,z ) \partial_x^3 = a^3 a'\partial_x^3 , \quad [ l_0^{z},l_1^{z}]^2 = a^6 { a'{}^2}\partial_x^6,\ ] ] and hence @xmath190 & = a^4 { a'{}^2}(x - z ) \partial_x^5 + b a^3 a ' \partial_x^4 , \\ [ l_0^{z},l_1^{z}]l_1^{z } & = a^4 { a'{}^2}(x - z)\partial_x^5 + \left ( b + 3 aa ' \right ) a^3 a ' \partial_x^4 . \end{split}\end{aligned}\ ] ] to compute the other commutators , we need the following lemma , which can be proved by induction using that @xmath191 = a[b , c ] + [ a , c]b$ ]
. in particular , @xmath192 = 2 \partial_x$ ] and @xmath193 = 2 + 4(x - z)\partial_x$ ] . for @xmath194 integers we have @xmath195
we therefore have : @xmath196 & = a^2 \left ( { a'{}^2}+ a { a'{}'}\right)(x - z ) \partial_x^3 , \\ & + a^2 \left ( b ' + { a'{}^2}/2 + a { a'{}'}/2 \right ) \partial_x^2 , \end{split}\end{aligned}\ ] ] and hence @xmath197 = a^4 ( { a'{}^2}+ a { a ' { } ' } ) , \partial_x^4,\ ] ] so that finally @xmath198 it follows that the approximation kernel of @xmath199 is given by the applications of a differential operator with polynomial coefficients to the green s function of @xmath157 .
if @xmath200 is a smooth function , we denote by @xmath201 the convolution operator with @xmath200 , then @xmath202 , which shows that the distribution kernel of @xmath203 is @xmath204 .
it is immediate to check that @xmath205 while @xmath206 . by remark [ remark.1 ] ,
the distribution kernel of @xmath153 is given by @xmath207 and hence @xmath153 is a convolution operator .
then , by @xmath208 , where @xmath209 and @xmath210 are the ( rescaled hermite ) polynomials satisfying @xmath211 and @xmath212 .
the polynomials @xmath210 are easily computed by induction as : @xmath213 using we have @xmath214 and @xmath215 ) e^{l_0 } \big ) ( x , y ) \\ & = \left ( b \partial_x + a a ' ( x - z ) \partial_x^2 + \frac{1}{2}a^3 a ' \partial_x^3 \right ) e^{l_0^z}(x , y)\\ & = \left ( bh_1(\theta ) + a a ' ( x - z ) h_2(\theta ) + \frac{1}{2}a^3 a ' h_3(\theta ) \right ) e^{l_0^z } \\ & = \frac{1}{\sqrt{2 \pi a^2 } } e^ { - \frac{|x - y|^2}{2 a^2 } } \left [ \left ( \frac{3 a a ' - 2 b}{2 a^2 } \right)(x - y ) - \frac{a'}{2a^3}(x - y)^3 \right . \\ & \left .
+ ( x - z ) \left ( \frac{(x - y)^2 - a^2 } { a^3 } \right ) \right ] . \end{split}\ ] ] we now carry out a similar calculation for the next ( and last ) term of our asymptotic expansion , namely @xmath216 with @xmath217 given by equation .
we finally have @xmath218 + \frac{1}{6 } [ l_0^{z } , [ l_0^{z } , l_{2,x}^{z } ] ] \right .
\\ & \left .
+ \frac{1}{2}l_1^{z,2 } + \frac{1}{3 } l_1^{z } [ l_0^{z } , l_1^{z } ] + \frac{1}{6 } [ l_0^{z } , l_1^{z } ] l_1^{z}+\frac{1}{8 } [ l_0^{z } , l_1^{z}]^2 \right ) e^{l_0^z}\\ & = \left ( p_0+\sum_{i=1}^6 p_ih_i ( \theta)\right ) e^{l_0^z } ( x , y)\ , .
\end{split}\end{aligned}\ ] ] where @xmath219 are polynomials in @xmath220 and @xmath221 with coefficients given in terms of the values of the functions @xmath20 , @xmath21 , and @xmath22 , and their derivatives , all evaluated at @xmath35 , as follows @xmath222,\\ & p_3=a(x - z ) ( a'b + \frac{1}{2 } a^2 { a'{}'}+ \frac{3}{2 } a { a'{}^2}),\\ & p_4=\frac{a^2}{3 } \big [ \frac{1}{2 } a^3 { a'{}'}+ 2 a^2 { a'{}^2}+ a \frac{3}{2 } a ' b + \frac{3}{2 } { a'{}^2}(x - z)^2 \big ] , \\ & p_5=\frac{1}{2 } a^4{a'{}^2}(x - z ) , \ \ p_6=\frac{1}{8 }
a^6 { a'{}^2}. \end{split}\ ] ] in particular , we obtain the following explicit formula .
[ ex.cev1 ] for the cev model given by equation , we have @xmath223 , @xmath224 , @xmath225 , @xmath35 , and hence , @xmath226 e^ { - \frac{|x - y|^2}{2 { \sigma}^2 z^{2{\alpha } } } } .\end{gathered}\ ] ] let us introduce the _ time dependent black - scholes - merton model _ to correspond to the operator @xmath227 thus the difference between the usual black - scholes - merton model and the time dependent black - scholes - merton model is that in the latter we allow @xmath10 and @xmath11 to depend on time .
then the asymptotic formula for the time dependent black - scholes - merton model is obtained by setting @xmath228 in the example [ ex.cev1 ] , since that formula does not contain time derivatives of the coefficients . at this stage
, we can allow the basepoint @xmath37 to vary with @xmath32 and @xmath33 . in section [ sec.basepoints ] below we compute the expansion for the basepoint @xmath65 and compare it in section [ sec.error ] .
different choices of basepoints @xmath37 may lead to more accurate and stable approximations . in future work
, we plan to study how to optimize the choice of @xmath37 .
[ def.admissible ] we call a function @xmath47 _ admissible _ if @xmath36 and all derivatives of @xmath37 are bounded . in @xcite , we rigorously prove error bounds for the remainder term in in sobolev spaces under the assumption that @xmath37 be admissible ( and all the coefficients of @xmath13 , together with their derivatives , be bounded functions , and @xmath13 be uniformly strongly elliptic ) .
the function @xmath47 can be thus quite general .
the choice @xmath230 yields a simplified expression for the approximation , since certain terms disappear , and the approximation yields the price of a european call option in closed form .
in fact , the convolution with the approximate green s function can be evaluated exactly and the price of a european call option given in closed form .
in particular there is no need for numerical quadrature in evaluating the integrals , thus improving the speed of our calculations . by setting @xmath230 in and evaluating all coefficient functions at @xmath231
, we obtain the first - order correction to the rescaled kernel @xmath199 in the form : @xmath232 similarly , the second - order correction to the rescaled kernel @xmath199 is obtained in the form : @xmath233 where @xmath234 , and @xmath235 are given by . for the time - dependent black - scholes - merton equation
, we have @xmath236 , @xmath237 and @xmath238 so that @xmath239,\end{aligned}\ ] ] where all coefficient functions are calculated at @xmath231 . the second - order correction to the rescaled kernel
is given by @xmath240 , \end{split}\end{aligned}\ ] ] where @xmath241 for the time - dependent cev model , @xmath242 and @xmath243 , @xmath244 with @xmath245 so that @xmath246 , \end{split}\end{aligned}\ ] ] and for the second order correction we have @xmath247 where @xmath248 @xmath249 and @xmath250 @xmath251 note that in the above two examples for the cev model , setting @xmath252 leads to the corresponding approximation for the bsm model .
in this section , we consider european call options . for european put options similar results can also be obtained , either directly from the definition or by using put - call parity @xcite . in
what follows , we will work with the expansion obtained by setting @xmath65 as the basepoint . in this case , we are able to compute the integrals defining the approximate option price @xmath253}$ ] from @xmath254}$ ] in closed form , which bypasses the need for more computationally intensive integration methods such as numerical quadrature , which are needed for more general basepoints @xmath37 . by and , the first - order approximate green s function in given by @xmath255}(x , y ) = \frac{1}{\sqrt{2\pi t}a}e^{-\frac{(x - y)^2}{2a^2 t}}\left(1 + \frac{3aa'- 2b } { 2a^2}(x - y ) - \frac{a'}{2a^3 t}(x - y)^3 \right ) .\ ] ] similarly , by and the second - order approximate green s function is given by we recall that we implicitly assume all coefficients are evaluated at @xmath231 .
@xmath256}(x , y)= \frac{\sqrt{t}}{\sqrt{2\pi } a}e^{-\frac{(x - y)^2}{2a^2 t } } \cdot \left ( p_6 h_6(\xi)+p_4h_4(\xi)+p_2 h_2(\xi)-r \right ) \ ] ] where @xmath257 , the functions @xmath258 are given by , and @xmath259 all the coefficient functions are evaluated at @xmath260 . for european call options with strike price @xmath6 , by the @xmath261- order approximated option price
is @xmath262}(t , x)=\int_0^\infty { { \mathcal g}}_t^{[n]}(x , y)(y - k)^+dy,\ ] ] where we only take @xmath66 here .
we recall that @xmath263 is the time to expiry @xmath7 and @xmath264 is real time .
so to be more precise , _ the @xmath261-order option pricing formula for european call options with expiry time
@xmath7 is @xmath265}(t-\mathfrak{t},x)$]_. we have already observed that the general form of the approximate kernel , when @xmath229 , is a product of polynomial functions against a rescaled gaussian . therefore , the integration in above can be carried out in terms of error functions .
explicitly , @xmath266}(t , x)&=&\frac{\sqrt{t}}{2\sqrt{2\pi } } e^{-\frac{(x - k)^2}{2a^2 t}}\left ( 2a - a'(x - k ) \right ) \\ & & + \frac{1}{2}\cdot \left ( { \rm erf } \left(\frac{x - k}{\sqrt{2 t}a}\right)+1\right ) \left(b\ , t+x - k\right),\end{aligned}\ ] ] and @xmath267}(t , x)=u^{[1]}(t , x)+\frac{rt}{2 } \left ( erf\left ( \frac{x - k}{a\sqrt{2 t } } \right)+1 \right ) ( k - x)\\ & + \frac{1}{2\sqrt{2\pi t}}e^{-\frac{(x - k)^2}{2a^2 t}}\cdot \left ( \frac{t^2}{6 } a^2{a'{}'}-a(r+a'^2/12)\ , t^2 \right.\\ & + ( t\ , \dot{a}+{a'{}'}(x - k)^2/3 ) \ , t+\frac{t}{a } ( t\ , b^2-a'^2(x - k)^2/6)\\ & \left.+\frac{t\ , b a ' ( x - k)^2}{a^2}+\frac{a'^2(x - k)^4}{4a^3 } \right ) , \end{split}\ ] ] note that in financial applications ( _ i.e. , _ in the risk free measure ) @xmath268 and @xmath237 . for the cev model we have @xmath269}(t , x)&=\frac{\sigma x^{{\alpha}- 1 } \sqrt{t}}{2\sqrt{2\pi } } e^{-\frac{(x - k)^2}{2 \sigma^{2}t\ , x^{2{\alpha}}}}\left ( ( 2 - { \alpha})x + \alpha k \right ) \\ & + \frac{1}{2}\cdot \left ( { \rm erf } \left(\frac{x - k}{\sqrt{2 t}\sigma x^{\alpha}}\right)+1\right ) \left ( ( 1 + r t)\,x - k \right ) .
\end{split}\end{aligned}\ ] ] and when @xmath252 , it reduces to the first order approximation for the black - scholes - merton model .
in this section , we discuss the accuracy and efficiency of our approximation for the black - scholes - merton and the cev model . we employ the black - scholes - merton model primarily as a didactic example , given that an exact kernel and option pricing formulas exists . for the cev model
, we compare our approximation to other solution formulas considered a benchmark in the literature , in particular the hagan - woodward scheme @xcite .
what we find in general is a very good agreement of the approximate pricing formulas we derive in this paper with those available in the literature , but with significant advantage in the computational efficiency . in particular , the agreement is good even for times that are not small . in section [ bootstrap ] , we propose a bootstrap scheme in time to improve the accuracy of our approximation for large time .
we start by discussing the black - scholes - merton model , and choose the parameters @xmath270 , @xmath271 and @xmath272 , and plot the exact and approximate solutions for @xmath273 .
we compare our formula with the black scholes exact solution formula for different times @xmath4 .
figure [ figure bs ] gives two different cases , which show that when @xmath4 is small the two solutions are in very good agreement with an absolute error of order @xmath274 .
we notice that even when @xmath4 is not small , the error is small .
tables [ table 1 ] and [ table 2 ] give a analysis of the _ pointwise _ error for the first order approximation with respect to the exact black - scholes formula .
throughtout this section , we _ fix the basepoint @xmath229 _ , so that we have closed - form approximate solution formulas , and we can better gauge the error introduced by the our method . for more general basepoints @xmath37 ,
further error is introduced by the numerical quadrature used for the integration and the truncation of the pay - off function @xmath5 at large @xmath32 ( this error is lower order , however , if @xmath5 is truncated at @xmath32 large enough with respect to @xmath6 ) . .
basepoint @xmath229 .
the left picture is for @xmath275 , and the right one is for @xmath276 .
note that the x - axis is scaled by 10 , i.e. the label 150 means the stock price is 15 .
, title="fig:",width=211 ] .
basepoint @xmath229 .
the left picture is for @xmath275 , and the right one is for @xmath276 .
note that the x - axis is scaled by 10 , i.e. the label 150 means the stock price is 15 .
, title="fig:",width=211 ] as predicted , we can increase the number of bootstrap steps to obtain arbitrary accuracy in the aproximation .
furthermore , for relatively large @xmath4 the number of bootstrap steps should be correspondingly large , so that the compound error from each bootstrap step is under control at the end .
for example ,
in our tests when @xmath279 , it is not enough to reduce the error by bootstrapping with @xmath280 steps .
using only @xmath280 steps in this case , in fact , introduces additional errors . for more detail , see @xcite .
in order to eliminate the effect of the truncation error , we shall work with a butterfly option in the rest of this section .
mathematically , a butterfly option corresponds to an intial pay - off given by a hat function , figure .
our method gives closed - form solution for butterfly options as well , by linearity .
figure shows the errors of a butterfly option within the black - scholes - merton model with @xmath281 , @xmath282 , @xmath283 obtained by our first order and second order approximation with or without bootstrap .
the benchmark is the true solution .
the parameters we were using are the same as we mentioned before .
again , we truncate the half line at @xmath284 . for a butterfly option ,
the truncation error is clearly very small , given that the data is compactly supported ( figure ) .
for the second order approximation with a bootstrap scheme , the error is almost zero .
it is in the scale of @xmath278 , while without the bootstrap the error is of the scale @xmath285 .
this coincides with our theoretical results .
we conclude by discussing the bootstrap scheme for the greeks . directly using the closed - form approximation formula to compute the greeks for very long maturity time ( @xmath286 )
is not advisable .
in fact , our closed - form approximation for call option oscillates near the strike price , and the oscillation grows with the time to expiry , as the overall error grows .
the appearance of the oscillation is due to the discontinuity of first derivative of the pay - off function at the strike price .
this phenomenon is clearly visible for butterfly options , where the first derivative of the payoff function has three discontinuities , figure .
this oscillation is amplified in the calculation of greeks .
the bootstrap scheme reduces this oscillation dramatically . for the numerical simulation
, we choose the same parameters as those in section .
the small time step ensures very good error control at each time step . also , we minimize the truncation error as before by truncating the integral at @xmath287 and comparing the approximations only on the interval @xmath288 $ ] near the strike price @xmath289 . the left graph of figure plots the true delta and our approximation in the same picture , and the right one plots the difference between these two curves , which shows that the difference between the true value and our approximation is in the order of @xmath290 with the biggest error around @xmath291
thus our approximation is quite accurate . for the gamma ,
we obtain similar results , see figure [ figure gamma bootstrap ] .
the difference is in the order of @xmath292 , and the biggest error is around @xmath293 . in both cases ,
there are no oscillations on the same scale of the solution . .
basepoint : @xmath229 .
the left graph plots the delta computed by our bootstrap method and the true delta .
the right graph plots their difference . in our numerical integration
, we truncate the half line at 400 . note that in the second figure the scale is @xmath290 . the x - axis scaled by 10 ,
i.e. the label 400 means the stock price is 40 .
, title="fig:",width=211 ] .
basepoint : @xmath229 .
the left graph plots the delta computed by our bootstrap method and the true delta .
the right graph plots their difference . in our numerical integration
, we truncate the half line at 400 .
note that in the second figure the scale is @xmath290 .
the x - axis scaled by 10 , i.e. the label 400 means the stock price is 40 .
, title="fig:",width=211 ] .
basepoint : @xmath229 .
the left graph plots the gamma computed by our bootstrap method and the true gamma .
the right graph plots their difference . in our numerical integration ,
we truncate the half line at 400
. note that in the second figure the scale is @xmath292 .
the x - axis is scaled by 10 , i.e. the label 400 means the stock price is 40.,title="fig:",width=211 ] .
basepoint : @xmath229 .
the left graph plots the gamma computed by our bootstrap method and the true gamma .
the right graph plots their difference . in our numerical integration ,
we truncate the half line at 400
. note that in the second figure the scale is @xmath292 .
the x - axis is scaled by 10 , i.e. the label 400 means the stock price is 40.,title="fig:",width=211 ] r. costantinescu , n. costanzino , a. l. mazzucato , and v. nistor .
approximate solutions to second order parabolic equations i : analytical estimates .
arxiv preprint 0910.1562v2 , i m a preprint 2248 . submitted .
|
we obtain new closed - form pricing formulas for contingent claims when the asset follows a dupire - type local volatility model . to obtain the formulas we use the dyson - taylor commutator method that we have recently developed in @xcite for short - time asymptotic expansions of heat kernels , and obtain a family of general closed - form approximate solutions for both the pricing kernel and derivative price .
a bootstrap scheme allows us to extend our method to large time .
we also perform analytic as well as a numerical error analysis , and compare our results to other known methods .
|
colloidal aggregation phenomena are intersting subjects of study for both theoretical and technological reasons . in systems with short range interactions the main aggregation features
are well understood @xcite .
diffusion limited cluster - cluster aggregation ( cca ) model @xcite and dynamic scaling theory @xcite explain well the scaling properties during aggregation .
it was found that these models , initialy developed for systems with short range interactions , can be used in systems where dipole - dipole interaction is dominant , for example aggregation of magnetic microspheres @xcite , aggregation of nanoparticles in magnetic fluid @xcite and aggregation of magnetic holes @xcite .
these experimental results show scaling of the significant parameters and features typical of cca . on the other hand
, the corresponding exponents may deviate slighly from the known models , and the reasons for this is still not understood .
our previous results @xcite served as motivation for this study .
aggregation of magnetic holes in constant magnetic fields was interpreted in frames of the cca model and dynamic scaling theory .
the scaling exponent @xmath13 for the cluster size dependence @xmath3 was found for particles of diameters @xmath14 and @xmath15 .
this value of the exponent is slighly lower than exponent values predicted by theory ( @xmath16 ) @xcite or exponents found by computer simulations ( @xmath16 for isotropic and @xmath17 for anisotropic aggregation ) @xcite . under certain experimental conditions ( i.e. , particles with larger diameter @xmath18 ) the exponent @xmath0 was close to or lower than the exponent value corresponding to a transition from 2d aggregation to 1d aggregation ( @xmath19 ) . based on these optical observations we know that for a constant magnetic field clusters move in 2d but they grow only in one dimension . the constant magnetic fields induced formation of long chains of particles @xcite .
to determine correct scaling exponents we may take into account hydrodynamic effects @xcite .
we proposed to apply rotating magnetic fields to ensure quasi isotropic properties inside the magnetic fluid ( mf ) sample . in this case clusters
can move in 2d space and can grow as 2d compact objects and thus the hydrodynamic correction is less important .
the dynamic properties of a few magnetic holes @xcite in rotating magnetic fields show interesting phenomena , for example nonlinear response of bound pairs of magnetic holes @xcite , complex braid dynamics @xcite , and equilibrium configurations of rotating particles without contact between particles @xcite . in a precessing magnetic field
paramagnetic particles dispersed in a drop of water self - assemble into two - dimensional viscoelastic small clusters @xcite . in the present study field induced aggregation of many magnetic holes has been observed . in sec .
[ sec : microscopic ] the experimental equipment and the methods used are described . the sec .
[ sec : results ] deals with the results concerning the determination of the scaling exponents and characterization of the diffusion behaviour of individual clusters by tracking of their motions . in sec .
[ sec : discussion ] we summarize the general features and try to explain the non - universal scaling exponents .
our conclusions follow in sec .
[ sec : conclusions ] .
the experimental setup shown in fig .
[ fig:1 ] consists of an optical microscope ( nikon optiphot ) , two pairs of coils , and a carefully prepared thin layer sample .
alternating curents were supplied to the coils in order to produce a magnetic field rotating in the horizontal plane of the sample .
microscopic observations were captured by a ccd camera ( q - imaging micropublisher 5 ) with resulution @xmath20 pixels .
( colour online ) experimental setup used to study aggregation of magnetic holes .
a rotating magnetic field @xmath21 is applied in the plane of the magnetic fluid .
, width=302 ] the sample size was about @xmath22 .
a layer of magnetic fluid of thickness approximately @xmath23 m was confined between two glass plates and sealed .
the kerosene based magnetic fluid @xcite had the following physical properies : density @xmath24 , susceptibility @xmath25 , saturation magnetization @xmath26 , and viscosity @xmath27 .
monodisperse polystyrene microspheres of diameter @xmath28 were dispersed in the mf layer in order to create magnetic holes in presence of magnetic fields .
optical micrographs of aggregation of nonmagnetic microspheres with diameter @xmath29 m in magnetic fluid at different times : ( a ) @xmath30 s , ( b ) @xmath31 s , and ( c ) @xmath32 s after the magnetic field was switched on .
the applied rotating magnetic field had an amplitude @xmath33 a@xmath34 and frequency @xmath35 hz . the optical view covers a sample area of about @xmath36m@xmath37@xmath38m.,width=302 ] optical micrographs of aggregation of nonmagnetic microspheres with diameter @xmath29 m in magnetic fluid at different times : ( a ) @xmath30 s , ( b ) @xmath31 s , and ( c ) @xmath32 s after the magnetic field was switched on .
the applied rotating magnetic field had an amplitude @xmath33 a@xmath34 and frequency @xmath35 hz . the optical view covers a sample area of about @xmath36m@xmath37@xmath38m.,width=302 ] optical micrographs of aggregation of nonmagnetic microspheres with diameter @xmath29 m in magnetic fluid at different times : ( a ) @xmath30 s , ( b ) @xmath31 s , and ( c ) @xmath32 s after the magnetic field was switched on .
the applied rotating magnetic field had an amplitude @xmath33 a@xmath34 and frequency @xmath35 hz . the optical view covers a sample area of about @xmath36m@xmath37@xmath38m.,width=302 ] without a magnetic field the particles are homogeneously dispersed in the layer and they can move freely . after some time a very low fraction of particles may randomly join to other particles and a few dimers were observed @xcite .
their volume fraction is very low in comparison with the volume fraction of single particles .
however , before the application of the rotating magnetic field , a short magnetic field pulse perpendicular to the sample ( coils are not shown in fig . [ fig:1 ] ) was applied in order to destroy these dimers and to create a monodisperse initial size distribution of particles .
this initial stage of the experiment is not shown in fig .
[ fig:2 ] .
the rotating magnetic field @xmath39 within the @xmath40 plane had the components : @xmath41 and @xmath42 .
the amplitude of the magnetic field was constant @xmath43 a@xmath34 , and angular velocity @xmath44 @xmath45 .
the temperature during the experiments was @xmath46 k. the effective volume susceptibility including the demagnetization correction for spherical magnetic holes was @xmath47 .
the dimensionless interaction strength parameter @xcite was @xmath48 . here , @xmath49 , where @xmath50 _ _ is maximal dipolar energy of two joined dipolar particles , @xmath51 is boltzmann s constant and @xmath52 is the temperature .
thus , dipole - dipole interaction among magnetic holes was dominant over the thermal fluctuations .
the rotating magnetic field induced an aggregation of the microspheres .
the process took place via the joining of single particles into dimers , trimers and formation of 2d clusters consisting of many particles .
these new clusters aggregate again and formed bigger clusters .
a typical aggregation dynamics is shown in fig .
[ fig:2 ] . a few samples with approximately the same layer thickness were investigated .
the volume fractions of particles were low , in the range @xmath53 . in order to analyze the digital images , a c programing language code and open graphical libraries were used .
several thousands of digital pictures have been analyzed in a distributed manner in a computational grid . we have analyzed the motion of individual particles during aggregation using our own tracking algorithms written in the python programing language .
the main advantage of the algorithm that was used is that it can track positions of new clusters which are results of the aggregation .
microspheres inside a magnetic fluid layer witout magnetic field behave as nonmagnetic particles dispersed in fluid .
they perform random _ _ brownian _ _ motion . in this case ,
aggregation events are rare due to the low particle concentration .
thus , in the initial stage of the experiments the microsperes are homogeneously dispersed in the layer of mf and the cluster size distribution is unimodal .
after application of the external magnetic field the microspheres begin to behave as interacting magnetic holes .
they have induced magnetic moments which are oppositely oriented to the external magnetic field .
when the energy of dipole - dipole interaction among two arbitrary spheres is larger than the thermal energy of the spheres , as quantified by the dimensionless interaction strength @xmath48 in the present case , field induced aggregation starts . during the aggregation complex motions of microsperes and clusters consisting of many microspheres were observed .
clusters containing regularly ordered particles were formed and small irregular clusters relatively quickly relax to highly ordered structures . based on the optical observation the complex modes of motion of microsperes and clusters may be classified as ; i ) joining of two clusters together followed by a very slow relaxation of the microsperes in the new cluster into a more ordered structure ; ii ) extremely slowly swivelling of all clusters in the same direction as the rotating magnetic field , followed by packing into a compact disk form ; and iii ) small random motions of the clusters induced by random forces resulting from interactions with the local cluster environment .
( color online ) ( a ) the total number of clusters @xmath54 and the mean ( weight average ) cluster size @xmath55 [ in units of number of spheres ] versus time .
( b ) the scaling function @xmath56 obtained from the cluster size distributions @xmath57 during the time interval @xmath58 s. ] we have observed that clusters of all sizes can join together and form bigger cluster which is the basic feature of cluster - cluster aggregation .
the cluster - cluster aggregation model @xcite predicts the scaling properties of the total number of clusters @xmath54 and mean cluster size @xmath55 .
the total number of clusters is defined as @xmath59 where @xmath57 is number of clusters of size @xmath7 at time @xmath60 .
the mean cluster size @xmath55 is defined as : @xmath61 where @xmath7 is cluster size . in our case
the cluster size @xmath7 is given by the number of particles which belong to the cluster .
the aggregation process in fig .
[ fig:2 ] was studied in more details . in fig .
[ fig:3 ] ( a ) we can see that the total number of clusters @xmath54 and mean cluster size @xmath55 show power law dependencies @xmath62 and @xmath63 . the power - law behaviour was found for the time interval @xmath58 s. the scaling exponents were determined as @xmath64 and @xmath65 . based on dynamic scaling theory @xcite all number of clusters @xmath57 observed during aggregation can be scaled into a single , universal curve or scaling function @xmath66 defined as : @xmath67 it is expected that @xmath68 for @xmath69 .
( color online ) ( a ) the total number of clusters @xmath54 and the mean ( weight average ) cluster size @xmath55 [ in units of number of spheres ] versus time .
( b ) the universal scaling function @xmath68 ( @xmath70 , @xmath71 ) calculated for the time intervall @xmath72 s. ] all the cluster number curves @xmath57 during the time interval @xmath58 s fall onto the single curve shown in fig . [ fig:3 ] . from that the characteristic scaling exponent @xmath73 was found .
we measured several samples , however , other samples behaved in a different manner .
the scaling exponents @xmath0 , @xmath1 , and crossover exponent @xmath5 were different from the results presented above .
typical results for a sample that shows a different type of behaviour are shown in fig .
[ fig:4](a ) .
here the scaling exponents were found to be @xmath74 and @xmath75 .
similarily to the case discussed above , the cluster numbers @xmath57 ( @xmath72 s ) that were measured for this sample could be scaled onto a single curve as shown in fig .
[ fig:4 ] ( b ) , but the scaling exponent @xmath5 was nearly twice as large is in the former case , @xmath76@xmath77 .
also in this case the visible dynamic behaviour was consistent with diffusion limited cluster - cluster aggregation but with clearly different scaling exponents from those above .
the results presented in figs .
[ fig:3 ] and [ fig:4 ] show that the scaling exponents for this system can not be universal . in order to understand this unexpected result
we have investigated the motions of individual clusters in more detail .
( color online ) tracks of particles and clusters in the experiment shown in figs .
[ fig:2 ] and [ fig:3 ] during the time interval @xmath78 s. red lines belong to short tracks that are shorter than 10 time steps ( the time step @xmath79s ) .
blue tracks are longer than 10 time steps .
the sample area is about @xmath36m@xmath37@xmath38m.,width=302 ] the complex motion of a cluster was simplified by considering only the motion of its central mass point .
there are effects that can change the position of the central mass point with nearly no motion of the cluster as a whole .
for example , after joining of two clusters a rearrangement of particles in the new cluster ( see the case i ) discussed above ) takes place .
we assume that these disturbing changes are smaller than the influence of random local forces ( case iii ) ) that essentially contribute to the cluster motions .
a very slow rotation of a cluster ( case ii ) ) does not change the position of the central mass point .
( color online ) tracks of particles and clusters during aggregation in the time interval @xmath78 s. red lines belong to short tracks that are shorter than 10 time steps ( the time step @xmath79s ) .
blue tracks are longer than 10 time steps .
the sample area is about @xmath36m@xmath37@xmath38m.,width=302 ] cluster tracks shown in fig .
[ fig:5 ] ( fig .
[ fig:6 ] ) were determined for experimental data shown in fig .
[ fig:3 ] ( fig .
[ fig:4 ] ) .
we see in figs .
[ fig:5 ] and [ fig:6 ] that the clusters moved in two directions , the tracks are complex and show features of _ _ brownian _ _ motion as expected . for brownian particles
it is characteristic that their motion is well described by @xmath80 where @xmath81 is the distance vector between an initial position and the position after time @xmath60 and @xmath82 is the diffusion coefficient .
we checked the validity of eq .
[ eq:3 ] for the cluster tracks and determined the relation @xmath83 for any cluster of the size @xmath7 . at each experiment we analyzed about @xmath84 tracks and found that this equation is valid with a cluster - size dependent diffusion coefficient @xmath6 .
( color online ) the diffusion coefficients vs. cluster size @xmath7 .
( a ) for the experimental data shown in figs .
[ fig:3 ] and [ fig:5 ] the diffusion coefficient follows @xmath85 .
( b ) for the experimental data shown in fig .
[ fig:4 ] and [ fig:6 ] the diffusion coefficient scales as @xmath86.,width=302 ] we found that @xmath6 clearly depended on the cluster size @xmath7 as shown in fig . [ fig:7 ] . for the case in fig .
[ fig:7 ] b ) the values of the diffusion coefficient @xmath6 fall in a broad range covering nearly three decades .
these ranges of values of @xmath6 are significantly larger than possible errors of measurement .
an average diffusion coefficient @xmath8 was calculated for any cluster size @xmath7 and the results ( fig .
[ fig:7 ] ) were fitted to a scaling law @xmath9 . for the data presented in figs .
[ fig:3 ] and [ fig:5 ] the diffusion scaling exponent @xmath11 was found . on the other hand , for the data in figs .
[ fig:4 ] and [ fig:6 ] the diffusion scaling exponent is clearly higher @xmath12 .
we have observed that application of a rotating magnetic field on a 2d magnetic many holes system causes field induced aggregation .
the clusters move and can grow in both dimensions , which is different from the case of a constant magnetic field where clusters are free to move in both dimensions but only grow in one dimension as determined by the external magnetic field .
the results show that the system is in a nonequilibrium state and its characteristic quantities , number of clusters @xmath54 and average cluster size @xmath55 , develope and show scaling according to a cluster - cluster aggregation model @xcite as was shown in figs . [
fig:3]a ) and [ fig:4]a ) .
both the scaling properties and the broad cluster size distributions found , as well as the existence of a scaling function @xmath66 , are main signatures of cluster - cluster aggregation and dynamic scaling theory @xcite . in many cases the cluster - cluster aggregation mechanism leads to formation of complex , fractal - like objects @xcite . however , in the present case the structure of the aggregates are simpler with a compact internal organization . in relatively strong magnetic fields ,
small clusters are regular 2d objects with well ordered structure of the particles inside the clusters .
extremely slow cluster rotations and rearrangement of particles inside the new , bigger clusters have been observed . as a consequence of these effects
the clusters are packed into regular objects with a nearly close - packed , triangular structure of the spheres .
the cluster diffusion coefficients do not depend on the direction in which the clusters move , i.e. , hydrodynamic corrections are not important as they are in case of constant magnetic fields @xcite . in the basic cluster - cluster aggregation model
@xcite it is assumed that the diffusion coefficient is @xmath87 and the corresponding scaling exponent is @xmath16 .
we have determined two distinct values for the diffusion coefficients @xmath10 and scaling exponents @xmath0 as a result of the two clearly different types of behaviour observed in our experiments .
the relationship between @xmath10 and @xmath0 for both values of @xmath10 follows the equation @xmath88 which has been found in other aggregation models . for @xmath11 ( @xmath12 )
we computed @xmath89 ( @xmath90 ) .
these scaling exponents agree well with exponents @xmath0 determined directly from the time dependence of @xmath55 , @xmath65 and @xmath91@xmath92 respectively .
unfortunately , at present we are not able to explain why similar experiments on approximately the same samples ( concentrations , layer thickness etc . ) show scaling exponents with values that come in two clearly separated ranges and diffusion exponents @xmath10 which are different from the expected value @xmath87 .
thus , the scaling exponent @xmath0 is either clearly lower or higher than the theoretically predicted value @xmath16 . in an earlier study of a similar system of magnetic holes in a constant magnetic fields
@xcite it was found that for small microspheres ( diameters @xmath93 m and interaction strength @xmath94 ) the scaling exponents @xmath0 and @xmath1 were approximately equal @xmath95and typically slightly lower than 0.5 : 0.38@xmath96 . however , for larger particles , @xmath97 m ( @xmath98 ) the values of @xmath0 and @xmath1 increased with the value of the dimensionless interaction strength @xmath99 from about 0.1 to 0.6 , and correspondingly the value of @xmath5 decreased from above @xmath100 to @xmath101 .
thus , depending on the particle size the scaling exponents changed from being nearly constant and near the theoretically expected values to being strongly non - universal .
although the present particles are within the diameter and @xmath99 ranges which showed nearly universal behaviour in ref .
@xcite , the magnetic interactions are very different ( anisotropic in the former and isotropic in the present ) and then the range of @xmath99 for which the behaviour in non - universal , seems to be changed .
it is unclear why the diffusion conditions , as quantified by the values of the diffusion coefficients @xmath10 , were so different in the two typical cases reported here
. it may possibly be related to fine details in the interaction between the microspheres and the glass plates confining the system . in principle
the magnetic holes should be repelled from the confining walls @xcite but if for some unknown reason a small fraction of the particles become attracted or even loosely attached to the walls , this would slow down the diffusion as shown by the anomalous value @xmath102 in one of the analyzed cases . extremely small values of @xmath8 could indicate that some of the particles are trapped in the sample volume or on the sample glass boundary .
diffusion limited cluster - cluster aggregation of magnetic holes has been induced by a rotating magnetic field .
the main features of the experimental results are well described by a diffusion limited cluster - cluster aggregation model and dynamic scaling theory .
the experimental conditions were designed in effort to have a well defined model of a low concentrated many body system where long range interactions are dominant . at present , the reason why two main aggregation regimes were observed is not clear . this resulted in scaling exponent values clearly different from those predicted by theory for systems with short range interactions .
this difference in behaviour was further confirmed by unusual values of the scaling exponent @xmath10 of the diffusion coefficient that were found by cluster tracking .
an open question remains : why do the isotropic , long - range particle - particle interactions suppress the diffusion regime where the size - dependence of the diffusion coefficient scales with @xmath87 ? this will hopefully be clarified in future studies .
the authors thank arne t. skjeltorp for many stimulating discussions .
the experimental part of this work has been done at the institute for energy technology ( ife , kjeller ) .
j.c . thanks for kind hospitality at the physics department at ife .
visual data processing was realized using the results of the projects : negroid and know arc .
we acknowledge financial support from the slovak ministry of education : grant .
this work was supported by the slovak research and development agency under the contract no .
rp eu-0006 - 06 .
|
we have experimentally investigated field induced aggregation of nonmagnetic particles confined in a magnetic fluid layer when rotating magnetic fields were applied . after application of a magnetic field rotating in the plane of the fluid layer , the single particles start to form two - dimensional ( 2d ) clusters , like doublets , triangels , and more complex structures .
these clusters aggregated again and again to form bigger clusters . during this nonequilibrium process
, a broad range of cluster sizes was formed , and the scaling exponents , @xmath0 and @xmath1 , of the number of clusters @xmath2and average cluster size @xmath3 were calculated .
the process could be characterized as diffusion limited cluster - cluster aggregation .
we have found that all sizes of clusters that occured during an experiment , fall on a single curve as the dynamic scaling theory predicts .
hovewer , the characteristic scaling exponents @xmath4 and crossover exponents @xmath5 were not universal .
a particle tracking method was used to find the dependence of the diffusion coefficients @xmath6 on cluster size @xmath7 .
the cluster motions show features of _ _ brownian _ _ motion .
the average diffusion coefficients @xmath8 depend on the cluster sizes @xmath7 as a power law @xmath9 where values of @xmath10 as different as @xmath11 and @xmath12 were found in two of the experiments .
|
warped geometries , such as the background used in randall - sundrum model @xcite , have provided many insights into general relativity and holography
. however , it is very difficult to find exact solutions to einstein s equations , so only a handful of warped geometries are known . it would be useful to have a tool for constructing theories that reproduce the features of the warped geometries without having to find and stabilize an appropriate gravitational source .
discrete gravitational dimensions could be such a tool @xcite .
if we work in the effective field theory framework with a cutoff , we may be able to learn a lot about general realtivity without needing exact solutions , as long as the discrete theory can be trusted .
of course , this will not tell us about the energy - momentum tensor to generate such a background , but we can use the discrete model as a tool for investigating stability of the system , the existence of ghosts , and the strong coupling scale , for example . in this paper , we study the the discretization of a single extra dimension in which we impose the exponential warp factor of the randall - sundrum model . although previous work discussed discrete nongravitational extra dimensions for rs @xcite , so far gravity has not been included in the discretized model .
we will find that many of the features of the continuum gravitational theory are correctly reproduced , and that some of the problems with flat space lattice gravity are ameliorated .
the discretized model has a number of sites , each of which has a separate four - dimensional metric , and , in the minimal case , has only nearest neighbor links . in the case of flat space , it has been shown that the there is a limit to how small the lattice spacing can be that depends on the overall size of the lattice .
this implies that there is no good continuum limit and it is impossible to reproduce 5d gravity all the way up to the five dimensional planck scale .
the absence of the continuum limit in the flat case could be connected to the absence of a simple local holographic description .
it seems reasonable that a four - dimensional discretized theory should exist for the warped case , which does have a dual lower - dimensional description .
the flat space strong coupling problems result from two properties of the kaluza - klein spectrum : there is a very light mode and that mode couples equally over the entire theory space .
both of these properties are the result of the discretized kinetic term , which introduces large mixing between the additional scalars that were originally associated with a single site .
the mixing leads to a light highly delocalized mode , whose strong interactions invalidate a low - energy description at a strong coupling scale that lies well below the top of the kk tower . this is a less severe problem in the warped geometry , because the strong warp - factor dependence in the kinetic term keeps the modes localized on only a few sites .
the absence of strong mixing means that the localized modes are not lighter than you would naively expect .
however , the warp factor only protects against mixing of modes that are localized on sites further apart than the ads curvature scale .
when the theory is discretized on smaller scales , the theory resembles flat space . within the ads curvature scale ,
modes do mix and get delocalized .
so although it is possible to reproduce many of the features of the continuum theory with the discretized version , it is still not possible to achieve the true continuum limit .
nevertheless , we find that the dangerous uv / ir problem is absent : the lattice is self - consistent for any number of sites .
we will be concerned with the poincare patch of @xmath2 , which is described by the metric @xmath3 as in rs , we impose orbifold boundary conditions at @xmath4 and @xmath5 .
the 5d gravitational lagrangian in this background is @xmath6 @xmath7 .\ ] ] we are interested in linear fluctuations so we expand @xmath8 . then , the terms quadratic in @xmath9 , after some integration by parts , are @xmath10 + e^{- 4 k y } \left [ ( \partial_y h_{\mu \nu } ) ^2 - ( \partial_y h ) ^2 \right ] \right\}\label{contlag } .\ ] ] to discretize the fifth dimension
, we choose points evenly spaced in the @xmath11-coordinate : @xmath12 where @xmath13 is the number of lattice sites and @xmath14 is the lattice spacing .
this is not the only possibility , and discretizing in a different coordinate may result in a drastically different lattice theory .
however , @xmath11 is in some sense the natural choice , since it is a geodesic coordinate for the poincare patch ( unlike , say , the conformal coordinate @xmath15 ) . but more importantly , we can justify using @xmath11 _ a posteriori _ because this discretization will lead to a lattice theory with some of the holographic features of ads that we expect to see .
we replace derivatives in the @xmath11 direction by differences . for example @xmath16
this brings out additional ambiguities related to the evaluation of the warping prefactor , and to boundary terms which appear if we integrate by parts before discretizing .
however , again , our choices will be justified _ a posteriori_. in any case ,
if the lattice theory is to be trusted , we should expect that these ambiguities are irrelevant in the continuum limit , which we discuss below .
thus we get @xmath17 , \ ] ] with @xmath18 the effective 4d planck scale on the sites . going to canonical normalization @xmath19
we get a standard lattice action @xmath20 with mass matrix @xmath21 .\ ] ] note that this is essentially the same mass matrix as for a gauge boson @xcite . at this point
it is helpful to think about the values for @xmath22and @xmath23 that we would like to study , which are determined by the continuum limit in which we are interested . as in flat space
, we would like to know whether we can have @xmath24 in such a way that the lattice matches the continuum between any two scales @xmath25 and @xmath26 . in the case of the compact randall - sundrum model relevant for the standard model , which has @xmath27 and @xmath28 , the limitation @xmath29 implies that @xmath30 in this particular case .
this is simply because in rs1 the fifth dimension is only about 30 planck units wide .
nonetheless , discretization on the scale of @xmath31 would be sufficient in principle to achieve a theory that is valid up to the scale @xmath23 , the same scale as the continuum limit .
in flat space , large @xmath13 allows you to make a theory that is valid deeper in the infrared .
this is also true for the rs model , but only by taking a larger volume " ( that is _
r _ ) . in this case
, @xmath26 would be smaller .
we will allow @xmath32 , @xmath13 , and @xmath33 to be free parameters , with the understanding that @xmath33 will always be greater than @xmath31 and @xmath13 will be taken as large as necessary to achieve any desired ir cutoff .
in this section , we will develop insight into the mass eigenstates of @xmath34 .
we are working in the regime with @xmath35 and @xmath36 so that the space is highly curved .
first , we make some quick observations about the mass matrix that will uncover the important qualitative features of the warped lattice .
we will afterwards perform a more careful analysis . since @xmath37 ,
the mass matrix is crudely approximated by @xmath38 \label{crude } .\ ] ] we have set @xmath39 but not @xmath40 , since @xmath41 may be large .
this mass matrix scales with position and is almost diagonal .
the only non - diagonal entries are next to the diagonal , because of the nearest - neighbor lattice structure we have assumed .
but since the position dependence of the warp factor is exponential , the mass eigenvalues will be geometrically spaced : @xmath42 we confirm this geometric spacing numerically in figure [ figevals ] .
this geometric spacing of the masses is very different from the linear spacing of eigenvalues for the kk modes in rs .
however , the linear rs spacing is deceptive , since there are only of order @xmath43 relevant at any site .
the remaining modes are above the local cutoff , and are in fact localized further in the uv .
that is , even though the classical kk tower of rs has an infinite number of linearly spaced modes , no calculation should ever involve more than @xmath43 of them .
in fact , the linearized rs regime does appear , and the geometric spacing is actually quite natural , reproducing the modes localized throughout the bulk with masses determined by the local ( warped ) ads scale . to see this , consider the elements of this matrix around @xmath44 for some fixed site @xmath45 . from eq .
( [ crude ] ) , @xmath46 . \label{flatapp}\ ] ] for @xmath47 the @xmath48 term is approximately @xmath49 .
then in this region the mass matrix looks like flat space .
so the eigenstates @xmath50 ( for this @xmath45 ) are roughly @xmath51 with eigenvalues @xmath52 for example , if we look at the lightest modes , with @xmath53 they are linearly spaced just as in rs .
this linear spacing breaks down around @xmath54 .
this is exactly what we expect because the @xmath55 mode has a mass of the local planck scale .
the linear spacing for the low modes is evident in figure ( [ figevals ] ) . in the approximation ( [ flatapp ] )
, the prefactor @xmath56 implies that there will be little mixing between modes from an expansion around different sites @xmath57 and @xmath58 if @xmath59 .
the linear spacing is reconciled with the geometric spacing which is transparent in the exponential prefactor , because the linear spacing appears only when the exponential is roughly constant .
we will see that the lack of mixing between modes localized more than of order @xmath60 apart is the reason for the larger regime of validity of the warped discretization , and allows @xmath13 and @xmath61 to be taken as large as desired .
the interesting thing about this way of approximating the modes is that it applies for modes centered around _
site @xmath45 .
of course , this is not surprising because of the conformal scaling of rs as we move from the ir to the uv .
but it is remarkable that we can get information about a dynamical feature of rs , that the relevant modes change with energy , from a fixed lattice .
indeed , the modes near a site @xmath57 and those near a site @xmath58 decouple , if @xmath62 .
this is very different from what happens to the kk modes @xmath63 of the continuum rs .
if we take two kk modes with masses matching the lattice modes , we find that they do have significant overlap in even if their masses are very different .
of course , the kk picture is not valid at all energies , so we should never be considering two widely spaced modes at the same time . on the lattice we can just use all the modes , and the widely spaced ones naturally decouple from each other .
although to use this rough flat - space approximation , we have to concentrate on a single site or a single mode , we will see that with a more careful analysis of the lattice eigenstates , the same qualitative features still hold . even though this flat space approximation tells us most of what we want to know about the lattice theory , it is instructive to have a representation of the mass eigenstates in the region away from where they have most of their support .
this both justifies our approximation , and will be used in the strong coupling calculations below .
we want to develop a better understanding of the mass eigenstates of the lattice theory .
we are looking for the eigenstates of @xmath21 , \ ] ] for @xmath64 .
let us define a matrix @xmath65 which goes from position eigenstates to mass eigenstates @xmath66 we have already observed from the flat space approximation , that @xmath67 for @xmath68 .
now we will confirm this and also approximate the @xmath69 in other regions .
one approach is to use the bulk kaluza - klein modes to guide the search for the @xmath70 .
the kk equation following from the lagrangian ( [ contlag ] ) , with @xmath71 , is @xmath72 it is not hard to see that the equation ( [ contkkeq ] ) maps directly onto the eigenvalue equation for @xmath34 .
the small parameter @xmath73 is the step size for the dimensionless continuum variable @xmath74 .
then , expanding the derivatives , ( [ contkkeq ] ) becomes @xmath75 - \frac{4 k } { ( k a ) } \frac{1}{2 } \left [ \chi ( \hat{y } + k a ) - \chi ( \hat{y } - k a ) \right ] + m_n^2 e^{2 \hat{y } } \chi ( \hat{y } ) = 0 .
\label{kkexp}\ ] ] for the first derivative , we have taken two steps to symmetrize with respect to @xmath76 . on the other hand , we can invert ( [ htobigh ] ) @xmath77 we have used the fact that @xmath70 and @xmath78 are canonically normalized , so @xmath79 is unitary
. then , hitting @xmath70 with @xmath34 , and projecting out the @xmath80 component we find @xmath81 - m_n^2 \eta^n_j = 0 \label{eveq}.\ ] ] for small @xmath73 this becomes @xmath82 + e^{2 k a ( j - 1 ) } m_n^2 \eta_j^n = 0 , \label{smallka}\ ] ] which is the same as ( [ kkexp ] ) for @xmath83 . in the continuum ,
the solutions to ( [ contkkeq ] ) are kk modes : @xmath84 the corresponding discrete modes are @xmath85 we can not really manipulate these discrete bessel functions , but we do expect them to satisfy approximately the same relations as the continuum bessel functions .
for example , the relation @xmath86 implies @xmath87 then it is rather trivial to show the modes are eigenstates
they must be because the equations match up .
but note that matching the continuum solutions onto the lattice works because we are using only the first term in the taylor series , in going from ( [ contkkeq ] ) to ( [ kkexp ] ) .
this is , in fact , all we can do , because we have only included nearest - neighbor links in the discretization . and
that means that the solution will break down when the first derivative is no longer a good approximation , i.e. when @xmath88 so these solutions should be good until @xmath89 .
in fact , the exact eigenstates of this matrix really look like the continuum kk bessel functions only in this limited regime . the lattice does not just take a selection of the kk modes , but presents a fundamentally different structure .
however , this structure actually reproduces the correct physics , as the heavier kk modes in the continuum theory would be cutoff by the local strong coupling scale .
the bessel solutions apply for small @xmath90 .
that is , they approximate the light kk modes at positions corresponding to energies above their mass .
the bessel functions peak and are localized when @xmath91 which defines @xmath45 .
so , the bessel approximation is good through the localization region .
we also see that the width of the peak is determined by the width of the exponential , that is @xmath92 , so the region where the bessel function peaks is the same as the region where the flat space approximation applies . in the flat space region
, we can take @xmath93 which satisfies the small @xmath94 equation , following from ( [ smallka ] ) : @xmath95 with the eigenvalues @xmath96 .
these are not quite the same flat space solutions as the ones derived in the previous subsection from looking at the modes around a particular site .
the difference is that in the rough approximation , the heavier modes have higher frequency , because they are the excited states of a box around a particular site .
we have seen here that the actual eigenstates all have the same frequency because each one is effectively the zero mode of a different box , centered around the site where the mode is localized . nonetheless , the important point is that modes of both approximations have the same qualitative features : oscillating behavior with roughly constant amplitude over around @xmath97 sites .
now we have solutions for small @xmath41 , in the bessel regime , and for @xmath98 , in the flat space regime . for large @xmath41
, so that @xmath99 the solutions are @xmath100 we can check @xmath101 @xmath102 so in this regime @xmath103 and thus ( [ eveq ] ) reduces to @xmath104 which is satisfied by our ansatz .
in summary , the solutions for the eigenstates are ( up to normalization ) @xmath105 to align and normalize the solutions , note that the bessel solution is exponentially growing @xmath106 and in the third , large @xmath41 , regime , @xmath69 dies as @xmath107 .
so in both of these regimes , the modes have most of their support towards the middle .
this means that the normalization is determined almost entirely from the flat space approximation .
thus we are led back to our original guess .
we only need the flat space approximation .
if we are looking near a site @xmath45 associated with a mode @xmath108 then @xmath109 some exact eigenvectors are plotted in figure [ figexact ] , and the approximation ( [ bigapprox ] ) is shown in figure [ figapprox ] .
now that we understand the eigenstates of the lattice theory , we can ask whether there is any limit in which the lattice is a good approximation to the continuum .
first , let us recall some difficulties that are encountered in flat space @xcite . to trust the predictions of the lattice theory , it is essential to know the scale at which it becomes nonperturbative and the effective theory breaks down . in a gravitational system , as in a gauge theory , the easiest way to find this scale is to introduce goldstone bosons restoring a non - linear symmetry , _ a la _ ccwz @xcite .
this procedure is explained in detail in @xcite , and we will only briefly review it here . for a massless graviton , this means restoring general coordinate invariance with vector ( @xmath110 ) and scalar ( @xmath111 ) goldstones : @xmath112 @xmath113 then the fierz - pauli mass term generates a kinetic term for the scalar @xmath114 the two four - derivative terms for @xmath111 cancel after integration by parts , leaving a proper two - derivative kinetic term for @xmath111 after the @xmath115 kinetic matrix is diagonalized .
the @xmath111 self - interactions are the strongest and indicate where the theory breaks down .
the scales are @xmath116 so the canonically normalized fields are @xmath117 and @xmath118 leading to @xmath119 from which we read that the cutoff is @xmath120 . on the flat space lattice
, there are many massive gravitons .
so we introduce many goldstones via @xmath121 .
then @xmath122 .\ ] ] barring any unforeseen cancellations the strong coupling scale is determined by the interactions of the lightest mode . on the lattice , the effective higher - dimensional modes ( the kk modes ) get contributions from all the site modes , so there is a corresponding @xmath123 suppression of the interactions .
for example , with @xmath124 @xmath125 so the planck scale associated with the sites is lower by a factor of @xmath126 than the planck scale of the modes .
for the gravity case , there is an additional @xmath13-dependence for the scalar longitudinal modes because they get their kinetic term from mixing with differences .
then , @xmath127 thus with @xmath128 we get @xmath129 .
this means that the strongest interaction , from the lightest mode , is @xmath130 where @xmath131 and @xmath132 .
thus the strong interaction for the flat space lattice is at a scale @xmath133 since this @xmath134 is the low energy planck scale , and @xmath135 is the physical mass , we can see that this scale is higher by a factor of @xmath136 than that of a single massive graviton .
the fact that the strong coupling scale is heavier than the mass of the light mode tells us that this is a good effective theory for the light mode .
but for a real lattice description of flat space , this must be a good effective theory all the way up to the 5d planck scale , @xmath137 .
since @xmath138 we can write the cutoff as @xmath139 . then for @xmath140 we would need @xmath141 .
but since @xmath142 this means that @xmath143 the lattice spacing has to be much smaller than the planck length .
but this does not make sense .
for the lattice to cut off the divergences from gravity , we must take @xmath144 . another way to understand
the difficulty is to observe that with the lattice spacing at its limit , @xmath145 , we get @xmath146 .
so there will be new effects at distances much much larger than the planck length .
these non - local effects are most apparent if we work directly in the continuum limit @xmath147 an integration by parts has been performed on the middle term .
this shows that @xmath148 is the propagating field , while @xmath111 without derivatives is producing the strong interactions . in terms of @xmath149
the lagrangian contains @xmath150 the long wavelength modes with @xmath151 interact at the scale @xmath152 we derived above .
but we can also now see that the strong interactions are really non - local in the extra dimension .
formally , as @xmath153 ( or @xmath154 ) , the strong coupling problem disappears , as it must if this lattice is to classically reproduce the continuum .
however , @xmath155 does not grow with @xmath13 fast enough to ensure that the effective theory is consistent . now let us see how these observations change in the warped background .
we introduce goldstones into the warped space lagrangian in the usual way @xmath156 .\ ] ] the warp factor in front of the kinetic term tells us that the planck scale on a site is @xmath157 this is the warping we expect from the continuum . for the kk modes
, we saw that to a good approximation @xmath158 modes have support at each site so we expect this parameter to play the role of @xmath13 in the previous section .
so the effective planck scale for the modes will be @xmath159 in agreement with the observation that @xmath158 in the warped case should play the role of @xmath13 .
in fact , because we are approximating the kk modes for warped space by mapping to a position - dependent set of flat space solutions , we can simply use the flat space results if we just complete the map . looking at the lagrangian , in the approximation of section [ secrough ] , we see that the warped space lagrangian for the @xmath97 modes around mode @xmath108 is equivalent to a flat space lattice with @xmath160 thus the strong coupling scale becomes @xmath161 where @xmath162 is the mass of the lowest mode in the expansion around site @xmath41 .
this result is superficially similar to that of flat space .
the strong coupling scale is a factor of @xmath163 above the strong coupling scale for a single massive graviton .
but there is a huge difference
the strong coupling scale does not depend on the size of the space we are latticizing .
there is no dependence on the total number of lattice sites @xmath13 , or equivalently on the ir scale @xmath164 .
in flat space , the strong coupling was determined by the lightest mode , but in warped space , it is determined by the lightest mode with support on the site @xmath41 . for a @xmath41 close to the uv brane , all the modes which live there are much heavier than the @xmath165 mode which would dominate if the space were flat .
however this is not the whole story . in warped space
, there is not a single strong coupling scale ; the strong coupling scale depends on the observer .
in fact , there is an important difference between the scale associated with a particular mode and the scale that an observer on a particular site would see . for a mode of mass @xmath166 ,
the strong coupling scale is the usual @xmath167 this scale is heavier than @xmath166 , so the mode is weakly coupled at energies near its mass .
however , an observer at site @xmath41 would see @xmath97 modes .
in particular , the mode of mass @xmath166 would be relevant even at energies as high as @xmath168 .
when looking at the strong coupling scale on a site we must use @xmath169 of equation ( [ lwarp ] ) .
the relevant planck scale for the observer is @xmath170 .
for example , with a lattice spacing @xmath171 we would find that @xmath172 so , if we only talk about modes , there is no strong - coupling problem in warped space .
but observers localized at some position in the bulk ( that is , on a particular lattice site ) must encounter strong coupling before the local planck scale . in asking about the lattice theory
, it makes sense to consider the site basis , since otherwise the cut - off is determined solely by the ir regime .
we can understand these results in continuum language as well . in the warped case , @xmath173 when we integrate by parts
, the @xmath174 term hits the warp factor , so there are two pieces @xmath175 initially , we might expect that for the long wavelength modes , with @xmath176 , the @xmath177 piece will dominate and prevent nonlocal effects .
however , this would work only if there were modes spread out over the whole space . in warped space
, there are not really any @xmath165 modes .
the wavelengths in the extra dimension are in fact limited by @xmath178 at each site , there are only @xmath97 modes .
so @xmath179 and we can basically ignore the @xmath177 term in ( [ k ] ) . using @xmath180 , and adding the @xmath181 volume factor , the lagrangian becomes @xmath182 the canonical propagating field is then @xmath183 and the cubic scalar coupling becomes @xmath184 for a mode of frequency @xmath185 this gives a strong coupling scale of @xmath186 if we look at the individual modes , then @xmath187 ( cf eq ( [ simpsum ] ) ) , and so the mode scale is @xmath188 which is the same as @xmath189 that we derived above .
an observer at position @xmath11 is sensitive to wavelengths as high as @xmath190 , which gives @xmath191 this is the same as the flat space interaction scale @xmath192 with @xmath193 the playing the role of the ir cutoff for the warped space .
we have shown that a straightforward discretization of the warped ads geometry produces a low - energy theory which is valid above the scale of local curvature at any site .
this is in fact sufficient for investigating many physical features , such as the renormalization group behavior of the theory as in @xcite .
however , the discretization does not allow us to reach the uv cutoff of the continuum theory . in the energy regime between the warped ads scale and the warped higher - dimensional planck scale ,
the theory acts like flat space and the same strong coupling problems as in flat space appear .
this is not surprising ; at energies above the local curvature , the theory approaches flat space , and therefore we do not expect manifestly holographic behavior .
but it is extremely interesting that some of the holographic behavior of warped space is also manifested on the lattice ; for example , we have found that the mass eigenvalues are globally geometrically spaced , with only @xmath43 modes localized near any particular site .
an important distinction from the flat space theory is that we can take the large volume limit because the uv and ir cutoffs are independent . in flat space
, the uv cutoff goes down as the size of the space , @xmath61 , is increased . in curved space , the uv cutoff only depends on the curvature scale , @xmath32 ; the cutoff in a particular region is completely ignorant of the total volume , or equivalently , the total number of sites .
this cutoff is still below the local planck scale , so even in the warped case , the lattice can not be used as a regulator
divergences must still be cutoff by hand , or new physics must enter at a scale below the lattice spacing .
as in any non - renormalizable theory with a dimensionful scale , there is a natural limit to the lattice spacing , and thus a natural limit to the number of sites on the lattice . in warped space , this is a particularly strong bound .
for example , in rs1 @xcite , the dimensionful scale is @xmath134 , but the size of the space is set by @xmath194 and so @xmath195 . nonetheless , on a larger warped space , approaching rs2 @xcite , @xmath196 effects can be parametrically ignored .
note that it is precisely because we can take the large volume limit with fixed lattice spacing that large @xmath13 is interesting .
in flat space , because the uv cutoff decreases as the volume grows , in the large volume limit the cutoff goes to zero . in warped space ,
it remains above the local curvature scale .
because the discretization of rs is sufficiently well behaved , we expect a similar discrete theory may be a useful tool for studying other geometries that exhibit holographic behavior . for example , the metric for de sitter space and some black holes can be written in a warp - factor notation @xcite . thus their discretizations should be similar and may help unravel their holographic features .
another interesting example is local localization @xcite , in which the warp factor decreases to a minimum and then grows again .
we expect that sites in the turnaround regime will have a low cutoff .
however , modes localized on these sites will have very little support in the region where four - dimensional gravity applies .
there is no reason to expect the massive mode of locally localized gravity that plays the role of the massive graviton to cause problems .
presumably locally localized gravity is far more general .
the discrete version of theories with non - monotonic warp factor could be a useful tool for studying different examples of locally localized gravity , even with more than one extra dimension .
we would like to thank j. gallicchio and i. yavin for discussing their results @xcite .
we would also like to thank y. shadmi and r. mahbubani for discussions .
after this work was published , we were informed that a related discussion can be found in @xcite .
the work of lr was supported in part by nsf award phy-0201124 .
rs1 l. randall and r. sundrum , phys .
lett . * 83 * , 3370 ( 1999 ) [ arxiv : hep - ph/9905221 ] .
l. randall and r. sundrum , phys .
lett . * 83 * , 4690 ( 1999 ) [ arxiv : hep - th/9906064 ] .
n. arkani - hamed , a. g. cohen and h. georgi , phys .
* 86 * , 4757 ( 2001 ) [ arxiv : hep - th/0104005 ] .
n. arkani - hamed , h. georgi and m. d. schwartz , annals phys . * 305 * , 96 ( 2003 ) [ arxiv : hep - th/0210184 ] .
n. arkani - hamed and m. d. schwartz , phys .
d * 69 * , 104001 ( 2004 ) [ arxiv : hep - th/0302110 ]
. m. d. schwartz , phys .
d * 68 * , 024029 ( 2003 ) [ arxiv : hep - th/0303114 ] .
a. falkowski and h. d. kim , jhep * 0208 * , 052 ( 2002 ) [ arxiv : hep - ph/0208058 ] .
a. katz and y. shadmi , jhep * 0411 * , 060 ( 2004 ) [ arxiv : hep - th/0409223 ] .
t. bhattacharya , c. csaki , m. r. martin , y. shirman and j. terning , arxiv : hep - lat/0503011 . o. andreev and w. siegel , phys .
d * 71 * , 086001 ( 2005 ) [ arxiv : hep - th/0410131 ] .
a. pomarol , phys .
b * 486 * , 153 ( 2000 ) [ arxiv : hep - ph/9911294 ] .
l. randall and m. d. schwartz , phys .
lett . * 88 * , 081801 ( 2002 ) [ arxiv : hep - th/0108115 ] .
l. randall and m. d. schwartz , jhep * 0111 * , 003 ( 2001 ) [ arxiv : hep - th/0108114 ] . w. d. goldberger and i. z. rothstein , phys .
lett . * 89 * , 131601 ( 2002 ) [ arxiv : hep - th/0204160 ] .
a. karch and l. randall , phys .
lett . * 87 * , 061601 ( 2001 ) [ arxiv : hep - th/0105108 ] .
|
we investigate the discretized version of the compact randall - sundrum model . by studying the mass eigenstates of the lattice theory ,
we demonstrate that for warped space , unlike for flat space , the strong coupling scale does not depend on the ir scale and lattice size .
however , strong coupling does prevent us from taking the continuum limit of the lattice theory .
nonetheless , the lattice theory works in the manifestly holographic regime and successfully reproduces the most significant features of the warped theory .
it is even in some respects better than the kk theory , which must be carefully regulated to obtain the correct physical results .
because it is easier to construct lattice theories than to find exact solutions to gr , we expect lattice gravity to be a useful tool for exploring field theory in curved space .
ucb - pth-05/18 + lbnl-57913 + * discretizing gravity in warped spacetime * lisa randall@xmath0 , matthew d. schwartz@xmath1 , and shiyamala thambyahpillai@xmath0 +
|
the pyrochlore titanates , with formula r@xmath0ti@xmath0o@xmath1 where r is a rare earth , have been the subject of experimental and theoretical studies for more than a decade @xcite .
the pyrochlore lattice where the r@xmath2 ( and the ti@xmath3 ) ions are located is formed by corner - sharing tetrahedra and leads indeed to a frustration of the exchange / dipolar interaction in some specific situations .
the best known consequence thereof is the existence of `` spin - ice '' materials , like ho@xmath0ti@xmath0o@xmath1 and dy@xmath0ti@xmath0o@xmath4 @xcite , where the rare earth moments have a strong ising character along the tetrahedron ternary axis .
the spin - ices remain in the paramagnetic phase down to the lowest attainable temperature , but the spin correlations are very strong and of the special `` two in - two out '' type , i.e. where two rare earth moments point `` outwards '' a given tetrahedron and two inwards . the excitations in spin - ices have been shown to be magnetic monopole quasi - particles @xcite . another pyrochlore material which remains paramagnetic down to at least 0.05k is tb@xmath0ti@xmath0o@xmath1 @xcite . the tb@xmath2 ion is a non - kramers ion with @xmath5=6 and a land factor @xmath6=3/2 . it is submitted to a trigonal symmetry crystal electric field ( cef ) which lifts the 13-fold degeneracy of the @xmath5=6 multiplet .
the peculiarity of the crystal field splitting in tb@xmath0ti@xmath0o@xmath1 has been early recognised @xcite , as consisting of two ground magnetic doublets separated by an energy of the order of 15k . for a non - kramers ion , each such doublet \{@xmath7 }
has an ising character , with a vanishing transverse matrix element of the total angular momentum : @xmath8=0 .
the exact wave - functions of these doublets were determined @xcite together with the exchange contribution to the paramagnetic curie temperature @xmath9 , which is negative and of the order of @xmath1010k @xcite , indicating dominant antiferromagnetic superexchange interactions .
the lack of magnetic order in tb@xmath0ti@xmath0o@xmath1 is therefore surprising since the ising antiferromagnet in the pyrochlore lattice is not frustrated , its ground state being the so - called `` all in / all out '' configuration .
early neutron scattering studies @xcite have shown that the tb moments are very short range correlated and fluctuate down to the 0.1k range . calculations of the diffuse scattering at 9k @xcite were successful in reproducing experimental data by taking into account the two ground doublets which are both appreciably populated at this temperature . at very low temperature , when the ground doublet alone is populated , it was recognised that ising - like wave - functions can not describe the physics in tb@xmath0ti@xmath0o@xmath1 @xcite , and that a non - zero transverse matrix element of * j * is needed .
in order to restore transverse spin fluctuations , a model was proposed which renormalises the low energy spin hamiltonian through quantum fluctuations between the two ground doublets @xmath11 virtual excitation of a third ion @xcite .
it will be referred to in the following as the `` virtual crystal field '' ( vcf ) model .
it results in an effective hamiltonian containing non - ising terms , but the ground state predicted by the vcf model can not be the true ground state of tb@xmath0ti@xmath0o@xmath1 since it retains magnetic ordering and thus is not a spin - liquid state @xcite . in zero magnetic field , the spin liquid state seems to be fragile : magnetic ordering appears under pressure @xcite , showing the great sensitivity of the spin liquid state to external stresses .
it is also sensitive to the exact tb stoichometry and may be destroyed by a slight excess in tb content @xcite : small bragg peaks are observed at ( @xmath12 ) positions and equivalents in non - stoichiometric powder samples , whereas diffuse maxima are seen at the same positions in single crystals @xcite .
long range magnetic order develops in the presence of a magnetic field . for a field along [ 1@xmath130 ] ,
a magnetically ordered phase with spin wave excitations is induced at 0.4k for fields of 2 t and above @xcite , and the evolution of the magnetic structure was more quantitatively explored in ref .. for a field along [ 111 ] , a similar behaviour is observed @xcite . of particular significance for the understanding of the spin - liquid ground state are the numerous experimental evidences of either a dynamic symmetry lowering , by low temperature x - ray diffraction @xcite , optical measurements @xcite and inelastic neutron scattering @xcite , or of a strong spin - phonon coupling : the giant magnetostriction @xcite , the low temperature divergence of the bulk modulus @xcite and of the elastic constants @xcite .
very recent high resolution neutron scattering experiments @xcite have revealed the presence of non - standard hybridisation between an acoustic phonon branch and the first cef transition , above an energy around 1.5mev , but with a vanishing low energy branch .
they have also confirmed the presence of dispersive inelastic excitations around an energy of 0.2mev , but only in the so - called `` spin ice channel '' .
finally , an anomalously strong phonon scattering at 0.3k has been reported @xcite , very likely linked with the @xmath14-electron - phonon hybridisation .
hence , the ground state in tb@xmath0ti@xmath0o@xmath1 appears to be of a quite complex and unusual kind and a full microscopic hamiltonian describing this state is still to be discovered .
this hamiltonian should include a coupling between the electronic and phononic degrees of freedom , so that the ground state be of mixed vibronic type .
previously , we proposed a model to account for the spin liquid behaviour in tb@xmath0ti@xmath0o@xmath1 , assuming the presence of a static tetragonal distortion at the tb site @xcite , i.e. of an off - diagonal crystal field term which lifts the degeneracy of the ground doublet into two cef singlets , coupled by the exchange / dipole interaction .
this model , akin to the bleaney model for cef singlets @xcite , yields a strong transverse matrix element of * j * and leads to a mean field paramagnetic phase , or spin liquid phase , down to the lowest temperature in a certain range of exchange integrals .
this picture has been questioned in ref .
, where it is claimed that no inelastic line corresponding to the splitting between singlets is observed , but the recent neutron data @xcite do show the presence of some kind of low energy inelastic excitation . on the one hand , this model satisfactorily reproduces the * q*-maps of inelastic @xcite and diffuse spin - flip and non - spin - flip @xcite neutron scattering at very low temperature . on the other hand , due to the non - magnetic character of its mean field single ion ground state
, this two singlet model fails to account for the strong intensity of the quasi - elastic neutron scattering @xcite and for the nuclear anomaly in the specific heat @xcite showing the presence of a hyperfine field of 142 t at the tb site .
seemingly , the vibronic ground state in tb@xmath0ti@xmath0o@xmath1 has some magnetic character . in this work ,
we are interested in the interpretation of the zero - field diffuse scattering and of the isothermal magnetisation for fields applied along [ 111 ] , [ 110 ] and [ 001 ] at very low temperature in tb@xmath0ti@xmath0o@xmath1
. we present zero - field and in - field 3d maps of the total diffuse neutron scattering at 0.16k and we use previously published magnetisation data @xcite .
we apply a generalisation of the two singlet model to a dynamic jahn - teller ( jt ) coupling , which seems better adapted , in first approximation , to describe the low temperature state of tb@xmath0ti@xmath0o@xmath1 since no static distortion has been evidenced . in a recent work @xcite , we have shown that the magnetic structure induced by a field along [ 111 ] can be explained in terms of such a dynamic jahn - teller coupling . here , we show that the model also satisfactorily reproduces : ( i ) the * q*-structure of zero - field diffuse neutron scattering maps at 0.16k and ( ii ) the magnetisation curves at 0.05k along the 3 cubic symmetry axes .
the trigonal crystal field acting on a tb@xmath2 ion is taken as in ref .. we emphasize that , contrary to theories describing the ground non - kramers doublet by a pseudo - spin 1/2 and using effective exchange parameters @xcite , we consider the whole cef level scheme and the actual momentum * j * , and we use the bare anisotropic exchange integrals @xmath15 .
we feel it important to recall that a non - kramers doublet @xmath16 presents the two properties , in the local frame : @xmath17 then , considering the projection of the cef hamiltonian onto the subspace spanned by @xmath16 can lead to some confusion since the z - components alone of the pseudo - spin behave as those of a true spin 1/2 , but its transverse components must be chosen as quadrupolar moments @xcite .
an important consequence of ( [ nonk2 ] ) is that transitions between @xmath18 and @xmath19 induced by the operator * j * are forbidden : there can be neither exchange / dipole induced fluctuations nor neutron quasielastic scattering within the doublet .
the magneto - elastic interaction describing the coupling of the electronic variables with the local strain variables @xcite has been shown to play an important role in tb@xmath0ti@xmath0o@xmath1 @xcite .
it can be written , in the local frame , up to second order in @xmath20 : @xmath21 where the @xmath22 variables are normalised strains , the @xmath23 variables are @xmath14 quadrupole operators and the @xmath24 are coupling coefficients , @xmath25 spanning the relevant symmetry allowed representations . in zero magnetic field , zero external strain and in the absence of any symmetry lowering transition , the mean equilibrium values @xmath26 vanish . in the presence of a field ,
the @xmath27 become non - zero and describe the parastriction .
the magneto - elastic coupling ( [ hmel ] ) is included in the calculation of the magnetisation curves @xcite described in section [ magh ] . as mentioned in the introduction ,
there is experimental evidence for hybridisation of a phonon branch with a cef transition in tb@xmath0ti@xmath0o@xmath1 at very low temperature @xcite .
the whole vibronic problem , with mixed phonon - electron wave - functions , should thus be treated by expanding in ( [ hmel ] ) the strain variables in terms of phonon operators . however , in this work , we approximate the vibronic problem by a dynamic jahn - teller ( jt ) effect , i.e. we introduce a dynamic distortion @xmath28 along the equiprobable tetragonal cubic axes [ 100 ] , [ 010 ] or [ 001 ] , hence conserving overall cubic symmetry .
a hamiltonian of type ( [ hmel ] ) is added to the cef interaction , defined by @xmath29 , where @xmath30 $ ] and @xmath31 is one of the three cubic \{100 } directions .
the associated hamiltonians @xmath32 are written in the local frame : @xmath33 \\ { \cal h}_{jt}^y & = & \frac{d_q}{3 } \ [ \frac{1}{2}\ q_{xx}+\frac{3}{2}\ q_{yy}+q_{zz}+\sqrt{3 }
\ q_{xy}\\ & & \hspace{15pt } -\sqrt{2 } \
q_{xz } -\sqrt{6 } \
q_{yz } ] \\ { \cal h}_{jt}^z & = & \frac{d_q}{3 } \ [ 2\ q_{xx}+q_{zz}+2\sqrt{2}\ q_{xz}].\end{aligned}\ ] ]
it is crucial to note that , contrary to the magnetic moment operators @xmath20 , the quadrupole moment operators @xmath34 generally couple the two states of the ground non - kramers doublet : @xmath35 , so that the dynamics of the system is deeply modified . for a given direction of the jt axis ,
the degeneracy of the ground doublet is lifted , and the new eigen - functions are close to the symmetric @xmath36 and antisymmetric @xmath37 combinations of the trigonal wave - functions @xmath18 and @xmath19 .
the ground state depends on the sign of the @xmath38 parameter , whose value @xmath39k was derived in our previous work @xcite from the energy of the lowest inelastic excitation . then , the ground singlet is @xmath40 .
we note that the single ion vibronic ( or jahn - teller ) coupling can lead to intersite quadrupole interactions , which could also play a role in defining the low temperature state of tb@xmath0ti@xmath0o@xmath1 .
the nearest neighbour exchange interaction is written in @xmath41 coupling , with the convention that a negative exchange integral corresponds to an af coupling : @xmath42 where @xmath43 means a summation over the first neighbour pairs .
the exchange tensor @xmath44 is chosen to be anisotropic , with its symmetric part diagonal in a frame linked to a r - r bond @xcite .
for instance , the frame associated with the bond between site 1 ( a/2,a/2,a/2 ) and site 2 ( a/4,a/4,a/2 ) has unit vectors : * a*@xmath45=[001 ] , * b*@xmath45=@xmath46 $ ] , * c*@xmath45 = @xmath47 $ ] .
the exchange tensor is identical in all these frames and is written : @xmath48 the @xmath49 , @xmath50 and @xmath51 components represent the symmetric part and @xmath52 the antisymmetric dzyaloshinski - moriya part of the exchange .
we use the anisotropic exchange tensor derived previously for tb@xmath0ti@xmath0o@xmath1 @xcite , slightly modified to better match the diffuse scattering maps ( see section [ diff ] ) . for computing the magnetisation curves for each field direction ,
we introduce the zeeman hamiltonian @xmath53 and we perform , for each site @xmath54 of a tetrahedron and for each direction @xmath55 of the jt axis , the diagonalisation of the following hamiltonian : @xmath56^i,\ ] ] where @xmath57^i$ ] is the part of the first - neighbour exchange and infinite range dipole - dipole couplings relative to site @xmath54 , which is treated in a mean field self - consistent way within the 4 sites of a tetrahedron .
the dipole sums are evaluated using the ewald summation method @xcite .
hence we limit our calculation to the case where the tb moments in each of the 4 fcc sublattices forming the pyrochlore lattice are collinear .
one thus obtains the magnetic structure ( and the magnetisation ) for a * k*=0 propagation vector only .
the diffuse neutron scattering cross section is computed in the paramagnetic phase in zero field using the 4-site rpa outlined in ref .. the starting points are the @xmath58 hamiltonian and the fourier transforms of the exchange and of the dipole - dipole interactions , the latter being obtained following ref ..
the components of the complex dynamic susceptibility @xmath59 are the solutions of a set of linear equations solved using lapack routines . for a given * q
* = * q*+*g * , where * g * is a vector of the reciprocal space of the fcc lattice and * q * belongs to the first brillouin zone , the diffuse scattering intensity is obtained from the real part of the susceptibility as : @xmath60 \ { \rm re } \
\chi^{\alpha,\beta}_{a , b}({\bf q},\omega),\end{aligned}\ ] ] where @xmath61 is the magnetic form factor of the tb@xmath2 ion , @xmath55 and @xmath62 label spatial coordinates and @xmath63 and @xmath64 label the 4 sites on a tetrahedron . since application of a magnetic field induces long range order in tb@xmath0ti@xmath0o@xmath1 @xcite , our rpa calculation of the diffuse scattering holds only in zero field , i.e. in the paramagnetic phase . in our simulations of the present single crystal data ,
since an external axis is defined ( the magnetic field or a particular plane of the reciprocal space ) , our assumption of a dynamic jahn - teller effect implies that an equal weight average over the three \{100 } directions of the jt axis must be performed .
diffuse scattering maps were measured at 0.16k on the super-6t2 diffractometer at the orphe reactor of the laboratoire lon brillouin , saclay ( france ) @xcite , in zero magnetic field and with fields of 1 and 4 t applied along [ 1@xmath130 ] .
data were collected using an area neutron detector ( @xmath65 = 2.35 ) covering a 26@xmath66 x 26@xmath66 angular region , by rotating the sample about the [ 1@xmath130 ] axis with 0.1@xmath66 step and recording a scattering pattern for two detector positions at 2@xmath67=17@xmath66 and 40@xmath66 .
in contrast to earlier diffuse scattering studies in tb@xmath0ti@xmath0o@xmath1 @xcite , the use of the area detector allows one to explore a large three - dimensional ( 3d ) segment of the reciprocal space by transforming a complete set of area detector images in reciprocal space .
the reconstructed volume was completed with its symmetrically equivalent orientations employing the laue symmetry of the structure . in order to strengthen the contrast of the diffuse scattering images ,
the regions of the nuclear and field induced magnetic bragg reflections were excluded after the reconstruction . a constant background signal was subtracted .
a 3d representation of an equal scattering intensity surface in zero field is shown in the left panel of fig.[diff160 ] . from this pattern , the conventional 2d cuts ( h , h , l ) and ( h+1,h@xmath101,l ) , perpendicular to the [ 1@xmath130 ] axis ,
were obtained ( upper central and right panels of fig.[diff160 ] ) .
the ( hhl ) cut is in reasonable agreement with the diffuse scattering data using polarised neutrons of ref .. it shows asymmetric `` butterfly - like '' structures at ( 002 ) and ( 00@xmath68 ) and triangular spots near ( 220 ) and ( @xmath690 ) , with strong intensity , and small pinch points at ( 111 ) etc .
the intensity at the zone center is weak .
the ( h+1,h@xmath101,l ) cut shows strong intensity spots at ( 200 ) and weak intensity `` butterfly - like '' structures at ( 1@xmath703 ) .
the maps were simulated in the presence of dynamic jahn - teller effect using the diagonal exchange tensor elements ( in k ) : @xmath71 as previously determined in ref .
( the absolute value of @xmath51 is 8% smaller that in ref . ) . the dzyaloshinski - moriya term @xmath52 is taken to be zero .
the simulated maps , shown in the lower panels of fig.[diff160 ] , are seen to capture the main features of the experimental data , especially the strong intensity butterfly - like structures at ( 002 ) etc . in the ( hhl ) plane , and the very low scattering `` corridor '' along ( 1,@xmath70,l ) and high intensity spots at ( 200 ) and ( 0@xmath680 ) in the ( h+1,h@xmath101,l ) plane .
application of magnetic field in the [ 1@xmath130 ] direction at 0.16k induces magnetic order with two propagation vectors * k*=0 and * k*=[001 ] , which leads to a strong decrease of diffuse scattering .
since a detailed description of the in - field af structure has been given in ref .
, we concentrate here exclusively on the diffuse scattering results . fig.[diffh160 ] shows in - field diffuse scattering maps with 1 and 4 t applied along [ 1@xmath130 ] , the 3d equal intensity surfaces in the top panels and the ( hhl ) cuts in the lower panels . for
these field values , our rpa approximation scheme , holding only in the paramagnetic phase , can not be applied , so we shall limit ourselves to a qualitative description of the phenomena .
we recall that the field - induced structure involves the so - called @xmath55 and @xmath62 chains along two perpendicular directions . the @xmath55 chains running along *
h * @xmath72 [ 1@xmath130 ] have their local anisotropy axis at 36@xmath66 from the field whereas the @xmath62 ones along [ @xmath73 have their easy axis perpendicular to the field . in first approximation
, the * k*=0 structure is related to the ordering of the @xmath55 chains and the * k*=[001 ] structure to the ordering of the @xmath62 ones . in a field of 1 t ,
only the long range ordered * k*=0 structure is stabilized ; the tb moments in the @xmath55 chains are easily aligned along their local axes , which leads to a strong increase of the intensities of the structurally allowed bragg reflections , accompanied by a decrease of diffuse scattering .
the `` butterfly - like '' patterns are replaced by a disk shaped diffuse scattering concentrated at the forbidden reflections corresponding to the * k * = [ 001 ] structure , like ( 110 ) , ( 112 ) and ( 001 ) .
this remaining diffuse scattering is due to the 1d short range order of the @xmath62 chains .
when the field is further increased to 4 t , a 3d ordering of the @xmath62 chains occurs giving rise to the appearance of the forbidden reflections which violate the extinction rules of the fcc lattice @xcite , and to the almost complete disappearance of the disk shape scattering .
diffuse scattering is however observed around the ( 111 ) reflection positions , possibly due to a small misorientation of the sample induced by the high applied field .
in the spin liquid tb@xmath0ti@xmath0o@xmath1 , application of a magnetic field induces magnetic order with * k*=0 for * h * // [ 110 ] @xcite and * h * // [ 111 ] @xcite , and probably also for * h * // [ 001 ] although no neutron diffraction data are available for this field direction . actually , for * h * // [ 110 ] , an af structure with * k * = [ 100 ] coexists with the * k*=0 structure above 2 t and below 1k @xcite , but it is not expected to contribute to the magnetisation .
therefore , the calculation sketched in section [ rpa ] can be applied to tb@xmath0ti@xmath0o@xmath1 with a magnetic field applied along [ 110 ] , [ 111 ] and probably [ 001 ] .
isothermal magnetisation curves in tb@xmath0ti@xmath0o@xmath1 were measured in the 0.05k range in refs . and
; they show a monotonic increase as the field increases ( solid symbols in fig.[mh3axe ] ) . for * h * // [ 111 ] , no magnetisation `` plateau '' predicted from the vcf model @xcite and akin to that observed in spin ices for this same field direction @xcite is observed down to 0.05k , and we see no precursor effect of it .
we recall however that , according to ref . , a plateau is expected to be clearly visible at a temperature ( 0.02k ) lower than that of these experiments ( @xmath740.050k ) . for the calculation of the magnetisation curves
, we use the values of the anisotropic exchange tensor as in the previous section .
we have also taken into account the standard magneto - elastic ( me ) interaction ( [ hmel ] ) , which yields giant magnetostriction effects @xcite .
we limited ourselves to me terms quadratic in the total angular momentum , following the formalism of ref . and using the me parameters values derived therein .
we have computed the 0.08k and 4k magnetisation curves for fields along [ 100 ] and [ 1@xmath130 ] , and the 0.057k and 4k curves for * h * // [ 111 ] up to 6 t . at 4k , we set @xmath38=0 , since at this temperature the off - diagonal crystal field probably has a quite small effect , if any .
the comparison between experimental data and our calculation is shown in fig.[mh3axe ] .
the overall agreement is reasonably good both at very low temperature and at 4k . in order to assess the importance of the me interactions , we have represented the calculated curves obtained without ( dashed lines ) and with ( solid lines ) me effects .
inclusion of the me interaction slightly modifies the magnetisation , especially at high fields .
it yields a small enhancement , which results in a better agreement with experiment for * h * // [ 100 ] and [ 1@xmath130 ] ( fig.[mh3axe ] * b * ) , but not for * h * // [ 111 ] ( fig.[mh3axe ] * a * ) . for this latter field direction ,
we have also reproduced the curve calculated using the vcf model at 0.05k ( blue line ) , taken from ref .. despite the limited field range , it is clear that it does not reproduce the data , which casts a doubt about the validity the vcf model for tb@xmath0ti@xmath0o@xmath1 , at least as far as the magnetisation is concerned .
furthermore , this model predicts a sizeable variation of the shape of the low field magnetisation curve between 0.02k and 0.1k , with the appearance of a clear plateau at the lowest temperatures ( 0.02k ) .
by contrast , the experimental data @xcite ( in agreement with our calculations ) show that the shape of the magnetisation curve does not appreciably change between 0.05k and 0.3k .
we believe the absence in tb@xmath0ti@xmath0o@xmath1 of the magnetisation plateau expected for ising spins is due to the vibronic nature of the ground state , which likely destroys the ising character at the rare earth site linked with the bare trigonal crystal field eigen - functions .
we note that these measurements of the isothermal magnetisation curves provide a _
direct _ test of the relevance of the vcf prediction .
in contrast , the recent experimental data at very low temperature @xcite presented in favor of the vcf model are only an _
indirect _ evidence .
they do not check the existence of the magnetization plateau in tb@xmath0ti@xmath0o@xmath1 but probe low temperature anomalies which could have another origin .
although our model correctly reproduces the overall magnetisation behaviour and its anisotropy , some deviation occurs at the lowest temperature at low field , around 1 t and below , mainly for * h * // [ 111 ] and [ 100 ] .
the curvature of the magnetisation as the field increases is not exactly reproduced . at higher fields , above 3 - 4 t , the calculated points lie somewhat above the data points , especially for * h * // [ 111 ] both at 0.08k and 4k . keeping in mind
that the uncertainty for such magnetic measurements is usually estimated to amount to a few percent , we can envisage various causes for these deviations .
first , our model would not capture all the details of the field - induced magnetic structure , especially at low fields .
second , the effect of a slight field misalignment with respect to the crystal axes can also play a role . for * h * // [ 1@xmath130 ] , indeed
, there occurs a `` spin melting '' near 1 t where the two moments lying on sites with their ternary axis perpendicular to the field ( @xmath62 sites ) vanish @xcite .
the occurrence of the `` spin melting '' and the configuration of the @xmath62 moments are very sensitive to the alignment of the field with respect to the crystal axis .
finally , we note that the anisotropy of the magnetisation curves is mainly a trigonal crystal field effect ; for fields above @xmath742 t , the jahn - teller interaction and its associated dynamic distortion play only a minor role .
so , at low fields , calculation of the magnetisation curves can suffer from the fact that the dynamic jahn - teller model is only a first approximation to describe the ground state of tb@xmath0ti@xmath0o@xmath1 .
the virtual crystal field approach of ref .
, which is a first order perturbation theory , does not seem to explain the ground state of tb@xmath0ti@xmath0o@xmath1 as observed experimentally .
it seems that a more profound change of the ground state wave function ( zeroth order perturbation ) is necessary , and the dynamic jt interaction developed here is a first attempt to introduce such a quantum mixing .
should then tb@xmath0ti@xmath0o@xmath1 be dubbed a `` quantum spin ice '' , as suggested by several authors ?
although this term was initially proposed for tb@xmath0ti@xmath0o@xmath1 in the context of the vcf approach @xcite , it is now more widely used when quantum fluctuations occur within a set of spin ice states @xcite .
indeed tb@xmath0ti@xmath0o@xmath1 , where tb@xmath2 is a non kramers ion , which shows dispersive excitations in a disordered ground state @xcite and where the likely presence of a vibronic coupling can induce an interaction between quadrupolar moments @xcite , is a good candidate .
however , the small energy gap to the first excited cef level renders the situation more complicated to treat theoretically than for the generic quantum spin ices with effective spin @xmath75 considered up to now .
the simple distortion or jahn - teller model for tb@xmath0ti@xmath0o@xmath1 implies that , for a certain range of exchange parameters , the mean field single ion ground state is a singlet , i.e. it is non - magnetic and thus could explain the absence of magnetic ordering down to the lowest temperature . however , energy resolved neutron scattering shows that elastic or quasi - elastic scattering is dominant at very low temperature @xcite , meaning that the ground state has a non - zero magnetic moment which is static at the time scale of the experiment .
the fact that our model can reasonably well reproduce the * q*-maps in reciprocal space of the inelastic @xcite and diffuse ( i.e. energy integrated ) scattering ( ref . for the spin - flip ( sf ) and non - spin - flip scattering and present work for the total scattering ) probably shows that it correctly represents the spin correlations .
however , it situates their characteristic energy in the inelastic channel rather than in the elastic one , since a non - magnetic ground state does not give rise to elastic scattering . in terms of the mixed eigen - states @xmath37 and @xmath36 introduced in section [ rpa ] , one can define a `` transition vector '' * t * = @xmath76 , whose square modulus is the intensity of the low energy inelastic mode .
the * t * vector is directed along the local trigonal axis since its only non - zero component is @xmath77 , and it is solely defined by the cef wavefunctions .
so the intensity of the calculated diffuse scattering involves the pure `` inelastic '' cef quantity @xmath78 , the subsequent energy integration yielding expression ( [ chipr ] ) , proportional to @xmath79 according to the kramers - kronig theorem , and its * q*-dependence is mainly due to the anisotropy of the exchange tensor . to emphasize the difference of the diffuse scattering in tb@xmath0ti@xmath0o@xmath1 with respect to those in pyrochlore systems with dominant spin - ice or antiferromagnetic spin correlations , we have computed the spin flip scattering maps in the ( hhl ) plane within the spin - liquid phase of our model @xcite ( see figure [ sfmap ] ) .
near the upper border of the sl phase with the ordered spin ice ( osi ) phase ( @xmath80k ) , the sf maps are very close to those of a spin ice @xcite , where the spin correlations are of the type `` two in - two out '' .
the corresponding bragg spots ( map with blue background in the osi phase ) are located at the pinch points of the sf map . near the border with the antiferromagnetic ( af ) phase ,
the sf scattering consists only in broadened peaks at positions ( 220 ) , ( 113 ) , etc .
these positions are those of the bragg peaks of the `` all in / all out '' spin structure , as shown in the map with blue background in the af phase .
the sketch of the sl phase in fig.[sfmap ] allows one to follow the progressive transformation of the sf maps as one scans the \{@xmath81 } plane .
it shows that the sf map is very specific to a given set of exchange integrals .
the map closest to the experimental data ( @xmath820.068k , @xmath830.090k ) , near the bottom left corner , is different from the two above described limiting cases and must therefore represent a special type of spin correlations occuring in tb@xmath0ti@xmath0o@xmath1 , probably consisting in a mixture of spin - ice and af correlations . within our model ,
tb@xmath0ti@xmath0o@xmath1 lies very close to the border with the long range ordered osi phase .
this could explain the presence of spin ice like features in the af spin correlations @xcite and the stabilisation of the osi phase in the sibling material tb@xmath0sn@xmath0o@xmath1 below 0.87k @xcite . in this latter compound ,
a nuclear schottky anomaly is present at low temperature and the hyperfine field derived from it is 180 t , corresponding to an effective magnetic moment of 4.5@xmath84/tb@xmath2 @xcite , using the @xmath85tb@xmath2 hyperfine constant of 40(4)t/@xmath84 @xcite .
this effective moment of 4.5@xmath84 is reduced with respect to the spontaneous moment 5.9(1)@xmath84 measured by neutron diffraction , which is attributed in ref . to persisting spin fluctuations in the osi phase .
more precisely , the hyperfine schottky anomaly is depleted in case the electronic spin fluctuation frequency @xmath86 is not much slower than the nuclear relaxation frequency 1/@xmath87 @xcite , i.e. if @xmath88 .
the presence of a similar nuclear schottky anomaly in the very low temperature specific heat of tb@xmath0ti@xmath0o@xmath1 , with a hyperfine field of 142 t @xcite , is rather unexpected in a material with no magnetic ordering .
it shows that there is a hyperfine field at the nucleus site , linked in this case with the short range dynamically correlated tb@xmath2 magnetic moments .
the moment value derived from the 142 t hyperfine field is 3.6(4)@xmath84 .
thus , from the point of view of the nuclear specific heat , tb@xmath0sn@xmath0o@xmath1 and tb@xmath0ti@xmath0o@xmath1 behave very similarly , although the former orders magnetically while the latter does not . as a further analogy , in order to account for the inelastic neutron scattering maps in the osi phase of tb@xmath0sn@xmath0o@xmath1 @xcite , it is necessary to introduce an off - diagonal crystal field term which was taken as a tetragonal distortion .
the @xmath89sr data in tb@xmath0ti@xmath0o@xmath1 @xcite are also very similar to those in tb@xmath0sn@xmath0o@xmath1 @xcite , revealing sizeable fluctuating moments in both compounds down to the lowest temperature
. inspection of the bulk of experimental data in tb@xmath0ti@xmath0o@xmath1 at very low temperature suggests the following : on the one hand , experiments probing the behaviour of tb@xmath0ti@xmath0o@xmath1 with a small characteristic time @xmath90 , like the quasi - elastic neutron scattering ( @xmath91 , where @xmath92 is the energy resolution , i.e. @xmath93s for @xmath94mev ) , the specific heat ( @xmath95 , where @xmath87 is the @xmath85 tb nuclear relaxation time ) , @xmath89sr spectroscopy ( @xmath96 , where @xmath97@xmath89s is the muon lifetime ) , reveal the short time behaviour of the tb@xmath2 ions .
this behaviour is that of an ion submitted to a dynamic non - zero exchange / dipolar field , whose magnitude ( @xmath741 t ) is enough , in the presence of the jt or vibronic coupling , to polarise the moment during the lifetime of the spin correlations . on the other hand , experiments with a long characteristic time , like the magnetisation measurements ( @xmath98 , where @xmath99s is the magnetometer measuring time ) or without well defined characteristic time , like the diffuse neutron scattering measurements , seem to be correctly described by the dynamic jt model , with its non - magnetic ground state
we have shown that it is possible to account for some quite different properties of the pyrochlore spin - liquid candidate tb@xmath0ti@xmath0o@xmath1 at very low temperature , i.e. the neutron diffuse scattering and the isothermal magnetisation , by a model which approximates the probably complex vibronic ground state of this material by a dynamic jahn - teller coupling within the ground electronic doublet .
the success of the model in reproducing the magnetisation curves is linked with the mixed character of the `` tunnel - like '' antisymmetric ground wave - function .
absence of this zero - order mixing automatically leads to a _
kagom_-ice state for intermediate field values along [ 111 ] and thus to a magnetisation plateau . as to the diffuse scattering ,
its calculated intensity reflects that of the transition between the mixed jahn - teller states , and the * q*-dependence of the maps reflects the anisotropy of the exchange tensor .
however , this approach does not provide a fully correct picture for tb@xmath0ti@xmath0o@xmath1 since it predicts a non - magnetic ground state , in contradiction with the observation of a sizeable quasielastic neutron scattering and of a magnetic hyperfine schottky anomaly in the specific heat .
we believe it could anyway be a starting point while awaiting for a more elaborate treatment of the peculiar coupling between the phononic and electronic degrees of freedom which seems to be at play in tb@xmath0ti@xmath0o@xmath1 at very low temperature .
we are indebted to b. z. malkin for his help with the magneto - elastic calculations .
we are grateful to j. robert and a.p .
sazonov for fruitful discussions , to c. decorse and g. dhalenne , and to p. dalmas de rotier and ch .
marin , for providing the tb@xmath0ti@xmath0o@xmath1 single crystals .
we thank e. lhotel and c. paulsen for communicating unpublished results .
99 j. s. gardner , m. j. p. gingras , and j. e. greedan , rev .
phys . * 82 * , 53 ( 2010 ) m. j. harris , s. t. bramwell , p. c. w. holdsworth , and j. d. m. champion , phys .
lett . * 81 * , 4496 ( 1998 ) a. p. ramirez , a. hayashi , r. j. cava , r. siddhartan , and b. s. shastry , nature * 399 * , 333 ( 1999 ) c. castelnovo , r. moessner , and s. l. sondhi , nature * 451 * , 42 ( 2008 ) j. s. gardner , s. r. dunsiger , b. d. gaulin , m. j. p. gingras , j. e. greedan , r. f. kiefl , m. d. lumsden , w. a. macfarlane , n. p. raju , j. e. sonier , i. swainson , and z. tun , phys .
. lett . * 82 * , 1012 ( 1999 ) i. v. aleksandrov , b. v. lidski , l. g. mamsurova , m. g. neygauz , k. s. pigalski , k. k. pukhov , n. g. trusevich , and l. g. shcherbakova , sov .
jetp * 62 * , 1287 ( 1985 ) m. j. p. gingras , b. c. den hertog , m. faucher , j. s. gardner , s. r. dunsiger , l. j. chang , b. d. gaulin , n. p. raju , and j. e. greedan , phys .
b * 62 * , 6496 ( 2000 ) i. mirebeau , p. bonville , and m. hennion , phys . rev .
b * 76 * , 184436(2007 ) j. s. gardner , b. d. gaulin , a. j. berlinsky , p. waldron , s. r. dunsiger , n. p. raju , and j. e. greedan , phys . rev .
b * 64 * , 224416 ( 2001 ) y. yasui , m. kanada , m. ito , h. harashina , m. sato , h. okumura , k. kakurai , and h. kadowaki , j. phys . soc
* 71 * , 599 ( 2002 ) y. j. kao , m. enjalran , a. del maestro , h. r. molavian , and m. j. p. gingras , phys .
b * 68 * , 172407 ( 2003 ) m. enjalran , and m. j. p. gingras , phys . rev .
b * 70 * , 174426 ( 2004 ) s. h. curnoe , phys . rev .
b * 75 * , 212404 ( 2007 ) h. r. molavian , m. j. p. gingras , and b. canals , phys
lett . * 98 * , 157204 ( 2007 ) h. r. molavian , p. a. mc clarty , and m. j. p. gingras , arxiv : 0912.2957 i. mirebeau , i. n. goncharenko , p.
cadavez - peres , s. t. bramwell , m. j. p. gingras , and j. s. gardner , nature * 420 * , 54 ( 2002 ) ; i. mirebeau , i. n. goncharenko , g. dhalenne , and a. revcolevschi , phys .
lett . * 93 * , 187204 ( 2004 ) t. taniguchi , h. kadowaki , h. takatsu , b. fk , j. ollivier , t. yamazaki , t. j. sato , h. yoshizawa , y. shimura , t. sakakibara , t. hong , k. goto , l. r. yaraskavitch , and j. b. kycia , phys . rev .
b * 87 * , 060408(r ) ( 2013 ) t. fennell , m. kenzelmann , b. roessli , m. k. haas , and r. j. cava , phys
lett . * 109 * , 017201 ( 2012 ) s. petit , p. bonville , j. robert , c. decorse , and i. mirebeau , phys .
b * 86 * , 174403 ( 2012 ) k. fritsch , k. a. ross , y. qiu , j. r. d. copley , t. guidi , r. i. bewley , h. a. dabkowska , and b. d. gaulin , phys . rev .
b * 87 * , 094410 ( 2013 ) k. c. rule , j. p. c. ruff , b. d. gaulin , s. r. dunsiger , j. s. gardner , j. p. clancy , m. j. lewis , h. a. dabkowska , i. mirebeau , p. manuel , y. qiu , and j. r. d. copley , phys . rev
* 96 * , 177201 ( 2006 ) a. p.
sazonov , a. gukasov , i. mirebeau , h. b. cao , p. bonville , b. grenier , and g. dhalenne , phys .
b * 82 * , 174406 ( 2010 ) a. p.
sazonov , a. gukasov , h. b. cao , p. bonville , e. ressouche , c. decorse , and i. mirebeau , phys .
b * 88 * , 184428 ( 2013 ) j. p. c. ruff , b. d. gaulin , j. p. castellan , k. c. rule , j. p. clancy , j. rodriguez , and h. a. dabkowska , phys .
* 99 * , 237202 ( 2007 ) t. t. a. lummen , i. p. handayani , m. c. donker , d. fausti , g. dhalenne , p. berthet , a. revcolevschi , and p. h. m. van loosdrecht , phys .
b * 77 * , 214310 ( 2008 ) k. c. rule , and p. bonville , j. phys . :
series * 145 * , 012027 ( 2009 ) j. p. c. ruff , z. islam , j. p. clancy , k. a. ross , h. nojiri , y. h. matsuda , h. a. dabkowska , a. d. dabkowski , and b. d. gaulin , phys .
* 105 * , 077203 ( 2010 ) l. g. mamsurova , k. s. pigalski , and k. k. pukhov , jetp letters * 43 * , 755 ( 1986 ) l. g. mamsurova , k. s. pigalski , n. g. pukhov , n. g. trusevich , and l. g. shcherbakova , sov .
jetp * 67 * , 550 ( 1988 ) y. nakanishi , t. kumagai , m. yoshizawa , k. matsuhira , s. takagi , and z. hiroi , phys . rev .
b * 83 * , 184434 ( 2011 ) y. luan , phd dissertation `` elastic properties of complex transition metal oxides studied by resonant ultrasound spectroscopy '' , university of tennessee , usa ( 2011 ) t. fennell , m. kenzelmann , b. roessli , h. mutka , j. ollivier , m. ruminy , u. stuhr , o. zaharko , l. bovo , a. cervellino , m. k. haas , and r. j. cava , arxiv : 1305.5405 s. guitteny , j. robert , p. bonville , j. ollivier , c. decorse , p. steffens , m. boehm , h. mutka , i. mirebeau , and s. petit , phys . rev . lett . *
111 * , 087201 ( 2013 ) q. j. li , z. y. zhao , c. fan , f. b. zhang , h. d. zhou , x. zhao , and x. f. sun , phys . rev .
b * 87 * , 214408 ( 2013 ) p. bonville , i. mirebeau , a. gukasov , s. petit , and j. robert , j. phys . :
series * 320 * , 012006 ( 2011 ) p.
bonville , i. mirebeau , a. gukasov , s. petit , and j. robert , phys .
b * 84 * , 184409 ( 2011 ) b. bleaney , proc . r. soc .
london a * 276 * , 19 ( 1963 ) b. r. cooper , and o. vogt , jour . physique ( paris )
* c1 * -958 ( 1971 ) b. d. gaulin , j. s. gardner , p. a. mcclarty , and m. j p. gingras , phys .
b * 84 * , 140402(r ) ( 2011 ) a. yaouanc , p. dalmas de rotier , y. chapuis , c. marin , s. vanishri , d. aoki , b. fk , l. p. regnault , c. buisson , a. amato , c. baines , and a. d. hillier , phys . rev .
b * 84 * , 184403 ( 2011 ) e. lhotel , c. paulsen , p. d. de rotier , a. yaouanc , c. marin , and s. vanishri , phys . rev .
b * 86 * , 020410(r ) ( 2012 ) e. lhotel , and c. paulsen , private communication h. b. cao , a. gukasov , i. mirebeau , p. bonville , c. decorse , and g. dhalenne , phys . rev . lett .
* 103 * , 056402 ( 2009 ) s. onoda , and y. tanaka , phys .
lett . * 105 * , 047201 ( 2010 ) g. a. gehring , and k. a. gehring , rep .
phys . * 38 * , 1 ( 1975 ) v. v. klekovkina , a. r. zakirov , b. z. malkin , and l. a. kasatkina , j. phys .
series * 324 * , 012036 ( 2011 ) b. z. malkin , t. t. a. lummen , p. h. m. van loosdrecht , g. dhalenne , and a. r. zakirov , j. phys . : condens .
matter * 22 * , 276003 ( 2010 ) p. bonville , s. petit , i. mirebeau , j. robert , e. lhotel , and c. paulsen , j. phys . : condens .
matter * 25 * , 275601 ( 2013 ) z. wang , and c. holm , j. chem . phys .
* 115 * , 6351 ( 2001 ) a. g. del maestro , and m. j. p. gingras , j. phys . :
matter * 16 * , 3339 ( 2004 ) a. p.
sazonov , a. gukasov , i. mirebeau , and p. bonville , phys .
b * 85 * , 214420 ( 2012 ) a. gukasov , a. goujon , j. l. meuriot , c. person , g. exil , and g. koskas , physica b * 397 * , 131 ( 2007 ) s. legl , c. krey , s. r. dunsiger , h. a. dabkowska , j. a. rodriguez , g. m. luke , and c. pfleiderer , phys .
lett . * 109 * , 047201 ( 2012 ) h. r. molavian , and m. j. p. gingras , j. phys . : condens .
matter * 21 * , 172201 ( 2009 ) k. matsuhira , z. hiroi , t. tayama , s. takagi , and t. sakakibara , j. phys . : condens .
matter * 14 * , l559 ( 2002 ) l. yin , j. s. xia , y. takano , n. s. sullivan , q. j. li , and x. f. sun , phys .
lett . * 110 * , 137201 ( 2013 ) o. benton , o. sikora , and n. shannon , phys .
b * 86 * , 075154 ( 2012 ) s. b. lee , s. onoda , and l. balents , phys . rev .
b * 86 * , 104412 ( 2012 ) m. j. p. gingras , and p. a. mcclarthy , arxiv : 1311.1817 h. takatsu , h. kadowaki , t. j. sato , j. w. lynn , y. tabata , t. yamazaki , and k. matsuhira , j. phys . : condens . matter * 24 * , 052201 ( 2012 ) the phase diagram in ref . is represented in the space of the molecular constants @xmath65 .
the relationship with the exchange constants used here is : @xmath100=@xmath101 , i.e. @xmath100(k)=1.51 @xmath65(t/@xmath102 ) . s. t. bramwell , and m. j. p. gingras , science * 294 * 1495 ( 2001 ) i. mirebeau , a. apetrei , j. rodriguez - carvajal , p. bonville , a. forget , d. colson , v. glazkov , j. p. sanchez , o. isnard , and e. suard , phys
94 * , 246402 ( 2005 ) the hyperfine field and the effective moment values in ref . have been erroneously quoted as 135 t and 3.35@xmath84/tb@xmath2 respectively .
their actual values in tb@xmath0sn@xmath0o@xmath1 are 180 t and 4.5@xmath84/tb@xmath2 respectively .
b. d. dunlap , mssbauer effect methodology ( plenum press ) * 7 * , 123 ( 1971 ) e. bertin , p. bonville , j. p. bouchaud , j. a. hodges , j. p. sanchez , and p. vulliet , eur
j. b * 27 * , 347 ( 2002 ) s. petit , p. bonville , i. mirebeau , h. mutka , and j. robert , phys .
b * 85 * , 054428 ( 2012 ) p. dalmas
de rotier , a. yaouanc , l. keller , a. cervellino , b. roessli , c. baines , a. forget , c. vaju , p. c. m. gubbens , a. amato , and p. j. c. king , phys .
96 * , 127202 ( 2006 ) f. bert , p. mendels , a. olariu , n. blanchard , g. collin , a. amato , c. baines , and a. d. hillier , phys .
lett . * 97 * , 117203 ( 2006 )
|
we present an interpretation of zero field diffuse neutron scattering and of high field magnetisation data at very low temperature in the frustrated pyrochlore system tb@xmath0ti@xmath0o@xmath1 .
this material has antiferromagnetic exchange interactions and it is expected to have ising character at low temperature .
contrary to expectations , it shows no magnetic ordering down to 0.05k , being thus labelled a `` spin liquid '' .
however , the ground state in tb@xmath0ti@xmath0o@xmath1 is not a mere fluctuating moment paramagnet but , as demonstrated by very recent experiments , a state where the electronic degrees of freedom are hybridised with the phononic variables in an unconventional way .
we show here that , by approximating this complex and still unraveled electron - phonon interaction by a dynamic jahn - teller coupling , one can account rather well for the diffuse neutron scattering and the low temperature isothermal magnetisation .
we discuss the shortcomings of this picture which arise mainly from the fact that the singlet electronic mean field ground state of the model fails to reproduce the observed strong intensity of the elastic and quasi - elastic neutron scattering .
|
( a ) conduction ( cb ) and valence states ( vb ) in a single particle picture . ( b )
the intensity of the neutral 1s a - exciton pl ( shown as gray solid line ) is recorded as a function of laser energy . in 1-photon ple ( red circles ) the 1s b - exciton state ( x@xmath2 ) is identified , in 2-photon ple ( open black squares ) the peaks
are assigned to the 2p a - exciton ( x@xmath3 ) and 2p b - exciton ( x@xmath4 ) .
( c ) maxima of the first derivative of the reflectivity allow to assign the 1s a - exciton and b - exciton state energies .
, scaledwidth=47.0% ] monolayers ( mls ) of the transition metal dichalcogenides ( tmdcs ) mos@xmath0 , mose@xmath0 , ws@xmath0 and wse@xmath0 ( abbreviated mx@xmath0 ) are semiconductors with a direct bandgap in the visible region @xcite .
mx@xmath0 mls are promising materials for ( opto-)electronics @xcite , non - linear optics @xcite and for exploring electron k - valley physics @xcite .
the mx@xmath0 ml materials share common characteristics : ( i ) their optical properties are dominated by excitons , strongly coulomb - bound electron hole pairs @xcite .
( ii ) in these tmdc mls crystal inversion symmetry breaking together with the strong spin - orbit ( so ) interaction leads to a coupling of carrier spin and k - space valley physics , i.e. , the circular polarization ( @xmath5 or @xmath6 ) of the absorbed or emitted photon can be directly associated with selective carrier excitation in one of the two non - equivalent k valleys ( k@xmath7 or k@xmath8 , respectively ) @xcite .
+ using non - resonant laser excitation in ml mos@xmath0 @xcite , wse@xmath0 @xcite and ws@xmath0 @xcite high values ( @xmath9 ) for the circular polarization @xmath10 of the stationary photoluminescence ( pl ) corresponding to successful valley polarization have been reported .
in contrast , the polarization reported for the promising material mose@xmath0 @xcite under similar experimental conditions is surprisingly very low ( @xmath11 ) @xcite .
experiments combing pl excitation ( ple ) and time resolved pl suggest either ultra fast polarization relaxation in mose@xmath0 in the sub - picosecond range or inefficient optical polarization generation due to anomalies in the bandstructure to be at the origin of this low valley polarization @xcite . +
our target is to investigate how an eventual bandstructure anomaly ( local extremum ) , can influence the nature of the optical transitions . to this aim
we combine linear and non - linear optical spectroscopy at t=4 k with calculations beyond standard density functional theory ( dft ) to study the electronic bandstructure of mose@xmath0 mls . in 2-photon ple
we detect 180 mev above the 1s state the 2p state of the a - exciton , well separated in energy from the 1s b - exciton emission recorded in both reflectivity and 1-photon ple .
the energy position of the neutral and charged exciton transitions is determined using second harmonic generation ( shg ) spectroscopy , as the shg signal is enhanced by more than 2 orders at these particular resonances , important for applications in non - linear optics .
+ the strong excitonic effects that dominate the 1- and 2-photon spectroscopy results in the investigated mose@xmath0 mls have important consequences not just in terms of the transition energy @xcite . due to the small exciton bohr radius ,
the extension in k - space is considerable and the exciton wavefunction will include contributions from states far away from the k - point of the brillouin zone where the direct free carrier ( electronic ) bandgap of the mx@xmath0 mls is situated . the position in energy and in k - space of the adjacent local minima in the conduction band ( cb ) and maxima in the valence band ( vb ) , respectively can be very different from one mx@xmath0 material to another @xcite .
our dft - gw calculations show in the cb of ml mose@xmath0 a local minimum that is energetically and in k - space close to the global minimum at the k - point @xcite .
we evaluate how the proximity of this local minimum along the @xmath1 direction can influence the polarization and energy of the excitonic states that govern the interband transitions , marking an important difference compared to mos@xmath0 and wse@xmath0 monolayers .
[ h ] the second harmonic generation ( shg ) signal is plotted as a function of twice the laser photon energy .
the local maxima corresponding to the neutral and charged exciton states are marked ( see text ) .
, title="fig:",scaledwidth=49.0% ]
similar to the hydrogen model , an electron @xmath12 hole pair in the tmdc ml interacts through an attractive coulomb potential and will form below the gap a series of excitonic rydberg - like states with definite parity , where @xmath13 states are of even , and @xmath14 states are of odd parity ( n is an integer ) @xcite .
the exciton binding energy @xmath15 of the order of 0.5 ev in ml mose@xmath0 can be estimated by determining the free carrier bandgap in scanning tunnelling spectroscopy ( sts ) and subtracting the neutral exciton energy ( corresponds to the 1s a - exciton state ) , see @xcite . here
we aim to uncover in addition to the 1s exciton ground state also higher excited exciton states , which were shown to dominate the linear and non - linear optical response in mos@xmath0 and wse@xmath0 mls @xcite . in fig .
[ fig : fig1]b we see a typical time - integrated pl spectrum at t=4 k for ml mose@xmath0 with two prominent peaks ( see appendix for information on samples and the experimental set - up ) .
the higher energy peak ( fwhm=10 mev ) at 1.667 ev has previously been attributed to the neutral a - exciton x@xmath16 @xcite . at 1.633 ev
we record the trion emission ( t ) corresponding to a binding energy of 34 mev , in agreement with previous measurements @xcite . here
we detect the intensity of the x@xmath16 pl emission as a function of the laser energy , shown in fig .
[ fig : fig1]b .
we find a clear maximum at 1.885 ev , which we attribute to the 1s b - exciton state . in order to confirm this
, we have performed reflectivity measurements using a white light source .
we find in reflectivity two prominent transitions in fig .
[ fig : fig1]c , one at 1.667 ev , which corresponds exactly to the neutral a - exciton emission energy measured in pl .
this indicates negligible localization of excitons in this sample ( no stokes shift ) .
the second transition in reflectivity at 1.885 ev gives the position of the 1s b - exciton .
the measured difference between 1s a and b - exciton is @xmath17 mev determined from both pl and reflectivity . from our dft calculations ( see section [ sec : calcul ] for details ) we obtain a spin splitting of @xmath18 mev in the valence band .
this is in very good agreement with the measured value of 180 mev from angle - resolved photoemission spectroscopy ( arpes ) measurements at t=40 k @xcite .
the cb spin splitting is predicted to have the opposite sign compared to the vb , as sketched in fig .
[ fig : fig1]a @xcite .
we obtain from our dft calculations a cb spin splitting of @xmath19 mev .
the measured difference @xmath20 of 220 mev is in close agreement with the calculated energy difference taking into account the vb and cb spin splitting of 183 + 30=213 mev .
in addition , we find in dft calculations in section [ sec : calcul ] that the effective masses in cb and vb are slightly different for different spin states , and hence also the exciton binding energies of a- and b - excitons are expected to differ and contribute to the measured value of @xmath20 .
an experimental signature of excited exciton states , in analogy with the @xmath13 and @xmath14 states of the hydrogen atom , have not been reported yet for mose@xmath0 mls , to the best of our knowledge .
the excited exciton states and the exact exciton binding energy @xmath15 in mx@xmath0 mls are currently debated in the literature for ws@xmath0 @xcite , mos@xmath0 @xcite and wse@xmath0 @xcite , with differences in the reported @xmath15 for the same material by factor of 2 , so additional experiments are important . in our 2-photon ple measurements we can directly address exciton states with p - symmetry and not with s - symmetry as in the 1-photon ple and reflectivity . in the 2-photon ple of fig . [
fig : fig1]b we find a well defined peak at 1.844 ev , which we assign to the 2p a - exciton transition , not obscured by the 1s b - exciton , which is parity forbidden here .
we measure an energy difference 1s-2p of 180 mev . using a binding energy of the order of @xmath21 ev @xcite in a simple 2d hydrogen model @xcite , the energy separation 1s-2s,2p
is expected to be @xmath22 mev , much larger than our measured value of 180 mev .
this shows for ml mose@xmath0 a strong deviation from a simple hydrogenic series , observed also in ml ws@xmath0 and wse@xmath0 , very likely due to a strong variation of the effective dielectric constant as a function of the spatial extension of the exciton state @xcite .
interestingly we observe a second , clear maximum at 2.06 ev , about 175 mev above the b - exciton 1s state .
we tentatively assign this peak in 2-photon ple to the 2p b - exciton state .
in addition to the 2-photon ple signal we can also plot the intensity of the shg signal as a function of laser energy in fig .
[ fig : fig2 ] .
the shg spectroscopy results in this ml material without crystal inversion centre also show clear resonances at the exciton energies .
interestingly we observe a strong shg signal at the transition energy of the charged exciton ( trion ) .
this hints at a substantial density of states of this complex .
+ the shg signal at the 1s b - exciton is about 120 times higher than in the region between 1.7 and 1.8 ev .
this finding shows that excitonic resonances dominate the non - linear optical response of ml mose@xmath0 .
the shg signal is non - zero for all laser energies , the shg peaks outside the exciton resonances between 1.7 and 1.8 ev are not clearly visible on the linear scale used in fig .
[ fig : fig2 ] .
we confirm the energy position of the 1s and 2p exciton levels a and b. note that resonances of 1s exciton states in shg are forbidden if strict electric dipole selection rules apply .
the prominent 1s features in our shg spectrum in fig .
[ fig : fig2 ] can come from the interplay of electric dipole with magnetic dipole transitions @xcite .
in addition strict electric dipole selection rules could be slightly relaxed if the overall symmetry of the crystal is lowered by extrinsic effects such as the substrate .
in the experiments on ml mose@xmath0 we find sharp excitonic features that are well defined as for the related ml material wse@xmath0 @xcite . despite similar exciton binding energy and excited state spectrum in ml wse@xmath0 and ml mose@xmath0 the optical valley polarization using non - resonant laser excitation
can be generated in the former but not in the latter @xcite .
one reason could be an anomaly in the bandstructure , which we aim to uncover in band structure calculations for mose@xmath0 . for comparison and to validate our computational approach we have performed in parallel calculations with the more thoroughly studied ml mos@xmath0 . the bandstructure of ml mose@xmath0 and mos@xmath0 can be compared in fig .
[ fig : fig3 ] , where striking differences in both valence and conduction band appear .
we extract the effective carrier masses , the band gaps and the exciton binding energies .
we discuss the most important features for optical transitions in detail , in particular the competition between direct and indirect exciton states in mose@xmath0 mls .
+ the comparison of quasi - particle band structure calculated by dft ( see appendix for computational details ) of ml mos@xmath0 and mose@xmath0 provides interesting insights : when replacing s with se , the spin - splitting is enhanced .
for mos@xmath0 the valence and conduction state the spin - splittings at the k@xmath23 point are @xmath24 mev and @xmath25 mev , respectively .
for ml mose@xmath0 we find @xmath18 mev and @xmath19 mev .
the direct band gap values ( neglecting excitonic effects ) at k@xmath23 , depicted by arrows in the left panels of fig .
[ fig : fig3 ] , are 2.13 and 2.31 ev for mose@xmath0 and mos@xmath0 , respectively .
the latter value is lower than previous theoretical studies as energies from 2.41 to 2.97 ev have been reported @xcite .
it is important to take into account that the uncertainty in this type of band gap calculations is in the hundreds of mev range , depending on the computational settings .
the obtained results critically depend on the choice of the @xmath26 methods used ( self - consistent or partially self - consistent scheme ) , the number of unoccupied states included , the vacuum height and the k - point sampling . interestingly using exactly the same computational settings ,
our mose@xmath0 direct band gap estimate is closer to recent reports : 2.33 ev for a @xmath27 calculation in ref .
, when ugeda _ et al _
@xcite propose 2.26 ev for a free standing mose@xmath0 ml . on the experimental side , using sts techniques the band gap of mose@xmath0 ml is 2.18 @xmath28 0.04 ev @xcite but it includes substrate screening effects . + if one compares the full band - structure of the two material systems another striking difference appears : in mos@xmath0 ml , the topmost valence bands in @xmath29 and k@xmath23 are only separated by 65 mev .
this feature appears to be essential in the mechanism of direct to indirect band gap transition and its evolution with layer thickness ( 1ml , 2ml ... ) @xcite and is expected to have a strong impact on the optically generated valley polarization as a function of laser energy @xcite .
the same trend for the gap evolution with layer thickness applies to mose@xmath0 @xcite , albeit with a much larger @xmath29 to k@xmath23 energy separation of 485 mev for 1ml . + for ml mose@xmath0 our calculations show a remarkable anomaly in the cb , which could play a key role for the valley polarization dynamics : in ml mose@xmath0 there is a small energy difference of only 46 mev between the cb minimum in k@xmath7 and in @xmath30 , as can be clearly seen in fig .
[ fig : fig3 ] .
the corresponding value for mos@xmath0 is 345 mev and therefore not in competition with direct optical transitions at the k points .
we note another important difference : the @xmath30 position along the k@xmath31 line for mos@xmath0 lies exactly at the @xmath32 coordinates .
in contrast , for mose@xmath0 it is slightly shifted towards the k point .
interestingly the spin - up and spin - down minima are not positioned at the same k - value .
note that the appearance of this local minimum has also been predicted recently in @xcite .
+ in the following we discuss how the proximity of the cb minimum at @xmath30 can potentially influence the optical transitions . excitons with large binding energies in ml mose@xmath0 have a very small bohr radius and as a result extend in k - space well beyond the k - point ( see fig . [
fig : fig3 ] ) . to allow for a more quantitative discussion , we need to extract the effective carrier masses from our calculated band structure . at vb and cb extrema ,
effective masses for holes and electrons can be extracted with a simple quadratic fitting procedure .
the extracted mass values will depend on the selected k - value range around the extrema . for consistency
, we have used the interval @xmath33 $ ] , with @xmath34 being defined as @xmath35 , where @xmath36 is the bohr radius of the exciton ground state in a wannier - mott picture of electron - hole pair @xcite .
the computed effective mass values are given in table [ tab1 ] .
for mos@xmath0 , the @xmath37 and @xmath38 values are only slightly different from previous studies @xcite , the difference being reduced in the electron - hole pair effective mass @xmath39 values .
there is a certain spread in values , especially for the mos@xmath0 hole effective mass with the value provided in ref . , or for mose@xmath0 values @xcite .
we obtain smaller values , but we attribute these differences to the choice of the extension in k - space around the extrema .
indeed , if one takes a fitting interval of length @xmath40 the effective masses are significantly increased by 30% in both material systems .
interestingly spin - up and spin - down excitons have different effective masses at k@xmath7 : 0.21 @xmath41 _ vs _ 0.24 @xmath41 for @xmath42 and @xmath43 respectively for mos@xmath0 , 0.25 @xmath41 and 0.28 @xmath41 in mose@xmath0 , due to larger values of the spin - down @xmath37 and @xmath38 , in both systems .
this confirms that the exciton binding energies for a- and b - excitons should be slightly different , which will in turn influence the energy difference @xmath20 between the corresponding optical transitions measured in fig .
[ fig : fig1 ] .
+ on top of @xmath27 calculations , exciton binding energies @xmath15 , given in table [ tab1 ] , have been extracted from the imaginary part of transverse dielectric constant after solving the bethe - salpeter equation ( bse ) .
our 0.58 ev estimate is in reasonable agreement with previous theoretical studies for the mos@xmath0 a - exciton binding energy : values from 0.55 @xcite , around 0.6 @xcite , or around 0.9 @xcite and up to 1.1 ev@xcite have been reported .
please note the computational details are extremely different for each of these values .
it is therefore difficult to extract precise trends .
alternatively , the spread in values should rather be used as a reasonable error bar for this type of calculations . for mose@xmath0 theoretical values for @xmath15 of 0.9 @xcite
, 0.78 @xcite and 0.65 ev @xcite can be found in the recent literature .
this indicates that the calculated exciton binding energy in ml mose@xmath0 is smaller than in mos@xmath0 , as observed experimentally .
+ .free carrier band - gap @xmath44 , exciton binding energy @xmath15 , cb effective masses @xmath37 and vb effective masses @xmath38 ( here given as positive values as in the hole picture ) and the corresponding relative static dielectric constant @xmath45 and exciton bohr radius @xmath36 for mos@xmath0 and mose@xmath0 monolayer ( see text ) .
[ cols="<,^,^,^,^,^,^,^,^,^,^,^,^,^ " , ] [ tab1 ] our target is now to estimate how far the exciton state extends in k - space around the k@xmath7 point .
this allows us to estimate the possible impact of the proximity of the local cb minimum at @xmath30 on the optical transitions and the comparatively weak valley polarization in ml mose@xmath0 .
this order of magnitude discussion is graphically represented in the right column of fig .
[ fig : fig3 ] .
starting from the standard 2d wannier - mott model @xcite , the exciton ground state energy writes simply as : @xmath46 with @xmath47 .
an estimate of the corresponding exciton bohr radius is thus given by @xmath48 combining these equations allows for the calculation of the relative dielectric constant , knowing the binding energy @xmath15 and the reduced mass of the electron - hole pair , extracted from @xmath27+bse calculations .
it yields values of 4.47 and 5.17 for mos@xmath0 and mose@xmath0 , respectively .
the main differences from previous theoretical results arise from our choice of a 2d model , the use of our own computed @xmath15 and @xmath49 .
these values are usually larger than the estimates found in ref . using @xmath50 ( 3.44 ) or by our own direct calculation of the static dielectric constant at the dft level including local field effect ( 2.61 ) for mos@xmath0 . using the same @xmath51 approximation
we find 2.80 for mose@xmath0 in comparison with 5.17 of table [ tab1 ] .
+ continuing our discussion based on a simple hydrogen - like model , the exciton ground - state wave function can be written as : @xmath52 where @xmath53 and @xmath54 is the coordinate of the relative electron - hole motion .
the corresponding fourier transform of this ground state provides a rough estimation of how far around the k@xmath23 point the exciton state is spread in k - space . @xmath55 our representation of @xmath56 for ml mose@xmath0 in fig .
[ fig : fig3 ] suggest the following scenario : the close proximity of the @xmath30 conduction band minimum could be at the origin of the low valley polarization degree of the a - exciton 1s state luminescence , as contributions to the exciton wavefunctions away from the k@xmath57 points do not obey the strict chiral valley selections rules .
it is important to underline in fig .
[ fig : fig3 ] the striking difference with ml mos@xmath0 , where the exciton state is mainly build with k@xmath7 electronic states as the contribution from the states around the @xmath30 point at much higher energy is negligible .
we probe the exciton states in ml mose@xmath0 in 1- and 2-photon ple . we see a clear signature of the 2p state of the a- and b - exciton about 180 mev above the respective 1s exciton state in 2-photon ple .
our post - dft calculations reveal in the mose@xmath0 ml conduction band a local minimum at the @xmath30 point only 46 mev above the global minimum at the k - points . as excitons with large binding energies ( @xmath58 ev ) involve k - states far away from the k - point , the contribution of states around the @xmath30 point to optical transitions is possible . as here the chiral optical selections rules are not applicable , these states could contribute to the overall low valley polarization reported for mose@xmath0 @xcite . to further test this hypothesis , experiments that modify the band structure via strain tuning
@xcite would provide useful information .
also a systematic study of valley polarization in mos@xmath0se@xmath0 alloy monolayers @xcite will help clarifying , why mos@xmath0 shows strong and mose@xmath0 shows very weak valley polarization in non - resonant pl experiments .
we thank misha glazov for fruitful discussions and acknowledge funding from erc grant no . 306719 , anr mos2valleycontrol and programme investissements davenir anr-11-idex-0002 - 02 , reference anr-10-labx-0037-next .
i. c. gerber thanks the calmip initiative for the generous allocation of computational times , through the project p0812 , as well as the genci - cines and genci - ccrt for the grant x2014096649 .
i. c. gerber also acknowledges the cnrs for financial support .
mose@xmath0 ml flakes are obtained by micro - mechanical cleavage of a bulk mose@xmath0 crystal on an sio@xmath0/si substrates using viscoelastic stamping @xcite .
the ml region is identified by optical contrast and very clearly in pl spectroscopy .
experiments at t=4 k are carried out in a confocal microscope optimized for polarized pl experiments @xcite .
the mose@xmath0 ml is excited by picosecond pulses generated by a tunable frequency - doubled optical parametric oscillator ( opo ) synchronously pumped by a mode - locked ti : sa laser .
the typical pulse and spectral width are 1.6 ps and 3 mev respectively ; the repetition rate is 80 mhz .
the laser average power is tunable from 2 to 200 @xmath49w .
the detection spot diameter is @xmath59 m , i.e. considerably smaller than the ml diameter . for time integrated experiments
, the pl emission is dispersed in a spectrometer and detected with a si - ccd camera .
the atomic structures , the quasi - particle band structures and optical spectra are obtained from dft calculations using the vasp package @xcite .
pbe functional @xcite is used as approximation of the exchange - correlation term .
it uses the plane - augmented wave scheme @xcite to treat core electrons .
fourteen and six electrons for mo and s , se respectively are explicitly included in the valence .
all atoms are allowed to relax with a force convergence criterion below @xmath60 ev / .
after primitive cell relaxation , the optimized lattice parameters are 3.22 and 3.32 for mos@xmath0 and mose@xmath0 respectively , these values being in good agreement ( 1% ) with previous studies@xcite and slightly larger than the bulk experimental values . to sample the brillouin zone a grid of 12@xmath6112@xmath611 k - points has been used , in conjunction with a vacuum height of 17 , to take benefit of error s cancellation in the band gap estimates @xcite .
this provides exciton binding energies in reasonable agreement with experiments as suggested in different works @xcite .
a gaussian smearing with a width of 0.05 ev is used for partial occupancies , when a tight electronic minimization tolerance of @xmath62 ev is set to determine with a good precision the corresponding derivative of the orbitals with respect to @xmath63 needed in quasi - particle band structure calculations .
spin - orbit coupling was also included non - self - consistently to determine eigenvalues and wave functions as input for the full - frequency - dependent @xmath26 calculations @xcite performed at the @xmath27 level .
the total number of states included in the @xmath26 procedure is set to 600 , after a careful check of the direct band gap convergence smaller than 0.1 ev .
we have used the wannier90 code @xcite and the vasp2wannier90 interface @xcite to interpolate the band structures on a finer grid .
optical absorption spectra have been calculated using bethe - salpeter equation in the tamm - dancoff approximation , including the six highest valence bands and the eight lowest conduction bands @xcite .
|
we combine linear and non - linear optical spectroscopy at 4 k with ab initio calculations to study the electronic bandstructure of mose@xmath0 monolayers . in 1-photon photoluminescence excitation ( ple ) and
reflectivity we measure a separation between the a- and b - exciton emission of 220 mev . in 2-photon ple
we detect for the a- and b - exciton the 2p state 180 mev above the respective 1s state . in second harmonic generation ( shg )
spectroscopy we record an enhancement by more than 2 orders of magnitude of the shg signal at resonances of the charged exciton and the 1s and 2p neutral a- and b - exciton .
our post - density functional theory calculations show in the conduction band along the @xmath1 direction a local minimum that is energetically and in k - space close to the global minimum at the k - point .
this has a potentially strong impact on the polarization and energy of the excitonic states that govern the interband transitions and marks an important difference to mos@xmath0 and wse@xmath0 monolayers .
|
pg1159 stars are hot hydrogen - deficient ( pre- ) white dwarfs ( @xmath0 between 75000 and 200000k , @xmath1=5.58 ; werner 2001 ) .
they are probably the outcome of a late he - shell flash , a phenomenon that drives the currently observed fast evolutionary rates of three well - known objects ( fg sge , sakurai s object , v605 aql ) .
flash - induced envelope mixing produces a h - deficient stellar surface ( herwig etal .
the photospheric composition then essentially reflects that of the region between the h- and he - burning shells in the precursor agb star .
the he - shell flash transforms the star back to an agb star ( `` born - again agb star '' ) and the subsequent , second post - agb evolution explains the existence of wolf - rayet central stars of planetary nebulae and their successors , the pg1159 stars .
pg1159 stars provide the unique possibility to study the chemistry in the intershell region between the h- and he - burning shells that is created after complicated and still poorly - understood burning and mixing processes during the agb phase .
usually the intershell material remains hidden within the stellar interior . during the third dredge - up on the agb , however , intershell material can get mixed into the convective surface layer and appears on the stellar surface , though in rather diluted abundances .
nevertheless , this process defines the role of agb stars as contributors of nuclearly processed matter to the galaxy .
our motivation to study pg1159 stars is based on the fact that these objects directly display their intershell matter . however , the quantitative interpretation of the abundance analyses is still premature because evolutionary calculations through a final he - shell flash including a full nuclear network are not available , yet . before the advent of fuse with its outstanding fuv capabilities we have performed spectral analyses of almost all known pg1159 stars based on optical and hst - uv spectra .
we determined @xmath0 , @xmath1 , and abundances of the dominant elements , namely he , c , and o. typical values are he=33% , c=50% , o=17%
( by mass ) .
traces of nitrogen ( 1% ) or considerable amounts of residual hydrogen ( about 25% ) were found in a few stars , and in a few cases , optical as well as chandra x - ray identifications of neon were successful ( 2% ) .
generally , fuse spectra of pg1159 stars show only few photospheric ( absorption ) lines , mainly from heii , civ , ovi , and nevii .
some of them show shallow nv lines and in many of them we see sulfur .
the svi 933/944 doublet in k1 - 16 suggests a solar abundance , which is in line with the expectation that s is not affected by the s - process .
we also identify silicon in some objects ( reiff etal .
2005 ) , but detailed abundance analyses remain to be done . the first surprising result of fuse spectroscopy was the detection of a significant iron deficiency ( 12 dex ) in the three best studied pg1159 stars ( miksa etal .
apparently , fe was transformed to heavier elements in the intershell region of the agb star by n - captures from the neutron source @xmath2c(@xmath3,n)@xmath4o ( herwig etal .
subsequently , several other studies have also revealed an fe - deficiency in [ wc ] type central stars ( see werner etal .
2003 ) , which matches our picture that these stars are immediate pg1159 star progenitors .
the next important result was accomplished by the identification of one of the strongest absorption lines seen in fuse spectra of most pg1159 stars , located at 973.3 .
it is a nevii line ( fig.1 ) that allowed to assess the neon abundance in a large sample of objects ( werner etal .
it turns out that neon is strongly overabundant , ( 2% , i.e. , 20 times solar ) .
this result clearly confirms the idea that pg1159 stars indeed exhibit intershell matter .
neon is produced in the he - burning environment by two @xmath3-captures of nitrogen , which itself resulted from previous cno burning : @xmath5n@xmath6f@xmath7o@xmath8ne .
there are still at least ten photospheric lines in the fuse spectra of pg1159 stars that remain unidentified .
some of them may stem from yet unknown nevii lines .
the latest identification is a feature at 1139.5 , which appears rather strong in some objects .
we found that it is a line from highly ionized fluorine ( fig.1 ) and derived large overabundances ( up to 250 times solar ) for a number of pg1159 stars .
we also could identify this line in `` normal '' h - dominated central stars and , in contrast , find about solar fluorine abundances ( werner etal .
this again is a clear proof that we see intershell matter on pg1159 stars . according to recent calculations by lugaro etal .
( 2004 ) , their stellar models show an effective fluorine production and storage in the intershell , leading to abundances that are comparable to the observed pg1159 abundances of fluorine .
the general problem for fluorine production is that @xmath9f , the only stable f isotope , is rather fragile and readily destroyed in hot stellar interiors by hydrogen via @xmath9f(p,@xmath3)@xmath4o and helium via @xmath9f(@xmath3,p)@xmath10ne .
the nucleosynthesis path for f production in he - burning environments of agb and wolf - rayet stars is @xmath5n(@xmath3,@xmath11)@xmath12f(@xmath13)@xmath12o(p,@xmath3)@xmath14n(@xmath3,@xmath11)@xmath9f .
this underlines that agb stars which dredge up material from the intershell are contributing to the galactic f content ( together with wolf - rayet stars and type ii sne ) .
this is completely in line with the detected f overabundances ( up to 30 times solar ) found in agb stars ( jorissen etal .
to what extent pg1159 stars themselves return f to the ism remains to be estimated .
the life time of a born - again agb star is short in comparison to a usual agb star , however , the f fraction in the mass lost by a wind of the former is much higher .
|
fuse spectroscopy has proved that extremely hot hydrogen - deficient post - agb stars ( pg1159 stars ) display matter on their surface that usually remains hidden in the region between the h- and he - burning shells of the former agb star .
hence , the spectral analysis of pg1159 stars allows to study directly the chemistry of this intershell region which is the outcome of complicated burning and mixing processes during agb evolution .
detailed abundance determinations provide constraints for these processes which are still poorly understood . with fuse
we have discovered high neon and fluorine overabundances . there is also a significant iron deficiency , which may be caused by s - process neutron capture transforming iron into heavier elements .
|
neumann and zagier @xcite study the variation of the volume function on a cusped hyperbolic 3manifold of finite volume using a decomposition of the manifold into hyperbolic ideal tetrahedra .
this is based on a construction by thurston @xcite , which associates to the underlying topological ideal triangulation a
_ parameter space _ of shapes of ideal hyperbolic tetrahedra satisfying certain polynomial equations .
such a parameter space can be associated to any ideal triangulation of a non - compact 3-manifold with torus cusps , and several authors have studied this affine algebraic set in this generality ( see , for instance , yoshida @xcite , francaviglia @xcite , tillmann @xcite , segerman @xcite ) . a key step in many applications is the construction of a so - called pseudo - developing map for a given point on the parameter space in order to produce a representation of the fundamental group of the manifold into the group of orientation preserving isometries of hyperbolic 3space .
the map is called a pseudo - developing map rather than a developing map because it is not necessarily locally injective .
more recently , luo @xcite initiated the study of the parameter space for arbitrary ideally triangulated manifolds . in luo s setting , they arise from ( possibly semi - simplicial ) triangulations of closed 3-manifolds by removing the vertices .
this raises the question of whether the existence of pseudo - developing maps , their associated holonomies and continuous extensions depend on topological hypotheses .
this note considers the most general setting , and has the following theorem as its core result .
the relevant definitions can be found in sections [ sec : definitions ] and [ sec : defo ] .
[ thm : essential edges ] let @xmath0 be a topologically finite , orientable 3manifold with ideal triangulation .
if there is a solution to the hyperbolic gluing equations , then all edges in the triangulation are essential .
[ cor : essential edges ] let @xmath0 be a topologically finite , orientable 3manifold with ideal triangulation .
if there is a solution to the hyperbolic gluing equations , then associated to it are a pseudo - developing map @xmath1 and a representation of @xmath2 into @xmath3 which makes this map equivariant .
luo @xcite also introduced _
generalised hyperbolic gluing equations _
, which state that the product of all shape parameters around an ideal edge is either @xmath4 or @xmath5 the sign may be different at different edges .
this can be generalised further by choosing , for each edge @xmath6 an element @xmath7 and requiring the product of all shape parameters around @xmath8 to equal @xmath9 the restriction of @xmath10 to @xmath11 is natural : if the shape parameters of all tetrahedra have positive imaginary parts , then this gives a ( possibly incomplete ) hyperbolic cone manifold structure on @xmath0 with cone angle @xmath12 ( for some @xmath13 ) around @xmath14 the most interesting case arises when each @xmath10 is a root of unity , but we will consider the general case throughout .
@xmath11valued " gluing equations are determined by the vector @xmath15 and are called the _
@xmath16hyperbolic gluing equations _ , since they depend on the choice of @xmath17 let @xmath18 where @xmath19 is the order of @xmath9 if one drills out the edges in @xmath20 one obtains a handlebody @xmath21 with a natural epimorphism @xmath22 we obtain a representation @xmath23 for each solution to the @xmath16hyperbolic gluing equations . if each order @xmath24 is finite , then we obtain a natural branched ( not necessarily finite ) cover @xmath25 of @xmath0 corresponding to the kernel of @xmath26 the branch locus is contained in the 1skeleton , and the cover has branch index @xmath24 at @xmath14 if some order is infinite , one can still define @xmath27 but this will have an edge of infinite degree and the points on such an edge of infinite degree are not manifold points . if all orders are equal to @xmath28 then @xmath29 theorem [ thm : essential edges ] is thus a special case of the following result : [ thm : branched cover ] let @xmath0 be a topologically finite , orientable 3manifold with ideal triangulation .
if there is a solution to the @xmath16hyperbolic gluing equations for @xmath20 then all edges in the induced ideal triangulation of @xmath25 are essential . treating each @xmath10 as an additional
, circle - valued variable gives the _ cone - hyperbolic gluing equations_. the existence of interesting solutions to them comes from the following observation : [ cor : existence ] given any ideally triangulated , topologically finite , orientable 3manifold @xmath20 there is a complete , non - compact hyperbolic cone - manifold structure on @xmath0 with singular locus contained in @xmath30 and having volume exactly the number of tetrahedra in @xmath0 times the volume of the regular hyperbolic ideal tetrahedron . *
outline : * triangulations , pseudo - manifolds and essential edges are discussed in section [ sec : definitions ] . the deformation variety and its generalisations are defined in section [ sec : defo ] , and yoshida s construction is recalled .
the handlebody construction as well as the proofs of the main results can be found in section [ sec : handlebody ] .
examples are given in section [ sec : zoo ] .
* acknowledgements : * the authors thank the referee for useful comments .
the second author thanks feng luo for helpful discussions .
this work was supported under the australian research council s discovery funding scheme ( project number dp1095760 ) .
an ideal triangulation @xmath31 of the topologically finite 3-manifold @xmath0 consists of a pairwise disjoint union of standard euclidean 3simplices , @xmath32 together with a collection @xmath33 of euclidean isometries between the 2simplices in @xmath34 termed _ face pairings _ , such that @xmath35 it is well - known that every non - compact , topologically finite 3manifold admits an ideal triangulation .
denote @xmath36 the associated _ pseudo - manifold _ ( or _ end - compactification _ of @xmath0 ) with quotient map @xmath37 let @xmath38 be a @xmath39simplex in @xmath40 then @xmath41 may be a singular @xmath39simplex in @xmath42 denote by @xmath43 the set of all ( possibly singular ) @xmath39simplices in @xmath42 an _
ideal @xmath39simplex _ is a @xmath39simplex with its vertices removed .
the vertices of the @xmath39simplex are termed the _
ideal vertices _ of the ideal @xmath39simplex .
similarly for singular simplices .
the standard terminology of ( ideal ) edges , ( ideal ) faces and ( ideal ) tetrahedra will be used for the singular simplices in @xmath0 and @xmath42 in this note , it is assumed throughout that @xmath0 is oriented and that all singular simplices in @xmath0 are given the induced orientations .
it follows that the link of each vertex in @xmath44 is an orientable surface .
the case in which each vertex link is a torus and @xmath0 has a complete , hyperbolic structure supported by the ideal triangulation is the most common setting for the study of thurston s hyperbolic gluing equations , see @xcite .
the case in which each vertex link is a sphere and @xmath44 is a closed hyperbolic 3manifold is treated in @xcite .
this note will not make any of these additional assumption .
it is often convenient to start with the 3dimensional , closed , orientable pseudo - manifold @xmath44 with ( possibly singular ) triangulation @xmath45 then @xmath46 is an ideally triangulated , non - compact , orientable , topologically finite 3manifold .
the _ degree of an edge _ @xmath8 in @xmath47 @xmath48 is the number of 1simplices in @xmath49 which map to @xmath14 given the edge @xmath8 in @xmath47 there is an associated _ abstract neighbourhood @xmath50 _ of @xmath14 this is a ball triangulated by @xmath51 3simplices , having a unique interior edge @xmath52 and there is a well - defined simplicial quotient map @xmath53 taking @xmath54 to @xmath14 this abstract neighbourhood is obtained as follows .
if @xmath8 has at most one pre - image in each 3simplex in @xmath34 then @xmath50 is obtained as the quotient of the collection @xmath55 of all 3simplices in @xmath49 containing a pre - image of @xmath8 by the set @xmath56 of all face pairings in @xmath33 between faces containing a pre - image of @xmath14 there is an obvious quotient map @xmath57 which takes into account the remaining identifications on the boundary of @xmath58 if @xmath8 has more than one pre - image in some 3simplex , then multiple copies of this simplex are taken , one for each pre - image . the construction is modified accordingly , so that @xmath50 again has a unique interior edge and there is a well defined quotient map @xmath59 complete details can be found in @xcite , section 2.3 . as above , let @xmath44 be a 3dimensional , closed , orientable pseudo - manifold with ( possibly singular ) triangulation @xmath60 and @xmath61 let @xmath62 be an open regular neighbourhood of @xmath63 in @xmath47 chosen in such a way that @xmath64 meets each singular 3simplex @xmath65 in @xmath44 in precisely four pairwise disjoint normal triangles , one at each of its corners . hence @xmath66 has the natural cone structure @xmath67 ) / \sim$ ] where @xmath68 if @xmath69 lie on a connected component of @xmath70 then @xmath71 is termed a _ compact core _ of @xmath72 it follows that @xmath44 can be viewed as obtained from @xmath73 by taking each connected component @xmath74 of @xmath75 and either collapsing it to a point or by attaching the cone over @xmath74 to a point .
neumann and yang @xcite call this the _ end - compactification _ of @xmath72 we have @xmath76 let @xmath77 be the space obtained from the universal cover @xmath78 of @xmath73 by attaching the cone over each connected boundary component to a point .
we then have natural inclusions @xmath79 and @xmath77 is termed the _ end - compactification of @xmath80 with respect to @xmath72 _ note that @xmath77 is also simply connected since adding cones over connected spaces does not increase the fundamental group .
it is hoped that the notation and terminology does not lead to confusion .
for instance , when @xmath0 is hyperbolic , then @xmath80 is an open ball and the natural compactification of this open ball ( without reference to @xmath0 ) is homeomorphic to the 3ball , whilst @xmath77 is @xmath80 with countably many points added . also , there are many examples where @xmath44 is simply connected ; for instance if @xmath73 is the complement of a knot or link in @xmath81 the space @xmath77 has a natural decomposition into 3simplices coming from the decomposition of @xmath78 into truncated 3simplices and the coning construction .
lifting the ideal triangulation of @xmath0 to @xmath80 gives a natural decomposition into ideal 3simplices .
it follows from the construction that the ideal triangulation of @xmath80 is precisely the restriction of the triangulation of @xmath77 to @xmath82 in particular , we have a well - defined simplicial map @xmath83 a triangulation @xmath31 of @xmath44 is said to be _ almost non - singular _ if no 3simplex has two of its edges identified . in this case , the only self - identifications of a 3simplex are at the vertices .
a triangulation @xmath31 of @xmath44 is _ non - singular _ if no 3simplex has any self - identifications .
a triangulation @xmath31 of @xmath44 is said to be _ virtually almost non - singular _ if the induced triangulation of @xmath84 is almost non - singular , and _ virtually non - singular _ if the induced triangulation of @xmath84 is non - singular .
note that a triangulation @xmath31 of @xmath44 is virtually almost non - singular if and only if every ideal 3simplex in @xmath80 is embedded .
an edge @xmath85 is _ null - homotopic _ if and only if there is a map @xmath86 such that @xmath87 equivalently , @xmath8 is null - homotopic in @xmath44 if and only if it represents the trivial element in @xmath88 the intersection @xmath89 is _ homotopic into @xmath75 _ if and only if there is an arc @xmath90 such that @xmath91 is homotopic to @xmath92 by a fixed - endpoint homotopy . to simplify terminology , we will say that the edge @xmath8 in @xmath44 is _ essential _ if @xmath93 is not homotopic into @xmath94 and it is _ not essential _ otherwise .
this reflects standard terminology for ideal edges in ideal triangulations .
[ lem : virt non sing iff essential ] @xmath31 is virtually non - singular if and only if every edge in @xmath44 is essential .
suppose the edge @xmath8 is not essential .
then @xmath93 is homotopic into @xmath95 the disc in @xmath73 lifts to a disc in @xmath78 with part of its boundary on a lift @xmath96 of @xmath8 and the remainder of its boundary on a connected component of @xmath97 hence @xmath96 is an edge in @xmath77 with both end - points at the same vertex .
thus , any tetrahedron containing @xmath96 is not embedded and therefore @xmath31 is not virtually non - singular .
conversely , suppose @xmath31 is not virtually non - singular .
there is a tetrahedron in @xmath98 with self - identifications and hence an edge @xmath8 with both end - points at the same vertex .
the end - points of @xmath99 lie on the same boundary component of @xmath100 and hence can be connected by an arc @xmath101 since @xmath78 is simply connected , the loop @xmath102 bounds an immersed disc in @xmath103 let @xmath104 be the covering map .
it follows that @xmath105 is homotopic to @xmath106 and hence the edge containing @xmath105 is not essential .
[ not_ess_implies_null_htpc ] if @xmath8 is not essential in @xmath47 then @xmath8 is _ null - homotopic _ in @xmath42 if @xmath89 is homotopic into @xmath107 then both of its end - points lie on the same boundary component , @xmath108 of @xmath109 hence the map @xmath110 can be extended to a continuous map @xmath111 by coning it over @xmath92 to the vertex of @xmath44 corresponding to @xmath112 the converse of lemma [ not_ess_implies_null_htpc ] is not true .
for instance the ideal triangulation of a knot or link in @xmath113 gives rise to a simply connected pseudo - manifold , but there are many edges which are not homotopic into the boundary . for instance , thurston s ideal triangulation of the figure eight knot complement yields a simply connected pseudo - manifold @xmath44 with the property that every edge is null - homotopic and essential .
let @xmath114 be the standard 3simplex with a chosen orientation .
suppose the edges from one vertex of @xmath114 are labeled by @xmath115 @xmath116 and @xmath117 so that the opposite edges have the same labeling .
then the cyclic order of @xmath115 @xmath116 and @xmath117 viewed from each vertex depends only on the orientation of the 3simplex .
it follows that , up to orientation preserving symmetries , there are two possible labelings , and we fix one of these labelings .
the labels are termed _
shape parameters_. suppose @xmath118 for each @xmath119 fix an orientation preserving simplicial map @xmath120 let @xmath121 and let @xmath122 be the number of edges in @xmath123 which have label @xmath124 for each @xmath125 define @xmath126 and for each @xmath127 let @xmath128 setting @xmath129 gives the _ parameter relations _ , and setting @xmath130 gives the _ hyperbolic gluing equations_. for a discussion and geometric interpretation of these equations , see @xcite .
the parameter relations imply that @xmath131 the _ deformation variety @xmath132 _ is the variety in @xmath133 defined by the hyperbolic gluing equations together with the parameter relations .
thurston s original parameter space is obtained by choosing a coordinate from each coordinate triple in the above . in calculations
, we will often use such a smaller coordinate system .
theorem [ thm : essential edges ] is equivalent to the statement that @xmath134 implies that all edges are essential .
let @xmath135 @xmath31 and @xmath136 be as defined above .
given @xmath137 each ideal tetrahedron in @xmath0 has edge labels which can be lifted equivariantly to @xmath138 following @xcite , we would like to define a continuous map @xmath139 which maps every ideal tetrahedron @xmath38 in @xmath80 to an ideal hyperbolic 3simplex @xmath140 such that the labels carried forward to the edges of @xmath141 correspond to the shape parameters of @xmath141 determined by its hyperbolic structure ; see @xcite for the geometry of hyperbolic ideal tetrahedra .
thus , _ we need to assume that @xmath31 is virtually almost non - singular _ , since no ideal hyperbolic simplex has edges identified with each other .
it will be necessary to have a consistent choice of parameterisation for the maps from ideal tetrahedra in @xmath80 to @xmath142 following thurston @xcite , we ensure this by assuming that the map @xmath143 always is a _
straight map_. see @xcite for the details , which play no role in the following .
each ideal 3simplex in @xmath80 inherits edge labels from @xmath144 choose a tetrahedron @xmath38 in @xmath80 and an embedding of @xmath38 into @xmath145 of the specified shape . for each tetrahedron of @xmath80 distinct from @xmath38 and which has a face in common with @xmath146
there is a unique embedding into @xmath145 which coincides with the embedding of @xmath38 on the common face and which has the shape determined by @xmath144 if the shape parameters of the new tetrahedron give an ideal hyperbolic tetrahedron of orientation opposite to that of the first one , then this map is not locally injective along the common face .
hence the map is called a _ pseudo - developing map _ rather than a developing map . starting with an embedding of @xmath147
there is a unique way to extend this to a well - defined , continuous map @xmath148 such that each ideal tetrahedron in @xmath80 is mapped to a hyperbolic ideal tetrahedron of the specified shape . since @xmath80 is simply connected and each abstract edge neighbourhood is embedded , it follows from the hyperbolic gluing equations that the map is well - defined .
the reader may find pleasure in doing this exercise or consult @xcite for a full treatment .
[ lem : special case ] if @xmath149 and @xmath31 is virtually almost non - singular , then @xmath31 is virtually non - singular ( hence all edges in @xmath44 are essential )
. let @xmath150 the yoshida map @xmath151 is well defined since @xmath31 is virtually almost non - singular .
suppose the ideal tetrahedron @xmath38 in @xmath80 has two ideal vertices at the same vertex of @xmath152 let @xmath153 and @xmath154 denote the normal triangles in @xmath155 dual to these vertices .
since @xmath153 and @xmath154 are in the link of the same vertex of @xmath156 there is a path in @xmath157 from @xmath153 to @xmath154 passing through finitely many normal triangles .
this path corresponds to a finite sequence @xmath158 of ideal tetrahedra in @xmath159 and a corresponding sequence @xmath160 of ideal vertices of the @xmath161 , dual to the sequence of normal triangles .
the map @xmath151 extends to map these ideal vertices into @xmath162 , and as we develop along the sequence of tetrahedra , we see that @xmath163 . but
@xmath164 and @xmath165 are two distinct vertices of @xmath38 , so the image of @xmath38 can not be an ideal hyperbolic 3simplex , and thus the yoshida map is not well - defined .
[ cor : yoshida depends on ideal points ] if the yoshida map @xmath166 is well - defined , then it extends to a continuous map @xmath167 suppose that the yoshida map is well - defined and , in particular , that @xmath31 is virtually almost non - singular .
we will see ( in corollary [ cor : essential edges ] ) that these hypotheses hold if @xmath134 . for each @xmath137 the yoshida map @xmath151
can be used to define a representation @xmath168 as follows ( see @xcite ) .
a representation into @xmath169 is an action on @xmath170 and this is the unique representation which makes @xmath151 @xmath2equivariant : @xmath171 for all @xmath172 @xmath173 thus , @xmath174 is well defined up to conjugation , since it only depends upon the choice of the embedding of the initial tetrahedron @xmath175 this yields a well defined map @xmath176 from the deformation variety to the @xmath169character variety .
it is implicit in @xcite that @xmath177 is algebraic ; see @xcite for details using a faithful representation of @xmath178 note that the image of each peripheral subgroup under @xmath174 has at least one fixed point on the sphere at infinity .
the representation associated to a solution of the hyperbolic gluing equations may be reducible , or even trivial .
for instance , the triangulation of @xmath113 with two tetrahedra obtained by identifying the boundary spheres of the tetrahedra in the natural way has a curve of solutions to the hyperbolic gluing equations , and the associated representations of the fundamental group of @xmath113 minus four points are all trivial .
the representation arising from yoshida s construction can be understood using _ elementary face pairings _ as follows .
an elementary face pairing of the hyperbolic ideal tetrahedron @xmath179 is an element of @xmath169 taking one face of @xmath179 to another .
if the deck transformation @xmath180 takes the ideal triangle @xmath181 in @xmath182 to the ideal triangle @xmath183 and @xmath184 is a sequence of tetrahedra with the property that @xmath185 @xmath186 and consecutive tetrahedra share a face distinct from @xmath181 and @xmath187 then there is an associated product of elementary face pairings , one for each @xmath188 such that @xmath189 is their product .
this fact is used in @xcite to show that @xmath177 is algebraic ; it will be used below for a different purpose . for each @xmath190
we introduce a variable @xmath191 and define @xmath192 setting @xmath193 gives the _ cone - hyperbolic gluing equation _ for @xmath194 since the product of all shape parameters associated to all edges equals @xmath28 we have the following consequence : @xmath195 we also define the _ holonomy around @xmath196 _ to be @xmath197 for a triangulation with @xmath198 tetrahedra and @xmath199 edges , the _ cone - deformation variety @xmath200 _ is the variety in @xmath201 defined by the cone - hyperbolic gluing equations together with the parameter relations .
the cone - deformation variety is non - empty for _ any _ triangulation . indeed ,
if @xmath202 then all of @xmath115 @xmath116 and @xmath117 are roots of unity and specify the regular hyperbolic ideal 3simplex .
this is the unique hyperbolic ideal 3simplex with the property that all shape parameters are roots of unity . assigning values to all shape parameters in this way
, one can solve @xmath203 for @xmath204 uniquely for each @xmath14 it turns out that this solution has particularly nice properties , giving us observation [ cor : existence ] .
identify each ideal 3simplex in @xmath205 with the regular ideal 3simplex .
gluing faces in pairs by hyperbolic isometries according to the face pairings gives a hyperbolic structure on @xmath0 minus edges . around the edges there is no shearing , and
the total angle around edge @xmath8 is its degree times @xmath206 hence there is a non - compact hyperbolic cone - manifold structure on @xmath0 with singular locus contained in @xmath207 near the ideal vertices of the tetrahedra , one can consistently choose horospherical triangles that match up and the structure is therefore complete .
denote the projections to the factors @xmath208 and @xmath209 given @xmath210 we let @xmath211 then @xmath212 may be empty , and @xmath213
consider the compact manifold @xmath214 , where @xmath215 is a set of pairwise disjoint open tubular neighbourhoods of the edges @xmath216 .
the manifold @xmath21 is a handlebody , and inherits a cell decomposition from @xmath31 into _ doubly truncated tetrahedra _ : truncated at the vertices and at the edges .
see figure [ trunc_tetra.pdf ] for a picture of a doubly truncated tetrahedron .
a doubly truncated tetrahedron has four _ boundary hexagonal faces _ on @xmath217 , six _ rectangular faces _ on @xmath218 and four _ interior hexagonal faces_. consider the decomposition of @xmath219 into the union of the _ vertex boundary _ @xmath220 ( composed of boundary hexagonal faces ) and the _ edge boundary _ @xmath221 ) ( composed of rectangular faces ) .
the edge boundary consists of a pairwise disjoint union of annulus components , one for each edge of @xmath73 . for any cover @xmath222 of @xmath21 , define a topological space @xmath223 as follows . lift the decomposition of @xmath21 to @xmath222 .
the boundary of @xmath222 decomposes into _ vertex boundary _ components made up out of boundary hexagonal pieces , and _ edge boundary _ components made up out of rectangular faces . for each edge boundary component of @xmath224
we fix a product structure which identifies it with @xmath225.$ ] then the edge boundary components of @xmath222 are either of the form @xmath225 $ ] or @xmath226,$ ] and the product structure is preserved by the deck transformations .
we form @xmath223 by first collapsing each edge boundary component of @xmath227 by projection to the @xmath228 $ ] factor , to form a topological space @xmath229 .
next we collapse each component of the boundary of @xmath229 to form @xmath223 . as in the construction of @xmath44 from @xmath73 , we can equivalently construct @xmath223 from @xmath222 by coning rather than collapsing .
note that @xmath230 .
also note that @xmath223 is a union of tetrahedra identified along faces .
unlike @xmath44 or @xmath84 , @xmath223 can have edges incident with infinitely many tetrahedra .
we will be interested in the universal cover @xmath231 of @xmath224 and in @xmath232 . since @xmath21 is a handlebody , @xmath231 is homotopy equivalent to a tree . notice that the construction gives a natural quotient map @xmath233 [ lem : non - singular in handlebody cover ] every tetrahedron in @xmath232 is non - singular .
if there is a tetrahedron with self - identifications of any kind then there will be some pair of its vertices that are identified , and thus an edge with its endpoints being the same point .
suppose that we have such an edge @xmath8 .
truncate all of the tetrahedra of @xmath232 at the vertices to form a topological space @xmath234 ( a `` core '' of @xmath232 , analogous to the compact core @xmath73 of @xmath44 ) .
both points @xmath235 are on the same component of @xmath236 , so we can choose a path @xmath92 contained in a single component of @xmath236 whose endpoints are @xmath237 and @xmath238 .
we may deform @xmath92 slightly so that it does not pass through any edges other than @xmath8 .
we can now drill out the neighbourhoods @xmath239 of the edges of @xmath234 to get back @xmath231 , with the curve @xmath92 contained in a component of @xmath240 , and meeting @xmath241 only at its endpoints , which are on the edge boundary near the two ends of @xmath8 .
we may assume that the path @xmath92 starts in some ( doubly truncated ) tetrahedron @xmath242 at an intersection between a boundary hexagonal face and a rectangular face , traverses through the tetrahedra of @xmath231 along @xmath240 , and then returns back to @xmath242 , ending at the other end of the rectangular face .
consider such a path @xmath92 that visits a minimal number of tetrahedra .
since the tetrahedra form a tree , there must be at least one `` leaf '' tetrahedron @xmath38 in the path .
that is , the path enters @xmath38 from one face gluing ( at the interior hexagonal face @xmath243 ) and exits at the same face gluing . the path @xmath92 is restricted to lie in @xmath240 , and so it enters @xmath38 at the edge of one of the three boundary hexagonal faces of @xmath38 adjacent to @xmath243 .
there is no path within @xmath38 along @xmath240 from one of the boundary hexagonal faces to any of the others , so it must exit @xmath38 at the same edge .
but then the part of @xmath92 in @xmath38 could be homotoped away , and the path was not minimal .
this gives a contradiction .
since there are no gluing consistency conditions to satisfy for the tetrahedra in @xmath232 ( because every edge of @xmath232 is of infinite degree ) , we can assign _ arbitrary _ shape parameters @xmath244 to the tetrahedra of @xmath232 and define a pseudo - developing map @xmath245 this follows as in the proof of lemma [ lem : non - singular in handlebody cover ] from the fact that @xmath232 is homotopy equivalent to a tree .
in particular , we can build the developing map by starting with a tetrahedron @xmath242 of @xmath232 and any ideal hyperbolic tetrahedron in @xmath145 , and then developing along any non - backtracking path of tetrahedra in @xmath232 starting with @xmath242 , using any non - degenerate shapes of ideal hyperbolic tetrahedra .
as we develop , we never have any consistency conditions to satisfy , because the tetrahedra of @xmath232 form a tree .
suppose that the parameters for the tetrahedra in @xmath232 are lifts of the parameters of the tetrahedra of @xmath246 then there exists a well - defined representation @xmath247 moreover , this representation makes @xmath248 equivariant if and only if for each edge @xmath8 in @xmath47 the holonomy around @xmath8 , @xmath249 is an element of @xmath250 the map @xmath251 has been defined as in the yoshida construction , and we would like to define @xmath252 in a similar fashion . consider a triangle @xmath253 of @xmath232 , and its image @xmath254 under the deck transformation @xmath255 .
the image of the three vertices of @xmath253 under @xmath251 give a triplet of distinct points on @xmath256 , and the three vertices of @xmath254 give another triplet .
we define @xmath257 to be the unique element of @xmath169 which maps the first triplet to the second .
since the parameters for the tetrahedra in @xmath232 are lifts of the parameters of the tetrahedra of @xmath21 , this definition is independent of the choice of @xmath258 this proves the existence of @xmath26 it follows from the construction that @xmath259 for all @xmath260 and @xmath261 _ except possibly _ for those @xmath262 that are contained in the 1skeleton . in @xmath263
we have a product structure on the lifts of the edge boundary components of @xmath224 which is preserved by the deck transformations .
consider a deck transformation @xmath264 which preserves the edge boundary component @xmath74 of @xmath265 now in @xmath266 is mapped to a 1simplex , @xmath267 and @xmath268 is a geodesic in @xmath269 composing @xmath251 with an isometry of @xmath170 we may assume that @xmath270.$ ] since @xmath271 preserves the product structure , @xmath272 for each @xmath273 . thus @xmath251 is equivariant with respect to @xmath252 if and only if @xmath257 acts on @xmath145 by fixing @xmath274 pointwise , i.e.acts as a ( possibly trivial ) rotation about @xmath275 in @xmath276 @xmath271 acts as a translation on the set of all 3simplices incident with @xmath277 a connected fundamental domain for this action consists of a finite number of 3simplices meeting in @xmath267 and their number equals the degree of the corresponding edge @xmath8 in @xmath42 let @xmath38 be a 3-simplex in this fundamental domain .
there is a unique isometry taking @xmath278 to @xmath279 and this is the rotation with eigenvalue precisely the product @xmath249 of all shape parameters at @xmath8 in @xmath42 hence @xmath257 is a rotation if and only if @xmath249 is an element in @xmath250 [ defn : cone - manifold cover ] given @xmath280 @xmath248 and @xmath281 as above , let @xmath282 and @xmath283 since @xmath284 acts simplicially , @xmath285 and @xmath25 have natural decompositions into simplices .
if @xmath24 is finite for each @xmath286 then each edge in @xmath25 has finitely many 3simplices incident with it ; namely if @xmath287 then @xmath288 in this case , the natural map @xmath289 is a ( not necessarily finite ) branched cover , with branch locus contained in the 1skeleton and group of deck transformations isomorphic to @xmath290 if @xmath24 is infinite for some @xmath291 then the points of @xmath25 mapping to @xmath8 are not manifold points .
we first give a proof of theorem [ thm : essential edges ] , and then modify it for the general case .
note that the inclusion map @xmath292 induces an epimorphism @xmath293 .
the kernel of this map is generated by certain loops around the annuli in @xmath294 .
for each edge @xmath8 we denote by @xmath295 a loop around the annulus corresponding to @xmath8 , where @xmath295 is a level set in the product structure , with an arbitrary orientation .
given lemma [ lem : special case ] , it suffices to assume for contradiction that there is a solution @xmath296 , but that @xmath31 is not virtually almost non - singular .
in particular , by lemma [ lem : virt non sing iff essential ] , some edge is inessential .
we work with the compact core @xmath73 of @xmath0 .
give the tetrahedra in @xmath231 the shape parameters inherited from @xmath297 ( which are first inherited by @xmath224 then lifted to @xmath231 and finally inherited by @xmath232 ) .
associated are a developing map @xmath298 and a representation @xmath299 recall that we specify a quotient map @xmath300 by collapsing edge boundary components @xmath226 $ ] to the second factor , then collapsing each vertex boundary component
. now let @xmath8 be an inessential edge in @xmath0 .
consider a path @xmath301\rightarrow h$ ] , which is the core curve of a rectangular face of a doubly truncated tetrahedron incident to the edge boundary annulus corresponding to @xmath8 .
let @xmath302\rightarrow h$ ] be the result of pushing @xmath303 slightly off the rectangle into the tetrahedron , keeping the endpoints on the boundary hexagonal faces .
then @xmath91 is parallel to @xmath8 , as in figure [ trunc_tetra.pdf ] . with the path @xmath91 parallel to the boundary rectangle corresponding to an edge of @xmath44.,scaledwidth=30.0% ]
since @xmath8 is inessential in @xmath73 , there is a homotopy of it into @xmath217 , fixing its endpoints . as @xmath91 ( viewed as a path in @xmath73 since @xmath304 ) is parallel to @xmath8
, we can use the same homotopy to homotope @xmath91 into @xmath217 , fixing its endpoints , by first homotoping across the rectangle with @xmath91 and @xmath8 as one pair of opposite sides , and the other sides on @xmath217 in the obvious way . viewing the homotopy as a map @xmath305
, we may deform it by a small amount to produce a homotopy @xmath306 transverse to @xmath216 , and moreover so that the intersection of @xmath307 with @xmath308 consists of a finite number of circles , each of which ( pulled back through @xmath306 ) bounds a disk in @xmath309 which is contained in a neighbourhood @xmath239 and intersects @xmath310 once , transversely .
thus each circle goes around the cylindrical part of @xmath217 corresponding to an edge of @xmath0 , and is homotopic to a @xmath295 .
in particular let @xmath311\rightarrow { \partial}c$ ] be the path after the homotopy , and by deforming if necessary we may assume that @xmath92 in fact maps into @xmath220 .
see figure [ homotopy_disk.pdf ] . between @xmath91 and @xmath92 .
the parts mapping into @xmath21 are shaded lightly , and the disks mapping into the neighbourhood of an edge in @xmath73 shaded more darkly .
, scaledwidth=60.0% ] we choose a base - point for @xmath284 at @xmath312 now @xmath313 is a sphere with a number of boundary components , where @xmath314 and all other boundary components correspond to circles @xmath315 \to h,$ ] which are fibres in the product structure of the edge boundary . for
each @xmath316 choose a non self - intersecting path @xmath317 on @xmath313 from @xmath318 to @xmath319 we can choose these paths to be disjoint apart from at @xmath318 . then @xmath313 gives the following relation between the elements in @xmath320 represented by these loops : @xmath321 = \prod [ \delta_k\gamma_k \delta_k^{-1}].\ ] ] choose a lift of @xmath318 to @xmath231 and lift @xmath91 accordingly .
let @xmath322 be the composition with the quotient @xmath323 then @xmath324\rightarrow \overline{\mathbb{h}}^3 $ ] is its image under the pseudo - developing map .
define @xmath325 , @xmath326 and @xmath327 similarly , choosing the lift @xmath325 to start from the same point as @xmath328 does . since the pseudo - developing map is well - defined and since the endpoints of @xmath8 are on the same components of @xmath220 as the respective ends of @xmath91 , @xmath329 and @xmath330 are distinct points on @xmath331 . with our choice of base - points ,
@xmath332)$ ] is an isometry of @xmath333 taking @xmath329 to @xmath334 we now claim that for each term @xmath335 in the right hand side of relation ( [ eq : disc relator ] ) , we have @xmath336)=1.$ ] indeed , @xmath336)$ ] can be expressed as a product of elementary face pairings .
now the products arising from @xmath317 and @xmath337 are inverses .
hence @xmath336)=1 $ ] if and only if the product corresponding to @xmath316 is trivial .
but @xmath316 gives a rotation with eigenvalue the product of all shape parameters around the edge corresponding to that loop . by hypothesis
, this product equals @xmath28 and hence the claim .
but then @xmath332)=1,$ ] contradicting the fact that it acts non - trivially on @xmath338 this completes the proof of theorem [ thm : essential edges ] . for the general case
, the above proof can be applied using a punctured sphere in @xmath339 the curves @xmath316 can only correspond to edges of @xmath25 which have finite degree , and the eigenvalue of the associated rotation is precisely of the form @xmath340 giving the same contradiction as above .
the main restriction in using yoshida s construction is the fact that ideal simplices in the universal cover are mapped to hyperbolic ideal simplices in hyperbolic 3space .
daryl cooper pointed out to us that the proofs in this section could be given using local arguments by subdividing the ideal tetrahedra in @xmath72 the authors feel that the approach using triangulations is more appropriate for this volume and the man it is dedicated to .
we give three examples of triangulations of @xmath113 with knotted or linked edges , which exhibit interesting features .
special hyperbolic cone - manifold structures on the first triangulation can be found in work by boileau - porti @xcite and hodgson @xcite , and the two concluding triangulations were provided by bus jaco .
consider the one - tetrahedron three - edges triangulation of @xmath113 shown in figure [ bph_cone_mfld.pdf](a ) .
we have @xmath341 since there are degree one edges . giving the parameter @xmath342 to the degree - one edges @xmath343 and @xmath344 , and letting @xmath345 be the degree - four edge ,
the cone - deformation variety is : @xmath346 there is one ideal point ( corresponding to @xmath347 ) and one flat solution ( corresponding to @xmath348 ) .
the degeneration @xmath347 corresponds to a normal surface which is a heegaard torus in @xmath349 and the flat solution will be analysed below using the face pairings .
the remaining structures come in pairs @xmath350 and it suffices to study the case @xmath351 here , @xmath352 implies that the arguments of @xmath353 are the angles of an isoceles triangle . letting @xmath91
denote the argument of @xmath115 the angle around @xmath345 is @xmath354 and the angle around each of @xmath343 and @xmath344 is @xmath355 this gives hyperbolic cone - manifold structures , with respective cone angles @xmath356 since the parameter @xmath342 at the degree - one edges is an element of @xmath357 it is easily verified that all of these hyperbolic cone - manifold structures are complete ; see figure [ tet_spine_horospheres1.pdf ] . in order to analyse the face pairings
, we normalise the developing map so that the degree one edges are mapped to the geodesics @xmath358 $ ] and @xmath359.$ ] the corresponding pairings for the faces incident with these edges are denoted @xmath360 and @xmath361 respectively .
the holonomy around the third edge is denoted @xmath362 the images of the group elements under the holonomy representation are determined by the following mbius transformations : @xmath363\mapsto[1,\infty,0 ] , \\ & \rho_z(\gamma_1):[0 , z '' , 1]\mapsto[0,z'',\infty ] , \\ & \rho_z(\gamma_2):[0 , \infty , 1]\mapsto[0,\infty , z^{-2}],\end{aligned}\ ] ] where @xmath364).$ ] this gives : @xmath365 letting @xmath366 we get the traces @xmath367 geometrically , as @xmath370 the fundamental domain degenerates to a quadrilateral , and the identification space is a sphere with cone angles @xmath371 . by construction
, this is a hyperbolic 2cone - manifold . as @xmath372
the limiting representation is infinite cyclic : @xmath373 the fact that @xmath374 is the generator of the image corresponds to the chosen normalisation : the edge @xmath359 $ ] pops off at infinity ( since @xmath375 and @xmath347 ) , giving @xmath376 let @xmath377 denote the minimal layered triangulation of @xmath113 with one degree - one edge @xmath344 and one degree - five edge @xmath378 as shown in figure [ knot_triangs_of_s3.pdf ] .
here , @xmath344 is the trefoil knot .
again , @xmath379 and @xmath380 where @xmath342 is the parameter given to @xmath381 the angle at @xmath344 is @xmath91 and the angle at @xmath345 is @xmath382 as above , there is an ideal point ( corresponding to @xmath347 ) , and one flat solution ( corresponding to @xmath348 ) .
they will be analysed below using the face pairings .
the degeneration @xmath347 corresponds to a thin edge - linking torus for @xmath385 and hence to the trefoil knot complement .
the remaining structures come in pairs @xmath350 and it can again be verified directly that they all give complete hyperbolic cone - manifold structures with singular locus consisting of the three edges ; see figure [ tet_spine_horospheres2.pdf ] .
these generate the modular group , and the quotient of @xmath394 is a sphere with 2 , 3 and @xmath395 cone points . see figure [ trefoil_collapse.pdf ] .
the flat solution does not appear to have a nice interpretation . in the previous examples
, we have analysed structures arising at ordinary points of the cone - deformation variety as well as a single ideal point which corresponding to a 1dimensional or 2dimensional collapse . in the last example
, we exhibit a surface of ideal points parameterising 3dimensional structures .
see figure [ knot_triangs_of_s3.pdf](b ) .
label the tetrahedra @xmath396 and @xmath397 at their upper and lower edges in figure [ knot_triangs_of_s3.pdf ] .
then the generalised gluing equations are : @xmath398 these simplify to : @xmath399 hence @xmath400 is parameterised by @xmath401 where @xmath402 and @xmath403 subject to @xmath404 from ( 12 ) and @xmath405 from ( 11 ) ( note that @xmath406 putting @xmath407 where @xmath408 , the second condition gives @xmath409 . letting @xmath410 for @xmath411 , this yields a single equation in @xmath412 and
the space of solutions is a surface in @xmath413 as shown in figure [ fig8_edge_in_s3_real_param_space2.jpg ] .
now @xmath400 is parameterised by @xmath414 and since @xmath402 is arbitrary , @xmath415 is the product of this surface with @xmath11 .
the only ideal points are at @xmath416 , parameterised by @xmath417 and the associated normal surface is the thin edge linking torus around @xmath222 , which splits the manifold into a solid torus and the figure 8 knot complement .
crushing along this surface identifies the single arrow edge with the double arrow edge , and gives us the canonical triangulation of the figure 8 knot complement .
the equations become : to get the complete structure on the figure eight knot complement , we need @xmath419 , which implies that @xmath420 and @xmath421 . then @xmath422 , so @xmath423 , and we get @xmath424 .
we also have @xmath421 , which implies that @xmath425 , so @xmath426 , and @xmath427 is the shape parameter of the regular hyperbolic ideal tetrahedron .
similarly for @xmath397 . note that solutions in a neighbourhood of this solution give 3dimensional hyperbolic cone - manifold structures on the figure eight knot complement with singular locus contained in the ideal edges .
feng luo , stephan tillmann and tian yang : _
thurston s spinning construction and solutions to the hyperbolic gluing equations for closed hyperbolic 3-manifolds _ , proceedings of the american mathematical society , in press ; arxiv:1004.2992v1 .
|
let @xmath0 be a topologically finite , orientable 3manifold with ideal triangulation . we show that if there is a solution to the hyperbolic gluing equations , then all edges in the triangulation are essential .
this result is extended to a generalisation of the hyperbolic gluing equations , which enables the construction of hyperbolic cone - manifold structures on @xmath0 with singular locus contained in the 1skeleton of the triangulation .
|
the role of turbulence in molecular cloud evolution has been a subject of much study in the literature ( see , for example , the excellent reviews of @xcite ) . observations of the properties of gas and dust motions in molecular clouds @xcite suggest that , indeed , turbulence is present .
it is clear that turbulent motion could influence the star formation rate and efficiency as well as the initial mass function @xcite . given all this there are several interesting questions which arise : 1 .
what is the source of molecular cloud turbulence ?
how fast does it decay ?
3 . how does it affect star formation ? 4 .
how does it affect the evolution of molecular clouds ?
the first two questions are clearly inter - related : if the turbulence decays very quickly then we need a lot of energy from its source in order to maintain it . indeed , to address the final two questions we must first gain insight into the first two .
in order to study turbulence in molecular clouds we must resort to numerical simulations .
quite simply , there are no satisfactory analytic techniques for addressing compressible magnetohydrodynamic ( mhd ) turbulence ( e.g. * ? ? ?
many authors have performed sophisticated numerical simulations in order to investigate both the qualitative nature of mhd turbulence and its decay @xcite .
most of this work has been carried out for the case where ideal mhd is valid - i.e. on relatively large length scales . when smaller length scales are considered ( e.g. lengths of significantly less than a parsec )
ambipolar diffusion becomes non - negligible in molecular clouds @xcite .
some authors @xcite have studied driven mhd turbulence in the presence of ambipolar diffusion .
all these authors find that ambipolar diffusion produces significant differences in the properties of the turbulence .
it has been suggested @xcite that although the hall resistivity is generally at least an order of magnitude lower than the ambipolar resistivity in molecular clouds , its effect should not be ignored due to the qualitative change it induces in the behavior of the magnetic field .
researchers working on reconnection and the solar wind have studied the hall effect in the context of turbulence and found that , although the overall decay rate appears not to be affected , the usual coincidence of the magnetic and velocity fields seen in mhd does not occur at small scales @xcite .
almost no work has been done on comparing the influences of this effect coupled with that of ambipolar diffusion on turbulence ( with the exception of low resolution simulations by @xcite ) . in particular , to our knowledge the work presented here represents the first systematic study of molecular cloud turbulence incorporating both the hall effect and ambipolar diffusion simultaneously .
the main aim of this work is to examine in detail the differences between the decay of ideal mhd turbulence and that of more realistic non - ideal mhd turbulence with a full tensor resistivity incorporating the effects of ambipolar diffusion , the hall effect and ohmic resistivity .
this work is new in two respects : no previous work has focussed on _ decaying _
( i.e. un - driven ) turbulence in the presence of non - ideal terms and , in addition , no previous work has addressed the issue of turbulence in the presence of both ambipolar diffusion and the hall effect simultaneously .
this is the first of a short series of papers describing a comprehensive study of multifluid mhd turbulence in a parameter regime appropriate to molecular clouds . in section [ sec
: num - method ] we outline the numerical techniques used in this work , as well as the initial conditions and general set - up for the simulations while in section [ sec : analysis ] we describe the methods used to analyze the simulation data . in section [ sec : results ] we present and discuss the results of our simulations of turbulent decay .
finally , section [ sec : conclusions ] contains a summary of our results .
in this work we use the code hydra @xcite to integrate the equations of non - ideal mhd ( see section [ subsec : eqns - algorithm ] ) .
we assume that the molecular cloud material we are simulating can be treated as isothermal and that initially the density and magnetic field are uniform . for this work
we assume spatially and temporally constant ohmic , ambipolar and hall resistivities ( see section [ subsec : eqns - algorithm ] ) .
we briefly outline the equations and assumptions in our model here but refer the reader to @xcite for a comprehensive description of both the full abilities of the hydra code and the assumptions underlying the equations used .
we assume that the cloud material can be treated as weakly ionized .
this is clearly valid for molecular clouds and allows us to ignore the inertia of the charged species @xcite .
the equations solved in this work are then @xmath5 where @xmath6 , @xmath7 , @xmath8 , @xmath9 , @xmath10 and @xmath11 are the neutral mass density , neutral velocity , sound speed , identity matrix , magnetic field and current density respectively . the electric field in the frame of the fluid , @xmath12 , is calculated from the generalized ohm s law for weakly ionized fluids ( e.g. @xcite ) and is given by @xmath13 where @xmath14 using the definitions @xmath15 , @xmath16 , @xmath17 , where @xmath18 , @xmath19 , @xmath20 .
@xmath21 , @xmath22 and @xmath23 are the ohmic , hall and ambipolar resistivities respectively . in this work
these resistivities are kept constant in both space and time .
we note that , physically , they depend on both the magnetic field and the density of the various charged species in the fluid ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and hence , in reality , do vary in both space and time .
treating the resistivities in such a simple way allows us to gain a deeper understanding of their influence on turbulence without having to consider the complicating effects of dynamically varying resistivities at the same time . as such , and as a first step away from the approximation of ideal mhd and a single form of magnetic diffusion we believe this is an interesting study .
having gained some insight into this simplified model a follow - up paper will address multifluid mhd turbulence under the influence of self - consistently calculated resistivities . as noted by @xcite and @xcite ,
the main difficulty with standard numerical techniques for integrating equation [ b_eqn ] lies with the hall term .
as this term becomes dominant the stable time - step goes to zero .
however , osullivan & downes ( @xcite ) presented a novel , explicit numerical method for integrating this term such that the limit on the stable time - step is not overly restrictive .
we use this `` hall diffusion scheme '' in this work .
of course , all explicitly differenced diffusion terms give rise to a stable time - step which is proportional to @xmath24 , where @xmath25 is the resolution of the simulation .
to ameliorate this we use standard sub - cycling of the hall terms and super time - stepping to accelerate the ambipolar diffusion terms ( see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
equations [ mass ] [ b_eqn ] are solved using a standard shock - capturing , second order , finite volume , conservative scheme .
equation [ divb ] is enforced using the method of @xcite .
the effects of the diffusive terms are then incorporated in an operator split fashion .
we examine the decay of supersonic mhd turbulence in conditions suitable for dense regions of molecular clouds . while the simulations presented here are , of course , scale - free we present the initial conditions used in standard astrophysical units for ease of reading .
the simulations are carried out in a cube of side @xmath26pc with periodic boundary conditions being enforced on all faces .
the sound speed is set to 0.55kms@xmath27and the initial density is chosen to be uniform with a value of @xmath28@xmath29 .
the magnetic field is also initially uniform in the @xmath30 direction with a magnitude of 1 mg . for these conditions ,
suitable conductivities are @xmath31 s@xmath27 , @xmath32s@xmath27 and @xmath33s@xmath27 ( see figure 1 , @xcite ) .
we choose these particular physical conditions with a view to maximizing the influence of the hall effect in our simulations @xcite . in this way we can use our simulations to find whether the hall effect is ever likely to be important in molecular cloud turbulence .
the initial velocity field is used to instigate the turbulence in these simulations .
each component of the velocity field is defined to be the sum of waves with 16 wave - vectors , each with random amplitude and phase - i.e. @xmath34 where @xmath35 defines the component ( @xmath36 , @xmath37 or @xmath38 respectively ) of the velocity , @xmath39 and @xmath40 are the random amplitudes and phases and @xmath41 where @xmath42 is the usual dirac delta function .
the inclusion of the term in parenthesis in the definition of @xmath43 restricts the initial velocity field to be solenoidal .
note that we perform all the analysis of these simulations at @xmath44 ( 1 flow crossing time ) at which time the effects of the precise initial conditions used should be negligible .
the nomenclature for the simulations is xx - ab - c where xx denotes the type of physics ( e.g. a standard molecular cloud run is `` mc '' , ideal mhd is `` mhd '' etc ) , ab is the initial rms mach number and c is the resolution used .
the initial root - mean - square ( rms ) of the field is chosen to be either mach 2.5 , 5 or 10 depending on the simulation in question .
these correspond to alfvnic mach numbers of approximately 0.96 , 1.9 and 3.85 respectively .
in addition to the non - ideal mhd simulations described we also run 4 further simulations .
the first is an ideal mhd simulation ( mhd-5 - 512 ) which we use for comparison purposes , another is a pure hydrodynamic simulation ( hd-256 - 0.5 ) , and the other two ( ambi-5 - 512 and hall-5 - 512 ) only incorporate one of ambipolar diffusion or the hall effect , respectively .
we use these latter simulations to separate out the effects of each of these diffusions to better understand the physics occurring .
table [ table : sim - defs ] contains definitions of the various simulations used in this work . lccl simulation & mach number & resolution & comment + mc-5 - 64 & 5 & @xmath45 & - +
mc-5 - 128 & 5 & @xmath46 & - + mc-5 - 256 & 5 & @xmath47 & - + mc-5 - 512 & 5 & @xmath0 & - + mc-2.5 - 512 & 2.5 & @xmath0 & - + mc-10 - 512 & 10 & @xmath0 & - + ambi-5 - 512 & 5 & @xmath0 & @xmath48 + hall-5 - 512 & 5 & @xmath0 & @xmath49 + mhd-5 - 512 & 5 & @xmath0 & ideal mhd + hd-5 - 256 & 5 & @xmath47 & hydrodynamic +
in this section we discuss the method of analysis of the output of the simulations described in section [ subsec : initial - conditions ] .
the main aim of this paper is to investigate the decay rate of supersonic turbulence in molecular clouds .
hence , the main analysis carried out on the simulation results is the calculation of the volume - averaged kinetic , magnetic and total energy in the simulation as a function of time .
these quantities are defined as @xmath50 where the angle brackets denote averaging over the computational domain .
note that @xmath51 is therefore the difference in the magnetic energy at the current time and the magnetic energy at @xmath52 ( see , for example , @xcite ) , given the assumption that no external electromotive force is applied and that our boundary conditions are periodic .
we also calculate the mass - weighted average mach number , defined by @xmath53 where @xmath8 is the sound speed and the mass - weighted velocity dispersions , @xmath54 , are defined by @xmath55 where @xmath56 is either @xmath36 , @xmath37 or @xmath38 and the angle brackets denote averaging over the computational domain ( see @xcite ) .
in section [ sec : power - spectra ] we present the power spectra for the velocity , density and magnetic field for each of the @xmath0 simulations .
these spectra are calculated by taking the power spectrum in the @xmath36 , @xmath37 and @xmath38 directions and then integrating the power for all @xmath57 satisfying @xmath58 for each @xmath59 with @xmath60 .
this gives us some insight into the scale of structures being formed by the turbulence for the various initial conditions and range of physics examined . finally , in section
[ sec : vel - disp ] we calculate the velocity dispersion as a function of length scale , @xmath61 .
for these purposes we define the velocity dispersion to be @xmath62 where @xmath63 where @xmath64 indicates an average taken over a cube of side @xmath61 in the simulation domain and @xmath65 indicates averaging of the quantity over all such non - overlapping cubes within the domain .
we now present the results of the simulations carried out .
each simulation was run for one sound crossing time , @xmath66yrs , of the simulation domain .
figure [ fig : mc - mhd - den ] shows the density distributions in a slice at @xmath67pc ( i.e. the mid - plane ) and times @xmath68 , @xmath69 and @xmath70 @xmath71 for the mc-5 - 512 and the mhd-5 - 512 simulations .
both the non - ideal and ideal mhd simulations show some anisotropy with respect to the projected initial magnetic field direction - i.e. the @xmath72 direction - with filaments both perpendicular and parallel to this direction .
this is what would be expected as the material will flow preferentially along the @xmath30 direction and hence we expect the shock fronts to be normal to it . this anisotropy becomes more evident as time progresses because the kinetic energy decays to below the energy of the mean magnetic field ( which is conserved with time ) - hence the flow evolves from being dominated by kinetic energy to dominated by magnetic energy over the lifetime of the simulations .
figure [ fig : hydro - den ] contains a plot of the same density distribution for hydro-5 - 256 at @xmath68@xmath71 for comparison .
it can be seen that this is qualitatively different to both the ideal and non - ideal simulations shown in figure [ fig : mc - mhd - den ] and does not display any signs of anisotropy .
there are obvious qualitative differences between the mc-5 - 512 and mhd-5 - 512 simulations at all times with the density distribution in the ideal case containing much more small scale structure than the non - ideal case .
this results from the ability of the magnetic field to move with respect to the fluid .
this means that it is not compressed as much as in the ideal case , and hence the same level of small structure is not formed .
since the coupling between the magnetic field and the flow field is strong , even though it is imperfect , this lack of small - scale structure in the magnetic field becomes mirrored in the density field .
the implications of this are discussed in more detail in section [ subsec : energy - decay ] .
clearly , in the non - ideal simulation energy is not as efficiently transported from large scales to smaller scales as in the ideal case .
we now continue our discussion of the results by considering a resolution study to demonstrate convergence of our numerical results .
we then go on to discuss the properties of the density , velocity and magnetic field distribution resulting from the turbulence , and finally the properties of the energy decay .
in order to be reasonably confident that the results we present in this work are converged we have performed a resolution study for the conditions used in the simulations .
we have run simulations which are identical in all respects except for the resolution , which varies from @xmath45 up to @xmath0 ( see table [ table : sim - defs ] ) .
we then analyze these in an identical fashion and consider the differences between our results for the different resolutions .
figure [ fig : res - energy ] shows plots of the kinetic energy normalized to its initial value as a function of time for each of the simulations in our resolution study .
the @xmath45 and @xmath46 plots do appear to be significantly different from the higher resolution simulations .
however , the @xmath47 and @xmath0 simulations are much more similar with a maximum difference in the total kinetic energy in the simulation at any one time being less than 10% .
table [ table : decay - exp ] contains least squares fits of the decay over the time - interval @xmath73 $ ] assuming it to be of the form @xmath74 for each of the kinetic energy ( @xmath75 ) , the magnetic energy ( @xmath76 ) and their total ( @xmath77 ) .
it can be seen that @xmath75 varies by approximately 4% over the entire range of the resolution study ( simulations mc-5 - 64 through mc-5 - 512 ) .
we feel , therefore , that we can be fairly confident of the value of this exponent when comparing it with the other simulations presented here .
lccc simulation & @xmath75 & @xmath76 & @xmath77 + mc-5 - 64 & 1.34 & 1.40 & 1.35 + mc-5 - 128 & 1.33 & 1.35 & 1.34 + mc-5 - 256 & 1.37 & 1.36 & 1.36 + mc-5 - 512 & 1.40 & 1.37 & 1.39 + ambi-5 - 512 & 1.40 & 1.35 & 1.38 + hall-5 - 512 & 1.25 & 1.18 & 1.22 + mhd-5 - 512 & 1.26 & 1.19 & 1.23 + hydro-5 - 256 & 1.10 & - & 1.10 + mc-2.5 - 512 & 1.21 & 1.29 & 1.23 + mc-10 - 512 & 1.42 & 1.39 & 1.41 + the decay of the energy in the magnetic field is more rapid than that in the kinetic energy with a difference between @xmath75 and @xmath76 of about 4% . at low resolution
( mc-5 - 64 ) @xmath76 is at its highest which is a sign that the low resolution is introducing sufficient numerical viscosity to induce large amounts of numerical reconnection . as the resolution
is increased @xmath76 reduces to approximately 1.36 and stays at about this value even up to the maximum resolution of @xmath0 .
we can be reasonably confident then that resolution is not affecting our estimate of the decay rate of the magnetic energy at our maximum resolution .
the decay of the total energy , being derived from the decay of kinetic and magnetic energy , is also reasonably well converged with a total change over the entire range of the resolution study of around 4% . note
that , while we can be confident from our results that the decay rate is converged , we have to be more careful when considering our power spectra results presented in sections [ sec : power - spectra ] and [ sec : mach - effect - power - spectra ] .
for example , @xcite , @xcite and @xcite present power spectra in ideal mhd which suggest that at resolutions of @xmath78 the turbulent inertial range is established over at most a decade in @xmath59 , while for resolutions of @xmath0 this falls to around half a decade .
we now discuss the behavior of the kinetic and magnetic energy in our non - ideal simulations and compare them with those for our ideal simulation .
figure [ fig : non - ideal - kin ] contains plots of the kinetic energy as a function of time for the mc-5 - 512 , ambi-5 - 512 , hall-5 - 512 and mhd-5 - 512 simulations .
it is apparent that the behavior of mhd-5 - 512 is rather similar to hall-5 - 512 and the behavior of ambi-5 - 512 is similar to that of mc-5 - 512 .
this indicates that , at least for the kinetic energy decay , the hall effect has little impact . however , there is a marked difference between those simulations incorporating ambipolar diffusion ( mc-5 - 512 and ambi-5 - 512 ) and those which do not ( hall-5 - 512 and mhd-5 - 512 ) .
ambipolar diffusion clearly increases the decay rate of the turbulence .
this can also be seen from the data presented in table [ table : decay - exp ] where the exponents of the simulations containing ambipolar diffusion are greater by about 10% than those without .
in general we expect that turbulence will decay more rapidly in systems which have higher viscosity than those without .
clearly the ambipolar diffusion , although it corresponds to a `` viscosity '' in the magnetic field , acts in a similar way to the usual viscous forces in a fluid when considering this system .
this can be seen in figure [ fig : mc - mhd - den ] where the density features are more smeared out in the non - ideal case as would be expected if we simply introduced a viscous term into the momentum equations .
it is important to note here that the non - ideal diffusive terms in the induction equation will lead to enhanced reconnection in the magnetic field .
since we assume an isothermal equation of state there is no path by which the energy released by reconnection can find its way to the kinetic energy of the system .
there may be significant differences between the effects of the non - ideal terms on the decay rate of turbulence in the isothermal and non - isothermal regimes in molecular clouds .
figure [ fig : mhd - mc - b ] contains plots of the energy of the magnetic perturbations induced by the turbulence ( normalized to the initial kinetic energy ) as a function of time for the mc-5 - 512 , ambi-5 - 512 , hall-5 - 512 and mhd-5 - 512 simulations .
a similar trend is evident in the results for magnetic energy decay as that already noted in section [ sec : kin - decay ] .
again , the main result is that ambipolar diffusion increases the energy decay rate while the hall effect does little to influence it .
this result is borne out by the data in table [ table : decay - exp ] in which the decay exponents are greater for the simulations incorporating ambipolar diffusion by about 14% than those which do not . to gain a little more insight into the interplay between the magnetic and kinetic energy in the system we now focus on simulations mc-5 - 512 and mhd-5 - 512 at early times .
figure [ fig : early - time - energy ] contains plots of the magnetic and kinetic energies of these two simulations as a function of time .
the magnetic energy grows initially as the flow converges and compresses the magnetic field in regions throughout the computational domain .
the kinetic energy gradually decays during this time .
once shocks form , equipartition between the total magnetic energy and kinetic energy is reached and the magnetic field begins to decay .
at this time the decay of the kinetic energy accelerates due to the dissipative effect of the shocks which have just formed .
this dependence on shocks forming is confirmed when the growth and decay of the magnetic energy is examined for simulations mc-2.5 - 512 and mc-10 - 512 and compared with mc-5 - 512 ( see figure [ fig : early - energy - mach ] ) .
here we can see that the peak in magnetic energy is reached at later times as the initial rms mach number is decreased .
the time at which shocks begin to form is determined by the initial rms mach number ( giving us a characteristic velocity ) and half the shortest wavelength in the initial velocity field . the approximate shock formation times are 0.0125@xmath71 , 0.025@xmath71 and 0.05@xmath71 for the mc-10 - 512 , mc-5 - 512 and mc-2.5 - 512 simulations respectively .
these times match up reasonably well , particularly given the temporal granularity of the simulation data , with the peak of the fluctuating part of the magnetic energy in these simulations . just after equipartition
is reached the decay of the kinetic energies of mc-5 - 512 and mhd-5 - 512 begin to behave differently .
this is unsurprising as it is only when the magnetic field perturbations are reasonably strong that non - ideal effects can play a dynamically significant role .
just after the magnetic energy peaks the kinetic energy of simulation mhd-5 - 512 goes through a short period during which it does not decay particularly rapidly - presumably a result of the transfer of energy from the compressed magnetic field back to the kinetic energy as the fluid expands after the initial compressions in the simulation .
this does not happen in simulation mc-5 - 512 since the magnetic field does not become so compressed in the first place due to non - ideal effects .
this can be seen from the fact that the magnetic energy peaks at a lower level in this simulation .
in addition , energy from this field will not be transferred efficiently back to the kinetic energy because of the imperfect coupling between the magnetic and velocity fields resulting from the non - ideal terms in the induction equation .
the picture which emerges is the following : initially kinetic energy is transferred into magnetic energy through compressions .
if there are diffusive terms present in the induction equation then this extra magnetic energy will be dissipated and there will be less energy available to transfer back to the kinetic energy as the compressed regions expand .
this process will be repeated throughout the simulation as shocks form and dissipate .
hence we expect a more rapid decay of both magnetic and kinetic energy when diffusive terms are present .
recalling that the hall effect does not actually diffuse magnetic energy , it is no surprise that simulation hall-5 - 512 is so similar in terms of energy decay to mhd-5 - 512 .
finally , we note that the peak of the fluctuating part of the magnetic energy occurs at a slightly later time in the mhd-5 - 512 simulation than the mc-5 - 512 one .
we explain this as follows . as noted in section [ sec : rho - power ] , the ambipolar diffusion appears to set a dissipation length scale corresponding to @xmath79 .
this is significantly larger than the dissipation scale in the ideal mhd simulation .
hence energy in the magnetic field need not cascade to such short length scales in order to be dissipated in the mc-5 - 512 and ambi-5 - 512 simulations as in the mhd-5 - 512 and hall-5 - 512 simulations .
since the energy cascade to shorter length scales takes time ( particularly at very early times when shocks are only beginning to form ) , the energy in the magnetic field can be removed from the former simulations at earlier times than from the latter simulations as it need not cascade so far .
we now turn to power spectra of the simulations .
these tell us about the scale of the structures formed and are useful in gaining some insight into the effects of the non - ideal physics incorporated .
it is important to stress that since these simulations are of decaying turbulence , rather than driven turbulence , we do not necessarily expect to get the often - quoted power - law dependence of power on wavenumber .
all the power spectra presented have been calculated at @xmath80 when we can be fairly confident that the turbulence is fully developed and the initial conditions have been effectively forgotten .
in addition , the concerns noted in section [ subsec : resolution - study ] should be borne in mind - i.e. the turbulent inertial range is likely to be of order half a decade in @xmath59 for our @xmath0 simulations .
we discuss the power spectra of density , magnetic field strength and velocity in turn .
figure [ fig : rho - power ] contains plots of the spherically integrated power spectrum of the density for the mc-5 - 512 , hall-5 - 512 , ambi-5 - 512 and mhd-5 - 512 simulations at @xmath81 .
it is clear that there is little difference between the mc-5 - 512 and ambi-5 - 512 simulations - again strongly indicating that the hall effect has little influence on the behavior of the simulations .
the mhd-5 - 512 and hall-5 - 512 simulations are rather similar , although there is less power in small - scale structures in the hall-5 - 512 simulation than the mhd-5 - 512 one .
we leave discussion of this until section [ sec : b - power ] . even at large length scales
there is a significant difference in the power spectra , indicating that dissipation on short length scales due to the non - ideal terms in the induction equation indeed affects the behavior of the density distribution at large scales .
the simulations with ambipolar diffusion have considerably steeper slopes than those without ( see table [ table : power - slopes ] ) in qualitative agreement with the results of @xcite . that the mc-5 - 512 simulation has less power at shorter length scales than mhd-5 - 512 can be seen by a cursory examination of figure [ fig : mc - mhd - den ] - the mc-5 - 512 images appear more `` blurry '' than those from mhd-5 - 512 - so this result is not a surprise .
lccc simulation & density & velocity & magnetic field + mc-5 - 512 & 2.09 , 4.06 & 1.47 & 2.17 , 4.96 + ambi-5 - 512 & 2.04 , 4.03 & 1.49 & 2.21 , 4.75 + hall-5 - 512 & 1.41 & 1.20 & 1.65 + mhd-5 - 512 & 1.45 & 1.17 & 1.59 + mc-2.5 - 512 & 2.37 , 3.32 & 2.14 & 2.86 , 3.80 + mc-10 - 512 & 2.20 , 4.20 & 1.44 & 2.59 , 4.70 + further structure in the power spectrum for mc-5 - 512 is apparent which is absent in the mhd-5 - 512 simulation : at @xmath82 there is a break to a steeper slope in the simulations containing ambipolar diffusion . for the mhd-5 - 512 simulation , at about @xmath83 numerical viscosity begins to affect the power spectrum , as evidenced by the roll over of the spectrum from a power - law to a steeper slope .
the same effect does not appear in the mc-5 - 512 simulation because the physical viscosity due to the non - ideal terms in the induction equation dominates the numerical viscosity up to much higher values of @xmath59 . while the power spectrum for the latter simulation does change slope at @xmath82
it then maintains a strong power law up to @xmath84 .
hence the break at @xmath85 for mc-5 - 512 and ambi-5 - 512 appears to be a physical result , rather than a numerical one .
this suggests that , apparently in contradiction with a result of @xcite , ambipolar diffusion can set a length scale in turbulence - in these simulations that length scale is about 0.02pc .
we discuss this further in section [ sec : b - power ] but note that we should be cautious about drawing general inferences given that the resistivities used in these simulations are constant in space and time . ultimately , at very short length scales ( @xmath86 or @xmath87pc ) all four simulations steepen significantly .
this is expected since at these values of @xmath59 , which correspond to lengths of less than about 5 grid zones in these simulations , numerical viscosity will certainly dominate structure generation / dissipation .
we now turn to the velocity power spectra .
figure [ fig : v - power ] contains plots of the spherically integrated power spectrum of the velocity for the mc-5 - 512 , hall-5 - 512 , ambi-5 - 512 and mhd-5 - 512 simulations at @xmath81 . in broad terms of the differences between the simulations we obtain the same result as for the density power spectrum ( cf figure [ fig : rho - power ] ) .
it is , however , obvious that the density and velocity power spectra differ quite substantially in qualitative terms .
at very low @xmath59 the velocity spectra have a rather shallow slope with little difference between any of the simulations up to @xmath88 - the maximum value of @xmath59 at which the initial velocity field was non - zero . in the range
@xmath89 both mhd-5 - 512 and hall-5 - 512 follow a similar power law , while mc-5 - 512 and ambi-5 - 512 attain a steeper slope in this range . in general , for @xmath90
the simulations with ambipolar diffusion have significantly steeper power spectra , as observed in the density power spectra and in qualitative agreement with the driven turbulence results of @xcite .
there is a further apparent break in the power spectrum at around @xmath84 in the simulations with ambipolar diffusion occurring where numerical viscosity begins to dominate the ambipolar diffusion ( see section [ sec : rho - power ] ) .
it is interesting to note that the series of breaks in the velocity power spectra in mc-5 - 512 and ambi-5 - 512 are not mirrored in the density power spectra indicating a level of decoupling between these two fields .
@xcite noted that under certain circumstances a power spectrum of @xmath91 could follow a power law with the classical kolmogorov slope of -5/3 even for compressible ( hydrodynamic ) turbulence .
figure [ fig : rho - third - v - power ] contains power spectra of this variable for the mc-5 - 512 and mhd-5 - 512 simulations .
it is apparent that the power spectra for this variable for these simulations is shallower than the -5/3 law , at least up to the value of @xmath59 where dissipative effects may be important . in the range @xmath92
the exponents are -1.289 and -1.14 for the mc-5 - 512 and mhd-5 - 512 simulations , respectively .
the theory under which these variables could be expected to have the kolmogorov slope has the assumption that the system has reached a statistical steady state - since we investigate decaying turbulence here this is unlikely to be the case .
it might be expected that if we do not continually supply energy at low @xmath59 then the slope after a turbulent crossing time when the turbulence is well established and has also decayed significantly would be too shallow as the power spectra would be `` too low '' at low @xmath59 .
figure [ fig : b - power ] contains plots of the power spectrum of the magnetic field at time @xmath81 .
once again these plots indicate that ambipolar diffusion is the dominant non - ideal effect in molecular cloud turbulence .
we can see that there are some differences between each of the simulations .
the mc-5 - 512 and ambi-5 - 512 simulations are rather similar and are rather similar to the density power spectra ( figure [ fig : rho - power ] ) . however ,
at high @xmath59 the ambi-5 - 512 simulation does have noticeably more power than mc-5 - 512 .
this indicates that the hall effect , present in the mc-5 - 512 simulation , is having some impact in the structure of the magnetic field at short length scales .
this effect is very small in the power spectra of both velocity and density and is not apparent in the energy decay rates either .
it would appear that when the ambipolar diffusion is strong in comparison to the hall effect , as is the case for molecular clouds , the hall effect may change the structure of the magnetic field but this change will not propagate into the rest of the fluid variables . the power spectra for mhd-5 - 512 and hall-5 - 512 are rather similar to each other , again with the exception of high @xmath59 where the hall-5 - 512 simulation has significantly less power .
we explain this as follows . while the hall effect does not actually diffuse the magnetic field it does re - orient it
. this process of re - orientation can give rise to a topology favoring some magnetic reconnection - particularly on the small scales at which the hall effect operates .
we therefore attribute the lower power at high @xmath59 in hall-5 - 512 to the interplay between the hall effect and numerical viscosity at these length scales . recall that since hall-5 - 512 contains no ambipolar diffusion and negligible parallel resistivity any reconnection which occurs must be almost entirely due to numerical viscosity .
the same argument holds for the difference between mc-5 - 512 and ambi-5 - 512 at high @xmath59 , but in this case the interplay is between the hall effect and ambipolar diffusion . the marked turn up in power for @xmath93 for the mhd-5 - 512
simulation is due to numerical effects .
overall the plots are rather similar to the density power spectra ( figure [ fig : rho - power ] ) : the density and magnetic field distributions appear to be fairly well coupled and different in nature to the velocity distribution .
it has been generally accepted that observations of line of sight velocity dispersion in molecular clouds exhibits a power law with the size of the field of view ( e.g. * ? ? ?
it is worth noting , however , that recent observational results call the so - called larson s law into question @xcite .
however , for completeness we feel it is worthwhile to examine how our non - ideal simulations behave in this regard . in this section
we examine the velocity dispersion as a function of characteristic length for each of the simulations .
figure [ fig : larson ] contains plots of the velocity dispersions at @xmath94 for each of the @xmath0 simulations .
once again , the results follow the general trend of indicating that the hall effect has almost no impact with the data for mhd-5 - 512 and hall-5 - 512 being almost indistinguishable .
at all length scales the velocity dispersion for the simulations containing ambipolar diffusion ( mc-5 - 512 and ambi-5 - 512 ) is lower , with the difference being larger at short length scales . from the power spectra study in section
[ sec : power - spectra ] we already know that the power in short scale variations of the velocity is decreased by the presence of ambipolar diffusion .
it is therefore no surprise that comparison of the results for the velocity dispersion shows that the dispersion is decreased by the presence of this diffusion .
there is no obvious power law in these results , in agreement with the simulations of driven , ideal mhd turbulence of @xcite .
as noted by these authors , the varying strength and direction of the magnetic field give rise to a large number of signal speeds within the domain , rather than the simple sound speed frequently assumed in hydrodynamic turbulence and hence there is no reason to expect a power law in the velocity dispersion .
finally we turn to the issue of the initial mach number chosen in the simulations . to study this we consider simulations mc-2.5 - 512 , mc-5 - 512 and mc-10 - 512 .
these correspond to initial sonic mach numbers of 2.5 , 5 and 10 , respectively and alfvnic mach numbers of 0.96 , 1.9 and 3.85 .
figure [ fig : mach - number - kin ] contains plots of the decay of kinetic energy for each simulation .
the energy in the plots has been normalized to the starting energy for each simulation .
it is clear that the higher mach number flows lose their energy more rapidly than the lower mach number ones .
this behavior is repeated in the magnetic energy and the total energy .
table [ table : decay - exp ] contains the power law indices of decay for the kinetic , magnetic and total energy in these simulations .
the exponents for the kinetic energy decay for the mc-2.5 - 512 , mc-5 - 512 and mc-10 - 512 simulations are 1.21 , 1.40 and 1.42 respectively , while those for magnetic energy are 1.29 , 1.37 and 1.39 respectively .
it is clear that , indeed , all the indices increase with initial rms mach number .
we explain this by noting that high mach number flows tend to have strong shocks which will dissipate energy more effectively than lower mach number flows .
figures [ fig : mach - effect - rho - power ] , [ fig : mach - effect - v - power ] and [ fig : mach - effect - b - power ] contain the power spectra for the mc-10 - 512 , mc-5 - 512 and mc-2.5 - 512 simulations for each of the density , velocity and magnetic field respectively taken at @xmath81 .
the density power spectrum does not appear to change much between the mc-10 - 512 and mc-5 - 512 simulations , although it is markedly shallower for the mc-2.5 - 512 simulation .
it is worth noting that the mc-2.5 - 512 simulation is initially sub - alfvnic and so is qualitatively different in nature to the other two .
in addition , the rms sonic mach number drops below 1 at @xmath95@xmath71 and hence we have rather well - evolved subsonic , decaying turbulence in this case .
the velocity power spectrum for the mc-10 - 512 and mc-5 - 512 simulations are also remarkably similar , while the mc-2.5 - 512 simulation is steeper at low @xmath59 and shallower for @xmath96 .
once again , we attribute this difference in behavior to the fact that the initial rms velocity of the mc-2.5 - 512 simulation is sub - alfvnic . the magnetic power spectrum follows a similar pattern in terms of differences between the mc-10 - 512 , mc-5 - 512 and mc-2.5 - 512 simulations as for the other two sets of power spectra .
figure [ fig : mach - effect - vel - disp ] shows plots of the velocity dispersion at @xmath81 for the mc-10 - 512 , mc-5 - 512 and mc-2.5 - 512 simulations .
again , as noted in section [ sec : vel - disp ] , no overall power - law is observed . at
short length scales ( less than about 0.01pc ) the slope of the relations are all approximately the same . above this scale
the slope is somewhat lower for mc-10 - 512 and mc-5 - 512 than for mc-2.5 - 512 .
this is to be expected since large velocity variations will be preferentially suppressed by strong shocks during the early evolution of the system , leading to each of these simulations having more similar velocity dispersion at large length scales .
hence we expect higher mach number simulations to retain somewhat higher velocity dispersions at all length scales than their lower mach number counterparts , but that the fractional differences in these dispersions will be lower at large length scales than at shorter ones .
we have presented the first study of turbulent decay in the presence of non - ideal terms in the induction equation and also the first simulations to incorporate both the hall effect and ambipolar diffusion simultaneously in simulations of molecular cloud turbulence .
this is the first stage in a comprehensive study of non - ideal mhd turbulence .
the non - ideal effects are therefore included in a simplistic way in this work with a view to developing a complete , intuitive understanding of their impact on turbulence as we continue the study and gradually add in more realistic components of these effects .
the results of this further stage of the study are the subject of a forthcoming paper currently in preparation .
we have used a resolution study to determine that our simulations are well - resolved for the purposes of the results which we present . in particular
, the turbulent energy decay is well resolved from a resolution of about @xmath47 and the results we present here are derived from @xmath0 simulations . we have found that the hall effect has little influence on kinetic energy decay when present at the levels found in molecular clouds .
however , ambipolar diffusion increases the rate of energy decay on length scales of @xmath97pc or less .
similar results are found for the behavior of magnetic energy .
ambipolar diffusion increases the rate of energy decay through diffusing away energy while it is stored in the magnetic field .
viewed in this way it is unsurprising that the hall effect does not influence kinetic energy decay strongly since it does not actually diffuse magnetic energy . the kinetic and magnetic energy decay is faster in higher mach number flows as would be expected both from this argument and from the fact that we are simulating isothermal flows which will lose energy more efficiently when stronger shocks are present .
the power spectra of the density for these simulations again suggest that the hall effect has little impact on the flows , while ambipolar diffusion can not be ignored . as might be expected from a diffusion term , the power spectrum is softer ( i.e. steeper ) when it is present . at a resolution of @xmath0 and an assumed length - scale of @xmath97pc
we appear to have resolved the length at which ambipolar diffusion begins to influence the flow .
ambipolar diffusion does appear to impose a characteristic length - scale on the turbulence .
however , we must be cautious about this interpretation since our assumption of spatially constant resistivities will have some impact on this result .
when comparing the velocity power spectra with the density power spectra we find that there appears to be a decoupling between the two fields with breaks in the power - laws not mirrored between the two sets of spectra when non - ideal effects are included .
the hall effect does have some impact on the magnetic power spectra at high @xmath59 .
it decreases the energy at high @xmath59 , probably due to re - orientation of the magnetic field at small scales in such a way as to favor reconnection ( either numerical or physical ) and hence destruction of structure on these scales .
calculations of the velocity dispersion as a function of length scale show that , again , ambipolar diffusion is the dominant diffusive term and that it has a significant impact .
it preferentially reduces the velocity dispersion at small scales .
we do not find a power law dependence between length scale and velocity dispersion , in agreement with @xcite .
comparisons of decaying turbulence with varying initial rms mach number show that higher mach number flows decay more quickly than their low mach number counterparts .
there are also differences in the power spectra with the mach 2.5 flow being significantly different to the mach 5 and mach 10 flows .
the mach 2.5 flow is slightly sub - alfvnic initially and this may explain the qualitative difference seen .
the velocity dispersions are lower at short scales for lower mach number flows , but are similar at large scales .
this is due to the tendency of strong shocks to decay very quickly and hence we do not expect even high mach number flows to maintain high velocity dispersion at large length scales for long .
the next step in this work is to study turbulent decay in the presence of resistivities determined consistently from the magnetic field and the density of charged species throughout the computational domain .
this will allow us to understand precisely the impact of spatially varying resistivities on turbulent decay .
we can expect , for example , that the behavior of the power spectra and velocity dispersions will be strongly affected .
in addition , such simulations will then incorporate all the non - ideal effects likely to be of importance in the study of molecular cloud turbulence .
alexiades v. , amiez g. , gremaud p. , 1996
, 12 , 31 ciolek g.e . , roberge , w.g . , 2002 ,
apj , 567 , 947 dedner , a. , kemm , f. , krner , d. , munz , c .- d . , schnitzer , t. , wesenberg , m. 2002 , j. comp .
phys . , 175 , 645 downes , t.p .
, osullivan , s. 2008 , numerical modeling of space plasma flows : astronum 2007 asp conference series , n.v .
pogorelov , e. audit , and g.p .
zank , 385 , 12 elmegreen , b.g .
1993 , , 419 , l29 elmegreen , b.g . , scalo , j. 2004 , ara&a , 42 , 211 falle , s.a.e.g .
2003 , , 344 , 1210 glover , s.c.o .
, mac low , m - m .
2007 , , 659 , 131 gustaffson , m. , brandenburg , a. , lemaire , j.l . , field , d. , 2006 , a&a , 454 , 515 heyer , m. , krawczyk , c. , duval , j. , jackson , j.m .
, 2009 , arxiv:0809.1397 klein , r.i .
, fisher , r.t .
, krumholtz , m.r . ,
mckee , c.f .
2003 , winds , bubbles and explosions : a conference to honour john dyson , s.j .
arthur , w.j .
henney , rev .
astrophys . , 15 , 92 kritsuk , a.g . , norman , m.l . , padoan , p. , wagner , r. , 2007 , apj , 665 , 416 kudoh , t. , basu , s. 2008 , , 679 , l97 larson , r.b .
1981 , , 194 , 809 lemaster , m.n . , stone , j.m .
2008 , , 682 , l97 lemaster , m.n . , stone , j.m .
2009 , , 691 , 1092 li , p.s . ,
mckee , c.f . ,
klein , r.i .
, fisher , r.t .
2008 , , 684 , 380 mac low , m .- m .
1999 , , 524 , 169 mac low , m .- m . , klessen , r.s .
2004 , rev .
phys . , 76 , 125 mac low , m .- m . ,
ossenkopf , v. 2000 , a&a , 353 , 339 mac low , m .-
m . , klessen , r.s .
, burkert , a. , smith , m.d .
1998 , 80 , 2754 matthaeus , w.h .
, dmitruk , p. , smith , d. , ghosh , s. , oughton , s. , 2003 , geophys .
30(21 ) , 2104 mininni , p.d . ,
alexakis , a. , pouquet , a. 2006 , journal of plasma physics , 73 , 377 oishi , j.s .
, mac low , m .- m .
2006 , , 638 , 281 ostriker , e.c . ,
stone , j.m . ,
gammie , c.f .
2001 , , 546 , 980 osullivan , s. , downes , t.p .
2006 , , 366 , 1329 osullivan , s. , downes , t.p .
2007 , , 376 , 1648 schmidt , w. , federrath , c. , hupp , m. , kern , s. , niemeyer , j.c .
2009 , a&a , 494 , 127 servidio , s. , carbone , v. , primavera , l. , veltri , p. , stasiewicz , k. 2007 , p&ss , 55 , 2239 vestuto , j.g . ,
ostriker , e.c . , stone , j.m .
2003 , , 590 , 858 wardle , m. 2004 , ap&ss , 292 , 317 wardle , m. , ng , c. 1999 , , 303 , 239
|
it is well known that non - ideal magnetohydrodynamic effects are important in the dynamics of molecular clouds : both ambipolar diffusion and possibly the hall effect have been identified as significant .
we present the results of a suite of simulations with a resolution of @xmath0 of turbulent decay in molecular clouds incorporating a simplified form of both ambipolar diffusion and the hall effect simultaneously .
the initial velocity field in the turbulence is varied from being super - alfvnic and hypersonic , through to trans - alfvnic but still supersonic .
we find that ambipolar diffusion increases the rate of decay of the turbulence increasing the decay from @xmath1 to @xmath2 .
the hall effect has virtually no impact in this regard .
the power spectra of density , velocity and the magnetic field are all affected by the non - ideal terms , being steepened significantly when compared with ideal mhd turbulence with exponents .
the density power spectra components change from @xmath3 to @xmath4 for the ideal and non - ideal simulations respectively , and power spectra of the other variables all show similar modifications when non - ideal effects are considered .
again , the dominant source of these changes is ambipolar diffusion rather than the hall effect .
there is also a decoupling between the velocity field and the magnetic field at short length scales .
the hall effect leads to enhanced magnetic reconnection , and hence less power , at short length scales .
the dependence of the velocity dispersion on the characteristic length scale is studied and found not to be power - law in nature .
|
rapid variability in intra - day variable ( idv ) sources is a long - standing problem as it implies apparent brightness temperatures in the radio regime which may exceed 10@xmath3 k , or requires relativistic beaming with extremely high doppler factors , coherent radiation mechanisms , or special geometric effects ( wagner & witzel 1995 ) . such high brightness temperatures would be well above the `` compton catastrophe '' limit @xmath4 k imposed by inverse - compton scattering ( kellermann & paulini - toth 1969 , slysh 1992 , kardashev 2000 ) when the photon energy density in the emission region exceeds the energy density in the magnetic field .
see krichbaum et al.(2002 ) and kedziora - chudczer et al .
( 2001 ) for recent reviews of idv sources .
the radio - loud quasar pks 0405 - 385 is an extreme example of an intra - day variable source with variations on timescales of @xmath5 d ( kedziora - chudczer et al .
1997 ) . making the assumption that the emission region subtends solid angle @xmath6 , where @xmath7 is the doppler factor and @xmath8 is the diameter distance to the source
, one can convert the observed 4.8 ghz flux to intensity , and obtain a variability brightness temperature of @xmath9 k. in this source , however , kedziora - chudczer et al .
( 1997 ) interpret the very short variability time as due interstellar scintillation requiring the angular diameter of the most compact component to be @xmath10 micro - arcsec , or smaller , corresponding to a solid angle subtended by the idv core @xmath11 where @xmath12 sr .
the fraction of the total flux they estimated to be associated with this compact idv core is @xmath13 , and the corresponding brightness temperature at 4.8 ghz is @xmath14 k. although much lower than @xmath15 , to reconcile this brightness temperature with the @xmath16 k limit would appear to require a very large doppler factor , @xmath17 , or an even larger angular diameter . to obtain a doppler factor as large as @xmath18 would require not only a jet lorentz factor of @xmath19 , but also very close alignment of the jet axis to our line of sight ( within @xmath20 ) .
the probability of such an alignment occurring by chance is then @xmath21 , in this case , which makes the very high doppler factor possibility unattractive .
the effective distance to the interstellar scintillation screen is crucial in determining the angular size of the source , and hence its brightness temperature .
the distance used by kedziora - chudczer et al .
( 1997 ) and walker ( 1998 ) is effectively the scale - height above the galactic plane of electron number density squared , i.e. @xmath22 , for which @xmath23 pc in the model of taylor and cordes ( 1993 ) for the free electron distribution in the galaxy .
however , that model was developed mainly for the consistent determination of pulsar distances from dispersion measures , and is most accurate at low galactic latitudes where the majority of pulsars are observed .
taylor and cordes ( 1993 ) themselves warn that one should be aware of uncertainties in their model associated with this
. the column density , @xmath24 , is fairly accurately determined , but the scale height of electron number density , @xmath25 , is less well determined , and @xmath26 is even less accurately known .
for example , in the recent model of gomez et al .
( 2001 ) for the free electron distribution in the galaxy @xmath27 pc .
it is interesting to note that beckert et al .
( 2001 ) suggest typical distances to the scattering medium of 200 pc , and that a scale height of about 100 pc seems to be required to explain idv in the case of 0917 + 624 .
krichbaum et al . (
2002 ) mention the possibility of an extremely clumpy ism . if this is the case , one could well question the use of an effective screen distance as it could be that the scintillation is due to an individual ionized cloud much nearer to us than the average distance .
in fact , if the distribution is highly peaked in the galactic plane the most probable , rather than average , distance is small .
the distance to the ionized cloud responsible for the extreme scattering event in 0954 + 658 is estimated to be @xmath28 pc ( cimo 2002 ) , while dennett - thorpe & de bruyn ( 2000 , 2002 ) suggest that the scattering region for idv in j1819 + 3845 may be located at about 20 pc , and possibly associated with the local bubble .
hence , i believe there could be considerable uncertainty in the distance to the scintillating material responsible for the extreme idv in pks 0405 - 385 . if , for example , the distance were smaller by a factor of 5 , this would translate into a factor of 25 reduction in the brightness temperature . a brightness temperature of @xmath29 k at @xmath30 ghz for the idv core in pks 0405 - 385 is still high , but
i will show that it can be achieved by standard electron synchrotron radiation using quite moderate doppler factors if one takes into account possible geometries of the emission region . in a recent paper ( protheroe 2002 )
i have explored the effect of emission region geometry on flux variability , and on the relationship between observed intensity and energy density for various source geometries for the case of optically thin emission and found that the average energy density in the source can be much less than one would estimate simply from the observed intensity .
although the radio emission from idv sources is usually assumed to be optically thick , if this is not the case then the above result may also have important implications for idv sources as the photon energy density responsible for causing the brightness temperature limit may actually be a few orders of magnitude lower than estimated from the intensity . in that case
, lower doppler factors would be required to avoid the compton catastrophe . in this paper
i shall explore the parameter space , including emission region geometry , of models able to reproduce the observed radio emission of the idv core of pks 0405 - 385 , and i shall model its spectral energy distribution ( sed ) from radio to gamma - ray frequencies .
in fig . [ fig : tb ] , i plot the available flux measurements of pks 0405 - 385 , divided by frequency squared , as gray symbols and gray vertical lines
. contemporaneous vla and atca data ( squares and triangles in the 1.4 ghz43 ghz range ) are also re - plotted as black symbols at the observed fluxes multiplied by @xmath31 , the fraction of the total flux assumed by kedziora - chudczer et al.(1997 ) to be associated with the idv core .
the right axis shows the brightness temperature of the idv core inferred by kedziora - chudczer et al .
( 1997 ) assuming the idv core subtends solid angle @xmath32 .
it is only the idv core component of these data that i am concerned with fitting .
nevertheless , it is interesting for comparison to include non - contemporaneous data at other wavelengths whose origin may or may not be associated with the same emission region . over the 1.4 ghz43 ghz range
, the intensity can be fitted well by @xmath33 ( @xmath34 ) .
such a spectrum would occur naturally if the emission were optically thin and if this frequency range were well below the critical frequency of the lowest energy electron .
note that for an xselectron with lorentz factor @xmath35 the critical frequency is @xmath36 where @xmath37 is the electron charge ( statcoulombs ) , @xmath38 is the component of magnetic field ( gauss ) perpendicular to the electron velocity , and @xmath39 is the electron mass ( grams ) . such an electron spectrum could occur naturally if the electrons were produced as secondaries of other particles , e.g. bethe - heitler pair - production by relativistic protons , or if explosive re - connection were responsible for electron acceleration .
for the purposes of this paper , i shall adopt a mono - energetic electron distribution .
for a mono - energetic isotropic electron distribution the synchrotron emission coefficient is given by @xmath40 where @xmath41 is the electron number density ( @xmath42 ) , @xmath43 and @xmath44 ( see e.g. rybicki and lightman 1979 ) . at low frequencies
@xmath45 @xmath46 the absorption coefficient for an isotropic electron distribution @xmath47 is @xmath48.\end{aligned}\ ] ] for the mono - energetic electron distribution considered here @xmath49 , and assuming @xmath50 , i find @xmath51\end{aligned}\ ] ] and this is plotted in fig .
[ fig : synch_alpha ] .
note that at low frequencies @xmath52 and this is plotted as the dotted line in fig .
[ fig : synch_alpha ] . from equations ( [ eq :
emiss_approx ] ) and ( [ eq : alpha_approx ] ) the source function at low frequencies is @xmath53 which is identical to the rayleigh - jeans approximation to a black body spectrum of temperature @xmath54 thus , in the optically - thick very - low frequency range the brightness temperature is constant at @xmath55 k , and then there is a transition to @xmath56 at frequency @xmath57 where the optical depth @xmath58 is unity , i.e.@xmath59 .
hence , @xmath60 can be used to estimate the lorentz factor of the electrons .
as noted earlier , the data appear to show @xmath61 over the 1.4 ghz43 ghz range corresponding to the contemporaneous idv observations . at other frequencies
it is not known whether the observed emission is due to the same compact component , or is due to a larger region , perhaps farther along the jet .
however , data at other frequencies may still be used to constrain the models .
since the brightness temperature at low frequencies is proportional to the electron lorentz factor , the brightness temperature problem is minimized by using the lowest possible lorentz factor .
i therefore take the highest frequency @xmath57 which is just consistent with the 1.4 ghz data , and adopt @xmath62 hz . apart from the normalization , for which i choose to take @xmath63 erg @xmath64s@xmath65hz@xmath65 ( after multiplying by @xmath13 ) , the other parameter determining the fit is the value of the critical frequency @xmath66 .
the lowest critical frequency @xmath66 which is consistent with the 43 ghz radio data is @xmath67 hz , and the highest critical frequency consistent with the r - band data is @xmath68 hz , and i shall adopt these two values in my modeling .
the resulting brightness temperature fits are shown by the solid curves in fig . [
fig : tb ] .
the challenge is now to find a combination of source parameters , i.e. doppler factor , jet - frame magnetic field , electron number density and emission region geometry , that gives the required intensity , and is physically possible .
if one views the emission region along the jet axis , one obtains the largest doppler boosting for any given jet lorentz factor , and i shall assume this to be the case for pks 0405 - 385 .
this may well be justified as such extreme idv radio quasars are extremely rare .
generally , the lower the energy density of synchrotron radiation photons in the emission region , the lower will be the doppler factor required to avoid the compton catastrophe on synchrotron photons .
however , as i shall show , very large doppler factors can cause a compton catastrophe on cosmic microwave background radiation ( cmbr ) photons .
i shall discuss the dependence of synchrotron photon energy density on the geometry of the emission region in the next section .
in a separate paper ( protheroe 2002 ) i have discussed the influence of emission region geometry on photon energy density in the emission region for the optically thin case . here
i shall extend this work to include optical depth effects and use a cylindrical emission region .
i shall consider the case of a cylinder of length @xmath69 and radius @xmath70 , with uniform emissivity @xmath71 and absorption coefficient @xmath72 , and determine the photon energy density @xmath73 averaged over the volume of the cylinder . at some arbitrary point @xmath74 within the cylinder
the intensity from direction defined by unit vector @xmath75 is @xmath76\}\end{aligned}\ ] ] where @xmath77 is the distance from @xmath74 to the boundary of the cylinder in direction @xmath75 .
the energy density at @xmath74 is @xmath78 using a monte carlo method one can sample a large number , @xmath79 , of directions @xmath80 , distributed isotropically , and then set @xmath81 the energy density averaged over the volume @xmath82 of the cylinder is then @xmath83 using a monte carlo method one can sample a large number , @xmath84 , of points @xmath85 , distributed uniformly throughout the volume of the cylinder , and then set @xmath86 hence , @xmath87\right\ } , \label{eq : uphot_tau}\end{aligned}\ ] ] and in this way one can calculate @xmath88 for various @xmath89 values , and absorption coefficients @xmath72 .
of course for emission in the jet frame , all the variables in equation ( [ eq : uphot_tau ] ) would be jet - frame variables . while the average energy density inside the cylinder is fixed for any set of @xmath90 , and @xmath70 , the intensity observed when viewing the cylinder can depend strongly on viewing direction . in order to obtain the highest observed intensity one would look in a direction such that the projected area of the cylinder is smallest . making the axis of the cylinder coincident with the jet axis ,
this would be achieved if one viewed the emission region at @xmath91 if the cylinder was short ( @xmath92 ) , and for @xmath93 if the cylinder was long ( @xmath94 ) , where primed coordinates correspond to jet - frame variables .
i shall consider the latter case , i.e.viewing the emission down the jet axis ( @xmath93 ) .
the jet - frame intensity is then simply given by @xmath95 $ ] where @xmath96 . in fig .
[ fig : uphot_tau ] i plot the ratio @xmath97 against @xmath98 for various values of @xmath99 .
as can be seen , the effect is very important where the emission is optically thin , e.g. for @xmath100 the average energy density is almost a factor @xmath18 lower than would be expected from the observed intensity ( note that the optical depth is lorentz invariant @xmath101 ) . in the next section
i shall consider various @xmath99 values when exploring the parameter space which could apply to the emission region for idv in pks 0405 - 385 .
the fits shown by the solid curves in fig . [ fig : tb ] correspond to @xmath63 erg @xmath64s@xmath65hz@xmath65 , @xmath62 hz , and @xmath102 hz or @xmath103 hz . in this section i shall determine the combinations of physical parameters which could give rise to these values , assuming a cylindrical emission region geometry , an isotropic mono - energetic electron distribution in the jet - frame , and observation along the jet axis .
the physical parameters describing the emission region are the perpendicular component of magnetic field @xmath104 as measured in the jet - frame , the doppler factor @xmath7 , the ratio @xmath99 of the emission region cylinder length to its radius as measured in the jet - frame , and an equipartition factor which gives the ratio of relativistic electron energy density to magnetic energy density @xmath105 where the ( 3/2 ) factor arises because @xmath104 is the perpendicular component of magnetic field , rather than its magnitude , and @xmath106 is the jet - frame electron lorentz factor . the solid angle subtended by the source is obtained from the angular radius of the source , i.e. @xmath107 , and the radius of the cylindrical emission region is obtained from the diameter distance and angular radius , i.e@xmath108 . note that for pks 0405 - 385 at redshift @xmath109 the diameter distance is @xmath110 cm for a @xmath111cdm model with @xmath112 and @xmath113 km s@xmath65 mpc@xmath65 ( here @xmath114 specifies the fraction of the cosmological closure density , whilst elsewhere in this paper @xmath114 is used for solid angles ) . noting that @xmath115 , and that in the optically thick region of the spectrum @xmath116 , from equation ( [ eq : source_fn ] ) one may obtain the lorentz factor the electrons would have if the emission took place in the observer frame , @xmath117 , and hence find the jet frame lorentz factor @xmath118 next , using equation ( [ eq : alpha_approx ] ) and that , by definition , the optical depth at jet - frame frequency @xmath119 must be unity @xmath120 substituting for @xmath121 from equation ( [ eq : equilib ] ) with @xmath106 from equation ( [ eq : gamma_prime ] ) and solving for @xmath104 i obtain @xmath122 then , using equation ( [ eq : nu_c ] ) for the jet - frame critical frequency , @xmath123 , i obtain @xmath124 finally , solving simultaneous equations ( [ eq : b_from_nu1 ] ) and ( [ eq : b_from_nuc ] ) i obtain @xmath125 @xmath126 fig .
[ fig : parameters ] shows the magnetic field
doppler factor parameter space for models fitting the radio intensity of the idv core of pks 0405 - 385 .
equations ( [ eq : d_model ] ) and ( [ eq : b_model ] ) define models which will give synchrotron spectra of mono - energetic electrons determined by @xmath127 , @xmath57 and @xmath66 , and these models are represented as solid lines in fig .
[ fig : parameters ] corresponding to either a fixed value of @xmath128 while varying @xmath129 as a parameter in these equations , or a fixed value of @xmath129 while varying @xmath128 .
all other variables are fixed at the values appropriate to pks 0405 - 385 , with fig .
[ fig : parameters](a ) being for @xmath102 hz and fig .
[ fig : parameters](b ) being for @xmath130 hz ( @xmath62 hz in both cases ) . if one allows the relativistic particle energy density to exceed the magnetic energy density , then very low doppler factors are possible . however ,
if the equipartition factor is greater than unity , this would be unstable .
it is therefore probably unrealistic for the relativistic particles to be too far from equipartition with the magnetic field , and so one should perhaps take more seriously results corresponding to @xmath131 , which i shall assume in what follows .
nevertheless , idv is truly a time - dependent problem , and episodes with @xmath132 are not ruled out completely .
if electrons are injected with jet - frame lorentz factor @xmath106 then the presence of dense radiation fields provides target photons for inverse compton scattering , and the generation of components in the sed at x - ray and gamma - ray frequencies which may or may not exceed observed x - ray flux and gamma - ray limit . if the electrons are accelerated in a quasi - continuous process such as diffusive shock acceleration by non - relativistic or mildly - relativistic shocks , then the presence of dense radiation fields may lead to excessive energy losses and prevent the electrons reaching the lorentz factor required to fit the observed radio spectrum . in either case , these effects may occur wherever the photon energy density in the emission region becomes comparable to or exceeds the energy density in the magnetic field .
in addition to the synchrotron emission itself , i shall assume that emission region is sufficiently far along the jet that the only other non - negligible field for inverse - compton scattering is the cosmic microwave background radiation ( cmbr ) .
if the emission region were close to the central engine then accretion disk radiation , broad line cloud emission , and torus infrared emission could also play a role ( donea & protheroe 2002 ) .
the average jet - frame energy density of the synchrotron emission inside the emission region can be estimated for given @xmath99 values from the observed intensity @xmath133 corresponding to @xmath134 plotted as the solid curves in fig . [
fig : tb ] . the energy density in the jet frame
is given by @xmath135 where @xmath136 i_\nu(\nu,\theta_{1/2 } ) d\nu,\end{aligned}\ ] ] @xmath137 from fig .
[ fig : synch_alpha ] , and @xmath138=\langle u'_{\nu'}\rangle /(4\pi i'_{\nu'}/c)$ ] given in fig .
[ fig : uphot_tau ] . the main contribution to the energy density for the spectrum fitted to
the radio observations of the idv core in pks 0405 - 385 comes from frequencies near to @xmath66 since this is where the sed , @xmath139 , peaks .
this spectrum is optically thin near @xmath66 and so i may make the approximation @xmath140 where @xmath141 for the spectra shown by the solid curves in fig . [
fig : tb ] , @xmath142hz@xmath143hz@xmath144 cm@xmath65sr hz and @xmath146hz@xmath147hz@xmath148 cm@xmath65sr hz . in the rest frame of the host galaxy
the cmbr would have had temperature @xmath149 , where @xmath150 k , and would have been isotropic .
i shall distinguish between the doppler factor for boosting from the jet frame to the host galaxy frame , @xmath151 , and the doppler factor for boosting from the jet frame to the observer frame , @xmath152 . in the jet frame ,
the black body temperature would depend on the angle @xmath153 with respect to the jet axis at which an observer in the jet frame looks @xmath154 where @xmath155 is the jet velocity as measured in the host galaxy frame , @xmath156 is its lorentz factor , and is given by @xmath157 for the case of observation of the agn jet at angle @xmath158 to its axis .
the energy density of the cmbr in the jet frame is then @xmath159 ^ 4 } \nonumber \\ & = & { 4\over 3}\gamma^2(1+z)^4a t_0 ^ 4 .\end{aligned}\ ] ] where @xmath160 and @xmath161 is the stefan - boltzmann constant .
note that whereas the jet frame energy density of synchrotron radiation decreases with assumed doppler factor , the energy density of the cmbr increases with assumed doppler factor .
this means that the compton catastrophe can not be avoided by using an arbitrarily large doppler factor .
in fact , very high doppler factors will result in a compton catastrophe due to inverse - compton scattering on the cmbr .
this has important consequences which i shall address next , and in subsequent sections . for efficient synchrotron radiation to take place the energy density in the magnetic field
should exceed the energy density in the radiation field @xmath162 .
\label{eq : b_min } \end{aligned}\ ] ] since the synchrotron photon energy density is large in the regime of low doppler factors , while the cmbr energy density is large in the regime of high doppler factors , i shall obtain separately the minimum magnetic field required in each case .
for pks 0405 - 385 being viewed along the jet axis , and using the approximation @xmath163 , valid for @xmath164 , the jet - frame energy density of the magnetic field is less than that of the cmbr when @xmath165 and this is independent of all model parameters except @xmath7 . this case
is shown by the shaded area on the right in figs .
[ fig : parameters](a)(b ) . in the case of synchrotron photons as targets for inverse compton scattering , i.e. the synchrotron self - compton ( ssc ) process , inverse compton losses dominate when @xmath166 because in this case the photon energy density depends on @xmath129 and @xmath99 , in addition to @xmath7 , the boundary between models ( @xmath167 ) is parametrized by substituting @xmath7 from equation ( [ eq : b_from_nuc ] ) into equation ( [ eq : b_min_ssc ] ) and solving for @xmath104 as a function of the parameters , and then using equation ( [ eq : b_from_nuc ] ) to obtain @xmath7 as a function of the parameters .
the resulting minimum value of @xmath104 is plotted against @xmath7 in figs .
[ fig : parameters](a)(b ) for various values of @xmath99 as the dotted lines ( labeled by @xmath99 ) bounding the shaded areas on the left .
the gyroradius , @xmath168 must be much smaller than the radius @xmath70 of the cylinder and so this provides , in principle , an additional lower limit to @xmath104 . however , this limit is well below the lower limit from equation ( [ eq : b_min ] ) already plotted , and therefore does not further constrain the models .
fifteen potential models which will be discussed later , i.e.combinations of @xmath38 , @xmath7 , @xmath129 and @xmath128 , are labeled as a to o in fig .
[ fig : parameters ] . before proceeding
, we should check if any of these models is ruled out by being optically thick to thomson scattering .
the thomson optical depth is @xmath169 where @xmath170 is the thomson cross section .
taking @xmath121 from equation ( [ eq : equilib ] ) with @xmath106 from equation ( [ eq : gamma_prime ] ) i obtain @xmath171 for the fifteen potential models the thomson optical depth ranges from @xmath172 ( model a ) to @xmath173 ( model o ) , and so thomson scattering can be neglected in this case .
i shall consider the following three emission processes : synchrotron radiation ( syn ) , inverse compton scattering of synchrotron photons ( ssc ) , and inverse compton scattering of cmbr photons ( icm ) .
assuming the inverse compton scattering is in the thomson regime , the rate of energy loss by either process is @xmath174 where @xmath175 for synchrotron losses , @xmath176 for ssc losses and @xmath177 for inverse - compton losses on the cmbr .
the fraction of emitted radiation by each process is then @xmath178 this does not mean , however , that the observed energy flux for the inverse - compton scattered cmbr is in this ratio to the other two components because , whereas the synchrotron emission and ssc emission is isotropic in the jet frame , the inverse - compton scattering of the cmbr is not because of the anisotropy of the cmbr in this frame .
one may well expect strong peaks in the sed due to inverse - compton scattering for models close to or within the shaded areas in figs .
[ fig : parameters](a)(b ) .
slysh ( 1992 ) also noted that gamma - ray emission may occur in idv sources . the intensity of synchrotron emission in direction @xmath158 in the observer frame is simply given by @xmath179\}\end{aligned}\ ] ] where @xmath180 . in the present paper
i shall assume that we view the emission emitted at angle @xmath181 to the jet axis .
the electrons are assumed to be isotropic and mono - energetic in the jet frame , and have lorentz factor @xmath182 , and velocity @xmath183 . for simplicity
, i shall also take the synchrotron target photons to be isotropic in the jet frame , while using the results from section 3 to normalize their spectrum . for isotropic target photons in the jet frame with frequency @xmath184 , the frequency of the scattered photons range between 0 and @xmath185 .
the jet - frame ssc emissivity ( erg @xmath42 s@xmath65 sr@xmath65
hz@xmath65 ) is then @xmath186 where @xmath187 is the average jet - frame synchrotron specific photon number density ( photons @xmath42 sr@xmath65 hz@xmath65 ) , @xmath188 and @xmath189 ( blumenthal and gould 1970 ) , giving @xmath190 finally , this is doppler boosted to the observer frame @xmath191 in the jet frame , the black body temperature would depend on the angle @xmath192 with respect to the jet axis at which the cmbr target photons propagate @xmath193 such that their specific photon number density ( photons @xmath42 sr@xmath65 hz@xmath65 ) is @xmath194 - 1}.\end{aligned}\ ] ] for target photons in the jet frame with frequency @xmath184 propagating at angle @xmath192 to the jet axis , the frequencies of the photons scattered by electrons propagating parallel to the jet axis are uniformly distributed between 0 and @xmath195 in the approximation that the scattered photons are isotropic in the electron rest frame .
the jet - frame emissivity ( erg @xmath42 s@xmath65 sr@xmath65 hz@xmath65 ) for ic on the cmbr in the jet direction is then @xmath196 where @xmath197 . finally , @xmath198
the angular radius inferred by kedziora - chudczer et al .
( 1997 ) is @xmath199 rad , i.e.@xmath200 . examining the parameter space for @xmath201 in fig .
[ fig : parameters](a ) , one sees that all models with this angular radius lie well inside the region where ic on the cmbr dominates the energy losses of electrons . for @xmath131 ( equal energy density in magnetic field and relativistic electrons ) and @xmath202
this would correspond to a point roughly mid - way between models a and b , and would require a doppler factor of @xmath203 . by choosing @xmath204 one may reduce the doppler factor , but
unrealistically large @xmath99 values are needed to reduce @xmath7 by a large factor . to reduce the doppler factor to a reasonable value , i.e. @xmath205 ,
would require the angular diameter to be larger by only a factor of @xmath30 which , as discussed earlier , i believe is quite possible ( for an angular diameter larger by a factor of 10 , even @xmath206 is possible ) .
because of the dominance of ic losses on the cmbr for models a d , one would expect the ssc peak at gamma - ray energies in the sed to exceed the synchrotron peak for these models .
the seds have been calculated for models a e and are shown in fig .
[ fig : sed](a ) and , as expected , show strong gamma - ray emission for models a d . non - contemporaneous x - ray and gamma - ray data are shown , and may be used with caution as upper limits .
i have added to fig .
[ fig : sed ] the sensitivity of the ibis gamma - ray detector on integral ( parmar et al .
2002 ) , and the expected sensitivities of glast ( gehrels and michelson 1999 ) , and hess ( hofmann et al .
2001 ) and cangaroo iii ( enomoto et al.2002 ) , the last two being southern hemisphere atmospheric cherenkov telescopes nearing completion .
taking the x - ray data as an upper limit does not rule out any of these models .
however , the egret upper limit obtained from figure 3 of hartman et al .
( 1999 ) appears to exclude models a c but be consistent with models d e ( note , however , that the egret data are not contemporaneous with the idv data ) .
models d e have @xmath207@xmath208 rad and @xmath20940 , respectively .
as expected from equation [ eq : d_model ] , increasing @xmath66 from @xmath210 hz to @xmath103 hz reduces @xmath7 by a factor of 10 if one keeps all other parameters fixed ( note , however , that this reduction in @xmath7 is accompanied by an increase in @xmath104 ) .
this is clearly seen in fig .
[ fig : parameters](b ) which is for @xmath130 hz , and thereby presents more opportunity for obtaining reasonable values of @xmath7 . for @xmath131 and @xmath202 ,
models f j have the same angular radius range ( @xmath211@xmath208 rad ) as models a e , but with considerably lower doppler factors , and these models range from being in the ic on cmbr dominated regime ( model f expect highest gamma - ray flux , lowest ssc flux ) to being in the ssc dominated regime ( model j expect highest ssc flux , lowest gamma - ray flux ) .
the calculated seds for models f j are shown in fig .
[ fig : sed](b ) and show the trends expected .
possibly models i and j may be ruled out by the x - ray data taken as an upper limit ( note , however , that the x - ray data is not contemporaneous with the idv data ) .
models f i predict high gamma - ray fluxes from mev energies to 100 gev energies , depending on model , which could be tested observationally by integral , glast , hess and cangaroo iii .
models f i have @xmath212@xmath213 rad and @xmath21420 , respectively . as also expected from equation [ eq : d_model ] ,
increasing @xmath99 from 1 to 316 results in a reduction in @xmath7 by a factor of @xmath215 ( plus a reduction in @xmath104 ) if one keeps all other variables fixed .
models k o in fig .
[ fig : parameters](b ) have all other parameters the same as models f j , and the resulting seds are shown in fig .
[ fig : sed](c ) . because of the reduction in synchrotron photon energy density associated with @xmath94 ,
although not identical , the seds of models k o are qualitatively similar to those of models f j , with model o being ruled out by both the gamma - ray limit and the x - ray data , although again i note that these data are not contemporaneuos with the idv data .
models k n have @xmath212@xmath216 rad and @xmath2178 .
model m , having @xmath129 only @xmath30 times that assumed by kedziora - chudczer et al . ( 1997 ) has a doppler factor of @xmath218 , and lower doppler factors are possible with , for example even larger @xmath99 .
bearing in mind that no attempt has been made to fit the ( non - contemporaneous ) x - ray and gamma - ray data , and that a simple mono - energetic electron spectrum has been used , it seems that models which are able to fit the idv radio data tend to predict observable fluxes of x - ray , and/or gamma ray emission at sub - gev and/or @xmath21910 gev energies .
use of a more sophisticated electron spectrum is unlikely to alter this conclusion .
hence , it may well be profitable for space gamma ray telescopes such as integral and glast , and southern hemisphere atmospheric cherenkov telescopes such as hess and cangaroo iii to look for emission from pks 0405 - 385 and other extreme idv radio galaxies .
the rapid intra - day variability in pks 0405 - 385 is almost certainly due to interstellar scintillation effects .
the radio flux does , however , also vary on longer timescales 10 per cent change in 2 months in july 1996 ( corresponding to @xmath220 yr for 50 per cent change ) , and has very intense for periods of about a month ( kedziora - chudczer et al .
1997 ) . a mechanism that may cause such variability would be the emission region moving along the jet and passing a region where external factors cause compression / expansion of the emission region , increasing / decreasing the magnetic field , and causing adiabatic acceleration / deceleration of relativistic charged particles , thereby affecting the observed intensity .
change can also occur as a result of the emission region passing through a bend in the jet causing , amongst other things , a change in viewing angle with respect to the motion of the emission region , and hence a change in doppler factor . for the examples
above , the observer - frame variability time depends on the geometry of the emission region ( see , e.g. , protheroe 2002 ) and , for viewing the assumed cylindrical emission region along its axis , is at least @xmath221 .
in general , arbitrarily low doppler factors are possible for arbitrarily high @xmath99 .
however , in this case this is clearly at the expense of having a time scale for variability which may be unreasonably large , and of course also requires the jet to be extremely well collimated over the length of the emission region .
alternatively a shock may pass through the emission region compressing the magnetic field and accelerating particles a pre - existing mono - energetic population of relativistic particles will receive an energy boost by a factor of @xmath222 and acquire a power - law tail as a result of diffusive shock acceleration . in the case of a plane shock moving at jet - frame speed @xmath223 along the jet ( see , e.g. protheroe 2002 ) ,
the observer - frame variability time in our case ( @xmath93 ) is at least @xmath224 which is longer ( shorter ) than @xmath221 if @xmath225 is less than ( greater than ) 0.5 for a forward - moving shock . except for a point - like emission region following a bent or helical trajectory such that the doppler boosting factor , changes rapidly with time , the variability time can not be shorter than the observer - frame energy - loss time scale , @xmath226 .
this is easily obtained from equation ( [ eq : dedt ] ) and is plotted ( thin curves ) against doppler factor in fig .
[ fig : dopp_tvar ] for the following cases : @xmath201 , @xmath131 and @xmath227 ( solid curves ) @xmath130 , @xmath131 and @xmath227 ( dotted curves ) ; and @xmath130 , @xmath131 and @xmath228 ( dashed curves ) . as noted above , the variability time will in general be longer than the energy - loss time scale due to the dimensions of the emission region . for the present geometry and viewing angle ,
i take the variability time to be @xmath229 ^ 2\right\}^{1/2},\end{aligned}\ ] ] but one should note that in the case of shock excitation that it can be longer than this for @xmath230 , or as short as @xmath226 if @xmath231 . in fig .
[ fig : dopp_tvar ] , i have also plotted @xmath232 vs.@xmath7 for the same three cases , and the values for models a o are indicated by the letters . of the models with doppler factors less than @xmath233 ,
models c e give variability times @xmath22030 years .
models f j give variability times of @xmath220 week to @xmath234 month , and are quite compatible with the long - term variability observed for pks 0405 - 385 .
models k o , with larger @xmath99 , give variability times of @xmath220 month to @xmath235 years , with model m having @xmath236 year .
to obtain low doppler factors , without invoking extreme variations from equipartition , requires a larger angular diameter for the idv core of pks 0405 - 385 than the @xmath10 micro - arcsec assumed by kedziora - chudczer et al .
( 1997 ) corresponding to a distance of @xmath237 pc to the ionized material responsible for the interstellar scintillation .
a further reduction in the minimum doppler factor needed to avoid the compton catastrophe can be obtained by having optically thin idv core emission , preferably originating from a region elongated along the jet , and observing it at a small viewing angle with respect to the jet axis .
an angular diameter a factor of @xmath2154 larger than assumed by kedziora - chudczer et al .
( 1997 ) and walker ( 1998 ) corresponds roughly to models d , h i and m
n , all of which are allowed by the present data , and have doppler factors @xmath238 , 8020 and 308 , respectively .
interstellar scintillation screen distances as small as 20 pc , possibly associated with material in the local bubble , have previously been invoked to explain idv in j1819 + 3845 ( dennett - thorpe & de bruyn 2000 , 2002 ) . by using a distance of 20 pc
, even lower doppler factors could easily fit the observations . in conclusion , if one observes optically thin emission along a long narrow emission region , the average energy density in the emission region can be significantly lower than @xmath239 .
if one takes into account the uncertainty in the distance to the ionized clouds responsible for interstellar scintillation causing rapid idv in pks 0405 - 385 the brightness temperature could be as low as @xmath2 k , or lower , at 5 ghz .
the radio spectrum can be fit by @xmath61 expected if the emission is optically thin and the electrons are mono - energetic , or have a minimum lorentz factor whose critical frequency is well above the frequency range of interest .
such a spectrum would occur naturally in models in which @xmath240 are produced as secondaries of interactions of protons , e.g. by bethe - heitler pair production , or through @xmath241 decay following collisions of with matter or low - energy target photons ( pion photoproduction ) by protons , or by neutrons themselves produced in pion photoproduction interactions ( e.g. @xmath242 ) perhaps closer to the central engine .
the combination of all these factors enables the compton catastrophe to be avoided , and also predicts that x - ray and gamma - ray emission at observable flux levels should be expected from compact cores of extreme idv sources such as pks 0405 - 385 .
my interest in this subject was stimulated by a workshop in september 1999 organized by the rcfta at the university of sydney .
i thank lucyna kedziora - chudczer , dave jauncey and stefan wagner for supplying data from observations of pks 0405 - 385 .
this research is supported by a grant from the australian research council .
|
i re - examine the brightness temperature problem in pks 0405 - 385 which is
an extreme intra - day variable radio quasar with an inferred brightness temperature of @xmath0 k at 5 ghz , well above the compton catastrophe limit of @xmath1 k reached when the synchrotron photon energy density exceeds the energy density of the magnetic field .
if one takes into account the uncertainty in the distance to the ionized clouds responsible for interstellar scintillation causing rapid intra - day variability in pks 0405 - 385 it is possible that the brightness temperature could be as low as @xmath2 k at 5 ghz , or even lower .
the radio spectrum can be fitted by optically thin emission from mono - energetic electrons , or an electron spectrum with a low - energy cut - off such that the critical frequency of the lowest energy electrons is above the radio frequencies of interest .
if one observes optically thin emission along a long narrow emission region , the average energy density in the emission region can be many orders of magnitude lower than calculated from the observed intensity if one assumed a spherical emission region .
i discuss the physical conditions in the emission region and find that the compton catastrophe can then be avoided using a reasonable doppler factor .
i also show that mev to 100 gev gamma - ray emission at observable flux levels should be expected from extreme intra - day variable sources such as pks 0405 - 385 .
galaxies : active galaxies : jets gamma - rays : theory radiation mechanisms : non - thermal quasars : pks 0405 - 385 .
|
although the proton is the simplest system in which the three colors of qcd neutralize into a colorless bound state , we still do not know how to describe the proton in terms of its fundamental quark and gluon degrees of freedom from basic principles .
the structure of the proton is rather complicated due to the nonperturbative and relativistic nature of the quark and gluon in the protons .
the complication also comes from the presence of sea quarks in the proton .
the sea flavor symmetry naively assumed in the gottfried sum rule @xcite , which is a symmetry between the light flavor @xmath3 and @xmath4 sea quarks inside the proton , was disproved by experiments of both deep inelastic scattering and drell - yan processes @xcite .
many theoretical attempts have been made to describe the origin of the nucleon sea and its antiquark asymmetry @xcite .
it is assumed that the primary mechanism to generate the sea is gluon splitting into @xmath5 and @xmath6 pairs .
field and feynman @xcite suggested that the extra valence @xmath3 quark in the proton could lead to a suppression of @xmath7 relative to @xmath8 via pauli blocking .
but a subsequent calculation @xcite found that the effects of pauli blocking are very small , and this result has been confirmed by another calculation@xcite .
thus , it is believed that there must be a nonperturbative origin .
for example , the meson - cloud inside the nucleon can account for such asymmetry @xcite and chiral quark models @xcite .
also the large-@xmath9 approach @xcite can explain the flavor asymmetry of the antiquark distribution . another attempt to understand
the sea flavor asymmetry of the proton is from a pure statistical consideration in a kinetic equilibrium model @xcite or `` statistical balance model '' as called in previous papers .
the idea is rather simple and perspicuous : while the sea quark - antiquark @xmath10 and @xmath11 pairs can be produced by gluon splitting with equal probabilities , the time - reversal invariant processes of the annihilation of the antiquarks with their quark partners into gluons are not flavor symmetric due to the net excess of @xmath3 quarks over @xmath4 quarks . as a consequence , the @xmath12 quarks have a larger probability to annihilate with the @xmath3 quarks than that of the @xmath13 quarks , and this brings an excess of @xmath13 over @xmath12 inside the proton . taking the proton as an ensemble of a complete set of quark - gluon fock states , and assuming the probability of ` arriving in ' one state from others equals to the probability of ` leaving ' it , one can obtain the probabilities of finding every fock state ( state density ) in the proton .
thus one can calculate the quark and gluon content of the nucleon from a pure statistical consideration .
it is interesting that the model gives a sea flavor @xmath12 and @xmath13 asymmetry as @xmath14 \sim 0.132 $ ] , which agrees with the experimental data the following diagram can describe the ` state shifting ' between states . assuming kinetic equilibrium , we have these kinetic equilibrium equations : @xmath15 where @xmath16 is @xmath17 state density
, @xmath18 is the non - normalized state - shift probability(nssp ) of @xmath19,@xmath20 is the total state number .
also there is the normalization condition @xmath21 if we know @xmath18 , we can derive state densities @xmath16 s by solving a system of @xmath20 linear algebraic equations when @xmath20 is a finite number .
if @xmath20 is infinite , we can get @xmath16 by asymptotic approach in some case if @xmath16 converges as @xmath22 .
actually , if we change @xmath18 to @xmath23 , where @xmath24 is a arbitrary constant , the result would be the same .
it means we only need the ratios of nssps @xmath18 s .
if only considering the particle numbers of quark , anti - quark and gluon , the proton state can be described as an ensemble of fock states @xmath25 because the u - quark number @xmath26,and @xmath27 , all fock state can be denoted with just three numbers as @xmath28 . in order to derive the state density @xmath29 we should know the probability of states shifting .
we introduce the rate @xmath30 as a quark splitting ability factor , there are @xmath31 quarks(including antiquarks ) in the initial state , so the nssp of @xmath32 is @xmath33 we also introduce the splitting rate @xmath34 and @xmath35 , so the nssp of @xmath36 and @xmath37 is @xmath38 and the nssp of @xmath32 is @xmath39 now , we consider the time - reversal process and assume those fusion rates @xmath40 for time - reversal invariance . hence , the nssp of @xmath41 is @xmath42 the nssp of @xmath43 is @xmath44 the nssp of @xmath45 is @xmath46 we can see that the probability of @xmath10 annihilation is larger than @xmath11 annihilation in all of the proton states because of valence quark asymmetry .
this is the origin of the sea quark flavor asymmetry .
it is assumed that all the splitting and fusion rates are the same in the previous papers @xcite .
if we get all the non - normalized state - shift probabilities @xmath18 , the state densities can be derived out if the particle numbers @xmath47 are finite .
we set an artificial limit @xmath48 and solve the finite linear equations .
the numeric state densities are then derived .
the sea quark flavor asymmetry can be written as : @xmath49=\sum_{\bar{u},\bar{d},g}(n_{\bar{d}}-n_{\bar{u}})\rho_{|n_{\bar{u}},n_{\bar{d}},n_{g}>}.\end{aligned}\ ] ] the sea quark flavor asymmetry converges to 0.133 when @xmath50 increases .
the result is consistent with experiment data @xcite .
some subsequent works @xcite followed the kinetic equilibrium principle to study the spin of nucleons and the parton distributions in the proton and pion , and obtained quite good results agreeing with the corresponding experimental values .
however , in the previous works , we assumed that all the splitting rates are the same as @xmath51 and did not estimate the `` error bound '' caused by the assumption .
as we can imagine , if the splitting - rates vary in different orders of magnitude , the convergence of flavor asymmetry will be bad .
it is necessary to solve large @xmath50 linear equations .
so we need a convenient numerical method to explore the effects of different splitting - rates and to study more complex hadronic states .
monte carlo simulation also can give the numeric state densities instead of solving algebraic equations even when the number of states is infinite . here , we want to explain some details about the monte carlo evolution on kinetic equilibrium and prove the equivalence between monte carlo evolution approach and solving algebraic equations .
let us start with an arbitrary initial state @xmath17 , and then let it make a possible shifting during each unit step .
the probability of the state @xmath17 shifting to @xmath52 is @xmath23 . here
, @xmath24 is an arbitrary large constant we introduced to ensure that the total shifting probability for each prior state is less than @xmath53 .
it is required that @xmath54 for all prior states @xmath17 , so the probability of staying in the prior state @xmath17 is @xmath55 the state evolves step - by - step as random walk , and we record the number of iteration steps as @xmath56 while the state @xmath17 is emerging . and after a large number of iteration steps @xmath57 , the normalized @xmath17 emerging probability is @xmath58 . for each step while the state is @xmath17 , the next step has the probability @xmath23 to be @xmath52 .
so there are the times @xmath59 of state shifting @xmath60 .
of course , other states also can shift to @xmath52 , meanwhile @xmath52 has chance to stay at @xmath52 .
that means the number of those steps @xmath52 emerging should be @xmath61 the equation can be reduced to @xmath62 the equation is independent of the constant @xmath24 .
the value of @xmath24 only determines the number of iteration steps needed to arrive at the equilibrium state after starting from an arbitrary initial state . we can find the above equation is just the kinetic equilibrium equation ( [ eq1 ] ) , if we consider that the normalized @xmath17 emerging probability @xmath58 is equivalent to the state density as @xmath63 and we also have the sum condition @xmath64 which is equal to the normalization condition eq([eq2 ] ) .
hence , we proved the equivalence of the monte carlo simulation approach and solving algebraic equations .
the monte carlo simulation approach provides a powerful method for solving kinetic equilibrium ensemble problems .
this method is error - controllable and very useful especially on complex multistate systems , such as the applications to other hadrons in the following sections .
we gain the same value of the sea quark flavor asymmetry @xmath65 in the proton as expected . here
, the error bar @xmath66 is the standard deviation of results with different random number series , and the deviation will decrease when computing time increases .
the fusion rate should be the same as the splitting rates for a time - reversal process .
in other words , the evolution in the proton should be time - reversal invariant .
but there is no principle requires that the quark and gluon splitting evolution abilities of @xmath67 and @xmath68 are equal .
therefore we should introduce three splitting - rates @xmath69,@xmath34 and @xmath35 to represent the quark and gluon splitting evolution abilities which are determined by the dynamics of quarks and gluons .
each rate enhances the corresponding splitting or fusion evolution probability . in previous works
, we assumed that all the splitting - rates are the same to be @xmath70 and did not estimate the `` error band '' caused by the assumption . in the present work
, we introduced a numerical monte carlo approach .
this new method is easy to apply to complex systems , and it is easy to put the variable splitting rates in evolutions and calculate the deviation caused by them . in the above section , we can see that the state densities or results are independent of the constant @xmath24 . the numerical value of @xmath71 , for example , is input as @xmath72 .
therefore the result does not depend on the absolute value of @xmath71 .
it means that the sea quark asymmetry does not depend on the absolute values of those splitting rates .
only two ratios between three splitting rates will affect the state densities and the value of sea quark flavor asymmetry .
so , we can fix the rate @xmath73 , and vary the other two ratios @xmath74 and @xmath75 as two parameters in the model .
ccccccccc @xmath14\times 100 $ ] & & & @xmath74 & & + @xmath76 & 100 & 10&1&0.1 & 0.01 & 0.001 + 0&@xmath77&@xmath78&@xmath79&@xmath80&@xmath81&@xmath82 + 1&@xmath83&@xmath84&@xmath85&@xmath86&@xmath87&@xmath88 + 2&@xmath89&@xmath90&@xmath91&@xmath92&@xmath93&@xmath94 + 5&@xmath95&@xmath96&@xmath97&@xmath98&@xmath99&@xmath100 + 10&@xmath101&@xmath102&@xmath103&@xmath104&@xmath105&@xmath106 + 100&@xmath107&@xmath108&@xmath108&@xmath109&@xmath110&@xmath111 + in table [ table1 ] , the values of sea quark asymmetry for different ratios of splitting rates are listed .
the previous result @xmath65 is reproduced when @xmath112 . from table
[ table1 ] , we can see that the asymmetry value @xmath14 $ ] is not sensitive to the model parameter @xmath74 , it is almost fixed when @xmath74 varies in a very large range over five order of magnitudes .
we also can find that the values of asymmetry are always larger than @xmath113 whatever the splitting rates vary in an arbitrary large range .
it reflects the principle of statistics contributes the dominant part of sea quark flavor asymmetry .
the asymmetry only has a variation @xmath114=(0.12 - 0.16)$ ] which is within @xmath115 when @xmath116 varies in the range @xmath117 , and still a small variation @xmath118=(0.12 - 0.18)$ ] even when @xmath119 varies in a larger magnitude range @xmath120 .
so the effect brought from details of the dynamics is small and within the bound of the experiments uncertainty . by
now , we do not consider the probability of @xmath121 splitting and @xmath122 recombination yet , because the probability is suppressed by coupling constant and `` three - body '' splitting kinematics .
@xmath121 can be regarded as two successive @xmath123 , and its effect is same as the effect of increasing @xmath35 , as we can see from table [ table3 ] .
however , the rate of three - body splitting @xmath124 must be much smaller than two - body splitting @xmath125 or @xmath126 , because the three - body phase space in perturbative qcd is suppressed by a factor of 2 - 3 order of magnitudes comparing with the two - body splitting . though the parton splitting in hadrons is a strong - coupling non - perturbative process
, we believe that we still can safely assume @xmath127 which only causes a very small enhancement as shown in table [ table3 ] .
the effect of the splitting @xmath124 is thus negligible .
ccccccc & & & @xmath128 & + 0&0.1&0.2&0.4&0.6&0.8&1.0 + @xmath129&@xmath130&@xmath131&@xmath132&@xmath133&@xmath134&@xmath135 + because the effect of the splitting @xmath124 and recombination @xmath122 is negligible and the asymmetry value of @xmath14 $ ] is almost independent of the parameter @xmath136 , there is only one parameter @xmath137 can vary the asymmetry .
this parameter is qcd relevant and it is the only input from dynamics .
if the parameter could be fixed by analysis of qcd , the deviation on sea quark asymmetry caused by the details of dynamics can be determined and the sea quark flavor asymmetry in proton is predictable .
these two splitting vertices are qcd vertices and have the same coupling constant .
the splitting kinematics of @xmath138 and @xmath139 are also similar .
so , the splitting rates of @xmath138 and @xmath140 should be in the same order of magnitude .
the assumption can be supported by the integrations of altarelli - parisi(a - p ) splitting functions . though these equations are valid in the perturbative region and the parton splitting in hadrons is a nonperturbative process
, the ratio of the total splitting rates is still inspirational .
the ratio parameter @xmath75 can be heuristically `` derived '' from altarelli - parisi splitting functions@xcite .
the a - p splitting functions are @xmath141 , \nonumber \\ & & p(g\to q(z)\bar q)=t_r[z^2+(1-z)^2 ] , \nonumber\end{aligned}\ ] ] where the color factors @xmath142 , @xmath143 and @xmath144 . the integrations of a - p splitting functions are assumed to be the total probabilities of quarks and gluons splitting .
so the splitting - rates directly to be : @xmath145 the rates @xmath146 and @xmath147 are logarithmic divergent when the integration limit @xmath148 , but fortunately the ratio between the two rates is not divergent , and thus we have the model parameter @xmath149 when @xmath150 .
the ratio parameter is not sensitive to the integration limit @xmath151 .
for example , when @xmath152 , the ratio is @xmath153 which is close to @xmath154 . such small deviation change on parameter @xmath137 dose not have effect on sea quark asymmetry .
considered the integration limit is relative to @xmath155 scale , then the model parameter @xmath75 and sea - quark asymmetry are not sensitive to @xmath155 scale .
we estimated the ratio parameter by the perturbative a - p splitting functions , it is just the ratio of color factors .
we assume the parameter value is still similar in the nonperturbative region .
the nonsensitive parameter @xmath156 also can be derived by above method .
but , it is relevant to the integration limit or @xmath155 scale .
the dependence can be extracted as @xmath157 when @xmath158 is small on the order of magnitude and becomes zero when @xmath150 .
for example , the value of parameter @xmath159 when @xmath160 , and the value is not sensitive to the magnitude of @xmath158 or @xmath155 scale because of its @xmath161 dependence .
we can see from table.[table1 ] , the sea - quark asymmetry is not sensitive to this parameter even it is so small . as discussed above ,
the ratio @xmath137 is almost fixed to ratio of color factors as @xmath154 and the asymmetry is independent of other details except the parameter @xmath137 .
therefore we arrived at the following conclusion : after considering the detail of qcd especially the color factors , we can predict the sea quark flavor asymmetry in proton is @xmath162 .
it is enhanced a little compared to the value given in the previous papers .
more precise measurement of @xmath114 $ ] is needed to examine the statistical balance model .
the @xmath163-dependent @xmath164 $ ] can be derived from deep inelastic scattering and drell - yan processes , and @xmath165 dx $ ] is given by extrapolating @xmath164 $ ] to @xmath166 and @xmath167 .
the sea quark asymmetry values from three collaborations are listed in table .
[ table4 ] , they are all consistent with the sea quark asymmetry value predicted above
. the value of e866 seems a little bit smaller compared to the prediction value , but the @xmath163 range of the e866 measurement is narrow and the uncertainty brought by extrapolating to small @xmath163 is out of control .
so , more precise measurements are needed to test the prediction .
ccc experiment & @xmath163 range & @xmath168 dx \;\;\;$ ] + e866 & @xmath169 & @xmath170 + nmc & @xmath171 & @xmath172 + hermes & @xmath173 & @xmath174 +
because the sea quark asymmetry value is not sensitive to details of dynamics and only depends on the parameter @xmath75 which is almost fixed as @xmath154 , then it should not only work for the proton , but also for the mesons and other baryons .
we suppose the statistical model also has validity on predicting sea quark asymmetry in other hadrons .
m. alberg , e. m. henley @xcite and c .- b .
yang @xcite derived the parton distributions of pions according the statistical model , but the sea quark asymmetry is zero because of the same valence quark number in pions .
while the valence quark numbers of the @xmath3 and @xmath4 quarks are different for the kaons , for example , @xmath175 has one @xmath3 valence quark and no @xmath4 valence quark .
the statistical balance model predicts the sea quark asymmetry value @xmath176 in @xmath177 , when @xmath178 . in the same way ,
the sea quark asymmetry value @xmath14=-0.275 $ ] in @xmath179 and @xmath180=-0.275 $ ] in @xmath181 , @xmath182 in @xmath183 .
these sea quark asymmetry values are also not sensitive to dynamics as shown in table [ table5 ] .
c|ccccccc @xmath14 $ ] & & @xmath74 & + @xmath76 & 100 & 10&1&0.1 & 0.01 + 0&@xmath184&@xmath185&@xmath186&@xmath186&@xmath187 + 0.1&@xmath186&@xmath188&@xmath189&@xmath190&@xmath190 + 1&@xmath191&@xmath192&@xmath193&@xmath194&@xmath195 + 5&@xmath196&@xmath197&@xmath198&@xmath199&@xmath200 + 10&@xmath201&@xmath202&@xmath202&@xmath202&@xmath203 + we can see from table [ table5 ] that the asymmetry @xmath14 $ ] is independent of @xmath136 and varies in a small range 0.263 - 0.31 as @xmath137 varies in a large range 0 - 10 .
we also use our statistical model to predict sea quark asymmetry for baryons . in a previous paper@xcite , l. shao _
et al_. derived the octet baryons sea quark asymmetry values by the method of solving linear equations .
they give @xmath14=0.41 $ ] in @xmath204 and @xmath14=0.276 $ ] in @xmath205 . in this paper , we get the same number by the monte carlo approach .
we can find that the sea quark asymmetry value in @xmath205 is almost the same as the meson @xmath175 because their @xmath3 and @xmath4 valence quark numbers are the same .
so , in the statistical model , the @xmath206 valence quark number in the hadron has a negligible effect on the @xmath14 $ ] sea quark asymmetry .
we also find the sea quark asymmetry values in the octet baryons are not sensitive to details of dynamics , they just depend on the valence quark numbers in those baryons . the asymmetries @xmath14 $ ] in @xmath204 and @xmath205 are enhanced a little to be @xmath207 and @xmath208 when @xmath209 .
besides octet baryons , we also derived @xmath2 baryons sea quark asymmetry value as : @xmath210 where , @xmath209 .
the sea quark asymmetry in @xmath211 is the same as in proton because of their same @xmath3 and @xmath4 valence quark numbers .
of course , the asymmetry in @xmath212 is the same as in neutron .
we also derived exotic baryons ( pentaquark states ) sea quark asymmetry values as : @xmath213 where , the sea quark asymmetry values are the same as in the proton because of their same @xmath214 and @xmath215 valence quark numbers
. if there is such a pentaquark state @xmath216 , then its sea quark asymmetry value would be @xmath14=0.21 $ ] derived by the statistical model .
in the previous works in the statistical balance model , the sea quark flavor asymmetry @xmath114\equiv\int dx(\bar d(x)-\bar u(x))$ ] in the proton was computed using the `` linear equations method '' .
because of the difficulty and limit of the linear equations method , it is hard to apply the method to more complex systems .
it is also assumed that all the splitting - rates are the same , @xmath217 in the previous works , and the `` error band '' caused by the assumption was not estimated . in the present work
, we introduced a numerical monte carlo approach .
this new method is easy to apply to complex systems , such as other mesons and baryons .
we also introduced the variable splitting rates representing details of the dynamics and studied the dependence on them .
we find the sea quark flavor asymmetry in the proton is always larger than @xmath113 whatever the splitting rates vary over an arbitrary large range .
it reflects the statistics principle contributes the dominant part of the asymmetry .
the asymmetry is almost independent of the model parameter @xmath218 and only changes within 30% when @xmath119 varies in the range @xmath219 .
so the effect caused by details of the dynamics is small and within the bound of the experiments uncertainty .
however , these two splitting vertices are qcd vertices and have the same coupling constant .
the splitting kinematics of @xmath138 and @xmath140 are also similar .
so the splitting rates of @xmath138 and @xmath140 should be in the same order of magnitude .
the assumption can be supported by the integrations of altarelli - parisi splitting functions .
though these equations are valid in the perturbative region , one may heuristically assume that the ratio of the total splitting rates obtained from them holds approximately also in the nonperturbative regime .
the parameter @xmath119 can be fixed to the ratio of color factors as @xmath154 by integrations of altarelli - parisi splitting functions . according to the above reasons
, we can conclude that the prediction only from a statistics principle has an accuracy @xmath220 . or , in other words , the details of the dynamics only bring less than 30% effect .
after considering the details of qcd especially the color factors , the sea quark flavor asymmetry in proton is enhanced to @xmath162 which is consistent with present experimental measurements and can be tested by more precise measurements .
the sea quark asymmetries are not sensitively dependent on the details of dynamics .
the sea - quark flavor asymmetry derived only from statistic principle contributes the dominant part of the asymmetry .
it strongly implies that the origin of the sea - quark flavor asymmetry of hadrons is the asymmetry of valence quarks .
we also applied this monte carlo approach of statistical model to predict the sea quark asymmetries in kaons , octet baryons @xmath0 , @xmath1 , and @xmath2 baryons , even in exotic pentaquark states .
all these asymmetries just only depend on the valence quarks number in those hadrons .
the sea - quark asymmetries for different @xmath3 and @xmath4 valence quark numbers are listed in table [ table6 ] .
these values can confirm the mechanism we proposed to explain the sea quark asymmetry in proton .
it can be observed from table [ table6 ] that the sea quark asymmetries are enhanced by the difference of corresponding valence quark numbers and suppressed by the sum of valence quark numbers . when the valence quark numbers @xmath221>[d_v]$ ] , the sea - quarks @xmath222 are easier to annihilated because of the existence of more @xmath3 valence quarks and it leads the sea quark asymmetry .
on the other hand , the larger total number of valence quark @xmath223 $ ] suppresses the relative difference of valence quarks and weakens the sea quark asymmetries even if @xmath224 $ ] remains the same .
these sea quark asymmetries for hadrons , except the proton , are listed purely for theoretical interest , as it is not known presently how to access this information in experiment .
c|ccccc asymmetry values & & @xmath3 valence quark number + @xmath4 valence quark number & 0 & 1&2&3 + 0&@xmath225&@xmath226&@xmath227&@xmath228 + 1&@xmath229&@xmath230&@xmath231&@xmath232 + 2&@xmath233&@xmath234&@xmath235&@xmath236 + 3&@xmath237&@xmath238&@xmath239&@xmath240 + * acknowledgment : * this work of b. z. is supported by the national science foundation of china under grant no . 10705017 and 11075086 .
z. is supported by liaoning education office scientific research project ( 2008288 ) new muon collaboration , p. amaudruz _ et al_. , phys .
* 66 * , 2712 ( 1991 ) ; m. arneodo _
et al_. , phys .
d * 50 * , r1 ( 1994 ) .
na51 collaboration , a. baldit _
et al . _ ,
* b 332 * , 244 ( 1994 ) .
eichten , i. hinchliffe , c. quigg , phys .
d * 45 * , 2269 ( 1992 ) .
cheng , l .- f .
li , phys .
74 * , 2872 ( 1995 ) .
brown , m. rho , phys . rep . * 363 * , 85 ( 2002 ) .
y. ding , r .-
xu , b .- q .
ma , phys .
* b 607 * , 101 ( 2005 ) ; y. ding , b .- q .
ma , phys .
d * 73 * , 054018 ( 2006 ) .
|
we suggested a monte carlo approach to simulate a kinetic equilibrium ensemble , and proved the equivalence to the linear equations method on equilibrium . with the convenience of the numerical method
, we introduced variable splitting rates representing the details of the dynamics as model parameters which were not considered in previous works .
the dependence on model parameters was studied , and it was found that the sea quark flavor asymmetry weakly depends on model parameters .
it reflects the statistics principle contributes the dominant part of the asymmetry and the effect caused by details of the dynamics is small .
we also applied the monte carlo approach of the statistical model to predict the theoretical sea quark asymmetries in kaons , octet baryons @xmath0 , @xmath1 , and @xmath2 baryons , even in exotic pentaquark states .
|
the universality of critical exponents is an important and remarkably elegant property of standard second order transitions , which has been explored in great detail through the renormalization group theory ( rgt ) .
the universality hypothesis states that for all systems within a universality class the critical exponents are rigorously identical and do not depend on the microscopic parameters of the model
. however , universality is not strictly universal ; there are known eccentricmodels which are exceptions and violate the universality rule in the sense that their critical exponents vary continuously as functions of a control variable .
the most famous example is the eight vertex model solved exactly by baxter @xcite ; there are other scattered cases , all in dimension two as far as we are aware . for ising spin glasses ( isgs ) ,
the form of the interaction distribution is a microscopic control parameter .
it has been assumed tacitly or explicitly that the members of the isg family of transitions obey standard universality rules , following the generally accepted statement that empirically , one finds that all systems in nature belong to one of a comparatively small number of universality classes@xcite .
however , we know of no formal proof that universality must hold in isgs ; it was found thirty years ago that the @xmath3-expansion for the critical exponents @xcite in isgs is not predictive since the first few orders have a non - convergent behavior and higher orders are not known .
this can be taken as an indication that a fundamentally different theoretical approach is required for spin glass transitions .
indeed `` classical tools of rg analysis are not suitable for spin glasses '' @xcite .
isg transition simulations are much more demanding numerically than are those on , say , pure ferromagnet transitions with no interaction disorder .
the traditional approach in isgs has been to study the temperature and size dependence of observables in the near - transition region and to estimate the critical temperature and exponents through finite size scaling relations after taking means over large numbers of samples .
finite size corrections to scaling should be allowed for explicitly which can be delicate . from numerical data , claims of universality
have been made repeatedly for isgs @xcite even though the estimates of the critical exponents are very sensitive to the precise value of the critical temperature and have varied over the years ( see ref .
@xcite for a tabulation of historic estimates ) .
we have estimated the critical exponents of the bimodal isg in dimension @xmath0 using complementary strategies .
first we use the standard finite size crossing points of the binder cumulant and other phenomenological couplings to obtain estimates for the critical temperature @xmath4 through finite size scaling @xcite .
we also register the size dependence of the peaks of thermodynamic derivatives which give independent estimates for @xmath5 and for @xmath2 @xcite .
we finally measure the temperature dependence of the thermodynamic limit ( thl ) isg susceptibility @xmath6 and second moment correlation length @xmath7 over a wide temperature range . using the scaling variable and scaling expressions appropriate for isgs as cited below @xcite together with the optimal @xmath5 from the above measurements , we estimate the critical exponents and the confluent corrections to scaling @xcite from data taken over almost the entire paramagnetic temperature range . the numerical data on different isgs in dimension @xmath0 show conclusively that the critical exponents depend on the form of the interaction distribution .
it is relevant that it has been shown experimentally that in heisenberg spin glasses the critical exponents depend on the strength of the dzyaloshinski - moriya interaction @xcite .
the hamiltonian is as usual @xmath8 with the near neighbor symmetric distributions normalized to @xmath9 .
the ising spins live on simple hyper - cubic lattices with periodic boundary conditions .
we have studied bimodal ( @xmath10 ) , gaussian , and laplacian distributions in @xmath0d . here
we will discuss the bimodal isg and will compare with published measurements on two other @xmath0d isgs @xcite .
the simulations were carried out using the exchange monte - carlo method for equilibration , on @xmath11 individual samples at each size .
data were registered after equilibration for the energy @xmath12 , correlation length @xmath13 , for the spin overlap moments @xmath14 , @xmath15 , @xmath16 , @xmath17 , and for the link overlap @xmath18 moments .
in addition the correlations between the energy and certain observables @xmath19 were also registered so that thermodynamic derivatives could be evaluated using the relation @xmath20 where @xmath21 is the energy @xcite .
bootstrap analyses of the errors in the derivatives as well as in the observables themselves were carried out . for the present analysis
we have observed the behavior of various `` phenomenological couplings '' , not only the familiar binder cumulant and correlation length ratio @xmath22 but also other observables showing critical behavior such as the kurtosis of the spin overlap distribution , the kurtosis of the absolute spin overlap distribution , and the variance and kurtosis of the link overlap distribution
. only part of these data are reported here .
near criticality in a ferromagnet the heights of the peaks of the thermodynamic derivative of many observables @xmath23 scale for large @xmath24 as @xcite @xmath25 and the temperature location of the derivative peak @xmath26 scales as @xmath27 the observables used for @xmath28 @xcite can be for instance the binder cumulant @xmath29 , the logarithm of the finite size susceptibility @xmath30 , or the logarithm of the absolute value of the spin overlap @xmath31 .
each of these data sets can give independent estimates of @xmath2 and @xmath4 without any initial knowledge of either parameter .
for the present analysis we note that both the minimum of the inverse derivative @xmath32 and the temperature location difference @xmath33 are proportional to @xmath34 to leading order .
hence @xmath32 plotted against @xmath35 with @xmath24 as an implicit variable must tend linearly to an intercept @xmath36 at @xmath37 for large @xmath24 .
all @xmath32 against @xmath35 plots should extrapolate consistently to the true @xmath4 . turning to spin glasses , for isgs with symmetric interaction distributions and a non - zero @xmath4
a general natural scaling variable is @xmath38 ( @xmath39 is also suitable for the bimodal case ) @xcite . in the isg context
@xmath40 replaces @xmath41 in the thermodynamic derivative scaling rules but otherwise the same methodology can be used as in the ferromagnet .
the thermodynamic derivative analysis , which as far as we are aware has not been used previously in spin glasses , provides reliable and precise estimates for @xmath42 .
these @xmath42 estimates are consistent with those from the traditional crossing point approach ; which method has the least sensitivity to finite size corrections depends on the individual system .
the thl sg susceptibility @xmath43 including the leading nonanalytic confluent correction term @xcite can be written @xmath44 where @xmath1 is the critical exponent and @xmath45 the wegner non - analytic correction exponent , both of which are characteristic of a universality class . following a protocol well - established in ferromagnets @xcite
one can define a temperature dependent effective thl exponent @xmath46 .
@xmath47 tends to the critical @xmath1 as @xmath48 and to @xmath49 as @xmath50 in simple [ hyper]-cubic lattices .
as long as samples of finite size @xmath24 are in the thl regime , @xmath51 , @xmath52 and other observables are independent of @xmath24 . working in the thl has a number of advantages : the temperatures studied are higher than the critical temperature so equilibration is facilitated , the sample to sample variations are automatically much weaker than at criticality , and there are no finite size scaling corrections to take into account although the confluent correction terms must be allowed for . in ferromagnets the thl susceptibility and correlation length data can be fitted accurately over the entire paramagnetic temperature range @xcite by including just one further effective correction term @xmath53 beyond the leading non - analytic term , which bundles together all the higher order correction terms .
we use this approximation also in the isgs .
hence @xmath54 a very effective method for analysing the thl susceptibility data is to plot @xmath55 against @xmath56 . with the two correction terms
the expression used to fit the thl regime data is : @xmath57 the critical intercept @xmath58 occurs when @xmath59 , and the initial slope starting at the intercept is @xmath60 .
the analogous natural scaling expression for the isg second moment correlation length @xmath61 is @xcite @xmath62 with a temperature dependent effective exponent defined as @xmath63 .
the reason for the factor @xmath64 arises from the generic form of the isg @xmath61 high temperature series @xcite .
the @xmath65 limit in isgs in simple hyper - cubic lattices of dimension @xmath66 is @xmath67 where @xmath68 is the kurtosis of the interaction distribution .
the derivative corresponding to eq . takes the form @xmath69 with the same @xmath70 and @xmath45 as for @xmath71 .
the @xmath58 intercept is again @xmath72 , with an initial slope at the intercept equal to @xmath73 .
for even @xmath24 @xmath0d bimodal interaction samples . symbol coding : blue squares @xmath74 , red circles @xmath75 , black triangles @xmath76 , down triangles @xmath77 , olive diamonds @xmath78 , purple left triangles @xmath79 .
the vertical red line corresponds to @xmath80.,width=336 ] [ fig:1 ] d bimodal isg thermodynamic derivative peak height @xmath81_{\max}$ ] as a function of size @xmath24 .
black squares : measured , red circles : fit.,width=336 ] [ fig:2 ] for @xmath0d bimodal interaction samples .
even @xmath24 data only are shown to avoid clutter .
symbol coding as in fig .
[ fig:1 ] .
large navy squares are the minima locations , odd and even @xmath24 .
the full red curve is the the thl calculated directly from htse @xcite .
blue curve : fit eq .
.,width=336 ] [ fig:3 ] for @xmath0d bimodal interaction samples .
even @xmath24 data only are shown to avoid clutter .
symbol coding as in fig .
[ fig:1 ] .
large navy squares are the minima locations , odd and even @xmath24 .
blue curve : fit eq .
.,width=336 ] [ fig:4 ] eq . for @xmath0d bimodal interaction samples .
even @xmath24 data only are shown to avoid clutter .
symbol coding as in fig .
[ fig:1 ] .
red curve : fit through the thl regime data .
red arrow : exact high temperature limit .
the estimate of ref .
@xcite for the critical value @xmath82 is @xmath83.,width=336 ] [ fig:5 ] high precision simulation measurements have been published on the @xmath0d gaussian isg , and on a @xmath0d bimodal isg with diluted interactions ( @xmath84 of the interactions having @xmath85 ) @xcite .
the critical temperature for the 4d gaussian isg was estimated from binder parameter and correlation length ratio measurements to be @xmath86 in full agreement with earlier simulation estimates @xmath87 @xcite and with the htse estimate @xmath88 .
the simulations gave essentially identical exponents for the two systems @xmath89 @xmath90 so indirectly @xmath91 .
present data on the @xmath0d gaussian ( not shown ) using the thermal derivative analysis as above lead to a @xmath70 in full agreement with that of ref @xcite , and @xmath92 , with very weak corrections .
it seems very reasonable to assume that if the present procedure were applied to the diluted bimodal system , it would confirm the conclusions of ref .
@xcite for that system also .
it can be noted that for this system the finite size correction to scaling in the binder cumulant is so small as to be unobservable . for the @xmath0d bimodal isg the htse critical temperature and exponent estimates
are @xcite @xmath93 , @xmath94 , and @xmath95 . from extensive domain wall
free energy measurements to @xmath96 hukushima gave an estimate @xmath97 @xcite .
inspection of the raw data show strong finite size corrections ; extrapolation to larger @xmath24 leads to an infinite size limit definitely greater than @xmath98 . from early simulation measurements up to @xmath96 a critical temperature @xmath99
was estimated @xcite using the binder parameter crossing point criterion .
however , finite size corrections to scaling were not allowed for .
recent simulations up to @xmath100 @xcite show large @xmath24 binder cumulant crossings up to @xmath101 .
our present binder cumulant data up to @xmath79 show very similar results , figure 1 . in this figure
crossing points for the largest @xmath24 can be seen to cluster around @xmath102 , which provides a lower limit on @xmath42 .
in addition , it can be seen that the value of the cumulant at the present high @xmath24 crossing points is @xmath103 , which is a lower limit on the infinite size critical @xmath104 . in comparison
@xmath104 has been estimated at @xmath105 and @xmath106 for the @xmath0d gaussian and diluted bimodal isg respectively @xcite .
the bimodal critical crossing point is thus 25 standard deviations above the diluted bimodal critical value .
as the critical @xmath107 is , for a given geometry , a parameter characteristic of a universality class , the bimodal and diluted bimodal systems are not in the same universality class ; neither are the bimodal and the gaussian isgs .
we will now turn to the inverse susceptibility derivative data , figures 2 , 3 and 4 . in figure 2 the thermodynamic derivative peak height @xmath108_{\max}$ ]
is fitted by @xmath109 with @xmath110 and @xmath45 fixed at @xmath111 , from thl data discussed below .
the value of the estimate for @xmath2 requires no information on the critical temperature @xmath5 .
it is significantly higher than the estimate @xmath112 given by ref .
. in figure 3 the @xmath113_{\min}$ ] points for different @xmath24 have a straight line limit at large @xmath24 which tends to an intercept at @xmath114 .
derivative minima plots of the same type for other observables ( not shown ) confirm this value for @xmath42 .
these estimates are very reliable as they come from parameter free straight line limit fits . the estimate of ref .
@xcite is @xmath115 ; this value is sensitive to the estimate for the correction to scaling exponent .
the curves for individual @xmath24 and the htse curve all lie on a size independent thl envelope curve to the left of the figure .
a satisfactory fit passing through all the thl data and the critical point can be made with a single correction term only .
the optimal fit parameters are @xmath116 , @xmath117 , @xmath118 .
a similar analysis made on the correlation length data , figure 4 , provides a completely consistent estimate for @xmath42 from the linear variation of the minima points .
the thl data are fitted with the parameters @xmath119 and @xmath120 together with the same @xmath121 as for the susceptibility fit .
an estimate of the exponent @xmath122 can be made from a plot of @xmath123 against @xmath124 , figure 5 .
extrapolating the thl regime data to criticality at @xmath125 leads to a direct estimate @xmath126 without needing any assumption concerning @xmath42 or finite size corrections .
the data clearly show that the estimate @xmath127 of ref .
@xcite is low .
the present exponents are reliable and much more accurate than the htse estimates principally because the uncertainty in @xmath42 is reduced by a factor of more than @xmath128 thanks principally to the thermal derivative peak simulation data .
the exponents can be compared to the values found @xcite for the @xmath0d gaussian and diluted bimodal systems which were almost identical to each other : @xmath129 and @xmath130 for the gaussian and @xmath131 , @xmath132 for the diluted bimodal .
the critical exponents of the @xmath0d bimodal isg are quite different from those of the @xmath0d gaussian and diluted bimodal isgs .
simulations on the @xmath0d bimodal isg up to size @xmath79 provide numerical data on finite size scaling observables , on the isg susceptibility and on the correlation length .
the critical temperature @xmath5 derived from the simulation data using a thermodynamic derivative technique @xcite is in full agreement with , but is considerably more precise than , the estimate from htse alone @xcite . because of the analysis techniques used it is also more reliable than previous numerical estimates .
data in the thermodynamic limit regime were analysed to obtain considerably improved critical exponent @xmath1 , @xmath2 and @xmath45 estimates together with the strengths of leading confluent correction terms .
the accurate estimates of @xmath1 and @xmath2 and the critical value of the binder cumulant show that the @xmath0d bimodal isg is in a different universality class from the @xmath0d gaussian or diluted bimodal isgs @xcite .
other results on isgs in dimension @xmath0 @xcite and in dimension @xmath133 @xcite confirm that spin glasses with different interaction distributions have different critical exponents . these results clearly demonstrate that the standard rgt universality rules do not apply in isgs .
we are very grateful to koji hukushima for comments and communication of unpublished data .
we thank amnon aharony for constructive criticism .
the computations were performed on resources provided by the swedish national infrastructure for computing ( snic ) at the high performance computing center north ( hpc2n ) .
20 r. baxter , phys .
* 26 * , 832 ( 1971 ) .
h. e. stanley , rev .
phys . * 71 * , s358 ( 1999 ) .
e. gardner , j. phys .
* 45 * , 1755 ( 1984 ) .
g. parisi , r. petronzio , and f. rosati , eur .
j. b * 21 * , 605 ( 2001 ) .
m. castellana , eur .
lett . * 95 * , 47014 ( 2011 ) .
m. c. angelini , g. parisi , and f. ricci - tersenghi , phys .
b * 87 * , 134201 ( 2013 ) r. n. bhatt and a. p. young , phys . rev .
b * 37 * , 3707 ( 1988 ) . h.
g. katzgraber , m. korner , and a. p. young , phys .
b * 73 * , 224432 ( 2006 ) .
m. hasenbusch , a. pelissetto , and e. vicari , phys . rev .
b * 78 * , 214205 ( 2008 ) .
t. jrg and h. g. katzgraber , phys .
b * 77 * , 214426 ( 2008 ) .
g. parisi , f. ricci - tersenghi , and j. j. ruiz - lorenzo , j. phys .
a * 29 * , 7943 ( 1996 ) .
m. ney - nifle , phys .
b * 57 * , 492 ( 1998 ) .
a. m. ferrenberg and d. p. landau , phys .
b * 41 * , 5081 ( 1991 ) .
m. weigel and w. janke , phys .
rev . lett.*102 * , 100601 ( 2009 ) .
we will use inverse temperatures @xmath134 throughout .
d. daboul , i. chang and a. aharony , eur .
j. b * 41 * , 231 ( 2004 ) .
i. a. campbell , k. hukushima , and h. takayama , phys .
lett . * 97 * , 117202 ( 2006 ) .
i. a. campbell and d. c. m. c. petit , j. phys .
japan , * 79 * , 011006 ( 2010 ) . f. wegner , phys .
b * 5 * , 4529 ( 1972 ) . i. a. campbell and p. butera , phys .
b * 78 * , 024435 ( 2008 ) .
i. a. campbell and p. h. lundow , phys .
b 83 , 014411 ( 2011 ) .
p. h. lundow and i. a. campbell , phys .
b * 83 * , 184408 ( 2011 ) . k. hukushima , phys .
e * 60 * , 3606 ( 1999 ) .
e. marinari and f. zuliani , j. phys .
a : math . gen . * 32 * , 7447 ( 1999 ) .
r. a. baos , l. a. fernandez , v. martin - mayor , and a. p. young , phys .
b * 86 * , 134416 ( 2012 ) .
the gaussian series was calculated to @xmath135 terms but explicit evaluation shows that there is a numerical error in the final term .
r. r. p. singh and s. chakravarty , phys .
lett . * 57 * , 245 ( 1986 ) .
j. kouvel and m. e. fisher , phys .
a * 136 * , 1626 ( 1964 ) .
p. butera and m. comi , phys .
b * 65*,144431 ( 2002 ) . p. h. lundow and i. a. campbell , phys .
e * 87 * , 022102 ( 2013 ) .
h. lundow and i. a. campbell , unpublished .
p. h. lundow and i. a. campbell , arxiv:1302.1100 .
|
extensive simulations are made on the bimodal ising spin glass ( isg ) in dimension four .
the transition temperature is established using a combination of standard finite size scaling and of thermodynamic derivative peak data .
measurements in the thermodynamic limit regime are analysed so as to estimate critical exponents and confluent correction terms .
comparisons with results on other @xmath0d isgs show that the susceptibility and correlation length critical exponents @xmath1 and @xmath2 depend on the form of the interaction distribution . from this observation
it can be deduced that critical exponents are not universal in isgs .
|
recently a new class of material , called topological insulator , has been paid much attention by condensed matter physicists @xcite .
topological insulator ( ti ) shows the conduction of electrons on the surface of 3d materials otherwise behaves as an insulator .
it is due to the time - reversal symmetry possessed by materials like bi@xmath1se@xmath2 , sb@xmath3te@xmath4 and bi@xmath3te@xmath4@xcite .
the conducting surface states of these material show a single dirac cone , in which spin is always locked perpendicular to it s momentum .
the angle resolved photoemission spectroscopy @xcite or scanning tunneling microscopy @xcite has been used to realize the single dirac cone in tis . in two - dimensional electron systems , under the presence of a perpendicular magnetic field , electron conducts along the boundary due to the circular orbits bouncing off the edges , leading to skipping orbits .
however , in 3d materials , even in absence of magnetic field electron conduction takes place on the surface . here , strong rashba spin - orbit coupling ( rsoc ) plays the role of magnetic field .
the rsoc originates from the lack of structural inversion symmetry of the sample @xcite .
+ though there have been several experimental works on the surface states of tis , one of the major obstacle in studying the transport properties of the surface is the unavoidable contribution of the bulk .
one of the best method to minimize this problem is to grow tis sample in the form of ultra - thin films , in which bulk contribution becomes relatively very small in comparison to the surface contribution@xcite .
the transition from 3d to 2d tis lead to several effects which have been studied for different thickness@xcite .
the ultra - thin ti not only reduces the bulk contribution , but also possesses some new phenomenon like possible excitonic superfluidity @xcite , unique magneto - optical response@xcite and better thermoelectric performances@xcite .
moreover , the small thickness leads to the overlap of the wave functions between top and bottom surfaces which introduces a new degree of freedom hybridization@xcite .
however , it happens to a certain thickness of five to ten quintuple layers@xcite i.e ; of the order of 10 nm .
the oscillating exponential decay of hybridization induced band gap with reducing thickness in bi@xmath3te@xmath4 has been also reported theoretically@xcite .
the formation of landau levels have been confirmed by several experiments@xcite in thin tis .
moreover , several theoretical study on low - temperature transport properties in a series of works@xcite have been reported .
+ thermoelectric properties of materials @xcite have always been interesting topic for providing an additional way in exploring the details of an electronic system .
when a temperature gradient is applied across the two ends of the electronic system , the migration of electrons from hotter to cooler side leads to the developement of a voltage gradient across these ends .
this voltage difference per unit temperature gradient is known as longitudinal thermopower .
in addition to this temperature gradient if a perpendicular magnetic field is applied to the system , due to lorentz force , a transverse electric field is also established and gives transverse thermopower . in conventional 2d electronic system ,
landau levels induced quantum oscillation ( shubnikov - de hass ) in thermoelectric coefficients has been reported theoretically as well as experimentally in a series of works@xcite . in 3d tis ,
improvement of thermoelectric performance without magnetic field have been predicted theoretically in a series of paper@xcite . in the
newly emerged relativistic - like 2d electron system - graphene , thermoelectric effects have been also studied@xcite . in this paper
, we study the effect of hybridization on the thermopower and the thermal conductivity of the ultra - thin tis in absence / presence of magnetic field .
we find thermopower increases and thermal conductivity decreases with increase of the hybridization energy when magnetic field is absent . in presence of perpendicular magnetic field ,
thermoelectric coefficients oscillate with inverse magnetic field .
the frequency of the quantum oscillations is strongly modified by the zeeman energy , and phase factor is determined by the product of the lande @xmath0-factor and the hybridization energy .
the analytical expressions of the thermoelectric coefficients are also obtained .
it is also shown that these transport coefficients oscillate with frequency depend on hybridization energy and lande @xmath0-factor .
this paper has following structure .
section ii briefly discusses energy spectrum and the density of states of the ultra - thin ti in absence and presence of magnetic field . in section iii , we have studied how the hybridization affects the thermoelectric coefficients for zero magnetic field . in section iv , a complete analysis of thermoelectric coefficients in present of magnetic field is provided with numerical and analytical results .
we provide a summary and conclusion of our work in section v.
let us consider a surface of an ultra - thin ti in @xmath5-plane with @xmath6 dimension , and carriers are dirac fermions occupying the top and bottom surfaces of the ti .
the quantum tunneling between top and bottom surfaces gives rise to the hybridization and consequently the hamiltonian can be written as the symmetric and anti - symmetric combination of both surface states as @xcite @xmath7,\ ] ] with @xmath8 .
here @xmath9 is the two - dimensional momentum operator , @xmath10 is the fermi velocity of the dirac fermion , @xmath11 are the pauli spin matrices and @xmath12 is the hybridization matrix element between the states of the top and bottom surfaces of the ti
. typical value of @xmath13 varies from @xmath14 mev depending on the thickness of the 3d ti @xcite . because of the block - diagonal nature
, the above hamiltonian can be written as @xmath15 where @xmath16 denotes symmetric and anti - symmetric surface states , respectively .
the energy spectrum of the dirac electron is given by @xmath17 here @xmath18 stands for electron and hole bands .
the density of states is given by @xmath19 in presence of magnetic field perpendicular to the surface , the hamiltonian for dirac electron with hybridization is @xmath20 where @xmath21 is the two - dimensional canonical momentum operator . using landau gauge @xmath22
, exact landau levels can be obtained very easily @xcite . for @xmath23 , there is only one energy level which is given as @xmath24 . when integer @xmath25 , there are two energy bands denoted by @xmath26 corresponding to the electron and @xmath27 corresponding to the hole with energy @xmath28 where @xmath29 is the cyclotron frequency with @xmath30 is the magnetic length , @xmath31 with @xmath0 is the lande @xmath0-factor .
the corresponding eigenstates for symmetric surface state are @xmath32 @xmath33 where @xmath34 is the normalized harmonic oscillator wave function , @xmath35 , @xmath36 and @xmath37 with @xmath38 $ ] .
the anti - symmetric surface state can be obtained by exchanging @xmath39 and @xmath40 .
we have derived approximate analytical form of density of states for @xmath41 , by using the green s function technique which is given as ( see appendix a ) @xmath42,\end{aligned}\ ] ] where @xmath43 and @xmath44 is the impurity induced landau level broadening .
in this section , we shall calculate thermoelectric coefficients of an ultra - thin ti in zero and non - zero magnetic fields . in this sub - section , the effect of hybridization on thermopower and thermal conductivity is presented .
we follow most conventional approach at low temperature regime .
the electrical current density @xmath45 and the thermal current density @xmath46 for dirac electrons can be expressed under linear response regime as @xmath47 and @xmath48 where @xmath49 is the electric field , @xmath50 is the temperature gradient and @xmath51 ( @xmath52 ) are the phenomenological transport coefficients .
the above equations describe the response of electronic system under the combined effects of thermal and potential gradient .
moreover , @xmath53 can be expressed in terms of an integral @xmath54 : @xmath55 , @xmath56 with @xmath57 ( e-\eta)^{(r ) } \sigma(e),\ ] ] where @xmath58 and @xmath59 $ ] is the fermi - dirac distribution function with @xmath60 is the chemical potential and @xmath61 . here , @xmath62 is the energy - dependent electrical conductivity . when circuit is open i.e ; for @xmath63 , thermopower can be defined as @xmath64 . by using sommerfeld expansion at low temperature regime , diffusive thermopower @xmath65 and thermal conductivity @xmath66
can be obtained from mott s relation and the wiedemann - franz law as @xmath67_{_{e = e_f}}\ ] ] and @xmath68 here , @xmath69 w@xmath70k@xmath71 is the lorentz number and @xmath72 is the electrical conductivity at the fermi energy .
classical boltzmann transport equation can be used to calculate zero magnetic field electrical conductivity , which is given by@xcite @xmath73v^{i}(k)v^{j}(k),\ ] ] where @xmath74 . for isotropic system @xmath75 . in our case , @xmath76.\ ] ] using these in eq.[[boltz ] ] , the energy dependent conductivity becomes as @xmath77.\ ] ] assuming the energy dependent scattering time to be @xmath78 , where @xmath79 is a constant depending on the scattering mechanism , @xmath80 is the fermi energy with @xmath81 . here ,
fermi vector @xmath82 .
substituting eq .
( [ con ] ) into eq .
( [ tp ] ) , the diffusion thermopower is obtained as @xmath83 /\sqrt{1+\big(\frac{\delta_h}{e_{f0}}\big)^2}.\ ] ] we plot thermopower versus hybridization for different carrier density in the upper panel of fig .
it shows that thermopower increases with increasing hybridization for a particular carrier density .
but for higher carrier density , this rate of enhancement with hybridization becomes very slow .
thermal conductivity can be directly obtained from wiedemann - franz law given in eq .
( [ ther ] ) where the electrical conductivity @xmath84 is given as @xmath85 here , @xmath86 is the drude conductivity without the hybridization constant for the dirac system .
thermal conductivity is plotted in the the lower panel of fig .
we note that the thermal conductivity is diminished with increasing hybridization .
however , unlike the case of thermopower , here thermal conductivity increases with carrier density . and for various carrier density.,width=340 ] in presence of magnetic field
, the classical approach can not explain the phenomenon depend on energy quantization . in this sub - section
we follow quantum mechanical approach , based on linear response theory , to study thermal transport coefficients .
thermoelectric coefficients for two - dimensional electron system in presence of magnetic field were derived by modifying the kubo formula in ref .
these phenomenological transport coefficients are @xmath87 @xmath88_{\mu\nu}\ ] ] @xmath89,\ ] ] where @xmath90 ( e-\eta)^{r } \sigma_{\mu\nu}(e).\ ] ] here , @xmath91 .
also , @xmath92 , @xmath93 and @xmath94 are the zero - temperature energy - dependent conductivity , thermopower and thermal conductivity tensors , respectively . generally , diffusive and collisional mechanism play major role in electron conduction .
the quantized energy spectrum of electrons results itself through shubnikov - de hass oscillation by collisional mechanism . in our case
, electron transport is mainly due to the collisional instead of diffusive .
the zero drift velocity of electron do not allow to have diffusive contribution .
in presence of temperature gradient , thermal transport coefficients can be expressed as @xmath95 and @xmath96 . in ref .
@xcite , the exact form of the finite temperature collisional conductivity has been calculated for the screened impurity potential @xmath97 under the limit of small @xmath98 with @xmath99 and @xmath100 being the inverse screening length and dielectric constant of the material , respectively . in this limit
, one can use @xmath101 with @xmath102 is the relaxation time , @xmath103 is the strength of the screened impurity potential and @xmath104 is the two - dimensional impurity density .
the exact form of the finite temperature conductivity@xcite can be reduced to the zero - temperature energy - dependent electrical conductivity as @xmath105 where @xmath106 $ ] .
here we have used @xmath107 $ ] . using eq .
( [ lmunu ] ) , the finite temperature diagonal ( @xmath108 ) and off - diagonal coefficients ( @xmath109 ) can be written as @xmath110_{e = e^{\tau_z}_n}\ ] ] and @xmath111 here , @xmath112 .
in our numerical calculations , the following parameters are used : carrier concentration @xmath113 m@xmath114 , @xmath115 , @xmath116 m s@xmath117 and @xmath118 k. these numerical parameters are consistent with ref.@xcite . in fig . [ 2 ] , the off - diagonal thermopower , @xmath119 , as a function of the inverse magnetic field for different values of the hybridization constant is shown .
similarly , the thermal conductivity , @xmath120 , versus inverse magnetic field for different values of the hybridization constant is shown in fig . [ 3 ] . careful observation of these two figures clearly show that both @xmath119 and @xmath120 oscillate with the same frequency which does not depend on the hybridization energy .
the hybridization energy only influences the phase of oscillations . versus inverse magnetic field for different values of the hybridization constant.,width=370 ] versus inverse magnetic field.,width=355,height=226 ] to determine the frequency and the phase of the quantum oscillations in the thermoelectric coefficients , we shall derive analytical expressions of these coefficients .
the components of the thermopower are given by @xmath121\ ] ] and @xmath122.\ ] ] here , @xmath123 .
the dominating term in the above two equations is the last term .
the analytical form of @xmath124 and @xmath119 can be obtained directly by deriving analytical form of the phenomenological transport coefficients .
the analytical form of density of states given in eq .
( [ dos_b ] ) allows us to obtain asymptotic expressions of @xmath125 and @xmath126 .
this is done by replacing the summation over discrete quantum numbers @xmath39 by the integration i.e ; @xmath127 , then we get @xmath128 and @xmath129,\ ] ] where @xmath130 , @xmath131 $ ] , the impurity induced damping factor is @xmath132 and the temperature dependent damping factor is the derivative of the function @xmath133 with @xmath134 . here , @xmath135 with @xmath136 is the critical temperature which depends on strength of hybridization through fermi energy .
note that @xmath137 is the temperature dependent damping factor for the electrical conductivity tensor .
the off - diagonal thermopower @xmath138 and the diagonal thermal conductivity @xmath124 are obtained as given by @xmath139^{1/2}\omega_{d } g^{\prime}(x)\nonumber\\ & \times & \sum_{\tau_z}\frac{u_{\tau_z}}{\sin^2(\bar{\theta}_{\tau_z } ) } f_{\tau_z}\sin\big(2\pi f / b - \tau_z \phi \big)\end{aligned}\ ] ] and @xmath140,\end{aligned}\ ] ] where the frequency @xmath141 is given as @xmath142 and the phase factor @xmath143 .
therefore , the thermopower and the thermal conductivity oscillate with the same frequency @xmath141 which is independent of @xmath12 , which can be shown from numerical result also .
the oscillation frequency is strongly reduced by the zeeman energy @xmath144 . on the other hand ,
the phase factor @xmath145 is related to the product of the lande @xmath0-factor and @xmath12 and it vanishes if either of them is zero .
although the frequency and the phase of @xmath146 and @xmath126 are the same but the damping factor and amplitude are different .
now we compare the numerical and analytical results of @xmath147 and @xmath148 in fig . [
4 ] . for better visualization ,
we have taken weak landau level broadening @xmath149 mev for @xmath119 and @xmath124 .
the analytical results , in particular the frequency @xmath141 , match very well with the numerical results .
we must mention here that for different values of @xmath44 analytical results may differ with numerical in the amplitude but the frequency and phase are always in good agreement .
solid and dashed lines stand for numerical and analytical results , respectively.,width=370 ] it is interesting to see from the analytical expressions of the thermopower and the thermal conductivity that these transport coefficients possess weak periodic oscillation with the hybridization energy for a given magnetic field .
these oscillation is shown in figure 5 .. the frequency and phase factor of these oscillations for a fixed @xmath150 are @xmath151 and @xmath152 , respectively . . here
, @xmath153 t is taken.,width=340 ] in presence of the magnetic field , these approximate analytical expression of the thermopower and thermal conductivity can be also used for monolayer graphene by putting @xmath154 .
there are several experimental results @xcite on thermoelectric properties of a graphene monolayer but analytical expressions are not available in the literature .
the approximate analytical expression of the thermopower and thermal conductivity can also be used for a monolayer graphene by setting @xmath154 .
we have presented a theoretical study on the thermoelectric coefficients of ultra - thin topological insulators in presence / absence of the magnetic field . in absence of the magnetic field ,
the thermopower and the thermal conductivity are modified due to the hybridization between top and bottom surface states .
the thermopower is enhanced and the thermal conductivity is diminished due to the hybridization .
the quantum oscillations in the thermopower and the thermal conductivity for different values of the hybridization constant are also studied numerically .
in addition to the numerical results , we obtained the analytical expressions the thermopower @xmath155 and the thermal conductivity ( @xmath124 ) .
the analytical results match very well with the numerical results .
we have also provided analytical expressions of the oscillation frequency and phase .
the oscillation frequency is the same for both the thermopower and thermal conductivity .
it is independent of the hybridization constant but strongly suppressed by the zeeman energy . on the other hand ,
the hybridization constant plays a very significant role in the phase as well as in the amplitude of the oscillations . from the analytical results , critical temperature @xmath156
is found to be reduced with increasing hybridization constant .
thermoelectric coefficients also show a very low - frequency oscillation with the hybridization constant for a given magnetic field
. moreover , our analytical expressions of the thermopower and thermal conductivity are also applicable for a graphene monolayer by setting @xmath157 .
this work is financially supported by the csir , govt.of india under the grant csir - srf-09/092(0687 ) 2009/emr f - o746 .
derivation of asymptotic analytical expression of density of states of a two dimensional electronic system in presence of impurity can be done by calculating self - energy @xcite which is given as @xmath158 imaginary part of the self - energy is related to density of states as @xmath159 $ ] . by writing the self - energy as the sum of real and imaginary part
, it can be further simplified as @xmath163\ ] ] here , @xmath164\ ] ] and @xmath165 . the imaginary part is @xmath166 .
now , this can be re - written by using the following standard relation as @xmath167 here , the most dominant term is for @xmath168 only .
we can write @xmath169.\ ] ] in the limit of @xmath170 , after first iteration , we have @xmath171 . substituting it in the earlier expression , we get @xmath172.\end{aligned}\ ] ] here , @xmath43 p. cheng , c. song , t. zhang , y. zhang , y. wang , j. f. jia , j. wang , y. wang , b. f. zhu , x. chen , x. ma , k. he , l. wang , x. dai , z. fang , x. xie , x. l. qi , c. x. liu , s. c. zhang , and q. k. xue , phys . rev .
lett * 105 * , 076801 ( 2010 )
|
thermoelectric coefficients of an ultra - thin topological insulator are presented here .
the hybridization between top and bottom surface states of a topological insulator plays a significant role . in absence of magnetic field ,
thermopower increases and thermal conductivity decreases with increase of the hybridization energy . in presence of magnetic field perpendicular to the ultra - thin topological insulator , thermoelectric coefficients exhibit quantum oscillations with inverse magnetic field , whose frequency is strongly modified by the zeeman energy and phase factor is governed by the product of the lande @xmath0-factor and the hybridization energy .
in addition to the numerical results , the low - temperature approximate analytical results of the thermoelectric coefficients are also provided .
it is also observed that for a given magnetic field these transport coefficients oscillate with hybridization energy , whose frequency depends on the lande @xmath0-factor .
|
massive galaxies and galaxy clusters at high redshift provide strong constraints on cosmological models and galaxy formation scenarios . the lyman break galaxies ( lbgs ) are a widespread population of high redshift galaxies selected by their rest - frame fuv colors @xcite .
lbgs at @xmath14 tend to reside in massive halos @xmath15 @xcite and are therefore believed to evolve into present - day massive elliptical galaxies .
although @xmath16 color - selected @xmath17 lbgs have spectroscopic confirmations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , only two systems , selected by their radio emission , have been spectroscopically identified as potential brightest cluster galaxies @xcite .
we report the discovery of a bright , m@xmath0 , rest - frame fuv color - selected lbg ( hereafter lbg-2377 ) at @xmath1 .
ground - based imaging and spectroscopy finds no evidence for agn activity or gravitational lensing making this the most luminous @xmath18 lbg discovered to date .
the deep images and high signal - to - noise ratio ( snr ) spectroscopy show evidence for multiple massive components indicative of a system in the late stage of merging .
we show that the component velocity dispersion and integrated fuv luminosity suggest that lbg-2377 resides in a very massive halo ( @xmath19 ) and is likely to evolve into a present - day massive brightest cluster galaxy ( [ sim ] ) .
therefore , lbg-2377 provides a rare opportunity to study the dynamics and properties of a massive lbg halo in the process of formation .
all @xmath18 lbgs with @xmath20 discovered to date are magnified by gravitational lensing . because all of these show ly-@xmath21 in absorption , they are representative of the spectral properties observed in @xmath225 - 50% of the lbg population @xcite .
lbg-2377 displays strong ly-@xmath21 in emission in two of its components and , because of its brightness , is the only system of its kind .
the luminosity of lbg-2377 enables high snr study of its individual components and thereby a means to constrain the observed range of lbg spectral properties .
we briefly describe the imaging and spectroscopic observations of lbg-2377 in [ obs ] , highlight the initial analysis of the data in [ analysis ] , and present a discussion in [ disc ] .
all magnitudes are in the ab @xcite magnitude system unless otherwise noted and we assume an @xmath22 , @xmath23 cosmology and use comoving units throughout .
the deep images of the qso pc1643 + 4631a field were obtained on 2001 april 18 using the low - resolution imaging spectrometer @xcite mounted on the keck i telescope as part of a multi - field survey for lbgs associated with damped ly-@xmath21 systems at @xmath13 @xcite .
specifically , the field containing lbg-2377 was imaged in the @xmath24bvr@xmath25 is the h. spinrad high - throughout night - sky r filter publicly available at the w. m. keck observatory .
the transformation from spinrad night - sky r@xmath25 into johnson vr is ( r@xmath25 - r ) = -0.004 - 0.072(v - r ) + 0.073(v - r)@xmath26 @xcite .
all r@xmath25 magnitudes presented here have been transformed into the johnson r bandpass.]i filters with seeing of @xmath27 arcsec and a depth of @xmath28 and bvri @xmath29 ( @xmath30 ) .
we discovered lbg-2377 as an r @xmath31 @xmath14 lbg candidate at r.a .
16:44:48.3 , dec.+46:27:08.2 ( j2000.0 ) toward one edge of the imaged field .
the isophotal magnitudes as determined using @xmath32 @xcite source extraction code are : @xmath33 , b @xmath34 , v @xmath35 , r @xmath36 , and i @xmath37 .
lbgs at @xmath13 have a half - light radius of @xmath38 ( e.g. , * ? ? ?
as illustrated in the contour plot shown in figure [ fig : contour ] , lbg-2377 spans @xmath39 and has a morphology that can be interpreted as a single extended system exhibiting multiple vigorous star forming regions or multiple separate near - point sources .
the ratio of the peaks of the three strongest identified components is approximately 10:5:1 with separations of @xmath2@xmath40 ( @xmath41 kpc , comoving ) .
inspection of the deep images shows that there are no significant lensing - source candidates within @xmath42 . a potential candidate @xmath43 from lbg-2377 ( r.a .
16:44:46.5 , dec + 46:27:04.1 ) does have an effective redshift for lensing ( @xmath44 ) , but would need a halo mass of @xmath45 to provide a significant ( @xmath46 ) magnification boost @xcite .
moreover , such a geometry would result in a large distortion of the lbg-2377 image .
we see no evidence for this in the deep ground - based images .
lbg-2377 was confirmed to be a @xmath1 galaxy from low - resolution discovery multi - object spectroscopy on 2004 february 18 @xcite .
the data were taken using the 300 line mm@xmath7 grism on lris with a spectral resolution of @xmath47 fwhm .
we acquired follow - up higher resolution , higher snr longslit spectroscopic observations with lris on 21 may 2007 using the 600 line mm@xmath7 grism blazed at 4000 with the blue arm and the 1200 line mm@xmath7 blazed at 7500 with the red arm .
we used five 1800s exposures in our final combined spectrum that resulted in a snr of @xmath48 and a resolution of @xmath49(figure [ fig:600 ] ) .
details of the observations will appear in a forthcoming paper .
the spectrum of lbg-2377 is dominated by o and b star continua and displays strong ly-@xmath21 emission and little reddening .
these features are consistent with the general properties of the @xmath14 lbg population that display ly-@xmath21 in emission @xcite .
although the ly-@xmath21 feature appears as a single emission line in the low - resolution discovery spectrum , the higher - resolution high snr spectroscopy reveals two strong peaks with two possible weak peaks more evident in the individual spectra .
the rest - frame fuv shows a complex series of interstellar atomic lines exhibiting gas absorption over velocities @xmath50 km s@xmath7 and evidence for @xmath5 components . over 20 fuv ism absorption lines
were used to identify and study individual components .
we identified @xmath515 photospheric lines ( to varying confidence levels ) of each component to estimate their systemic redshifts .
these lines include : siiii @xmath52 , 1296.73 , 1417.24 , ciii @xmath53 , 1427.85 cii @xmath541323.93 , niii @xmath55 , oiv @xmath56 , and sv @xmath57 .
the strongest components have redshifts @xmath58 , and 3.0244 .
we find a 1-d velocity dispersion of @xmath59 km s@xmath7 for these three components and @xmath60 km s@xmath7 for the five components .
the details of the identified components are listed in table [ comp ] .
we note that the ly-@xmath21 emission and ism absorption lines of the lbg population show observed velocity offsets of @xmath61 and @xmath62 km sec@xmath7 from their systemic velocities , respectively @xcite .
the ly-@xmath21 and ism offsets in the identified components are consistent with that of the lbg population .
the accepted interpretation is that these offsets are caused by the presence of an expanding galactic - scale shell of gas and dust driven by supernovae and stellar winds .
ly-@xmath21 photons are largely absorbed in the approaching shell but are resonantly scattered off the receding portion and survive . in this scenario , the two prominent ly-@xmath21 peaks observed in the spectrum of lbg-2377 would trace two distinct expanding galactic - scale shells with a systemic velocity difference of @xmath63 km sec@xmath7 .
our fit to the components in table [ comp ] is shown in figure [ fig : spectra ] .
although the spectra of the components are superposed , the ism absorption - line strengths and behavior of each component appear internally consistent with being separate systems .
the high luminosity and complex structure of lbg-2377 suggest that this system may be a brightest cluster galaxy in formation . here
we investigate this possibility in the context of @xmath8cdm .
we estimate the dark matter halo mass of lbg-2377 using both its measured kinematics and luminosity .
encouragingly , both estimates suggest a group - scale mass @xmath64 .
if we assume an nfw halo , our measured @xmath6 km s@xmath7 1-d velocity dispersion corresponds to a halo maximum circular velocity of @xmath65 , or @xmath66 km s@xmath7 @xcite . at @xmath67
this implies @xmath68 ( e.g. * ? ? ?
* ) . alternatively ,
if we assume a simple monotonic relationship between galaxy luminosity and dark matter halo @xmath69 ( e.g. * ? ? ?
* ) and use the @xmath13 lbg luminosity function @xcite , the number density of objects similar to lbg-2377 ( m@xmath0 ) is @xmath70 mpc@xmath71 .
this number density corresponds to halos with @xmath72 km s@xmath7 at @xmath10 @xcite .
the agreement between these two estimates suggests that lbg-2377 does indeed sit within a massive @xmath66 km s@xmath7 halo .
halos that are this massive at @xmath10 almost certainly contain substructure .
we illustrate this explicitly by examining the high - resolution 80 @xmath73 mpc@xmath74 @xmath8cdm n - body simulation described in @xcite .
there are three halos in this box with @xmath75 km s@xmath7 at @xmath10 ( specifically , @xmath76 and @xmath77 km sec@xmath7 ) .
these objects have between five and eight @xmath78 sub - halos within their ( physical ) virial radii @xmath79 kpc @xmath80 .
note that this substructure mass is typical of that expected for lbg hosts at @xmath13 ( @xmath8110@xmath82 * ? ? ?
if we look at simulated halos that are significantly smaller than our estimate , we find that only two out of 16 objects with @xmath83 km sec@xmath7 in the simulation have no sub - halos with @xmath84 .
we conclude that it would be very unlikely for an object like lbg-2377 to exist _ without _ multiple components . by following the merger trees of our three @xmath85 km sec@xmath7 halos
, we find that they evolve from @xmath86 halos at @xmath10 to cluster - mass systems with @xmath87 at @xmath12 .
these are among the @xmath88 most massive halos in the box at @xmath12 and contain @xmath89 sub - halos larger than @xmath90 .
this analysis suggests that lbg-2377 will evolve into the dominant galaxy within a large group or cluster by the present day .
the double peak nature of the ly-@xmath21 feature in lbg-2377 was only detected with higher resolution and snr .
however , multiple ly-@xmath21 peaks are not unique to lbg-2377 . in the lbg spectroscopic survey of @xcite
, we find @xmath91% of the low - resolution spectra to display double ly-@xmath21 emission peaks and resolved double peaks in their deep image profiles .
we can only assume that this fraction increases with higher resolution or snr .
one example of this effect may be found in the sample of @xcite .
of the nine deep high snr lbg spectra with ly-@xmath21 in emission , three have double peaks .
two of the remaining six have redshifts within @xmath92 and a separation of @xmath93 and are consistent with a double - peak system at a larger angular separation .
interestingly , upon inspection of the average internal velocity differences between the ly-@xmath21 emission and ism absorption features ( @xmath94 ) , the single peak systems are reported to have @xmath95550 km sec@xmath7 , whereas the double peak systems have @xmath96 km sec@xmath7 .
km sec@xmath7 . if the absorption features of the lower redshift systems are weak or narrow in this interpretation , adopting a typical @xmath97 for the that system would result in @xmath98 km sec@xmath7 for the assumed two components . ]
one plausible interpretation of this discrepancy is the result of the velocity separations , and consequent spectral blending , of two individual interacting systems . confirming separate components in lbgs showing only ly-@xmath21 in absorption is more difficult .
however , there are indications of separate components in the intermediate - resolution , high snr spectrum of ms1512-cb58 @xcite .
the prediction implied is that with higher snr and/or resolution , a significant fraction , possibly as high as @xmath99% , of lbgs with ly-@xmath21 in emission have multiple massive components and may indeed be separate interacting or merging systems .
this result would be consistent with the fraction of expected major mergers at @xmath14 from numerical simulations ) , and with galaxy formation theories that suggest mergers are an important contributor to the star formation activity in lbgs @xcite .
what is the nature of lbg-2377 ?
possible interpretations are that lbg-2377 represents either ( 1 ) a single extended lbg with @xmath100 exhibiting multiple regions of vigorous star formation , ( 2 ) a single , extended lbg with mass more typical of lbgs ( @xmath101 ) that appears over - luminous due to exceptionally high bursts of star formation , ( 3 ) multiple unbound @xmath102 lbgs in near proximity , or ( 4 ) multiple lbgs in a @xmath103 halo observed in a late stage of merging .
we discuss each scenario below and show that the evidence gathered to date favors interpretation 4 .
if the component stellar photospheric lines are misidentified and the complex nature of the ism absorption and ly-@xmath21 emission lines are ascribed to violent mixing of the ism and discrete star forming regions separated by @xmath104 kpc , interpretation 1 may hold .
however , a single @xmath105 halo with no massive sub - halos is in conflict with simulation results . in this interpretation
, lbg-2377 is indeed a massive and extended system that appears to be experiencing a monolithic burst of star formation . under interpretation 2
, lbg-2377 would be in less conflict with simulations yet the above assumptions for stellar and ism lines must hold . in this case
, lbg-2377 is a single system undergoing an unusually strong star formation episode at a level that has not been witnessed in any other lbg .
in addition , this exceptional star formation must be occurring in multiple regions separated by @xmath104 kpc .
if the velocity offsets are interpreted as redshifts , as in interpretation 3 , the systems would have separations of @xmath106 mpc comoving .
an event of three or more unbound systems with @xmath101 , @xmath107 angular separations , and an lbg density of @xmath108 mpc@xmath109 has a vanishingly low probability , even if the systems were given reasonable peculiar velocities .
the most likely scenario is interpretation 4 because it is consistent with the extended nature and multiple peaks seen in the images , the superposed spectra of multiple identified components , the observed velocity dispersion , and the expectations from the simulation .
in addition , this interpretation provides a natural driver for the observed star formation and bright integrated luminosity .
furthermore , the double ly-@xmath21 emission peak of lbg-2377 is also seen in other lbgs which exhibit two distinct components in their deep images and ly-@xmath21 emission - line velocity differences consistent with close galaxy pairs .
lbg-2377 is a fortuitous discovery that enables detailed study of the individual components of a candidate massive galaxy in the late stage of merging .
study of the relationships between the spectral and spatial properties of this single system will constrain the formation processes of the lbg population and lend insight into the fraction of lbgs detected as a result of merger induced star formation .
future high - resolution ground - based ifu data on existing instruments , such as osiris on keck , have the capability to resolve the velocity structure of lbg-2377 and hence its true nature . j.c . is supported in part by a gary mccue postdoctoral fellowship through the center for cosmology at the university of california , irvine .
e.j.b . and j.s.b .
are supported by the center for cosmology at the university of california , irvine and j.s.b . and k.s .
are supported by nsf grant ast-0507816 .
the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community .
we are most fortunate to have the opportunity to conduct observations from this mountain .
adelberger , k. l. , steidel , c. c. , shapley , a. e. , & pettini , m. 2003 , apj , 584 , 45 adelberger , k. l. , steidel , c. c. , pettini , m. , shapley , a. e. , reddy , n. a. , & erb , d. k. 2005 , apj , 619 , 697 berrier , j. , bullock , j. s. , zentner , a. r. , guenther , h. , barton , e. j. , kravtsov , a. v. , & wechsler , r. h. 2006 , apj , 652 , 56 bertin , e. & arnouts , s. 1996 , a&as , 117 , 393 bullock , j. s. , kolatt , t. s. , sigad , y. , somerville , r. s. , kravtsov , a. v. , klypin , a. a. , primack , j. r. , & dekel , a. 2001 , , 321 , 559 cooke , j. , wolfe , a. m. , gawiser , e. , & prochaska , j. x. 2005 , apj , 621 , 596 cooke , j. , barton , e. , barbour , e. gawiser , e. , & wolfe , a. 2008 in preparation fukugita , m. , ichikawa , t. , gunn , j. e. , doi , m. , shimasaku , k. , & schneider , d. p. 1996 , aj , 111 , 1748 gardner , j. p. , brown , t. m. , & ferguson , h. c. 2000 , apj , 542l , 79 klypin , a. ; gottlber , s. , kravtsov , a. v. , & khokhlov , a. m. 1999 , apj , 516 , 530 kolatt , t. s. , bullock , j. s. , somerville , r. s. , sigad , y. , jonsson , p. , kravtsov , a. v. , klypin , a. a. , primack , j. r. , faber , s. m. , & dekel , a. 1999 , apj , 523 , 109 law , d. r. , steidel , c. c. , erb , d. k. , pettini , m. , reddy , n. a. , shapley , a. e. , adelberger , k. l. , & simenc , d. j. 2007 , apj , 656 , 1 miley , george k. , overzier , roderik a. , zirm , andrew w. , ford , holland c. , kurk , jaron , pentericci , laura , blakeslee , john p. , franx , marijn , illingworth , garth d. , postman , marc , rosati , piero , rttgering , huub j. a. , venemans , bram p. , & helder , eveline 2006 , apj , 650 , 29 minor , q. 2007 , private communication .
oke , j. b. , cohen , j. g. , carr , m. , cromer , j. , dingizian , a. , harris , f. h. , labrecque , s. , lucinio , r. , schaal , w. , epps , h. , & miller , j. 1995 , pasp , 107 , 375 pettini , m. , steidel , c. c. , adelberger , k. l. , dickinson , m. , & giavalisco , m. 2000 , apj , 528 , 96 pettini , m. , rix , s. a. , steidel , c. c. , adelberger , k. l. , hunt , m. p. , & shapley , a. e. 2002 , apj , 569 , 742 ryan , r. e. , cohen , s. h. , windhorst , r. a. , & silk , j. 2007 , arxiv:0712.0416 sawicki , m. & thompson , d. 2006 , apj , 642 , 653 shapley , a. e. , steidel , c. c. , adelberger , k. l. , & pettini , m. 2003 , apj , 588 , 65 shapley , a. e. , steidel , c. c. , pettini , m. , adelberger , k. l. , & erb , d. k. 2006 , apj , 651 , 688 somerville , r. s. , primack , j. r. , & faber , s. m. 2001 mnras , 320 , 504 steidel , c. c. , giavalisco , m. , pettini , m. , dickinson , m. , & adelberger , k. l. 1996 , apj , 462 , 17 steidel , c. c. , adelberger , k. l. , giavalisco , m. , dickinson , m. , pettini , m. , & kellogg , m. 1998 , , 492 , 428 steidel , c. c , adelberger , k. l. , giavalisco , m. , dickinson , m. , & pettini , m. 1999 , apj , 519 , 1 steidel , c. c. , adelberger , k. l. , shapley , a. e. , pettini , m. , dickinson , m. , & giavalisco , m. 2003 , apj , 592 , 728 steidel , c. c. , shapley , a. e. , pettini , m. , adelberger , k. l. , erb , d. k. , reddy , n. a. , & hunt , m. p. 2004 , apj , 604 , 534 stern , d. , dey , a. , spinrad , h. , maxfield , l. , dickinson , m. , schlegel , d. , & gonzlez , r. a. 1999 , aj , 117 , 1122 stewart , k. r. , bullock , j. s. , wechsler , r. h. , maller , a. h. , & zentner , a. r.2007 , arxiv:0711.5027 wechsler , r. h. , somerville , r. s. , bullock , j. s. , kolatt , t. s. , primack , j. r. , blumenthal , g. r. , & dekel , a. 2001 , apj , 554 , 85 zirm , a. w. , overzier , r. a. , miley , g. k. , blakeslee , j. p. , clampin , m. , de breuck , c. , demarco , r. , ford , h. c. , hartig , g. f. , homeier , n. , illingworth , g. d. , martel , a. r. , rttgering , h. j. a. , venemans , b. , ardila , d. r. , bartko , f. , bentez , n. , bouwens , r. j. , bradley , l. d. , broadhurst , t. j. , brown , r. a. , burrows , c. j. , cheng , e. s. , cross , n. j. g. , feldman , p. d. , franx , m. , golimowski , d. a. , goto , t. , gronwall , c. , holden , b. , infante , l. , kimble , r. a. , krist , j. e. , lesser , m. p. , mei , s. , menanteau , f. , meurer , g. r. , motta , v. , postman , m. , rosati , p. , sirianni , m. , sparks , w. b. , tran , h. d. , tsvetanov , z. i. , white , r. l. , & zheng , w. 2005 , apj , 630 , 68 cccccc a & 3.0385 & 3.0354 & ( 3.0488 ) & -230 & ( 765 ) + b & 3.0343 & 3.0308 & & -260 & + c & 3.0289 & 3.0258 & 3.0416 & -230 & 945 + d & 3.0244 & 3.0209 & 3.0341 & -260 & 725 + e & ( 3.0161 ) & ( 3.0143 ) & ( 3.0263 ) & ( -135 ) & ( 760 ) +
|
we report the discovery of a very bright ( m@xmath0 ) lyman break galaxy at @xmath1 that appears to be a massive system in a late stage of merging .
deep imaging reveals multiple peaks in the brightness profile with angular separations of @xmath2@xmath3 ( @xmath225 @xmath4 kpc comoving ) .
in addition , high signal - to - noise ratio rest - frame uv spectroscopy shows evidence for @xmath5 components based on stellar photospheric and ism absorption lines with a velocity dispersion of @xmath6 km s@xmath7 for the three strongest components .
both the dynamics and high luminosity , as well as our analysis of a @xmath8cdm numerical simulation , suggest a very massive system with halo mass @xmath9 .
the simulation finds that all halos at @xmath10 of this mass contain sub - halos in agreement with the properties of these observed components and that such systems typically evolve into @xmath11 halos in groups and clusters by @xmath12 .
this discovery provides a rare opportunity to study the properties and individual components of @xmath13 systems that are likely to be the progenitors to brightest cluster galaxies .
|
understanding the structure of the proton and its excited states is one of the key questions in hadronic physics . known as the missing - baryon problem ,
quark models based on three constituent quark degrees of freedom predict many more states than have been observed experimentally .
baryon resonances are broad and widely overlap , especially at higher energies , imposing challenges on the interpretation of experimental data in terms of resonance contributions . without precise data from many decay channels , it will be difficult or even impossible to accurately determine the properties of well established resonances , or to confirm or rule out the existence of weakly established resonances or new , so - far not observed states . of particular importance
are well - chosen decay channels which can help isolate contributions from individual excited states and clarify their importance .
photoproduction of @xmath0 and @xmath1 mesons offers the distinct advantage of serving as an _ isospin filter _ for the spectrum of nucleon resonances and thus , simplifies data interpretations and theoretical efforts to predict the excited states contributing to these reactions .
since the @xmath0 and @xmath1 mesons have isospin @xmath6 , the @xmath7 and @xmath8 final states can only originate from intermediate @xmath9 nucleon states .
data on @xmath0 photoproduction off the free proton were obtained and studied at many different laboratories over a wide kinematic range @xcite .
a review of the main data sets and a corresponding comparison of their coverage in energy and solid angle can be found in @xcite .
almost all analyses found that the @xmath10 nucleon resonance dominates @xmath0 photoproduction at threshold , though there are models which do not need the @xmath10 resonance to describe threshold production of @xmath0 mesons @xcite .
the @xmath10 state is well - known for its large @xmath7 coupling , whereas other resonances couple only weakly to @xmath7 .
small contributions from the @xmath11 resonance via interference with the @xmath12 resonance have been determined from data on photon beam asymmetries @xcite and angular distributions .
data from target polarization experiments @xcite revealed surprising effects concerning the phase relations of the @xmath13- and @xmath14-wave amplitudes @xcite . despite its four - star assignment by the particle data group ( pdg ) @xcite , the role and nature of the @xmath10 is still not well understood .
surprisingly , the @xmath15 nucleon resonance has the same quantum numbers , but shows no strong @xmath7 coupling .
many different arguments have been discussed to explain this observation ; the two @xmath12 states can have appreciable mixing @xcite , for instance .
the @xmath10 resonance can also be a dynamically - generated state of the @xmath16 system @xcite or more generally , a dynamically generated resonance coming from the interaction of the octet of pseudoscalar mesons with the ground - state octet of baryons @xcite .
recent efforts at jefferson laboratory have concentrated on describing the @xmath17 transitions by the interaction of the photon with the core of the resonance including meson - cloud effects in the low @xmath18 region @xcite .
the agreement of the model predictions for the helicity amplitude @xmath19 with experimental data is good for some lower - lying @xmath20 states to fair for the @xmath10 not ruling out alternative explanations .
the importance of contributions from the @xmath15 resonance to @xmath0 photoproduction has been discussed further in conjunction with its photoproduction off the neutron .
recently , the neutron data have attracted interest due to the observation of a bump - like structure at 1.67 gev/@xmath21 @xcite , which has not been seen in the cross section off the proton . in @xcite , it has been shown that a strong interference between @xmath22 , @xmath23 , and a non - resonant background can provide a good description of these data . a partial wave analysis ( pwa ) of recent cb - elsa data in the framework of the bonn - gatchina ( bnga ) model @xcite , which included data on other reactions and from other experiments , found the dominance of three nucleon resonances in @xmath0 photoproduction : @xmath10 , @xmath24 , and a proposed new state , @xmath25 @xcite .
the large @xmath7 coupling of the @xmath24 was surprising .
solutions of the @xmath0-maid model @xcite in this mass range assign much of the intensity to the @xmath26 instead .
current efforts with regard to the extraction of double - polarization observables will help shed light on this controversy .
the data presented here cover the full angular range and within the framework of the bonn - gatchina model , are still consistent with the dominance of the three nucleon resonances , @xmath10 , @xmath24 , and @xmath25 .
data on @xmath1 photoproduction is scarce .
analyses published before 2005 observed fewer than 300 @xmath1 events @xcite and an interpretation in terms of resonance contributions was difficult .
data from jefferson laboratory significantly improved the world database @xcite .
they observed @xmath27 events , which allowed the extraction of differential cross sections .
though more precise than previous measurements , the clas data are still limited in their angular coverage . in the model by nakayama and haberzettl ( @xcite and ref .
[ 12 ] therein ) , the @xmath8 final state couples to @xmath10 and @xmath26 .
the authors claim the importance of @xmath28 states ( @xmath29 , @xmath30 , @xmath31 ) in the process , which are useful to obtain the correct shape of the differential cross sections for energies from 1.728 gev to 1.879 gev . in this paper , we present total and differential cross sections for the reactions : @xmath32 the data cover an incoming photon energy range up to 2.55 gev and show the full angular coverage .
the paper has the following structure .
section [ section : experimentalsetup ] gives a brief introduction to the cbelsa / taps experimental setup .
the data reconstruction and selection is discussed in section [ section : dataanalysis ] and the extraction of differential and total cross sections is described in section [ section : determinationcs ] .
experimental results are finally presented in section [ section : results ] .
the experiment was carried out at the electron accelerator facility elsa @xcite at the university of bonn using a combination of the crystal - barrel @xcite and taps @xcite detectors .
the experimental setup is shown in fig .
[ figure : experiment ] .
electrons with an energy of 3.175 gev were extracted from elsa via slow ( resonant ) extraction .
the bremsstrahlung - tagger photon - beam facility at elsa delivered unpolarized tagged - photon beams in the energy range from 0.5 to 2.9 gev by passing the electron beam through a thin copper radiator with a thickness of @xmath33 ( radiation length ) .
electrons are deflected in the field of the tagger dipole magnet according to their energy loss in the bremsstrahlung process ; the remaining energy is determined in a tagger detector consisting of 480 scintillating fibers above 14 scintillation counters ( tagger bars ) in a configuration with adjacent paddles partially overlapping .
the corresponding energy of an emitted photon is @xmath34 .
electrons not undergoing bremsstrahlung are deflected at small angles and guided into a beam dump located behind the tagger detectors .
the energy resolution is about 2 mev for the high - energy photons and 25 mev for the low - energy part of the bremsstrahlung spectrum . for the energy calibration of the tagger
, a polynomial was determined in simulations using the measured field map of the bending magnet and the known positions of the fibers .
the calibration was cross - checked by measurements with the elsa electron beam at two different energies . at 600 and 800 mev ,
a low - current beam was guided directly into the tagger , while the magnetic field was slowly varied .
these results provided corrections to the initial polynomial @xcite .
the photons hit the liquid hydrogen target in the center of the crystal - barrel ( cb ) calorimeter .
the target cell ( 5 cm in length , 3 cm in diameter ) was surrounded by a scintillating - fiber detector @xcite , which provided an unambiguous impact point for charged particles ( due to the arrangement of its three layers ) leaving the target .
the cb - calorimeter in its cbelsa / taps configuration of 2002/2003 consisted of 1290 csi(tl ) crystals with a length of 16 @xmath35 .
the modules have an excellent photon detection efficiency ; a detailed description can be found in @xcite . for this series of experiments , the ( downstream ) rings 11 - 13
were removed to combine the detector with taps in the forward direction .
the cb - calorimeter covered the complete azimuthal angle and polar angles from @xmath36 to @xmath37 .
all crystals are of trapezoidal shape pointing to the center of the target ( fig .
[ figure : cb - luzy - h2 ] , top ) .
the taps detector consisted of 528 hexagonal baf@xmath38 crystals with a length of about 12 @xmath35 .
it was configured as an hexagonal wall serving as the forward end cap of the crystal - barrel calorimeter ( fig .
[ figure : cb - luzy - h2 ] , bottom ) .
taps provided a high granularity in the forward direction covering polar angles between @xmath39 and @xmath36 ( full @xmath40 coverage ) .
a 5 mm thick plastic scintillator in front of each taps module allowed the identification of charged particles .
the combination of the crystal - barrel and taps calorimeters covered 99% of the @xmath41 solid angle and served as an excellent setup to detect multi - photon final states .
the fast response of the taps modules provided the first - level trigger .
the second - level trigger was based on a cellular logic ( face ) , which determined the number of clusters in the barrel .
the trigger required either two hits above a low - energy threshold in taps ( led - low ) or one hit above a higher - energy threshold in taps ( led - high ) in combination with at least one face cluster .
the shape of the logical segmentation for the taps trigger is shown in ( fig .
[ figure : cb - luzy - h2 ] , bottom ) . the beam - monitor placed at the end of the beam line provided valuable information on the beam intensity ( photons not interacting in the h@xmath38 target ) used for the determination of the photon flux .
this total absorption @xmath42erenkov counter consisted of an array of 9 lead glass crystals .
data presented here were accumulated from october 2002 until november 2002 in two run periods with elsa beam energies of 3.175 gev .
these data were used to extract differential and total cross sections for a variety of final states @xcite . the event reconstruction and selection of the two @xmath0 photoproduction channels ( [ reactioneta1 ] ) and ( [ reactioneta2 ] ) as well as the @xmath1 photoproduction channel ( [ reactionetap ] ) with incident photon energies up to 2.55 gev is presented in this section . the total number of @xmath43 @xmath0 events was observed ( @xmath44 for @xmath2 and @xmath45 @xmath46 ) covering invariant masses from 1510 to 2380 mev/@xmath21 . for the @xmath1 channel , @xmath47 events
were observed covering invariant masses from 1920 to 2380 mev/@xmath21 .
the @xmath1 threshold region , @xmath48 $ ] mev/@xmath21 , was analyzed separately using a finer energy binning to better study the threshold behavior .
events with at most one proton and with two or six photons were selected , respectively .
the charged clusters were identified in taps by using the plastic scintillators mounted in front of each baf@xmath38 crystal .
the efficiencies of these ( photon)-veto detectors were determined and modelled in the monte - carlo ( mc ) program . in the crystal - barrel reconstruction , a cluster
is assigned to a charged particle if the trajectory from the target center to the barrel hit forms an angle of less than @xmath49 with a trajectory from the target center to a hit in the scintillating fiber detector .
proton identification is only used to remove it from the list of photon candidates .
the proton momentum is then reconstructed from event kinematics in `` missing - proton '' kinematic fitting .
proton clusters are on average much smaller than photon clusters and provide worse angular resolution . the proton momentum direction reconstructed from kinematic fitting had to be consistent again with a calorimeter hit when a charged cluster was identified .
our monte - carlo studies of reaction ( [ reactioneta1 ] ) show that if a proton was observed in the event , the overall misidentification probability is less than 3% .
.kinematic fits and constraints used in the analysis .
the proton is treated as missing particle .
its momentum is determined from the kinematic fit . for the @xmath1 reconstruction ,
only one pion mass was imposed in the kinematic fit ( 3c fit ) . [ cols="^,<,<,^",options="header " , ] the target area density , i.e. the number of atoms in the target material per cross - sectional area ( orthogonal to the photon beam ) , is given by @xmath50 where @xmath51 g/@xmath52 is the density and @xmath53 g / mol the molar mass of liquid h@xmath38 .
@xmath54 mol@xmath55 is the avogadro number and @xmath56 cm the length of the target cell .
the factor of two accounts for the molecular composition of hydrogen ( h@xmath38 ) .
the cross sections were extracted independently for both @xmath0 decay modes , @xmath2 as well as @xmath46 , and then averaged ( weighted with their errors ) based on the observed good agreement ( fig .
[ ratiodecaymodes ] ) .
the number of events in an ( @xmath57,cos@xmath58 ) bin comprises events with two or three final - state particles ( at least @xmath59 s ) or six or seven `` particles '' ( at least @xmath60 s ) , respectively .
the proton can be `` missing '' , but events with and without detected proton are treated in the same way in the event reconstruction . at threshold , the event kinematics requires that the proton is used in the ( taps ) trigger .
thus , the threshold function for the detection of low - energy protons needs to be reasonably well understood .
the interval of the solid angle is given by @xmath61 with @xmath62 describing the width of the angular bins .
it was chosen to be @xmath63 for the @xmath0 data presented here .
energy bins were defined by considering statistics and ensuring a good comparability with other experiments .
a total of 34 bins is presented in energy steps of 50 mev for @xmath64 $ ] mev . for the @xmath1 data ,
@xmath62 was chosen to be 0.2 .
a total of 20 bins is presented in energy steps of 50 mev for @xmath65 $ ] mev .
the number of observed @xmath0 and @xmath1 mesons needs to be corrected for unseen decay modes .
partial - decay branching fractions used to correct the measured cross - sections are taken from @xcite : br@xmath66 , br@xmath67 , and br@xmath68 .
the tagging hodoscope consisted of 480 scintillating fibers above 14 partially overlapping scintillation counters ( tagger bars ) .
the photon flux was measured directly in the experiment and determined according to @xmath69 where @xmath70 are the _ free _ hardware counts for the individual fibers corrected for the life - time of the detector .
the parameter @xmath71 accounts for the ( fiber)-cluster reconstruction in the tagger , which has to be performed in the same way as for real hadronic events . the photon definition probability or @xmath72 denotes the probability that a real photon is emitted along the beam axis in the tagger and traverses the liquid hydrogen target .
the scalers are recorded in _
events , which were accumulated with a _ minimum - bias _ trigger at a rate of 1hz during the regular data taking .
this trigger required only a hit in the tagger and was thus independent of hadronic cross sections .
the parameter @xmath72 is determined from tagger - or - runs @xmath73 separate data runs utilizing a _ minimum - bias _ trigger . the total error of the photon flux is assumed to be dominated by the @xmath72 error and depends strongly on the efficiency of the beam monitor at the end of the beam line ( fig . [ figure : experiment ] ) .
@xmath72 was determined to @xmath74 by varying the background subtraction of non - coincident tagger - beam monitor hits .
this value is consistent with determinations from multiple tagger - or - runs at different incoming photon rates .
an overall error of 10% has been assigned to the photon flux determination .
the statistical errors are determined from the number of events in each ( @xmath57,cos@xmath58 ) bin .
statistical errors are shown for all data points ; systematic uncertainties are given as error bands at the bottom of each plot .
sources of the systematic errors are uncertainties in the exact position of the liquid hydrogen target and a possible offset of the photon beam .
the position of the target cell was determined from kinematic fitting by comparing the off - zero displacement of different pull distributions to monte - carlo simulations .
it was found to be shifted upstream by 0.65 cm @xcite .
the corresponding systematic errors were determined by varying the target position in the monte carlo @xmath75 and evaluating changes in the re - extracted differential cross sections .
the errors show an angular dependence , but are 2 - 3% on the average and @xmath76% at most around cos@xmath77 .
the photon beam was assumed to be shifted by less than 2 mm off axis at the target position .
the errors of the decay branching fractions are negligible .
the uncertainty of the proton trigger has been determined from the small disagreement of the differential @xmath0 cross sections using the @xmath78 and @xmath3 decay channels for @xmath79 gev and cos @xmath80 ( fig .
[ figure : etadcs_1 ] ) .
the reconstruction of neutral mesons and the identification of final states requires a sequence of cuts including the use of kinematic fitting .
as discussed in the following section , the reconstruction of @xmath0 mesons from final states with two and six photons leads to compatible results .
this fact emphasizes a good understanding of the detector response to multi - photon final states .
an overall @xmath81 error is assigned to the reconstruction efficiency as determined in @xcite .
an additional 3% systematic error accounts for the slightly different effects of confidence - level cuts on data and monte - carlo events .
all these errors are added quadratically to give the total systematic error .
moreover , the @xmath1 systematic error receives an additional contribution from an alternative way of reconstructing events via @xmath82 ( sec .
[ subsection : eventreconstruction ] ) .
[ figure : etadcs_1 ] shows the @xmath86 differential cross sections for the two different @xmath0 decay modes ( [ reactioneta1 ] ) and ( [ reactioneta2 ] ) .
we have excluded those data points in the analysis showing a monte - carlo acceptance of .
the data sets show excellent agreement .
we have checked the consistency of the two cross - section measurements by forming the ratio of partial widths of the two @xmath0 decay modes . since the cross sections in fig .
[ figure : etadcs_1 ] are corrected for the decay branching ratios , we determine @xmath87_{\eta\rightarrow 3\pi^0}\biggl[\frac{\gamma_{\eta\rightarrow 3\pi^0}}{\gamma_{total}}\biggr]_{\rm{pdg } } } { \biggl[{\rm \frac{d\sigma}{d\omega}}(e_{\gamma},\cos\theta_{\rm{cm}})\biggr]_{\eta\rightarrow 2\gamma}\biggl[\frac{\gamma_{\eta\rightarrow 2\gamma}}{\gamma_{total}}\biggr]_{\rm{pdg}}}\ ] ] for each ( @xmath57,cos@xmath58 ) bin .
the values are weighted with their squared inverse errors and histogrammed ( fig .
[ ratiodecaymodes ] ) .
we derive a peak position of @xmath88 where the systematic error is derived from considering all systematic uncertainties discussed in section [ subsection : systematicuncertainties ] .
we do not claim a new measurement of @xmath0 branching fractions here , but have rather used this number to check our reconstruction efficiency .
the particle data group gives two values for this branching ratio @xcite : @xmath89 is the mean value of all direct measurements while a combined fit to all partial decay widths yields a value of @xmath90 .
both pdg values are consistent with our result . since the differential cross sections for the two @xmath0 decay modes are consistent , we have calculated error - weighted mean values .
these are presented in fig .
[ figure : etadcs_2 ] as functions of energy and the @xmath0 production angle .
since the cross sections change rather smoothly , only few resonances are likely to contribute to the process .
the @xmath10 state is known to dominate the threshold region resulting in a flat distribution in cos@xmath91 ; interference with other amplitudes leads to deviations from that flat distribution . at higher energies , above @xmath92
gev , the development of a forward peak indicates important contributions from @xmath93-channel @xmath94 and @xmath95 exchange .
[ figure : etadcs_2 ] also shows a comparison of our new @xmath0 results to published results from cb - elsa @xcite .
the agreement between the two data sets is very good at lower energies .
however , the differential cross sections reported by cb - elsa show somewhat larger discrepancies at higher energies and forward angles . above @xmath96 gev and in the forward most angle bins , cb - elsa results are approximately 30% larger than our new cbelsa / taps results at these energies , but are still consistent within the errors .
we believe that the discrepancy is due to underestimated background in the cb - elsa data in the low - statistics forward most bins where fewer than 10 events were observed .
[ figure : etapdcs ] shows the differential cross sections for the reaction @xmath98 ( reaction [ reactionetap ] ) using cos@xmath99 bins of width 0.2 .
the data cover the full angular range . very similar to @xmath0 photoproduction
, a rather flat angular distribution is observed at low energies suggesting @xmath13-channel resonance production near threshold .
both data sets also show a continuing increase in slope at forward angles , which becomes more prominent at higher energies .
this forward peak is most likely due to @xmath93-channel exchange mechanisms .
moreover , our new data indicate a decrease in the forward most bin , which has not been observed before . above 2 gev in photon energy , growth at backward angles could be indicative of @xmath100-channel contributions .
the overall agreement between the cbelsa / taps and clas data is good at threshold to fair above @xmath101 mev .
the s - wave behavior of the reaction close to the reaction threshold is apparent from the experimental data .
[ figure : etapwqthreshold ] shows the differential cross sections for @xmath102 at and close to the reaction threshold using 5 cos@xmath103 bins of width 0.4 .
the cross sections have been determined for individual fibers of the tagging system and cover the full angular range .
the data points suffer from low statistics , but are overall consistent with flat angular distributions .
the expected energy dependence of the reaction at the threshold is given by @xcite : @xmath104 where @xmath105 denotes @xmath106 waves , etc . for @xmath13-wave dominance
, a linear energy dependence of the squared total cross section is thus expected close to the reaction threshold .
the differential cross sections shown in fig .
[ figure : etapwqthreshold ] have been used to determine the total @xmath1 cross section and to study the energy dependence . since the data cover the full angular range , no extrapolation is needed .
[ figure : etapthreshold ] shows the incoming photon energy plotted versus the squared total cross section .
a linear dependence is clearly observed and a fit determines the energy threshold to @xmath107 mev , which is compatible with the value of @xmath108 mev derived from the @xmath1 mass listed in the pdg @xcite .
the mass of the @xmath0 meson was determined by the taps / a2 collaboration in a very similar procedure @xcite .
[ fig : p11p13 ] shows the total cross section for @xmath0 photoproduction . due to the complete solid angle coverage
, no extrapolation is required and the data points are truly experimental . in the low - energy range , the @xmath12
partial wave dominates the cross section .
the solid line represents our previous pwa solution and is not a fit to these data ; the two states @xmath24 and @xmath25 saturated the total cross section @xcite . a new pwa solution including the new cbelsa / taps data presented here and other data sets is in preparation .
it is clear that single- and double - polarization variables are required to firmly establish resonance contributions . coupled channel fits to many reactions can also help ; in particular , when three - body final states are included , the phase of two - particle partial - wave amplitudes is tested in the crossed channel . a small anomaly is observed at 1.73 gev/@xmath21 in the total @xmath0 cross section . as discussed earlier , recent data off the neutron show a pronounced bump - like structure at 1.68 gev @xcite , which has been suggested to signal the existence of a narrow baryon state with @xmath109 mev/@xmath21 .
in particular , the possibility that this state could be the @xmath110 , non - strange member of an anti - decuplet of pentaquarks is certainly interesting @xcite .
[ fig : yields ] shows the total number of @xmath2 events for the tagger channels 419 - 431 .
the data have been fitted using a polynomial to indicate the smooth behavior of the distribution .
the data point at 1.73 gev/@xmath21 in the total cross section is based on the photon energy interval @xmath111 $ ] mev defined by the tagger channels @xmath112 $ ] .
no statistically significant enhancement is observed in fig .
[ fig : yields ] over the small energy range under investigation to explain the anomaly and a narrow state compatible with the observation in the total cross section can be ruled out .
we believe that this structure is an instrumental effect originating from tagger cluster - size corrections of low - rate tagger - or - data used in the photon - flux determination , which occurs only for those channels . in fig .
[ fig : etaptot ] , the total cross section for @xmath1 photoproduction is displayed .
again , due to the complete solid angle coverage , no extrapolation is required . at 2 gev in the invariant mass and above
, the cross section for @xmath1 production is about 50% of that for @xmath0 production assuming that the processes are dominated by @xmath113-exchange resonances .
for a pseudoscalar mixing angle of @xmath114 , the non-@xmath115 components of @xmath1 and @xmath0 differ by a factor 1/2 .
the similarity of the two numbers suggests that the dynamics of @xmath0 and @xmath1 photoproduction is similar .
the analysis of previous cb - elsa data on photoproduction of @xmath0-mesons @xcite revealed two surprises : \(1 ) the analysis suggested a new resonance with spin and parity @xmath116 , @xmath25 @xcite .
the total cross section was nearly saturated with three resonances , the well - known @xmath10 , the @xmath24 , and the new resonance .
\(2 ) the strong coupling of @xmath117 was also unexpected ; in maid @xcite , the @xmath118 was very significant while @xmath24 hardly contributed to @xmath0 photoproduction .
the pattern of states contributing most to @xmath0 photoproduction , @xmath10 , @xmath24 , @xmath25 , was interpreted as a sequence of quark model states with total intrinsic angular momenta @xmath119 and @xmath120 coupling to @xmath121 .
the regularity of this pattern was used to argue that the seed of @xmath10 should be of a three - quark nature ; due to the presence of @xmath122-wave thresholds , the state may attract large @xmath7 and @xmath123 molecular components .
the relative strength of the two nucleon excitations at an incoming photon energy of 1700 mev in @xmath86 remains disputed .
the gieen group @xcite found - like maid - @xmath118 to provide a significant contribution .
other coupled channel analyses confirmed the dominance of @xmath24 ( relative to @xmath26 ) . in @xcite and @xcite ,
a large variety of @xmath0 production data was fitted using an effective lagrangian approach ; in both analyses , the @xmath24 contribution was considerably larger than that of @xmath26 .
a chiral quark model approach complemented with a one - gluon exchange model @xcite arrived at the same conclusion .
restricted to photoproduction data , the best solution for the new data presented here ( in terms of its @xmath124 value ) still supports the dominance of the three nucleon resonances , @xmath10 , @xmath24 , and @xmath25 in @xmath0 photoproduction .
however , this solution is incompatible with data on @xmath125 .
this is presently investigated further and will be subject of a forthcoming publication of the bonn - gatchina partial wave analysis group .
in summary , we have presented data on the photo - produced @xmath0 and @xmath1 cross sections from the reactions @xmath86 with @xmath3 as well as @xmath78 and from the reaction @xmath102 with @xmath126 .
the continuous beam from the elsa accelerator and the fiber detector of the tagging system provided tagged - photons in the energy range from 850 to 2550 mev .
the results are in very good agreement with previous measurements , but extend over the full angular range in @xmath4 of the @xmath0 and @xmath1 meson .
the inclusion of the new @xmath0 data into a multi - channel partial wave analysis is in preparation .
the threshold behavior of the @xmath1 data indicate @xmath13-wave dominance .
we thank the technical staff at elsa and at all the participating institutions for their invaluable contributions to the success of the experiment .
we acknowledge financial support from the national science foundation ( nsf ) , deutsche forschungsgemeinschaft ( dfg ) within the sfb / tr16 and from schweizerischer nationalfond . the collaboration with st .
petersburg received funds from dfg and the russian foundation for basic research .
b. krusche _ et al .
_ , phys .
lett . * 74 * , 3736 ( 1995 ) .
j. ajaka _ et al .
_ , phys .
lett . * 81 * , 1797 ( 1998 ) .
a. bock _ et al .
_ , phys .
* 81 * , 534 ( 1998 ) . c. s. armstrong _ et al . _
[ jefferson lab e94014 collaboration ] , phys .
d * 60 * , 052004 ( 1999 ) .
r. thompson _
[ clas collaboration ] , phys .
* 86 * , 1702 ( 2001 ) .
f. renard _ et al . _
[ graal collaboration ] , phys .
b * 528 * , 215 ( 2002 ) [ arxiv : hep - ex/0011098 ] .
m. dugger _
[ clas collaboration ] , phys .
lett . * 89 * , 222002 ( 2002 ) [ erratum - ibid . *
89 * , 249904 ( 2002 ) ] .
v. crede _ et al . _
[ cb - elsa collaboration ] , phys .
lett . * 94 * , 012004 ( 2005 ) [ arxiv : hep - ex/0311045 ] .
d. elsner _
_ [ cbelsa collaboration and taps collaboration ] , eur .
j. a * 33 * , 147 ( 2007 ) . h. denizli _ et al . _ [ clas collaboration ] , phys .
c * 76 * , 015204 ( 2007 ) [ arxiv:0704.2546 [ nucl - ex ] ] .
o. bartholomy _ et al . _
[ cb - elsa collaboration ] , eur .
phys . j. a * 33 * , 133 ( 2007 ) .
j. denschlag , l. tiator and d. drechsel , eur .
j. a * 3 * , 171 ( 1998 ) [ arxiv : nucl - th/9802063 ] .
l. tiator , g. knochlein and c. bennhold , pin newslett .
* 14 * , 70 ( 1998 ) [ arxiv : nucl - th/9802064 ] . c. amsler
_ et al . _
[ particle data group ] , phys .
b * 667 * , 1 ( 2008 ) .
n. isgur and g. karl , phys .
b * 72 * , 109 ( 1977 ) .
n. kaiser , p. b. siegel and w. weise , phys .
b * 362 * , 23 ( 1995 ) [ arxiv : nucl - th/9507036 ] .
d. jido , j. a. oller , e. oset , a. ramos and u .- g .
meiner , nucl .
phys . a * 725 * , 181 ( 2003 ) [ arxiv : nucl - th/0303062 ] .
d. jido , m. doering and e. oset , phys .
c * 77 * , 065207 ( 2008 ) [ arxiv:0712.0038 [ nucl - th ] ] .
b. julia - diaz , t. s. lee , a. matsuyama , t. sato and l. c. smith , phys .
c * 77 * , 045205 ( 2008 ) .
v. kuznetsov [ graal collaboration ] , workshop on the physics of excited nucleons ( nstar 2004 ) , grenoble , france , 24 - 27 march 2004 , arxiv : hep - ex/0409032 . i. jaegle _ et al . _ [ cbelsa / taps collaboration ] , phys .
lett . * 100 * , 252002 ( 2008 ) .
anisovich , i. jaegle , e. klempt , b. krusche , v.a .
nikonov , a.v .
sarantsev and u. thoma , arxiv:0809.3340 [ hep - ph ] .
a. v. anisovich , a. sarantsev , o. bartholomy , e. klempt , v. a. nikonov and u. thoma , eur .
j. a * 25 * , 427 ( 2005 ) [ arxiv : hep - ex/0506010 ] . w. t. chiang , s. n. yang , l. tiator , m. vanderhaeghen and d. drechsel , phys .
c * 68 * , 045202 ( 2003 ) .
abbhhm collaboration , phys .
rev . * 175 * , 1669 ( 1968 ) .
r. plotzke _ et al . _
[ saphir collaboration ] , phys .
b * 444 * , 555 ( 1998 ) .
m. dugger _
_ , phys .
lett . * 96 * , 062001 ( 2006 ) [ erratum - ibid .
* 96 * , 169905 ( 2006 ) ] .
w. hillert , eur .
phys . j. a * 28s1 * , 139 ( 2006 ) .
e. aker _
et al . _ ,
instrum .
a * 321 * ( 1992 ) 69 .
a. r. gabler _ et al .
_ , nucl .
instrum .
a * 346 * , 168 ( 1994 ) .
r. novotny [ taps collaboration ] , ieee trans .
* 38 * , 379 ( 1991 ) .
frank klein , ph.d .
thesis ( in preparation ) , university of bonn ; igor horn , ph.d .
thesis , university of bonn .
g. suft _ et al .
_ , nucl .
instrum .
a * 538 * , 416 ( 2005 ) .
r. castelijns _ et al . _
[ cbelsa / taps collaboration ] , eur .
j. a * 35 * , 39 ( 2008 ) [ arxiv : nucl - ex/0702033 ] .
m. nanova _ et al . _
[ cbelsa / taps collaboration ] , eur . phys .
j. a * 35 * , 333 ( 2008 ) [ arxiv:0803.2146 [ nucl - ex ] ] . h. van pee _
[ cb - elsa collaboration ] , eur .
phys . j. a * 31 * , 61 ( 2007 ) [ arxiv:0704.1776 [ nucl - ex ] ] .
c. amsler _ et al .
_ [ crystal barrel collaboration ] , z. phys .
c * 58 * , 175 ( 1993 ) .
r. a. arndt , w. j. briscoe , i. i. strakovsky and r. l. workman , phys .
c * 66 * , 055213 ( 2002 ) .
b. krusche and s. schadmand , prog .
phys . * 51 * , 399 ( 2003 ) [ arxiv : nucl - ex/0306023 ] .
b. krusche _ et al .
_ , z. phys . a * 351 * , 237 ( 1995 ) .
d. diakonov , v. petrov and m. v. polyakov , z. phys .
a * 359 * , 305 ( 1997 ) [ arxiv : hep - ph/9703373 ] .
m. v. polyakov and a. rathke , eur .
j. a * 18 * , 691 ( 2003 ) [ arxiv : hep - ph/0303138 ] .
d. diakonov and v. petrov , phys .
d * 69 * , 094011 ( 2004 ) [ arxiv : hep - ph/0310212 ] .
l. tiator , int .
j. mod .
a * 22 * , 297 ( 2007 ) .
v. shklyar , h. lenske and u. mosel , phys .
b * 650 * , 172 ( 2007 ) [ arxiv : nucl - th/0611036 ] .
r. shyam and o. scholten , arxiv:0808.0632 [ nucl - th ] . k. nakayama , y. oh and h. haberzettl , arxiv:0803.3169 [ hep - ph ] .
j. he , b. saghai and z. li , phys .
c * 78 * , 035204 ( 2008 ) [ arxiv:0802.3816 [ nucl - th ] ] .
i. horn _ et al . _
[ cb - elsa collaboration ] , arxiv:0806.4251 [ nucl - ex ] .
i. horn _ et al . _
[ cb - elsa collaboration ] , phys .
* 101 * , 202002 ( 2008 ) [ arxiv:0711.1138 [ nucl - ex ] ] .
a. lleres _ et al . _
[ graal collaboration ] , arxiv:0807.3839 [ nucl - ex ] .
a. anisovich , e. klempt , a. sarantsev and u. thoma , eur .
j. a * 24 * , 111 ( 2005 ) [ arxiv : hep - ph/0407211 ] .
a. v. anisovich , v. v. anisovich , e. klempt , v. a. nikonov and a. v. sarantsev , eur .
j. a * 34 * , 129 ( 2007 ) .
a. v. anisovich and a. v. sarantsev , eur . phys .
j. a * 30 * , 427 ( 2006 ) [ arxiv : hep - ph/0605135 ] .
r. a. arndt , w. j. briscoe , i. i. strakovsky and r. l. workman , phys .
c * 74 * , 045205 ( 2006 ) [ arxiv : nucl - th/0605082 ] .
|
total and differential cross sections for @xmath0 and @xmath1 photoproduction off the proton have been determined with the cbelsa / taps detector for photon energies between 0.85 and 2.55 gev .
the @xmath0 mesons are detected in their two neutral decay modes , @xmath2 and @xmath3 , and for the first time , cover the full angular range in @xmath4 of the @xmath0 meson .
these new @xmath0 photoproduction data are consistent with the earlier cb - elsa results .
the @xmath1 mesons are observed in their neutral decay to @xmath5 and also extend the coverage in angular range .
|
the nature of the bright x - ray sources found in nearby galaxies , the ultraluminous x - ray sources ( ulxs ) , remains a major astrophysical puzzle .
the fundamental conundrum is that some of these objects have x - ray luminosities uncomfortably high to be stellar - mass black holes ( bh ) without violating standard eddington limit arguments .
three different solutions have been proposed for the luminosity problem .
1 ) the objects are intermediate - mass bhs ( colbert & mushotzky 1999 ) .
2 ) they are stellar - mass bhs with , in some cases , substantial beaming of their x - ray radiation ( king et al .
2001 ) , or , 3 ) they are stellar - mass bhs emitting above their eddington limit ( begelman 2006 ) .
it is possible that some ulxs appear ultraluminous because of a combination of all three factors ( moderately higher mass , mild beaming and mild super - eddington emission ) .
it may also be that these objects make up an inhomogeneous population , comprised of both a sub - sample of intermediate - mass bhs and moderately beamed stellar bhs ( for recent reviews see fabbiano & white 2006 ; miller & colbert 2004 ) . due to their extragalactic nature
the study of ulx counterparts at other wavebands has been difficult , and has so far precluded the use of the familiar methods of dynamical astronomy to weigh them .
however , recent work has resulted in mass measurements for some local group stellar bhs , including ic 10 x-1 ( prestwich et al . 2007 ; silverman & filippenko 2008 ) , and m33 x-7 ( orosz et al . 2007 ; liu et al .
2008 ) , but these are not ulxs .
a substantial body of work now supports the idea that timing properties can be used to constrain the masses of bhs .
for example , mchardy et al . (
2006 ) have shown that when luminosity variations are taken into account , both stellar - mass and supermassive bhs populate a `` variability plane '' linking their broad band variability time - scales , luminosities and masses ( see also krding et al .
casella et al ( 2008 ) applied this relation in order to estimate the masses of two ulxs , m82 x-1 and ngc 5408 x-1 , concluding that the masses of ngc 5408 x-1 , and m82 x-1 lie in the range ( in solar units ) @xmath6 , and @xmath7 , respectively .
however , in this case the scaling was not direct in that casella et al .
( 2008 ) had to estimate the relevant break time - scale from the observed qpo frequencies in m82 x-1 and ngc 5408 x-1 .
they did this by applying a scaling relation between the two quantities derived only from observations of stellar - mass bhs .
recently , gierlinski , nikolajuk & czerny ( 2008 ) have shown that the strength of high frequency variability
parameterized as the root mean squared ( rms ) amplitude integrated above the break in the power spectrum
scales approximately linearly with bh mass .
they show that a broad correlation exists when comparing stellar and supermassive bhs , but the relation is not tight enough to predict the mass of a stellar system by comparison , for example , with another stellar - mass bh of known mass .
it is now firmly established that galactic bhs accreting at high rates often classified as the intermediate state ( is ) or steep power - law state ( spl)show strong correlations between their spectral and temporal parameters . in these states
a significant fraction of the x - ray luminosity is in a power - law component whose spectral index correlates well with the frequency of a qpo ( so - called type c qpos ) that is also commonly detected in such states ( see sobczak et al .
2000 ; vignarca et al . 2003
; kalemci et al . 2005 ; remillard et al . 2002 ; casella et al . 2004 ) .
other spectral parameters also correlate with qpo frequency , such as the disk flux .
recent work by shaposhnikov & titarchuk ( 2007 ; 2009 ; hereafter st07 and st09 ) has demonstrated rather convincingly that the qpo frequency - spectral index scaling for type c qpos can be used as an empirical bh mass estimator .
previous efforts have been made to use this scaling argument to constrain the masses of ulxs ( fiorito & titarchuk 2004 ; dewangan , titarchuk & griffiths 2006 ; strohmayer et al .
2007 , hereafter paper 1 ) , however , in these previous studies , there was no direct evidence to indicate that the qpo properties seen in the ulx sources actually correlates with the spectral parameters as in stellar - mass systems .
in this paper we present the results of new timing and spectral measurements that show for the first time that ngc 5408 x-1 behaves very much like a galactic stellar - mass bh system with the exception that its characteristic x - ray time - scales are @xmath8 times longer , and its luminosity is greater by a roughly similar factor .
we argue that these new findings provide strong evidence that this system contains a few thousand solar mass bh .
paper 1 summarizes results from xmm - newton observations obtained in 2006 january in which x - ray qpos were detected from ngc 5408 x-1 for the first time ( we sometimes also refer to this as observation 1 ) . in order to look for correlated timing and spectral behavior we sought and obtained additional xmm - newton observations in 2008 january .
these new observations began on january 13 , 2008 , and continued for @xmath9 ksec ( observation 2 ) .
we used the standard sas version 8.0.0 tools to filter and extract images and event tables for both the pn and mos cameras .
we extracted events in an 18 radius around the source in both the pn and mos cameras . due to its higher count rate and better time resolution
we began our timing study using the pn data .
there was some background flaring present during our observation which broke up the exposure into several useful intervals .
we work with the four longest intervals , labeled a through d in time order , and with exposures of 15.9 , 23.0 , 20.0 and 13.6 ksec , respectively .
figure 1 shows the pn lightcurve in 78 s bins for the longest interval ( interval b ) .
we began our timing analysis by making average power spectra by combining all good intervals .
based on the energy dependent variability behavior exhibited in observation 1 ( see paper 1 ) we made power spectra in several energy bands ; 0.2 - 8 kev , 1 - 8 kev , and 2 - 8 kev .
we found a prominent peak at 10 mhz in the @xmath10 kev power spectrum .
figure 2 shows the average power spectrum binned at 0.64 mhz in which we found the peak . to assess the significance of the peak we first rescaled the power spectrum .
we did this by fitting a bending power - law continuum to the spectrum , excluding the two highest bins in the peak to avoid biasing the fit to higher values .
we then divided the power spectrum by the best - fitting continuum model .
this procedure is important when searching for qpos against an intrinsic , broad - band noise component as is present in ngc 5408 x-1 .
we estimate the significance as the chance probability of obtaining the highest observed peak in the rescaled power spectrum using for the noise power distribution the @xmath11 distribution with 176 degrees of freedom ( dof ) .
this is the expected distribution given the number of independent frequency bins that were averaged ( 88 ) .
we compared the distribution of powers in our observed power spectrum with this expected distribution and confirmed they are consistent .
this then gives a chance probability , per trial , of @xmath12 .
we searched out to 0.5 hz and this was the 2nd power spectrum searched , giving a total of @xmath13 independent trials , resulting in a significance of @xmath14 , a strong detection . to further quantify the variability we fit the same continuum model but now including a lorentzian to account for the qpo .
this model fits well , with a best fitting @xmath15 for 459 degrees of freedom , and is shown in figure 2 . removing the qpo component from the fit results in an increase in @xmath11 of 49 . for the addition of three additional qpo parameters ,
this increase in @xmath11 gives a chance probability by the f - test of @xmath16 , further supporting the detection .
the qpo has a centroid frequency of @xmath17 mhz , and a coherence of @xmath18 .
the qpo is strong , with an average amplitude ( rms ) in the @xmath1 kev band of @xmath19 .
we do not detect the qpo at energies below 1 kev , with a @xmath20 upper limit on the amplitude ( rms ) of @xmath21 , demonstrating that the amplitude of the qpo is a strong function of energy .
the band - limited noise is also quite strong , with an integrated amplitude ( rms ) of @xmath22 ( 1 - 100 mhz ) .
similarly to the qpo the broad - band noise amplitude is also stronger at higher energies .
this behavior is typical of galactic bhs as well ( nowak et al .
1999 ; belloni et al .
2005 ) .
we next examined the intervals separately . for the longest interval ( b ) we again detect the 10 mhz qpo , but we also find evidence for a second qpo at 13.4 mhz .
figure 3 ( upper curve ) shows the power spectrum from interval b in the @xmath10 kev energy band .
the 13.4 mhz feature is significant at a bit better than the @xmath23 level , so we consider it robust .
interestingly , these two peaks have frequencies consistent with a ratio of 4:3 .
a pair of qpos with a similar frequency ratio were also detected in observation 1 ( see paper 1 ) .
these similarities provide added confidence that the 13.4 mhz qpo is significant . combining the two longest intervals ( b and c ) above 1 kev gives a power spectrum with a strong detection of the 10 mhz feature , as well as two other candidate features ( see figure 3 , lower curve ) .
the higher frequency qpo at 13.4 mhz is still present , though it is less prominent than in interval b alone , and a third feature is suggested at 6 mhz .
we modeled this power spectrum with the same bending power - law continuum and included up to three qpo components .
we find a statistically acceptable fit with @xmath24 ( 497 degrees of freedom ) .
the 10 mhz qpo is absolutely required , as removing it from the fit increases @xmath11 by 96 , better than an @xmath25 detection based on the f - test . removing the lowest and highest frequency qpo components one at a time results in increases in @xmath11 of 17.7 and 29.2 , respectively .
these correspond to f - test probabilities of @xmath26 and @xmath27 , respectively , for the additional components .
while this provides rather strong evidence for the 13.4 mhz qpo , we regard the 6 mhz feature as tentatively detected . in summary , our new observations of ngc 5408 x-1 reveal a strong qpo at 10 mhz , and very strong `` flat - topped '' band - limited noise breaking to a power law , with the break close to the qpo .
this behavior is qualitatively similar to results from our 2006 observations , but with the qpo and break frequencies shifted down in frequency by about a factor of two , and with stronger rms variability .
previous spectral studies have shown that ngc 5408 x-1 has a cool thermal disk component with @xmath28 kev and a power - law extending to higher energies with a slope of @xmath3 ( kaaret et al . 2003 ; soria et al . 2004 ; paper 1 ) .
we obtained a pn spectrum by extracting an 18 region around the source .
background was obtained from a nearby circular region free of sources .
we began by fitting the present spectrum with the same model components used to describe our 2006 data ( paper 1 ) ; a relativistic disk ( _ diskpn _ in xspec ) , a power - law , and a thermal plasma ( _ apec _ in xspec ) .
these were modified by successive photoelectric absorption components , one fixed at the best galactic value ( @xmath29 @xmath30 ) , the other left free to account for possible local absorption .
this model provides an acceptable fit , and from a qualitative standpoint the spectral parameters are similar to those derived from our 2006 observations .
the disk temperature of @xmath31 kev is consistent within the errors to that derived from the 2006 data .
the power - law index , at @xmath32 , is slightly smaller than the previous measurement , but only different at about the 2@xmath33 level .
we searched for line emission in the fe band but did not detect any significant features .
however , the limits are not that constraining , with an @xmath20 confidence upper limit on the equivalent width at 6.4 kev of @xmath34 ev .
the thermal plasma model parameters and flux are consistent with the 2006 data , contributing a 0.3 - 10 kev flux of @xmath35 erg @xmath30 s@xmath36 in each observation .
examination of _ chandra _ images of ngc 5408 indicate the existence of extended emission co - spatial with ngc 5408 x-1 .
this emission is unresolved in the xmm observations of ngc 5408 x-1 , and we suggest that this spectral component is likely associated with star formation in ngc 5408 and not intrinsic to the ulx
. we will compare and constrast the spectral results from the two epochs in more detail below .
we can now compare the timing and spectral properties from the two epochs , and determine whether or not ngc 5408 x-1 shows patterns of behavior similar to that seen in galactic bhs .
we begin with a more detailed comparison of the energy spectrum and fluxes during the two observations . for consistency
we re - extracted the spectrum from the 2006 observations to include essentially all of the good exposure , and in each epoch the spectrum represents the total accumulated over the entire observation .
we produced count rate spectra for each observation and figure 4 shows the difference spectrum as a function of energy ( observation 1 - observation 2 ) .
one can see from the figure that the first observation ( 2006 ) had a higher count rate , and that most of the difference is in the @xmath37 kev band , which is dominated by the disk flux .
this demonstrates that the 2006 observation had relatively more counts in the soft band ( @xmath37 kev ) , and less in the power - law component extending to higher energy .
table 2 provides a detailed comparison of the spectral parameters derived from the two epochs .
the unabsorbed flux ( 0.3 - 10 kev ) in 2008 was @xmath38 , about @xmath39 less than the 2006 observation , and the power - law component represents a greater fraction of the total flux than in the 2006 data .
the two most important conclusions are that the disk flux was about @xmath40 higher in observation 1 , and that the power - law component contributes a greater fraction of the total flux in observation 2 .
we next compared power spectra accumulated during each observation .
figure 5 shows power spectra characteristic of each epoch .
the upper curve is data from our 2008 observations , and the lower curve is from 2006 ( see also paper 1 , figure 3 ) .
these power spectra were not accumulated over the exact same energy bands . because the variability has a significant energy dependence , and because a primary goal is to compare the characteristic time - scales in each epoch , we choose to compare the spectra in the energy bands where the signal to noise ratio of the respective qpos is largest .
for the 2006 data this corresponds to 0.2 - 8 kev , whereas for the 2008 data we use , of the bands searched , the 1 - 8 kev band .
table 1 compares relevant timing properties for the two observations .
timing properties were derived from the model fits shown in figure 5 using the bending power - law continuum and lorentzian components for the qpos .
we note that the bend frequency in the @xmath41 kev 2006 data we used to compare the 2006 and 2008 qpo parameters is not well constrained .
a more representative value is the bend frequency of @xmath42 mhz derived from the @xmath43 kev 2006 data ( paper 1 ) . in general , the measurement of the qpo frequency in both epochs is much more precise than the bend frequency , which is why we emphasize it in our comparisons of the two epochs .
closer examination of figure 5 reveals several important conclusions concerning the variability in ngc 5408 x-1 .
we see that the strongest qpo has shifted from 20 to 10 mhz , over the same time that the disk flux dropped by @xmath40 .
additionally , the photon power - law index decreased modestly as the qpo frequency dropped .
these behaviors are entirely consistent with what has been observed in galactic bhs with so - called type c qpos ( see , for example , sobczak et al . 2000 ;
vignarca et al .
2003 ; kalemci et al . 2005 ) .
another strikingly evident feature in figure 5 is the larger amplitude of the variability ( broad - band and qpo ) in the 2008 observation ( top ) .
this quantity is proportional to the integral of the power spectrum above the poisson level ( here a value of 2 ) , which is clearly larger in the 2008 observations .
the rms amplitude also has a scaling with the square root of the count rate , but this was smaller in the 2008 observations , so this effect simply makes the apparent difference larger .
we also compared the rms amplitudes _ in the same energy band _ , and the conclusion is unchanged .
again , the higher rms variability at the epoch of lower disk flux is strikingly similar to what is observed in galactic systems . here
the primary conclusion is that the variability is mostly carried by the power - law component , and not the disk flux .
so , when the power - law becomes more dominant , then the rms amplitude increases .
finally , we note that the modest drop in the power - law index with the decrease in qpo frequency is also consistent with the correlations seen in galactic systems . in this case
the relatively small change in the index accompanying a larger change in qpo frequency suggests that the power - law index in ngc 5408 x-1 may be near `` saturation '' of the scaling ( see for example , figure 5 in vignarca et al .
2003 ; st09 ) , that is , both the qpo frequency and index are near their upper ranges in the correlation .
this conclusion is also supported by the behavior of the rms amplitude .
the strong drop in the rms amplitude at relatively constant power - law index is consistent with the qpo frequency being at the higher end of the observed correlations ( see , for example , the behavior of xte j1550 - 564 illustrated in figure 5a from sobczak et al .
our new observations of ngc 5408 x-1 reveal correlated variations in its timing and spectral properties very much like a stellar - mass black hole . indeed , the evidence is now quite compelling that the strong low frequency qpo detected on several occasions from ngc 5408 x-1 is a type - c qpo analogous to those seen in galactic systems .
it varies in frequency and amplitude with changes in flux and spectrum in a manner entirely consistent with the behavior in galactic bhs such as xte j1550 - 564 and grs 1915 + 105 .
whereas these stellar - mass systems have characteristic qpo frequencies of a few hz , ngc 5408 x-1 shows qpo frequencies lower by a factor @xmath8 while simultaneously radiating an x - ray luminosity larger by a roughly similar factor . given greater confidence in the identification of and scaling properties of the strong qpo seen in ngc 5408 x-1
we can use it to derive a mass estimate .
we essentially follow the scheme outlined by st09 .
we use the qpo frequency - power - law index correlations measured for five systems ; grs 1915 + 105 ( vignarca et al . 2003 ) , xte j1550 - 564 , xte j1859 + 226 , h 1743 - 322 and gx 339 - 4 ( st09 ) .
we use these systems as our primary sources for mass estimates because they have both measured power - law indices that overlap the range of observed indices in ngc 5408 x-1 , and reasonable bh mass estimates from dynamical measurements .
since we do not yet have a correlation from ngc 5408 x-1 with many points to scale from we carry out a simplified procedure . for our reference galactic systems
we find the range of qpo frequencies with power - law indices between 2.4 and 2.6 .
we choose this index range as representative and conservative .
it is 2@xmath33 below the lower best fit value ( from 2008 ) and 2@xmath33 above the higher best - fit value ( from 2006 ) for ngc 5408 x-1 . in the case
that a range of qpo frequency exists at either end of our index range , we take the largest range of observed frequencies . to be specific ,
we take the lowest frequency measured at an index of 2.4 , and the highest frequency found at an index of 2.6 .
we then find the multiplicative scale factor , @xmath44 , required to bring the lower and upper qpo values measured for ngc 5408 x-1 ( 10 and 20 mhz ) into agreement with the measured frequency range for our reference systems . in the case that a constant scale factor can not match both the lower and higher frequencies of the observed range , we find the scale factor that aligns the centers of the two ranges ( we call this estimate @xmath45 ) . the derived mass estimate for ngc 5408 x-1 is then the mass of the reference stellar system times the scale factor .
in one case , h 1743 - 322 , frequency measurements exist only near the low end of our target index range ( 2.4 ) . in this case
we derive the scale factor by simply scaling to the lower qpo frequency measured in ngc 5408 x-1 .
the derived scale factors and mass estimates based on comparison with these five reference systems are summarized in table 3 .
the quoted uncertainty on the best mass estimate in table 3 reflects the uncertainty in the masses of the reference systems ( where available ) , that is , assuming a `` correct '' scale factor .
the qpo frequency and power - law index measurements for ngc 5408 x-1 described here are rather precise , so that if all the assumptions made with regard to the mass scaling arguments above are accurate , then the amount of statistical error in the estimates is modest .
this leaves systematic uncertainties , that is , how accurate is the derived scale factor ?
now , some galactic bhs do show variations in their qpo frequency spectral index correlations .
these are primarily associated with apparent changes in the value of the spectral index at which the correlation saturates ( see , for example , the behavior of xte j1550 - 564 in figure 7 from st09 ) , or simply that the correlation is not exact , and that at a given index value , there can be a range of measured qpo frequencies .
we took into account such variations in defining the range of qpo frequencies to scale to for each reference system .
however , the existence of a range of measured frequencies at a given spectral index suggests a conservative way to bound the scale factor sytematic error , by defining a minimum and maximum scale factor given the observed qpo frequency range .
we do this by scaling the lowest observed qpo frequency in our reference systems with the highest observed qpo frequency in ngc 5408 x-1 , and vice versa . for example , for xte j1550 - 564 , which has observed qpo frequencies from 2.5 - 6.5 hz , the minimum and maximum scale factors would be ( 2.5 hz ) / ( 0.02 hz ) = 125 , and ( 6.5 hz ) / ( 0.01 hz ) = 650 . based on these limits on the scale factor
we also derive minimum and maximum mass estimates by multiplying the maximum and minimum scale factors by the @xmath46 mass limits for each reference system .
both the scale factor ranges and minimum and maximum mass estimates are also given in table 3 .
we emphasize that we think these represent rather conservative limits . while the derived mass ranges in table 3 are rather large and we emphasize that this is not a precision mass measurement technique there is substantial overlap among the estimates from all the different sources .
perhaps more interesting is that all the scaling estimates suggest a bh with a mass comfortably greater than @xmath47 , that is , much greater than the current known mass range for stellar bhs .
we note that the candidate bh system 4u 1630 - 47which is not a dynamically confirmed bh also has qpo frequency measurements that overlap our target range of spectral index ( see figure 9 in vignarca et al .
2003 ) . in this source
the frequency range is @xmath48 hz , suggesting a best scale factor of 400 , which falls in the range derived for our other reference systems .
while it s mass is not known , a typical value of @xmath49 @xmath4 would also suggest a mass for ngc 5408 x-1 in the range of several thousand @xmath4 , consistent with the other sources .
one concern with regard to qpo frequency scaling arguments has been which qpo frequency to scale to .
for example , both xte j1550 - 564 and grs 1915 + 105 can sometimes show qpo frequencies below 1 hz , approaching 0.1 hz in the case of xte j1550 - 564 .
however , our observations of qpos from ngc 5408 x-1 at different epochs , and with different frequencies and rms amplitudes goes along way towards alleviating this concern .
in fact , the behavior of the qpo rms amplitude in ngc 5408 x-1 is completely reversed to what one would expect for a scaling with the lowest qpo frequencies observed in xte j1550 - 564 and grs 1915 + 105 .
results from xte j1550 - 564 show that at low qpo frequency , the rms increases with an increase in qpo frequency ( see figure 5a in sobczak et al .
2000 ) , but this is exactly opposite to the behavior seen in ngc 5408 x-1 .
the behavior of the qpos in both xte j1550 - 564 and grs 1915 + 105 when the power - law index and frequency are both high is a much better match to the behavior seen in ngc 5408 x-1 .
another concern has been that while qpo frequency scalings may give one result , a simple scaling of luminosities gives another , and that the two may not be in agreement .
here one must try to be consistent and compare luminosities under the same conditions . in the case of our scaling arguments
this would be to match luminosities at the appropriate power - law index and scaled qpo frequency .
an important issue here is that in many cases the distances to bhs in the galaxy are relatively less well known than the distances to nearby galaxies hosting ulxs .
nevertheless , we can attempt to compare the luminosity of ngc 5408 x-1 to some of our reference bhs . for xte 1550 - 564 , st09 report x - ray spectral and qpo frequency measurements . for power - law index and qpo frequency appropriate to our scaling
the observed x - ray flux from xte j1550 - 564 ( 2 - 20 kev ) was in the range 3 - 5 @xmath50 erg @xmath30 s@xmath36 .
the distance to this source is not very well constrained , with estimates ranging from 2.5 - 6 kpc . taking a flux of @xmath51 erg @xmath30 s@xmath36 as representative
, we have a luminosity of @xmath52 .
this compares with a representative luminosity from our 2006 observation of about @xmath53 erg @xmath30 s@xmath36 ( 0.2 - 10 kev ; assuming a distance of 4.8 mpc ) .
the luminosity ratio is then in the range of 317 - 55 , for the distance range of 2.5 - 6 kpc for xte j1550 - 564 . simply scaling up the @xmath46 mass range for xte j1550 - 564 by these limits gives a mass estimate for ngc 5408 x-1 of @xmath54 @xmath4 .
interestingly , st09 use an observed correlation in the power - law index and bulk motion comptonization model normalization ( model _ bmc _ in xspec ) to estimate distances as well as bh masses by scaling to a reference system .
interestingly , they favor a distance closer to 3 kpc than 6 kpc for xte j1550 - 564 , which would favor the high end of the derived mass range . for grs 1915 + 105 ,
again the distance is rather uncertain , likely being in the range from @xmath55 kpc ( dhawan et al .
2007 ) . for the qpo frequency range above a representative flux range is about 2 - 5 @xmath50 erg @xmath30 s@xmath36 ( muno et al .
1999 ) , giving a range of luminosity ratio of @xmath56 .
similar scaling as for xte j1550 - 564 would imply a mass range from @xmath57 @xmath4 .
while this range appears systematically smaller than the mass range inferred from the qpo scaling , it still overlaps with the minimum mass estimate for grs 1915 + 105 .
recently , gierlinski et al . (
2008 ) have argued that the amplitude of x - ray variability at high frequencies can be used as an estimator of bh mass .
their hypothesis is that there is a `` universal '' power spectral shape for bhs at high frequency . here ,
high frequency means above the break in the power spectrum .
this universal form is roughly a power law with index of 1.5 - 2.0 .
they explore the notion that this part of the power spectrum is nearly constant for a given source , but scales with bh mass .
they find that there is a roughly linear correlation extending from the stellar - mass bhs to agn ( using type 1 seyferts ) .
the relation is not exact , and the scatter is such to make it difficult to estimate the mass of a stellar bh by scaling from another stellar - mass system , however , for order of magnitude estimates the method seems reasonably robust .
we used our power spectrum continuum models ( above the bend or break in the spectrum ) to estimate the parameter @xmath58 , which is simply the normalization of the high frequency power - law component at 1 hz .
we integrated our best fitting continuum models from the bend frequency to 0.5 hz in order to estimate @xmath58 ( see eqn . 2 in gierlinski et al .
we find @xmath59 ranges of @xmath60 , and @xmath61 for the 2008 and 2006 observations , respectively .
these values correspond to mass ranges of 1686 - 2612 , and 4435 - 5714 @xmath4 , respectively , for the best fitting correlation derived by gierlinski et al . , and 3737 - 4704 , and 9828 - 12661 @xmath4 using their soft - state relation derived from cyg x-1 . while the possible mass range from this method is large , the @xmath58 values for ngc 5408 x-1 fall almost midway between the stellar - mass systems and ngc 4395 , a low mass agn ( see figure 7 in gierlinski et al .
2008 ) , and comfortably within the mass range consistent with intermediate mass bhs .
our new xmm - newton observations have revealed for the first time that ngc 5408 x-1 exhibits correlated x - ray timing and spectral properties quite analogous to those exhibited by galactic stellar - mass bhs in the `` very high '' or `` steep power - law '' state . its longer observed time - scales ( qpo and power spectral break frequencies ) at higher luminosity compared to galactic stellar - mass bhs can be understood if the mass of ngc 5408 x-1 is a few thousand solar masses .
we have arrived at this conclusion by several independent arguments . given our present understanding of the timing properties of bhs at all mass scales
, it seems to us hard to escape the conclusion that ngc 5408 x-1 is an intermediate mass bh .
we have again found evidence for a pair of sharp , closely spaced qpos in ngc 5408 x-1 with a frequency ratio consistent with 4:3 .
we thus conclude that this is an intrinsic feature of the system and not some artifact or coincidence associated with having only a single observation of the source . as we noted in paper 1 , it is possible that this results from detection of both a type c qpo ( the stronger feature ) , and a type b qpo ( casella et al . 2004 ) .
if the source transitions from one to another , we do not have sufficient signal to noise ratio data to `` watch '' such transitions occur .
rather , we simply see average detections of each qpo . we have argued that the observed variations in qpo frequency , accompanied by spectral changes consistent with behavior seen in galactic systems , allows the mass of ngc 5408 x-1 to be estimated by scaling the observed qpo frequencies to match those observed in stellar - mass systems of known mass .
essentially the same method has now been used on a good number of galactic systems and seems empirically robust ( st09 ) .
these arguments give mass estimates for ngc 5408 x-1 in a broad range from @xmath62 @xmath4 , however , we find that none of the qpo scaling estimates are consistent with masses below @xmath5 @xmath4 .
independent mass constraints based on the amplitude of high frequency variability also appear consistent with this range .
additional observations would help to map out the timing - spectral correlations for ngc 5408 x-1 more clearly , and thus allow more rigorous estimates . as noted in paper 1 , the measured disk temperature in ngc
5408 x-1 is also consistent with a few thousand @xmath4 bh if the theoretical accretion disk scaling of @xmath63 holds .
importantly , none of these new estimates appears consistent with a mass as low as even a few hundred solar masses .
we think these new findings provide strong evidence that ngc 5408 x-1 harbors an intermediate mass bh .
figure 1 : lightcurve of ngc 5408 x-1 ( 0.2 - 15 kev band ) from xmm - newton epic / pn observations showing the longest contiguous time interval used in our power spectral analysis ( interval b ) .
the bin size is 78.125 seconds .
a characteristic error bar is also shown .
figure 2 : average power spectrum of ngc 5408 x-1 from the @xmath1 kev epic / pn data ( histogram ) and the best fitting model ( solid ) .
the frequency resolution is 0.64 mhz , and each bin is an average of 44 independent power spectral measurements .
the effective exposure is @xmath64 ksec .
a characteristic error bar is also shown .
see the text for a detailed discussion of the model , and table 1 for model parameters .
figure 3 : average power spectrum of ngc 5408 x-1 from @xmath1 kev epic / pn data .
the upper curve shows data from interval b with a frequency resolution of 0.434 mhz .
each bin is an average of 10 independent measurements .
the bottom curve is an average of the two longest intervals ( b and c ) at a frequency resolution of 0.78 mhz .
each bin is an average of 32 independent measurements .
see the text for additional discussion .
figure 4 : difference of the count rate spectra between observation 1 and observation 2 ( i.e. obs1 - obs 2 ) .
this shows that observation 1 was brighter , and that most of the difference was associated with @xmath65 kev photons , that is , the disk component .
figure 5 : average power spectra and best fitting models characteristic of both observations of ngc 5408 x-1 from epic / pn data .
the upper curve is from the 2008 data ( intervals b + c ) , while the bottom curve is from the 2006 data .
see the text for additional details and table 1 for a summary of the model parameters .
a & @xmath66 & @xmath67 & @xmath68 + @xmath69 & @xmath70 & @xmath71 & @xmath72 + @xmath73 & @xmath74 & @xmath75 & @xmath76 + @xmath77 ( mhz ) & @xmath78 & @xmath79 & @xmath80 + @xmath81 ( @xmath82 ) & @xmath83 & @xmath84 & @xmath85 + @xmath86 ( mhz ) & @xmath87 & @xmath88 & @xmath89 + @xmath90 @xmath91 & @xmath92 & @xmath93 & @xmath94 + @xmath95 ( mhz ) & na & @xmath96 & @xmath97 + @xmath98 @xmath91 & na & @xmath99 & @xmath100 + @xmath101 ( mhz ) & na & @xmath102 & na + @xmath103 @xmath91 & na & @xmath104 & na + @xmath11 ( dof ) & 463 ( 459 ) & 518.8 ( 497 ) & 152.2 ( 180 ) + ccc + n@xmath105 & @xmath106 & @xmath107 + t@xmath108 & @xmath109 & @xmath110 + kt & @xmath111 & @xmath112 + @xmath113 & @xmath114 & @xmath32 + @xmath11/dof & 814/673 & 776/673 + f@xmath115 ( 0.3 - 10 , kev ) & @xmath116 & @xmath117 + f@xmath118 ( 0.3 - 10 , kev ) & @xmath119 & @xmath120 + f@xmath121 ( 0.3 - 10kev ) & @xmath122 & @xmath123 + cccccc xte j1550564 & @xmath124 & 2.5 - 6.5 & 125 , 300 , 650 & 1050 , @xmath125 , 6890 & 1 , 2 , 3 + grs 1915 + 105 & @xmath126 & 2.1 - 3.5 & 105 , 185 , 350 & 1050 , @xmath127 , 6300 & 2 , 3 , 4 + h 1743322 & @xmath128 & 3 - 5 & 150 , 267 , 500 & 1650 , 2937 , 5500 & 3 , 5 + xte j1859 + 226 & 7.6 - 12.0 & 7.5 & 375 , na , 750 & 2850 , na , 9000 & 3 , 6 , 7 + gx 3394 & @xmath129 & 5.9 - 7.8 & 295 , 457 , 780 & 1770 , 2742 , 4680 & 3 , 8 , 9 + 4u 163047 & na & 4 - 8 & 200 , 400 , 800 & na & 2 +
|
we report the discovery with xmm - newton of correlated spectral and timing behavior in the ultraluminous x - ray source ( ulx ) ngc 5408 x-1 .
an @xmath0 ksec pointing with xmm / newton obtained in january , 2008 reveals a strong 10 mhz qpo in the @xmath1 kev flux , as well as flat - topped , band limited noise breaking to a power law .
the energy spectrum is again dominated by two components , a @xmath2 kev thermal disk and a power - law with an index of @xmath3 .
these new measurements , combined with results from our previous january 2006 pointing in which we first detected qpos , show for the first time in a ulx a pattern of spectral and temporal correlations strongly analogous to that seen in galactic black hole sources , but at much higher x - ray luminosity and longer characteristic time - scales .
we find that the qpo frequency is proportional to the inferred disk flux , while the qpo and broad - band noise amplitude ( root mean squared , rms ) are inversely proportional to the disk flux .
assuming that qpo frequency scales inversely with black hole mass at a given power - law spectral index we derive mass estimates using the observed qpo frequency - spectral index relations from five stellar - mass black hole systems with dynamical mass constraints .
the results from all sources are consistent with a mass range for ngc 5408 x-1 from 1000 - 9000 @xmath4 .
we argue that these are conservative limits , and a more likely range is from 2000 - 5000 @xmath4 . moreover , the recent relation from gierlinski et al .
that relates black hole mass to the strength of variability at high frequencies ( above the break in the power spectrum ) is also indicative of such a high mass for ngc 5408 x-1 .
importantly , none of the above estimates appears consistent with a black hole mass less than @xmath5 @xmath4 for ngc 5408 x-1 .
we argue that these new findings strongly support the conclusion that ngc 5408 x-1 harbors an intermediate mass black hole .
|
wide band - gap nitride semiconductors are currently used in full - color displays , white light sources , and ultraviolet laser diodes for high - density storage systems.@xcite such semiconductors combine group - v nitrogen with elements of group iii such as boron , aluminium , gallium , and indium .
recently , the well - known nitride compound gan has been extensively investigated in the form of quantum dots , both with wurtzite@xcite and zinc - blende@xcite crystal structures .
the typical phenomena appearing in quantum dots are the discretization of the electronic spectra and the blue shift of the fundamental gaps.@xcite moreover , the gan nanocrystals can be doped with diluted magnetic impurities such as manganese .
in fact , ref .
12 shows ( ga , mn)n quantum dots prepared under solvothermal conditions in the wurtzite phase . these particles seem to show a ferromagnetic signal in the ground state,@xcite like bulk ( ga , mn)n as calculated for diluted mn spins.@xcite we must note that many experiments on bulk ( ga , mn)n suggest that mn spins are not diluted but forming clusters which give the observed ferromagnetism.@xcite a comparative _ ab - initio _ study between ( ga , mn)n nanocrystals in wurtzite and zinc - blende phases is nevertheless missing . in this work
we investigate wurtzite and zinc - blende gan quantum dots doped with one or two mn impurities within density functional theory .
the doping of nanocrystals with mn atoms that replace host cations ( mn@xmath0 ) is actually possible , as already confirmed by several experiments on dots larger than ours.@xcite the experimental doping of small nanocrystals of about 1 nm in size has not yet been reported .
however , it is expected from evidence of 2 nm undoped nanoparticles already synthesized . hence
, our calculations anticipate future experimental work with small doped dots .
the theoretical and computational details are given in sec .
[ computational ] .
the case of ( ga , mn)n nanoparticles doped with a single mn impurity is studied in sec .
[ nanosingle ] .
we show in this section that the doping reaction is endothermic and requires high temperatures as already confirmed by the experiments.@xcite the case of ( ga , mn)n crystallites doped with two mn dopants is investigated in sec .
[ nanotwo ] , where we show that the ground - state mn impurities are antiferromagnetically aligned .
the antiferromagnetic order of the two mn spins is related to the different chemical environment around the empty state induced by the mn dopant close to surface .
hereafter , we call `` holes '' these empty mn - induced states which have mainly mn @xmath1 character , but also n.@xcite this hole state lies within the nanocrystal gap near the conduction region and therefore does not contribute effectively to the ferromagnetic alignment of the two mn impurities
. the interesting role of the mn holes near the surface is explained in sec .
[ nanosingle ] and sec .
[ nanotwo ]
. then , ( ga , mn)n quantum dots of about 2 nm in size would be likely antiferromagnetic .
this possibility is interesting to be taken into account for spintronic applications based on such iii - v mn - doped nanostructures .
of course , the antiferromagnetic behavior of such small dots will also affect the overall magnetic character of granular solids@xcite formed by ( ga , mn)n nanocrystals .
we calculate ( ga , mn)n quantum dots within density functional theory , following the kohn - sham scheme and the projector augmented - wave method , as implemented in vasp ( vienna ab - initio simulation package).@xcite apart from the @xmath2 valence states inherent in semiconductors , we also take into account the mn @xmath3 states . these latter electrons are responsible for the spin - splitted states at the gap edges through the @xmath2-@xmath1 hybridization . for the exchange - correlation potential@xcite in the kohn - sham equations we use the generalized - gradient approximation @xmath4 ( gga+@xmath5),@xcite in which @xmath5 and @xmath6 are special parameters that account for the strong coulomb and exchange interactions between the mn 3@xmath1 electrons .
we take @xmath7 and @xmath8 as calculated for mn when doping bulk gan.@xcite we investigate quasi - spherical nanocrystals of @xmath9 in diameter centered in a ga position both with wurtzite and zinc - blende structures .
we passivate the surface dangling bonds with pseudohydrogens ( h@xmath10)@xcite as an approach to quantum dots synthesized in colloidal solutions and also grown in semiconductor matrices .
these fictitious atoms also prevent the appearance of surface states in the near - gap spectrum.@xcite through passivation , every ga ( 4@xmath114@xmath12 ) dangling bond at the dot surface is attached to a pseudohydrogen with a fractional charge of @xmath13 , and every n ( 2@xmath112@xmath14 ) dangling bond to a pseudohydrogen with a fractional charge of @xmath15 .
dot surfaces are perfectly saturated and free of defects so that we can avoid any perturbation and clearly investigate mn - mn magnetic interactions within nanocrystals .
anyhow , the influence of surface defects in the magnetic properties of mn - doped dots is an interesting issue for further studies .
the doped ( ga , mn)n nanoparticles contain one or two mn atoms which substitute for one or two ga cations .
we use the supercell approximation@xcite to calculate wurzite and zinc - blende crystallites which are infinitely repeated in space .
the size of the supercell is fixed to 22 so that surfaces of adjacent dots become separated by 10 and total energies are converged to mevs .
this supercell size thus permits an accurate enough description of mn - mn magnetic interactions within nanocrystals .
the atomic positions are fully relaxed until the forces on the atoms are small enough ( @xmath16 ev / ) . the input ga - n bond lengths are taken from bulk gan in the relaxed wurtzite and zinc - blende structures , @xmath17 .
for wurtzite gan , our calculated lattice constants are @xmath18=3.24 and @xmath19=5.29 ; for zinc - blende gan , our lattice constant is @xmath18=4.59 .
the cut - off energy in the plane - wave basis set is fixed to 500 ev in order to converge the total energies of bulk gan and ferromagnetic mnn below 1 mev .
( color online ) studied ( ga , mn)n nanocrystals with ( a ) wurtzite and ( b ) zinc - blende structures .
the ga atoms are drawn with large spheres ( pink ) , the n atoms with medium spheres ( green ) , and the @xmath20 pseudohydrogens with small spheres ( yellow ) .
the considered positions for the substitutional mn impurities in ga sites are the centers labeled `` i '' , the off - center sites labeled `` ii '' , and for zinc - blende dots the sites labeled `` iii '' . ]
in this section we study gan nanocrystals doped with a single mn atom .
the dopant concentrations@xcite are @xmath21 for wurtzite quantum dots and @xmath22 for zinc - blende dots .
these concentrations are in the same order of those used in the experiments.@xcite the relaxed geometries are depicted in fig .
[ fig : fig1 ] , where we see that the mn impurity can occupy the central site labeled `` i '' , the off - center sites labeled `` ii '' , and in zinc - blende particles also the sites labeled `` iii '' . positions ii are near the crystal surface , not exactly at the surface . since results concerning site iii are dependent on the passivation species , in the following discussions we will focus on sites i and ii .
as we relax the atomic positions in ( ga , mn)n quantum dots , we comment the influence of mn doping on the crystal geometries . in the undoped nanoparticles the ga - n bond lengths calculated for wurtzite and zinc - blende structures
are @xmath23 for the central ga cations and @xmath24 around the off - center ga ions placed in site ii .
the smaller bond distances for position ii as compared with position i are typically due to quantum dot surfaces . in the doped nanoparticles , the mn - n bonds measure @xmath25 for the central - mn case and @xmath26 around the off - center mn dopants placed in site ii
therefore , as compared with the undoped structures , the n shell around mn expands by @xmath27 , in close accordance with the 2@xmath28 calculated expansion in wurtzite and zinc - blende bulk ( ga , mn)n . we note that this expansion was not observed in ii - vi ( cd , mn)te quantum dots,@xcite and is even contrary to the contraction around mn dopants in iii - v ( in , mn)p
nanowires.@xcite we focus now on the energies involved in the formation and doping of ( ga , mn)n nanocrystals . the cohesive energy of the undoped nanoparticle is defined in relation to the free atoms , @xmath29 , where @xmath30 is the energy of the @xmath31th free atom , @xmath32 is the total ground - state energy , and @xmath33 is the number of atoms in the dot .
single - atom total energies are also obtained within the supercell approximation .
the computed values are 3.607 ev / atom for the wurtzite structure and 3.593 ev / atom for the zinc - blende one . in undoped nanocrystals
the wurtzite phase is thus more stable than the zinc - blende phase . however , since the growth of quantum dots ( qds ) also depends on kinetics and other chemical potentials apart from those used in the cohesive energy , the undoped gan crystallites can actually be synthesized both with wurtzite and zinc - blende structures.@xcite therefore , we are going to study mn - doped gan nanoparticles in both kind of geometries , wurtzite and zinc blende .
( color online ) substitutional reaction energy for a single mn impurity . squares ( blue ) indicate wurtzite structure and circles ( green ) indicate zinc - blende structure .
dotted lines are drawn to guide the eye .
the most stable position for the mn dopant is site ii irrespective of the crystal geometry . ] anyhow , the energetics of mn doping for both structures is studied by the following reaction , @xmath34 in which the mn@xmath35 dopant substitutes a ga@xmath36 cation in the quantum dot .
reaction ( 1 ) points to the main energy difference in the process of solvothermal growth during substitutional doping .
the cationic form for the mn dopant participates in reactions with the solvent which are beyond the scope of this work .
for instance , these reactions may include precursors such as gacl3 and mncl2 in the presence of hexamethyldisilazane ( hdms).@xcite total energies of ions stem from the calculation of total energies of neutral atoms to which the first and second ionization energies ( ies ) are summed up , that is , @xmath32(mn@xmath35)@xmath37(mn)+1@xmath38 ie + 2@xmath39 ie , where ie values are taken from the literature .
the substitutional energy @xmath40 involved in reaction ( [ ener ] ) is calculated against the mn position and plotted in fig .
[ fig : fig5 ] . since this energy is positive , reaction ( [ ener ] ) becomes endothermic and its activation requires high temperatures .
this finding is in agreement with the experiments in which ( ga , mn)n nanocrystals are prepared under solvothermal conditions at about 350 @xmath41c.@xcite in addition , the substitutional energy @xmath40 is smaller for site ii than for site i. this indicates that position ii near the surface is more stable than position i , in accordance with previous calculations for substitutional mn impurities embedded in ii - vi qds@xcite and also in iii - v nanowires.@xcite ( color online ) ( a ) spin density in zinc - blende quantum dots doped with a central mn impurity .
the isosurfaces correspond to constant densities @xmath42@xmath43 , with light gray ( red ) for the positive value and dark gray ( blue ) for the negative one .
( b ) the shown spin density is in a plane which includes the mn dopant and two of its four n neighbors .
the density at the mn site is chopped for the sake of clarity .
see how the spherical integration of the spin density around mn and n atoms yields local magnetic moments antiferromagnetically coupled . ] when doping ( ga , mn)n nanocrystals with a single mn impurity , one @xmath3 electron of the free mn atom ( @xmath44 ) transfers to its neighbor n anions and yields a @xmath45 configuration at the dopant position with a localized mn hole.@xcite nevertheless , the value of the local magnetic moment at the mn site is different from 4@xmath46 due to the @xmath2-@xmath1 hybridization .
this local moment is 3.99@xmath46/4.10@xmath46 when the mn atom occupies the center of wurtzite / zinc - blende quantum dots and 3.88@xmath46/3.88@xmath46 when it is placed off - center in position ii .
moreover , the mn dopant induces the magnetic polarization of its neighbor n anions . in fig .
[ fig : fig2 ] we show for zinc - blende nanoparticles the spin density around the central mn impurity and its four n neighbors .
the integration of this density within spheres of wigner - seitz radii centered in the n atoms results in n local magnetic moments of -0.06@xmath46/-0.09@xmath46 for wurtzite / zinc - blende quantum dots .
small modifications of these radii would lead to roughly the same local moments with the same signs , as already seen for other mn - doped nanocrystals.@xcite the exchange coupling between mn and n magnetic moments is thus antiferromagnetic , as also calculated for bulk wurtzite and bulk zinc - blende ( ga , mn)n .
we can compare the mn local magnetic moments obtained for quantum dots and for bulk ( ga , mn)n with similar mn concentration . for bulk
wurtzite @xmath47 , the computed mn magnetic moment is 3.99@xmath46 ; for bulk zinc - blende @xmath48 , the mn magnetic moment is 4.03@xmath46 .
these bulk values are hence similar to the previous mn magnetic moments centered in wurtzite and zinc - blende ( ga , mn)n nanocrystals . however , they show larger differences with respect to those moments calculated for mn in site ii near the surface .
we shall see in next sec .
[ gap- ] the local densities of states for further explanation about such larger differences .
( color online ) local densities of states ( ldos ) for wurtzite quantum dots doped with a single mn impurity placed ( a ) in the dot center and ( b ) in site ii off - center .
the densities of states are projected onto the mn @xmath3 states in black ( blue ) , the neighbor n @xmath49 states with solid black lines , and the neighbor n @xmath50 states in gray ( orange ) .
the vertical dotted lines indicate the fermi energy .
see for mn off - center the hole level placed within the gap at around 2 ev near the conduction edge .
the zinc - blende densities of states are analogous to the wurtzite ones with the only difference of a triply degeneracy in the spin - up valence edge state ( homo ) for the central - mn case .
insets widen the conduction edge states ( ulumos ) .
note the mn hole level closer to the conduction region in panel ( b ) as compared with panel ( a ) , and the different energy shifts for the up and down ulumos . ]
we investigate the hole linked to the mn impurity by looking at the local densities of states for mn in site i and site ii close to surface . in fig .
[ fig : fig3 ] we show the local projections of the crystal states onto the mn @xmath3 states and n @xmath2 states around the mn dopant . the wurtzite case is given in fig .
[ fig : fig3](a ) for a centered mn impurity and in fig .
[ fig : fig3](b ) for an off - center mn placed in site ii . from the densities of states
we conclude that ( i ) the hole level associated with the central mn dopant lies at around 0.6 ev near the fermi energy ; ( ii ) nevertheless , the hole level associated with the off - center mn impurity lies at around 2 ev near the conduction region .
the zinc - blende densities of states are similar to the wurtzite densities apart from the central - mn case which shows a triply degeneracy in the spin - up valence edge state .
such zinc - blende case is described in detail in fig .
[ fig : fig4 ] . due to the @xmath51 crystal field ,
the five 3@xmath1 states of the central mn atom are divided in two groups ; one is composed of three @xmath52-symmetry states and the other of two @xmath53-symmetry states .
the @xmath52 states of the mn impurity hybridize with the @xmath12-like @xmath52 states of its neighbor n atoms and yield the formation of bonding and antibonding states .
the latter states are degenerate and only partially occupied due to the mn hole . for zinc - blende nanoparticles with off - center mn dopants and also for wurtzite dots , the crystal symmetry is @xmath51-like .
( color online ) branch diagram describing the @xmath2-@xmath1 hybridization in a zinc - blende quantum dot doped with a centered mn impurity . on the left
we show the splitted mn @xmath3 states in the @xmath51-symmetry crystal field , in the middle the hybridized @xmath54 bonding and antibonding nanocrystal states , and on the right the spin - up 2p states of the four n atoms surrounding the mn impurity in the td - symmetry crystal field . the mn hole is indicated as an open rectangle in the shadowed valence region . ]
the @xmath2-@xmath1 hybridization between the mn 3@xmath1 states and the @xmath2 host states yields an effective mn - quantum dot exchange interaction that splits the crystal states at the gap edges.@xcite note that we are interested in the change of iii - v dot states by mn impurities .
the @xmath55-@xmath56 energy splittings are explicitly given here and also rewritten in terms of @xmath2-@xmath1 constants , @xmath57 and @xmath58 .
thereby , the rescaled splittings are interesting not only for theoreticians but also for experimentalists , both working on diluted magnetic compounds in the bulk@xcite and in quantum dots.@xcite the @xmath2-@xmath1 exchange constants are defined with the following mean - field theory expressions:@xcite @xmath59 here @xmath60 is the number of cations per unit volume;@xcite @xmath61 is the splitting of the up and down undoped lowest unoccupied molecular orbitals ( ulumos ) ; @xmath62 is the splitting of the up and down highest occupied molecular orbitals ( homos ) ; and @xmath63 is the average mn spin . the splittings @xmath64 and @xmath65 can be extracted from the local densities of states depicted in fig . [
fig : fig3 ] .
the @xmath57 and @xmath58 exchange constants are calculated from eq .
( [ eq : exchange ] ) and presented in fig .
[ fig : fig7 ] together with @xmath66 for different mn positions .
both wurtzite and zinc - blende @xmath67 values are larger for the central site i than for site ii off - center , as it also happens for ii - vi quantum dots doped with mn.@xcite ( color online ) the @xmath2-@xmath1 exchange constants and gap - edge splittings for single mn - doped nanocrystals as a function of the impurity position .
the open symbols stand for @xmath57 exchange constant and the closed symbols stand for @xmath58 . squares ( blue ) refer to wurtzite structure and circles ( green ) refer to zinc - blende structure .
dotted lines are drawn to guide the eye .
the bulk exchange constants are indicated on the left axis with similar notation .
the @xmath67 values for mn in the dot center ( site i ) are larger than the bulk ones due to the different roles of the mn holes . ] the dependence of the @xmath58 exchange constant with the mn position can be explained in more detail by looking at the valence states in the gap region . in fig .
[ fig : fig10 ] we show these states for two mn sites , dot center and off - center position ii .
for the central - mn case , the up - down homo splitting is large both for wurtzite and zinc - blende geometries .
however , when mn is moved off - center the spin - up valence levels decrease in energy and the spin - down levels slightly increase .
these shifts are due to a smaller mn - quantum dot exchange interaction caused by a smaller host charge density around position ii .
the up - down homo splittings for these mn sites are consequently reduced as compared with those for the dot centers , and the @xmath58 exchange constants become hence smaller .
we note that for the central - mn case the charge of the up states is mainly distributed over the mn impurity and its n neighbors , but for the down states it suffers a deplection around the mn dopant that we ascribe to the fact that mn and n are antiferromagnetically coupled .
in addition , for mn in position ii the charge in the up homos spreads perpendicularly to the @xmath68 axis and globally shows @xmath69-like character .
( color online ) valence levels in the gap region for ( a ) wurtzite and ( b ) zinc - blende quantum dots doped with a single mn impurity located in the central site i with open rectangles and in an off - center site ii with solid lines .
the charge densities next to the valence states show their global @xmath70- or @xmath71-like character .
the geometries with the reference axis centered in the dot are oriented so that position ii lies on the @xmath68 axis .
crosses indicate site ii whenever visible . ]
the @xmath2-@xmath1 exchange constants are different for nanoparticles than for bulk ( ga , mn)n . in order to assess this difference
, we calculate @xmath57 and @xmath58 for two bulk compounds with mn concentrations which are similar to those in the studied crystallites . for bulk
we obtain @xmath72 ev and @xmath73 ev ; for bulk zinc blende @xmath48 we obtain @xmath74 ev and @xmath75 ev . for comparison ,
these bulk constants are indicated in fig .
[ fig : fig7 ] as marks on the left axis .
the @xmath57 values for mn in positions i and ii are comparable to the @xmath57 ones calculated for bulk wurtzite and bulk zinc blende .
moreover , the @xmath67 values are larger than the bulk ones for mn in the central site i than in site ii near the surface . due to confinement
, the hole level associated with the central mn impurity is closer to the valence levels within the nanocrystal gap than in the bulk .
the stronger exchange interaction between this mn hole with @xmath1-like character and the valence edge states increases the splitting of the up and down homos and consequently the @xmath67 values . for position ii
as compared with position i , the off - center mn hole lying near the conduction levels indicates a smaller interaction with the valence states , and hence smaller @xmath67 values .
in this section we investigate ( ga , mn)n nanoparticles doped with two substitutional mn impurities in the ferromagnetic and antiferromagnetic configurations . the ferromagnetic state is calculated for a total magnetic moment in the quantum dot ( qd ) of 8@xmath46 ; the antiferromagnetic state is calculated for a null magnetic moment in the qd . by doping with two mn ions we replace two ga cations as described by the following reaction : @xmath76 the required energy for this double substitution is referred to as @xmath77 and plotted in fig .
[ fig : fig6](a ) for different mn - mn positions and magnetic alignments . as commented previously ,
the positive substitutional energies indicate that reaction ( 3 ) is endothermic and thus activated by increasing the temperature , as it occurs in the experiments.@xcite fig .
[ fig : fig6 ] also shows that the most stable mn impurities are aligned antiferromagnetically and placed in sites i - ii close to surface .
the calculated antiferromagnetic ground state was unexpected , since it is different from the ferromagnetic alignment of mn spins in bulk ( ga , mn)n.@xcite we relate it to the different role of the hole linked to the mn dopant that approaches the surface .
the energy of this hole lies within the nanocrystal gap near the conduction levels .
therefore , this mn hole does not contribute effectively to the ferromagnetic order of the two mn spins . as a consequence ,
the mn - mn coupling becomes antiferromagnetic in the ground state as it occurs for two mn spins in ii - vi ( cd , mn)te qds.@xcite ( color online ) ( a ) substitutional reaction energy for two mn dopants .
( b ) interaction energy between two mn impurities .
squares ( blue ) on the left stand for wurtzite ( wz ) structure and circles ( green ) on the right stand for zinc - blende ( zb ) structure .
closed symbols refer to ferromagnetic mn spins and open symbols refer to antiferromagnetic .
dotted lines are drawn to guide the eye .
we note from ( a ) that the most stable mn impurities are aligned antiferromagnetically and located in positions i - ii both for wurtzite and zinc - blende geometries . ] we next study the interaction energy between the two mn dopants which is defined as @xmath78 mn@xmath79 stands for the first mn atom and mn@xmath80 stands for the second mn atom .
the calculated interaction energies are plotted in fig .
[ fig : fig6](b ) as a function of the positions and magnetic couplings of the two mn spins .
the interaction energy @xmath81 quantifies the relative stability of nanoparticles doped with one or two mn impurities .
positive interaction values indicate that the two mn dopants tend to occupy the same nanocrystal and negative values indicate that they tend to dope two different quantum dots individually .
for instance , from fig .
[ fig : fig6](b ) it can be seen that one crystallite with two antiferromagnetic mn spins in positions i - ii is more stable than two nanoparticles with two single mn spins placed in sites i and ii . we stress again that in doping reactions kinetics and other chemical potentials different from those used in the cohesive energy could also play an important role and may modify previous results concerning the stability of the nanostructures .
since we are dealing with two mn impurities , we calculate their local magnetic moments and total energies in the ferromagnetic and antiferromagnetic states .
they are not explicitely given here but rewritten in terms of an effective mn - mn exchange interaction , quantified by @xmath82 , which is interesting not only for theoreticians@xcite but also for experimentalists@xcite working on diluted magnetic semiconductors .
the @xmath82 exchange constant stems from the heisenberg - like hamiltonian@xcite @xmath83 and it is thus defined as @xmath84 where @xmath85 is the total energy of the ferromagnetic state , @xmath86 is the total energy of the antiferromagnetic state , and @xmath87 is the local magnetic moment at the mn@xmath88 site in bohr magnetons .
the @xmath82 exchange constants are calculated and plotted in fig .
[ fig : fig9 ] as a function of the mn - mn positions .
negative @xmath82 values mean antiferromagnetic alignments between mn spins and positive @xmath82 values , ferromagnetic alignments . fig
[ fig : fig9 ] shows also that in the most stable positions i - ii the two mn atoms are antiferromagnetically ordered unlike mn spins in bulk ferromagnetic ( ga , mn)n . ( color online ) the @xmath82 exchange constants for two mn impurities as a function of the mn - mn positions . squares ( blue ) denote wurtzite structure and circles ( green ) denote zinc - blende structure .
dotted lines are drawn to guide the eye .
for the most stable positions i - ii the negative @xmath82 values indicate that the two mn spins are unexpectedly ordered antiferromagnetically , unlike mn spins in bulk ferromagnetic ( ga , mn)n . ]
( color online ) the @xmath2-@xmath1 exchange constants and gap - edge splittings for quantum dots doped with two mn impurities .
squares ( blue ) on the left refer to wurtzite quantum dots in the ferromagnetic configuration of mn spins and circles ( green ) on the right refer to zinc - blende dots .
open symbols denote @xmath57 exchange constant and closed symbols denote @xmath58 .
the triangles stand for antiferromagnetic mn spins .
dotted lines are drawn to guide the eye . for comparison , the exchange constants for bulk ferromagnetic ( ga
, mn)n are indicated on the two left axis with similar notation .
the @xmath67 values are larger for ferromagnetic nanocrystals than for bulk ( ga , mn)n . ] to look at the modification of the iii - v dot states by mn impurities , we now investigate the @xmath57 and @xmath58 exchange constants obtained from the calculated spin splittings at the gap edges and eq .
( [ eq : exchange ] ) .
the computed values are shown in fig .
[ fig : fig8 ] as a function of the positions and magnetic couplings of the two mn spins .
the main results are the following : \(i ) @xmath89 and @xmath67 values are larger for ferromagnetic mn spins in positions i - ii than for farther apart mn spins in positions ii - ii .
this decrease is similar to that observed for two mn spins in ii - vi ( cd , mn)te quantum dots.@xcite \(ii ) @xmath57 values are negative for ferromagnetic mn spins . the hole levels associated with the off - center mn impurities lie at high energies within the nanocrystals gaps and push the spin - up ulumos above the spin - down ulumos .
\(iii ) positive @xmath57 values are larger for antiferromagnetic mn spins placed in sites i - ii than @xmath57 for a doped nanoparticle with a central mn atom . the hole level linked to the spin - down mn dopant located in site ii pushes upward in energy the spin - down ulumo and thereby increases the spin splitting of the ulumos corresponding to the single - mn case .
\(iv ) comparing fig .
[ fig : fig7 ] and fig .
[ fig : fig8 ] we see that the largest @xmath67 values are for wurtzite and zinc - blende quantum dots doped with a central mn impurity . as in the previous discussion of sec .
[ gap- ] , we now compare the @xmath57 and @xmath58 exchange constants for nanocrystals and for bulk ( ga , mn)n .
we calculate the bulk ferromagnetic compounds wurtzite @xmath90 and zinc - blende @xmath91 .
the mn concentration in bulk @xmath90 is similar to @xmath92 in wurtzite quantum dots ; that of bulk @xmath91 is similar to @xmath93 in zinc - blende dots . for bulk
wurtzite we obtain @xmath94 ev and @xmath95 ev ; for bulk zinc blende we obtain @xmath96 ev and @xmath97 ev .
these exchange values are shown for comparison in fig .
[ fig : fig8 ] as marks on the left axis . for wurtzite nanoparticles in the ferromagnetic configuration of mn spins ,
the @xmath89 values are smaller than in the bulk ; on the contrary , for zinc - blende quantum dots in the ferromagnetic state , the @xmath89 values are significantly larger than in the bulk . for ferromagnetic nanocrystals ,
the wurtzite and zinc - blende @xmath67 values are both larger than the corresponding @xmath67 bulk constants .
these differences can also be explained by the different roles played by the mn holes in bulk structures and in quantum dots ( as seen in previous sec .
[ gap- ] ) .
it seems that for nanostructures , the situation of the dopant hole within the gap must be analyzed in detail in order to understand their basic magnetic properties .
( color online ) local densities of states ( ldos ) projected onto the mn 3@xmath1 states for wurtzite and zinc - blende quantum dots made of ( ga , mn)as .
the two mn impurities are closely placed in positions i - ii and ferromagnetically aligned .
note that the two mn holes approach each other and the valence edge unlike in ( ga , mn)n nanoparticles . ]
so far , the antiferromagnetic alignment of mn in gan quantum dots has been the main result of our discussion : role of mn hole in gan nanostructures in sec .
[ gap- ] and magnetic order of two mn spins in the last section .
in fact , the mn atom close to the crystal surface induces impurity states near the conduction levels .
now , to make contact with the antiferromagnetic coupling of such mn atoms , it is essential to consider other iii - v nanoparticles . experimentally , ( ga , mn)as nanocrystals can be created by mn implantation on gaas followed by thermal treatment,@xcite and also by annealing ( ga , mn)as thin films grown by molecular beam epitaxy.@xcite the self - organized nanoclusters are analyzed by x - ray spectroscopy,@xcite microscopic techniques,@xcite and also by using squid magnetometry.@xcite magnetic force microscopy measurements on ( ga , mn)as nanoprecipitates show ferromagnetic features at room temperature.@xcite this experimental result and the reported antiferromagnetic behavior of ( ga , mn)n quantum dots motivates us to search for a change in the magnetic order also in ( ga , mn)as nanocrystals . to illustrate this other compound ,
( ga , mn)as , without taking into consideration all the positions for mn , let us focus on i - ii sites , i.e. the case shown to be clearly antiferromagnetic ( afm ) in ( ga , mn)n .
we calculate ( ga , mn)as nanoparticles with two mn dopants which replace two ga atoms in the close interacting sites i - ii .
we study both wurtzite and zinc - blende geometries of about 1 nm in diameter as those given in fig .
[ fig : fig1 ] .
the computational details in this case are similar to those already explained in sec . [ computational ] .
as compared with the undoped structures , in the doped ones the as shell around the mn atom close to surface is expanded by @xmath98 , thus it is more expanded than the n shell in ( ga , mn)n nanoparticles . moreover , the computed quantum dots are ferromagnetic ( fm ) in the ground state , with a fm - afm exchange interaction of 122 mev in the wurtzite phase and 104 mev in the zinc - blende phase . these exchange energies are roughly half the bulk value for mn atoms sitting in close ga positions , @xmath99200 mev,@xcite in agreement with the same decreasing tendency already calculated for other magnetic nanostructures such as ( cd , mn)te nanocrystals.@xcite it should be clear from the previous discussion of mn in gan quantum dots that for ferromagnetic mn spins the so - called mn hole levels must be close to the gaas valence states . indeed , the local densities of states projected onto the mn 3@xmath1 states in fig .
[ fig : fig11 ] show that the mn holes in gaas nanoparticles are close together in the nearby of the valence region .
these mn holes behave hence as in bulk ( ga , mn)as@xcite and mediate the ferromagnetic alignment of the two mn spins .
looking at the mn hole levels in its relation to the valence ( conduction ) region , we can see a ferromagnetic ( antiferromagnetic ) behavior of mn impurities in doped nanostructures .
the difference in this hole position must be found in the smaller bond compression around mn for ( ga , mn)as quantum dots .
in summary , we have investigated wurtzite and zinc - blende ( ga , mn)n quantum dots doped with one or two substitutional mn impurities within density functional theory . we have obtained that wurtzite and zinc - blende structures show similar results as they can not be distinguished up to second neighbors .
anyhow , there are small differences between them commented in the text when appropiate . for a single mn dopant in the dot center , the calculated @xmath67 values are larger than in bulk ( ga , mn)n with similar mn concentration . for two mn dopants ,
the most stable magnetic state is antiferromagnetic , and this was unexpected since bulk ( ga , mn)n exhibits ferromagnetism in the ground state . we ascribe this surprising effect in ( ga , mn)n nanoparticles to the holes linked to the mn impurities placed close to surface . these holes do not contribute effectively to the ferromagnetic order of the two mn spins .
we show that the ferromagnetic behavior of bulk ( ga , mn)n can be changed by reducing the crystal size . from the antiferromagnetic result on small ( ga , mn)n nanoparticles , it seems possible that larger dots and other nanostructures such as thin films with mn dopants close to surface could also be antiferromagnetic in the ground state .
therefore , we hope that our results concerning antiferromagnetic ( ga , mn)n quantum dots will encourage further _ ab - initio _ calculations and experiments on mn impurities buried near surfaces of semiconductor nanostructures .
indeed , recent in - progress work on ( ga , mn)n nanolayers suggests the same antiferromagnetic behavior for mn spins .
|
we investigate the electronic and magnetic properties of ( ga , mn)n nanocrystals using the density functional theory . we study both wurtzite and zinc - blende structures doped with one or two substitutional mn impurities . for a single mn dopant placed close to surface ,
the behavior of the empty mn - induced state , hereafter referred to as `` mn hole '' , is different from bulk ( ga , mn)n . the energy level corresponding to this off - center mn hole lies within the quantum dot gap near the conduction edge . for two mn dopants , the most stable magnetic configuration is antiferromagnetic , and this result was unexpected since ( ga , mn)n bulk shows ferromagnetism in the ground state .
the surprising antiferromagnetic alignment of two mn spins is ascribed also to the holes linked to the mn impurities that approach the surface . unlike ( ga , mn)n bulk
, these mn holes in confined ( ga , mn)n nanostructures do not contribute to the ferromagnetic alignment of the two mn spins .
|
the non - interacting electronic state in bilayer graphene with ab ( bernal ) stacking , whose lattice structure is illustrated in fig . 1 , has a pair of degenerate valence and conduction bands at two momentum points @xmath0 in the brillouin zone .
the energy varies quadratically with the momentum about these points , in contrast with the linear dispersion in the single - layer graphene .
@xcite with hopping energy @xmath1 between the stacking carbon atoms in different layers , the other two pairs of conduction and valence bands have energy at @xmath2 at the degeneracy points .
such a state with chemical potential at the charge neutrality point is unstable to infinitesimal electron - electron interactions . in weak - coupling approximation ,
in which the interaction energies are small compared to @xmath1 , one may look for instabilities restricting the hamiltonian to the set of lowest energy conduction and valence bands .
in such a reduced basis , a wide variety of symmetry - breaking states have been proposed as possible ground states @xcite including nematic state , @xcite anomalous quantum hall effect ( ahe ) state , @xcite and layered quantum antiferromanget ( afm ) . @xcite .
however , the conductance experiments , @xcite done with high mobility samples and with both a top and a bottom gate to ensure that the chemical potential is at the charge neutrality point , @xcite show an insulating state with characteristics which are not met by the proposed states .
@xcite the two most important experimental findings can be summarized as follows : + ( 1 ) the state at charge neutrality shows a gap in two - terminal conductance measurements with a conductance @xmath3 smaller than the limit of measurement @xmath4 for voltage @xmath5 , above which @xmath6 with @xmath7 , indicating an insulating state .
+ ( 2 ) in small magnetic fields , the gap increases monotonically with a field ; the conductance is consistent with a gap @xmath8 with @xmath9 , @xmath10 the free - electron mass , indicating an orbital effect .
the afm state has a gap above which the conductance has the form of eq .
( [ cond ] ) , but its gap is insensitive to the external finite magnetic field in the mean - field approximation . a self - consistent calculation by interpolating the @xmath11 insulating state to a high-@xmath12 quantum hall ferromagnet state could give a linear @xmath12 dependence of the gap at high magnetic fields @xcite ( this has also been pointed out in ref .
but the extension to low fields produces additional features which are not observed in experiments .
@xcite the authors argue that this may relate to experimental limitations .
the ahe state has a gap in the _ bulk _ whose variation with field is also of the form ( [ deltah ] ) ( refs . ; see also fig .
[ fig : gapvsb ] ) .
in addition , it has a surface band with a quantized hall conductance of @xmath13 .
this contributes to the two probe conductance measurements and therefore is not consistent with the observations . and
@xmath14 represent the bonds between the nearest neighbors ( nn ) and the next - nearest neighbors ( nnn ) , respectively . ] in this paper , we propose a magnetoelectric ( me ) state , with spontaneously generated current loops from next - nearest - neighbor electron interaction ( @xmath15 ) , which has properties consistent with the above experimental observations .
this state [ see figs .
[ fig : flux_pattern](d)-[fig : flux_pattern](f ) ] breaks the time - reversal symmetry and has an odd parity with respect to two layers . as the product of time - reversal and inversion symmetries is preserved , it does not have the topologically protected surface bands as in the ahe state .
in contrast , the loop current state which has an even parity with respect to two layers [ see figs . [
fig : flux_pattern](a)-[fig : flux_pattern](c ) ] , is an ahe state .
@xcite however , in weak coupling , the me state is a semi - metal with hole ( electron ) pockets at @xmath16 ( @xmath17 ) points .
we show that when the coupling is larger ( @xmath18 ) , the me state becomes gapped and is energetically favored over the ahe state .
this regime demands the full four band basis and is beyond the weak - coupling approaches .
in addition to showing that the dependence of the gap with magnetic field is consistent with the experimental observations , we also illustrate new properties of this state and propose experiments to differentiate it from other theoretical proposals .
the rest of paper is organized as follows .
we introduce a theoretical model for bilayer graphene in sec .
[ sec : model ] , which includes a long - range ( next - nearest neighbor ) coulomb interaction term in addition to the tight - binding hamiltonian .
we show by a mean - field analysis , in sec .
[ sec : meanfield ] , time - reversal symmetry - breaking states with ordered loop currents arising due to the long - range interaction .
they can further be classified by the inversion symmetry with respect to two layers as ahe and me states .
we further study the energetics , topological properties , as well the properties under an external magnetic field of these states by numerical analysis , in sec .
[ sec : numeric ] .
some details of the calculations as well as additional results are presented in the appendixes .
we consider the hamiltonian for bilayer graphene , @xmath19 where @xmath20 is the layer index taking the values @xmath21 , and @xmath22 labels the honeycomb lattice sites .
@xmath23 is the tight - binding hamiltonian on each layer , with nearest ( @xmath24 ev ) and next - nearest ( @xmath25 ev ) neighbor hoppings .
@xmath26 to @xmath27 are the three vectors connecting the nearest - neighbor sites on each layer and @xmath28 to @xmath29 are the 6 vectors connecting the next - nearest - neighbor sites [ see fig.1(b ) ] .
@xmath30 is the tight - binding part between two layers , for which we only keep the hopping between stacked atoms @xmath31 ev . to focus on the time - reversal symmetry - breaking states through orbital loop currents , we only consider the next - nearest - neighbor interaction @xmath15 in @xmath32 .
@xmath33 or @xmath34 is the charge density operator .
we assume that other interactions such as onsite interaction are not strong enough to generate an order and only renormalize the parameters for the states we consider .
in particular , we show in appendix [ sec : vnn ] that the nearest - neighbor interaction @xmath35 need not be considered even though @xmath36 .
we also assume the degeneracy in spin degrees of freedom and therefore the spin index is dropped .
the mean - field analysis of the model starts with the decomposition of the diagonal in the spin part of the density interactions between sites as was done to derive time - reversal breaking me states in cuprates @xcite @xmath37 the one - particle terms can be dropped .
now a mean - field approximation is made with @xmath38 .
this leads to a mean - field hamiltonian in tight - binding form but with complex hopping between next - nearest - neighbor sites : @xmath39 as well as an energy term @xmath40 .
@xmath41 is determined by minimization of energy .
when @xmath41 is finite and has the same sign ( loop currents ) along the triangular loop of the next - nearest - neighbor bonds for one sublattice , an ordered flux pattern is formed with alternating orientations ( signs ) between neighboring enclosed triangular areas ( see fig . [
fig : flux_onelayer ] ) .
classified by the sign combinations of two sublattices on two layers , four kinds of ordered loop current states can be generated through @xmath15 which break time reversal without breaking translational symmetry .
( 1 ) within each layer , the triangles of the a sublattice and the b sublattice have the same or opposite signs of flux ( see fig . [
fig : flux_onelayer ] ) .
( 2 ) between two layers , the triangles of the unstacked atoms ( b1 and a2 ) centered at the stacked atoms ( a1 and b2 ) have the same or opposite signs of flux as well [ see figs . [
fig : flux_pattern](a ) , [ fig : flux_pattern](b ) , [ fig : flux_pattern](d ) , and [ fig : flux_pattern](e ) ] . of these the two states which break inversion through having flux in the stacked and unstacked triangle of atoms in a given layer in opposite direction [ fig .
[ fig : flux_onelayer](b ) ] always have a higher ground - state energy than the other two constructed from fig .
[ fig : flux_onelayer](a ) and will not be considered further .
we are then left with two possibilities , with flux in the stacked and unstacked triangle of atoms in a given layer in the same direction .
they can be further classified by the second condition , and are + ( i ) the haldane or ahe state of the bilayer in which the orientations of the flux in the loops atop each other in the two layers are the same .
the flux patterns are shown in figs .
[ fig : flux_pattern](a)-[fig : flux_pattern](c ) . + ( ii ) the me phase with the opposite orientations of the flux in the loops atop each other in the two layers , figs . [ fig : flux_pattern](d)-[fig : flux_pattern](f ) . in both cases
the net flux through a unit cell is zero . :
the triangles enclosed by bonds connecting a1 and b1 sublattices and with the same center have the same ( a ) or opposite ( b ) signs of fluxes .
( a ) corresponds to the haldane ahe flux pattern for a single - layer honeycomb lattice ( ref . ) . here
, the inversion symmetry is preserved because the center of the honeycomb lattice is the inversion point .
however , the inversion symmetry is always broken for ( b ) . ] , the bottom layer [ ( b ) and ( e ) ] , and the net fluxes [ ( c ) and ( f ) ] are shown .
the color representation is the same as in fig [ fig : flux_onelayer ] . in ( f ) , we use light red ( blue ) to represent weaker strengths of fluxes compared to the regions with regular red ( blue ) colors . ]
we calculate numerically the ground - state energy for the mean - field hamiltonian in the space of all four bands .
we assume the magnitude of @xmath41 to the same for both stacked and unstacked atoms in each layer .
taking @xmath42 and @xmath43 , we find the leading dependence in energy at zero temperature on the order parameter @xmath41 for the ahe and the me states to be , respectively , @xmath44 equation [ eq : gsenergy ] gives the interesting result that for @xmath45 , where the saddle - point values of @xmath41 are finite for both states , the me state has a lower ground state energy in such mean - field calculation . as the critical value of @xmath15 is smaller for the ahe state , it also suggests that as temperature is decreased , the first state to arise through a transition with ising symmetry is the ahe state ( with a first order transition ) , and then to the me state at a lower temperature .
we calculate the temperature dependence of the ground - state energy , from which we estimate the transition temperature to the me state is @xmath46 ( for @xmath47 ) .
the transition temperature to the ahe state is @xmath48 higher .
the value of the transition temperature is expected to be depressed from these estimates when fluctuations are included .
and @xmath17 . the other two high - energy bands are split with a scale of @xmath1 . in the ahe state ,
band gaps are opened at both @xmath16 and @xmath17 due to the order parameter @xmath41 , @xmath49 . in the me state , there are also direct band gaps at @xmath16 and @xmath17 , @xmath50 .
but the center of the direct gaps shift in different directions of energy for @xmath16 and @xmath17 . when @xmath51 , an indirect gap opens . for illustration purposes ,
we take @xmath42 and @xmath52 here , which are slightly bigger than realistic values.,title="fig : " ] and @xmath17 .
the other two high - energy bands are split with a scale of @xmath1 . in the ahe state ,
band gaps are opened at both @xmath16 and @xmath17 due to the order parameter @xmath41 , @xmath49 . in the me state
, there are also direct band gaps at @xmath16 and @xmath17 , @xmath50 .
but the center of the direct gaps shift in different directions of energy for @xmath16 and @xmath17 . when @xmath51 , an indirect gap opens . for illustration purposes , we take @xmath42 and @xmath52 here , which are slightly bigger than realistic values.,title="fig : " ] we proceed to examine the properties of me and ahe states from band structures , density of states , and topological structures .
the calculational details are provided in appendixes [ sec : nummethod ] and [ sec : addresults ] .
the bandstructures of these two states are shown in fig .
[ fig : bandstructure ] . for the ahe state
, any finite order parameter @xmath53 will open up a gap @xmath49 . as indicated by the crossing behaviors of the edge modes
, it has a quantized hall conductance at zero magnetic field .
we verify this by calculating the chern numbers for the four bands , which are @xmath54 in sequence of energy for @xmath55 [ for @xmath56 , the chern numbers are @xmath57 instead ] .
this gives a quantized hall conductance @xmath58 .
while there is a direct gap at @xmath16 ( or @xmath17 ) for infinitesimal @xmath41 , the me state has an indirect gap ( see the bottom panel of fig .
[ fig : bandstructure ] ) only for @xmath59 . for small @xmath41
, one obtains a semi - metal .
the insulating me state is not accessible within the two - band low - energy approximation . as shown in fig .
[ fig : bandstructure ] , the edge modes do not cross ; so the me state does not carry a net hall current at zero field .
this is again verified by the chern number calculations .
the density of states ( dos ) of the me state is shown in fig .
[ fig : dos_me ] . the vanishing dos near @xmath60
is associated with the indirect bandgap .
we expect that the density of states of the me state at the edge of the gap becomes much sharper because the excitonic effects are not included in our calculation .
there are two jumps of dos at high energies ( @xmath61 and @xmath62 in fig .
[ fig : dos_me ] ) .
this feature is associated with the direct band gap which is the order of @xmath1 and is a signature of the me state . and @xmath52 .
the zero - field dos has been multiplied by a factor 5 .
the magnetic field strength is specified by the flux per unit cell @xmath63 , where @xmath64 is an integer , proportional to the lattice size . ] .
@xmath65 , and @xmath64 has been chosen as @xmath66 , @xmath67 , @xmath68 , @xmath69 . ]
we now examine the property of the me state in a finite magnetic field .
the density of states is shown in fig .
[ fig : dos_me ] . unlike the ahe state , where the lowest landau level is pinned to the edge of the gap , @xcite the lowest landau level(s ) shifts up in energy , leading to a larger gap .
the magnetic field dependence of the gap is shown in fig .
[ fig : gapvsb ] , and agrees with experiments when extrapolated to low fields of the experiments .
in summary , we have shown that a magneto - electric state , with loop currents ordered in each layer as in haldane s phase but with odd - parity combination of two layers , is the ground state of interacting bilayer graphene .
it has a gap which increases monotonically with the applied magnetic field , which is in agreement with the experimental observations on a gapped state in bilayer graphene .
we have also considered the possibility of the actual realization of the haldane ahe state in a single layer graphene by our mean - field procedure . again
the nearest - neighbor interactions are ineffective and need a very large value to get any change in symmetry , but a next - nearest - neighbor interaction @xmath15 larger than about 0.1@xmath70 leads to the haldane state .
this requires a dielectric constant to be smaller than about 10 , which appears to be close to what experiments report .
we discuss further experiments which test the applicability of our proposed me state , to differentiate from other proposed states from weak - coupling .
( 1 ) the me state is characterized by not only an indirect gap ( across @xmath16 and @xmath17 points ) , which is related to the observed gap @xmath71 mev , but also a direct gap ( at @xmath16 or @xmath17 point ) , proportional to the loop current order parameter , of the order of the interlayer hopping .
this is different from other weak - coupling proposals where only a low - energy gap exists .
the direct gap leads to sharp features at high energies , such as the jumps in the density of states around @xmath1 , as shown in fig .
[ fig : dos_me ] .
we suggest optical absorption measurements on high mobility , dual - gated suspended bilayer graphene samples to check these features .
( we note , however , previous optical absorption measurements of bilayer graphene on a substrate , which do not realize an insulating state , have not observed a direct band gap at the nodes .
@xcite ) ( 2 ) the me state breaks not only time - reversal symmetry through loop current patterns in each layer , but also inversion symmetry across the layers .
the former leads to a monotonically increasing gap in a magnetic field similar to the ahe state which has even parity across the layers .
we note that an inversion symmetry - breaking term , such as an electric field perpendicular to the layers , introduces a linear coupling of ahe and me order parameters ( @xmath72 ) .
this promotes the ahe state when the me order parameter is finite .
the two - terminal probe indeed showed a finite conductance @xmath73 when a finite perpendicular electric field is applied , @xcite which may be interpreted as a signature of the ahe state .
we suggest a four - terminal experiment to verify this mechanism , and therefore , provide further proof for the existence of the me state at charge neutrality .
we acknowledge extensive discussions with c. n. lau .
l.z . also acknowledges discussions with d .-
this research was partially supported by nsf grant no .
dmr-0906530 ( l.z . and c.m.v . ) , and ucr initial complement(v.a . ) .
for the nearest - neighbor interaction @xmath74 , the generated loop currents flow around the perimeter of a hexagonal cell with net flux .
if we allow flux ( positive or negative ) in any unit cell , we run into the problem of frustration of an ising model in a triangular lattice .
this in general is not a possible ground state because unlike with spins , alternate arrangements with lower energy are possible .
the simplest is a @xmath75 structure of hexagons depicted in fig .
[ fig : kekule ] in which a hexagonal cell with zero net flux is surrounded by six hexagonal cells , each with a kekule pattern of currents in the links with alternate net positive and negative flux .
this state breaks the translational symmetry while the unit cell is enlarged to include @xmath76 unit cells of a translational invariant honeycomb lattice . .
only one layer is shown .
the red , blue and white colors represent the net positive ( clockwise ) , negative ( anticlockwise ) , and zero fluxes in a hexagonal cell . ]
we assume that the flux pattern in the top layer follows fig .
[ fig : kekule ] , which takes a sequence ( @xmath77,@xmath78,@xmath79 ) from left to right .
the bottom layer could follow the same pattern , or takes another pattern ( @xmath78,@xmath77,@xmath79 )
. therefore , there are two types of ordered loop current states ( the other combinations are equivalent to either of these two states by rotation ) .
we carry out a mean - field calculation on the ground - state energy for these states , and find that the states with different patterns for two layers are lower in energy .
this lower energy state is allowed only if @xmath80 , i.e. , be comparable to the bandwidth .
the reason is that such a state does not have a nesting " periodicity and therefore does not use any singularity in the joint density of states .
the mean - field hamiltonian of the ordered loop current states can be readily diagonalized for each momentum point ( @xmath81,@xmath82 )
. however , to examine their properties under an applied magnetic field , as well as the topological properties such as the edge modes , we also carry out a real - space calculation .
-@xmath83 axis specified in the figure , the lattice can be represented by a @xmath84 rectangle lattice with four atoms in each lattice site . ] following a common numerical practice for honeycomb lattices , we describe the bilayer graphene by a @xmath84 lattice , with four atoms on each lattice site ( see fig . [
fig : lattice ] ) .
the real - space coordinations of each atom ( and the momentum ) in the orthogonal @xmath85-@xmath83 axis can be easily obtained from this `` deformed''-lattice representation .
the effect of an external magnetic field can be captured by imposing a phase to the hopping term to each bond @xmath86 .
we adopt the landau gauge @xmath87 to take advantage of the translational invariance along the @xmath83 direction . for the periodic boundary condition
, we commonly take the flux due to the external magnetic field in a unit cell to be @xmath88 , where @xmath64 is an integer proportional to @xmath89 .
@xmath90 is equivalent to a magnetic field strength @xmath91 for graphene systems . in general
, we choose periodic boundary conditions along both directions ( a torus ) .
this provides a verification of our momentum - space calculations for the zero field . for the edge modes calculation
, we choose an open boundary along the @xmath85 axis , which corresponds to a cylinder with the zigzag edge .
in fig . [ fig : dos ] , we show the dos for the me state with various values of the order parameter @xmath41 .
we also show the result for the non - interacting case as well . for @xmath92 ,
the dos at @xmath60 is finite , which is due to the parabolic dispersion @xmath93 .
there is a sudden jump of dos at @xmath94 , which is the energy scale of the gap between the top two valence bands ( bottom conduction bands ) .
when @xmath41 is small , the conduction band at the @xmath16 point and the valence band at the @xmath17 point have overlaps in energy .
therefore , it remains a semimetal state .
a full gapped state happens when @xmath95 . in the main paper
, we have shown the topological properties of the ahe and me states by showing their edge modes , calculated by assuming a periodic boundary condition along the @xmath83 axis and an open boundary condition along the @xmath85 axis .
judging from the crossed edge modes , we learn the ahe state indeed has a finite hall conductance at zero field .
similarly , the me state does not have any hall effect at zero field .
this topological property can also be checked from the chern numbers . the chern number for a given band is defined as @xmath96 where @xmath97 is the wavefunction for the @xmath98-th band .
the hall conductance is given by @xmath99 .
we calculate the chern numbers in a momentum - space algorithm , by summing the berry curvature in small areas @xmath100 . for the ahe state
, any finite order parameter @xmath41 will open up a gap .
we find the chern numbers for the four bands are @xmath54 in sequence of energy , i.e , the valence and conduction bands away from the fermi energy do not carry any chern numbers . however , when @xmath101 , the chern numbers become @xmath57 instead . in either case ,
a finite hall conductance is a robust feature . for the me state
, an indirect gap is open only for a finite order parameter @xmath102 .
however , as long as the four bands do not touch each other ( having direct band gaps ) , the chern number calculation is robust .
we find the chern numbers for the four bands of the me state are all zero , indicating a non - quantum hall state .
when @xmath103 , we find the chern numbers become @xmath104 . in this limit , each layer is a haldane state while @xmath1 only acts as a perturbation , slightly splitting the states from two layers .
still , there is no hall effect near the fermi energy . in the main text ,
we have shown the edge modes for the ahe and me states at zero magnetic field . for completeness
, we also show the edge modes in a finite external magnetic field , in fig .
[ fig : edgemodeb ] .
we concentrate here on the experiments reported in refs . and because the four other references , refs . , have similar results but with a higher conductance in the gap .
one group ( ref . ) obtains rather different results which might be consistent with the gapless nematic order .
the origin of the difference from those of the other four groups is not clear .
e. a. henriksen , z. jiang , l .- c .
tung , m. e. schwartz , m. takita , y .- j .
wang , p. kim , and h. l. stormer , phys .
lett . * 100 * , 087403 ( 2008 ) ; l. m. zhang , z. q. li , d. n. basov , m. m. fogler , z. hao , and m. c. martin , phys .
b * 78 * , 235408 ( 2008 ) ; z. q. li , e. a. henriksen , z. jiang , z. hao , m. c. martin , p. kim , h. l. stormer , and d. n. basov , phys .
lett . * 102 * , 037403 ( 2009 ) ; a. b. kuzmenko , e. van heumen , d. van der marel , p. lerch , p. blake , k. s. novoselov , and a. k. geim , phys .
b * 79 * , 115441 ( 2009 ) .
|
while single - layer graphene shows extraordinary phenomena which are stable against electronic interactions , the non - interacting state of bilayer graphene is unstable to infinitesimal interactions leading to one of many possible exotic states .
indeed a gapped state is found in experiments but none of the states proposed so far can provide full accounts of its properties . here
we show that a _ magnetoelectric _ ( me ) state is consistent with the experimental observations .
this state breaks time - reversal symmetry through a pair of spontaneously generated current loops in each layer , and has odd - parity with respect to the two layers .
we also suggest further experiments to check whether the me state is indeed the gapped state found in experiments .
|
over an year ago , we studied an ising model on a cubic lattice @xcite by use of the @xmath0 expansion @xcite .
the study is based upon the high temperature expansion and the resulting series rewritten in the inverse of a mass squared @xmath1 is transformed by the @xmath0-expansion which played a crucial role in the estimation of critical quantities . however the foundamentals of the estimation protocol , which will be explicitly stated in the next section , have not been thoroughly investigated so far .
this paper deals with the issue in enough depth via the study of ising models at one- and two - dimensions where the models are solved and high temperature expansion is available to very large orders .
another purpose of this paper is to improve the estimation protocol of critical quantities employed in ref .
we will give a renewed procedure and confirm it works better in the low dimensional models , where the estimation of @xmath2 , @xmath3 and @xmath4 will be presented on the square lattice . the new protocol
would be applicable to high dimensional models and even field theoretic models . before the application to these more interesting models , it needs a detailed test on well understood models .
the ising model we consider is a pure one without applied forces , specified by the action@xmath5 where @xmath6 denotes the inverse temperature and @xmath7 means the pair of nearest neighbor sites .
it is well known that a one - dimensional model is in the one phase and the square lattice model undergoes a second order phase transition at @xmath8 .
one often discusses critical behaviors of thermodynamic quantities in terms of @xmath9 @xcite where @xmath4 stands for the critical temperature . in the high temperature region , in contrast , thermodynamic quantities are expanded in @xmath6 . as the temperature approaches from a hot region to the vicinity of the critical point , the effective parameter changes from @xmath6 to @xmath10 .
we feel that this exchange in the basic descriptive parameter is somewhat strange .
there exists , however , a sole parameter effective in the high - temperature phase .
we note that the second order transition is activated by the divergence of the correlation length .
the physics is controlled by correlation length @xmath11 .
the critical point is specified as the point where @xmath11 is divergent and this means that , irrespective of specific value of @xmath4 which is of non - universal , the critical point is universally specified by the point of @xmath12 or the massless point @xmath13 where @xmath1 should be understood as suitably related to @xmath11 .
if one describes the system in @xmath1 , even the temperature is a function of @xmath1 ( one can obtain @xmath14 by the inversion of @xmath15 ) and the critical temperature is given by the limit @xmath16 the high - temperature expansion corresponds to the @xmath17 expansion . from the large mass side , we access the critical point with the crucial help of the @xmath0-expansion @xcite . by making use of the fact that the critical point is at the massless point , we will explore a new way of handling high - temperature expansion .
the paper is organized as follows : in sec .
2 , the @xmath0-expansion is reviewed .
we posit a base on which the estimation protocol is constructed . in sec . 3
, a one - dimensional case is taken up to analytically study the @xmath0-expansion to arbitrary orders . in sec .
4 , a two - dimensional model on the square lattice is numerically revisited . the parametric extension proposed in ref .
@xcite is re - formulated by the linear differential equation ( lde ) with constant coefficients .
the spectrum of exponents in the series expansion in the vicinity of the critical point is given by the spectrum of roots of the characteristic equation of the lde satisfied by the thermodynamic quantities .
the critical temperature and the amplitudes appear as the integration constants . studies to quite large orders tell us that the computational base under the @xmath0-expansion is numerically confirmed .
then , under a new protocol to be proposed , we demonstrate that the estimation becomes more accurate . in the final section ,
we give concluding remarks .
the @xmath0-expansion originates in the dilatation of the scaling region . with the dilatation rate @xmath18 @xmath19 , the change of the parameter from @xmath1 to @xmath20 by @xmath21 in a given thermodynamic quantity @xmath22 induces a dilatation transform as follows : consider the large mass expansion truncated at order @xmath23 , @xmath24
then , truncate the expansion in @xmath0 in @xmath25 such that the sum of the expansion order of @xmath0 and the order of @xmath26 is less than or equal to @xmath23 but not over @xmath23 .
this specific truncation rule yields @xmath27 then , taking the limit @xmath28 which means formally the infinite rate dilatation , we obtain @xmath29 where @xmath30 denotes the binomial coefficient @xmath31 thus , the results of the @xmath0 expansion denoted by @xmath32 $ ] or @xmath33 are obtained as a function of @xmath34 , @xmath35=\bar\omega_{n}(t)=\sum_{n=0}^{n } c_{n , n}a_{n}t^n.\ ] ] on the other hand , for a general power of @xmath1 we use the analytic extension of the @xmath36 function giving @xmath37=c_{n,-\alpha}t^{-\alpha}=\frac{\gamma(n+1)}{\gamma(-\alpha+1)\gamma(n+\alpha+1)}t^{-\alpha}. \label{delta2}\ ] ] one can deduce two useful results : by setting @xmath38 , we have @xmath37=0,\quad \alpha=1,2,3,\cdots . \label{regular}\ ] ] by the expansion of ( [ delta2 ] ) in @xmath39 , it follows , for example at order @xmath40 , @xmath41=-\log t-\sum_{n=1}^{n}\frac{1}{n}. \label{deltalog}\ ] ] by comparison of both sides of ( [ delta2 ] ) at order @xmath42 , one has @xmath43 $ ] .
further , if one sets @xmath44 with infinitesimal @xmath45 , expansion in @xmath45 provides a similar formula for @xmath46 times logarithms . assuming the power - like behavior of thermodynamic function @xmath22 in the scaling region ,
let us explain the effect of @xmath0-expansion .
series here is @xmath17 expansion truncated at order @xmath23 .
then , our task is to obtain the critical information of the supposed behavior @xmath47 here , @xmath48 and @xmath49 .
there are other types of critical behaviors such that @xmath50 @xmath51 and @xmath52 .
these are typical in the study of phase transition as these two represent the divergences of thermodynamic functions .
let us explain one of the advantages in the @xmath0-expansion , taking the scaling form ( [ scaling1 ] ) as an example .
the construction via @xmath0-expansion gives @xmath53=\omega_{0}+c_{1}c_{n,-\lambda_{1}}t^{-\lambda_{1}}+c_{2}c_{n,-\lambda_{2}}t^{-\lambda_{2}}+\cdots .
\label{delta_scale}\ ] ] note that order @xmath23 of large mass expansion is included in the coefficients in ( [ delta_scale ] ) .
the first term is most important and the rest acts as the corrections to the first one . in the limit @xmath54 ,
respective terms of the corrections tend to vanish since @xmath55 this represents the scaling of the amplitude @xmath56 at large enough order .
thus , as @xmath57 , @xmath58 and the @xmath59th - order correction tends to zero for fixed @xmath34 .
this is quite convenient for the estimation of @xmath60 .
the above result ( [ scale_c ] ) suggests us that for @xmath61 where @xmath23 denotes the order of large @xmath1 expansion , @xmath62(t)=\omega_{0 } , \label{delta_limit}\ ] ] _ in some region @xmath63 of @xmath34_. as mentioned in the previous section , this manifests a major characteristic property of @xmath0 expansion and plays a foundamental role in estimating the critical quantities . though formal reasoning supports ( [ delta_limit ] ) , if we inspect the terms in the large mass expansion , we find @xmath64 which diverges to infinity as @xmath57 even when @xmath34 is small . though previous work @xcite provided confirmed results on the point , the study is limited to @xmath65th order and it is as yet non - trivial whether the fundamental result ( [ delta_limit ] ) remains valid to large enough orders .
we shall deeply investigate the point in two solvable models in the following sections .
in the one - dimensional ising model , it is well known that the exact correlation length @xmath11 is given by @xmath66 .
then , the mass squared @xmath1 defined by the zero momentum limit of correlation function is obtained by @xmath67 .
we note that @xmath1 and @xmath11 do not obey the relation @xmath68 but satisfy @xmath69 ( @xmath70 ) @xcite and inversion gives @xmath71 where @xmath72 the function @xmath73 has a branch point at @xmath74 , and the large mass expansion , @xmath75 is valid within the convergence region @xmath76 .
here we have introduced the notation @xmath77 , which stands for the inverse temperature at small @xmath78 . at large @xmath78
, we make a division such as @xmath79 and obtain the expansion , @xmath80 where the notation @xmath81 means the expansion of @xmath6 near the critical point and @xmath82 denotes the regular part given by @xmath83 since @xmath6 shows logarithmic divergence in the @xmath84 limit , @xmath85 .
we call the behavior ( [ critical1 ] ) a critical behavior . from ( [ deltalog ] ) , we find @xmath86=\frac{1}{4}\big(\log(4t)+\sum_{k=1}^{n}\frac{1}{k}\big)+d_{n}[r]=:\bar\beta_{n<}.\ ] ] now , we consider @xmath87=\bar\beta_{n>}$ ] where @xmath88 stands for the truncation of @xmath77 to @xmath23th order such that @xmath89 .
we then have @xmath90 the above integral representation shows that the region where @xmath91 remains valid in the @xmath57 limit is @xmath92 ( @xmath93 ) . note that the center of the region @xmath94 is @xmath95 .
this indicates that the correct expansion should be inside region @xmath63 and not at @xmath96 on the boundary of @xmath63 . from ( [ beta_closed ] ) ,
we obtain the expansion around the center @xmath95 : @xmath97 from the comparison of ( [ critical1 ] ) and ( [ beta_expansion ] ) , we find @xmath98=\sum_{n = n+1}^{\infty}\frac{(1 - 4t)^n}{n}.\ ] ] thus we can not use @xmath99=d_{n}[m^n]=0\,(n=1,2,3,\cdots)$ ] inside the infinite sum , though the terms of the form @xmath100 are absent in the right - hand - side ( the expansion of the right - hand - side around @xmath101 is impossible ) .
the result of @xmath0-expansion reads @xmath102=o(\epsilon^{n+1})$ ] where @xmath103 and the correction to the logarithmic scaling shows rapid suppression at large @xmath23 in the neighborhood of @xmath104 . in other words , as @xmath23 tends to infinity , @xmath102 $ ] becomes negligible and @xmath105 .
thus , we are convinced that the divergence of the coefficient in the @xmath57 reflects that the point @xmath96 is just on the boundary of the convergence region @xmath94 .
the @xmath57 limit of @xmath91 diverges for any @xmath106 , and this is due to the presence of the logarithmic singularity in the original function as @xmath84 .
this case shows a variant of ( [ delta_limit ] ) in the case of a logarithmically divergent limit .
we are able to capture the behavior of the logarithmic divergence in @xmath91 via numerical analysis .
see fig . 1 where @xmath91 is depicted at @xmath107 .
= \bar\beta_{10>}$ ] against @xmath108 .
the broken gray line indicates the exact asymptotic behavior @xmath109 . ] by study of the derivative @xmath110 , we find further evidence of logarithmic behavior and the coefficient of the log , @xmath111 : we find directly from ( [ beta_closed ] ) by the differentiation with respect to @xmath112 , @xmath113 and the exact validity of ( [ delta_limit ] ) is stated as @xmath114 for convenience , we show the plots of @xmath115 at @xmath116 and @xmath117 in fig .
2 . at @xmath118 . the dashed gray line indicates the limit @xmath111 . at every orders , the unique stationary point gives the exact limit . ] in accord with ( [ 3_10 ] ) , the functions imply the step - like function , which is almost flat over the region @xmath94 . at finite @xmath23 ,
the unique stationary point is always realized at @xmath95 and gives an exact value of @xmath119 , which agrees with the coefficient of @xmath120 .
it is also instructive to note that the higher order derivatives satisfy @xmath121 @xmath122 .
in a one - dimensional ising model , the scaling behavior is simple and allows exact analysis . in the two - dimensional model , a complexity related to
the critical behaviors appears .
most of our investigation is based on the exact solution , and the model serves as a good testing ground for understanding large order aspects of @xmath0 expansion . the exact result on which our study based is a relation between the correlation length and inverse temperature @xcite , @xmath123 where @xmath11 stands for the correlation length extracted in the large separation limit of a two - point function along one of the two axes . from above , we obtain a well - known result : @xmath124 in addition we have @xmath125 where @xmath126 , and we find @xmath127 the mass squared @xmath1 corresponding to @xmath11 is given by the transformation @xmath67 @xcite . from the straightforward expansion @xmath128 we obtain @xmath129 and
arrive at @xmath130 where @xmath82 denotes the regular part . as well - known , the two - dimensional model has no confluent singularity , and the structure of the expansion is rather simple . in our analysis
, we do not intend to use the above simplicity .
we just assume the power - like expansion @xmath131 where @xmath132 . the large mass expansion of @xmath6 is obtained to an arbitrary large order from ( [ 2dxi ] ) . to the first several orders it reads @xmath133 we briefly comment on the analytic extension of original @xmath134 on the complex @xmath78-plane using pad
approximates : the poles and zeros line up side by side on the negative real axis .
the pole of the largest absolute value extends toward @xmath135 , and the pole of the smallest absolute value seems to be converging to @xmath136 . from conventional knowledge , this kind of set of poles and zeros indicates the cut .
the implied convergence radius of @xmath17 expansion is @xmath137 and this agrees with the simple study of the coefficient ratio . in the one - dimensional model , the convergence region of original @xmath17 expansion is @xmath138 and the @xmath0-expanded counterpart reads @xmath139 .
it is as if the region of convergence is shifted right by @xmath111 .
if @xmath34 is negative and small in the absolute value , @xmath140 diverges in the @xmath57 limit .
this may sound strange , since the series in powers of @xmath78 is effective around @xmath141 and then the expansion in @xmath34 would ordinarily include the origin within the convergence circle . before proceeding to estimate critical quantities in the two - dimensional model , we would like to elucidate the point in the square lattice . from ( [ 4_10 ] )
, we obtain @xmath142 at orders @xmath143th , @xmath144th and @xmath145th .
dashed line indicates @xmath4 .
( b)the plots of @xmath91 ( solid line ) and @xmath146 ( dashed gray line ) at @xmath147 . here
we note that @xmath148.,title="fig : " ] at orders @xmath143th , @xmath144th and @xmath145th .
dashed line indicates @xmath4 .
( b)the plots of @xmath91 ( solid line ) and @xmath146 ( dashed gray line ) at @xmath147 . here
we note that @xmath148.,title="fig : " ] we first show in fig .
3 the plots of @xmath91 at various orders .
as seen in the upper graph ( a ) , @xmath91 shows slow convergence to @xmath4 .
this is due to the presence of the corrections , especially the leading one @xmath149 .
actually , the simultaneous plot of @xmath91 and @xmath146 ( @xmath150 and ( [ delta2 ] ) are used ) at @xmath147 shown in fig .
3 ( b ) reveals that the two graphs are in excellent agreement .
thus , we find that @xmath91 precisely approximates @xmath151 in its simplest form and , as long as @xmath152 in the @xmath57 limit , @xmath153 in the region @xmath154 .
figure 4 shows the plots of @xmath155 at various orders .
the coefficient @xmath156 multiplied to @xmath112 is chosen from the asymptotic form of @xmath6 near the critical point @xmath157 ( see ( [ crit_2d_a ] ) ) , i. e. , the inverse of the critical exponent @xmath158 is used .
the value makes the leading correction , the term @xmath159 , be canceled between @xmath6 and @xmath160 .
thus the correction is highly suppressed in the combination . as a result , @xmath155 shows a clear convergence trend over the whole region of @xmath154 . at @xmath161 , @xmath144 and @xmath145 .
dashed line indicates @xmath4 . ]
we can say that the region supporting ( [ delta_limit ] ) thus indeed exists for the functions of interest .
explicitly , the region is confirmed to be @xmath162 approximately .
the point is , as in a one - dimensional model , that the origin @xmath96 seems to be excluded from this region . to confirm it in a more conclusive ,
we compute @xmath163 at large orders @xmath164 and @xmath165 over the region @xmath166 .
the result is plotted in fig .
5 . the effective region where @xmath91 takes a finite value is limited between @xmath167 and @xmath168 .
in other regions , we find @xmath169 .
( dotted ) at @xmath164 ( black ) and @xmath165 ( gray ) as functions of @xmath34 . the filled blobs signal points @xmath170 and @xmath171 at which @xmath172 . ]
the above feature is investigated from another angle by using pad approximants .
since @xmath91 and the derivatives converge respectively to @xmath4 and zero , the diagonal approximants denoted as @xmath173 $ ] ( @xmath174 ) are the most appropriate choice in the pad table .
6 ( a)-(b ) show the poles and zeros of the @xmath173 $ ] element of @xmath91 and @xmath175 at @xmath176 , @xmath177 and @xmath178 . from fig .
6 , we find , in spite of coefficients of @xmath179 and @xmath180 being dependent on the order , that both shapes of distributed poles and zeros seem to converge to circular ones , though the one for poles is chipped .
the shape of the pole set is apparently recognized as the circle even at rather low orders and the convergence is not slow .
in contrast , the shape of the zero set changes very slowly for @xmath91 though the reason behind it is not clear to us . at @xmath164 for @xmath91 ,
the points at the left end of pole sets have the coordinate @xmath181 , and the points at the right chipped end have @xmath182 . as for @xmath175 at the same order , the points at the left end of pole sets have the coordinate @xmath183 and the points at the right end have @xmath184 .
the respective norms are @xmath185 and @xmath186 for @xmath91 and @xmath187 and @xmath188 for @xmath175 .
the deflection angles for the upper right end points are , respectively , @xmath189 and @xmath190 for @xmath91 and @xmath175 . from these results and those at lower orders
, it is strongly suggested that in the infinite order the both sets of poles for @xmath91 and @xmath175 are dense and form the chipped circle with the radius @xmath137 , the center being at the origin and upper end points having deflection angle @xmath191 . both zero - point sets in the @xmath57 limit form dense sets of the circle with radius @xmath137 and center at @xmath192 .
then , if a @xmath34-series obtained by the @xmath0-expansion would be a series expanded around @xmath96 , it must be effective `` inside '' of the chipped pole - circle and not effective for @xmath193>1/8 $ ] .
but it is not actually the case as manifest in fig .
, we conclude that effective region of @xmath0-expansion is shown by the zero - circle , not by the pole - circle .
the origin @xmath96 is actually at the boundary of the effective region and this is in accord with the fact that the coefficient of @xmath194 diverges as @xmath57 ; that is , the @xmath0-expansion does not provide expansion around @xmath96 . in the complex @xmath34 plane .
stars , crosses and dots in black for poles at @xmath176 , @xmath177 and @xmath178 respectively .
green triangles , red diamonds and small blue circles for zeros at @xmath176 , @xmath177 and @xmath178 respectively .
( b ) same plots for @xmath195,title="fig : " ] in the complex @xmath34 plane .
stars , crosses and dots in black for poles at @xmath176 , @xmath177 and @xmath178 respectively .
green triangles , red diamonds and small blue circles for zeros at @xmath176 , @xmath177 and @xmath178 respectively .
( b ) same plots for @xmath195,title="fig : " ] from the argument so far , though it is hard to analytically proves the convergence in the large @xmath23 limit , we can say that there exists enough evidence of ( [ delta_limit ] ) , stated in this case as @xmath196 similarly , we conclude that @xmath197 and hence for arbitrary @xmath198 and parameters @xmath199 , @xmath200 an example of the above kind of combinations is @xmath201 considered before corresponding @xmath202 and @xmath203 .
the speed of convergence as @xmath23 grows depends on the value of @xmath199 . from ( [ crit_2d_a ] ) , we find that the suitable set is @xmath204 but the spectrum of the true exponents is the target itself in our study and not allowed to be used as an input . in the next section we discuss how to use ( [ fundamental_theorem ] ) for the estimation of critical temperature and exponents . figure 3(b ) clearly exhibits that @xmath91 recovers the scaling behavior .
this provides us the possibility of estimating critical quantities from @xmath91 effective at small @xmath34 . in this subsection
, we explicitly carry out estimations of critical quantities @xmath4 and @xmath2 .
in fact , @xmath4 estimation is able to carry through with the protocol presented in ref . @xcite . though we shall revise the protocol later , we first present the protocol and the result below .
we start with the scaling form written in @xmath78 , @xmath205 where @xmath82 denotes the analytic part ( see ( [ crit_2d ] ) ) .
consider the simplest ansatz to the first correction @xmath206 . from the argument on the @xmath0-expansion of regular series presented in the one - dimensional model , we neglect
@xmath102 $ ] and employe the simple ansatz , @xmath207 the terms have exponents @xmath208 , @xmath209 and the above ansatz satisfies second order lde @xmath210 , and then provides after the integration over @xmath120 @xmath211\bar\beta_{n<}=\beta_{c}. \label{lde1}\ ] ] the inverse critical temperature appears as the integration constant .
then , we assume the basic result of @xmath0-expansion which states @xmath212\bar\beta_{n>}=\beta_{c}\quad t\in i. \label{base1}\ ] ] note that @xmath91 effective at small @xmath34 is used in the place of @xmath151 . now , see fig . 7 where @xmath91 and its derivatives are plotted at @xmath147 .
we confirm from fig .
3(a ) and fig . 7 that approximate scaling behaviors for @xmath213 and @xmath214 can be observed in the region @xmath215 .
( we see from ( [ ansatz ] ) that @xmath216 .
the decreasing trend of the first derivative is clearly exhibited in @xmath217 ) and its derivatives @xmath218 for @xmath219 at @xmath147 .
the thick gray line represents @xmath8 .
though the scaling of @xmath220 has just begun to be observed , other derivatives exhibit scaling after the last bumps .
this allows us , at @xmath65th order , to use a three - parameter ansatz . a four- or more multiparameter ansatz needs higher order terms above @xmath65th .
] then , we make use of the approximate case of ( [ base1 ] ) without the @xmath57 limit ; that is , we identify the value of @xmath221\bar\beta_{n>}$ ] at a stationary point @xmath222 in the plateau ( see the plots in fig .
the result can be written as @xmath211\bar\beta_{n>}\big |_{t = t^{*}}=\beta_{c}. \label{recipe1}\ ] ] then , the task remaining is to find most reliable or optimal set of @xmath34 and @xmath223 , which is expected to be close to @xmath224 .
for this purpose , the extension of the principle of minimum sensitivity ( pms ) due to stevenson @xcite was used .
it suggests that the optimal set of @xmath223 and @xmath34 is determined by the following simultaneous equations @xmath225\bar\beta_{n>}^{(1 ) } & = & 0 \label{pms1}\\ \,\big[1+p_{1}^{-1}\frac{d}{d\log t}\big]\bar\beta_{n>}^{(2 ) } & \sim&0 .
\label{pms2}\end{aligned}\ ] ] here the symbol @xmath226 means the left - hand side must be vanishing or a local minimum in the absolute value
. some comments would be in order : in general , the solution of ( [ pms1 ] ) and ( [ pms2 ] ) is not unique at a given order , and among those solutions , the best one must be selected somehow .
let us explain this issue by focusing on typical order cases , @xmath227 . from ( [ pms1 ] ) , we obtain @xmath223 as a function of a stationary point @xmath34 as @xmath228 .
for instance , at @xmath229 , one has @xmath230 substitution of @xmath231 into the left - hand - side of ( [ pms2 ] ) gives @xmath232\bar\beta_{n>}^{(2)}(t)$ ] which is plotted in fig . 8 .
\bar\beta_{n>}^{(2)}$ ] at @xmath233 .
the dotted vertical lines indicate the singularities . ]
the optimal solution is determined in the following way : at @xmath229 we have just one solution at the cross point and this gives optimal @xmath234 . at @xmath107
there is one more solution depicted by the black blob .
the peak between the two zeros shows the turning point to the scaling and therefore the blob point is plausible to be selected as the optimal one , giving @xmath235 .
similarly at @xmath236 , the local minimum represented by blob point corresponds to the optimal one and @xmath237 .
then we obtain the optimal estimation of @xmath223 by @xmath238 and the estimation of @xmath4 is given by ( [ recipe1 ] ) .
we use the term `` proper solution '' for those depicted by the filled blobs at @xmath107 and @xmath239 . just from these orders , the proper approximation sets in . at @xmath229 ,
the set @xmath240 is optimal but not proper
. we can also carry out multi - parameter or @xmath241th order lde analysis from @xmath242\bar\beta_{n>}=\beta_{c}. \label{cond}\ ] ] extending the pms idea , we require that @xmath91 satisfies at some @xmath34 that @xmath243\bar\beta_{n>}^{(1)}&= & 0,\label{cond6}\\ \cdots & \cdots & \nonumber\\ \prod_{i=1}^{k}\big[1+p_{i}^{-1}\frac{d}{d\log t}\big]\bar\beta_{n>}^{(k)}&= & 0,\label{cond7}\nonumber\\ \prod_{i=1}^{k}\big[1+p_{i}^{-1}\frac{d}{d\log t}\big]\bar\beta_{n>}^{(k+1)}&\sim & 0.\label{cond8}\end{aligned}\ ] ] the first @xmath241 condition allows a solution @xmath223 , @xmath244 , @xmath245 , @xmath246 as functions of @xmath34 which specifies the estimation point .
then , the last condition ( [ cond8 ] ) gives the optimal or proper solution for @xmath222 , and the substitution gives the estimate @xmath247 .
estimation of @xmath4 can then be done with ( [ cond ] ) .
lccccccccc & @xmath248 & @xmath249 & @xmath143 & @xmath65 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath177 + @xmath254 & 0.43451427 & 0.43768742 & 0.43905467 & 0.43952324 & 0.43989324 & 0.44004633 & 0.44020306 & 0.44027202 & 0.44035468 + @xmath255 & 0.44270440@xmath256 & 0.43995506 & 0.44045465 & 0.44057056 & 0.44062720 & 0.44064773 & 0.44066261 & 0.44066888 & 0.44067449 + @xmath257 & & & 0.44064208@xmath256 & 0.44065581 & 0.44067626 & 0.44068146 & 0.44068414 & 0.44068514 & 0.44068583 + @xmath258 & & & & & 0.44068347 & 0.44068540 & 0.44068627 & 0.44068653 & 0.44068666 + @xmath259 & & & & & & 0.44068684@xmath256 & 0.44068662 & 0.44068672 & 0.44068677 + in general , as many terms are included into the ansatz , the accuracy of estimate is increased , though the onset of proper estimation is delayed to higher orders .
see fthe numerical results summarized in table i. an approximate assessment at which order @xmath91 begins to show that scaling is found from the plot of the function and its derivatives .
for instance , at @xmath147 , we notice that , though the seventh - order derivative which enters the job at a three - parameter ansatz shows rough scaling , all the derivatives can be said to be exhibiting the scaling behaviors .
it is also notable that from the plot one can guess the value of proper solution @xmath260 . on these grounds
, one may expect that the three parameter ansatz is available at @xmath147 .
more precisely , we consider that reliable estimation begins with the emergence of the proper stationary point . and , at a given order of large @xmath1 expansion , the best approximation of @xmath4 is obtained in an ansatz with @xmath241 parameters where @xmath241 denotes the maximal number of parameters providing a proper solution @xmath260 .
this is the summary of the protocol used in ref .
@xcite .
the estimate of an exponent @xmath261 by @xmath262 accepts slightly less accuracy .
for instance at @xmath147 , a three - parameter ansatz meets just proper case and yields @xmath263 , @xmath264 . at @xmath265 , a proper six - parameter ansatz yields @xmath266
, @xmath267 , and @xmath268 . the value of @xmath244 at @xmath177th order proves that the next order term is @xmath269 and not @xmath270 , which is contained in @xmath82 ( see ( [ crit_2d ] ) ) .
the value of @xmath271 at @xmath265 indicates the next - to - the - next order is @xmath272 .
all of these results are in accord with ( [ regular ] ) , ( [ beta_scale1 ] ) , and ( [ beta_scale2 ] ) . in the previous protocol , critical exponents
are estimated in the relation with @xmath4 . here
we try to estimate the leading exponent @xmath223 in the manner independent of @xmath4 estimation .
the point is simple .
it just suffices to differentiate @xmath273 by @xmath120 to remove @xmath4 and then construct a quantity which converges to @xmath223 in the @xmath274 limit .
we will demonstrate that the ratio of the derivatives @xmath275 is convenient for the aim and the estimation accuracy will be improved .
one remark is in order here . in working @xmath73
, we assumed that the regular part @xmath82 is negligible and the numerical results on the exponents @xmath244 and @xmath271 at high orders support that it is the case .
however , in @xmath276 , a regular part plays a role of effective corrections as we can see below : writing @xmath277 , we find @xmath278 the first part implies the expansion result comes solely from non - analytic part of @xmath73 , while the last part involves @xmath82 .
we find that @xmath82 dependent part has expansions in the fractional powers of @xmath78 . on the other hand ,
actually , the first part is a regular series , and the @xmath0-expansion would make it be negligible except for the first term @xmath209 .
thus @xmath82 plays a major role in @xmath276 . neglecting higher order corrections
, @xmath276 satisfies @xmath279\bar f_{\beta n<}= 0,\label{dif1}\ ] ] where @xmath280 , indicating the constant term @xmath209 in ( [ ratio2d ] ) .
another exponent @xmath281 @xmath282 stands for the power of terms involved in the second part of ( [ ratio2d ] ) . here
, @xmath283 stands for the number of terms included in @xmath284 . integration with respect to @xmath34 and the replacement @xmath285
would then give @xmath286\bar f_{\beta n>}= -p_{1}. \label{dif2}\\\ ] ] this is the equation we start with .
conditions to estimate @xmath209 are obtained by the differentiation , giving @xmath287\bar f_{\beta n>}^{(1)}&= & 0 , \label{cond9}\\ \prod_{i=2}^{k_{\nu}}\big[1+q_{i}^{-1}t\frac{d}{dt}\big]\bar f_{\beta n>}^{(2)}&= & 0,\label{cond10}\\ \cdots & \cdots & \nonumber\\ \prod_{i=2}^{k_{\nu}}\big[1+q_{i}^{-1}t\frac{d}{dt}\big]\bar f_{\beta
n>}^{(k)}&\sim & 0.\label{cond11}\end{aligned}\ ] ] the estimation details are same as those in the estimation of @xmath4 .
the only thing worth mentioning is that the series of @xmath288 becomes somewhat less - qualified compared to the series of @xmath213 itself .
this depletion is reflected in the derivatives of @xmath288 , which is depicted in fig .
9 . for the @xmath289-lde case
, a proper sequence appears at eighth order for even orders and @xmath239th for odd orders . for the @xmath156-lde sequence ,
a proper estimate begins from @xmath290th for even orders and @xmath291th for odd orders . for the 3-lde sequence
, we have @xmath292th and @xmath293th orders , and for the @xmath294-lde sequence , @xmath295th and @xmath296th orders .
the result of the estimate is summarized in table ii and depicted in fig .
10 . at @xmath147 .
the seventh - order derivative shows behavior far from scaling and the sixth - order derivative is just about setting in the scaling .
the thick gray line indicates @xmath297 . ] via @xmath6 and @xmath276 .
`` @xmath298 '' means the result from the @xmath241-parameter ansatz of @xmath81 .
similarly , `` @xmath299 '' means the result from the @xmath241-parameter ansatz of @xmath300 .
plot ( a ) shows the results via the one- and two - parameter ansatz and plot ( b ) via the three- and four - parameter ansatz .
the dotted lines indicate @xmath301 . both of ( a ) and ( b )
show that the estimations via @xmath302 are superior to those via @xmath303.,title="fig : " ] via @xmath6 and @xmath276 .
`` @xmath298 '' means the result from the @xmath241-parameter ansatz of @xmath81 .
similarly , `` @xmath299 '' means the result from the @xmath241-parameter ansatz of @xmath300 .
plot ( a ) shows the results via the one- and two - parameter ansatz and plot ( b ) via the three- and four - parameter ansatz .
the dotted lines indicate @xmath301 . both of ( a ) and ( b )
show that the estimations via @xmath302 are superior to those via @xmath303.,title="fig : " ] lcccccc & @xmath249 & @xmath143 & @xmath65 & @xmath250 & @xmath251 & @xmath252 + @xmath304 & 1.0072333 & 1.0039272 & 1.0027581 & 1.0018766 & 1.0014966 & 1.0011272 + @xmath255 & 0.9829895 & 0.9928665 & 0.9958010 & 0.9974954 & 0.9981953 & 0.9987589 + @xmath305 & 1.0051863@xmath256 & 1.0015675 & 1.0002212 & 1.0001129 & 1.0000743 & 1.0000465 + @xmath257 & & & 0.9986069 & 0.9994266 & 0.9996737 & 0.9998178 + @xmath306 & & & 1.0001608@xmath256 & 1.0000539 & 1.0000069 & 1.0000033 + @xmath258 & & & & 0.9997888 & 0.9998998 & 0.9999569 + @xmath307 & & & & & 1.0000048@xmath308 & 1.0000018 + in table ii and fig .
10 , the results of the @xmath223 estimate by the previous protocol in the 2-lde ansatz are also plotted . as a typical case ,
compare 2-lde results with the improved 2-lde ( including @xmath309 and @xmath310 ) results .
we see from table ii that the improved 2-lde results are superior to the original 2-and 3-lde results when those are available at a given order .
the only drawback in the former is the delay of the emergence of the proper sequence .
the correction exponents can be also estimated : at @xmath265 , a proper five - parameter ansatz is available and yields @xmath311 , @xmath312 , and @xmath313 .
this result serves as evidence that @xmath314 , @xmath315 , and @xmath316 .
note that the integer power is not detected .
ccccccc & @xmath249 & @xmath143 & @xmath65 & @xmath250 & @xmath251 & @xmath252 + @xmath255 & 0.44104211 & 0.44079548 & 0.44071637 & 0.44070377 & 0.44069666 & 0.44069360 + @xmath257 & & & 0.44069402 & 0.44068944 & 0.44068762 & 0.44068726
+ @xmath258 & & & 0.44069045@xmath256 & 0.44068811 & 0.44068702 & 0.44068689 + @xmath259 & & & & & 0.44068692 & 0.44068685 + by using the @xmath223-estimate obtained in this manner , the improved estimate of @xmath4 becomes possible as we can see below : the new protocol tells that , at a given order , one should use the ansatz of @xmath317 with maximal number of parameters , say @xmath241 .
then we bias @xmath4 estimation via @xmath318 with @xmath241-lde by the substitution of @xmath319 obtained in the @xmath283-lde estimation via @xmath276 , where the order @xmath283 stands for the available maximal order of lde .
for instance , at @xmath65th order a three - parameter ansatz is available ( see table i ) .
then , we can use @xmath305 proper estimate of @xmath223 via @xmath276 . on the other hand , at @xmath292th order , we deal with the same three - parameter ansatz with @xmath306 proper estimate of @xmath223 . at each orders , we substitute @xmath223 obtained in respective orders into the set of conditions ( [ cond6 ] ) - ( [ cond8 ] ) .
since the @xmath223 exponent is thus fixed , one can discard one condition .
the last condition ( [ cond8 ] ) involves the highest derivatives among them and suitable to be discarded . in this way
, we redo the estimation of @xmath4 , and the results are summarized in table iii and fig . 11 . in the naive and improved protocols .
`` @xmath320 '' means the naive result from the @xmath241-parameter ansatz of @xmath81 .
similarly , `` @xmath321 '' means the improved result from @xmath241-parameter ansatz of @xmath81 under the bias of estimated @xmath2 via @xmath276 .
the dotted line indicates @xmath4 .
for a three - parameter ansatz , non - proper results are also plotted . ] at @xmath65th order , for example , the naive estimation gives @xmath322 in @xmath323-lde . in the improved estimation
, we have @xmath324 in @xmath323-lde .
the improved result is more accurate than the naive one .
the accuracy improvement occurs to all orders we have examined to @xmath177th order .
we conclude that the new formula improves the accuracy of the @xmath4 estimate .
the exact and closed expression of the magnetic susceptibility defined by @xmath326 is not obtained yet . however
, its mathematical structure has been vigorously investigated for some years .
references @xcite provide good guidance .
it is now known that the behaviors of @xmath327 near the transition point read @xcite , @xmath328 where @xmath329 and @xmath330 use of the expansion of @xmath331 in @xmath1 , @xmath332 then gives the behavior of @xmath327 described in @xmath78 , @xmath333 compared with the critical behavior of @xmath73 , @xmath327 shows more complicated structure .
the spectrum of exponents comes in two families of @xmath334 and @xmath335 .
in addition the later family allows logarithms which are considered as the remnant of the exponents with double , triple , etc degeneracies .
the similar but more complicated structure is observed in the ratio function @xmath336 . though @xmath337 is more complicated than @xmath327 , its leading term is just a constant @xmath3 and convenient for its estimation .
we therefore work with @xmath337 .
first of all , we do not make use of the detailed result ( [ chi_scaling ] ) , since our purpose here is to assess the power of our approach to estimate the @xmath327-exponent @xmath338 under simple assumptions .
we start in a robust manner with the supposed power - like behavior with analytic back ground , @xmath339 where @xmath340 .
then , we have @xmath341 where @xmath342 and @xmath343 .
the lde to be satisfied by @xmath344 to the first @xmath241 expansion reads @xmath345\big[p_{1}+\frac{d}{d\log t}\big]\prod_{n=2}^{k}\big[r_{n}+\frac{d}{d\log t}\big]\bar f_{\chi n<}=0.\ ] ] integration over @xmath34 and the replacement @xmath346 then gives @xmath211\prod_{n=2}^{k}\big[1+r_{n}^{-1}\frac{d}{d\log t}\big]\bar f_{\chi n > } \sim \frac{\gamma}{2\nu}. \label{lde_fchi}\ ] ] the behaviors of @xmath347 @xmath348 at @xmath147 are depicted in fig .
12 . from the plots ,
we can understand roughly at which region of @xmath34 , the estimation should be carried out in a given order . at @xmath147 for example , all of @xmath347 from @xmath349 show approximate scaling around @xmath350 and @xmath260 would be close to that
. it would be interesting to see whether the @xmath0-expansion plays an expected role in this case , i. e. , whether @xmath351 recovers the true complicated critical behavior .
we have numerically studied both functions @xmath351 and @xmath352 at several @xmath23 and confirmed it is the case .
for instance , fig .
13 shows the plots of those at @xmath147 .
the two functions are almost degenerate over a region of @xmath34 ranging from @xmath353 to @xmath354 .
the agreement between them is confirmed also for the first and second derivatives at @xmath147 .
( @xmath355 ) at @xmath147 .
the thick gray line indicates @xmath356 . ] and @xmath352 at @xmath147 .
the dashed line indicates @xmath356 . ] in ref .
@xcite , an unbiased estimation of @xmath338 via large mass expansion was done .
however , we must say that the result is not as good in the accuracy . here
we note that the estimation of @xmath4 was improved under the bias of @xmath2 estimated order by order via @xmath276 .
this would apply to the @xmath338 estimation since the leading correction comes in @xmath357 .
we therefore employ the same procedure : we substitute @xmath358 estimated in the previous subsection into the left - hand - side of ( [ lde_fchi ] ) and extended pms conditions that follow from ( 4.28 ) to ( 4.30 ) under the replacement @xmath359 in them . the result of @xmath3 estimation to @xmath177th order is shown in table iv and fig . 14 .
it is apparent that the use of @xmath319 has made the accuracy better .
here we notice that the used value of @xmath319 is the one obtained in the @xmath289-lde for @xmath276 .
the @xmath360-lde with larger @xmath360 by which @xmath319 is supplied does not bring about any net improvement on @xmath3 . in the naive and improved protocols .
`` @xmath361 '' means the unbiased result from the @xmath241-parameter ansatz of @xmath362 .
similarly , `` @xmath363 '' means the improved ( biased ) result from the @xmath241-parameter ansatz of @xmath362 .. ] cccccccc & @xmath143 & @xmath65 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath177 + @xmath254 & 0.87663881 & 0.87601885 & 0.87575751 & 0.87552991 & 0.87542884 & 0.87531581 & 0.87526554 + @xmath255 & 0.87444212 & 0.87481600 & 0.87490090 & 0.87492284 & 0.87493726 & 0.87494082 & 0.87494692 + @xmath257 & & 0.87496338 & 0.87495516 & 0.87495917 & 0.87496394 & 0.87496794 & 0.87497436 + our biased protocol may be said to be working basically well up to the moderate orders
. however , a complicated phenomenon occurs at higher orders .
for instance , a `` second '' proper solution appears about @xmath364 and @xmath365 , respectively , in three- and four - parameter ansatze .
the region where the second proper solution lives is separated from the region of first proper solution by the pole of the relevant functions .
after the appearance of the second proper solution , the latter region gradually shrinks and the new region develops .
thus , we come to be aware that the new solution must be kept as the true proper solution , and the old one should be called a pseudo proper solution . as for
one- and two- parameter ansatze , the second proper solution if it exists does not appear up to @xmath366 , which is the maximum order we have studied .
however , a seemingly irregular phenomenon still occurs at higher orders .
for instance , using the exact value @xmath367 , we found that @xmath368\bar f_{\chi n > } ^{(1)}$ ] changes behavior at @xmath369 .
see the change shown in fig .
\bar f_{\chi n > } ^{(1)}$ ] at @xmath370 ( a ) and @xmath371 ( b).,title="fig : " ] \bar f_{\chi n > } ^{(1)}$ ] at @xmath370 ( a ) and @xmath371 ( b).,title="fig : " ] at @xmath370 , @xmath368\bar f_{\chi n > } ^{(1)}$ ] approaches zero from below .
in contrast , at @xmath366 , the same function goes across the zero - line at @xmath372 . as plotted in fig .
16 , from the exact critical behaviors to order @xmath373 ( the exponents of terms included are @xmath208 , @xmath374 , @xmath375 , @xmath376 , @xmath377 , @xmath378 , @xmath379 , @xmath380 ) , we can observe similar behavior as in the case @xmath366 of @xmath368\bar f_{\chi n > } ^{(1)}$ ] . though the matching of the plotted two functions is not good yet
, the gray curve representing @xmath368\bar f_{\chi n < } ^{(1)}$ ] at @xmath366 passes across the zero line at @xmath381 .
the value of the across point slightly increases as the number of terms included grows .
for instance , when the terms to @xmath382 are included , we have @xmath383 for the crossing point . on the other hand
, the evaluation of the crossing point in @xmath368\bar f_{\chi n > } ^{(1)}$ ] proves that @xmath384 at @xmath385 , respectively .
this would signal the reliability of the behavior of @xmath368\bar f_{\chi n > } ^{(1)}$ ] at quite large orders .
we emphasize that the discrepancy between the two functions is in fact small . indeed , the peak values are of the order @xmath386 , and the discrepancy is also of the same order .
the speed of approaching zero of @xmath368\bar f_{\chi n > } ^{(1)}$ ] is extremely slow .
this slow convergence to zero is a signal that this anomalous behavior originates from the logarithmic terms .
\bar f_{\chi n > } ^{(1)}$ ] at @xmath366 ( black ) and @xmath368\bar f_{\chi n < } ^{(1)}$ ] to @xmath387 ( gray ) , in which @xmath388 is included . ] in the improved protocols at large orders .
`` @xmath363 '' means the result from the @xmath241-parameter ansatz of @xmath362 . ]
the sequences of one- and two- parameter ansatze seem to imply slightly smaller value of convergence limit ( see fig .
however , we have checked that over @xmath389 , the estimate shows a slow increasing trend with small oscillation .
the reason of the appearance of an increasing trend even in the peak of the logarithmic effect is not seen in @xmath368\bar f_{\chi n > } ^{(1)}$ ] is that the coefficients of the logarithmic terms decrease as @xmath23 increases ( see ( [ scale_c1 ] ) ) .
the same phenomenon does not occur in @xmath2 and @xmath4 estimations up to @xmath366 .
this would be because @xmath390 and @xmath391 near @xmath4 are of the power - like structure .
in the one - dimensional ising model , the basic result ( [ delta_limit ] ) is analytically proved . in the square model , numerical work to large orders confirmed that ( [ delta_limit ] ) is valid . based upon ( [ delta_limit ] )
, we performed the estimation of @xmath2 , @xmath338 , and @xmath4 and demonstrated that the new protocol works better than the previous one @xcite . also important
is the confirmation of the convergence region of the @xmath0-expanded large @xmath1 expansion .
the existence is proved in one - dimension and verified numerically in the square model .
we emphasize in both cases that the convergence region is of the type @xmath392 , and the origin @xmath96 is excluded .
the presented approach has a characteristic property that one can estimate critical exponents directly from the high - temperature expansion , with no bias from the critical temperature . under the new protocol
, we understand that @xmath2 can be estimated first and then , @xmath338 and @xmath4 in the bias of estimated @xmath2 .
it is logically natural that universal quantities are directly estimated by the large @xmath1 expansion .
the exponent @xmath2 has a special role in our approach because any typical thermodynamic quantity scales as @xmath393 and the first correction has the extra power @xmath224 due to the relation @xmath394 near the critical point . in the traditional method such as dlog pad @xcite and differential approximants methods
@xcite , @xmath4 is involved in the estimation equations .
many estimation data are supplied in a given order and the statistical processing of those data is employed and the average becomes accurate compared to the raw data themselves . in our method , in contrast , we have essentially a unique estimate .
thus , there is no room for the statistical processing of the data .
the accuracy is not superior to those in the traditional approaches , but the logic is alternative and would be interesting since a direct path to the exponent is opened .
99 h. yamada , arxiv : 1303.3714 [ hep - lat ] ( 2013 ) .
h. yamada , phys .
d * 76 * , 045007 ( 2007 ) . on the choice of parameter describing the thermal behavior , see ; i.a .
campbell , k. hukushima and h. takayama , phys .
lett . * 97 * , 117202 ( 2006 ) ; i.a .
campbell , k. hukushima and h. takayama , phys . rev .
* b76 * , 134421 ( 2007 ) ; i.a .
campbell and p. butera , phys . rev .
* b78 * , 024435 ( 2008 ) . at small @xmath11 , the two variable has connection @xmath395 and @xmath1 and @xmath396 has no similarity .
actualy , at small @xmath11 we have expansion of @xmath6 such as @xmath397 .
the @xmath0-expansion is not applicable to this kind of series .
the squared mass @xmath1 can not be replaced by @xmath396 .
montroll , r.b .
potts and j.c .
ward , j. math .
phys . * 4 * , 308 ( 1963 ) .
b. tarko and m. e. fisher , phys
. rev . * b11 * , 1217 ( 1975 ) .
p. m. stevenson , phys .
rev . * d23 * , 2916 ( 1981 ) .
b. m. mccoy , advanced statistical mechanics , oxford university press , 2010 .
w. p. orrick , b. nickel , a. j. guttmann and j. h. h. perk , j. stat .
* 102 * , 795 ( 2001 ) .
s. boukraa , a. j. guttmann , s. hassani , i. jensen , j .-
maillard , b. nickel and n. zenine , j. phys .
a * 41 * , 455202 ( 2008 ) . ;
y. chan , a.j .
guttmann , b.g .
nickel and j.h.h .
perk , j. stat .
145 * , 549 ( 2011 ) .
we have made use of the data of high temperature series opened to the public at http://www.ms.unimelb.edu.au/@xmath226iwan/ising/ising_ser.html .
h. yamada , cit report * 61 * , 91 , 2014 ( arxiv : 1311.4741 ) .
g. a. baker , phys . rev . * 61 * , 768 , 1961 .
a. j. guttmann and g. s. joyce , j. phys . *
a5 * , l81 , 1972 ; m. e. fisher and h. au - yang , j. phys . * a12 * , 1677 , 1979 ; d. l. hunter and g. a. baker , phys . rev . *
b19 * , 3808 , 1979 .
a. j. guttmann , in _ phase transitions and critical phenomena _ ,
* 13 * , 1989 , c. domb and j. lebowitz eds .
( academic press , new york ) .
|
we investigate the large order aspects of the @xmath0-expansion under the estimation procession of the critical quantities . as illustrative examples ,
we revisit the one - dimensional ising model for the analytic study and the two - dimensional square ising model in the high - temperature phase for the numerical experiment to large orders . in both models ,
the proposed fundamental base on which the estimation protocol should be constructed is investigated in details and confirmed to be valid . in the square lattice model
, we present a new protocol for the estimation of critical exponents and temperature .
|
we start with a few background facts about kernel - based methods .
details can be retrieved from the monographs @xcite and the surveys @xcite .
let @xmath3 be a nonempty set , and let @xmath4 be a positive definite and symmetric kernel on @xmath5 .
associated with @xmath6 there is a unique _ native space _
@xmath7 , that is a separable hilbert space of functions @xmath8 where k is the reproducing kernel .
this means that @xmath9 is the riesz representer of the evaluation functional @xmath10 , i.e. , @xmath11 where we use the notation @xmath12 , without subscript , to denote here and in the following the inner product of @xmath7 .
also the converse holds : any hilbert space on @xmath5 where the evaluation functionals @xmath10 are continuous for any @xmath13 is the native space of a unique kernel . one way to construct
the native space is as follows .
first one considers the space @xmath14 and then equips it with the positive definite and symmetric bilinear form @xmath15 the native space @xmath7 then is the closure of @xmath16 with respect to the inner product defined by this form
. given a finite set @xmath17 of distinct points , the interpolant @xmath18 of a function @xmath19 on @xmath20 is uniquely defined , since the _ kernel matrix _
@xmath21_{i , j=1}^n\ ] ] is positive definite , the kernel being positive definite . letting @xmath22 be the subspace spanned by the kernel translates @xmath23 , @xmath24 , @xmath25 , the interpolant @xmath18 is the projection of @xmath26 into @xmath27 with respect to @xmath28 . the usual way to study
the approximation error is to consider the _ power function _
@xmath29 of @xmath20 at @xmath30 , that is the @xmath7-norm of the pointwise error functional @xmath31 at @xmath30 , but we will consider also other measurements of the interpolation error , and other projectors .
we make the additional assumptions that @xmath5 is a compact set in @xmath32 and the kernel is continuous on @xmath33 .
this ensures that the native space has a continuous embedding into @xmath34 .
indeed , using the reproducing property we have @xmath35 where the integral of the kernel is finite .
this allows to define a compact and self - adjoint integral operator @xmath36 which will be crucial for our goals . for @xmath37
we define @xmath38 it can be shown that the range @xmath39 of @xmath40 is dense in @xmath7 , and @xmath41 the operator @xmath40 can be defined using any positive and finite measure @xmath42 with full support on @xmath5 ( see @xcite ) and the same properties still hold , but we will concentrate here on the lebesgue measure . the following theorem
( see e.g. ( * ? ? ?
5 ) ) applies to our situation , and provides a way to represent the kernel as an _ expansion _ ( or _ hilbert - schmidt _ or _ mercer _ ) kernel ( see e.g. @xcite ) .
[ mercer ] if @xmath6 is a continuous and positive definite kernel on a compact set @xmath5 , the operator @xmath40 has a countable set of positive eigenvalues @xmath43 and eigenfunctions @xmath44 with @xmath45 .
the eigenfunctions are orthonormal in @xmath34 and orthogonal in @xmath7 with @xmath46 .
moreover , the kernel can be decomposed as @xmath47 where the sum is absolutely and uniformly convergent .
from now on we will call @xmath48 the _ eigenbasis _ , and use the notation @xmath49 .
we recall that it is also possible to go the other way round and define a positive definite and continuous kernel starting from a given sequence of functions @xmath50 and weights @xmath51 , provided some mild conditions of summability and linear independence .
see @xcite for a detailed discussion about this construction , and note that eigenfunction expansions play a central role in the rbf - qr technique dealt with in various papers @xcite recently .
furthermore , eigenexpansion techniques are a central tool when working with kernels on spheres and riemannian manifolds @xcite . in this paper we shall study the eigenbasis in detail and compare the eigenspaces to other @xmath2-dimensional subspaces of @xmath7 .
general finite dimensional subspaces and their associated @xmath34- and @xmath7- projectors are treated in section [ sec : pro ] .
in particular , the approximation error is bounded in terms of generalized power functions that turn out to be very useful for the rest of the paper .
the determination of error - optimal @xmath2-dimensional subspaces is the core of section [ sec : minerrsub ] , and is treated there by @xmath2-widths , proving that eigenspaces are optimal under various circumstances .
in addition to the case of kolmogorov @xmath2-width as treated in @xcite , we prove that eigenspaces minimize the @xmath34 norm of the power function .
in section [ sec : restriction ] we move towards numerical calculation of eigenbases by focusing on ( possibly finite - dimensional ) closed subspaces . in particular , we want to use subspaces @xmath27 spanned by kernel translates @xmath52 , @xmath24 , @xmath53 for point sets @xmath20 to calculate approximations to the eigenbasis . by means of power functions
we can bound the error committed by working with finite dimensional subspaces .
section [ sec : asympeig ] focuses on the decay of eigenvalues .
we recall the fact that increased smoothness of the kernel leads to faster decay of eigenvalues .
we prove that using point - based subspaces @xmath27 we can approximate the first @xmath2 eigenvalues with an error connected to the decay of @xmath54 .
section [ sec : alg ] describes the numerical algorithms used for the examples in section [ sec : num ] . in particular
, we use a greedy method for selecting sets @xmath20 of @xmath2 points out of @xmath55 given points such that eigenvalue calculations in @xmath27 are stable and efficient .
finally , section [ sec : num ] shows that our algorithm allows to approximate the eigenvalues for sobolev spaces in a way that recovers the true decay rates , and by results of section [ sec : asympeig ] we have bounds on the committed error . for brownian bridge kernels
the eigenvalues are known , and our algorithm approximates them very satisfactorily .
as mentioned in the introduction , the interpolation problem in @xmath7 is well defined , and the interpolation operator is a @xmath7-projector into the spaces generated by translates of the kernel .
but we want to look also at fully general subspaces @xmath56 of @xmath7 , generated by any set of @xmath2 basis functions , and at other linear approximation processes defined on such spaces , e.g. approximations in the @xmath34 norm .
for instance , we consider the two projectors @xmath57 where the @xmath58 and @xmath59 are @xmath34- and @xmath7- orthonormal basis functions of @xmath56 , respectively .
the first projector is defined on all of @xmath34 and can be analyzed on all intermediate spaces .
we want to see how these projectors are connected .
the two projectors do not coincide in general , but there is a special case . for the sake of clarity we present here the proof of the following lemma , even if it relies on a result that is proven in section [ sec : restriction ] .
[ thevn ] if the projectors coincide on @xmath7 for an @xmath2-dimensional space @xmath56 , then @xmath56 is spanned by @xmath2 eigenfunctions .
the converse also holds .
we start with the converse . for each fixed @xmath60 , thanks to , we have @xmath61 then @xmath62 assume now that the projectors coincide on @xmath7 . we can choose a basis @xmath63 of @xmath56 which is @xmath34-orthonormal and @xmath7-orthogonal ( see lemma [ doubleorth ] ) , with @xmath64 .
since @xmath65 for any @xmath19 , necessarily @xmath66 and in particular for @xmath67 , @xmath13 . consequently , @xmath63 and @xmath68 are eigenfunctions / eigenvalues of @xmath40 .
we are now interested in an error analysis of the approximation by functions from these general subspaces , and we want to allow both of the above projectors .
[ defgenpow ] for a normed linear space @xmath69 of functions on @xmath5 and a linear operator @xmath70 on @xmath69 such that all the functionals @xmath71 are continuous for some norm @xmath72 , the _ generalized power function _ in @xmath13 is the norm of the error functional at @xmath30 , i.e. , @xmath73 the definition fits our situation , because we are free to take @xmath74 with @xmath75 , the normed linear space @xmath69 being @xmath7 . in the following , when no confusion is possible , we will use the simplified notation @xmath76 or just @xmath77 to denote the power function of @xmath78 with respect to @xmath79 . to look at the relation between generalized power functions , subspaces and bases we start with the following lemma .
[ lemsumex ] if a separable hilbert space @xmath69 of functions on @xmath5 has continuous point evaluation , then each @xmath69-orthonormal basis @xmath63 satisfies @xmath80 conversely , the above condition ensures that all point evaluation functionals are continuous .
the formula @xmath81 holds in the sense of limits in @xmath69 .
if point evaluations are continuous , we can write @xmath82 in the sense of limits in @xmath83 . since the sequence @xmath84 is in @xmath85 and can be an arbitrary element of that space , the sequence @xmath86 must be in @xmath85 , because the above expression is a continuous linear functional on @xmath85 . for the converse , observe
that for any @xmath87 , @xmath13 and @xmath88 the term @xmath89 is bounded above by @xmath90 , which is finite for any @xmath13 .
hence , for any @xmath13 , @xmath91 is uniformly bounded for @xmath92 .
[ lemprojpow ] for projectors @xmath93 within separable hilbert spaces @xmath69 of functions on some domain @xmath5 onto finite - dimensional subspaces @xmath56 generated by @xmath69-orthonormal functions @xmath94 that are completed , we have @xmath95 provided that all point evaluation functionals are continuous .
the pointwise error at @xmath30 is @xmath96 and , thanks to the previous lemma , we can safely bound its norm as @xmath97 with equality if @xmath98 .
this framework includes also the usual power function .
[ lempointproj ] let @xmath20 be a set of @xmath2 points in a compact domain @xmath5 , and let @xmath27 be spanned by the @xmath20-translates of @xmath6 .
then the above notion of @xmath99 coincides with the standard notion of the interpolatory power function wrt .
@xmath20 .
the proof follows from the fact that the interpolation operator coincides with the projector @xmath100 and the power function is in both cases defined as the norm of the pointwise error functional .
in this section we present the main results of this paper .
our goal is to analyze the behavior of the approximation error , considered from different points of view , depending on the @xmath2-dimensional subspace @xmath56 . roughly speaking
, we will see that it is possible to exactly characterize the @xmath2-dimensional subspaces of minimal error , both considering the @xmath34 norm of the error and the the @xmath34 norm of the pointwise error . the way this problem
is addressed in approximation theory is the study of _ widths _ ( see e.g. the comprehensive monograph @xcite , and in particular chapter 4 for the theory in hilbert spaces ) .
we will concentrate first on the @xmath2-width of kolmogorov .
the kolmogorov @xmath2-width @xmath101 of a subset @xmath102 in an hilbert space @xmath69 is defined as @xmath103 it measures how @xmath2-dimensional subspaces of @xmath69 can approximate a given subset @xmath102 .
if the infimum is attained by a subspace , this is called an _ optimal subspace_. the interest is in characterizing optimal subspaces and to compute or estimate the asymptotic behavior of the width , usually letting @xmath102 to be the unit ball @xmath104 of @xmath69 . the first result that introduces and studies @xmath2-widths for native spaces was presented in @xcite .
the authors consider the @xmath2-width @xmath105 , simply @xmath106 in the following , and prove that @xmath107 and the unique optimal subspace is @xmath108 .
this result is the first that exactly highlights the importance of analyzing the expansion of the operator @xmath40 to better understand the process of approximation in @xmath7 . in the following
we will try to deepen this connection .
our main concern will be to replace the @xmath34 projector @xmath109 by the @xmath7 projector @xmath78 , while still keeping the @xmath34 norm to measure the error .
the @xmath7 projector is closer to the standard interpolation projector , and it differs from the @xmath34 projector unless the space @xmath56 is an eigenspace , see lemma [ thevn ] .
we consider the @xmath34 norm of the error functional in @xmath7 for the projection @xmath78 into a subspace @xmath110 , i.e. , @xmath111 and we look for the subspace which minimizes this quantity . in other words , following the definition of the kolmogorov @xmath2-width , we can define @xmath112 we recall in the next theorem that @xmath113 is equivalent to @xmath106 , i.e. , the best approximation in @xmath34 of @xmath114 with respect to @xmath115 can be achieved using @xmath7 itself and the projector @xmath78 .
the result can be found in @xcite .
[ thekolwidth ] for any @xmath116 we have @xmath117 and the unique optimal subspace is @xmath108 .
since @xmath118 and since @xmath119 is the best approximation from @xmath56 of @xmath19 wrt .
@xmath115 , we have @xmath120 on the other hand , since @xmath121 on @xmath7 ( lemma [ thevn ] ) , @xmath122 since @xmath108 is optimal for @xmath106 .
hence @xmath123 .
we now move to another way of studying the approximation error . instead of directly considering the @xmath34 norm of approximants ,
we first take the norm of the pointwise error of the @xmath124 projector and then minimize its @xmath34 norm over @xmath5 .
this means to find a subspace which minimizes the @xmath34 norm @xmath125 of the power function @xmath77 among all @xmath2-dimensional subspaces @xmath110 . using the definition of the generalized power function
, we can rephrase the problem in the fashion of the previous results by defining @xmath126 in the following theorem , mimicking ( * ? ? ?
* theorem 1 ) , we prove that , also in this case , the optimal @xmath2-dimensional subset is @xmath108 , and @xmath127 can be expressed in terms of the eigenvalues .
[ l2minpow ] for any @xmath116 we have @xmath128 and the unique optimal subspace is @xmath108 . for a subset @xmath56
we can consider a @xmath7-orthonormal basis @xmath129 and complete it to an orthonormal basis @xmath130 of @xmath7 .
we can move from the eigenbasis to this basis using a matrix @xmath131 as @xmath132 where @xmath133 .
hence , the power function of @xmath56 is @xmath134 and , defining @xmath135 , we can compute its norm as @xmath136 now we need to prove that @xmath137 .
let @xmath138 .
we split the cumulative sum over the @xmath139 into integer ranges @xmath140 then @xmath141 can be infinite , but @xmath142 is finite , and since @xmath143 we get stepwise @xmath144 for @xmath145 , using @xmath146 . since the sequence of the eigenvalues is non negative and non increasing , this implies @xmath147 if we take @xmath148 and @xmath149 as its basis , the matrix @xmath102 in is the infinite identity matrix .
thus equality holds in the last inequality .
the previous results motivate the interest for the knowledge and study of the eigenbasis . but from a practical point of view there is some limitation : the eigenbasis can not be computed in general , and it can not be used for truly scattered data approximation , since there exists at least a set @xmath150 such that the collocation matrix of @xmath108 on @xmath20 is singular ( by the mairhuber - curtis theorem , see e.g. ( * ? ? ?
* theorem 2.3 ) ) . to overcome this problem we consider instead subspaces of @xmath7 of the form @xmath151 , where @xmath152 is a possibly large but finite set of points in @xmath5 .
the basic idea is to replace @xmath7 by @xmath153 in order to get a good numerical approximation to the true eigenbasis with respect to @xmath7 . to this end , we repeat the constructions of the previous section for a finite - dimensional native space , i.e. , the problem of finding , for @xmath154 , an @xmath2-dimensional subset @xmath56 which minimizes the error , in some norm , among all the subspaces of @xmath153 of dimension @xmath2 , and that can now be exactly computed .
one could expect that the optimal subset for this restricted problem is the projection of @xmath108 into @xmath27 .
in fact , as we will see , this is not the case , but the optimal subspace will still approximate the true eigenspaces in a near - optimal sense . the analysis of such point - based subspaces can be carried out by looking at general closed subspaces of the native space . it can be proven ( see ( * ? ? ?
1 ) ) that , if @xmath155 is a closed subspace of @xmath7 , it is the native space on @xmath5 of the kernel @xmath156 , with inner product given by the restriction of the one of @xmath7 . the restricted operator @xmath157 defined as @xmath158 maps @xmath34 into @xmath155 .
then theorem [ mercer ] applied to this operator gives the eigenbasis for @xmath155 and @xmath159 on @xmath5 and the corresponding eigenvalues , which will be denoted as @xmath160 , @xmath161 .
they can be finitely or infinitely many , depending on the dimension of @xmath155 .
we will use the notation @xmath162 if @xmath163 .
this immediately proves the following lemma that was already used in section [ sec : pro ] .
[ doubleorth ] any closed subspace @xmath155 of the native space has a unique basis which is @xmath7-orthogonal and @xmath34-orthonormal .
uniqueness is understood here like stating uniqueness of the eigenvalue expansion of the integral operator defined by the kernel , i.e. , the eigenspaces are unique . before analyzing the relation between the approximation in @xmath155 and in @xmath7
we establish a connection between the corresponding kernels and power functions . if @xmath164 is closed , @xmath165 moreover , if @xmath166 are closed , the power functions are related as @xmath167 the power function is the norm of the error functional . for @xmath19 , @xmath168 , and @xmath13
we have @xmath169 with equality if @xmath26 is the normalized difference of the kernels .
this proves , and the relation between the power functions easily follows .
since @xmath155 is a native space itself , the results of the previous section hold also for @xmath155 .
we can then define in @xmath155 the analogous notions of @xmath106 , @xmath113 and @xmath127 , and by theorems [ thekolwidth ] , [ l2minpow ] we know that they are all minimized by @xmath170 , with values @xmath171 , @xmath171 , and @xmath172 , respectively .
these results deal with the best approximation of the unit ball @xmath173 , but allow also to face the problem of the constrained optimization in the case of @xmath127 , i.e. , the minimization of the error of approximation of @xmath114 using only subspaces of @xmath155 .
indeed , thanks to lemma [ powandkers ] , we know that for any @xmath174 and for any @xmath13 , the squared power functions of @xmath7 and of @xmath155 differ by an additive constant .
this means that the minimality of @xmath170 does not change if we consider the standard power function on @xmath7 .
moreover , @xmath175 this proves the following corollary of theorem [ l2minpow ] .
[ l2minvm ] let @xmath164 be a closed subspace of @xmath7 , and let @xmath176 .
for any @xmath2-dimensional subspace @xmath177 we have @xmath178 and @xmath170 is the unique optimal subspace .
in particular @xmath179 this corollary has two consequences .
at one hand , if we want to have a small power function , we need to choose a subspace @xmath155 which provides a good approximation of the true eigenvalues . on the other hand , when dealing with certain point based subspaces , we can control the decay of the power function depending on the number of points we are using , and this bound will provide a bound also on the convergence of the discrete eigenvalues to the true one . the last fact will be discussed in more detail in section [ sec : asympeig ] .
we remark that there is a relation between @xmath108 and @xmath170 : as mentioned before , the optimal subspace @xmath170 is not the projection of @xmath108 into @xmath155 , but is near to be its optimal approximation from @xmath155 . to see this ,
observe that the operator @xmath159 is the projection of @xmath40 into @xmath155 .
in fact , given @xmath180 , we have for any @xmath37 @xmath181 this means that the couples @xmath182 are precisely the bubnov - galerkin approximations ( see e.g. ( * ? ? ?
18.4 ) ) of the solutions @xmath183 of the eigenvalue problem for the restricted operator @xmath184 ( which is still a compact , positive and self - adjoint operator ) .
we can then use the well known estimates on the convergence of the bubnov - galerkin method to express the distance between @xmath108 and @xmath170 .
the following proposition collects convergence results which follow from ( * ? ? ?
* th . 18.5 , th .
18.6 ) .
[ prop : conv ] let @xmath185 be a sequence of closed subspaces which become dense in @xmath7 . for @xmath186 we have a. @xmath187 , b. let @xmath188 be the multiplicity of @xmath189 and + @xmath190 . for @xmath2
sufficiently large , @xmath191 and there exists @xmath192 s.t .
@xmath193 equation proves in particular that @xmath170 is an asymptotically optimal approximation of @xmath108 .
indeed , under the assumptions of the last proposition , we have @xmath194 with @xmath195 as @xmath196 . to conclude this section we point out that , in addition to the point based sets , there is another remarkable way to produce closed subspaces of the native space .
namely , if @xmath197 is any lipschitz subdomain of @xmath5 , @xmath198 is a closed subset of @xmath7 .
this implies that the eigenvalues are decreasing with respect to the inclusion of the base domain ( as can by proven also by the min / max characterization of the eigenvalues ) .
we established a relation between the approximation error and the eigenvalues of @xmath40 .
this allows to use the well known bounds on the approximation error to give corresponding bounds on the decay of the eigenvalues .
these results were already presented in @xcite , but we include them here for completeness and add some extensions .
we consider a set of asymptotically uniformly distributed points @xmath20 , such that the fill distance @xmath199 behaves like @xmath200 where @xmath201 is independent of @xmath2 .
if the kernel is translational invariant and fourier transformable on @xmath32 and @xmath5 is bounded and with a smooth enough boundary , there are standard error estimates for the error between @xmath19 and its interpolant @xmath18 on the points @xmath20 ( see @xcite ) . for kernels @xmath202 with finite smoothness , we have that @xmath203 for @xmath204 , and @xmath205 while for infinitely smooth kernels we have @xmath206 both bounds are in fact bounds on the @xmath207 norm of the power function . if instead one considers directly the @xmath34 error , for kernels with finite smoothness the estimate can be improved as follows :
@xmath208 this immediately leads to the following theorem .
[ theeigdecay ] under the above assumptions on @xmath6 and @xmath5 , the eigenvalues decay at least like @xmath209 for a kernel with smoothness @xmath210 , and at least like @xmath211 for kernels with unlimited smoothness .
the constants @xmath212 are independent of @xmath2 , but dependent on @xmath6 , @xmath5 , and the space dimension .
it is important to notice that the asymptotics of the eigenvalues is known for the kernels of limited smoothness , and on @xmath32 .
if the kernel is of order @xmath210 , its native space on @xmath32 is norm equivalent to the sobolev space @xmath213 . in these spaces
the @xmath2-width , and hence the eigenvalues , decay like @xmath214 ( see @xcite ) .
this means that in sobolev spaces one can recover ( asymptotically ) the best order of approximation using kernel spaces . this can be done also with point based spaces .
the following statement follows from corollary [ l2minvm ] , applying the same ideas as before .
observe that , in this case , we need to consider the bound .
[ theeigapprox ] under the above assumptions on @xmath6 and @xmath5 , we have @xmath215 for a kernel with smoothness @xmath210 , and at least like @xmath216 for kernels with unlimited smoothness . the constants @xmath212 are independent of @xmath2 , but dependent on @xmath6 , @xmath5 , and the space dimension .
this corollary and the previous theorem proves that , using point based sets with properly chosen points , one can achieve at the same time a fast decay of the true eigenvalues and a fast convergence of the discrete ones . both results in this section raise some questions about the converse implication . from theorem [ theeigdecay ]
we know that the smoothness of the kernel guarantees a fast decay of the eigenvalues .
but we can also start from a given expansion to construct a kernel .
is it possible to conclude smoothness of the kernel from fast decay of the eigenvalues ?
corollary [ theeigapprox ] tells us that uniformly distributed point based sets provide a good approximation of the eigenvalues .
we will see in section [ sec : alg ] and [ sec : num ] that one can numerically construct point based sets whose eigenvalues are close to the true ones .
is it possible to prove that these sets are necessarily asymptotically uniformly distributed ?
if we consider a closed subspace @xmath217 spanned by translates of the kernel on a set @xmath218 of @xmath55 points in @xmath5 , it is possible to explicitly construct @xmath219 , @xmath220 .
the number @xmath55 of points should be large enough to simulate @xmath34 inner products by discrete formulas .
the first method we present aims at a direct construction of the eigenspaces and gives some insight on the structure of the basis , but it is numerically not convenient since it involves the computation of two gramians . for stability and computability reasons , we shall then focus on lower - dimensional subspaces of @xmath153 .
there are various ways to construct these , and we present two .
we will see in section [ sec : num ] that these approximation are very effective . from now on we will use the subscript @xmath55 in place of @xmath153 to simplify the notation , keeping in mind that there is an underlying set of points @xmath221 and the corresponding subspace @xmath153 .
we use at first the fact that the eigenbasis is the unique set of @xmath55 functions in @xmath153 which is orthonormal in @xmath34 and orthogonal in @xmath7 , where uniqueness is understood in the sense of uniqueness of the eigendecomposition of the integral operator .
given any couple of inner products @xmath222 and @xmath223 on @xmath153 , it is always possible to build a basis @xmath224 of @xmath153 which is @xmath225-ortho_normal _ and @xmath226-ortho_gonal _ with norms @xmath227 .
let @xmath228_{i , j=1}^n,\\ b = & [ ( k(\cdot , x_i),k(\cdot , x_j))_b]_{i , j=1}^n\end{aligned}\ ] ] be the gramians with respect to the two inner products of the standard basis @xmath52 , @xmath24 , @xmath229 of @xmath6 translates .
following the notation of @xcite , to construct the basis we need to construct an invertible matrix @xmath230 of change of basis to express this new basis with respect to the standard basis . to have the right orthogonality , we need @xmath231 and @xmath232 , where @xmath233 is the diagonal matrix having on the diagonal the @xmath226-norms of the new basis .
this means to simultaneously diagonalize the two gramians , and since they are symmetric and positive definite this is always possible , e.g. in the following way : * @xmath234 be a cholesky decomposition , * define @xmath235 ( which is symmetric and positive definite ) , * let @xmath236 be a svd decomposition , * define @xmath237 .
observe that , for practical use , it is more convenient to swap the role of @xmath102 and @xmath238 . in this way
we construct the basis @xmath239 , which is @xmath7-orthonormal hence more suitable for approximation purposes , and , moreover , we obtain directly the eigenvalues of order @xmath55 as @xmath240 . in both ways
we are just computing a generalized diagonalization of the pencil @xmath241 , up to a proper scaling of the diagonals . in our case
@xmath102 is the usual kernel matrix , while @xmath242 .
thus , provided we know @xmath238 , we can explicitly construct the basis .
this can be done , for a general kernel , using a large set of points to approximate the @xmath34 inner product .
numerical experiments ( see section [ sec : num ] ) suggest that this construction of the eigenspaces is highly unstable also in simple cases , since it requires to solve an high dimensional matrix eigenvalue problem .
another way to face the problem consists of greedy procedures as considered next . instead of directly constructing the subspace @xmath243 via @xmath244 matrix eigenvalue problems as described before
, we can first select @xmath2 points in @xmath218 such that working on these points is more stable than working with the full original matrix , and then solve the problem in @xmath27 .
the selection of the set @xmath20 is performed with a greedy construction of the newton basis ( see @xcite ) .
first we show how to construct the eigenspaces via the newton basis .
assume that @xmath94 is the newton basis for @xmath27
. then @xmath245 and if @xmath246 is an eigenvalue with eigenfunction @xmath247 then @xmath248 implies @xmath249 thus the coefficients of the eigenbasis with respect to the newton basis are the eigenvectors of the @xmath34 gramian of the newton basis .
experimentally , the newton basis is nearly @xmath34 orthogonal .
thus the above procedure should have a nice gramian matrix , provided that the @xmath34 inner products that are near zero can be calculated without loss of accuracy . to select the points we use two similar greedy strategies , based on maximization of @xmath207 and @xmath34 norms of the power function .
the first point is chosen as @xmath250 denoting by @xmath251 the already chosen points , the @xmath252-th point is selected as @xmath253 the point sets selected by the two strategies are different , but we will see in the next section that they provide similar results in terms of approximation of the eigenspaces .
in these spaces the asymptotic behavior of the kolmogorov width , hence of the eigenvalues , is known as recalled in section [ sec : asympeig ] .
we assume here that the same bounds hold in a bounded domain ( the unit disk ) , and we want to compare it with the discrete eigevalues of point based sets , which can be computed with one of the methods of the previous section . to this aim , we start from a grid of equally spaced points in @xmath256 ^ 2 $ ] restricted to the unit disk , so that the number of points inside the disk is @xmath257 .
we use this grid both for point selection and to approximate the @xmath34 inner products as weighted pointwise products , i.e. , @xmath258 we select @xmath259 points by the greedy @xmath207 maximization of the power function in the unit disk .
the results are shown in figure [ fig : matorder ] . as expected , for any order under consideration there
exist a positive constant @xmath201 such that the discrete eigenvalues decay with the same rate of the sobolev best approximation .
moreover , we expect that the discrete eigenvalues converge to the true ones with a rate that can also be controlled by @xmath210 . indeed , according to corollary [ theeigapprox ]
, we have @xmath260 to verify this , we instead look at the decay of @xmath261 since we can exactly compute the first term .
indeed , since the kernels are radial , we have @xmath262 results are presented in figure [ fig : matdiff ] . from the experiments it seems that we can improve the convergence speed somewhat , and in fact obtain a rate of order @xmath263 instead of @xmath264 .
this may be another instance of the `` gap of @xmath265 '' already observed in @xcite . sometimes , observed convergence rates are by @xmath265 better than proven ones .
w. pogorzelski . .
translated from the polish by jacques j. schorr - con , a. kacner and z. olesiak . international series of monographs in pure and applied mathematics , vol .
pergamon press , oxford , 1966 .
r. schaback .
native hilbert spaces for radial basis functions i. in m. buhmann , d. h. mache , m. felten , and m. mller , editors , _ new developments in approximation theory _ ,
number 132 in international series of numerical mathematics , pages 255282 .
birkhuser verlag , 1999 .
r. schaback and h. wendland .
approximation by positive definite kernels . in m.
buhmann and d. mache , editors , _ advanced problems in constructive approximation _ ,
volume 142 of _ international series in numerical mathematics _ ,
pages 203221 , 2002 .
|
kernel - based methods in numerical analysis have the advantage of yielding optimal recovery processes in the native hilbert space @xmath0 in which they are reproducing . continuous kernels on compact domains have an expansion into eigenfunctions that are both @xmath1-orthonormal and orthogonal in @xmath0 ( mercer expansion ) .
this paper examines the corresponding eigenspaces and proves that they have optimality properties among all other subspaces of @xmath0 .
these results have strong connections to @xmath2-widths in approximation theory , and they establish that errors of optimal approximations are closely related to the decay of the eigenvalues .
though the eigenspaces and eigenvalues are not readily available , they can be well approximated using the standard @xmath2-dimensional subspaces spanned by translates of the kernel with respect to @xmath2 nodes or centers .
we give error bounds for the numerical approximation of the eigensystem via such subspaces .
a series of examples shows that our numerical technique via a greedy point selection strategy allows to calculate the eigensystems with good accuracy . * _ keywords : _ * mercer kernels , radial basis functions , eigenfunctions , eigenvalues , @xmath2-widths , optimal subspaces , greedy methods + * _ 2000 msc : _ * 41axx , 41a46 , 41a58 , 42cxx , 42c15 , 42a82 , 45p05 , 46e22 , 47a70 , 65d05 , 65f15 , 65r10 , 65t99
|
constrained systems are those having constraints on their degrees of freedom .
when a constraint is imposed on spatial coordinates , the constraint is called a `` holonomic constraint''@xcite .
constrained systems are useful and simple , because of which they are widely used as model systems .
an example of such a model system is a freely jointed chain @xcite , which is a model composed of one - dimensional chain ( 1d ) of @xmath0 masses such that the distances between adjacent masses are constant .
the freely jointed chain is known as a simplified model of polymers . in computational software packages such as charmm
@xcite and amber @xcite for molecular dynamics calculations , the shake and rattle algorithms enable one to treat model systems as constrained ones by setting distances between atoms as constant . as a result , the computational task can be made much easier , and then , physically important results can be obtained . in molecular dynamics calculations ,
water molecules are often treated as having a fixed shape , where the length of bonds between hydrogen and oxygen atoms is fixed @xcite . in both cases , we replace the bond between atoms with a rigid link when the frequency of bond vibration is extremely high .
it is known that in a constrained system , the equipartition of energy occurs in a somewhat complicated way , and the average kinetic energies of particles @xmath1 can take different values for particles located at different places in the system @xcite . in the generalized form of equipartition
@xcite , what is equal among degrees of freedom is not @xmath1 but @xmath2 , where @xmath3 is the momentum conjugate to the generalized coordinate @xmath4 of the @xmath5th degree of freedom , and @xmath6 is the kinetic energy of the system . in constrained systems , @xmath6 depends on coordinates and @xmath3 is no longer equal to @xmath7 ; therefore @xmath8 , whose average takes the same value for all @xmath5 , is not equal to @xmath9 .
recently , we found that for a chain - type system , termed a planar chain model , the average kinetic energy of each particle differs systematically ; that is , particles near both the ends of the chain have relatively large average kinetic energies @xcite .
this model consists of masses connected by rigid links ; since the distances between adjacent masses are fixed in this model , it is a constrained system .
it is the constraint that induces the nonuniformity of average kinetic energies
. the abovementioned energetic motion of end particles observed in the model would be useful in understanding the dynamical behavior of chain - type systems such as polymers @xcite , dnas , proteins @xcite , and some artificial objects such as manipulator arms of spacecraft .
thus far , we have described the behavior of constrained systems . however , strictly speaking , a constrained system or rigid link does not exist in the real world .
when the potentials of a system are somewhat steep , i.e. , spring constants are quite large and the frequency of bond vibration is reasonably high , we approximate the bonds with rigid links . a rigid link or holonomic constraint
is an idealized limit of a stiff spring . however ,
if we replace the rigid links with springs , say spring - chain model , the usual expression of equipartition of energy , i.e. , @xmath10 , holds regardless of the magnitude of the spring constant @xmath11 , where @xmath12 denotes the spatial dimension .
in other words , although a spring - chain system , made by replacing rigid links in the planar chain system with springs , appears to behave like the planar chain model when the spring constants are large , the behavior of energetic motion of end particles can not be reproduced by applying equilibrium statistical mechanics to the model
. then it would be interesting to know whether the large average kinetic energies of end particles observed in the planar chain model can also be found in the real world or whether it is an artifact observed only in mathematical models and is never observed in the real world .
if the former is true , then we can expect to observe an interesting feature that the energy distribution of many - body systems shows nonuniform behavior .
since it is natural to consider that the dynamical behavior of a stiff spring is similar to that of a rigid link at least for a finite time , the solution to the problem will be the knowledge of the relaxation properties to equilibrium . with this background ,
the aim of this study is to examine the property of relaxation to equipartition for the spring - chain system , particularly for a large value of the spring constant @xmath11 .
we measure the relaxation time @xmath13 and investigate its relation with the spring constant @xmath11 .
further , we estimate @xmath13 by the boltzmann - jeans theory and compared it with the value measured .
this paper is organized as follows . in sec.ii
we introduce the model , the spring - chain system . in sec.iii
we briefly describe the method of numerical computation .
the results are shown in sec.iv .
the final section is devoted to the summary and discussion .
we now introduce the spring - chain system .
it is composed of @xmath0 particles ( masses ) connected by @xmath14 massless springs .
the masses can rotate smoothly on the @xmath15-plane , as shown in fig .
[ fig : spring - chain ] .
the system is defined by the following lagrangian @xmath16 : @xmath17 where @xmath18 is the mass of the @xmath5th particle , @xmath19 represents the position of the @xmath5th particle , @xmath20 and @xmath21 are the spring constant and natural length of the @xmath5th spring , respectively .
@xmath22 is an external potential . in this paper
we consider the case of @xmath23 , @xmath24 , and @xmath25 for all @xmath5 .
we set @xmath26 and @xmath27 .
the spring - chain model is a kind of beads - type models , which are used as models of polymer and protein @xcite .
we integrate the equation of motion of the model by a fourth - order symplectic integrator that is the composition of three successive second - order symplectic integrators .
external potential @xcite @xmath28 is applied in order to break the rotational symmetry and thus prevent the conservation of angular momentum .
here @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 . throughout this paper ,
the following parameters and initial condition are set @xmath36 , @xmath37 for all @xmath5 and @xmath38 , @xmath39 for all @xmath5 , respectively .
the values of system size and initial momentum for each simulation set will be defined in the subsequent sections .
next , we briefly summarize the relation between our numerical simulation and thermal equilibrium . under most of the initial conditions considered for the present simulations , the system undergoes chaotic motion .
since the model has no conserved quantities other than the total energy , one may think that the states of the system attained in the course of a long - duration are well approximated by a microcanonical distribution . in that case
, the distribution of the state of each particle in the chain can be approximated by a canonical distribution at a certain temperature , by considering the other particles in the chain as a heat bath . that is , the long - term average of kinetic energy of each particle , @xmath40 , is equal to thermal average @xmath41 .
then , according to the principle of equipartition of energy , the average kinetic energy of each particle is the same : @xmath42 where @xmath43 , @xmath44 , @xmath45 is the boltzmann constant , and @xmath46 is the temperature . since our aim is to investigate the property of relaxation to equipartition , we define the following quantity in order to measure how close the system is to equipartition : @xmath47 ^ 2 \ , .\ ] ] if @xmath48 , then @xmath49 for all @xmath5 .
similar quantities have been used to measure the degree of equipartition in studies on a supercooled liquid @xcite , self - gravitating systems @xcite , and proteins @xcite . vs. @xmath50 for @xmath51 .
plots are measured at 10 successive times : @xmath53 ( blue circles ) , @xmath54 ( red squares ) , @xmath55 ( yellow diamonds ) , @xmath56 ( green triangles ) , and @xmath57 ( blue inverted triangles ) .
the time step of integration @xmath58 .
the initial conditions are @xmath59 , @xmath60 , @xmath61 , and @xmath62 , @xmath63 @xmath64 . ]
figure [ fig : ki - vs - t - for - k-1e4 ] shows the time evolution of the profile of @xmath65 .
it is clearly observed that in the initial stage of the time evolution , the average kinetic energy of all particles is not equal ; rather , particles near both the ends of the chain have a larger average @xmath66 .
the profile is similar to that of the rigid link , i.e. , the planar chain model @xcite .
then , as time progresses , differences in average @xmath66 among particles gradually decrease and tend to zero , and equipartition is achieved . for @xmath67 ( blue circles ) , @xmath68 ( red squares ) , @xmath69 ( yellow diamonds ) , and @xmath70 ( green triangles ) .
the other parameters and initial conditions are the same as those mentioned in the caption of fig .
[ fig : ki - vs - t - for - k-1e4 ] .
the dashed line represents @xmath71 . ]
figure [ fig : delta - vs - t - k1e4 ] shows the relaxation of @xmath72 [ eq .
( [ eq : delta ] ) ] for the data considered in fig .
[ fig : ki - vs - t - for - k-1e4 ] .
we observe that the system relaxes to equilibrium with the progress of time .
the physical process of relaxation can be understood by examining the kinetic energy in greater detail .
we rewrite the hamiltonian as @xmath73 where @xmath74 ( blue circles ) and @xmath75 ( red squares ) .
@xmath52 , @xmath76 , and @xmath77 .
@xmath78 the initial conditions are @xmath79 , @xmath60 , @xmath61 , @xmath80 ( @xmath81 ) , and @xmath63 @xmath82 . ] if equipartition is achieved , @xmath83 , because the model has the same number of springs and angles .
figure [ fig : delta - kvib - krot - vs - t ] shows the temporal evolution of @xmath84 and @xmath85 for @xmath86 . here ,
the time average with two arguments , @xmath87 , is defined as @xmath88 and @xmath89 ^ 2 \ , .\ ] ] we find that the system relaxed to equipartition on a timescale similar to that on which the rotational energy @xmath90 transformed into vibrational energy @xmath91 .
as mentioned earlier , the aim of this study is to examine the property of relaxation to equipartition of energy for the model expressed in eq .
( [ eq : spring - chain ] ) .
first , for each sample orbit starting from different initial condition , we define the relaxation time @xmath92 as the time required for @xmath72 to decay below a critical value @xmath93 .
we define the average relaxation time over @xmath94 orbits as @xmath95 this time is a measure of the relaxation time to equipartition .
dependence of average relaxation time @xmath13 . at each @xmath11 value
, 15 samples are taken .
threshold value @xmath96 .
the time step @xmath97 of integration is scaled as @xmath98 .
the error bars represent standard deviations of @xmath92 for 15 samples .
the dashed line shows @xmath99 obtained by fitting . ]
figure [ fig : log - trelax - vs - sqrt - k ] shows the plot of the dependence of @xmath13 on the spring constant @xmath11 .
we observe that as the stiffness of the spring increases , the relaxation time increases rapidly .
that is , systems with hard springs or a steep potential show rigid - like behavior of energetic particles near the chainends for a very long time , as shown in fig .
[ fig : ki - vs - t - for - k-1e4 ] , before relaxing to equipartition . here
, we mention a technical detail about the numerical integration used for obtaining the plot in fig .
[ fig : log - trelax - vs - sqrt - k ] . with increasing spring constant @xmath11 , the period of bond - stretching vibration decreases in proportion to @xmath100 .
therefore , for large values of @xmath11 , the magnitude of the time step of numerical integration should be reduced .
we confirmed that @xmath13 converges at @xmath101 for @xmath102 .
thus , we used @xmath98 for each @xmath11 . on changing the initial conditions , the relaxation time @xmath103 changes ,
and we obtain a distribution of @xmath92 , denoted as @xmath104 . for @xmath105 ( blue circles ) and @xmath86 ( red squares )
for each value of @xmath11 , the numbers of samples are 200 .
both distributions are shown on a semi - log scale .
the dashed lines show @xmath106 .
for @xmath105 , @xmath107 and @xmath108 . for @xmath86 , @xmath109 and @xmath110 .
] figure [ fig : hist - trelax ] shows the distribution of @xmath13 for @xmath111 and @xmath105 . for both cases ,
the histograms show exponential decay expressed as @xmath112 which suggests the existence of a characteristic timescale for the relaxation .
next , we analyze the results of the abovementioned calculation using the concept of the boltzmann - jeans theory ( also known as boltzmann - jeans conjecture ) @xcite .
the essence of this theory is roughly described as follows .
( for a detailed description of the theory please refer to @xcite . )
suppose we have a system described by a hamiltonian , which has two subsystems @xmath113 and @xmath114 , their typical time scale being @xmath115 and @xmath116 , respectively . here , subscripts @xmath117 and @xmath118 denote `` fast '' and `` slow , '' respectively .
let us call @xmath113 and @xmath114 as a `` fast subsystem '' and `` slow subsystem , '' respectively .
if the timescales of the fast and slow subsystems differ greatly , i.e. , @xmath119 , then the timescale for the occurrence of energy exchange between these two subsystems is on the order of @xmath120 that is , energy exchange occurs after a long time . in the case of the spring - chain system [ eq .
( [ eq : spring - chain ] ) ] with a large spring constant @xmath11 , the fast and slow subsystems correspond to bond vibration and relative rotation , respectively .
since the typical timescale of bond vibration is on the order of @xmath121 and that of rotation is assumed to be constant , we have @xmath122 since the relaxation to equipartition occurs by energy transfer from rotation to vibration ( as we observed before ) , we can consider that @xmath123 mentioned above is essentially the same as the relaxation time @xmath13 : @xmath124 now , we examine whether @xmath13 obtained by the simulation obeys eq .
( [ eq : timescale - relax ] ) . the result is already shown in fig . [ fig : log - trelax - vs - sqrt - k ] .
@xmath125 is proportional to @xmath126 ; therefore the interpretation by the boltzmann - jeans theory is appropriate .
this theory can also be used for interpreting the histogram of @xmath92 .
since @xmath13 is defined from the average of a number of samples , we have @xmath127 where @xmath128 is the distribution of @xmath92 . if we adopt the exponential form for the distribution @xmath129 [ eq . ( [ eq : hist - exp ] ) ] , then @xmath130 where @xmath131 is the coefficient that appears in eq .
( [ eq : hist - exp ] ) . combining eqs .
( [ eq : timescale - relax ] ) and ( [ eq : t - relax - alpha ] ) , the relation between the coefficient @xmath131 and the spring constant should be @xmath132 a comparison between the estimation [ eq .
( [ eq : alpha - vs - k ] ) ] and data is shown in figure [ fig : alpha - vs - k ] , from which we find that they are in good agreement .
vs @xmath11 , obtained from 300 samples .
the dashed line shows @xmath133 , where @xmath134 and @xmath135 . ]
thus , the fact that relaxation to equilibrium takes quite a long time to occur can be interpreted as the outcome of the boltzmann - jeans theory .
in this paper , we have numerically shown the occurrence of energetic motion of end particles for a 1d chain of point masses connected by hard springs .
the timescale at which the energetic motion is observed depends on the spring constant , and this timescale lengthens with increasing the spring constant .
relaxation to equilibrium occurs as rotational kinetic energy is converted into vibrational kinetic energy .
the timescale of relaxation is estimated using the boltzmann - jeans theory , which describes the energy exchange rate in a system in which fast and slow motions coexist . in the case of our model
, fast motion corresponds to bond vibration by a hard spring and slow motion corresponds to rotation and deformation of the chain .
the result of our numerics is that the timescale is estimated as an exponential of the square root of the spring constant , which coincides well with simulation data .
the energetic motion of end particles is also observed in other systems , e.g. , a planar chain system @xcite and multiple pendulum @xcite .
these systems consist of masses connected by rigid links , and therefore , they are constrained systems .
if we consider a rigid link as a limiting case of a hard spring when the spring constant @xmath11 is large ( @xmath136 ) , then the spring - chain system examined in this study becomes a planar chain system when @xmath137 .
therefore , it is natural that the spatial energy distribution of the spring - chain system considered in this study resembles those of systems with rigid links .
however , the equilibrium behaviors of spring - chain system and rigid - link chain system are quite different . in the spring - chain system ,
equipartition achieved is of the usual form ; that is , the average kinetic energy of each particle is the same , regardless of the magnitude of @xmath11 .
then , chain - end particles in thermal equilibrium should not exhibit the energetic behavior .
in contrast , in the rigid - link - chain system , chain - end particles behave energetically even in thermal equilibrium @xcite .
the energetic behavior of end particles in the spring - chain system is a transient behavior before the system relaxes to thermal equilibrium .
what is important here is that the relaxation time increases with increasing the spring constant @xmath11 and that the relaxation time eventually diverges at @xmath138 .
therefore , a large @xmath11 value provides a good opportunity to observe energetic behavior in spring - chain systems .
the relaxation time to equipartition changes if we change the initial conditions .
we found that the the relaxation time is distributed according to the exponential form for large relaxation time .
this implies that relaxation occurs almost randomly .
that is , the system moves on the energy surface in a random way and happens to encounter at which the system can divert toward the equipartition state .
this situation is in contrast to the process of slow dynamics often observed in many hamiltonian systems , where the relaxation time is often distributed according to the power law .
slow relaxation is often observed in hamiltonian systems , and in most cases , it is accompanied by strong temporal correlation caused by sticky or stagnant motion around kam tori ( regular orbits ) and their remnants and @xmath139-type fluctuations . such slow relaxation is observed in area - preserving mappings and some other high - dimensional systems @xcite . such slow relaxation is explained by a hierarchical structure generally found in nearly integrable hamiltonian systems . in this sense , it is quite common to find slow relaxation in hamiltonian systems , if the hamiltonian is nearly integrable . however , the spring - chain model considered in this study is not nearly integrable .
in addition , the systems which show slow relaxation in real world are not always nearly integrable .
in contrast , the boltzmann - jeans theory is still applicable to hamiltonian systems that are not nearly integrable and is able to explain the slow relaxation .
similar to the research by shudo et al .
@xcite , the result of our study also implies that the boltzmann - jeans theory can describe slow relaxation in more general systems other than the nearly integrable ones .
we used the boltzmann - jeans theory @xcite to estimate the relaxation time .
although the concept of the theory dates back to the 19th century @xcite , its importance is not very familiar and not many examples of the application of this theory have been demonstrated @xcite .
because the results of this study show that there is good agreement between numerical data and theoretical estimation , this study can be a good example of the applicability of the boltzmann - jeans theory .
in this study , we showed that chain - end particles behave energetically even for systems with a finite but large spring constant .
this implies that similar behavior can be observed for natural chain - type systems that are not made of rigid links but whose intrachain potential between elements is very steep .
the results of this study are expected to have many useful applications to polymers , proteins and some artificial objects .
|
in this study , one - dimensional systems of masses connected by springs , i.e. , spring - chain systems , are investigated numerically .
the average kinetic energy of chain - end particles of these systems is larger than that of other particles , which is similar to the behavior observed for systems made of masses connected by rigid links .
the energetic motion of the end particles is , however , transient , and the system relaxes to thermal equilibrium after a while , where the average kinetic energy of each particle is the same , that is , equipartition of energy is achieved .
this is in contrast to the case of systems made of masses connected by rigid links , where the energetic motion of the end particles is observed in equilibrium .
the timescale of relaxation estimated by simulation increases rapidly with increasing spring constant .
the timescale is also estimated using the boltzmann - jeans theory and is found to be in quite good agreement with that obtained by the simulation .
|
the summary of the recent target and horn studies for the cern to frjus neutrino beam is presented in this paper .
the main design and the physics reach of the super beam project are described in @xcite .
the optimization procedure for the horn shape and layout - geometry to achieve optimum physics , the study of a target able to withstand a multi - mw proton - beam power , multi - physics simulation to investigate heat transfer , cooling and mechanical stress for the horn and safety aspects are discussed here .
a 4-mw proton - beam from cern s spl is foreseen to be separated by a series of kicker magnets into four beam lines .
then each beam will be focused by a series of quadruples and correctors to a four horn / target assembly ( figure [ fig : 4horn ] ) . in this way , each horn / target assembly is able to accommodate better the multi - mw power and thus increasing its lifetime @xcite .
a 0.25 mm thick beryllium beam window has been studied as the interface between each 1 mw proton - beam line and each horn / target assembly .
maximum temperatures as high as 180 @xmath0c and ( 109 @xmath0c ) and von mises stresses as high as 50 mpa and ( 39 mpa ) are developed respectively for water and helium cooling : these are well below the beryllium strength limit .
a packed - bed target with ti6al4v - spheres and helium transverse cooling has been chosen as the baseline target option @xcite .
it is placed inside the upstream part of horn s inner conductor .
the advantages of the packed - bed target are among others a@xmath1 large surface area for heat transfer with coolant able to access areas with highest energy deposition b@xmath1 minimal thermo - mechanical and inertial stresses and c@xmath1 potential heat removal rates at the hundreds kilowatt level with high helium flow rate .
advantages of the helium transverse cooling are a@xmath1 almost beam neutral b@xmath1 no generation of stress wave in coolant and c@xmath1 low activation of coolant with no corrosion problems .
alternatively , a pencil - like geometry of solid beryllium has been studied @xcite .
this pencil - like geometry gives steady - state thermal stress within acceptable range for beryllium .
pressurized helium cooling appears feasible but center proton - beam effects could be problematic because of the stress induced : this point needs further thermo - mechanical studies .
horn with cooling system . ]
horn with cooling system . ]
the end - design consists of an inner conductor with a cylindrically shaped upstream part to decrease the transverse momentum of the low - energy charged mesons , followed by a trapezoidal shaped middle part to select a special particle energy spectrum ( for optimum physics ) and finally a convex downstream plate to de - focus wrong - sign mesons that contribute to the background neutrino spectra .
this configuration has been selected as the best compromise between physics performance and reliability under 1 to 1.3 mw proton - beam power @xcite .
the detailed drawing for a horn is shown in figure [ fig : horn ] .
the horn parameters as well as the geometrical parameters of the decay tunnel ( length and radii ) are optimized for the best achievable sensitivity limit on sin@xmath2 .
the beam parameters are initially scanned broadly and then restrictly in three iterations in order to minimize the cp - violation averaged 99% c.l .
sensitivity limit on sin@xmath2 @xcite .
the aluminum alloy al-6061t6 horn is subjected to cyclic deformation due to a pulsed magnetic pressure load . in addition
, the temperature field creates a thermal static stress . for a non uniform cooling with maximal temperature of 60 @xmath0c
, the maximal static thermal stress is calculated about @xmath3 mpa and is located in the upstream corner and downstream top part of the horn . if a uniform temperature is achieved everywhere , the horn is expanding , and the maximum thermal static stress is 6 mpa @xcite .
the combined stresses in the inner conductor due to the magnetic pulses and the thermal stress for a uniform achieved temperature of @xmath3 @xmath0c are around @xmath4 mpa . for the inner conductor horn ,
the magnetic pressure pulse creates a peak von mises stress of about @xmath5 mpa .
the stress and the deformation are shown in figures [ fig : h_stress ] and [ fig : h_dis ] .
u@xmath6=2.4 mm . ]
u@xmath6=2.4 mm . ]
to remove a total power of about @xmath3 kw and maintain a temperature of about @xmath7 c a water - jet cooling system is being studied .
this system will be made of 5 rows of 6 nozzles ( figure [ fig : horn ] ) and to spray water toward the inner conductor of the horn .
the estimated water flow rate is calculated between 60 to 120 l / min per horn depending the design . to minimize the thermal static stress , the nozzle size and disposition
should be properly located in order to achieve the most uniform temperature everywhere inside the horn @xcite .
there is no fatigue limit available for aluminum alloy , which means that fatigue data can only give a probability of failure for a determined level of stress and number of cycles . according to @xcite ,
the fatigue strength limit is 20 mpa for 10@xmath8 pulses with a mean static stress due to thermal dilatation . for the weld junction
a limit of @xmath9 mpa should be respected to maximize horn lifetime @xcite .
a one - half sinusoid current waveform with a 350 ka maximum current and pulse - length of 100 @xmath10s at 12.5 hz frequency is needed for each horn . in order to compensate for these extreme conditions a power supply with 10 modules branched in parallel with 35 ka current each running at 50 hz frequency
is being studied .
also , the power supply components lifetime is very important as 0.216 10@xmath11 pulses for each horn s switch and a total of 0.864 10@xmath11 pulses per 200 days for the power supply are expected .
thus , the following designs are being investigated : a power supply with energy recovery made by a ) a self that implies huge stresses on the switches and the capacitor due to high reverse voltage and current and b ) a diode that implies less stress on the switches and the capacitor , but increase the electric consumption and energy thermal dissipation .
their comparison in terms of manufacturing , reliability and electric costs will define the final solution .
the future design of multi - mega watt sources facility has to take to account the significant amount of radiation produced during beam operation and the radio - activation of the surrounding environment .
the design of the shielding should reduce the dose equivalent rate to a minimal level . in order to reach these dosimetry objectives the alara ( as low as reasonably achievable )
approach will be used in the design of the facility .
alara consists of an iterative process between three phases : a ) preparation , design of the facility , dose equivalent rate map , study the intervention procedures for workers b ) execution , engineering phase check / improve the dosimetry objectives and c ) analyse and feedback on safety from others experiments .
activity distribution in facility . ]
activity distribution in facility . ]
the super beam infrastructure consists of : a ) proton driver line , b ) experimental hall ( target station , decay tunnel , beam dump ) , c ) spare area room , d ) hot cell , e ) service galleries ( power supply , cooling system ) f ) air - ventilation system and g ) waste area ( figure [ fig : newlayout ] ) .
as example of safety monte - carlo studies with fluka , the activation of the whole structure and of molasse rock surroundings have been evaluated after 200 days of irradiation : there is minimum rock ( molasse ) activation surrounding the super beam facility due to concrete shielding with 5.5 m of thickness as shown in figure [ fig : activity ] .
monte - carlo , thermo - mechanical and dynamical stress finite - elements analysis studies show that the four - horn / target system can be feasible under the extreme 4 mw proton - beam power conditions . furthermore , r&d is needed for the target and the horn in order to study the fatigue , cooling , power supply designs and radiation degradation .
we acknowledge the financial support of the european community under the european commission framework programme 7 design study : euronu , project number 212372 .
the ec is not liable for any use that may be made of the information contained herein .
9 longin a 2010 _ a new design for the cern - frjus super beam _ ( euronu - wp2 - 10 - 04 ) http://www.euronu.org
|
in the framework of the euronu design study , a new design for the cern to frjus neutrino beam based on the spl is under development by the wp2 group .
the main challenge of this project lies with the design of a multi - mw neutrino beam facility .
the horn and the decay tunnel parameters have been optimized to maximize any potential discovery .
the target design , thermo - mechanical analysis , and power supply design of the horn system as well as any safety issues are being studied to meet the mw power requirements for the proton - beam . _
( contribution to nufact 11 , xiiith international workshop on neutrino factories , super beams and beta beams , 1 - 6 august 2011 , cern and university of geneva ( submitted to iop conference series ) ) _
|
the tremendous progress in the development of high performance computing systems , especially expecting drastically new exascale computing systems , including the challenges in architecture , scale , power and reliability , gives new opportunities for the mathematical modeling of important physical phenomena in the present and future .
nevertheless the complexity of the challenges in science and engineering continues to outpace our ability to adequately address them through impressively growing computational power .
a feature of the present is that the development of technologies and computer systems architecture are well ahead of software development .
the software problems are primarily associated with the complexity of the algorithms adaptation for the differential equations of mathematical physics to high performance computing systems architecture .
in particular they refer to one of the important requirements as the accuracy in combination with the correctness of the initial mathematical models .
another requirement for the methods is their logical simplicity and high efficiency at the same time .
the numerical algorithms should be simple and transparent from a logical point of view .
one of the important directions to overcome these problems is the development of a nontraditional approach to initial mathematical models and computational algorithms . in the present study for the solution of the multidimensional gas dynamics and magnetohydrodynamics problems kinetic difference scheme
is proposed .
it is convenient from the physics point of view to define the gas dynamics and magnetohydrodynamics quantities from close relations between the kinetic and gas dynamics description of physics processes @xcite .
another aspect is the study of the explicit finite difference schemes , which seem to be preferable for future high performance parallel computing , especially in terms of their simplicity and well adaptability to parallel program realization , including hybrid high performance parallel computing systems .
the weakness of explicit schemes is a strictly limited time step that ensures computational stability .
this restriction becomes critical with the growing number of nodes and the reduction in the step of a spatial mesh .
the advanced explicit kinetic finite difference schemes have a soft stability condition giving the opportunity to enhance the stability and to use very fine meshes @xcite .
the mentioned aspects are used for the development of the framework for the study of the dynamics of the conducting gas media in strong magnetic fields at high performance parallel computing systems .
the kinetic theory describes the gas dynamics by the boltzmann differential equation through the evolution of the distribution function @xmath0 @xcite : + f(,,t)=c(f ) [ boltzmanneq ] where @xmath1 is a nonlinear integral operator which describes the collisions between gas molecules .
this evolution equation follows naturally from the relations between the kinetic and the gas dynamics description of continuous media .
the macroscopic observables such as density , momentum , energy flux as a function of @xmath2 and @xmath3 are obtained from the moments of the distribution function with respect to the macroscopic velocity .
the evolution equations for these gas dynamics quantities are obtained by integrating eq .
( [ boltzmanneq ] ) over molecular velocities @xmath4 with summational invariants @xmath5,@xmath6,@xmath7 .
the computational interest in kinetic formulations of the gas dynamics is high due to the linearity of the differential operator on the left side of eq .
( [ boltzmanneq ] ) .
nonlinearity is confined by the collision term , which is generally local in @xmath2 and @xmath3 .
an important feature is that the collision integral vanishes in the equilibrium state when the local boltzmann distribution function @xmath8 is a maxwellian : f(x , , t ) = exp \ { - ( -u(t , x))^2 } this leads to the use of this model for numerical methods and possible generalizations in order provide a natural kinetic description of the system of conservation laws .
this approximation is sufficient for the gas dynamics processes and is called the kinetic approach @xcite . in @xcite it was shown that electromagnetic fields do not destroy the validity of the boltzmann equation and this opened the way for the implementation of the electromagnetic term in the boltzmann - like distribution function . from the vector nature of the electromagnetic interaction
, the distribution function should taking to account the vector behavior and provide correct formulation for the evolution of the magnetic field , i. e. the magnetic field should be generally defined as the momentum of the boltzmann - like distribution function . a few useful attempts to formulate the vector boltzmann - like distribution function
can be found in @xcite , but physical meaning was not clear defined .
we propose an evaluation of the electromagnetic processes in the context of the distribution function , taking to account the axial nature of the magnetic field .
the electromagnetic field is considered as a complex vector field as proposed in @xcite : f=+i for the purposes of magnetohydrodynamics , the effect , which a magnetic field exerts on a certain volume , is obtained by integrating the electromagnetic stress tensor over the surface of that volume and the correspondent propagation velocity can be defined as a complex vector of velocity : v_em=_em+iw_em at first approximation the term defined of electric forces could be neglected and the magnetic term could be defined through the tension of the magnetic field line and shows a similarity to the alfven wave mechanism : w_em= using the above definitions we define the local complex boltzmann maxwellian distribution function of magnetohydrodynamics with drift velocity @xmath9 in magnetic field @xmath10 at the equilibrium : f_m(,,t ) = \ { - |(-(,t ) ) -i w_em |^2 } [ fmhd ] the first term on the right - hand side of ( [ fmhd ] ) includes the internal energy and the second term is the magnetic field energy .
the hydrodynamics observables are real scalars and vectors .
the complex components include the dynamics of the macroscopic observables introduced by the evolution of the magnetic field , keeping their specific pseudo - vectorial nature .
the magnetogasdynamics observables are obtained as integrals of the distribution function ( [ fmhd ] ) with the summational invariants @xmath11 .
the integration is performed on the path @xmath12 with respect to the molecular velocities @xmath4 in the complex plane correspondent to the value of @xmath10 in the imaginary space .
the relations are obtained for the real and imaginary terms : ( , t)=_m f_m d^3 ( , t)=_m f_m d^3 e(,t)=_m ^2 f_m d^3 ( , t)=-_m ^ f_m d^3 [ hydq ] the proposed complex boltzmann maxwell like distribution function contains the hydrodynamics terms and the electromagnetic terms . thus by using this distribution function to calculate the mass , momentum , energy and magnetic field fluxes , most of the electromagnetic contributions are calculated directly , i.e. one does not have to solve the hydrodynamics and magnetic force components separately or differently , as will be shown below .
to provide the first step of the formulation of the mhd conservation laws equation , the equilibrium state is considered with the proposed distribution function .
the mhd system of equations is obtained by the integration of ( [ boltzmanneq ] ) with vanishing collision integral with the summational invariants following the definition in ( [ hydq ] ) : _ m+_mf d^3=0 _
m+_mf d^3=0 _ m^2+_m^2 f d^3=0
_ m^+_m^f d^3=0 [ mhd3d ] the result obtained , set of eq .
( [ mhd3d ] ) , is the ideal magnetohydrodynamics system of equations : + u_i=0 u_i + = 0 + = 0 + = 0 [ mhd3d ] in addition an equation for @xmath13 is obtained as the imaginary part of the path integral of the summational invariant @xmath14 with respect to the velocities @xmath4 : _ m+_mf d^3=0 [ divb ] = 0 [ divb ]
the model of the kinetic differential schemes is based on the discrete model of evolution of the distribution function .
kinetic schemes are obtained directly from the boltzmann kinetic equation by using the principle of total approximation .
consider the local volume of gas ( cell @xmath15 ) with the distribution function in time @xmath16 . by using the splitting method of particle flow for the cell @xmath15 , the evolution of the distribution function by first order differential scheme for the kinetic boltzmann equation
can be written as : = - + c(f^j ) [ balance_sc1 ] as mentioned before the collision of particles leads to the establishment of the equilibrium state which is adequately described by the single - particle maxwell distribution function with vanishing of the collision integral in the right part of the balance relations . the time evolution of the distribution function can be represented as the time evolution of the local maxwellian distribution function in discrete moments : * at time @xmath16 , on each cell , the locally constant one - particle maxwellian distribution function is defined : + f_m = \ { - |(- ) -i|^2 } + where the magnetohydrodynamics parameters @xmath17,@xmath18,@xmath19,@xmath10 are not varied on the cell . * during the time interval @xmath20 collisionless processes of the gas dynamics occurs , * at time @xmath21 the distribution function is instantaneously maxwellised * for the time @xmath22 these processes are repeated .
the kinetic difference scheme in this case can be written : = - [ balance_sc2 ] or in more general form for the multidimensional case : + _ if_^j_i=|_i|x_i _ i [ balance_sc3 ] where : + @xmath23 is the surface element @xmath24 perpendicular to the direction @xmath25 , + @xmath26 is the value of the distribution function at the surface @xmath27 between the two volume elements @xmath28 and @xmath29 , + @xmath30 the distribution function derivative at the surface @xmath27 between the two volume elements .
the sum in eq .
( [ balance_sc3 ] ) is extended to the 6 surface elements at the boundary of the 3-dimensional rectangular volume element .
the kinetic scheme of the conservation laws of the macroscopic observables for 3d magnetohydrodynamics processes can be obtained by integrating the balance relation ( [ balance_sc3 ] ) with the summational invariants @xmath31 , using the same integration rules as in eq .
( [ mhd3d ] ) : + ( u_i)__i=_|x_ix_i + _ _ k= + _ |x_kx_k + _ _ i = + _ |x_ix_i + ( u_kb_i - u_ib_k)__k = _ |x_kx_k [ mhdscheme ] where @xmath32 , @xmath33 in addition to eq . ( [ mhdscheme ] )
the condition is obtained as the complex part of the path integral of the summational invariants @xmath14 with respect to the molecular velocities @xmath4 in the complex plane : ( b_i)__i=_|x_ix_i dissipative terms appear in the time evolution of the magnetic field which does not preserve the condition @xmath34 and require a specific treatment .
the kinetic quasi magnetohydrodynamics system of equations is closely related to the kinetic scheme and represents a differential form notation for the numerical algorithms .
the balance relation in eq .
( [ balance_sc3 ] ) can be rewritten as : + _ _
if_d=_|_i| x_id=__i^2 d[balance2a ] and using the gauss - ostrogradsky formula it is possible to transform eq .
( [ balance2a ] ) to the differential form : + ( f^j_m ) = _
i_k f^j_m [ balance1 ] here the quasi magnetohydrodynamics system of equation involves explicitly two @xmath35 parameters .
hydrodynamics processes are introduced by the quantity @xmath35 that corresponds to the time of free distance flight of particles , or the characteristic time of particle collisions . by analogy
the quantity @xmath36 is introduced as the characteristic time of propagation of magnetohydrodynamics by electromagnetic processes .
the characteristic time values @xmath35 and @xmath37 are defined respectively for hydrodynamics and electromagnetic processes : = _
m=_m where : + @xmath38 is the size of the computational cell , + @xmath39 , @xmath40 are the sound speed and alphen speed in the computational cell . the introduction of the physical meaning of characteristic times @xmath35 and @xmath36 provides an important contribution to the understanding of the processes and the simplification of the numerical scheme .
the evolution equations for the gas dynamics parameters and for the magnetic field are obtained from eq .
( [ balance1 ] ) by integration with the summation invariants @xmath41 over the molecular velocities , under the assumption : = the integration is performed as in eq .
( [ mhd3d ] ) and eq .
( [ mhdscheme ] ) .
the gas dynamics and magnetic field quantities are obtained respectively as the real and imaginary path of the integral in the complex space .
the compact form of the kinetic quasi magnetohydrodynamics system of equations can be written as : + u_i= + _
ik=_ik^d+w_iu_k + = + ^d_iku_k+(+ ) w_i + m^b_ik=_ik^db [ qgde1 ] the left - hand part of the system of eq .
( [ qgde1 ] ) corresponds to the euler system of equations .
the right - hand of the kinetic quasi mhd eq .
( [ qgde1 ] ) contains dissipative terms . in comparison with other methods ,
the dissipative terms are obtained not by phenomenology with some assumption about magnetohydrodynamics processes but in consistency with the difference scheme of the boltzmann equation .
@xmath42 is the momentum flux density tensor for a perfect gas in magnetic field : _ ik=(p+)_ik+u_iu_k - b_ib_k @xmath43 is the heat transfer flux of a perfect gas in magnetic field : f_i= @xmath44 is the asymmetric product between velocity @xmath9 and magnetic field flux @xmath45 : m^b_ik = u_kb_i - u_ib_k the right hand part of the system of eq .
( [ qgde1 ] ) includes the dissipative terms : w_i==_ik ^d_ik= + + + + + + + _ ik + _ ik + + + q_i^d= + + + + + + + + + + + _
ik^db= + + + + + + + + + + [ dissipative ] the dissipative terms appear because the construction of the quasi magnetohydrodynamics system is based on the assumption that the distribution function slightly changes over the distance between neighborhood cells , what is related to the characteristic times @xmath35 and @xmath36 .
it was shown in @xcite that the dissipative terms of the quasi gas dynamics system are small in comparison with the convective terms with the condition of cell size equivalent to the free path they converge to the viscous terms of the corresponding navier - stokes equations .
the corresponding dissipative terms are associated with real physics processes .
an important remark is that in this case the gas dynamics parameters such as viscosity and heat conductivity are obtained from the kinetic theory .
the navier - stokes viscosity is identified as the first term of eq .
( [ dissipative]b ) : ^ns_ik== where the bulk viscosity component is neglected and the shear viscosity @xmath46 is related to the gas pressure @xmath47 and the characteristic time @xmath35 as @xmath48 .
the navier - stokes thermal flux vector is identified as the first term of eq .
( [ dissipative]d ) : q_i==k with @xmath19 gas temperature and @xmath49 thermal coefficient expressed as @xmath50 , with @xmath51 prandl number .
a similar analysis of the dissipative terms of the electromagnetic processes gives the estimation of their smallness . with correct conditions for the size of cells
the equation converges to the correct representation of magnetic viscosity .
the gas resistivity is identified as the first term of eq .
( [ dissipative]c ) and also appeares as a result of the kinetic formulation : ^b_ik== with the resistivity @xmath52 .
the computational algorithm uses a cartensian , staggered , divergence free mesh configuration .
a detailed description is presented in @xcite , in order to preserve the condition @xmath53 .
[ fig:3d_domain ] shows the four neighbour to the cell ( @xmath54 ) used in evaluations of the hydrodynamics and electromagnetic variables .
the hydrodynamics observables - mass density , momentum and energy density are defined at the cell center .
the components of the magnetic field are defined at the face centers of the cells .
a duality is established between the electric field and the fluxes .
this duality is utilized to obtain the electric field at the edges of the computational cell through a reconstruction process that is applied directly to the properly upwinded fluxes .
the electric field is then utilized to make an update of the magnetic fields that preserves the solenoidal nature of the magnetic field and ensures that the magnetic field in a magnetohydrodynamics model remains strictly solenoidal up to discretization errors .
generally the explicit numerical scheme is used model , considering that it is perspective for the modern high performance computing systems due to the logical simplicity and efficiency of the algorithms .
the finite volume method is used to update the conserved observables , mass , momentum and energy , by calculating the fluxes of this observables across the cell face .
updating the magnetic field is a more complicated procedure and is performed via electric field integration along the edge of the cells , as showh on fig .
[ fig:3d_domain ] .
the distribution function method is proposed in the calculation as described above .
the explicit scheme is used in the time evolution for the integration of the quasi magnetohydrodynamics system of equations .
the code uses a variable time step .
the time step in an explicit scheme is controlled by a courant type conditions on the time step estimation @xcite .
the computational framework is created on the basis of fortran 90 and c++ , with parallel implementation on mpi .
the demonstration of the performance of the method is performed on the basis of the solution of the spherical expansion problem of ionised gas and the solution of the expansion of an ionised gas in strong magnetic field .
the simulations are performed for a cartesian rectangular mesh @xmath55 in the physics domain [ 0,1 ] .
the initial conditions consist of a sphere with radius 0.1 placed in the center of the physical region with pressure of 100 in comparison to the overall represented area with pressure 1 . for the study of ionized gas in a strong magnetic field the uniform magnetic field aligned with the @xmath56 coordinate is added to the initial conditions .
[ fig:3d_exp],[fig:3d_b_exp ] present the state of the 3d simulation of the processes for relative time 0.03 . on the 3d pictures
the arrows represent the velocities of the ionised gas and the color represents the density of gas .
3d figures clearly show the confinement of the ionized gas in the cylindrical area along @xmath56 due to the magnetic field . fig .
[ fig:2d_proj_1],[fig:2d_proj_2 ] represent the @xmath57 projections of the density , pressure and kinetic energy of the gas expansion without magnetic field and the @xmath58 density profile for these condition .
[ fig:2d_b_proj_1],[fig:2d_b_proj_2 ] represents the @xmath57 projections of the density , pressure , magnetic pressure and kinetic energy for the gas expansion problem of the ionized gas with initial magnetic field of @xmath59 and fig .
[ fig:1d_b_proj_1 ] shows the @xmath58 profile of the density for these conditions .
[ fig:2d_b1_proj_1],[fig:2d_b1_proj_2 ] represent the @xmath57 projections of the density , pressure , magnetic pressure and kinetic energy for the gas expansion of the ionized gas in the strong magnetic field of @xmath60 at time 0.01 and fig .
[ fig:1d_b1_proj_1 ] represents the 1d profile of the density for these conditions .
view of the conductive gas expancion in strong magnetic field , scaledwidth=50.0% ] view of the conductive gas expancion in strong magnetic field , scaledwidth=50.0% ] kinetic energy and 1d density profile , title="fig:",scaledwidth=35.0% ] kinetic energy and 1d density profile , title="fig:",scaledwidth=35.0% ] kinetic energy and 1d density profile , title="fig:",scaledwidth=35.0% ] kinetic energy and 1d density profile , title="fig:",scaledwidth=35.0% ] density profile in the magnetic field @xmath59 , title="fig:",scaledwidth=35.0% ] density profile in the magnetic field @xmath59 , title="fig:",scaledwidth=35.0% ] density profile in the magnetic field @xmath59 , title="fig:",scaledwidth=35.0% ] density profile in the magnetic field @xmath59 , title="fig:",scaledwidth=35.0% ] density profile in the magnetic field @xmath59 , scaledwidth=35.0% ] projection of the density in the magnetic field @xmath60 , title="fig:",scaledwidth=35.0% ] projection of the density in the magnetic field @xmath60 , title="fig:",scaledwidth=35.0% ] projection of the density in the magnetic field @xmath60 , title="fig:",scaledwidth=35.0% ] projection of the density in the magnetic field @xmath60 , title="fig:",scaledwidth=35.0% ] projection of the density in the magnetic field @xmath60 , scaledwidth=35.0% ] similar studies of the problem of spherical explosion including the conditions with magnetic field are presented in @xcite .
the comparison of the results shows a reasonable agreement and will be analysed further .
a new 3d kinetic algorithm has been developed for the solution of the magnetohydrodynamics problems .
the novel feature of the method is that the local complex boltzmann - like distribution function incorporated most of the electromagnetic processes terms .
the fluxes of mass , momentum and energy across the cell interface as well as the magnetic field are calculated by integrating a local complex boltzmann - like distribution function over the velocity space .
thus by using this distribution function to calculate the mass , momentum and energy fluxes , most of the electromagnetic contributions are calculated directly , i.e. one does not have to solve the hydrodynamics and magnetic force components separately or differently .
|
the impressive progress of the kinetic schemes in the solution of gas dynamics problems and the development of effective parallel algorithms for modern high performance parallel computing systems led to the development of advanced methods for the solution of the magnetohydrodynamics problem in the important area of plasma physics .
the novel feature of the method is the formulation of the complex boltzmann - like distribution function of kinetic method with the implementation of electromagnetic interaction terms .
the numerical method is based on the explicit schemes . due to logical simplicity and its efficiency ,
the algorithm is easily adapted to modern high performance parallel computer systems including hybrid computing systems with graphic processors .
magnetohydrodynamics ( mhd ) , kinetic scheme , high performance computing
|
smoothness plays an important role in the theory of codimension one foliations of 3-manifolds .
reeb constructed the first @xmath3 foliation on @xmath4 as the union of two foliated solid tori , or reeb components @xcite .
this work of reeb , together with work of alexander @xcite and wallace @xcite , led to the proofs by lickorish @xcite and novikov and zieschang @xcite that any closed 3-manifold has a @xmath3 codimension one foliation . on the other hand
, haefliger @xcite showed that no foliation of @xmath4 can be analytic .
this was greatly improved by novikov @xcite to show that any @xmath1 foliation of @xmath4 must have reeb components , and these never exist in analytic foliations .
the qualitative nature of foliation theory and its impact on the ambient 3-manifold was considerably advanced by thurston s introduction of the norm on the homology of a 3-manifold , and in particular the minimizing properties of leaves of taut , transversely oriented , @xmath1 foliations @xcite .
a foliation is _ taut _ if closed smooth transversals to the foliation pass through every point of the manifold .
this is also known as _ everywhere taut _ to distinguish it from the more familiar notion of _ smoothly taut _ in which closed smooth transversals are only required to intersect every leaf of the foliation . for a discussion of these and other notions of tautness ,
why they are different for @xmath5 foliations , the same for @xmath6 foliations , and interchangeable up to @xmath7 approximation and isotopy of foliations , see @xcite .
foliations as a tool for understanding problems in 3-dimensional topology came to the fore as a result of gabai s constructions of both @xmath8 and often less smooth , but finite depth , taut foliations @xcite .
the success of gabai s applications of foliation theory led to many constructions of taut codimension one foliations .
often these foliations are constructed using denjoy blowup techniques that yield foliations that are only @xmath9 ; that is , leaves are smoothly immersed , but transversely , their tangent plane fields vary only continuously .
the impetus for our work starts with the eliashberg - thurston approximation theorem @xcite .
they showed that a taut , co - oriented codimension one @xmath1 foliation of a 3-manifold can be @xmath7 approximated by a pair of symplectically fillable contact structures .
this allows non - trivial heegaard - floer invariants to be assigned to any manifold that supports a taut foliation @xcite .
this is , consequently , one of the pillars of the conjectural relationship between @xmath10-spaces , taut foliations , and left orderability of the fundamental group . for details ,
see , for example @xcite and @xcite .
in @xcite and @xcite we extended the eliashberg - thurston approximation theorem to the class of all @xmath11 , co - oriented taut foliations , thereby extending its reach to manifolds carrying the new constructions of foliations mentioned above .
similar results can be found in @xcite . in doing so
, we found that many of the standard tools for working with foliations had either not been developed for foliations with lesser smoothness than originally intended , or had not been developed with an eye towards @xmath7 approximation theory in which it is often necessary to produce a new foliation while only moving the tangent planes of the original foliation slightly .
this paper includes enhancements to well - known results of calegari @xcite on smoothing leaves , dippolito @xcite on denjoy blowup of leaves , and tischler @xcite on approximations by fibrations .
it is possible that some of our results can be obtained by `` reading between the lines '' of the original source .
however , as is well - known , subtleties , sometimes fatal , arise when smooth objects are replaced by objects that are merely continuous .
( see , for example , @xcite . )
these are foundational results in foliation theory , and proofs of these theorems , in the generality in which they are used , do not exist in the literature .
an advantage to the flow box decomposition approach we use is that each of these results can be proved directly with a single inductive strategy .
the methods developed in this paper are not intrinsically 3-dimensional techniques , and we expect they can be adapted to prove corresponding results for codimension one foliations in @xmath2-dimensional manifolds . basic definitions ( codimension one foliation , flow , @xmath12 compatible , @xmath7 close , and @xmath7 small ) are given in 2 . in
3 we recall the definition of flow box decomposition , define regular neighborhood structure , and prove a sequence of useful local smoothing results .
the main result of 4 is a proof that any @xmath11 foliation is isotopic to a @xmath7 close @xmath9 foliation . in 5 we prove that any @xmath11 measured foliation is isotopic to a @xmath7 close smooth measured foliation .
basic facts from @xcite about holonomy neighborhoods are recalled in 6 .
we give dippolito s definition @xcite of denjoy blowup in 7 and prove that particularly nice , @xmath7 close , denjoy blowups of a @xmath11 codimension one foliation always exist . throughout this paper ,
unless stated otherwise , @xmath0 will denote a 3-manifold that is either smooth or smooth with corners . when @xmath13 , it is often useful to think of @xmath0 as a sutured manifold , not necessarily orientable , in the sense of @xcite . recall that any topological 3-manifold admits a smooth structure , unique up to diffeomorphism @xcite .
we begin by defining foliations in 3-manifolds with empty boundary . near the end of this section , we extend these definitions to 3-manifolds with nonempty boundary that are smooth or smooth with corners ; namely , manifolds locally modelled by open sets in @xmath14 .
[ folndefn1 ] let @xmath0 be a smooth 3-manifold with empty boundary .
let @xmath15 and @xmath16 be non - negative integers or infinity with @xmath17 .
both @xmath18 and @xmath19 _ codimension one foliations _
@xmath20 are decompositions of @xmath0 into a disjoint union of @xmath18 immersed connected surfaces , called the _ leaves _ of @xmath20 , together with a collection of charts @xmath21 covering @xmath0 , with @xmath22 a homeomorphism , such that the preimage of each component of a leaf intersected with @xmath21 is a horizontal plane .
the foliation @xmath20 is @xmath18 if the charts @xmath23 can be chosen so that each @xmath24 is a @xmath18 diffeomorphism .
the foliation @xmath20 is @xmath19 if for all @xmath25 and @xmath26 , 1 .
the derivatives @xmath27 , taken in any order , on the domain of each @xmath24 and each transition function @xmath28 are continuous for all @xmath29 , and @xmath30 , and 2 .
if @xmath31 , @xmath24 is a @xmath32 diffeomorphism . the smoothness conditions on both the charts and the transition functions are to ensure that the smooth structure on the leaves is compatible with the smooth structure on @xmath0 .
in particular , @xmath33 exists and is continuous if and only if @xmath20 is @xmath11 .
also notice that @xmath19 foliations are @xmath34 , but not conversely .
two @xmath5 foliations @xmath20 and @xmath35 of @xmath0 are called _
@xmath5 equivalent _
if there is a self - homeomorphism of @xmath36 that maps the leaves of @xmath20 to the leaves @xmath35 , and is @xmath18 when restricted to any leaf of @xmath20 .
we use the terms _ transverse , transversal , _ and _ transversely _ in the smooth sense ; that is , they refer to smooth objects intersecting so that the associated tangent spaces intersect minimally . in contrast , a curve is _ topologically transverse _ to @xmath20 if no nondegenerate subarc is isotopic , relative to its endpoints , into a leaf of @xmath20 . given a codimension one foliation @xmath20 , it is useful to fix a one dimensional foliation @xmath37 transverse or topologically transverse to @xmath20 .
such a @xmath37 always exists and can be realized as the union of curves @xmath38 of continuous local flows @xmath39 .
when @xmath20 is transversely oriented , @xmath37 can be realized as the union of curves @xmath38 of a global flow @xmath40 .
when @xmath20 is @xmath7 , this is proved in @xcite ( theorems 1.1.2 and 1.3.2 ) .
when @xmath20 is @xmath11 , @xmath37 can be chosen to be smooth ; in fact , in this case , it is elementary to see that @xmath37 exists and consists of the integral curves of a smooth line field transverse to @xmath33 .
see , for example , lemma 5.1.1 of @xcite .
* conventions : * unless otherwise stated , throughout the rest of this paper , `` foliation '' will mean a codimension one transversely oriented foliation of a 3-manifold @xmath0 .
such @xmath0 will be assumed to be compact and oriented .
since the foliations studied will be @xmath11 we can assume without loss of generality that a smooth transverse flow to the foliation is chosen . to simplify the exposition ,
we will abuse language , and refer to a one dimensional foliation transverse to a codimension one foliation as a flow .
when a foliation @xmath20 is understood , a submanifold of positive codimension in @xmath0 is called _ horizontal _ if each component is a submanifold of a leaf of @xmath20 and _ vertical _ if each component is transverse to @xmath20 .
when both a foliation @xmath20 and a transverse flow @xmath37 in @xmath0 are understood , a submanifold of positive codimension in @xmath0 is called _ vertical _ if and only if it can be expressed as a union of subsegments of the flow @xmath37 .
a codimension-0 submanifold @xmath41 of @xmath0 is called _ @xmath12-compatible _ if its boundary is piecewise horizontal and vertical , and hence @xmath20 and @xmath37 restrict naturally to foliation and flow on @xmath41 .
if @xmath41 is @xmath12-compatible , let @xmath42 denote its vertical boundary , and let @xmath43 denote its horizontal boundary .
suppose @xmath41 is an @xmath12-compatible submanifold of @xmath0 , where possibly @xmath44 .
an isotopy of @xmath41 which maps each flow segment of @xmath45 to itself is called a _ flow compatible _ , or _
@xmath37 compatible , isotopy_. note that a flow compatible isotopy of @xmath41 fixes @xmath43 pointwise . by allowing the foliation atlas to include boundary charts
, definition [ folndefn1 ] naturally extends to the case that @xmath0 has nonempty boundary that is either smooth or smooth with corners .
smooth boundary components must either be a leaf of @xmath20 , and hence horizontal , or transverse to @xmath20 , and hence vertical .
a boundary component with corners must decompose along its corners into smooth subsurfaces , where if two subsurfaces share a corner , one is horizontal and one is vertical .
such an @xmath0 is a sutured manifold , in the sense of @xcite .
thus , if @xmath13 and we double @xmath46 along @xmath47 , @xmath48 is a foliation of @xmath49 with all components of @xmath50 leaves of @xmath20 .
we restrict attention to flows @xmath37 that meet @xmath51 in a similarly constrained way .
a flow is required to be either everywhere transverse or everywhere tangent to a smooth component of @xmath51 . and
if @xmath52 is a boundary component with annular sutures @xmath53 , a flow is required to be transverse to @xmath54 and tangent to @xmath53 . in particular , if @xmath37 is a flow transverse to @xmath20 , it is possible to double @xmath37 along @xmath47 so that @xmath0 is a @xmath55-compatible submanifold of @xmath49 .
the terms @xmath7 close and @xmath7 small both refer to distances between tangent planes . more formally ,
suppose that a metric , @xmath56 , has been chosen on the set of continuous sections of the grassmann bundle of 2-planes in @xmath57 .
given a section , typically the tangent bundle of a foliation , @xmath33 , we say that another section , @xmath58 , can be found _
close _ to @xmath33 , if for all @xmath59 a @xmath35 exists such that @xmath60 . for brevity , this is stated as , @xmath35 can be found _
close _ to @xmath20 .
an isotopy @xmath61 of @xmath20 is called _ @xmath7 small _ if it can be chosen so that at all times @xmath61 is @xmath7 close to @xmath20 .
an isotopy can be found _
@xmath7 close to the identity _ if given any @xmath59 an isotopy can be found that keeps every section within @xmath62 of its starting position . throughout the paper , @xmath63
will be used to denote the closed interval @xmath64 $ ] .
flow box decompositions were introduced and shown to exist in @xcite . in the definition given below , an extra condition , ( 5 ) , is added that is particularly useful for inductive arguments .
@xcite [ flowboxdefn ] let @xmath20 be either a @xmath18 or @xmath19 foliation , and let @xmath37 be a smooth transverse flow .
a _ flow box _ ,
@xmath65 , is an @xmath12 compatible closed chart , possibly with corners .
that is , it is a submanifold diffeomorphic to @xmath66 , where @xmath67 is either a closed @xmath18 disk or polygon ( a closed disk with at least three corners ) , @xmath37 intersects @xmath65 in the arcs @xmath68 , and each component of @xmath69 is embedded in a leaf of @xmath20 .
the components of @xmath70 give a family of @xmath18 graphs over @xmath67 .
in the case that @xmath67 is a polygon , it is often useful to view the disk @xmath67 as a 2-cell with @xmath71 the cell complex obtained by letting the vertices correspond exactly to the corners of @xmath67 .
similarly , it is useful to view the flow box @xmath65 as a 3-cell possessing the product cell complex structure of @xmath66
. then @xmath72 is a union of two ( horizontal ) 2-cells and @xmath73 is a union of @xmath74 ( vertical ) 2-cells , where @xmath74 is the number of corners of @xmath67 . in the case
that @xmath67 has no corners , we abuse language slightly and consider @xmath72 to be a union of two ( horizontal ) 2-cells and @xmath73 to be a single vertical face , where the face is the entire vertical annulus @xmath75 .
suppose @xmath76 is either empty or else a compact , @xmath12 compatible , codimension-0 submanifold of @xmath0 .
a _ flow box decomposition of @xmath0 rel @xmath76 _ , or simply _ flow box decomposition of @xmath0 _ , if @xmath77 , is a decomposition of @xmath78 as a finite union @xmath79 where 1 .
each @xmath80 is a flow box , 2 .
@xmath81 is a union , possibly empty , of horizontal subsurfaces and vertical 2-cells of @xmath82 , together possibly with some 0- and 1-cells , 3 .
the interiors of @xmath80 and @xmath83 are disjoint if @xmath84 , 4 . if @xmath85 and @xmath86 is nonempty , it must be homeomorphic to a point , an interval , or a disk that is wholly contained either in @xmath87 or in a single face in each of @xmath88 and @xmath89 , and 5 . if @xmath90 is a vertical 2-cell of @xmath91 and the interior of @xmath90 intersects a vertical 2-cell @xmath92 of some @xmath82 with @xmath93 , then @xmath94 .
most of the results proved in this paper use flow box decompositions relative to an empty codimension-0 submanifold .
the general definition is particularly useful for approximating foliations by contact structures , as described in @xcite , and it appears in support of that work in corollary [ theorem5.2 ] . [ existence of fbd ] suppose @xmath20 is either a @xmath18 or a @xmath19 foliation of a compact manifold @xmath0 , and let @xmath37 be a smooth flow transverse to @xmath20 .
suppose @xmath76 is either empty or else a compact , @xmath12 compatible , codimension-0 submanifold of @xmath0 .
then @xmath0 has a flow box decomposition rel @xmath76 .
moreover , any flow box decomposition of @xmath76 can be extended to a flow box decomposition of @xmath0 .
conditions ( 1)(4 ) follow from proposition 4.4 of @xcite .
thus it is enough to show that a flow box decomposition satisfying ( 1)(4 ) can be inductively subdivided so that ( 5 ) is satisfied . to do this , consider the union , @xmath41 , of all vertical 2-cells contained in an @xmath82 with @xmath93 that intersect the interior of some vertical 2-cell of @xmath95 .
split @xmath95 along a finite collection of leaves of @xmath96 that contain @xmath97 , and let @xmath98 be the resulting components . redefine the polygonal structure on each @xmath98 by decreeing that , in addition to the original vertical edges , every component of @xmath99 is also a vertical edge .
replacing @xmath95 by the @xmath98 completes the inductive step of the construction .
a flow box decomposition is called _ @xmath76-transitive _ , or _ transitive _ , when @xmath77 , if @xmath100 , @xmath101 , and for @xmath102 , * @xmath103 contains a vertical 2-cell of @xmath82 . condition ( 6 ) is used in @xcite where flow boxes were needed to laterally propagate an approximating contact structure from @xmath76 to the rest of @xmath0 .
[ transitiveflowbox ] if @xmath0 is compact and each point in @xmath0 can be reached from @xmath76 by a path in a leaf of @xmath20 , then there is a transitive flow box decomposition of @xmath0 rel @xmath76 .
if @xmath104 is a flow box decomposition of @xmath0 rel @xmath76 , an isotopy of @xmath0 is _
@xmath105 compatible _ if it is @xmath37 compatible and , in addition , maps each cell of each flow box of @xmath105 to itself setwise . by condition ( 5 ) , the set of vertical faces of the flow boxes @xmath82 is partially ordered by set containment ; if @xmath106 and @xmath107 are vertical faces of @xmath82 and @xmath108 , respectively , and their interiors have nonempty intersection , then @xmath109 or @xmath110 .
call a vertical face _
maximal _ if it is maximal with respect to this partial ordering ; namely , if it is not properly contained in any vertical face .
let @xmath111 be a listing of the maximal faces
. it will sometimes be helpful to consider a regular neighborhood of @xmath112 of the following sort .
[ rnd defn ] let @xmath113 be a listing of the flow boxes of a flow box decomposition @xmath114 . a _ regular neighborhood structure _ @xmath115 for @xmath105 is a tuple of the form @xmath116 where 1 .
@xmath111 is a listing of the maximal faces of @xmath105 , 2 .
each @xmath117 is a flow box that properly contains @xmath118 , 3 .
@xmath119 is a @xmath12 compatible regular neighborhood of the union of the vertical 1-cells of the maximal faces @xmath118 , 4 .
@xmath119 decomposes as a finite union of flow boxes @xmath120 , where @xmath121 , and @xmath122 for each @xmath123 , 5 . if @xmath124 , then @xmath125 is contained in the interior of @xmath119 , 6 .
@xmath126 , and 7 .
@xmath127 is a deformation retract of @xmath128 .
figure [ regnbdsection ] illustrates a horizontal cross section of a regular neighborhood structure in a neighborhood of a single flow box .
note that a @xmath105 compatible isotopy takes a regular neighborhood structure @xmath129 to a regular neighborhood structure @xmath130 .
a standard method of proof is to work inductively with a cell complex , working first with 0-cells , and then extending over the 1-cells , followed by the 2-cells , and finally the 3-cells . when smoothness is a priority , it is often useful to work instead with regular neighborhoods of the cells .
regular neighborhood structures provide a vocabulary for this approach in the context of flow box decompositions ; namely , establish a property first on @xmath119 , then on the union @xmath131 , and finally extend this property over the 3-cells complementary to @xmath132 . a flow box decomposition @xmath105 is _ smooth sided _ if the interior of every vertical face of every flow box @xmath82 of @xmath105 is a smooth surface .
the flow box decomposition is called _ smooth _ if it is smooth - sided and every horizontal face has a neighborhood in the leaf it is contained in that is smoothly embedded .
[ smooth - sided ] let @xmath0 be compact .
if @xmath20 is @xmath11 and @xmath37 is a smooth transverse flow in @xmath0 , then there exists a smooth - sided flow box decomposition of @xmath0 .
if @xmath20 is @xmath9 and @xmath37 is a smooth transverse flow in @xmath0 , then there exists a smooth flow box decomposition of @xmath0 .
the proof of proposition 4.4 of @xcite starts by choosing initial flow boxes and these may be taken to be smooth sided .
the rest of the construction involves transversality of vertical intersections and splitting along leaves , and both of these operations work with smooth vertical faces .
[ smooth fbd ] if @xmath20 is @xmath11 , @xmath37 is a smooth transverse flow in @xmath0 , and @xmath105 is a smooth - sided flow box decomposition , then there exists a flow compatible isotopy that takes @xmath20 to a @xmath7 close @xmath11 foliation and takes @xmath105 to a smooth flow box decomposition .
let @xmath133 be the union of small neighborhoods , in leaves of @xmath20 , of each of the horizontal faces of all @xmath134 .
then @xmath133 is a @xmath32 embedded surface .
this may be isotoped , while preserving flow lines of @xmath37 and keeping @xmath20 @xmath7 @xmath62 close to itself , to a smoothly embedded surface .
applying this isotopy to @xmath105 produces the desired smooth flow box decomposition .
suppose @xmath105 is a smooth @xmath12 flow box decomposition , where @xmath20 is a @xmath5 foliation for some @xmath135 , and @xmath37 is a smooth flow transverse to @xmath37 .
each flow box describes a smooth chart for @xmath0 in which the flow restricts to the union of vertical segments @xmath136 and the leaves of @xmath20 restrict to a @xmath7 family of @xmath18 graphs . after fixing a point @xmath137 ,
an index @xmath138 may be chosen so that the leaf containing @xmath139 is given by the graph @xmath140 .
we now give several elementary and frequently used smoothing operations that will be used in a neighborhood of a surface .
to streamline statements , let @xmath141 be a surface , possibly with boundary , and let @xmath142 .
strictly horizontal foliation of @xmath143 _ is the foliation with leaves @xmath144 .
an _ almost horizontal foliation _ of @xmath143 is a foliation transverse to the @xmath63 fibers which contains @xmath145 as leaves .
a product submanifold , @xmath143 of @xmath0 , is called an _ @xmath12 compatible product _
if the restriction of @xmath20 to @xmath143 is almost horizontal , and the @xmath63-fibers @xmath146 are flow segments of @xmath37 .
[ nbd notation ] if @xmath141 is a proper subsurface of a leaf of a given foliation , @xmath147 will denote a regular neighborhood of @xmath141 in its leaf .
denote by @xmath148 an extension of the product structure on @xmath149 .
if @xmath143 is a @xmath12 compatible product , @xmath150 denotes a @xmath12 compatible product . for some estimates , it is necessary to consider a smooth damping function @xmath151 that vanishes on @xmath152 and is identically @xmath153 on @xmath141 .
the results established in the rest of this section are elementary , useful , and substantially similar .
they can be grouped as existence of approximations , existence of extensions , or uniqueness results , and they all occur in the context of @xmath7 approximation .
these results are typically applied when @xmath141 is a surface and @xmath143 is a subset of @xmath0 . given a @xmath154 foliation on all , or just a portion of @xmath149 , we ask if it can be extended to all of @xmath149 , if it can be approximated by a smoother foliation , and to what extent the approximating foliation is unique . complicating the statements , though
not the proofs , many of the results when applied require a relative version .
vertically , when @xmath141 has a boundary , the goal is to not change the given foliation near @xmath155 .
horizontally , there may be leaves on which a given foliation is as smooth as needed and should not be changed . in the next proposition
, we establish that @xmath11 almost horizontal foliations of @xmath143 are @xmath7 close to @xmath8 almost horizontal foliations . [ smoothing a product foliation ] [ smooth interpolation ]
let @xmath141 be a compact smooth surface , and let @xmath156 be a @xmath11 almost horizontal product foliation on @xmath149 .
then @xmath156 can be @xmath7 deformed to a @xmath7 close , smooth , almost horizontal product foliation @xmath157 on @xmath149 .
if @xmath156 is smooth on some compact @xmath158 compatible submanifold , then we may choose the deformation to fix this submanifold pointwise .
in addition , if some finite number of leaves of @xmath20 are smoothly embedded , then the deformation can be chosen to fix these leaves pointwise , and if @xmath159 are subsurfaces of leaves @xmath160 of @xmath20 so that regular neighborhoods @xmath161 of @xmath162 in @xmath163 are smoothly embedded in @xmath0 , then the deformation can be chosen to fix each @xmath163 pointwise .
pick a metric on the bundle of tangent two planes to @xmath149 .
fix a point @xmath164 in @xmath141 , and denote the leaf of @xmath156 that contains @xmath165 by @xmath166 . given @xmath167 and @xmath168 ,
let @xmath169_x$ ] denote the subinterval of @xmath170 with boundary points in @xmath171 .
each @xmath166 is the graph of a @xmath32 function @xmath172 .
it is enough to deform the continuously varying family @xmath173 to a smoothly varying family whose graphs foliate @xmath149 .
let @xmath174 .
choose a partition @xmath175 of @xmath64 $ ] with the property that for all @xmath168 , for each i , the tangent planes to leaves of @xmath156 at each point of @xmath176_x$ ] are all within @xmath62 of each other . perform a @xmath7 small isotopy so that @xmath177 is smoothly embedded .
if some finite number of leaves of @xmath178 are smoothly embedded or contain subsurfaces smoothly embedded in @xmath0 , the partition can be chosen so that these leaves appear as @xmath179 .
let @xmath180 \to [ 0,1]$ ] be a smooth bijection that vanishes to infinite order at the endpoints . for @xmath181 $ ]
define @xmath182 then @xmath183 is a smooth family of functions whose graphs give a smooth foliation @xmath157 of @xmath143 .
since @xmath183 is a linear combination of @xmath184 and @xmath185 , it is easily checked using local coordinates on @xmath141 that the normal vector to the graph of @xmath183 is a linear combination of normals to @xmath184 and @xmath185 .
it follows that tangent planes to @xmath183 @xmath7 approximate the tangent planes to @xmath173 .
since @xmath173 and @xmath183 are graphs , there is an @xmath63 fiber preserving deformation of @xmath156 to @xmath157 .
sometimes , the `` smoothing '' of a horizontal foliation is required to preserve an existing structure , and @xmath9 smoothing is the best that can be hoped for .
the proof of proposition [ smooth interpolation ] modifies easily to yield the following .
[ @xmath9 smoothing a product foliation ] [ smoothleaf interpolation ] let @xmath141 be a compact smooth surface , and let @xmath156 be a @xmath11 almost horizontal product foliation on @xmath149 .
then @xmath156 can be @xmath7 deformed to a @xmath7 close , @xmath9 , almost horizontal product foliation @xmath157 on @xmath149 .
if @xmath156 is @xmath9 on some compact @xmath158 compatible submanifold , then we may choose the deformation to fix this submanifold pointwise .
in addition , if some finite number of leaves of @xmath20 are smoothly embedded , then the deformation can be chosen to fix these leaves pointwise , and if @xmath159 are subsurfaces of leaves @xmath160 of @xmath20 so that regular neighborhoods @xmath161 of @xmath162 in @xmath163 are smoothly embedded in @xmath0 , then the deformation can be chosen to fix each @xmath163 pointwise .
[ local product smoothing in @xmath0][local smoothing ] let @xmath20 be a @xmath11 ( respectively , @xmath9 ) foliation of @xmath0 , and @xmath37 a smooth flow transverse to @xmath20 .
suppose that @xmath143 is a @xmath12 compatible product , smoothly embedded in @xmath0 , and fix @xmath147 .
then there is a @xmath7 small , @xmath37-compatible , isotopy of @xmath0 that is the identity outside @xmath150 , and takes @xmath20 to a @xmath7 close @xmath11 ( respectively , @xmath9 ) foliation that is @xmath8 on @xmath187 .
if @xmath105 is a smooth flow box decomposition of @xmath0 , then this isotopy can be chosen to be @xmath105 compatible .
if some finite number of leaves of @xmath20 are smoothly embedded , then the deformation can be chosen to fix these leaves pointwise . moreover ,
if @xmath159 are subsurfaces of leaves @xmath160 of @xmath20 so that regular neighborhoods @xmath161 of @xmath162 in @xmath163 are smoothly embedded in @xmath0 , then the deformation can be chosen to fix each @xmath162 pointwise .
if @xmath188 , this follows immediately from proposition [ smooth interpolation ] . restrict attention , therefore , to the case that @xmath141 is a proper subsurface of a leaf of @xmath20 . fix a riemannian metric on @xmath0 . choose a regular neighborhood @xmath147 of @xmath141 , and product structure
@xmath150 extending the product structure @xmath143 , so that the restriction of @xmath20 to @xmath150 remains almost horizontal .
the restriction of @xmath20 to @xmath150 is therefore a product foliation , @xmath156 say .
let @xmath151 be a fixed smooth damping function that vanishes on @xmath152 and is identically @xmath153 on @xmath141 . by proposition [ smooth interpolation ]
, @xmath156 can be @xmath7 deformed to a @xmath7 close , @xmath8 , almost horizontal product foliation @xmath157 on @xmath148 .
if some finite number of leaves of @xmath20 are smoothly embedded , then the deformation can be chosen to fix these leaves pointwise . moreover ,
if @xmath159 are subsurfaces of leaves @xmath160 of @xmath20 so that regular neighborhoods @xmath161 of @xmath162 in @xmath163 are smoothly embedded in @xmath0 , then the deformation can be chosen to fix each @xmath162 pointwise . using the notation found in the proof of proposition [ smooth interpolation ] , let @xmath173 and @xmath183 , @xmath189 , be functions over @xmath147 whose graphs give the leaves of @xmath156 and @xmath157 , respectively . if @xmath173 describes the graph of the leaf @xmath163 for some @xmath25 , then @xmath190 along @xmath162 .
note that given any @xmath191 , the proof of proposition [ smooth interpolation ] guarantees that we may choose the smoothly varying family @xmath183 so that both the graphs @xmath183 and @xmath173 and their tangent plane fields are @xmath62 @xmath7 close , for each @xmath192 $ ] .
let @xmath193 $ ] and define @xmath194 by @xmath195 for each @xmath196 , points of @xmath148 are uniquely expressible as @xmath197 , thus @xmath198 defines a fiber preserving isotopy of @xmath148 that is the identity in a neighborhood of @xmath199 and takes @xmath156 to @xmath157 on @xmath143 .
since we can choose @xmath157 to be arbitrarily @xmath7 close to @xmath156 , we can guarantee that this isotopy is @xmath7 small , and the resulting foliation is @xmath7 close to @xmath20 . to see this ,
let @xmath193 $ ] and consider @xmath200 suppose @xmath201 is given locally on @xmath141 by real coordinates @xmath202 .
taking the partial with respect to @xmath203 gives @xmath204 a symmetric equation gives the derivative with respect to @xmath205 .
we start with a fixed @xmath206 , and , for any @xmath191 , we may choose the smoothly varying family @xmath183 so that both the graphs @xmath183 and @xmath173 and their tangent plane fields are @xmath62 @xmath7 close , for each @xmath192 $ ] .
therefore , we can guarantee that both the graphs @xmath198 and @xmath173 and their tangent plane fields are @xmath62 @xmath7 close , for each @xmath207 $ ] .
[ local smoothing - uniqueness ] [ local smoothing unique ] let @xmath20 be a @xmath11 foliation of @xmath0 , and @xmath37 a smooth flow transverse to @xmath20 .
suppose that @xmath143 is a @xmath12 compatible product , smoothly embedded in @xmath0 , and fix @xmath147 .
suppose @xmath157 is an almost horizontal product foliation of @xmath149 .
there is an isotopy of @xmath20 , fixing the complement of @xmath148 , that takes @xmath156 to @xmath157 .
the distance that a tangent plane to @xmath20 moves during this isotopy depends on the distance between the tangent planes to @xmath20 and @xmath157 on @xmath148 and the maximum of the derivative of the damping function used .
this will follow from a more general statement , proposition [ uniqueness in a product ] that allows non - trivial holonomy .
let @xmath20 be an almost horizontal foliation on @xmath149 .
if @xmath208 is an arc with @xmath209 and @xmath210 , let @xmath211 denote the homeomorphism from @xmath212 given by lifting @xmath213 to leaves of @xmath20 .
more precisely , given such an arc @xmath213 , let @xmath214 be the path in a leaf of @xmath20 that ends at @xmath215 and projects to @xmath213 , and define @xmath216 .
let @xmath217 $ ] be the path homotopy class of @xmath213 , and define @xmath218)=\rho_\mathcal f(\beta)$ ] .
restricting attention to loops in @xmath141 based at @xmath219 , we obtain the _ holonomy representation _ @xmath220 .
[ uniqueness in a product][uniqueness in a product ] let @xmath20 and @xmath35 be foliations of @xmath0 that restrict to almost horizontal foliations of @xmath148 . if the holonomy representations of @xmath20 and @xmath35 agree on @xmath149 , then there is a @xmath7 isotopy of @xmath0 that preserves @xmath63 bundle fibers , is supported in a neighborhood on @xmath148 , and on @xmath149 takes @xmath20 to @xmath35 . the distance that a tangent plane to @xmath20 moves during this isotopy depends on the distance between the tangent planes to @xmath20 and @xmath35 on @xmath148 and the maximum of the derivative of the damping function used .
fix @xmath219 in @xmath141 . for each @xmath168 , choose a path @xmath213 in @xmath141 from @xmath219 to @xmath201 , and define a homeomorphism @xmath221 of @xmath170 by @xmath222 .
then @xmath221 gives a well defined @xmath63 fiber preserving homeomorphism of @xmath143 taking @xmath20 to @xmath35 that can be used to produce the desired straight line isotopy . in the next two propositions , we focus attention on flow boxes of the form @xmath223 , where @xmath224 , and establish leafwise smoothing subject to a monodromy constraint .
let @xmath225\to d$ ] be the arc given by @xmath226 .
given an almost horizontal foliation @xmath20 of @xmath65 , we have @xmath227 [ @xmath9 smoothing a product foliation with holonomy constraint ] [ leafsmoothingholconstraints ] let @xmath224 , with core curve @xmath228 , oriented as above .
let @xmath37 denote the 1-dimensional foliation by @xmath63 fibers @xmath229 .
suppose @xmath156 is a @xmath11 almost horizontal foliation on @xmath66 that is @xmath9 on the neighborhood @xmath230 of @xmath231 , where @xmath232 and @xmath233 are nondegenerate closed intervals containing @xmath234 and @xmath153 respectively .
then @xmath156 can be @xmath7 deformed to a @xmath7 close , @xmath9 , almost horizontal foliation @xmath35 on @xmath235 such that @xmath35 agrees with @xmath156 on @xmath119 and @xmath236 . if @xmath237 are smoothly embedded leaves of @xmath156 , then the deformation can be chosen to fix these leaves pointwise . moreover , if @xmath159 are subsurfaces of leaves @xmath160 of @xmath156 so that regular neighborhoods @xmath161 of @xmath162 in @xmath163 are smoothly embedded in @xmath66 , then the deformation can be chosen to fix each @xmath162 pointwise .
let @xmath238 be an open neighborhood of @xmath119 on which @xmath156 is @xmath9 . pick a metric on the bundle of tangent 2-planes to @xmath66 .
denote the leaf of @xmath156 that contains @xmath239 by @xmath166 .
each @xmath166 is the graph of a @xmath32 function @xmath240 .
fix a smooth damping function @xmath241 that vanishes on @xmath242 and is identically @xmath153 on the complement of @xmath243 . by proposition [ smoothleaf interpolation ] ,
there is a @xmath9 , almost horizontal , foliation @xmath157 of @xmath66 that is @xmath7 close to @xmath156 and agrees with @xmath156 on @xmath119 . denote the leaf of @xmath157 that contains @xmath239 by @xmath244 .
each @xmath244 is the graph of a @xmath8 function @xmath245 .
if @xmath173 describes the graph of the leaf @xmath162 for some @xmath25 , then choose @xmath157 so that @xmath246 along @xmath162 .
it may be helpful , even though not necessary , to recall that by the construction of @xmath157 found in the proof of proposition [ smooth interpolation ] , we may assume that that there is a partition @xmath247 , such that for each @xmath25 , @xmath248 , and each leaf @xmath249,$ ] has tangent plane field @xmath7 close to the tangent plane field of each of @xmath179 and @xmath250 .
define a homeomorphism @xmath251 by @xmath252 the point of the definition of @xmath253 is that a leaf of @xmath157 that contains @xmath254 will intersect the leaf of @xmath156 that contains @xmath255 at the point @xmath256 . since @xmath156 and @xmath157 agree on @xmath257 , @xmath258 on all of @xmath257 .
since @xmath156 and @xmath157 are @xmath7 close , @xmath221 is @xmath7 close to the identity map , and , for each @xmath259 , the graphs @xmath260 and @xmath261 are @xmath32 close on the rest of @xmath67 .
we obtain a @xmath9 foliation @xmath35 approximating @xmath156 and satisfying the holonomy constraint @xmath262 as follows .
define @xmath263 then for a fixed @xmath138 , @xmath264 and @xmath265 by choosing @xmath157 sufficiently close to @xmath156 , we guarantee that the @xmath9 foliation @xmath35 with leaves the graphs of the functions @xmath266 is @xmath7 close to @xmath156 . since @xmath267 on @xmath119 ,
the choice of @xmath206 implies @xmath268 on @xmath119 . and ,
as has now been argued many times , this foliation @xmath35 is obtained by a @xmath7 small deformation of @xmath156 . a similar argument yields the following . [ @xmath8 smoothing a product foliation with smooth holonomy constraint ] [ smoothingholconstraints ]
let @xmath224 , with core curve @xmath228 , oriented as above .
let @xmath37 denote the 1-dimensional foliation by @xmath63 fibers @xmath229 .
suppose @xmath156 is a @xmath11 almost horizontal foliation on @xmath66 that is smooth on the neighborhood @xmath230 of @xmath231 , where @xmath232 and @xmath233 are nondegenerate closed intervals containing @xmath234 and @xmath153 respectively .
suppose that the holonomy map @xmath269 is smooth .
then @xmath156 can be @xmath7 deformed to a @xmath7 close , smooth , almost horizontal foliation @xmath35 on @xmath235 such that @xmath236 .
moreover , @xmath35 can be chosen to agree with @xmath156 on @xmath119 . if leaves @xmath237
smoothly embedded leaves of @xmath156 , then the deformation can be chosen to fix these leaves pointwise . moreover ,
if @xmath159 are subsurfaces of leaves @xmath160 of @xmath156 so that regular neighborhoods @xmath161 of @xmath162 in @xmath163 are smoothly embedded in @xmath66 , then the deformation can be chosen to fix each @xmath162 pointwise .
once again , local smoothing can be realized by a flow compatible , @xmath7 small , isotopy of @xmath0 .
this is made precise in the following corollary .
[ leafsmoothingcor ] let @xmath20 be a @xmath11 ( respectively , @xmath9 ) foliation of @xmath0 , and @xmath37 a smooth flow transverse to @xmath20 .
suppose that @xmath223 , @xmath224 , is a smooth flow box in @xmath0 , and that @xmath20 is @xmath9 in a neighborhood @xmath119 of @xmath270 .
then there is a @xmath7 small , @xmath37-compatible , isotopy of @xmath0 that is the identity outside a smooth flow box @xmath271 , and takes @xmath20 to a @xmath7 close @xmath11 foliation @xmath35 that is @xmath9 on @xmath66 , agrees with @xmath20 on @xmath119 , and satisfies @xmath272 .
if @xmath105 is a smooth flow box decomposition of @xmath0 , then this isotopy can be chosen to be @xmath105 compatible .
for the rest of this section , given a flow box @xmath273 , @xmath274 will denote a smooth regular collar neighborhood of @xmath71 in @xmath67 , and @xmath275 will denote a regular neighborhood of @xmath274 in @xmath67 .
the next proposition shows that a smooth annular foliation on @xmath276 that approximates @xmath20 can be extended to a smooth foliation on @xmath65 that approximates @xmath20 .
[ @xmath7 close damped coning][damped coning ] let @xmath20 be a @xmath11 foliation of @xmath0 , and @xmath37 a smooth flow transverse to @xmath20 .
let @xmath223 be a @xmath8 flow box , and suppose that @xmath277 is a smooth almost horizontal product foliation of @xmath278 that is @xmath7 close to the restriction of @xmath20 to @xmath278 .
then there is a smooth foliation @xmath35 of @xmath65 that is @xmath7 close to @xmath20 and restricts to @xmath277 on @xmath276 . by corollary
[ local smoothing ] , there is a smooth foliation @xmath279 of @xmath65 which is @xmath7 close to @xmath20 . then @xmath277 , @xmath20 , and @xmath279 are all @xmath7 close on @xmath280
. since @xmath277 and @xmath279 are both product foliations on @xmath280 , the result follows by applying uniqueness of local smoothing , corollary [ local smoothing unique ] , to @xmath274 and @xmath275 .
this result can be strengthened so that the approximating foliation is a smooth extension of @xmath20 near the vertical boundary . [
damped coning ] [ extendsmoothoverthreecell ] let @xmath20 be a @xmath11 foliation of @xmath0 , and @xmath37 a smooth flow transverse to @xmath20 .
let @xmath223 be a @xmath8 flow box , and suppose that the restriction of @xmath20 to @xmath278 is a @xmath8 foliation .
then there is a @xmath7 small , flow compatible , isotopy of @xmath65 that is the identity on @xmath276 and takes @xmath20 to a @xmath7 close @xmath8 foliation @xmath35 of @xmath65 .
if @xmath20 is smooth in a neighborhood of @xmath281 in @xmath65 , then the isotopy can be chosen so that @xmath282 in a slightly smaller neighborhood of @xmath281 .
if @xmath105 is a smooth flow box decomposition of @xmath0 , then this isotopy can be chosen to be @xmath105 compatible .
let @xmath277 be the restriction of @xmath20 to @xmath278 .
the result then follows immediately from proposition [ damped coning ] .
[ leafwise smooth damped coning][extendleafsmoothoverthreecell ] let @xmath20 be a @xmath11 foliation of @xmath0 , and @xmath37 a smooth flow transverse to @xmath20 .
let @xmath223 be a @xmath8 flow box , and suppose that the restriction of @xmath20 to @xmath278 is a @xmath9 foliation .
then there is a @xmath7 small flow compatible isotopy of @xmath65 that is the identity on @xmath276 and takes @xmath20 to a @xmath7 close @xmath9 foliation @xmath35 of @xmath65 .
if @xmath20 is smooth in a neighborhood of @xmath281 in @xmath65 , then the isotopy can be chosen so that @xmath282 in a slightly smaller neighborhood of @xmath281 .
if @xmath105 is a smooth flow box decomposition of @xmath0 , then this isotopy can be chosen to be @xmath105 compatible .
recall that @xmath274 denotes a smooth regular collar neighborhood of @xmath71 in @xmath67 , and @xmath275 denotes a smooth regular neighborhood of @xmath274 in @xmath67 .
decompose @xmath283 as a union of smooth annuli with pairwise disjoint interiors and @xmath284 disjoint from @xmath285 . by assumption ,
the restriction of @xmath20 to @xmath278 , and hence to @xmath286 , is a @xmath9 product foliation by almost horizontal annuli .
apply proposition [ local smoothing ] with @xmath287 : there is a @xmath7 small , flow compatible isotopy of @xmath65 that is the identity outside a small neighborhood @xmath288 and takes @xmath20 to a @xmath7 close @xmath9 foliation @xmath35 that is @xmath8 on @xmath286 . by choosing @xmath289 disjoint from @xmath274
, we may assume this isotopy is the identity on @xmath276 .
it follows that the restriction of @xmath35 to @xmath276 agrees with the restriction of @xmath20 to @xmath276 .
next , apply proposition [ damped coning ] to the flow box @xmath290 .
the result is a smooth foliation @xmath291 on @xmath292 that is @xmath7 close to @xmath20 and agrees with @xmath35 on @xmath286 .
it therefore extends as @xmath35 to give a @xmath9 foliation of @xmath65 that is @xmath7 close to @xmath20 and agrees with @xmath20 on @xmath276 .
the next theorem adds @xmath7 approximation to a theorem of calegari , @xcite .
it is applied and cited as theorem 2.10 in @xcite . [ calegarikr ]
suppose @xmath20 is a @xmath11 foliation of a compact manifold @xmath0 .
then there is a @xmath7 small isotopy of @xmath0 taking @xmath20 to a @xmath9 foliation @xmath35 that is @xmath7 close to @xmath20 . if @xmath37 is a smooth flow transverse to @xmath20 , the isotopy may be taken to be flow compatible .
let @xmath37 be a smooth flow transverse to @xmath20 , and apply lemma [ smooth - sided ] to obtain a smooth - sided @xmath12-flow box decomposition @xmath293 of @xmath0 .
let @xmath105 denote the smooth flow box decomposition and @xmath294 the foliation that results from applying the isotopy of lemma [ smooth fbd ] to @xmath293 and @xmath20 respectively .
let @xmath295 be a listing of the maximal vertical faces of @xmath105 , and choose a regular neighborhood structure @xmath296 for @xmath105 . by corollary
[ local smoothing ] , there is a @xmath7 small , @xmath105 compatible isotopy of @xmath0 that takes @xmath294 to a @xmath7 close @xmath11 foliation @xmath297 that is smooth on @xmath119 . by corollary [ leafsmoothingcor ] ,
there is a @xmath7 small , @xmath105 compatible isotopy of @xmath0 that takes @xmath297 to a @xmath7 close @xmath11 foliation @xmath298 that is smooth on @xmath119 and @xmath9 on each @xmath299 .
finally , by proposition [ extendleafsmoothoverthreecell ] , there is a @xmath7 small , @xmath105 compatible isotopy of @xmath0 that takes @xmath298 to a @xmath7 close @xmath9 foliation @xmath35 .
[ limitsmooth ] any @xmath11 foliation is a limit of @xmath9 foliations .
a _ transverse measure _ on a codimension one foliation @xmath20 is a _ continuous , non - degenerate , invariant _ measure , @xmath300 , on each arc transverse to @xmath20 .
it is continuous in the sense that if @xmath301 is smoothly parametrized as @xmath302 $ ] , then @xmath303)$ ] is continuous in @xmath201 .
non - degenerate means that @xmath300 is positive on every open interval .
invariant , in this context , means that the measure of a transverse arc is unchanged under isotopies of the arc that keep each point on the same leaf of @xmath20 .
[ smoothtransversal ] let @xmath304 be a @xmath9 measured foliation in @xmath0 .
suppose @xmath301 is a smoothly embedded arc or closed curve which is everywhere transverse to @xmath20 .
then there is a @xmath7 small isotopy of @xmath0 which is the identity outside some small regular neighborhood @xmath128 of @xmath301 and takes the measured foliation @xmath304 to a @xmath9 measured foliation @xmath305 such that @xmath306 is smooth in a neighborhood of @xmath301 and the measure , @xmath307 , restricted to @xmath301 is smooth . if @xmath300 is smooth on a closed submanifold @xmath274 of @xmath301 , then the isotopy can be chosen so that @xmath308 on @xmath274 .
if , in addition , @xmath20 is smooth in a @xmath12 compatible regular neighborhood @xmath309 of @xmath274 , then the isotopy can be chosen to be the identity on @xmath309 .
it suffices to consider the case that @xmath301 is a smoothly embedded arc .
regard @xmath301 as a smooth map @xmath310 .
then @xmath311)$ ] is a homeomorphism , @xmath312 , onto its image .
approximate @xmath221 by a diffeomorphism , relative to end points , @xmath313 .
the goal is to make a small , continuous change of coordinates on @xmath301 so that @xmath300 is smooth in the new coordinates . in other words
, we must choose a homeomorphism @xmath314 so that @xmath315 is smooth .
this is accomplished by defining @xmath316 .
the next step is to use this reparametrization to describe a @xmath7 small isotopy of @xmath0 which is the identity outside a small neighborhood of @xmath301 and takes the measure @xmath300 to a measure @xmath307 that on @xmath301 satisfies @xmath317 = \mu ( \tau\circ f[0,t])$ ] for each @xmath192 $ ] .
in particular , @xmath307 restricted to @xmath301 is smooth . to accomplish this ,
let @xmath318 and @xmath319 be small smoothly embedded tubular neighborhoods of @xmath301 satisfying @xmath320 .
choose these neighborhoods small enough so that @xmath20 meets each in a foliation by meridian disks .
parametrize these disks by their intersection with @xmath321 $ ] and so that @xmath318 is identified with the smooth family of meridian disks @xmath322 .
next isotope @xmath20 in @xmath319 so that the disk @xmath323 is taken to the disk @xmath324 .
if we choose @xmath318 small enough and @xmath325 sufficiently close to the identity , we may choose these isotopies to be as close as desired to the identity .
define @xmath307 along @xmath301 so that is invariant and agrees with @xmath300 away from @xmath301 .
[ smoothtransversal2 ] let @xmath20 be a @xmath9 foliation in @xmath0 .
suppose @xmath301 is a smoothly embedded arc or closed curve which is everywhere transverse to @xmath20 .
then there is a @xmath7 small isotopy of @xmath0 that is the identity outside some small regular neighborhood of @xmath301 and takes @xmath20 to a @xmath9 @xmath306 such that @xmath306 is smooth in a neighborhood of @xmath301 .
if @xmath20 is smooth in a @xmath12 compatible regular neighborhood @xmath309 of a closed subset of @xmath301 , then the isotopy can be chosen to be the identity on @xmath309 .
choose a small regular neighborhood @xmath128 of @xmath301 so that @xmath20 meets it in a product foliation by disks .
use distance along @xmath301 to define a smooth transverse measure on the restriction of @xmath20 to @xmath128 .
the result now follows immediately from lemma [ smoothtransversal ] .
the next lemma shows how the existence of a transverse measure allows the foliation to be smoothed near a compact portion of a leaf .
[ smoothing in the neighborhood of a compact subsurface of a leaf ] [ smoothhorizontal ] let @xmath304 be a @xmath9 measured foliation in @xmath0 .
suppose @xmath141 is a compact subsurface of a leaf of @xmath20 .
then there is a @xmath7 small isotopy of @xmath0 which is the identity outside some small regular neighborhood of @xmath141 in @xmath0 and takes the measured foliation @xmath304 to a @xmath9 measured foliation @xmath305 such that @xmath306 is smooth in a neighborhood of @xmath141 and the measure @xmath307 restricted to this neighborhood is smooth .
let @xmath10 be the leaf of @xmath20 containing @xmath141 .
if @xmath326 , let @xmath327 .
otherwise , let @xmath147 be the closure of a regular neighborhood of @xmath141 in @xmath10 .
use the measure , @xmath300 , to give a continuous parametrization of the flow @xmath37 in a neighborhood of @xmath328 . to avoid confusion ,
let @xmath329 denote this reparametrized restriction of @xmath37 .
choose this parametrization so that @xmath330 for all @xmath331 and @xmath332))=t - s$ ] for @xmath167 sufficiently close to @xmath234 .
for some @xmath191 , @xmath333\to m$ ] is a topological embedding , and for each @xmath334 $ ] , @xmath335 is a compact subsurface of a leaf of @xmath20 , necessarily isotopic to @xmath147 .
since @xmath20 is @xmath9 and @xmath329 is smooth when restricted to a leaf , @xmath336 ) $ ] is a smooth codimension 0 submanifold , possibly with corners .
use proposition [ smooth interpolation ] to @xmath7 isotope @xmath20 in @xmath336 ) $ ] so that it is a smooth foliation by surfaces isotopic to @xmath147 .
the resulting measured foliation @xmath305 and the measure @xmath307 are necessarily smooth on the neighborhood @xmath337 of @xmath147 .
the next theorem is applied and cited as theorem 8.10 in @xcite : [ msredtosmooth0 ] suppose @xmath20 is a transversely orientable @xmath11 measured foliation in @xmath0 .
then there is an isotopy of @xmath0 taking @xmath20 to a @xmath8 measured foliation which is @xmath7 close to @xmath20 . if @xmath37 is a smooth flow transverse to @xmath20 , the isotopy may be taken to be flow compatible . by theorem [ calegarikr ]
we may assume @xmath20 is @xmath9 .
from lemma [ smooth - sided ] it follows that @xmath0 has a smooth flow box decomposition , @xmath338 .
using lemma [ smoothhorizontal ] , we may assume , after a @xmath7 small isotopy , that @xmath20 and @xmath300 are smooth in a small regular neighborhood @xmath339 of @xmath340 . the vertical 1-skeleton of @xmath105 is a disjoint union of transversals to @xmath20 , and hence , by applying lemma [ smoothtransversal ] , we may assume , after a @xmath7 small isotopy , that @xmath20 and @xmath300 are smooth on @xmath341 , where @xmath119 is a union of flow boxes that form a @xmath12 compatible regular neighborhood of the union of the vertical 1-cells of @xmath342 .
let @xmath295 be a listing of the maximal vertical faces of @xmath105 , and let @xmath343 be a regular neighborhood structure for @xmath105 .
choose @xmath344 , so that @xmath345 is also a regular neighborhood structure for @xmath105 .
let @xmath346 and @xmath347 denote the two vertical edges of @xmath348 .
let @xmath349 denote the homeomorphism obtained by following leaves of @xmath20 across @xmath348 . since the measure @xmath300 is smooth on @xmath350 , and @xmath349 preserves @xmath300 , @xmath349 is smooth . by proposition [ smoothingholconstraints ] ,
there is a @xmath7 small , @xmath105 compatible isotopy of @xmath0 that is the identity on @xmath351 and outside the union @xmath352 , and takes @xmath20 to a foliation @xmath35 that is smooth on @xmath128 and @xmath7 close to @xmath20 .
finally , we apply damped coning , corollary [ extendsmoothoverthreecell ] , to extend @xmath35 to a smooth foliation that is @xmath7 close and isotopic , by a @xmath7 small isotopy , to @xmath20 .
since @xmath300 is defined on the vertical boundary of every flow box , it extends to all of @xmath35 .
the next corollary now follows from theorem 8.11 in @xcite , which uses theorem [ msredtosmooth0 ] in its proof .
alternatively , it also follows from theorem [ msredtosmooth0 ] together with tischler s theorem @xcite , which states that any transversely oriented , measured @xmath1 foliation on a compact n - manifold can be @xmath8 approximated by a smooth fibration over @xmath353 .
[ tischler ] a @xmath11 , transversely oriented , measured foliation on a compact 3-manifold is @xmath7 close to a smooth fibration over @xmath353 .
in this section , we recall some definitions and results used in @xcite , giving those proofs which , for clarity of exposition , were deferred to this paper .
let @xmath354 be an _ oriented _ simple closed curve in a leaf @xmath10 of @xmath20 , and let @xmath123 be a point in @xmath354 .
we are interested in the behavior of @xmath20 in a neighborhood of @xmath354 .
let @xmath355 be a holonomy map for @xmath20 along @xmath354 , and let @xmath356 and @xmath301 be small closed segments of the flow @xmath37 which contain @xmath123 in their interiors and satisfy @xmath357 .
choose @xmath301 small enough so that @xmath358 is a closed segment and not a loop .
notice that @xmath359 is necessarily a closed segment containing @xmath123 in its interior .
there are three possibilities : 1 .
@xmath360 , 2 .
one of @xmath356 and @xmath301 is properly contained in the other , or 3 .
@xmath359 is properly contained in each of @xmath356 and @xmath301 .
we will need to consider very carefully a regular neighborhood of @xmath354 which lies nicely with respect to both @xmath20 and @xmath37 . to this end , restrict attention to foliations @xmath20 which are @xmath9 and transversely oriented , and transverse flows @xmath37 which are smooth , and suppose that @xmath354 is smoothly embedded in @xmath10 .
let @xmath274 be the closure of a smooth regular neighborhood of @xmath354 in @xmath10 ; so @xmath274 is a smoothly embedded annulus in @xmath10 .
[ lemma 3.1 of @xcite ] [ holnbddefn ] suppose @xmath20 is @xmath9 and transversely oriented , and @xmath37 is smooth .
if @xmath301 and @xmath274 are chosen to be small enough , there is a compact submanifold @xmath76 of @xmath0 , smoothly embedded with corners , which satisfies the following : 1 .
@xmath76 is homeomorphic to a solid torus , 2 .
@xmath361 is piecewise vertical and horizontal ; namely , @xmath361 decomposes as a union of subsurfaces @xmath362 , where @xmath363 is a union of flow segments of @xmath37 and @xmath364 is a union of two surfaces @xmath365 and @xmath366 , each of which is either a disk or an annulus , contained in leaves of @xmath20 , 3 .
each flow segment of @xmath367 runs from @xmath365 to @xmath366 , 4 .
@xmath301 is a component of the flow segments of @xmath367 , and 5 .
@xmath274 is a leaf of the foliation @xmath368 .
cover a small open neighborhood of @xmath354 by finitely many smooth flow boxes . by passing to a smaller @xmath301 and @xmath274 as necessary
, we may suppose that @xmath274 is covered by two flow boxes with union , @xmath76 , satisfying the properties ( 1)-(5 ) .
[ notationattracting ] denote the neighborhood @xmath76 of lemma [ holnbddefn ] by @xmath369 .
notice that if @xmath370 , then @xmath369 is diffeomorphic to @xmath276 , where @xmath63 is a nondegenerate closed interval . otherwise , there is a unique smooth vertical rectangle , @xmath371 say , so that the result of cutting @xmath369 open along @xmath371 , and taking the metric closure , is diffeomorphic to a solid cube .
[ notationq ] let @xmath372 denote any smooth vertical rectangle embedded in @xmath369 such that the result of cutting @xmath369 open along @xmath371 , and taking the metric closure , is diffeomorphic to a solid cube .
when @xmath369 is not diffeomorphic to a product , @xmath372 is uniquely determined .
let @xmath373 denote the resulting solid cube ; so @xmath374 note that if @xmath354 is essential , then @xmath373 can be viewed as a @xmath375 flow box , where @xmath375 is the lift of @xmath12 to the universal cover of @xmath0 .
the neighborhood @xmath369 is called the _ holonomy neighborhood determined by @xmath376 _ , and is called an _ attracting neighborhood _ if @xmath377 is contained in the interior of @xmath356 .
[ complete attracting ] let @xmath20 be a transversely oriented , @xmath9 foliation .
a set of holonomy neighborhoods @xmath378 for @xmath20 is _ spanning _ if each leaf of @xmath20 has nonempty intersection with the interior at least one @xmath379 .
[ vcompatible ] let @xmath76 be the union of pairwise disjoint holonomy neighborhoods @xmath378 for @xmath20 .
a transversely oriented , @xmath9 foliation @xmath35 in @xmath0 is called _
@xmath76-compatible with @xmath20 _ , ( or simply _
@xmath76-compatible _ if @xmath20 is clear from context ) if each @xmath379 is a holonomy neighborhood for @xmath35 , with @xmath76 spanning for @xmath35 if it is spanning for @xmath20 .
fix a set of pairwise disjoint holonomy neighborhoods @xmath380 @xmath381 for @xmath20 , and let @xmath76 denote their union .
let @xmath382 , and let @xmath371 denote the union of the @xmath383 .
for each @xmath384 , fix a smooth open neighborhood @xmath385 of @xmath383 in @xmath386 . choose each @xmath385 small enough so that its closure , @xmath387 , is a closed regular neighborhood of @xmath383 .
let @xmath388 denote the union of the @xmath385 .
now , given @xmath76 , @xmath371 and @xmath388 , we further constrain the set of foliations @xmath20 ( that we need to approximate by smooth contact structures ) to @xmath9 foliations which are smooth on @xmath388 .
the following lemma , applied and cited as lemma 3.7 in @xcite , establishes that we can do this with no loss of generality .
[ smoothaboutr ] let @xmath20 be a transversely oriented , @xmath9 foliation , and let @xmath37 be a smooth flow transverse to @xmath20 .
let @xmath76 denote the union of a set of pairwise disjoint holonomy neighborhoods for @xmath20 and fix @xmath388 as above .
there is an isotopy of @xmath0 taking @xmath20 to a @xmath9 foliation which is both @xmath7 close to @xmath20 and smooth on @xmath388 .
this isotopy may be taken to preserve @xmath76 and be flow compatible .
this follows from the next lemma [ stronglycompatiblef ] .
next we describe a preferred product parametrization on a closed set containing @xmath76 . in this paper , we express @xmath353 as the quotient @xmath389/\sim$ ] , where @xmath390 is the equivalence relation on @xmath391 $ ] which identifies @xmath392 and @xmath153 .
[ prodnbddefn ] let @xmath20 be a transversely oriented , @xmath9 foliation , and let @xmath37 be a smooth flow transverse to @xmath20 .
let @xmath76 denote the union of pairwise disjoint holonomy neighborhoods @xmath393 for @xmath20 , and fix @xmath388 as above .
suppose @xmath20 is smooth on @xmath388 .
then for each @xmath384 , there is a pairwise disjoint collection of closed solid tori @xmath394 such that @xmath394 contains @xmath386 and there is a diffeomorphism @xmath395\times s^1\times [ -1,1]$ ] which satisfies the following : 1 . the flow segments @xmath396 are identified with the segments @xmath397 $ ] , 2 .
@xmath398 is identified with @xmath391\times s^1\times \{0\}$ ] , 3 .
@xmath399 is identified with @xmath400 , 4 .
@xmath383 is identified with @xmath391\times \{1\sim -1\}\times [ -1,1]$ ] , and 5 .
the restriction of the diffeomorphism to @xmath385 maps leaves of @xmath20 to horizontal level surfaces @xmath401 , where @xmath402\times ( ( [ 1/2,1 ] \cup [ -1,-1/2])/\sim)$ ] .
since @xmath386 is homeomorphic to a solid torus , it is contained in a solid torus which is diffeomorphic to @xmath391\times s^1\times [ -1,1]$ ] , where the diffeomorphism can be chosen to identify @xmath274 with @xmath391\times s^1\times \{0\}$ ] and the flow segments @xmath403 with the segments @xmath397 $ ] .
moreover , since the restriction of @xmath20 to @xmath404 is a smooth product foliation transverse to vertical fibers , and there is a unique such up to diffeomorphism , this diffeomorphism @xmath405\times s^1\times [ -1,1]$ ] can also be chosen so that the restriction of the diffeomorphism to @xmath388 maps leaves of @xmath406 to the horizontal level surfaces @xmath407 .
[ productnbddefn ] fix @xmath76 and @xmath388 as above .
let @xmath394 and @xmath395\times s^1\times [ -1,1]$ ] be as given in lemma [ prodnbddefn ] .
abuse notation and use the diffeomorphism to identify @xmath394 with @xmath391\times s^1\times [ -1,1]$ ] . let @xmath408 be the product foliation of @xmath394 with leaves @xmath409\times s^1)\times \{t\}$ ] , and call such a foliated solid torus , @xmath410 , a _ product neighborhood of @xmath411 .
_ letting @xmath178 denote the union of the @xmath394 and @xmath156 denote the union of the @xmath408 , call @xmath412 a _ product neighborhood of @xmath413 . _
[ stronglycompatible ] let @xmath20 be a transversely oriented , @xmath9 foliation and @xmath76 the union of pairwise disjoint , holonomy neighborhoods @xmath414 for @xmath20 .
let @xmath371 denote the union of the @xmath415 , and let @xmath388 be an open regular neighborhood of @xmath371 in @xmath76 .
let @xmath412 be a product neighborhood of @xmath413 .
the foliation @xmath20 is _
strongly @xmath416 compatible _ if 1 .
@xmath417 , and 2 . in the coordinates inherited from @xmath178 ,
@xmath368 is a product foliation @xmath391\times \mathcal f_0 $ ] , where @xmath418 is a @xmath9 foliation of @xmath419)$ ] ( i.e. , @xmath368 is @xmath201-invariant ) .
given @xmath76 , @xmath371 and @xmath388 , we will further constrain the set of foliations @xmath20 to @xmath9 foliations which are strongly @xmath416 compatible for some choice of product neighborhood @xmath412 . the following lemma ,
applied and cited as lemma 3.11 in @xcite , establishes that we can do this with no loss of generality ; namely , after a small perturbation of @xmath20 , it is possible to rechoose the diffeomorphisms @xmath395\times s^1\times [ -1,1]$ ] so that @xmath420 on @xmath421 and @xmath20 is invariant under translation in the first coordinate .
[ stronglycompatiblef ] let @xmath20 be a transversely oriented , @xmath9 foliation and let @xmath37 be a smooth flow transverse to @xmath20 .
let @xmath76 denote the union of a set of pairwise disjoint holonomy neighborhoods for @xmath20 and fix @xmath388 as above .
there is a @xmath7 small , flow compatible , isotopy of @xmath0 that takes @xmath20 to a @xmath9 foliation that is @xmath7 close to @xmath20 and strongly @xmath416 compatible for some choice of product neighborhood @xmath412 of @xmath413 .
this isotopy may be taken to preserve @xmath76 .
the method of proposition [ uniqueness in a product ] can be applied here .
that is , on @xmath76 there exists a foliation @xmath35 with the smoothness , restriction , and invariance properties specified in the lemma .
moreover , such a @xmath35 can be constructed to have the same holonomy representation as @xmath20 with respect to each horizontal annulus of @xmath76 and any vertical fiber of @xmath422 .
thus the method of proposition [ uniqueness in a product ] gives the desired isotopy .
in @xcite , denjoy gave examples of @xmath32 foliations on @xmath423 with exceptional minimal sets . in @xcite ,
dippolito generalized denjoy s method to @xmath9 codimension one foliations of @xmath2-manifolds .
this generalized construction is commonly referred to as _ denjoy blowup _ , and is defined precisely as follows .
[ maindefinition ] let @xmath10 be a countable ( finite or countably infinite ) union of leaves of a @xmath5 foliation @xmath20 of @xmath0 with @xmath135 , and let @xmath37 be a smooth flow transverse to @xmath20 .
a @xmath5 foliation , @xmath306 , is a _ denjoy blowup of @xmath20 along @xmath10 _ if there is an open subset @xmath424 and a continuous collapsing map @xmath425 satisfying the following properties : 1 .
@xmath306 is transverse to @xmath37 , 2 .
there is an injective map @xmath426 such that @xmath427 is a @xmath18 immersion and @xmath428 , 3 . for each @xmath429 , @xmath430
is contained in a flow line of @xmath37 , 4 .
@xmath431 and @xmath432 are leaves of @xmath306 , 5 .
@xmath433 is a point if @xmath434 and equals @xmath430 if @xmath435 , 6 .
@xmath436 is @xmath37 compatible and maps leaves of @xmath306 to leaves of @xmath20 , 7 .
@xmath436 is @xmath18 when restricted to any leaf of @xmath306 , and 8 .
@xmath436 is the time one map of a @xmath37 compatible @xmath18 isotopy @xmath437 .
if @xmath438 is a @xmath5 almost horizontal foliation of @xmath439 , and the pullback of the denjoy blowup @xmath306 to @xmath439 is @xmath18 equivalent to @xmath438 , then @xmath306 is a denjoy blowup of @xmath20 along @xmath10 _ by @xmath438 . _
[ canfirstclean ] suppose @xmath20 and @xmath35 are @xmath11 foliations of @xmath0 transverse to a common smooth flow @xmath37 .
if @xmath20 and @xmath35 are @xmath37 compatible isotopic , then a denjoy blowup of @xmath35 is a denjoy blowup of @xmath20 .
the following result extends dippolito s generalization of denjoy s construction ( theorem 7 , @xcite ) in two ways .
first , it allows for foliations which are not @xmath9 .
second , it shows that the resulting foliation , @xmath306 , can be constructed arbitrarily @xmath7 close to @xmath20 .
[ blowup ] let @xmath20 be a @xmath11 foliation in a compact 3-manifold @xmath0 .
suppose that @xmath20 is transverse to a smooth flow @xmath37 .
let @xmath10 be a countable collection of leaves of @xmath20 , and let @xmath438 be a @xmath11 almost horizontal foliation of @xmath439
. then there exists @xmath9 @xmath306 arbitrarily @xmath440 close to @xmath20 that is a denjoy blowup of @xmath20 along @xmath10 by @xmath438
. moreover , if @xmath20 is @xmath9 and @xmath105 is a @xmath12-flow box decomposition of @xmath0 , with flow boxes @xmath113 , and @xmath10 is disjoint from @xmath441 , then the denjoy blowup can be chosen to be strongly @xmath105 compatible in the following sense : the restriction of @xmath306 to each @xmath82 is the denjoy blowup of the restriction of @xmath20 to @xmath82 . by theorem [ calegarikr ] , a @xmath5 foliation on a compact 3-manifold can be isotoped by a @xmath7 small @xmath37 compatible isotopy to a @xmath7 close , @xmath9 foliation .
thus , by lemma [ canfirstclean ] , it is enough to prove the theorem with the assumption that @xmath20 is @xmath9 and @xmath438 is @xmath5 .
while it may be possible to conclude @xmath7 proximity of @xmath306 from dippolito s original proof , the method of flow box decompositions gives a direct and elementary proof .
[ one flow box ] let @xmath20 be a strictly horizontal foliation of a @xmath8 flow box @xmath223 .
let @xmath10 be a countable union of leaves of @xmath20 , and let @xmath438 be a @xmath5 almost horizontal foliation of @xmath439 , for some @xmath135 .
then there exists a @xmath9 foliation @xmath306 arbitrarily @xmath440 close to @xmath20 that is a denjoy blowup of @xmath20 along @xmath10 by @xmath438 .
we begin by describing the denjoy blowup of @xmath63 along the points @xmath447 .
let @xmath448 denote a summable sequence of positive numbers , with sum @xmath449 .
cut @xmath63 at each @xmath447 and insert an interval @xmath450 of length @xmath448 .
the result is a new interval of length @xmath451 .
the left inverse of this operation is a cantor function ; denote this cantor function by @xmath452\to [ 0,1]$ ] .
let @xmath453\to [ 0,1]$ ] denote the function obtained by composing the function @xmath74 with the linear scaling @xmath454\to [ 0,1+w]$ ] ; so @xmath455 .
these function are illustrated in figure [ blowupfig ] .
let @xmath456=s^{-1}(j_i)$ ] , and notice that @xmath457 .
set @xmath458 , and let @xmath459 be the strictly horizontal lamination with leaves @xmath460 .
let @xmath461 , and @xmath462 , the open set @xmath463 .
let @xmath306 denote the foliation obtained by taking the union of @xmath459 with @xmath474 .
now fix @xmath191 .
since @xmath475 , and @xmath20 is strictly horizontal , the @xmath448 can be chosen so that @xmath306 is @xmath62 @xmath7 close to @xmath20 . properties ( 1)(7 ) of definition [ maindefinition ] then follow immediately .
the isotopy @xmath476 of property ( 8) is given by the straight line , @xmath37 compatible , isotopy from the identity map to @xmath436 . by theorem [ calegarikr ]
, we may isotope the resulting @xmath11 denjoy blowup to a @xmath7 close @xmath9 denjoy blowup . finally , if @xmath442 is a listing of finitely many leaves of @xmath20 that are disjoint from @xmath10 , cut @xmath65 open along along each @xmath442 , and perform denjoy blowup , as just described , on each resulting flow box . by theorem [ calegarikr ] , @xmath20 is @xmath37 compatible isotopic to a @xmath7 close @xmath9 foliation . by lemma [ canfirstclean ]
therefore , we may restrict attention to the case that @xmath20 is @xmath9 .
let @xmath477 , be a listing of the maximal vertical faces of @xmath105 , and let @xmath478 be a regular neighborhood structure for @xmath105 . by corollary [ smoothtransversal2 ] and lemma [ canfirstclean ] , it suffices to further restrict attention to the case that @xmath20 is @xmath9 , and smooth when restricted to @xmath119 .
recall that @xmath119 is a union of flow boxes , @xmath480 , satisfying conditions ( 3 ) and ( 4 ) of definition [ rnd defn ] .
rechoose the @xmath481 , if necessary , so that @xmath482 has empty intersection with @xmath340 .
this can be achieved by cutting each @xmath481 open along any horizontal level that has nonempty intersection with @xmath340 .
let @xmath483 be a @xmath481 that has nonempty intersection with @xmath10 .
since the restriction of @xmath20 to @xmath284 is smooth , there is a smooth parametrization @xmath484 of @xmath284 such that the restriction of @xmath485 to @xmath284 is strictly horizontal and the restriction of @xmath37 to @xmath284 has flow lines the vertical line segments @xmath486 . by lemma [ one flow box ] , therefore , there is a denjoy blowup @xmath487 of the restriction of @xmath20 to @xmath284 , and hence functions @xmath488 and @xmath489 satisfying the conditions of definition [ maindefinition ] .
repeat this process for each @xmath481 that has nonempty intersection with @xmath10 . and let @xmath490 on the remaining @xmath481 .
thus , we get a @xmath9 denjoy blowup @xmath491 of the restriction of @xmath20 to @xmath119 that is strongly compatible with @xmath105 , together with functions @xmath492 and @xmath493 satisfying the conditions of definition [ maindefinition ] . the foliation @xmath494 can be chosen to be @xmath7 close to the restriction of @xmath20 to @xmath119 .
next , let @xmath356 be any maximal vertical face of @xmath105 , and let @xmath495 be the vertical edges of @xmath356 .
let @xmath496 , respectively @xmath497 , denote the component of @xmath498 that contains @xmath499 , respectively @xmath500 .
the denjoy blowup , @xmath491 , is defined on @xmath119 , and hence on each @xmath501 .
let @xmath502 be a listing of the components of @xmath503 .
set @xmath504 , and set @xmath505 .
writing @xmath506 , where @xmath224 , @xmath507 , and @xmath508 .
orient @xmath509 so that @xmath510 . on each @xmath511 $ ] ,
define @xmath512 . on the complement of @xmath513 $ ] , where @xmath514 is single valued , define @xmath515 . by corollary [ leafsmoothingcor ] , the foliation @xmath491 defined on @xmath119 extends to a @xmath9 denjoy blowup of @xmath20 on @xmath516 that is @xmath105 compatible , @xmath7 close to @xmath20 on @xmath299 and satisfies @xmath517 . repeating this construction for each maximal face
@xmath348 extends the definition of the @xmath7 close denjoy blowup of @xmath20 to @xmath128 .
finally , the denjoy blowup @xmath518 defined on @xmath128 extends to a strongly @xmath105 compatible denjoy blowup @xmath306 , @xmath7 close to @xmath20 , on each flow box @xmath519 of @xmath105 by damped coning , as described in proposition [ extendleafsmoothoverthreecell ] .
moreover , since at each step we are extending over a flow box @xmath519 , the resulting packet , @xmath520 , of inserted leaves will be @xmath5 equivalent to @xmath438 on @xmath519 .
similarly , the extension of the collapsing map @xmath521 to @xmath519 is uniquely determined by the properties that it maps leaves of @xmath306 to leaves of @xmath20 and maps each @xmath63 fiber to itself .
[ theorem5.2 ] let @xmath20 be a transversely oriented , @xmath5 foliation with @xmath135 that is transverse to a smooth flow @xmath37 .
let @xmath10 be a countable collection of leaves of @xmath20 , and let @xmath294 be a @xmath5 foliation of @xmath439 transverse to the @xmath63 coordinate that contains @xmath522 as leaves .
then there exists a @xmath5 @xmath306 arbitrarily @xmath7 close to @xmath20 that is a denjoy blowup of @xmath20 along @xmath10 , and such that the pullback of @xmath306 to @xmath439 is equivalent to @xmath294 . moreover , if @xmath76 is the union of a set of pairwise disjoint holonomy neighborhoods for @xmath20 , @xmath523 is a product neighborhood of @xmath76 , and @xmath20 is strongly @xmath524 compatible , then @xmath306 can be chosen to be strongly @xmath524 compatible . using the notation of [ notationq ] ,
@xmath76 can be cut open along @xmath371 into a cube @xmath525 .
parametrize @xmath526 so that @xmath527\cup [ 0,1/4))\times i,\ ] ] and decompose @xmath525 along @xmath528 into two flow boxes .
thus @xmath76 is realized as a union of two flow boxes , and , by corollary [ existence of fbd ] , this flow box decomposition of @xmath76 extends to a flow box decomposition @xmath105 of @xmath0 .
moreover , this extension , @xmath105 , can be chosen so that each vertical face of @xmath76 , except for the proper subface of @xmath371 , is maximal .
let @xmath529 denote the maximal face @xmath371 . 1 .
@xmath306 is strictly horizontal in the flow boxes @xmath533 of @xmath119 that contain vertical edges of @xmath371 , 2 .
@xmath306 is strictly horizontal in @xmath534 , and 3 . in the coordinates inherited from @xmath535 , the fixed product neighborhood of @xmath76
, @xmath536 is @xmath201-invariant .
|
transverse one dimensional foliations play an important role in the study of codimension one foliations . in @xcite ,
the authors introduced the notion of flow box decomposition of a 3-manifold @xmath0 .
this is a combinatorial decomposition of @xmath0 that reflects both the structure of a given codimension one foliation and that of a given transverse flow , and that is amenable to inductive strategies . in this paper , flow box decompositions
are used to extend some classical foliation results to foliations that are not @xmath1 .
enhancements of well - known results of calegari on smoothing leaves , dippolito on denjoy blowup of leaves , and tischler on approximations by fibrations are obtained .
the methods developed are not intrinsically 3-dimensional techniques , and should generalize to prove corresponding results for codimension one foliations in @xmath2-dimensional manifolds .
|
in the early days of string theory , one way to obtain amplitudes for the scattering of an arbitrary number of strings was by using the factorization property , what means that the scattering amplitude of @xmath0 strings may be interpreted as the scattering amplitudes of a smaller number of strings sewn together .
this made it possible to build the @xmath0 string scattering amplitude by knowing the expression for three string scattering amplitudes .
even though they were very ingenious and successful , those calculations did nt take into account the ghost structure of the vertices , and that is what is done here . in @xcite
, it was shown how to sew tree vertices without ghosts using the group theoretic approach to string theory @xcite in order to obtain a new , composite vertex .
following the same procedure , we shall calculate the scattering amplitude of @xmath0 strings taking account the ghost structure .
we shall start with a short review of how to sew tree vertices without ghosts .
what we must do is sew two legs of two vertices , one leg from each vertex .
what we have in the beginning are two vertices @xmath1 and @xmath2 with @xmath3 and @xmath4 legs , respectively ( figure 1 ) .
0.4 cm figure 1 : individual vertices
. 0.2 cm we now sew leg @xmath5 from @xmath1 with the adjoint of leg @xmath6 from @xmath2 .
what we have now is the substitution of the two sewn legs by a propagator ( figure 2 ) . when this propagator is written in parametric form , it is an integration of one of the variables ( in order to cancel one spurious degree of freedom ) and a conformal factor @xmath7 which contains terms of @xmath8 s acting on leg @xmath5 only .
0.4 cm figure 2 : sewn vertices .
0.2 cm so the resulting vertex @xmath9 ( called the _ composite vertex _ ) has the generic form @xmath10 where the hermitian conjugate of @xmath2 is for the sewn leg @xmath6 only and @xmath11 where @xmath12 is a suitable variable . in what follows , we shall often write @xmath7 instead of @xmath13 , calling attention to the integration when necessary . when the two vertices are sewn together , we identify legs @xmath5 and
@xmath6 . we also identify the koba - nielsen variable @xmath14 with one of the koba - nielsen variables of vertex @xmath15 , and the koba - nielsen variable @xmath16 we identify with one of the koba - nielsen variables of vertex @xmath1 . in @xcite ,
this identification is made in the following way : @xmath14 may be identified with @xmath17 or @xmath18 , and @xmath16 may be identified with @xmath19 or @xmath20 .
so there are four possible combinations : a ) @xmath21 ; b ) @xmath22 ; c ) @xmath23 ; d ) @xmath24 .
the two original vertices satisfy some overlap identities and so shall do the composite vertex .
one particular overlap identity is given by considering the operator @xmath25 with conformal weight @xmath26 defined by @xcite @xmath27 where @xmath28 are bosonic oscillators with commutation relations @xmath29=-\eta ^{\mu \nu } \delta ^{ij}\delta _ { n ,- m}\ , \
n , m\neq 0\ , \\ \label{eq4.3 } & & [ \alpha ^{\mu i}_0,a^{\nu j}_n]=0\ , \
\forall n\ .\end{aligned}\ ] ] the overlap identity is given by @xmath30 \ .\ ] ] because we are considering the adjoint of leg @xmath6 in vertex @xmath2 , we must see what the adjoint of these overlap identities are .
first , by the definition of @xmath25 , we have that @xmath31 so , the adjoint of the overlap equations is given by @xmath32 v^\dagger = 0\ .\ ] ] we now take the overlap identity considering the effects of the operator @xmath33 on the vertex @xmath1 on a generic leg @xmath34 and on leg @xmath5 or @xmath6 . ]
( figure 3 ) : 0.7 cm figure 3 : overlap identity for @xmath1 .
@xmath35 = 0\ .\ ] ] 0.2 cm we may then insert the unit operator @xmath36 and multiply by @xmath7 without altering the result : @xmath37 { \cal p}\ .\ ] ] since the conformal operator @xmath7 acts only on leg @xmath5 , it will have no effect on @xmath33 . in order to compute the effect of @xmath7 on @xmath38
, we must know that , for a conformal transformation @xmath39 acting on a conformal operator @xmath40 of weight @xmath41 , @xmath42 since @xmath38 has conformal weight @xmath26 , we have @xmath43 and ( figure 4 ) 0.7 cm figure 4 : overlap identity for @xmath44 .
@xmath45 = 0\ .\ ] ] 0.2 cm the second term in the expression above is facing now leg @xmath6 of vertex @xmath15 , or best , its hermitian conjugate . considering that the hermitian conjugate of @xmath25 ( given by ( [ eq4.1a ] ) ) , we then have the following overlap identity between legs @xmath34 and @xmath6 : @xmath46 = 0\ .\ ] ] we can then make a cycling transformation in order to obtain the correct factor for an arbitrary leg @xmath47 of vertex @xmath15 .
the only term that will be affected is the term depending on leg @xmath6 : @xmath48 doing this , the overlap identity for the composite vertex @xmath9 ( figure 5 ) can be written as 3.1 cm figure 5 : overlap identity for @xmath9 .
@xmath49 = 0\ , \ ] ] 0.2 cm which is the overlap equation between two arbitrary legs @xmath34 and @xmath50 of the composite vertex @xmath9 .
but the overlap identity for the composite vertex , since none of legs @xmath34 or @xmath50 involves the propagator , must be given by @xmath51 = 0\ ] ] and so in order for the equation we have obtained for the overlap of the composite vertex @xmath9 to be true we must have @xmath52 which implies that the propagator is given by @xmath53 in order to give an explicit expression for the propagator , we will now choose @xmath54 to be of the form @xmath55 this choice is called the `` simple cycling '' @xcite and it is the one that simplifies our calculations the most . in this choice , the propagator is given by @xmath56 or in terms of the @xmath57 operators ] ] , @xmath58 this form works for all choices for the composite vertex discussed before . ] .
the true propagator is given by expression ( [ eq4.8 ] ) integrated over a suitable variable .
choosing this variable to be @xmath59 , we then have @xmath60 before going any further , we must discuss another aspect of the theory that depends on the particular way in which the legs are identified during the sewing procedure .
let us consider the more general case of an arbitrary cycling @xmath61 .
this kind of cycling may depend on other coordinates that are not @xmath62 . as an example , let us suppose that we are identifying coordinate @xmath14 of leg @xmath5 with coordinate @xmath17 of vertex @xmath15 and coordinate @xmath16 of leg @xmath6 with coordinate @xmath17 of vertex @xmath1 .
the overlap identity between legs @xmath34 and @xmath63 on vertex @xmath1 is @xmath64\ ] ] where we are calling @xmath65 the cycling transformation on leg @xmath50 @xmath66 .
these cycling transformations may depend on the other legs .
as an example when this happens , we take another choice of the cycling transformations @xmath54 that is not as trivial as ( [ eq4.7 ] ) but gives a simpler formula for the propagator .
this choice is given by @xcite @xcite @xmath67 which is the transformation that takes @xmath68 , @xmath62 and @xmath69 to @xmath70 , 0 and 1 , respectively .
its inverse is given by @xmath71 in the case of the cycling given by ( [ eq4.10 ] ) , the cycling for leg @xmath63 will depend on leg @xmath5 , which is not present in the composite vertex . on this vertex , the overlap between leg @xmath63 and an arbitrary leg @xmath50 reads @xmath72 \ .\ ] ] this overlap equation involves terms that depend on leg @xmath5 , which is non - existent in the composite vertex @xmath9 .
the correct overlaps should be given by @xmath73 \
, \ ] ] where the cycling transformations @xmath74 and @xmath75 do not depend on legs @xmath5 or @xmath6 .
so , in order to restore the correct cycling transformation for the composite vertex , a conformal transformation must be made on the cycling transformations on vertex @xmath1 .
these are given by @xmath76 where @xmath77 in this definition , we consider implicit that the transformation @xmath78 since the conformal transformations on leg @xmath5 will not be part of the composite vertex and so need not be modified .
considering the general case , we have that the overlap equation for @xmath1 ( figure 6 ) that will lead to the correct composite vertex is now obtained from the original overlap 0.8 cm figure 6 : overlap identity for @xmath1 . @xmath79 = 0\ , \ ] ] where @xmath80 and @xmath81 . by inserting conformal transformation ( [ eq4.11 ] ) ,
we obtain ( figure 7 ) 0.8 cm figure 7 : overlap identity for @xmath82 .
@xmath83 = 0\ .\ ] ] inserting now the propagator , we obtain ( figure 8) 0.8 cm figure 8 : overlap identity for @xmath84 .
@xmath85 = 0\ ] ] and we expect the composite vertex to have a different form ( given shortly ) than in ( [ eq4.1 ] ) in order to amount for the contributions of the conformal transformations .
the second term of the overlap is now facing leg @xmath6 of vertex @xmath15 so that we have the following overlap between legs @xmath34 and @xmath6 : @xmath86 = 0\ .\ ] ] we are now facing the conformal transformation @xmath87 , @xmath88 may depend on the variable @xmath14 .
] , defined by @xmath89 which is necessary in order to change @xmath90 . introducing this transformation
we obtain ( figure 9 ) 0.8 cm figure 9 : overlap identity for @xmath91 .
@xmath92 = 0\ .\ ] ] making a cycling transformation from leg @xmath6 to leg @xmath50 , we then obtain ( figure 10 ) 0.8 cm figure 10 : overlap identity for @xmath93 .
@xmath94 = 0\ .\ ] ] once again , a conformal transformation must be introduced because of the cycling transformations @xmath65 .
this is defined by @xmath95 so that we now have ( figure 11 ) 0.8 cm figure 11 : overlap identity for @xmath96 .
@xmath97 = 0\ .\ ] ] the composite vertex must be defined in terms of the new cycling transformations and so it must now include the conformal transformations that perform this change .
so , it will now be defined by @xmath98 considering this , the overlap identity for the composite vertex @xmath9 can be written as @xmath49 = 0\ .\ ] ] since the correct overlap identity for the composite vertex is given by @xmath51 = 0\ ] ] we must have @xmath99 which implies once again that @xmath100 for the cycling transformation ( [ eq4.10 ] ) , it is only necessary to do conformal transformations on legs @xmath63 , @xmath5 , @xmath101 , @xmath102 , @xmath6 and @xmath103 , depending on the particular way the variables associated with these legs are identified with the variables associated to legs @xmath5 and @xmath6 . in this particular example ( which is case @xmath104 seen before ) ,
the conformal transformations are given by @xmath105 where @xmath106 for the cycling ( [ eq4.10 ] ) , the propagators obtained for the four possible combinations discussed before are given by @xmath107 where can be connected with the coefficient @xmath108 in reference @xcite by @xmath109 . ]
@xmath110 the true propagators are obtained when we integrate the expressions above multiplied by a suitable constant .
the results are : @xmath111 it is now necessary to verify the effect of the gauge transformations @xmath112 and @xmath113 on the composite vertex as given by formula ( [ eq4.14 ] )
. we shall do it by verifying the effect of @xmath112 on vertex @xmath1 . in order to do this
we need the explicit expression for the bosonic oscillator vertex @xmath1 , given by @xcite @xcite @xmath114\ ] ] where @xmath115 and @xmath116 are cycling transformations involving leg @xmath5 and the oscillators @xmath117 have commutation relations given by ( [ eq4.2 ] , [ eq4.3 ] ) .
matrices @xmath118 are defined in the following way @xcite : @xmath119 ^n\ , \\
d_{nm}(\gamma ) & = & \left .
\sqrt{\frac{m}{n}}\frac{1}{m!}\frac{\partial ^m } { \partial z^m}\left [ \gamma ( z)\right ] ^n\right| _ { z=0}\ , \\ d_{00}(\gamma ) & = & \left .
\frac{1}{2}\ln \left [ \frac{d}{dz}\gamma ( z)\right ] \right| _ { z=0}\end{aligned}\ ] ] and have the following multiplication property : @xmath120 in order to calculate the effects of the conformal transformation @xmath112 on these oscillators we must make use of the following conformal operator @xcite @xmath121 which has conformal weight one , what means it transforms like @xmath122 an oscillator @xmath123 can be expressed in terms of this conformal operator in the following way : @xmath124 acting on it with the conformal transformation @xmath112 , we have @xmath125 making now a change of variables @xmath126 , we have @xmath127 expanding @xmath128 in terms of @xmath118 matrices and @xmath129 in terms of the oscillators , we obtain @xmath130 \alpha ^{\mu i}_p \nonumber \\ & & + \frac{1}{\sqrt{n}}\oint _ { \xi _ i=0}d\xi _ i\ \left [ \sqrt{n}d_{n0}\left ( v_{0i}^{-1}v_i \right ) ( \xi _ i)^{-1 } + \sum_{m=1}^\infty \sqrt{\frac{n}{m}}d_{nm}\left ( v_{0i}^{-1}v_i\right ) ( \xi _ i)^{m-1}\right ] \alpha ^{\mu i}_0\ .\end{aligned}\ ] ] performing the integrations we then obtain @xmath131 using the same process for @xmath132 , we obtain @xmath133 using these transformation properties , the multiplication rules of matrices @xmath118 ( equation ( [ eq4.20 ] ) ) and the property @xcite @xmath134 we can show that the effect of the conformal transformation @xmath112 on vertex @xmath1 is given by @xmath135 \ , \ ] ] i.e. the effect of @xmath112 on vertex @xmath1 is to change @xmath136 and @xmath137 thus eliminating the dependence of the cycling transformations of every leg except leg @xmath5 on the latter .
the same can be done to obtain the effect of conformal transformation @xmath113 on vertex @xmath15 , with the same results .
so , the effect of these transformations is to eliminate from the cycling transformations of the composite vertex all dependence on the sewn legs @xmath5 and @xmath6 .
we now introduce ghosts in the vertex so that what we must sew now are two vertices with some ghost variables attached to them , i.e. we will be considering the physical vertices @xcite which have the correct ghost number . in this case , in addition to satisfying the overlap identities with the conformal operator @xmath25 , the two physical vertices and the composite vertex must also satisfy some overlap identities with the conformal operators @xmath138 and @xmath139 , given by @xcite @xmath140 where @xmath141 and @xmath142 are ghost anticommuting oscillators with anticommutation relations @xmath143 these operators have , respectively , conformal weights @xmath144 and @xmath145 , what means that they transform like @xmath146 the overlap identities for a vertex @xmath39 with these operators are given by @xmath147 = 0\ , \\ \label{eq4.25 } & & v\left [ c^i(\xi _ i)-\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1}c^j(\xi _ j)\right ] = 0\ .\end{aligned}\ ] ] we shall be working here with overlap identities for the physical vertex @xmath148 @xcite , which has the correct ghost number , instead of the overlap identities for vertex @xmath39 .
the physical vertex is given by @xcite @xmath149 where @xmath150 are any three legs of the vertex and the matrix @xmath151 is given by @xmath152 where the cycling transformations are now defined on the complete generators @xmath153 of the conformal algebra of the bosonic oscillators and of the ghost oscillators .
these vectors @xmath151 have the following property : @xmath154 in order to derive the overlap identity for the physical vertex @xmath148 , we must multiply the overlap identity for @xmath39 by the same factor as in equation ( [ eq4.26 ] ) , @xmath155 \left ( \prod _ { \scriptstyle k=1\atop \scriptstyle k\neq a , b , c}^n\sum_{l=1}^n \sum_{n=-1}^\infty e_n^{kl}b_n^l\right ) = 0\ , \\ & & v\left [ c^i(\xi _ i)-\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1}c^j(\xi _ j)\right ] \left ( \prod _ { \scriptstyle k=1\atop \scriptstyle k\neq a , b , c}^n\sum_{l=1}^n \sum_{n=-1}^\infty e_n^{kl}b_n^l\right ) = 0\end{aligned}\ ] ] and pass it through the overlap identities , obtaining @xmath156 = 0\ , \\ & & u\left [ c^i(\xi _ i)-\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1}c^j(\xi _ j)\right ] \nonumber \\ \label{eq4.29 } & & + v\sum_{\scriptstyle p=1\atop p\neq \scriptstyle a , b , c}^n(-1)^p\prod _ { \scriptstyle k=1 \atop { \scriptstyle k\neq a , b , c\atop \scriptstyle k\neq p}}^n\sum_{l=1}^n\sum_{q=-1}^\infty e_q^{kl}b_q^l\sum_{n=-1}^\infty \left [ e_n^{pi}(\xi _ i)^{n+1}-\left ( \frac{d\xi _ j}{d\xi _ i } \right ) ^{-1}e_n^{pj}(\xi _ j)^{n+1 } \right ] = 0\ .\end{aligned}\ ] ] from ( [ eq4.29 ] ) we can see that there will be an anomalous term in the @xmath139 overlap of the physical vertex @xmath148 unless both legs @xmath34 and @xmath50 are precisely those legs ( @xmath104 , @xmath157 or @xmath108 ) that do not have any ghosts attached to them .
these ghosts which are attached to all the other legs are responsible for the anomalous terms . before going any further ,
it is necessary to make some considerations on the ghost number of the composite vertex . as we shall see shortly , in the case where we perform the sewing with ghosts included , using the physical vertices , the resulting composite physical vertex will not have the correct ghost number unless we insert some extra ghosts in vertex @xmath158 before the sewing takes place . considering this
, we shall define the composite vertex to be given by @xmath159 where @xmath160 are some extra ghosts that will be introduced in order to make vertex @xmath161 have the correct ghost number and @xmath13 is the propagator ( in its integrated form ) .
we must now analyze the ghost number of the composite vertex and of its parts in order to calculate the ghost number that the extra ghosts @xmath160 must have . in order to do this
, we shall use the ghost number operator @xmath162 . for a vertex with @xmath0 legs ,
the ghost number operator is defined by @xmath163 the reason why the ghost number operator is a sum from @xmath164 to @xmath165 is because there are @xmath0 vacua that will annihilate the operators corresponding to each one of them .
when acting on the physical vertex , this operator gives a ghost number @xmath0 , what is the correct ghost number for a tree vertex with @xmath0 legs . in the case of the composite vertex ,
it has @xmath166 legs ( because it does not have legs @xmath5 and @xmath6 , which have been sewn together ) and so it must have ghost number @xmath167 . for this vertex
, the ghost number operator @xmath168 can be divided into two parts : @xmath169 where @xmath170 this ghost number operator will have the following effect on the composite vertex : @xmath171 given formula ( [ eq4.30 ] ) for the composite vertex , we then have @xmath172 pu_2^\dagger + u_1gpu_2^\dagger n_2^{\rm gh}\ .\ ] ] in order to calculate this , we must pay some attention to terms one and three of the right hand side of the expression above .
we know that @xmath173 where @xmath174 taking the hermitian conjugate of equation ( [ eq4.34 ] ) only on leg @xmath6 , we obtain @xmath175 since for @xmath176 ( and for any arbitrary ghost number operator ) @xmath177 we then obtain , substituting ( [ eq4.33 ] ) , ( [ eq4.35 ] ) and ( [ eq4.36 ] ) into equation ( [ eq4.32 ] ) , @xmath178 pu_2^\dagger -u_1n_e^{\rm gh}gpu_2^\dagger + u_1gpn_f^{\rm gh}u_2^\dagger \ .\ ] ] passing @xmath179 through the extra ghosts @xmath160 , we then obtain @xmath180 pu_2^ \dagger \nonumber \\ \label{eq4.37 } & & -u_1\left [ n_e^{\rm gh},g\right ] pu_2^\dagger -u_1gpn_e^{\rm gh}u_2^\dagger + u_1gpn_f^{\rm gh}u_2^\dagger \ .\end{aligned}\ ] ] we must now remember that , in the composite vertex , we identify every operator on leg @xmath5 with operators on leg @xmath6 so that @xmath181
. doing this , the last two terms in ( [ eq4.37 ] ) cancel and we obtain the following result : @xmath182 pu_2^\dagger \ .\ ] ] the fact that @xmath161 has ghost number @xmath166 then implies that @xmath183 = g\ , \ ] ] i.e. the extra ghosts that must be introduced in vertex @xmath158 must have ghost number 1 . and @xmath6 to have or not to have ghosts attached to them .
] there is an infinite number of combinations of ghosts that have ghost number 1 .
we could have any linear combination of ghosts of the type @xmath157 , @xmath184 , @xmath185 , etc .
but it will prove to be simpler to choose @xmath160 to be a combination of @xmath186 only so that we may represent it as @xmath187 where @xmath188 are the coefficients of the linear combination . in order to determine the correct linear combination , we must use some other conditions , like brst invariance of the scattering amplitude .
this we shall see next .
we must now impose that the scattering amplitude obtained from the composite vertex is brst invariant and check whether this condition is strong enough to determine @xmath160 .
the scattering amplitude @xcite is obtained by acting with the composite vertex @xmath189 on a certain number of physical states ( latexmath:[$|\chi _ 1\rangle |\chi _
2\rangle \dots @xmath62 ( @xmath191 ; @xmath192 ) : latexmath:[\[w=\int \prod_{\scriptstyle i=1\atop \scriptstyle i\neq e , f}^{n_1+n_2}dz_i\ u_1gpu_2^\dagger @xmath13 is the propagator in its integrated form and @xmath160 are the extra ghosts to be inserted in @xmath158 .
the action of the brst charge @xmath194 on this scattering amplitude is given by @xmath195 gpu_2^\dagger |\chi _
1\rangle |\chi _ 2\rangle \dots |\chi
_ n\rangle \nonumber \\ & & + \int \prod_{\scriptstyle i=1\atop \scriptstyle i\neq e , f}^{n_1+n_2}dz_i\ u_1 \left [ g , q\right ] pu_2^\dagger |\chi _
1\rangle |\chi _ 2\rangle \dots |\chi _ n\rangle + \int \prod_{\scriptstyle i=1\atop \scriptstyle i\neq e , f}^{n_1+n_2}dz_i\ u_1g\left [ p , q\right ] u_2^\dagger |\chi _ 1\rangle |\chi _
2\rangle \dots |\chi _
n\rangle \nonumber \\ & & + \int \prod_{\scriptstyle i=1\atop \scriptstyle i\neq e , f}^{n_1+n_2}dz_i\ u_1gp \left [ u_2^\dagger , q\right ] |\chi _ 1\rangle |\chi _ 2\rangle \dots
|\chi _ n\rangle \ .\end{aligned}\ ] ] the first and third terms will result in total derivatives that give zero when one integrates over some variables @xcite and so what remain are just the second and third terms . the commutator @xmath13 is given by a pure conformal transformation , and it is a function of the generators @xmath196 only .
as the brst charge commutes with all @xmath197 s , i.e. @xmath198 = 0\ ] ] we have @xmath199 = 0\ .\ ] ] considering now that @xmath200 = { \cal l}_n^i\ , \ ] ] we then have , for @xmath160 given by ( [ eq4.38 ] ) , @xmath201 = u_1\sum_{i=1}^{n_1}\sum_{n=-1}^\infty \alpha _ n^i{\cal l}_n^i\ .\ ] ] in order for the scattering amplitude @xmath202 to be brst invariant , expression ( [ eq4.39 ] ) must be zero or a total derivative ( that can be integrated out to become a null surface term ) . at the same time , we want these extra ghosts to place ( talking in terms of the simple cycling ) a ghost on one of the legs in @xmath158 that do not have any ghosts attached to them .
if we now remember property ( [ eq4.27 ] ) , we see that we can satisfy these constraints in a nice way by choosing @xmath160 to be given by @xmath203 where @xmath204 is one of the legs of vertex @xmath158 that does not have ghosts attached to it . inserting these ghosts in vertex @xmath158 , we have @xmath205 using formula ( [ eq4.40 ] ) for the extra ghosts @xmath160 , we then have @xmath206 & = & u_1(-1)^{n_1+a } \sum_{j=1}^{n_1 } \sum_{n=-1}^\infty e_n^{aj } \left [ b_n^j , q\right ] = ( -1)^{n_1 + a}u_1\sum_{j=1}^{n_1 } \sum_{n=-1}^\infty e_n^{aj } { \cal l}_n^j \nonumber \\ & = & ( -1)^{n_1+a } \frac{\partial v_1}{\partial z_a } \prod_{\scriptstyle i=1\atop \scriptstyle i\neq a , b , c}^{n_1}\sum_{j=1}^{n_1 } \sum_{n=-1}^\infty e_n^{ij}b_n^j\end{aligned}\ ] ] what is a total derivative that will vanish when one does the integration over @xmath207 . in references @xcite and @xcite , the extra ghosts have been placed in the propagator .
although this can be done , there is no way one can derive a formula for the ghosts in the propagator for a general cycling . in that case , the extra ghosts must be derived and brst invariance has to be checked for each particular cycling .
also , the resulting composite vertex obtained in that case is not similar in its ghost structure to an ordinary tree vertex , although it has the correct ghost number .
we must now use the overlap identities to determine the propagator that satisfies them . in order to do this
we shall start with vertex @xmath208 , which is the vertex with cycling transformations @xmath209 which involve leg @xmath5 .
considering equations ( [ eq4.28 ] ) and ( [ eq4.29 ] ) , the overlap identities for vertex @xmath208 between an arbitrary leg @xmath34 and leg @xmath5 are given by ( figure 12 ) 0.8 cm figure 12 : overlap identity for @xmath208 .
@xmath210 = 0\ , \\ & & u_1 ^ 0\left [ c^i(\xi _ { 0i})-\left ( \frac{d\xi _ { 0e}}{d\xi _ { 0i}}\right ) ^{-1}c^e(\xi _ { 0e } ) \right ] \nonumber \\ \label{eq4.43 } & & + v^0_1\sum_{\scriptstyle p=1\atop p\neq \scriptstyle a , b , c}^{n_1}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq a , b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{l=1}^{n_1}\sum_{q=-1}^\infty e_q^{kl}b_q^l \sum_{n=-1}^\infty \left [ e_n^{pi}(\xi _ { 0i})^{n+1}-\left ( \frac{d\xi _ { 0e}}{d\xi _ { 0i}}\right ) ^{-1}e_n^{pe } ( \xi _ { 0e})^{n+1 } \right ] = 0\ .\end{aligned}\ ] ] the extra ghosts must then be inserted in vertex @xmath208 so that the composite vertex will have the correct ghost number .
multiplying expressions ( [ eq4.42 ] ) and ( [ eq4.43 ] ) by the extra ghosts @xmath160 ( given by ( [ eq4.40 ] ) ) and passing them through the overlaps we obtain ( figure 13 ) 0.8 cm figure 13 : overlap identity for @xmath211 .
@xmath212 = 0\ , \\ & & u_1 ^ 0g\left [ c^i(\xi _ { 0i})-\left ( \frac{d\xi _ { 0e}}{d\xi _ { 0i}}\right ) ^{-1 } c^e(\xi _ { 0e})\right ] + u^0_1(-1)^{n_1+a}\sum_{n=-1}^\infty \left [ e_n^{ai}(\xi _ { 0i})^{n+1 } -\left ( \frac{d\xi _ { 0e}}{d\xi _ { 0i}}\right ) ^{-1}e_n^{ae}(\xi _ { 0e})^{n+1 } \right ] \nonumber \\ & & + v_1 ^ 0\sum_{\scriptstyle p=1\atop p\neq \scriptstyle a , b , c}^{n_1}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq a , b , c\atop \scriptstyle k\neq p}}^{n_1 } \left ( \sum_{l=1}^{n_1}\sum_{q=-1}^\infty e_q^{kl}b_q^l\right ) \left ( \sum_{j=1}^{n_1 } \sum_{n=-1}^\infty e_n^{aj}b_n^j\right ) \nonumber \\ \label{eq4.45 } & & \times \sum_{n=-1}^\infty \left [ e_n^{pi}(\xi _ { 0i})^{n+1}-\left ( \frac{d\xi _ { 0e}}{d\xi _ { 0i}}\right ) ^{-1}e_n^{pe}(\xi _ { 0e})^{n+1 } \right ] = 0\ .\end{aligned}\ ] ] the second and third terms of equation ( [ eq4.45 ] ) can be combined so that it becomes @xmath213 \nonumber \\
\label{eq4.46 } & & + v_1 ^ 0\sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod
_ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \left ( \sum_{l=1}^{n_1}\sum_{q=-1}^\infty e_q^{kl}b_q^l\right ) \sum_{n=-1}^\infty \left [ e_n^{pi}(\xi _ { 0i})^{n+1}-\left ( \frac{d\xi _ { 0e}}{d\xi _ { 0i}}\right ) ^{-1 } e_n^{pe}(\xi _ { 0e})^{n+1 } \right ] = 0\ .\end{aligned}\ ] ] at this point , we must introduce conformal transformations of the type of @xmath112 , given by ( [ eq4.11 ] ) in order to have at the end the correct cycling transformations for the composite vertex . in order to do this
we need to use matrices @xmath214 , defined by @xcite @xmath215 \right| _ { z=0}\ ] ] which have the following properties : @xmath216 the action of the operator @xmath112 on the ghosts @xmath142 can then be calculated in the following way : first we write @xmath142 in terms of an integral over the conformal operator @xmath217 @xmath218 then we insert the operator @xmath112 : @xmath219 after a change of variables @xmath220 we have is a polynomial in @xmath221 ( with no constant term ) , then @xmath222 . ]
@xmath223 using matrices @xmath214 , we then may expand @xmath221 in terms of @xmath54 . if we also expand @xmath224 , we then obtain @xmath225 performing the integration we then have @xmath226 using ( [ eq4.52 ] ) in equations ( [ eq4.44 ] ) and ( [ eq4.46 ] ) and multiplying ( [ eq4.44 ] ) by @xmath227 and ( [ eq4.46 ] ) by @xmath228 , we then have ( figure 14 ) figure 14 : overlap identity for @xmath229 . @xmath230 = 0\ , \\ & & u_1 ^ 0gc_1^{-1 } \left [ c^i(\xi _ i ) -\left ( \frac{d\xi _ e}{d\xi _ i}\right ) ^{-1 } c^e(\xi _ e)\right ] + v_1 ^ 0c_1^{-1 } \sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c
\atop \scriptstyle k\neq p}}^{n_1 } \left [ \sum_{l=1}^{n_1 } \sum_{q , t=-1}^\infty e_q^{kl } e_{qt}\left ( v_{0l}^{-1}v_l\right ) b_t^l\right ] \nonumber \\ \label{eq4.56 } & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}\left ( v_{0i}^{-1}v_i\right ) ( \xi _ i)^{m+1 } -\left ( \frac{d\xi _ e}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}\left ( v_{0e}^{-1 } v_e\right ) ( \xi _ e)^{m+1 } \right ] = 0\ .\end{aligned}\ ] ] before going further , some words must be said about the effects of @xmath231 on vertex @xmath208 with the extra ghosts @xmath160 . this is given explicitly by @xcite@xcite @xmath232 \nonumber \\ & & \times \prod_{r=-1}^1 \sum_{i=1}^{n_1 } \sum_{s=-1}^1 e_{rs}(v_{0i})b_s^i \times \prod_{\scriptstyle i=1\atop \scriptstyle i\neq b , c}^{n_1 } \sum_{j=1}^n \sum_{n=-1}^\infty e_n^{ij}b_n^j\ .\end{aligned}\ ] ] making use of matrices @xmath233 , defined by @xcite @xmath234 ^{n-2}\left [ \frac{\partial } { \partial z}\gamma ( z)\right ] ^{-1 } \right\ } \right| _ { z=0}\ ] ] we may calculate in a similar way as we did for the @xmath142 ghosts the effect of @xmath112 on the @xmath141 ghosts , obtaining @xmath235 using this together with the property @xmath236 and equation ( [ eq4.52 ] ) , we may then show that the result of acting with @xmath231 on @xmath211 is @xmath237 \nonumber \\ \label{eq4.58 } & & \times \prod_{r=-1}^1 \sum_{i=1}^{n_1 } \sum_{s=-1}^1 e_{rs}(v_i)b_s^i \times \prod_{\scriptstyle
i=1 \atop \scriptstyle i\neq b , c}^{n_1 } \sum_{j=1}^n \sum_{n , m=-1}^\infty e_n^{ij}e_{nm}(v_{0j}v_j)b_m^j\ .\end{aligned}\ ] ] so we can see that in this case the action of @xmath231 on @xmath211 is not just to change @xmath238 . because of the peculiar nature of @xmath151 , it transforms as @xmath239 only in one particular group of cycling transformations ( as we shall see later ) will this be just equivalent to changing @xmath238 .
we shall call from now on @xmath240 and @xmath241 .
the calculation for vertex @xmath242 will be similar to the one we have just made for @xmath229 . having done this , we must insert the propagator @xmath7 into the overlap identities ( [ eq4.55 ] ) and ( [ eq4.56 ] ) in the same way as in the case with no ghosts .
but now we must take extra care since there are terms depending on @xmath243 in the second term of equation ( [ eq4.56 ] ) . using equation ( [ eq4.52 ] ) as a guideline
, we have @xmath244 so that the result of inserting @xmath7 into overlaps ( [ eq4.55 ] ) and ( [ eq4.56 ] ) is ( figure 15 ) 0.8 cm figure 15 : overlap identity for @xmath245 . @xmath246 = 0\ , \\ & & u_1{\cal p}\left [ c^i(\xi _ i)-\left ( \frac{d}{d\xi _ i}{\cal p}^{-1}\xi _
e\right ) ^{-1}c^e \left ( { \cal p}^{-1}\xi _
e\right ) \right ] \nonumber \\ & & + v_1{\cal p}\sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1 } e_q^{kl } e_{qt}(v_{0l}^{-1}v_l)b_t^l + e_q^{ke}e_{qt}(v_{0e}^{-1}v_e{\cal p})b_t^e\right ] \nonumber \\ \label{eq4.60 } & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}(v_{0i}^{-1}v_i)(\xi _ i)^{m+1 } -\left ( \frac{d\xi _ e}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}(v_{0e}^{-1}v_e)(\xi _ e)^{m+1 } \right ] = 0\ .\end{aligned}\ ] ] the operators of the overlap equations are now facing leg @xmath6 of vertex @xmath247 .
in order to obtain the overlap identities for this leg , we must now identify the operators of leg @xmath5 with the ones of leg @xmath6 , which are adjoint operators : @xmath248 first , as @xmath249 and @xmath250 are conformal operators with weights 2 and @xmath145 , respectively , we have @xmath251 then , for @xmath252 and @xmath253 , we obtain @xmath254 then , we must also make the change @xmath255 so , the overlap equations become @xmath256 = 0 \ , \\ & & u_1{\cal p}\left [ c^i(\xi _ i)+\left ( \frac{d}{d\xi _ i}{\cal p}^{-1}\xi _
e\right ) ^{-1 } \left ( { \cal p}^{-1}\xi _
e\right ) ^2c^f\left ( \gamma { \cal p}^{-1}\xi _
e\right ) \right ] \nonumber \\ & & + v_1{\cal p}\sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod _
{ \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}e_q^{kl } e_{qt}\left ( v_{0l}^{-1}v_l\right ) b_t^l + e_q^{ke}e_{qt}\left ( v_{0e}^{-1}v_e{\cal p}\right ) b_{-t}^f\right ] \nonumber \\ \label{eq4.62 } & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}\left ( v_{0i}^{-1}v_i\right ) ( \xi _ i)^{m+1 } -\left ( \frac{d\xi _ f}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}\left ( v_{0e}^{-1 } v_f\right ) ( \xi _ f)^{m+1}\right ] = 0\ .\end{aligned}\ ] ] we are then facing the conformal transformation @xmath87 that takes @xmath257 into @xmath258 . inserting this transformation we obtain ( figure 16 ) : 0.8 cm figure 16 : overlap identity for @xmath259 .
@xmath260 = 0 \ , \\ & & u_1{\cal p}c_f\left [ c^i(\xi _ i)+\left ( \frac{d}{d\xi _ i}{\cal p}^{-1}\xi _
e\right ) ^{-1 } \left ( { \cal p}^{-1}\xi _
e\right ) ^2 \left ( \frac{dv_{0f}^{-1}v_f\gamma { \cal p}^{-1}\xi _ e } { d\gamma { \cal p}^{-1}\xi _
e}\right ) ^{-1}c^f\left ( v_{0f}^{-1}v_f\gamma { \cal p}^{-1}\xi _ e \right ) \right]\nonumber \\ & & + v_1{\cal p}c_f\sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}e_q^{kl } e_{qt}\left ( v_{0l}^{-1}v_l\right ) b_t^l -e_q^{ke}e_{qt}\left ( v_{0e}^{-1}v_e{\cal p}\gamma v_f^{-1}v_{0f}\right ) b_t^f\right ] \nonumber \\ \label{eq4.64 } & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}\left ( v_{0i}^{-1}v_i\right ) ( \xi _ i)^{m+1 } -\left ( \frac{d\xi _ { 0f}}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}\left ( v_{0e}^{-1 } v_{0f}\right ) ( \xi _ { 0f})^{m+1}\right ] = 0\ .\end{aligned}\ ] ] we are facing now vertex @xmath261 .
this vertex satisfies the following overlap identity @xcite : @xmath262 using equation ( [ eq4.50 ] ) , we have @xmath263 all terms on @xmath264 with @xmath265 get annihilated by the conjugate vacuum @xmath266 , while we may use identity ( [ eq4.65 ] ) to substitute the terms in @xmath267 , @xmath268 . doing this , equation ( [ eq4.64 ] )
becomes @xmath269 \nonumber \\ & & + v_1{\cal p}c_f\sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod _
{ \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}e_q^{kl } e_{qt}\left ( v_{0l}^{-1}v_l\right ) b_t^l+\sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2 } e_q^{ke}e_{qt}\left ( v_{0e}^{-1}v_e{\cal p}\gamma v_f^{-1}v_{0l}\right ) b_t^l\right ] \nonumber \\ \label{eq4.66 } & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}\left ( v_{0i}^{-1}v_i\right ) ( \xi _ i)^{m+1 } -\left ( \frac{d\xi _ { 0f}}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}\left ( v_{0e}^{-1}v_{0f}\right ) ( \xi _ { 0f})^{m+1}\right ] = 0\ .\end{aligned}\ ] ] equations ( [ eq4.63 ] ) and ( [ eq4.66 ] ) are the overlap identities between legs @xmath34 and @xmath6 . in order to obtain the overlaps between leg @xmath34 of vertex @xmath158 and an arbitrary leg @xmath50 of vertex @xmath247 , we must now perform a cycling transformation that will take the operators from leg @xmath6 to leg @xmath50 .
the effect of this transformation on @xmath270 is @xmath271 then , we must also write @xmath272 doing this , overlap equations ( [ eq4.63 ] , [ eq4.66 ] ) become ( figure 17 ) 0.8 cm figure 17 : overlap identity for @xmath273 .
@xmath274 = 0\ , \\ & & u_1{\cal p}c_fv_2^{0\dagger } \left [ c^i(\xi _ i)+\left ( \frac{d}{d\xi _ i}{\cal p}^{-1 } \xi _ e\right ) ^{-1}\left ( { \cal p}^{-1}\xi _
e\right ) ^2 \left ( \frac{dv_{0j}^{-1}v_f\gamma { \cal p}^{-1}\xi _ e}{d\gamma { \cal p}^{-1}\xi _
e}\right ) ^{-1}c^f\left ( v_{0j}^{-1}v_f\gamma { \cal p}^{-1}\xi _
e\right ) \right ] \nonumber \\ & & + v_1{\cal p}c_fv_2^{0\dagger } \ !
\sum_{\scriptstyle p=1\atop p\neq \scriptstyle b , c}^{n_1 } ( -1)^p\ ! \ ! \ ! \ ! \prod
_ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \ !
\sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l \neq e}^{n_1}e_q^{kl}e_{qt}\left ( v_{0l}^{-1}v_l\right ) b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2}e_q^{ke}e_{qt}\left ( v_{0e}^{-1}v_e{\cal p}\gamma v_f^{-1}v_{0l } \right ) b_t^l\right ] \nonumber \\ \label{eq4.68 } & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}\left ( v_{0i}^{-1}v_i\right ) ( \xi _ i)^{m+1 } -\left ( \frac{d\xi _ { 0j}}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}\left ( v_{0e}^{-1}v_{0j}\right ) ( \xi _ { 0j})^{m+1}\right ] = 0\ .\end{aligned}\ ] ] the operators are now facing the ghosts that surround vertex @xmath275 ( like in equation ( [ eq4.26 ] ) ) : @xmath276 so that we must insert these ghosts into the expressions for the overlaps . before doing that , we must notice that the extra ghosts acting on vertex @xmath261 have at their left both the conformal transformation @xmath87 and the propagator @xmath7 so that we must first pass them through in order to reach vertex @xmath158 : @xmath277 then , identifying legs @xmath5 and @xmath6 , we have the following expression for the ghosts : @xmath278 { \cal p}\ .\ ] ] we may now pass it through the conformal transformation @xmath112 , obtaining @xmath279 { \cal p}\ .\ ] ] this is now facing vertex @xmath280 , which satisfies the following overlap identity @xcite : @xmath281 using this identity , we may then write the extra ghosts as @xmath282 .\ ] ] passing it back through @xmath112 we then have @xmath283 .\ ] ] we now insert these ghosts into expressions ( [ eq4.67 ] ) and ( [ eq4.68 ] ) .
we do so by multiplying them by ( [ eq4.69 ] ) and passing it through the first term of the overlaps .
what we obtain is ( figure 18 ) 0.8 cm figure 18 : overlap identity for @xmath284 .
@xmath285 = 0\ , \\ & & u_1{\cal p}c_fu_2^{0\dagger } \left [ c^i(\xi _ i)+\left ( \frac{d}{d\xi _ i}{\cal p}^{-1 } \xi _ e\right ) ^{-1}\left ( { \cal p}^{-1}\xi _
e\right ) ^2 \left ( \frac{dv_{0j}^{-1}v_f\gamma { \cal p}^{-1}\xi _ e}{d\gamma { \cal p}^{-1}\xi _
e}\right ) ^{-1}c^f\left ( v_{0j}^{-1}v_f\gamma { \cal p}^{-1}\xi _
e\right ) \right ] \nonumber \\ & & + u_1{\cal p}c_fv_2^{0\dagger } \sum_{p=1}^{n_2}(-1)^p\prod_{\scriptstyle k=1\atop { \scriptstyle k\neq d , g , h\atop \scriptstyle k\neq p}}^{n_2}\left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2}\sum_{q=-1}^\infty e_q^{kl}b_q^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}\sum_{q , t=-1}^\infty e_q^{kf}e_{qt}\left ( v_{0f}^{-1}v_f\gamma { \cal p}^{-1}v_e^{-1}v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{n=-1}^\infty \left [ \sum_{m=-1}^\infty e_n^{pf}e_{nm}\left ( v_{0f}^{-1}v_f \gamma { \cal p}^{-1}v_e^{-1}v_i\right ) ( \xi _ i)^{m+1 } \right .
\nonumber \\ & & \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \left .
+ \left ( \frac{d}{d\xi _ i}{\cal p}^{-1 } \xi _ e\right ) ^{-1}\left ( { \cal p}^{-1}\xi _
e\right ) ^2 \left ( \frac{dv_{0j}^{-1}v_f \gamma { \cal p}^{-1}\xi _ e}{d\gamma { \cal p}^{-1}\xi _
e}\right ) ^2e_n^{pj}\left ( v_{0j}^{-1}v_f\gamma { \cal p}^{-1}\xi _
e\right ) ^{n+1}\right ] \nonumber \\ & & + v_1{\cal p}c_fv_2^{0\dagger } \sum_{\scriptstyle p=1\atop \scriptstyle p\neq b , c}^{n_1 } ( -1)^p\ ! \ ! \ ! \ ! \prod
_ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k \neq p}}^{n_1}\sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l \neq e}^{n_1}e_q^{kl}e_{qt}(v_{0l}^{-1}v_l)b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l \neq f}^{n_2}e_q^{ke}e_{qt}\left ( v_{0e}^{-1}v_e{\cal p}\gamma v_f^{-1}v_{0l}\right ) b_t^l \right ] \nonumber \\ & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}(v_{0i}^{-1}v_i)(\xi _ i)^{m+1}-\left ( \frac{d\xi _ { 0j}}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}(v_{0e}^{-1}v_e)(\xi _ { 0j})^{m+1}\right ] \nonumber \\ \label{eq4.71 } & & \times \prod_{\scriptstyle k=1\atop \scriptstyle k\neq d , g , h}^{n_2}\left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2}\sum_{q=-1}^\infty e_q^{kl}b_q^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}\sum_{q , t=-1}^\infty e_q^{kf}e_{qt } \left ( v_{0f}^{-1}v_f\gamma { \cal p}^{-1}v_e^{-1}v_l\right ) b_t^l\right ] = 0\ .\end{aligned}\ ] ] we are then facing the last term of this composite vertex : the conformal transformation @xmath113 on vertex @xmath261 . inserting it into equations ( [ eq4.70 ] ) and ( [ eq4.71 ] ) , we obtain ( figure 19 ) 0.8 cm figure 19 : overlap identity for @xmath286 . @xmath287 = 0\ , \\ & & u_1{\cal p } u_2^\dagger \left [ c^i(\xi _ i ) + \left ( \frac{d}{d\xi _ i}{\cal p}^{-1 } \xi _ e \right ) ^{-1 } \left ( { \cal p}^{-1}\xi _
e\right ) ^2 \left ( \frac{dv_j^{-1 } v_f\gamma { \cal p}^{-1 } \xi _ e}{d \gamma { \cal p}^{-1 } \xi _ e}\right ) ^{-1 } c^f\left ( v_j^{-1 } v_f \gamma { \cal p}^{-1 } \xi _ e\right ) \right ] \nonumber \\ & & + u_1 { \cal p } v_2^\dagger \sum_{p=1}^{n_2}(-1)^p \prod_{\scriptstyle k=1 \atop { \scriptstyle k\neq d , g , h \atop \scriptstyle k\neq p}}^{n_2 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1 \atop \scriptstyle l\neq f}^{n_2 } e_q^{kl } e_{qt}(v_{0l}^{-1}v_l ) b_t^l+ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1 } e_q^{kf}e_{qt}\left ( v_{0f}^{-1 } v_f \gamma { \cal p}^{-1 } v_e^{-1}v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{n=-1}^\infty \left [ \sum_{m=-1}^\infty e_n^{pf}e_{nm } \left ( v_{0f}^{-1 } v_f \gamma { \cal p}^{-1 } v_e^{-1 } v_i\right ) ( \xi _ i)^{m+1 } \right .
\nonumber \\ & & \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \left .
+ \left ( \frac{d}{d\xi _ i}{\cal p}^{-1 } \xi _ e\right ) ^{-1}\left ( { \cal p}^{-1 } \xi _ e\right ) ^2 \left ( \frac{dv_j^{-1 } v_f\gamma { \cal p}^{-1}\xi _ e}{d\gamma { \cal p}^{-1}\xi _
e}\right ) ^2e_n^{pj}\left ( v_j^{-1 } v_f \gamma { \cal p}^{-1}\xi _
e\right ) ^{n+1}\right ] \nonumber \\ & & + v_1{\cal p}v_2^\dagger \sum_{\scriptstyle p=1\atop \scriptstyle p\neq \scriptstyle b , c}^{n_1}(-1)^p \prod _
{ \scriptstyle k=1 \atop { \scriptstyle k\neq b , c \atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1 \atop \scriptstyle l\neq e}^{n_1}e_q^{kl } e_{qt}(v_{0l}^{-1}v_l ) b_t^l + \sum_{\scriptstyle l=1 \atop \scriptstyle l\neq f}^{n_2 } e_q^{ke}e_{qt } \left ( v_{0e}^{-1}v_e{\cal p } \gamma v_f^{-1}v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}(v_{0i}^{-1}v_i ) ( \xi _ i)^{m+1}- \left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1 } e_n^{pe}e_{nm}(v_{0e}^{-1}v_e)(\xi _ j)^{m+1 } \right ] \nonumber \\ \label{eq4.73 } & & \times \prod_{\scriptstyle k=1 \atop \scriptstyle k \neq d , g , h}^{n_2 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2 } e_q^{kl } e_{qt}(v_{0l}^{-1}v_l ) b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1 } e_q^{kf } e_{qt } \left ( v_{0f}^{-1}v_f \gamma { \cal p}^{-1}v_e^{-1 } v_l\right ) b_t^l\right ] = 0\end{aligned}\ ] ] where we have called @xmath288 and @xmath289
. we must now extract the ghosts from vertex @xmath158 in the second term of equation ( [ eq4.73 ] ) : @xmath290 { \cal p } \nonumber \\ & = & v_1{\cal p}\prod_{\scriptstyle k=1\atop \scriptstyle k\neq b , c}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1 } e_q^{kl}e_{qt}(v_{0l}^{-1}v_l)b_t^l+e_q^{ke}e_{qt}(v_{0e}^{-1}v_e{\cal p})b_t^e\right ] \ .\end{aligned}\ ] ] identifying @xmath291 with @xmath292 and passing now these ghosts through the conformal transformation @xmath87 , we obtain @xmath293 \ .\ ] ] using now overlap identities ( [ eq4.65 ] ) for vertex @xmath15 and inserting the conformal transformation @xmath113 , we then obtain @xmath294 .\ ] ] substituting ( [ eq4.74 ] ) into equation ( [ eq4.73 ] ) we then obtain the overlap identities between legs @xmath34 and @xmath50 : @xmath295 = 0\ , \\ & & u_c\left [ c^i(\xi _ i ) + \left ( \frac{d}{d\xi _ i } { \cal p}^{-1}v^{-1}_ev_i\xi _
i \right ) ^{-1 } \left ( { \cal p}^{-1}v_e^{-1 } v_i\xi _ i\right ) ^2 \right .
\nonumber \\ & & \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \ \left .
\times \left ( \frac{d}{d\xi _ i } { \cal p}^{-1 } v_e^{-1}v_i\xi _ i\right ) ^{-1 } c^j \left ( v_j^{-1 } v_f\gamma { \cal p}^{-1 } v_e^{-1}v_i\xi _ i\right ) \right ] \nonumber \\ & & + v_c \prod _ { \scriptstyle k=1 \atop { \scriptstyle k\neq b , c \atop \scriptstyle k \neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle
l=1 \atop \scriptstyle l \neq e}^{n_1 } e_q^{kl}e_{qt } ( v_{0l}^{-1}v_l)b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2 } e_q^{ke}e_{qt } \left ( v_{0e}^{-1 } v_e{\cal p } \gamma v_f^{-1 } v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{\scriptstyle p=1 \atop \scriptstyle p\neq d , g , h}^{n_2}{-1}^p \prod_{\scriptstyle k=1 \atop { \scriptstyle k \neq d , g , h \atop \scriptstyle k \neq p}}^{n_2 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1 \atop \scriptstyle l\neq e}^{n_1 } e_q^{kf } e_{qt } \left ( v_{0f}^{-1 } v_f \gamma { \cal p}^{-1 } v_e^{-1 } v_l \right ) b_t^l + \sum_{\scriptstyle l=1 \atop \scriptstyle l\neq f}^{n_2 } e_q^{kl } e_{qt } ( v_{0l}^{-1}v_l ) b_t^l\right ] \nonumber \\ & & \times \sum_{n , m=-1}^\infty \left [ e_q^{pf } e_{nm}\left ( v_{0f}^{-1 } v_f{\cal p}^{-1 } v_e^{-1}v_i\right ) ( \xi _ i)^{m+1 } + \left ( \frac{d}{d\xi _ i}{\cal p}^{-1 } v_e^{-1}v_i\xi _ i \right ) ^{-1}\left ( { \cal p}^{-1}v_e^{-1 } v_i\xi _ i\right ) ^2\right .
\nonumber \\ & & \times \left .
( \frac{dv_j^{-1 } v_f\gamma { \cal p}^{-1}\xi _ e}{d \gamma { \cal p}^{-1 } \xi _ e}\right ) ^2 e_n^{pj } e_{nm}(v_{0j}^{-1}v_j ) \left ( v_j^{-1}v_f\gamma { \cal p}^{-1}v_e^{-1 } v_i\xi _ i\right ) ^{m+1}\right ] \nonumber \\ & & + v_c\sum_{\scriptstyle p=1\atop \scriptstyle p\neq b , c}^{n_1}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1 } e_q^{kl } e_{qt}(v_{0l}^{-1}v_l ) b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2 } e_q^{ke } e_{qt}\left ( v_{0e}^{-1}v_e{\cal p } \gamma v_f^{-1}v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi } e_{nm}(v_{0i}^{-1}v_i ) ( \xi _ i)^{m+1 } -\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1 } e_n^{pe } e_{nm } \left ( v_{0e}^{-1}v_j\right ) ( \xi _ j)^{n+1 } \right ] \nonumber \\ & & \times \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq d , g , h\atop \scriptstyle k \neq p}}^{n_2 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l \neq e}^{n_1 } e_q^{kf}e_{qt}\left ( v_{0f}^{-1}v_f \gamma { \cal p}^{-1}v_e^{-1}v_l\right ) b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2 } e_q^{kl}e_{qt}(v_{0l}^{-1 } v_l)b_t^l\right ] = 0\ .\end{aligned}\ ] ] if we now impose that these are the correct overlap equations between legs @xmath34 and @xmath50 of the composite vertex we then must have : @xmath296 what fixes the propagator as @xmath297 which is the same form of the propagator for the bosonic part , but now with the cyclings defined on the complete generators @xmath153 .
the overlap equations now read @xmath298
= 0\ , \\ & & u_c\left [ c^i(\xi _ i)-\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1}c^j(\xi _ j)\right ] + v_c\prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}e_q^{kl } e_{qt}(v_{0l}^{-1}v_l)b_t^l+\sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2}e_q^{ke } e_{qt}\left ( v_{0e}^{-1}v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{\scriptstyle p=1\atop \scriptstyle p\neq d , g , h}^{n_2}(-1)^p \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq d , g , h\atop \scriptstyle k\neq p}}^{n_2 } \sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}e_q^{kf } e_{qt}\left ( v_{0f}^{-1}v_l\right ) b_t^l+\sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2 } e_q^{kl}e_{qt}(v_{0l}^{-1}v_l)b_t^l\right ] \nonumber \\ & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pf}e_{nm}\left ( v_{0f}^{-1}v_i\right ) ( \xi _ i)^{m+1}-\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1}e_n^{pj}e_{nm}(v_{0j}^{-1}v_j ) ( \xi _ j)^{m+1 } \right ] \nonumber \\ & & + v_c\sum_{\scriptstyle p=1\atop \scriptstyle p\neq b , c}^{n_1}(-1)^p\prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq b , c\atop \scriptstyle k\neq p}}^{n_1}\sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l\neq e}^{n_1}e_q^{kl}e_{qt}(v_{0l}^{-1}v_l ) b_t^l + \sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2}e_q^{ke}e_{qt}\left ( v_{0e}^{-1 } v_l\right ) b_t^l\right ] \nonumber \\ & & \times \sum_{n , m=-1}^\infty \left [ e_n^{pi}e_{nm}(v_{0i}^{-1}v_i)(\xi _ i)^{m+1}-\left ( \frac{d\xi _ j}{d\xi _ i}\right ) ^{-1}e_n^{pe}e_{nm}\left ( v_{0e}^{-1}v_j\right ) ( \xi _ j)^{n+1 } \right ] \nonumber \\ & & \times \prod _ { \scriptstyle k=1\atop { \scriptstyle k\neq d , g , h\atop \scriptstyle k \neq p}}^{n_2}\sum_{q , t=-1}^\infty \left [ \sum_{\scriptstyle l=1\atop \scriptstyle l \neq e}^{n_1}e_q^{kf}e_{qt}\left ( v_{0f}^{-1}v_l\right ) b_t^l+\sum_{\scriptstyle l=1\atop \scriptstyle l\neq f}^{n_2}e_q^{kl}e_{qt}(v_{0l}^{-1}v_l)b_t^l\right ] = 0\ .\end{aligned}\ ] ] these are the overlap identities between legs @xmath34 and @xmath50 of the composite vertex @xmath161 . from these overlap equations
it is then possible to derive the form of the composite vertex .
it is given by @xmath299 \times \prod_{r=-1}^1\sum_{\scriptstyle i=1\atop \scriptstyle i\neq e , f}^{n_1+n_2}\sum_{s=-1}^1e_{rs}(v_i)b_s^i \nonumber \\ & & \times \prod_{\scriptstyle i=1\atop \scriptstyle i\neq b , c}^{n_1}\sum_{n , m=-1}^\infty \left [ \sum_{\scriptstyle j=1\atop \scriptstyle j\neq e}^{n_1}e_n^{ij}e_{nm}(v_{0j}^{-1}v_j ) b_m^j + \sum_{\scriptstyle j=1\atop \scriptstyle j\neq f}^{n_2}e_m^{ie}e_{mn}\left ( v_{0e}^{-1}v_j \right ) b_n^j\right ] \nonumber \\ \label{eq4.75 } & & \times \prod_{\scriptstyle i=1\atop \scriptstyle i\neq d , g , h}^{n_2}\sum_{n , m=-1}^\infty \left [ \sum_{\scriptstyle j=1\atop \scriptstyle j\neq e}^{n_1}e_n^{if}e_{nm}\left ( v_{0f}^{-1 } v_j\right ) b_n^j + \sum_{\scriptstyle j=1\atop \scriptstyle j\neq f}^{n_2}e_n^{ij}e_{nm } ( v_{0j}^{-1}v_j)b_m^j\right ] \
.\end{aligned}\ ] ] although this is the correct composite vertex for a general cycling , its ghost structure is not very apparent .
we may use the explicit expression for the vectors @xmath151 @xcite : @xmath300 where @xmath301 and @xmath302 if we assume now that the cycling transformations @xmath303 and @xmath304 do not depend on the variables @xmath62 of vertex @xmath158 or @xmath305 of vertex @xmath247 , and if we have in mind that @xmath306 for any leg @xmath34 of vertex @xmath158 and a leg @xmath50 of vertex @xmath247 ( or vice - versa ) , then we have @xmath307 what simplifies things considerably .
we shall call all cyclings that have such properties `` simple cycling - like '' . in this kind of cyclings ,
each leg has its own ghost attached to it , with the exception of three of the legs which have no ghosts attached to them .
we will now consider three cases separately : one in which none of the legs @xmath5 or @xmath6 have ghosts attached to them , one in which one of these legs ( say @xmath5 ) has a ghost attached to it , and a case where both legs ( @xmath5 and @xmath6 ) have ghosts attached to them .
the composite vertex for the case where neither @xmath5 nor @xmath6 ( we choose @xmath308 and @xmath309 ) have ghosts attached is given by @xmath310 where @xmath311 is the vertex given by the first two terms of equation ( [ eq4.75 ] ) . for the case where leg @xmath5 has a ghost attached to it , but not leg @xmath6 ( we shall call @xmath309 ) , the composite vertex is given by @xmath312 in the last case , where both @xmath5 and @xmath6 have ghosts attached to them , we then have @xmath313 we can see that , for each case , the composite vertex has the correct ghosts number ( @xmath166 ) .
using overlap identities , two vertices were sewn together in order to become a composite vertex .
the calculations have been done with the correct ghost numbers for each vertex and the result has both brst invariance and the correct ghost counting .
this work has been done in the department of mathematics , king s college london , university of london , under the supervision of professor p. c. west , and finished at the university of so paulo ( usp ) .
the author wishes to thank professor west for suggesting , guiding and correcting this work and mrs .
nanci romero for helping with the figures .
my gratitude to capes and cnpq , brazil , for financial support . g.e.d . !
|
it is shown how to sew string vertices with ghosts at tree level in order to produce new tree vertices using the group theoretic approach to string theory .
it is then verified the brst invariance of the sewn vertex and shown that it has the correct ghost number .
pacs number : 11.25 0.5 cm
|
large - scale wave - like perturbations in the solar atmosphere have been observed for many years .
for example , the h@xmath2 moreton wave @xcite , the 10830 wave @xcite , the extreme - ultraviolet ( euv ) wave @xcite and the soft x - ray ( sxr ) wave @xcite . the h@xmath2 moreton wave , which manifests as a propagating dark / white front in the h@xmath2 off - band dopplergrams , has been recognized as a chromospheric surface wave observed immediately following an impulsive flare @xcite .
observations have indicated that the 10830 and sxr waves are consistent with the chromospheric moreton wave and thereby they were interpreted as fast - mode waves and were thought to be the counterparts of the moreton wave at different heights @xcite . for the euv waves ,
significant controversy remains over their physical natures and origins .
so far , there are several competing interpretations for the euv waves , including the fast - mode wave model @xcite , the slow - mode wave model @xcite , and the non - wave models which are related to a current shell or successive restructuring of field lines caused by coronal mass ejections ( cmes ) @xcite . in addition
, a few authors proposed that both the wave and non - wave models should be required to explain the complex euv waves ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
detailed observational characteristics and various theoretical explanations of the euv waves could be found in several recent reviews ( e.g. , * ? ? ?
* ; * ? ? ?
. at present both the wave and non - wave models could not fully explain the observed characteristics of the euv waves .
therefore , investigating the temporal and spatial evolutions of the euv waves using high resolution , multiwavelength observations should be an effective way to clarify their physical natures and origins . in this letter
, we present the observations of an euv wave on 2011 august 9 .
although this event has been reported by @xcite , they only studied the wave in h@xmath2 and 193 observations . in this letter , for the first time we report the wave signatures observed simultaneously in the photosphere , chromosphere , transition region as well as the corona , using the ultraviolet ( uv ) and euv observations taken by the atmospheric imaging assembly ( aia ; * ? ? ? * ) onboard the _ solar dynamics observatory_. we find that the wave signatures at different height of the solar atmosphere were caused by a fast magnetosonic shock wave propagating in the corona .
the full - disk h@xmath2 images are obtained by the solar magnetic activity research telescope ( smart ; * ? ? ?
* ) at hida observatory , kyoto university , japan .
the smart provides full - disk h@xmath2 images in seven channels : h@xmath2 center and six off - bands ( @xmath3 , @xmath4 , and @xmath5 ) .
its cadence is 2 minutes , and the pixel size is @xmath6 .
the aia onboard the _ sdo _ has high time resolution of up to 12 ( 24 ) s for the euv ( uv ) channels , and the images have a pixel resolution of @xmath7 . in this letter , all the aia s uv ( 1700 and 1600 ) and euv ( 94 , 131 , 171 , 193 , 211 , 304 , and 335 )
observations are used to analysis the wave kinematics at different heights of the solar atmosphere .
all images are differentially rotated to a reference time ( 08:05:00 ut ) , and the solar north is up , west to the right .
the wave event on 2011 august 9 was accompanied by a _ goes _ x6.9 flare in noaa ar11263 ( n18w80 ) and a halo cme with a average speed ( acceleration ) of 1610 km s@xmath0 ( -40 m s@xmath1 ) .
the flare started at 07:48 ut and peaked at 08:05 ut , which is the most powerful flare observed so far in the current solar cycle 24 . in this letter
, we mainly investigate the kinematics and the spatial correlation of the wave at different heights of the solar atmosphere .
figure [ fig1 ] shows the morphological evolution of the wave on the 1600 and 1700 , and h@xmath2 center base - difference images , in which the thick curves outline the wavefronts at different times .
the wavefront determined from the 1600 image at 08:03:53 ut is overlaid on the h@xmath2 image at 08:04:03 ut ( white curve in figure [ fig1](e ) ) .
one can see that the wavefronts at the two moments showed a similar shape , which indicates the synchronous evolution of the wave in different atmosphere layers .
the wavefront seeing on the 1700 images was relatively weak .
the end times of the wave in the 1700 , h@xmath2 , and 1600 observations were at 08:04:55 , 08:10:03 , and 08:06:41 ut , respectively . assuming the wave started at 08:02:00 ut , the corresponding lifetimes of the wave at different lines should be 175 ( 1700 ) , 483 ( h@xmath2 ) , and 281 s ( 1600 ) ( see animations 1 3 ) . the wave in euv observations is shown in figure [ fig2 ] , using the 211 base - difference images .
it was first observed as a semicircular sharp emission at 08:02:48 ut , and the position of the wavefront coincided well with those observed at h@xmath2 and uv lines ( see the overlaid dotted and dashed curves in figure [ fig2](a )
( c ) and animation 4 ) .
moreover , a dome - like structure , which extended from the sharp bright wavefront , can be observed off the disk limb ( see the white arrow in figure [ fig2](c ) ) , which is thought to be a shock wave traveling in the corona ( also see , * ? ? ?
therefore , the sharp bright wave could be considered as the intersection of the shock wave with the corona .
after 08:07:12 ut , the sharp bright wavefront became more and more diffuse ( see figure [ fig2](d ) ( e ) ) .
it is interesting that a dimming region was observed behind the bright wavefront , which did not expand to a large distance as the bright wavefront did .
this result supports the scenario that the dimming region maps the cme footprint on the solar surface , while the euv wave is a shock wave driven by the associated cme @xcite . to reveal the magnetic topology of the coronal condition where the wave propagated , we extrapolate the three - dimensional coronal magnetic field using the potential field source surface ( pfss ; *
* ) model based on the helioseismic and magnetic imager ( hmi ; * ? ? ?
* ) magnetograms .
the extrapolated field lines are overlaid in figure [ fig2](d ) , from which we find that a remote region with open fields was situated on the propagation path of the wave ( see the red lines in figure [ fig2](d ) ) .
the measurements of the wave kinematics along cuts c1 c4 are shown in figure [ fig3 ] , using the time - distance diagrams obtained from the 211 and 193 running difference images . in each time - distance diagram , the propagating wave could be identified as a bright stripe of positive slope
. a linear fit to it yields the wave speed in the plane of the sky .
obviously , the wave propagated with different speeds along different cuts ( 428 756 km s@xmath0 , see figure [ fig3 ] ) . it is interesting that another wave stripe of negative slope was observed in the time - distance diagrams obtained from cuts c2 and c3 . as it can be seen in figure [ fig2](d ) ,
cuts c2 and c3 are passing through the region with open magnetic fields .
this suggests that the stripe of negative slope represents the reflected wave from the open fields region .
the speed of the reflected wave was 300 452 km s@xmath0 , which indicates that the primary wave has been decelerated significantly when it reached the open fields region .
more importantly , the reflection effect manifested the wave nature of the primary euv wave .
to compare the wave kinematics at different heights of the solar atmosphere , the time - distance diagrams ( along cut c2 ) obtained from the running difference images of all the uv and euv channels are shown in figure [ fig4 ] . in the 1700 and 1600 time - distance diagrams ,
the propagating wave is identified as a faint stripe that could only be traced to about 200 mm from the flare kernel ( see the black arrows in figure [ fig4](a ) and ( c ) ) . the speeds ( accelerations ) of the wave measured from the 1700 and 1600 diagrams are 967 km s@xmath0 ( -485 m s@xmath1 ) and 893 km s@xmath0 ( -334 m s@xmath1 ) , respectively . in the time - distance diagrams obtained from the euv observations ,
the wave can be traced to about 400 mm from the flare kernel , and the wave speed was 398 656 km s@xmath0 , which averaged at 605 km s@xmath0 . due to the influence of the flare diffraction , we do not take the the speeds measured from 193 , 94 , and 131 observations into account . to compare the wave speed in the euv and uv observations ,
we calculate the wave speed in the euv observations within a distance of 100 200 mm from the flare kernel , and find that the wave propagated at a speed of 967 1049 km s@xmath0 ( averaged at 999 km s@xmath0 ) within this distance , similar to the wave speed measured from the 1600 and 1700 observations .
this result indicates that the wave signature observed at different heights of the solar atmosphere resulted from the same physical origin and evolved synchronously .
the acceleration of the wave measured from the euv observations was -334
-520 m s@xmath1 ( averaged at -424 m s@xmath1 ) . meanwhile , the average acceleration of the wave during the initial stage was -533 m s@xmath1 , larger than that measured within the whole lifetime of the wave , which suggests the rapidly deceleration of the wave during its initial stage .
it is important to note that a slower wave is observed in the euv time - distance diagrams ( see figure [ fig4](b ) , ( d ) , ( g ) and ( h ) ) .
it propagated at a speed of 211 231 km s@xmath0 and with a small acceleration of -68 -146 m s@xmath1 .
the average speed of the slower wave was 220 km s@xmath0 , about three times smaller than that of the primary fast wave .
we think that this slower wave was probably caused by the reconfiguration of the low coronal fields due to the eruption of the associated cme , as it has been predicted in the field - line stretching model @xcite .
all the wavefronts determined from the 1700 ( yellow ) , 1600 ( purple ) , and 211 ( green ) and h@xmath2 ( red ) observations are plotted in figure [ fig5](a ) .
one can see that all of them showed a semicircular shape and propagated synchronously on the solar surface , and the wavefronts determined on 1700 , 1600 , and h@xmath2 images coincided well with the sharp bright wavefront observed in the 211 observations .
on the other hand , the wave stripes in the time - distance diagrams along c2 at different wavelengths are all plotted in figure [ fig5](b ) .
it can be seen that the primary fast wave has a similar speed along cut c2 in all uv and euv channels , which suggests a common mechanism of the wave in the different layers of the solar atmosphere .
the perturbation profiles of the wavefront at different times are plotted in figure [ fig5](c ) , which well reflected the evolution of the wave .
it can be seen that the steepness of the perturbation profile weakened quickly with increasing time , while the amplitude first showed an increase and then decreased significantly within several minutes .
the amplitude reached the highest value of 3.96 at 08:03:36 ut and dropped to 1.75 at 08:08:48 ut . assuming the intensity enhancement is primarily due to plasma compression and the wave propagation was perpendicular to the magnetic fields
, we can obtain the magnetosonic mach number of the wave @xcite , which is 1.85 ( 1.25 ) at 08:03:36(08:08:48 ) ut .
on the other hand , the width of the perturbation profile of the wavefront broadened a lot with increasing time .
all these results including the deceleration of the magnetosonic mach number , amplitude decrease , and wavefront broadening are consistent with the shock wave scenario @xcite .
in addition , the evidence for the appearance of the shock wave was also observed as a type - ii radio burst in the metric radio spectrogram , which was observed from 08:02:40 to 08:06:30 ut and had a derived speed of 850 km s@xmath0 ( see ; * ? ? ? * for details ) . the start time and
the speed of the type - ii radio burst were in agreement with those observed in the uv and euv observations .
for the first time , we report an euv wave that was observed simultaneously in the different layers of the solar atmosphere . with the high temporal and spatial resolution h@xmath2 , uv and euv observations taken by the smart and the _ sdo_/aia , we investigate the kinematics and the spatial correlation of the wave at different heights of the solar atmosphere .
the main results are summarized as follows . 1 .
the wave could be observed simultaneously in the photosphere , chromosphere , transition region and the low corona .
the lifetime of the wave determined from the 1700 , 1600 , and h@xmath2 observations were 175 , 483 , and 281 s , respectively .
the wave signatures observed in the photosphere , chromosphere , transition region and the sharp bright wavefront observed in the euv observations evolved synchronously in space and time , which suggests that these waves signature at different layers of the solar atmosphere had a common origin and the same physical mechanism .
2 . in the euv observations ,
the average speed ( acceleration ) of the wave was 605 km s@xmath0 ( -424 m s@xmath1 ) . during the initial stage ,
the wave had a similar speed of about 1000 km s@xmath0 at all uv , euv , h@xmath2 , and radio wavelength bands . moreover , the wave speed during the initial stage decreased faster than that during the whole wave lifetime .
the wave kept propagating after the following dimming region has stopped to expand .
this phenomenon suggests that the wave was driven by the associated cme @xcite and thereby rules out the pseudo - wave models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in addition , the wave nature of the euv wave was also manifested by the reflection of the euv wave from the remote open fields region and the type - ii radio burst observed in the metric radio spectrogram .
a slower wave with a speed ( acceleration ) of 220 km s@xmath0 ( -93 m s@xmath1 ) was observed behind the primary fast wave , which was probably resulted from the reconfiguration of the low coronal fields caused by the associated cme @xcite .
5 . during the initial stage ,
the steepness of the wavefront s perturbation profile weakened quickly and the width broadened significantly , and the amplitude first increased a lot and then decreased significantly within several minutes .
the highest mach number of the wave is 1.85 , which quickly decreased to a mach number of unity .
all these results are consistent with the physical properties of a fast magnetosonic shock wave @xcite .
in addition , the evidence for appearance of the shock wave was also identified , such as the dome - like structure observed in the euv observations and the type - ii radio burst in the metric radio spectrogram .
base on our analysis results , we conclude that the euv wave analyzed in this letter should be a fast magnetosonic shock wave , which was firstly driven by the associated cme and then propagated freely in the corona .
the wave signatures observed in the photosphere , chromosphere , transition region layers and the sharp bright wavefront observed in the corona were the intersections of the coronal shock wave with these lower layers of the atmosphere . as the shock propagates ,
its energy is gradually dissipated into the ambient plasma medium by either plasma oscillations ( that are subsequently damped ) or plasma microinstabilities @xcite .
this energy dissipation and the expansion of the shock front will cause the decrease the perturbation amplitude , and consequently its speed and mach number .
this also explains why the deceleration rate decreases with increasing time and distance .
eventually , the shock will decay to an ordinary fast - mode wave with mach number of 1 @xcite . in the field - line stretching model @xcite , as the cme flux rope rises , a piston - driven shock wave is formed preceding the envelope of the expanding cme , which sweeps the solar surface at a super - alfvnic speed .
this mechanism implies that the legs of the shock wave would produce the surface wave phenomena in the solar lower atmosphere layers . in the meantime , a slower wave - like feature at a speed of about three times smaller than that of the shock wave could be identified behind the fast shock wave .
this structure is thought to be produced by the successive stretching of closed field lines during the launch of the cme .
this model also predicts the generation of type - ii radio bursts from the top of the shock wave .
our observational results are not only consistent with this field - line stretching model , but also indicate that the legs of the shock wave can extend downward into the photosphere .
further investigations involving high temporal and spatial resolution observations of uv , euv , and sxr would be helpful to fully understand the physical natures of the waves in the solar atmosphere .
is a mission for nasa s living with a star ( lws ) program .
we thank the smart for data support .
the authors thank an anonymous referee for many constructive suggestions .
this work is supported by the natural science foundation of china under grants 10933003 , 11078004 , and 11073050 , and the national key research science foundation ( 2011cb811400 ) .
asai , a. , ishii , t. , isobe , h. , et al .
2012 , , 745 , l18 athay , r. g. , & moreton , g. e. 1961 , , 133 , 935 attrill , g. d. r. 2010 , , 718 , 494 attrill , g. d. r. , harra , l. k. , van driel - gesztelyi , l. , & dmoulin , p. 2007
, , 656 , l101 balasubramaniam , k. s. , cliver , e. w. , pevtsov , a. , et al .
2010 , , 723 , 601 balasubramaniam , k. s. , pevtsov , a. a. , & neidig , d. f. 2007 , , 658 , 1372 chen , p. f. , fang , c. , & shibata , k. 2005 , , 622 , 1202 chen , p. f. , & wu , y. 2011 , , 732 , l20 chen , p. f. , wu , s. t. , shibata , k. , & fang , c. 2002 , , 572 , l99 cohen , o. , attrill , g. d. r. , manchester iv , w. b. , & wills - davey , m. j. 2009 , , 705 , 587 delanne , c. 2000 , , 545 , 512 delanne , c. , hochedez , j .- f . , & aulanier , g. 2007 , , 465 , 603 delanne , c. , trk , t. , aulanier , g. , & hochedez , j .- f .
2008 , , 247 , 123 downs , c. , roussev , i. i. , van der holst , b. , et al .
2011 , , 728 , 2 gallagher , p. t. , & long , d. m. 2011 , , 158 , 365 gilbert , h. r. , daou , a. g. , young , d. , tripathi , d. , & alexander , d. 2008 , , 685 , 629 gilbert , h. r. , & holzer , t. e. 2004a , , 610 , 572 gilbert , h. r. , holzer , t. e. , thompson , b. j. , & burkepile , j. t. 2004b , , 607 , 540 gopalswamy , n. , yashiro , s. , temmer , m. , et al .
2009 , , 691 , l123 hudson , h. s. , khan , j. i. , lemen , j. r. , nitta , n. v. , & uchida , y. 2003 , , 212 , 121 khan , j. i. , & aurass , h. 2002 , , 383 , 1018 lemen , j. r. , title , a. m. , akin , d. j. , et al .
2012 , , 275 , 17 liu , w. , nitta , n. v. , schrijver , c. j. , title , a. m. , & tarbell , t. d. 2010 , , 723 , l53 long , d. m. , gallagher , p. t. , james mcateer , r. t. , & bloomfield , s. 2008 , , 680 , l81 moreton , g. e. 1960 , , 65 , 494 muhr , n. , vrnak , b. , temmer , m. , veronig , a. m. , & magdaleni , j. 2010 , , 708 , 1639 narukage , n. , hudson , h. s. , morimoto , t. , et al .
2002 , , 572 , l109 narukage , n. , ishii , t. t. , nagata , s. , et al .
2008 , , 684 , l45 narukage , n. , morimoto , t. , kadota , m. , et al .
2004 , , 56 , l5 ofman , l. , & thompson , b. j. 2002 , , 574 , 440 patsourakos , s. , & vourlidas , a. 2009 , , 700 , l182 priest , e. r. 1982 , solar magneto - hydrodynamics ( dordrecht : reidel ) schmidt , j. m. , & ofman , l. 2010 , , 713 , 1008 schou , j. , borrero , j. m. , norton , a. a. , et al .
2012 , , 275 , 327 schrijver , c. j. , & de rosa , m. l. 2003 , , 212 , 165 temmer , m. , veronig , a. m. , gopalswamy , n. , & yashiro , s. 2011 , , 273 , 421 thompson , b. j. , plunkett , s. p. , gurman , j. b. , et al .
1998 , geophys .
25 , 2465 uchida , y. 1968 , , 4 , 30 ueno , s. , nagata , s. , kitai , r. , kurokawa , h. , & ichimoto , k. 2004 , , 5492 , 958 veronig , a. m. , gmry , p. , kienreich , i. w. , et al .
2011 , , 743 , l10 veronig , a. m. , muhr , n. , kienreich , i. w. , temmer , m. , & vrnak , b. 2010 , , 716 , l57 vrnak , b. , warmuth , a. , braja , r. , & hanslmeier , a. 2002 , , 394 , 299 wang , y .- m .
2000 , , 543 , l89 warmuth , a. 2010 , adv .
space res .
, 45 , 527 warmuth , a. , mann , g. , & aurass , h. 2005 , , 626 , l121 warmuth , a. , vrnak , b. , aurass , h. , & hanslmeier , a. 2001 , , 560 , l105 warmuth , a. , vrnak , b. , magdaleni , j. , hanslmeier , a. , & otruba , w. 2004 , , 418 , 1117 wills - davey , m. j. , deforest , c. e. , & stenflo , j. o. 2007 , , 664 , 556 wu , s. t. , zheng , h. , wang , s. , et al .
2001 , , 106 , 25089 zhukov , a. n. , & auchre , f. 2004 , , 427 , 705
|
for the first time , we report a large - scale wave that was observed simultaneously in the photosphere , chromosphere , transition region and low corona layers of the solar atmosphere . using the high temporal and high spatial resolution observations taken by the solar magnetic activity research telescope at hida observatory and the atmospheric imaging assembly ( aia ) onboard _ solar dynamic observatory _ , we find that the wave evolved synchronously at different heights of the solar atmosphere , and it propagated at a speed of 605 km s@xmath0 and showed a significant deceleration ( -424 m s@xmath1 ) in the extreme - ultraviolet ( euv ) observations . during the initial stage ,
the wave speed in the euv observations was 1000 km s@xmath0 , similar to those measured from the aia 1700 ( 967 km s@xmath0 ) and 1600 ( 893 km s@xmath0 ) observations .
the wave was reflected by a remote region with open fields , and a slower wave - like feature at a speed of 220 km s@xmath0 was also identified following the primary fast wave .
in addition , a type - ii radio burst was observed to be associated with the wave .
we conclude that this wave should be a fast magnetosonic shock wave , which was firstly driven by the associated coronal mass ejection and then propagated freely in the corona . as the shock wave propagated , its legs swept the solar surface and thereby resulted in the wave signatures observed in the lower layers of the solar atmosphere .
the slower wave - like structure following the primary wave was probably caused by the reconfiguration of the low coronal magnetic fields , as predicted in the field - line stretching model .
|
the chromospherically active binaries are detached binary systems with cool components characterized by strong chromospheric , transition region , and coronal activity .
the rs cvn systems have at least one cool evolved component whereas both components of the by dra binaries are main sequence stars ( fekel et al .
1986 ) . in this series of papers we try to study the chromosphere of this kind of extremely active stars using the information provided by several optical spectroscopic features that could be used as chromospheric activity indicators .
the simultaneous observations of different lines , that are formed at different height in the chromosphere ( from the region of temperature minimum to the higher chromosphere ) , are of special interest for stellar activity studies since they provide very useful information about this stellar region .
ideally , simultaneous observations should be performed at all wavelengths in order to develop a coherent 3-d atmosphere model . in practice ,
simultaneous observations of several activity indicators are rare and tend to focus on the same small number of extremely active systems .
the best way to obtain the active - chromosphere contribution to some spectral line in the chromospherically active binaries is to subtract the underlying photospheric contribution using the spectral subtraction technique ( subtraction of a synthesized stellar spectrum constructed from artificially rotationally broadened , radial - velocity shifted , and weighted spectra of inactive stars chosen to match the spectral types and luminosity classes of both components of the active system under consideration ) the emissions in the ca ii h & k resonance lines are the most widely used optical indicators of chromospheric activity , since their source functions are collisionally controlled and represent an extremely important cooling mechanism . in chromospheric active binaries the subtraction of the photospheric flux in this spectral region has been recently applied using the spectral subtraction ( see montes et al .
1995c , 1996a and references therein ) .
the h@xmath0 line is also an important chromospheric activity indicator , but it is only in emission above the continuum in very active stars , and in less active star only a filled - in absorption line is observed .
so , to infer chromospheric activity level the spectral subtraction is needed ( see montes et al . 1994 ; 1995a ; b ; d , and references therein ; lzaro & arvalo 1996 ) .
a similar behaviour is observed in the other balmer lines ( hall & ramsey 1992 ; montes et al .
1995d ) .
recently , the spectral subtraction technique has been used in other lines as the ca ii irt , mg i b , na i d@xmath1 , d@xmath2 , and he i d@xmath3 lines ( gunn & doyle 1996 ; gunn et al .
the ca ii irt lines are formed deeper in the atmosphere and are thus sensitive probes of the temperature minimum region .
the na i d@xmath1 , d@xmath2 lines are collision dominated and are good indicators of changes in the lower chromosphere .
the mg i b triplet lines are formed in the lower chromosphere and the region of temperature minimum and they are good diagnostics of photospheric activity ( basri et al .
the he i d@xmath3 line has been largely ignored as activity indicator ; however it could be a valuable probe of stellar activity and the observation of this line in emission supports the detection of flare like events ( zirin 1988 ) . in this first paper
we focus our study on the analysis of the extensively used h@xmath0 chromospheric activity indicator together with simultaneous observations of the less studied he i d@xmath3 and na i d@xmath1 , d@xmath2 spectral features in a sample of 18 northern active binary systems selected from `` a catalog of chromospherically active binary stars ( second edition ) '' ( strassmeier et al .
1993 , hereafter cabs ) . by using the spectral subtraction technique ,
we have determined the excess emission in these lines and we have computed absolute chromospheric fluxes in h@xmath0 .
the primary aim of this study is analyse in detail the excess h@xmath0 emission and to study the subtracted h@xmath0 line profile , especially in some extremely active stars which exhibit broad wings .
moreover , we try to understand the behaviour of the he i d@xmath3 and na i d@xmath1 , d@xmath2 lines as chromospheric activity indicators taking into account the advantage that we simultaneously know the behaviour of the chromospheric excess h@xmath0 emission in these systems . in forthcoming papers we will analyze in detail several optical spectroscopic features using echelle spectroscopy in order to determine the effects of stellar activity on spectral lines originating at different heights in the chromosphere .
another of our goals is to obtain information about the presence of extended matter ( prominence - like structures ) in the chromospheric active binaries using simultaneous h@xmath0 and h@xmath4 observations at near - eclipse orbital phases . in sect .
2 we give the details of our observations and data reduction . in sect .
3 we describe the individual results of h@xmath0 , na i d@xmath1 , d@xmath2 , and he i d@xmath3 line observations of our sample . finally in sect . 4
we discuss our results .
lcccccccccccc name & & & & & + & date & ut & @xmath5 & s / n & & date & ut & @xmath5 & s / n & & h@xmath0 & ca ii + bd cet & 1995/09/15 & 00:43 & 0.569 & 330 & & 1995/09/15 & 01:03 & 0.569 & 361 & & - & 95c,96a + ay cet & 1995/09/15 & 02:03 & 0.797 & 385 & & 1995/09/15 & 02:20 & 0.797 & 398 & & - & 95c,96a + ar psc & 1995/09/13 & 01:48 & 0.373 & 387 & & 1995/09/13 & 01:27 & 0.372 & 141 & & 94,95a & 94,96a + `` & 1995/09/14 & 01:42 & 0.443 & 361 & & 1995/09/14 & 02:00 & 0.443 & 369 & & - & - + '' & 1995/09/15 & 03:55 & 0.519 & 392 & & 1995/09/15 & 04:23 & 0.520 & 399 & & - & - + `` & 1995/09/15 & 05:34 & 0.524 & 305 & & 1995/09/15 & 05:28 & 0.523 & 359 & & - & - + hd 12545 & 1995/09/15 & 02:59 & 0.401 & 354 & & 1995/09/15 & 02:25 & 0.400 & 361 & & - & 95c,,96a + ux ari & 1995/09/13 & 02:08 & 0.419 & 346 & & & & & & & 95b & 95c,96a + '' & 1995/09/13 & 05:08 & 0.438 & 202 & & 1995/09/13 & 04:52 & 0.437 & 494 & & - & - + `` & 1995/09/14 & 02:24 & 0.576 & 276 & & 1995/09/14 & 02:09 & 0.574 & 293 & & - & - + '' & 1995/09/15 & 03:09 & 0.736 & 352 & & 1995/09/15 & 03:32 & 0.739 & 410 & & - & - + v711 tau & 1995/09/13 & 05:15 & 0.922 & 360 & & 1995/09/13 & 05:36 & 0.927 & 370 & & 94,95a & 94,96a + `` & 1995/09/14 & 04:19 & 0.261 & 374 & & 1995/09/14 & 04:35 & 0.265 & 423 & & - & - + '' & 1995/09/14 & 05:35 & 0.280 & 332 & & 1995/09/14 & 05:50 & 0.283 & 386 & & - & - + `` & 1995/09/15 & 03:50 & 0.606 & 342 & & 1995/09/15 & 03:44 & 0.605 & 332 & & - & - + '' & 1995/09/15 & 06:11 & 0.641 & 314 & & 1995/09/15 & 06:05 & 0.639 & 326 & & - & - + v833 tau & 1995/09/13 & 06:13 & 0.762 & 184 & & 1995/09/15 & 05:47 & 0.752 & 325 & & 95b & - + `` & 1995/09/14 & 06:07 & 0.319 & 424 & & 1995/09/14 & 05:55 & 0.314 & 417 & & - & - + '' & 1995/09/15 & 04:56 & 0.851 & 394 & & 1995/09/15 & 05:18 & 0.859 & 322 & & - & - + v1149 ori & 1995/09/15 & 06:17 & 0.439 & 311 & & 1995/09/15 & 06:29 & 0.439 & 335 & & 95b & 95c,96a + mm her & 1995/09/12 & 20:56 & 0.498 & 274 & & 1995/09/12 & 21:31 & 0.501 & 329 & & 94,95a & 94,96a + `` & 1995/09/13 & 22:06 & 0.630 & 332 & & 1995/09/13 & 20:45 & 0.623 & 384 & & - & - + '' & 1995/09/14 & 20:09 & 0.745 & 257 & & 1995/09/14 & 20:41 & 0.748 & 314 & & - & - + v815 her & 1995/09/12 & 20:43 & 0.978 & 360 & & 1995/09/12 & 20:21 & 0.970 & 341 & & 94,95a & 94,96a + `` & 1995/09/13 & 20:15 & 0.520 & 375 & & 1995/09/13 & 20:22 & 0.523 & 341 & & - & - + '' & 1995/09/14 & 21:24 & 0.099 & 354 & & 1995/09/14 & 21:44 & 0.107 & 394 & & - & - + by dra & 1995/09/13 & 22:58 & 0.684 & 387 & & 1995/09/13 & 23:15 & 0.686 & 372 & & - & 94,96a + `` & 1995/09/14 & 21:56 & 0.839 & 398 & & 1995/09/14 & 21:05 & 0.838 & 404 & & - & - + v775 her & 1995/09/14 & 22:03 & 0.394 & 342 & & 1995/09/14 & 21:53 & 0.392 & 373 & & 94,95a & 94,96a + v478 lyr & 1995/09/14 & 22:14 & 0.953 & 326 & & 1995/09/14 & 22:33 & 0.959 & 412 & & - & 94,96a + hk lac & 1995/09/12 & 22:56 & 0.067 & 343 & & 1995/09/12 & 22:42 & 0.067 & 357 & & 94,95a & 94,96a + '' & 1995/09/13 & 00:42 & 0.070 & 130 & & 1995/09/13 & 00:25 & 0.070 & 193 & & - & - + `` & 1995/09/13 & 23:44 & 0.110 & 346 & & 1995/09/13 & 23:58 & 0.110 & 343 & & - & - + '' & 1995/09/14 & 01:37 & 0.113 & 332 & & 1995/09/14 & 01:21 & 0.113 & 264 & & - & - + `` & 1995/09/14 & 22:53 & 0.149 & 321 & & 1995/09/14 & 23:09 & 0.149 & 339 & & - & - + '' & 1995/09/15 & 01:26 & 0.153 & 329 & & 1995/09/15 & 01:21 & 0.153 & 358 & & - & - + ar lac & 1995/09/12 & 22:16 & 0.405 & 360 & & 1995/09/12 & 21:57 & 0.399 & 421 & & 94,95a & 94,96a + `` & 1995/09/13 & 23:39 & 0.939 & 346 & & 1995/09/13 & 23:23 & 0.933 & 415 & & - & - + '' & 1995/09/14 & 22:49 & 0.425 & 339 & & 1995/09/14 & 22:42 & 0.423 & 407 & & - & - + kz and & 1995/09/14 & 23:29 & 0.145 & 336 & & 1995/09/14 & 23:51 & 0.150 & 367 & & - & 94,95c,96a + kt peg & 1995/09/13 & 00:07 & 0.693 & 221 & & 1995/09/13 & 00:13 & 0.694 & 271 & & - & 95c,96a + `` & 1995/09/15 & 01:19 & 0.024 & 313 & & 1995/09/15 & 01:13 & 0.023 & 368 & & - & - + ii peg & 1995/09/12 & 23.02 & 0.575 & 300 & & 1995/09/12 & 23:23 & 0.577 & 298 & & - & - + '' & 1995/09/13 & 00:58 & 0.587 & 138 & & 1995/09/13 & 01:18 & 0.589 & 240 & & - & - + `` & 1995/09/14 & 00:57 & 0.735 & 265 & & 1995/09/14 & 01:14 & 0.737 & 324 & & - & - + '' & 1995/09/14 & 03:07 & 0.749 & 315 & & 1995/09/14 & 02:51 & 0.747 & 328 & & - & - + `` & 1995/09/14 & 05:00 & 0.760 & 266 & & 1995/09/14 & 05:16 & 0.762 & 263 & & - & - + '' & 1995/09/14 & 23:22 & 0.874 & 318 & & 1995/09/14 & 23:15 & 0.873 & 324 & & - & - + `` & 1995/09/15 & 01:56 & 0.890 & 332 & & 1995/09/15 & 01:49 & 0.889 & 352 & & - & - + '' & 1995/09/15 & 04:38 & 0.907 & 298 & & 1995/09/15 & 04:31 & 0.906 & 311 & & - & - + l l c
c c c c l l c hd & name & t@xmath6 & sb & r & d & v - r & p@xmath7 & p@xmath8 & vsin_i _ + & & & & ( r@xmath9 ) & ( pc ) & & ( days ) & ( days ) & ( km s@xmath10 ) + 1833 & bd cet & k1iii & 1 & @xmath1110 & 71 & 0.81 & 35.1 & 34.46 & 15 + 7672 & ay cet & wd / g5iii & 1 & 0.012/15 & 66.7 & 0.69 & 56.824 & 77.22 & 4 + 8357 & ar psc & g7v / k1iv@xmath12 & 2 & /@xmath111.5 & 17 & 0.74@xmath12 & 14.30226@xmath12 & 12.245 & 6.5/3.5@xmath12 + 12545 & xx tri & k0iii & 1 & @xmath118 & 310 & [ 0.62 ] & 23.9824 & 24.3 & 17 + 21242 & ux ari & g5v / k0iv & 2 & 0.93/@xmath114.7 & 50 & 0.70/0.54 & 6.43791 & @xmath13p@xmath7 & 6/37 + 22468 & v711 tau & g5iv / k1iv & 2 & 1.3/3.9 & 36 & 0.62/0.75 & 2.83774 & 2.841 & 13/38 + 283750 & v833 tau & dk5e & 1 & @xmath110.22 & 16.7 & 0.69 & 1.7878 & 1.797 & 6.3 + 37824 & v1149 ori & k1iii & 1 & @xmath1111 & [ 164 ] & 0.90 & 53.58 & 54.1 & 11 + 341475 & mm her & g2/k0iv & 2 & 1.58/2.83 & 190 & [ /0.64 ] & 7.960322 & 7.936 & 10/18 + 166181 & v815 her & g5v/[m1 - 2v ] & 1 & 0.93:/ & 31 & 0.54 & 1.8098368 & 1.8 & 27/ + 234677 & by dra & k4v / k7.5v & 2 & 1.2 - 1.4/ & 15.6 & 1.10 & 5.975112 & 3.827 & 8.0/7.4 + 175742 & v775 her & k0v/[k5-m2v ] & 1 & 0.85/ & 24 & 0.80 & 2.879395 & 2.898 & 15/ + 178450 & v478 lyr & g8v/[dk - dm ] & 1 & @xmath110.9 & 26 & 0.65 & 2.130514 & 2.185 & 21/ + 209813 & hk lac & f1v / k0iii & 1 & - & 150 & 0.75 & 24.4284 & 24.4284 & /15 + 210334 & ar lac & g2iv / k0iv & 2 & 1.8/3.1 & 47 & 0.77 & 1.98322195 & 1.98322195 & 46/81 + 218738 & kz and & dk2/dk2 & 2 & @xmath110.74/ & [ @xmath1323 ] & [ 0.74/0.74 ] & 3.032867 & 3.03 & 12.3/11.6 + 222317 & kt peg & g5v / k6v & 2 & 0.93/0.72 & 25 & [ 0.54/ ] & 6.20199 & 6.092 & 8/5 + 224085 & ii peg & k2 - 3v - iv & 1 & 2.2 & 29.4 & 0.89 & 6.724183 & 6.718 & 21 + @xmath12 parameters from fekel ( 1996 )
observations in the h@xmath0 and na i d@xmath1 , d@xmath2 , he i d@xmath3 line regions have been obtained during three nights ( 1995 september 13 - 15 ) with the isaac newton telescope ( int ) at the observatorio del roque de los muchachos ( la palma , spain ) using the intermediate dispersion spectrograph ( ids ) with grating h1800v , camera 500 and a 1024 x 1024 pixel tek3 ccd as detector .
the reciprocal dispersion achieved is 0.24 / pixel which yields a spectral resolution of 0.48 @xmath14 and a useful wavelength range of 250 @xmath14 centered at 6563@xmath14 ( h@xmath0 ) and 5876@xmath14 ( hei d@xmath3 ) respectively .
the spectra have been extracted using the standard reduction procedures in the iraf package ( bias subtraction , flat - field division , and optimal extraction of the spectra ) .
the wavelength calibration was obtained by taking spectra of a cu - ar lamp . finally , the spectra have been normalized by a polynomial fit to the observed continuum . in table
[ tab : obslog ] we give the observing log . for each star we list the date ,
ut , orbital phase ( @xmath5 ) and signal to noise ratio ( s / n ) obtained for each observation in both spectral regions .
where appropriate , we also give the reference of our previous observation of these systems in the h@xmath0 and ca ii h & k lines . in table
[ tab : par ] we show the hd number , name and the adopted stellar parameters ( from cabs or the references given in the table ) for the 18 chromospherically active binary systems selected .
we have obtained the chromospheric contribution in h@xmath0 na i d@xmath1 , d@xmath2 , and he i d@xmath3 lines using the spectral subtraction technique described in detail by montes et al .
( 1995a , c ) .
the synthesized spectra were constructed using artificially rotationally broadened , radial - velocity shifted , and weighted spectra of inactive stars chosen to match the spectral types and luminosity classes of both components of the active system under consideration .
the reference stars used have been observed in this campaign and previous observational seasons with similar spectral resolution ( see the spectral library of montes et al .
1997 ) . in some case , the difference spectrum obtained appears noisier than expected from the observation s / n ratio ( @xmath13 300 ) due to small differences in spectral type between active and reference star , or to non appropriate evaluation of the rotational broadening and/or of the doppler shift .
in addition , in some spectra telluric lines also appear in the difference spectrum .
this noise in the the difference spectrum have been evaluated as the mean standard deviation ( @xmath15 ) in the regions outside the chromospheric features .
we have obtained values of @xmath15 in the range 0.01 - 0.03 which could be important in low active star but in the more active stars the errors in the excess h@xmath0 ew are small .
we have considered as a clear detection of excess emission or absorption in h@xmath0 , na i d@xmath1 , d@xmath2 , and he i d@xmath3 only when these features in the difference spectrum are larger than 3 @xmath15 . table [ tab : measures ] gives the h@xmath0 line parameters , measured in the observed and subtracted spectra of the sample .
column ( 2 ) of this table gives the orbital phase ( @xmath5 ) for each spectrum , and in column ( 3 ) , h and c mean emission belonging to hot and cool component respectively , and t means that at these phases the spectral features can not be deblended . column ( 4 ) gives the contributions for the hot and cool component to the total continuum ( s@xmath16 and s@xmath17 ) . column ( 5 )
describes the observed h@xmath0 profile , i.e. if the line is in absorption ( a ) in emission ( e ) or totally filled by emission ( f ) .
columns ( 6 ) , ( 7 ) , ( 8) give the following parameters measured in the observed spectrum : the full width at half maximum ( w@xmath18 ) ; the residual intensity , r@xmath19 ; and the h@xmath0 core flux , f(1.7 ) , measured as the residual area below the central 1.7 @xmath14 passband .
the last four columns give he following parameters measured in the subtracted spectrum : the full width at half maximum ( w@xmath20 ) , the peak emission intensity ( i ) , the excess h@xmath0 emission equivalent width ( ew ( h@xmath0 ) ) , and absolute fluxes at the stellar surface logf@xmath21(h@xmath0 ) obtained with the calibration of pasquini & pallavicini ( 1991 ) as a function of ( v - r ) , very similar values of f@xmath21(h@xmath0 ) are obtained using the more recently calibration of hall ( 1996 ) as a function of ( v - r ) and ( b - v ) . for a more detailed description of the parameters given in this table see our previous study of the excess h@xmath0 emission in active binaries ( montes et al .
1995a ) .
in table [ tab : measures_nb ] we list the parameters ( i , fwhm , ew ) of the broad and narrow components used in the two gaussian components fit to the h@xmath0 subtracted emission profile , which we have performed in the stars that present broad wings .
see the comments for each individual star in sect .
3 and the interpretation of these components given in sect . 4
in the following we describe the h@xmath0 , na i d@xmath1 , d@xmath2 , and he i d@xmath3 spectra of the stars of this sample .
the line profiles of each chromospherically active binary system are displayed in fig .
1 to 19 the name of the star , the orbital phase , and the expected positions of the features for the hot ( h ) and cool ( c ) components are given in each figure .
for each system we plot the observed spectrum ( solid - line ) , the synthesized spectrum ( dashed - line ) , the subtracted spectrum , additively offset for better display ( dotted line ) and the gaussian fit to the subtracted spectrum ( dotted - dashed line ) .
single - lined spectroscopic binary classified as k1iii + f by bidelman & macconnell ( 1973 ) .
it presents strong ca ii h & k emission lines centered at the absorption line ( montes et al .
1995c ) and the h@xmath0 line as moderate absorption ( fekel et al . 1986 ) .
we have obtained one spectrum of this system at the orbital phase 0.569 ( see fig . [
fig : plot_bdcet_hahe ] ) .
the h@xmath0 subtracted spectrum shows a weak excess emission . in the na
i line region no detectable filling - in of the d@xmath1 and d@xmath2 lines is present .
a clear absorption in the he i d@xmath3 line appears in the subtracted spectrum . in both spectral regions the spectrum
is matched using a k2iii as reference star .
ay cet is a single - lined binary composed of a spotted g5iii primary and a white dwarf secondary .
it presents strong ca ii h & k emission lines ( montes et al .
1995c ) and a filled in absorption h@xmath0 line ( fekel et al .
1986 ; strassmeier et al .
1990 ) .
we present here one observation of this system at the orbital phase 0.797 ( see fig .
[ fig : plot_aycet_hahe ] ) . in the h@xmath0 line region
we have used a g8iii reference star to perform the spectral subtraction and the subtracted spectrum obtained shows a weak excess emission .
no detectable filling - in of the na i d@xmath1 and d@xmath2 lines has been found , however , a weak absorption in the he i d@xmath3 line appears in the subtracted spectrum .
lcccccccccccc & & & & & & + & @xmath5 & e & s@xmath16/s@xmath17 & h@xmath0 & w@xmath18 & r@xmath22 & f(1.7 ) & & w@xmath20 & i & ew & @xmath23 + & & & & & ( ) & & & & ( ) & & ( ) & + bd cet & 0.569 & - & - & a & 1.80 & 0.452 & 0.928 & & 0.82 & 0.221 & 0.193 & 5.75 + ay cet & 0.797 & - & - & a & 1.62 & 0.446 & 0.940 & & 1.08 & 0.214 & 0.245 & 6.02 + ar psc & 0.373 & - & 0.20/0.80 & e & - & 1.065 & 1.715 & & 1.47 & 0.599 & 1.071 & 6.58 + `` & 0.443 & - & 0.20/0.80 & e & - & 1.063 & 1.680 & & 1.61 & 0.653 & 1.403 & 6.70 + '' & 0.519 & - & 0.20/0.80 & e & - & 1.059 & 1.736 & & 1.46 & 0.676 & 1.211 & 6.64 + `` & 0.524 & - & 0.20/0.80 & e & - & 1.070 & 1.755 & & 1.47 & 0.687 & 1.263 & 6.66 + xx tri & 0.401 & - & - & e & 2.68 & 1.387 & 2.301 & & 1.80 & 1.089 & 2.599 & 6.92 + ux ari & 0.419 & c & 0.30/0.70 & e & 1.26 & 1.225 & 1.966 & & 1.89 & 0.722 & 1.798 & 6.85 + '' & 0.438 & c & 0.30/0.70 & e & 1.07 & 1.101 & 1.799 & & 1.82 & 0.626 & 1.405 & 6.74 + `` & 0.576 & c & 0.30/0.70 & e & 1.41 & 1.159 & 1.818 & & 1.83 & 0.609 & 1.476 & 6.76 + '' & 0.736 & c & 0.30/0.70 & e & 3.17 & 1.533 & 2.504 & & 2.51 & 0.996 & 3.286 & 7.11 + v711 tau & 0.922 & c & 0.16/0.84 & e & 3.36 & 1.315 & 2.084 & & 2.48 & 0.837 & 2.710 & 6.96 + `` & 0.261 & c & 0.16/0.84 & e & 3.63 & 1.264 & 2.095 & & 2.26 & 0.759 & 2.524 & 6.93 + '' & 0.280 & c & 0.16/0.84 & e & 3.51 & 1.255 & 2.081 & & 2.24 & 0.763 & 2.421 & 6.92 + `` & 0.606 & c & 0.16/0.84 & e & 2.93 & 1.342 & 2.185 & & 2.20 & 0.842 & 2.641 & 6.95 + '' & 0.641 & c & 0.16/0.84 & e & 2.88 & 1.328 & 2.164 & & 2.24 & 0.812 & 2.382 & 6.91 + v833 tau & 0.762 & - & - & e & 1.61 & 1.087 & 1.808 & & 1.37 & 0.754 & 1.101 & 6.37 + `` & 0.319 & - & - & e & - & 1.078 & 1.738 & & 1.35 & 0.709 & 1.018 & 6.34 + '' & 0.851 & - & - & e & 1.56 & 1.085 & 1.790 & & 1.35 & 0.745 & 1.070 & 6.36 + v1149 ori & 0.439 & - & - & a & 2.00 & 0.476 & 0.953 & & 0.91 & 0.269 &
0.259 & 5.75 + mm her & 0.498 & c & 0.59/0.41 & a & 1.65 & 0.561 & 1.118 & & 1.42 & 0.320 & 0.590 & 6.38 + `` & 0.630 & h & 0.50 & a & - & 0.588 & 1.101 & & - & 0.000 & 0.000 & 0.00 + & & c & 0.50 & a & - & 0.846 & 1.490 & & 1.46 & 0.302 & 0.541 & 6.34 + '' & 0.745 & h & 0.50 & a & - & 0.603 & 1.134 & & - & 0.000 & 0.000 & 0.00 + & & c & 0.50 & a & - & 0.831 & 1.476 & & 1.41 & 0.270 & 0.494 & 6.30 + v815 her & 0.978 & h & - & f & - & 0.878 & 1.508 & & 1.75 & 0.535 & 1.279 & 6.95 + `` & 0.520 & h & - & a & - & 0.739 & 1.314 & & 1.65 & 0.396 & 0.875 & 6.79 + '' & 0.099 & h & - & a & - & 0.787 & 1.390 & & 1.80 & 0.445 & 1.095 & 6.89 + by dra & 0.684 & t & 0.70/0.30 & e & 1.39 & 1.202 & 1.919 & & & & 1.357 & + & & h & 0.70 & e & & & & & 1.36 & 0.540 & 0.780 & 6.20 + & & c & 0.30 & e & & & & & 1.70 & 0.319 & 0.577 & 5.73 + `` & 0.839 & t & 0.70/0.30 & e & 1.70 & 1.270 & 2.039 & & & & 1.455 & + & & h & 0.70 & e & & & & & 1.38 & 0.548 & 0.803 & 6.22 + & & c & 0.30 & e & & & & & 1.89 & 0.324 & 0.652 & 5.78 + v775 her & 0.394 & h & - & e & - & 1.044 & 1.719 & & 1.39 & 0.745 & 1.103 & 6.51 + & & c & - & e & - & 1.027 & 1.690 & & 1.02 & 0.138 & 0.149 & - + v478 lyr & 0.953 & h & - & a & 3.36 & 0.763 & 1.353 & & 1.53 & 0.446 & 0.727 & 6.55 + hk lac & 0.067 & c & - & a & 1.66 & 0.611 & 1.178 & & 1.31 & 0.366 & 0.765 & 6.43 + '' & 0.070 & c & - & a & 1.75 & 0.610 & 1.170 & & & & & + `` & 0.110 & c & - & a & 1.61 & 0.657 & 1.243 & & 1.32 & 0.411 & 0.887 & 6.49 + '' & 0.113 & c & - & a & 1.61 & 0.647 & 1.237 & & & & & + `` & 0.149 & c & - & a & 1.61 & 0.664 & 1.257 & & 1.37 & 0.420 & 0.929 & 6.51 + '' & 0.153 & c & - & a & 1.57 & 0.661 & 1.244 & & & & & + ar lac & 0.405 & h & 0.36 & a & - & 0.787 & 1.401 & & 1.92 & 0.083 & 0.169 & 5.98 + & & c & 0.64 & a & - & 0.774 & 1.335 & & 2.037 & 0.114 & 0.248 & 6.00 + `` & 0.425 & h & 0.37 & a & - & 0.696 & 1.241 & & 1.44 & 0.062 & 0.095 & 5.73 + & & c & 0.63 & a & - & 0.675 & 1.167 & & 0.92 & 0.018 & 0.018 & 4.86 + '' & 0.939 & h & 0.27 & a & - & - & - & & - & - & - & - + & & c & 0.73 & a & - & - & - & & - & 0.050 & 0.104 & 5.59 + kz and & 0.145 & 1 & 0.58 & a & 1.81 & 0.780 & 1.404 & & 1.06 & 0.209 & 0.237 & 5.94 + & & 2 & 0.42 & a & - & 0.884 & 1.517 & & 1.22 & 0.207 & 0.269 & 5.99 + kt peg & 0.693 & c & 0.95/0.05 & a & 1.85 & 0.349 & 0.786 & & - & 0.000 & 0.000 & 0.00 + `` & 0.024 & c & 0.95/0.05 & a & 1.77 & 0.354 & 0.792 & & - & 0.000 & 0.000 & 0.00 + ii peg & 0.575 & - & - & e & 1.69 & 1.456 & 2.316 & & 1.62 & 1.134 & 2.127 & 6.66 + '' & 0.587 & - & - & e & 1.69 & 1.456 & 2.302 & & 1.62 & 1.135 & 2.146 & 6.66 + `` & 0.735 & - & - & e & 2.04 & 1.407 & 2.311 & & 1.69 & 1.102 & 2.362 & 6.70 + '' & 0.749 & - & - & e & 2.02 & 1.393 & 2.295 & & 1.66 & 1.087 & 2.486 & 6.72 + `` & 0.760 & - & - & e & 2.63 & 1.505 & 2.484 & & 1.80 & 1.198 & 3.692 & 6.88 + '' & 0.874 & - & - & e & 1.87 & 1.279 & 2.102 & & 1.57 & 0.956 & 1.963 & 6.62 + `` & 0.890 & - & - & e & 1.91 & 1.267 & 2.084 & & 1.57 & 0.942 & 1.972 & 6.62 + '' & 0.907 & - & - & e & 1.85 & 1.288 & 2.096 & & 1.66 & 0.959 & 1.961 & 6.62 + this extremely active rs cvn system is a double - lined spectroscopic binary consisting of a k1iv primary and a g7v secondary ( fekel 1996 ) . in our previous observations of this system we have found strong ca ii h & k and h@xmath24 emissions from the cool component ( ffmcc ) , the h@xmath0 line of the active component in emission above the continuum ( ffmcc and montes et al .
1995a , b ) and an important filling - in by chromospheric emission in the h@xmath4 line ( montes et al .
1995d ) .
now we have analysed four spectra of this system in both spectral regions at the orbital phases from 0.37 to 0.52 ( see fig .
[ fig : plot_arpsc_hahe ] ) .
the h@xmath0 line of the active component appears in emission above the continuum , with a profile that changes with the orbital phase owing to the different amounts of overlapping with the absorption of the other component . by subtracting the synthesized spectrum , constructed with g6iv and k0iv reference stars and a relative contribution of 0.2/0.8
we have found a large excess h@xmath0 emission which is well matched using a two - components gaussian fit . in the na
i d@xmath1 and d@xmath2 lines an important excess emission is present in the subtracted spectra , however , the he i d@xmath3 line does not appear in absorption .
this extremely active rs cvn binary is a single - lined spectroscopic binary of spectral type k0iii .
it has very strong ca ii h & k and h@xmath24 emission lines and the h@xmath0 line in emission above the continuum ( strassmeier et al . 1990 ; bopp et al .
1993 ; montes et al .
this system shows the largest amplitude of light variation from spots yet recorded ( hampton et al .
1996 ) . our observed h@xmath0 spectrum at the orbital phase 0.401 ( fig .
[ fig : plot_xxtri_hahe ] ) shows a strong and very broad h@xmath0 emission above the continuum .
the synthesized spectrum has been constructed with a reference star of spectral type k0iii obtaining a satisfactory fit .
the subtracted spectrum exhibits an asymmetric profile that is not well matched using a gaussian fit , therefore we fit the profile by means two gaussian components . in the na i line region a clear excess emission in the d@xmath1 and d@xmath2 lines is present , in agreement with the strong activity of this system .
the behaviour of the he i d@xmath3 line is not clear in this spectrum , but a weak absorption seems to appear .
the description of the simultaneous h@xmath0 , na i d@xmath1 , d@xmath2 , and he i d@xmath3 observations and the detection of a flare on this system can be found in montes et al .
( 1996b ) .
v711 tau , one of the most active of the rs cvn binaries , is a double - lined spectroscopic binary whose components have spectral types g5iv and k1iv . our ca ii h & k analysis of this system ( ffmcc ) showed that both components present emissions .
the cool component is the more active star in the system , and also presents the h@xmath24 line in emission .
this system shows the h@xmath0 line in emission above the continuum and the spectral subtraction ( montes et al .
1995b ) reveals that the k1 star is responsible for most of the excess h@xmath0 emission . recently uv observations obtained with the hst s ghrs ( wood et al .
1996 ; dempsey et al .
1996b , c ) indicate that the transition region lines of v711 tau are emitted almost entirely by the k1 star , and the g star contributes 14% to the chromospheric mg ii h & k lines flux .
we report here five new observations of this systems in the h@xmath0 line region at the orbital phases 0.92 , 0.26 , 0.28 , 0.61 and 0.64 ( 1995 september 13 - 15 ) ( see fig .
[ fig : plot_v711tau_hahe ] ) that confirm the results obtained in our previous observations taken in 1988 , 1986 and 1992 .
the h@xmath0 subtracted profiles presents at all the epochs broad wings and are not well matched using a single - gaussian fit .
these profiles have therefore been fitted using two gaussian components .
the parameters of the broad and narrow components used in the two - gaussian fit are given in table [ tab : measures_nb ] and the corresponding profiles are plotted in fig .
[ fig : plot_v711tau_hahe_nb ] .
in contrast to the behaviour found in other systems in which we have used a two - gaussian fit , the broad component of v711 tau dominates the line profile ( the average contribution of the broad component to the total h@xmath0 ew is 77% , whereas in other stars the contribution ranges between 35% and 63% ) .
wood et al .
( 1996 ) and dempsey et al .
( 1996c ) also found this behaviour in the broad component of the chromospheric and transition region lines of v711 tau in comparison with the broad components found in the less active stars au mic and capella . in fig .
[ fig : plot_v711tau_hahe_nb ] we can also see that , the contribution of the broad component is larger at orbital phases near 0.2 .
( 80 - 87% ) than at larger orbital phases ( 63 - 72% ) and this behaviour remains at different epochs ( see fig .
[ fig : plot_v711tau_hahe_nb ] ) .
these changes indicate not only that the broad component is the result of microflaring in the chromosphere , as in other stars , but that large and long - lived chromospheric flares , that take place in this system ( dempsey et al .
1996c ) , could also produced enhanced emission in the extended line wings .
variable excess emission appears in the na i d@xmath1 and d@xmath2 lines of the corresponding subtracted spectra , however , no detectable absorption is observed in the he i d@xmath3 line .
v833 tau is a by dra system ( dk5e ) and one of the hottest known flare stars . in our previous observations of this system we have found a moderate h@xmath0 emission above the continuum that presents a variable little central self - reversal ( montes et al .
1995a , b ) and a strong h@xmath4 excess emission ( montes et al .
1995d ) .
we analyse here three new observations of this systems in the h@xmath0 line region at orbital phases 0.762 , 0.319 , and 0.851 ( fig .
[ fig : plot_v833tau_hahe ] ) that confirm the behaviour previously noted .
the self - reversal feature changes its wavelength - position across the emission h@xmath0 line profile , however , it appears always centered with the corresponding h@xmath0 absorption line of the synthesized spectrum . the na i d@xmath1 and d@xmath2 lines exhibit very broad wings in the observed spectra , as correspond to a star of spectral type as later as k5 .
although the observed and synthesized spectra are not well matches , due to problems in the normalization , a clear excess emission in the d@xmath1 and d@xmath2 lines can be seen in the subtracted spectra .
the he i d@xmath3 line is not detected in this star . a single - lined spectroscopic binary classified as k1iii .
our previous observations reveal a clear excess h@xmath0 emission ( montes et al .
1995a , b ) , strong ca ii h & k and h@xmath24 emission lines ( montes et al .
1995c ) .
we present here one spectrum at the orbital phase 0.439 ( see fig .
[ fig : plot_v1149ori_hahe ] ) .
the difference with respect to a k0iii reference star reveals a lower excess h@xmath0 emission than the obtained in our previous observations in 1992 ( montes et al .
the narrow absorption that appears in the red wing of the excess h@xmath0 emission in the subtracted spectra could be attributed to a telluric line .
the spectral subtraction reveals that this star has no measurable filling - in of the na i d@xmath1 and d@xmath2 lines .
however , a clear absorption in the he i d@xmath3 line appears in the subtracted spectrum .
double - lined spectroscopic binary ( g2/k0iv ) with partial eclipses .
this system has ca ii h & k emission lines from both components and excess h@xmath0 emission from the cool component ( ffmcc , montes et al .
1995a ) .
we have obtained three new spectra ( fig .
[ fig : plot_mmher_hahe ] ) of this system in the h@xmath0 and na i d@xmath1 and d@xmath2 line regions at orbital phases 0.498 , 0.630 , and 0.745 . at phase 0.498
the hot component hides a 0.30 fraction of the cool one , ( see the diagram of fig .
[ fig : plot_mmher_hahe ] ) and the absorption lines from both components appear overlapped in the observed spectrum . however , at phases 0.630 and 0.745 there is not eclipse and the lines are clearly wavelength - shifted .
the more intense h@xmath0 absorption observed in these spectra corresponds to the hot component and the less intense and blue - shifted absorption corresponds to the cool one .
the spectral subtraction , using g2iv and k0iv as reference stars and a relative contribution of 0.5/0.5 , indicates that the excess h@xmath0 emission arises only from the cool component .
the excess h@xmath0 emission line profiles obtained are well matched using a two - component gaussian fit . in the subtracted spectra of the na i d@xmath1 and d@xmath2 lines region we can also see an excess emission in these lines from the cool component .
the narrow absorption that appears near the position of the hot component could be due to deficient sky correction in these spectra .
the he i d@xmath3 line is not present .
single - lined spectroscopic binary .
our previous ca ii h & k and h@xmath0 observations ( ffmcc , montes et al .
1995a ) indicate that the hot star is the active component .
multiwavelength observations of this system have been recently reported by dempsey et al .
( 1996a ) .
we have taken three spectra of this system at orbital phases 0.978 , 0.520 , 0.099 ( see fig . [
fig : plot_v815her_hahe ] ) . in the h@xmath0 line region
the spectra show a filling - in absorption line with noticeable night to night changes .
the excess h@xmath0 emission obtained with the spectral subtraction of a g5 v reference star is larger at the orbital phases near to 0.0 .
the subtracted h@xmath0 profile presents broad wings and is well matched using a fit with two gaussian components ( narrow and broad ) .
the narrow component is more important when the excess h@xmath0 emission is larger ( see table [ tab : measures_nb ] ) .
a small filling - in is observed in the na i d@xmath1 and d@xmath2 lines at the three orbital phases , and only at phase 0.523 absorption in detected in the he i d@xmath3 line . the prototype of the by@xmath25dra stars .
our previous ca ii h & k observations of this system ( ffmcc ) clearly show that both components are active with the hot component having the stronger ca ii emission .
the two components also show h@xmath24 in emission .
we present in this paper simultaneous h@xmath0 and na i d@xmath1 and d@xmath2 observations of this system at orbital phases 0.684 and 0.839 ( see fig .
[ fig : plot_bydra_hahe ] ) .
these spectra show strong h@xmath0 emission above the continuum and the na i d@xmath1 and d@xmath2 lines with very broad wings corresponding to the later spectral type of the components of this system ( k4v / k7.5v ) . by applying the spectral subtraction technique we have obtained an asymmetric excess h@xmath0 emission line profile which has contributions from both components . a two - gaussian fit has been used to deblend the contribution of the hot and cool components to the line profile .
this fit reveals that the hot component have the stronger excess h@xmath0 emission ew in agreement with the behaviour observed by us in the ca ii h & k lines .
the subtracted spectra in the na i d@xmath1 and d@xmath2 line region reveal that the excess emission in these lines also arises from both components .
the he i d@xmath3 line is not detected in these spectra .
single - lined spectroscopic binary ( k0v/[k5-m2v ] ) with strong ca ii h & k emission lines from the hot component ( ffmcc ) .
the h@xmath0 feature may change from a weak absorption feature to emission above the continuum on times scales of hours ( xuefu and huisong 1984 ) . in our h@xmath0 spectrum at the orbital phase 0.394 ( fig .
[ fig : plot_v775her_hahe ] ) we can see the h@xmath0 line of the hot component totally filled - in by emission and a small emission bump red - shifted in relation to the absorption lines . by subtracting the synthesized spectrum , constructed with a k0iv reference star
, we have obtained a strong excess h@xmath0 emission coming from the hot component and a small excess emission , red - shifted 1.9 @xmath14 with respect to the emission of the hot component .
this small excess perhaps could be attributed to the cooler star of the system , whose assumed spectral type is k5-m2v and whose contribution to the observed spectra is negligible . in the na i lines region the spectral subtraction points out a filling - in of the d@xmath1 and d@xmath2 lines and not detectable absorption in the he i d@xmath3 line .
this by@xmath25dra system is a single - lined spectroscopic binary with strong ca ii h & k emissions from the hot component ( ffmcc ) and a filled - in h@xmath0 absorption line ( fekel 1988 ) .
the h@xmath0 spectrum of this system exhibits a strong filling - in absorption line ( see fig .
[ fig : plot_v478lyr_hahe ] ) . by subtracting the synthesized spectrum constructed with a g8v star we have obtained strong excess h@xmath0 emission , a small excess emission in the na i d@xmath1 and d@xmath2 lines and a clear absorption in the he i d@xmath3 line .
hk lac is a single - lined spectroscopic binary ( f1v / k0iii ) with very strong ca ii h & k emissions , the h@xmath24 line in emission and an important excess h@xmath0 emission ( ffmcc ; montes el at .
1995a ) .
this system shows a very variable h@xmath0 profile ( from filled - in absorption to moderate emission ) and flares ( see catalano & frasca 1994 ) .
however , in our six h@xmath0 spectra taken in three consecutive nights , with orbital phases from 0.067 to 0.153 , we always observe filled - in absorption line with small night to night variations in the excess h@xmath0 emission from the cool component . in fig .
[ fig : plot_hklac_hahe ] a spectrum of each night is showed .
the observed spectra are well matched using a k0iii as reference star .
the subtracted spectra show an important excess emission in the na i d@xmath1 and d@xmath2 lines and a clear absorption in the he i d@xmath3 line .
the total eclipsing binary ar lac is one of the best studied rs@xmath25cvn systems .
both components of the system ( g2iv / k0iv ) are active ( cabs ) , however , due to the orbital phase , in our previous ca ii h & k and h@xmath0 observations ( ffmcc ) it was not possible to separate the contribution from each component .
we analyse here h@xmath0 observations of this system at three orbital phases . in fig .
[ fig : plot_arlac_hahe ] we have superimposed to each spectrum the corresponding geometrical position of the cool and hot components . at orbital phase 0.405 there is not eclipse and the subtracted spectrum exhibits excess h@xmath0 emission from both components , being a slightly larger that corresponding to the cool one . at phase 0.425
the hot component hides a 4 @xmath26 of the cool one and the excess emission obtained is now slightly larger in the hot component .
this change in the excess emission observed in both components could be attributed to the hot component hiding of prominence - like material or other active regions responsible for the h@xmath0 emission of the cool component . in the observation at the orbital phase 0.939 ,
when a 33 @xmath26 of the hot component is hidden by the cool component , we detect an excess h@xmath0 absorption located at the wavelength position corresponding to the hot component .
this excess absorption indicates the presence of prominence - like extended material seen off the limb of the cool component that absorbs the photospheric continuum radiation of the star behind .
similar prominence - like structures have been found in other eclipsing rs cvn systems ( see hall & ramsey 1992 ) and recently , siarkowski et al .
( 1996 ) have found geometrically extended structures in the corona of ar lac .
at these three orbital phases the na i d@xmath1 and d@xmath2 lines also present excess emissions from both components and the he i d@xmath3 shows absorption features in the subtracted spectra . at orbital phase 0.933
the he i d@xmath3 line presents a large excess absorption at the wavelength position corresponding to the hot component , which could also be attributed to the prominence - like material .
kz and is the component b of the visual pair ads 16557 , and is a double - lined spectroscopic binary with spectral types dk2/dk2 .
this system presents ca ii h & k and h@xmath24 emissions and excess h@xmath4 emission from both components ( ffmcc , montes et al .
1995c , d ) . by applying the spectral subtraction technique we have found that also both components present excess emission in the h@xmath0 line and in the in the na i d@xmath1 and d@xmath2 lines , with very similar intensities ( see fig . [
fig : plot_kzand_hahe ] ) .
the he i d@xmath3 line is not present in the subtracted spectrum . the synthesized spectrum has been constructed with two k2v stars and with a contribution of each component to the total continuum of 0.58/0.42 . this system is a double - lined spectroscopic binary ( g5v / k6v ) with small ca ii h & k emissions from both components ( montes et al .
we analyse here two spectra at the orbital phases 0.693 and 0.024 ( see fig .
[ fig : plot_ktpeg_hahe ] ) .
we have constructed the synthesized spectrum with two reference stars of spectral types g2iv and k3v taking into account that the hot component has the larger contribution to the continuum . in both subtracted spectra no detectable excess emission is observed in the h@xmath0 line nor in the na i d@xmath1 and d@xmath2 lines .
the he i d@xmath3 line appears in absorption .
this is a single - lined spectroscopic binary ( k2 - 3v - iv ) with strong ca ii h & k emission lines and variable h@xmath0 emission above the continuum .
we have reported a strong excess emission in the h@xmath4 line ( montes et al .
1995d ) .
we present here eight spectra of this system taken in three consecutive nights and with orbital phases ranging from 0.58 to 0.91 .
the strong h@xmath0 emission present in the subtracted spectra in all the phases shows remarkable night to night changes ( see fig . [
fig : plot_iipeg_hahe ] ) . in the second night
the spectrum at orbital phase 0.760 shows a remarkable excess h@xmath0 emission enhancement with regard to the other phases , which is much more noticeable in comparison with the spectra of the first and third nights .
the excess h@xmath0 emission ew increases in a factor of 1.9 in an interval of 1 day .
this fact suggests that we have detected a flare since enhancements of this amount during flares are typical of chromospheric emission lines ( simon et al .
1980 , catalano & frasca 1994 ) and has also been observed by us in the flare detected in ux ari ( montes et al .
1996b ) .
the he i d@xmath3 line appears in emission in the observed spectrum only at orbital phase 0.760 ( see fig . [
fig : plot_iipeg_hahe ] ) corresponding to the increase of the emission observed in the h@xmath0 line .
this fact supports the detection of a flare - like event , since in the sun the he i d@xmath3 line appears as absorption in plage and weak flares and as emission in strong flares ( zirin 1988 ) . at the other orbital phases of the first and second night an small emission in he i d@xmath3
is also present in the subtracted spectrum .
however , in the third night where the h@xmath0 ew present the lower values the he i d@xmath3 emission is not observed .
the subtracted spectra show that the h@xmath0 profile presents broad wings , which are much more remarkable in the flare spectrum . in fig .
[ fig : plot_iipeg_hahe_nb ] we have represented the subtracted spectrum at phases 0.587 , 0.749 , 0.760 and 0.890 and the corresponding narrow and broad components used to perform the two - component gaussian fit .
the changes in the ew(h@xmath0 ) and fwhm(h@xmath0 ) of the total subtracted spectra and of the corresponding narrow and broad components during the three nights can be seen in fig .
[ fig : iipeg_ew_fwhm ] . note the strong change in the fwhm of the broad component during the flare .
the inspection of the subtracted h@xmath0 spectra shows that in some stars the emission line profile has very broad wings , and is not well matched using a single - gaussian fit .
these profiles have therefore been fitted using two gaussian components : a narrow component having a fwhm of 45 - 90 km s@xmath27 and a broad component with a fwhm ranging from 133 to 470 km s@xmath27 . in table
[ tab : measures_nb ] we list the parameters ( i , fwhm , ew ) of the broad and narrow components . as can be seen in this table the average contribution of the broad component to the total ew of the line ranges from 35% in ar psc to 77% in v711 tau .
we have observed this behaviour in the chromospheric h@xmath0 line only in the most active systems of the sample , the stars with intense h@xmath0 emission above the continuum ( ar psc , xx tri , ux ari , v711 tau ( fig .
[ fig : plot_v711tau_hahe_nb ] ) , ii peg ( fig .
[ fig : plot_iipeg_hahe_nb ] ) ) and also in systems with important excess h@xmath0 emission ( mm her , v815 her , hk lac ) .
furthermore , a revision of the h@xmath0 spectra of a large sample of chromospherically active binaries previously analysed by us ( montes et al .
1995a , b ) indicates that the very active systems dm uma and vv mon also have broad components and that some systems also studied in this paper ( v711 tau , v815 her , hk lac ) exhibit this behaviour at different epochs .
this parameterization of the subtracted h@xmath0 profile using a narrow and a broad component have been only reported until now for the rs cvn system dm uma by hatzes ( 1995 ) who suggests that the broad component could result from large - scale motions or winds in the chromosphere .
similar broad components have also been found in several transition region lines of the dm0e star au mic and the rs cvn systems capella and v711 tau using high - resolution uv observations obtained with the hst s ghrs ( linsky & wood 1994 ; linsky et al .
1995 ; wood et al .
1996 ; dempsey et al .
1996b , c ; robinson et al .
the broad components in the transition region lines are interpreted by linsky & wood ( 1994 ) as arising from microflaring , because these broad profiles are reminiscent of the broad profiles observed in solar transition region explosive events , which are thought to be associated with emerging magnetic flux regions where field reconnection occurs .
the microflares are frequent , short - duration , energetically weak disturbances , i.e. they are the low - energy extension of flares , and therefore have large - scale motions associated that could explain the broad wings observed in these lines .
the microflaring activity could occur not only in the transition region but also in the chromosphere of very active stars as indicates the detection of broad components in the chromospheric mg ii h & k lines of v711 tau ( wood et al .
1996 ) but not in the chromospheric lines of the less active star capella ( linsky et al .
1995 ) .
our detection of broad wings in the chromospheric h@xmath0 line of the most active systems of our sample allows us to conclude that microflaring occurs also in the chromosphere and that it is much more important in extremely active stars .
furthermore , within the group of stars that present this phenomenon a correlation between the contribution of the broad components and the degree of stellar activity seems to be present , as can be seen in fig .
[ fig : fwhm_b_ew_t ] left - panel where we have plotted for each star the ew of the broad component versus the total excess h@xmath0 ew .
this correlation was also noted by wood et al .
( 1996 ) when compared the dominant broad component of v711 tau with the smaller broad component of the less active stars au mic and capella . on the other hand , when we plot the fwhm of the broad component versus the total h@xmath0 ew ( fig .
[ fig : fwhm_b_ew_t ] right - panel ) a general trend is not observed .
however , in this case the relation appears for individual star , i.e. there is an increase in the fwhm when the activity level increases , which is consistent with the hypothesis that microflaring is responsible for the broad component emission .
we have also found that the larger changes in the excess h@xmath0 emission in the stars analyzed appear to occur predominantly in the broad component , as have been already noted in the case of dm uma by hatzes ( 1995 ) .
an extreme case of this behaviour is the strong change in the broad component that occurs during the flares detected in ux ari and ii peg ( see fig .
[ fig : iipeg_ew_fwhm ] and [ fig : fwhm_b_ew_t ] ) . since large scale mass motions do occur in solar flares ,
a large flare in these two systems may explain the increase of the h@xmath0 emission and the very broad wings observed in these spectra .
lcccccccccc & & & & + & @xmath5 & i & fwhm & ew@xmath28 & ew@xmath28/ew@xmath29 & & i & fwhm & ew@xmath30 & ew@xmath30/ew@xmath29 + & & & ( ) & ( ) & ( % ) & & & ( ) & ( ) & ( % ) + ar psc & 0.373 & 0.089 & 3.570 & 0.338 & 31.6 & & 0.541 & 1.272 & 0.732 & 68.3 + & 0.443 & 0.132 & 4.473 & 0.625 & 44.5 & & 0.534 & 1.358 & 0.772 & 55.0 + & 0.519 & 0.108 & 3.362 & 0.388 & 32.0 & & 0.587 & 1.319 & 0.823 & 68.0 + & 0.524 & 0.107 & 3.652 & 0.417 & 33.0 & & 0.603 & 1.316 & 0.845 & 67.0
+ xx tri & 0.401 & 0.392 & 3.785 & 1.581 & 60.8 & & 0.712 & 1.132 & 1.007 & 38.7 + ux ari & 0.419 & 0.302 & 3.720 & 1.194 & 66.4 & & 0.431 & 1.317 & 0.605 & 33.6 + & 0.438 & 0.297 & 2.917 & 0.922 & 65.6 & & 0.344 & 1.316 & 0.483 & 34.4 + & 0.576 & 0.145 & 5.153 & 0.778 & 50.8 & & 0.468 & 1.512 & 0.753 & 49.2 + & 0.736 & 0.405 & * 5.414 * & 2.229 & 67.8 & & 0.580 & 1.723 & 1.069 & 32.4 + v711 tau ( 1995 ) & 0.922 & 0.335 & 4.822 & 1.700 & 62.7 & & 0.502 & 1.852 & 0.990 & 36.5 + & 0.261 & 0.425 & 4.563 & 2.046 & 81.1 & & 0.344 & 1.254 & 0.459 & 18.2 + & 0.280 & 0.417 & 4.406 & 1.947 & 80.4 & & 0.344 & 1.265 & 0.463 & 19.1 + & 0.606 & 0.318 & 5.176 & 1.707 & 64.6 & & 0.509 & 1.635 & 0.887 & 33.6 + & 0.641 & 0.402 & 4.063 & 1.735 & 72.8 & & 0.406 & 1.486 & 0.642 & 27.0 + v711 tau ( 1992 ) & 0.130 & 0.546 & 4.022 & 2.303 & 88.3 & & 0.234 & 1.226 & 0.305 & 11.7 + v711 tau ( 1986 ) & 0.200 & 0.570 & 4.489 & 2.660 & 87.8 & & 0.251 & 1.381 & 0.369 & 12.2 + & 0.260 & 0.576 & 4.513 & 2.570 & 85.2 & & 0.311 & 1.486 & 0.448 & 14.8 + v711 tau ( 1988 ) & 0.880 & 0.362 & 5.063 & 1.821 & 70.5 & & 0.410 & 1.743 & 0.761 & 29.5 + mm her & 0.498 & 0.076 & 3.367 & 0.272 & 46.1 & & 0.248 & 1.203 & 0.318 & 53.9 + & 0.630 & 0.044 & 3.612 & 0.167 & 30.9 & & 0.274 & 1.283 & 0.374 & 69.1 + & 0.745 & 0.064 & 3.222 & 0.221 & 44.7 & & 0.217 & 1.180 & 0.273 & 55.3 + v815 her ( 1995 ) & 0.978 & 0.142 & 4.622 & 0.687 & 53.7 & & 0.397 & 1.377 & 0.582 & 45.5 + & 0.520 & 0.134 & 3.577 & 0.509 & 58.2 & & 0.267 & 1.288 & 0.366 & 41.8 + & 0.099 & 0.113 & 4.775 & 0.561 & 51.2 & & 0.334 & 1.467 &
0.521 & 47.6 + v815 her ( 1989 ) & 0.520 & 0.092 & 4.420 & 0.416 & 60.1 & & 0.200 & 1.294 & 0.276 & 39.9 + hk lac ( 1995 ) & 0.067 & 0.089 & 4.890 & 0.458 & 59.9 & & 0.291 & 0.991 & 0.307 & 40.1 + & 0.110 & 0.095 & 5.286 & 0.515 & 58.1 & & 0.329 & 1.065 & 0.373 & 42.0 + & 0.149 & 0.089 & 5.951 & 0.522 & 56.2 & & 0.339 & 1.129 & 0.408 & 43.9 + hk lac ( 1989 ) & 0.100 & 0.142 & 5.654 & 0.726 & 58.8 & & 0.346 & 1.384 & 0.509 & 41.2 + ii peg & 0.575 & 0.211 & 3.379 & 0.758 & 35.6 & & 0.952 & 1.351 & 1.369 & 64.4 + & 0.587 & 0.247 & 3.352 & 0.882 & 41.1 & & 0.913 & 1.301 & 1.264 & 58.9 + & 0.735 & 0.271 & 3.880 & 1.118 & 47.3 & & 0.856 & 1.366 & 1.244 & 52.7 + & 0.749 & 0.210 & 5.000 & 1.119 & 45.0 & & 0.898 & 1.430 & 1.368 & 55.0 + & 0.760 & 0.245 & * 7.984 * & 2.077 & 56.3 & & 0.952 & 1.594 & 1.616 & 43.7 + & 0.874 & 0.223 & 3.697 & 0.876 & 44.6 & & 0.750 & 1.359 & 1.086 & 55.3 + & 0.890 & 0.197 & 4.049 & 0.849 & 43.1 & & 0.761 & 1.386 & 1.121 & 56.8 + & 0.907 & 0.242 & 3.563 & 0.918 & 46.8 & & 0.743 & 1.319 & 1.042 & 53.1 + dm uma & 0.400 & 0.207 & 10.36 & 0.810 & 27.8 & & 0.999 & 1.971 & 2.107 & 72.2 + & 0.530 & 0.246 & 8.607 & 1.135 & 38.0 & & 0.918 & 1.893 & 1.850 & 62.0 + vv mon & 0.710 & 0.177 & 5.382 & 0.918 & 68.2 & & 0.252 & 1.596 & 0.428 & 31.8 + the he i d@xmath3 @xmath315876 and he i @xmath3110830 triplets are known to be activity indicators in the sun and late type stars ( zirin 1988 , shcherbakov et al .
in the sun , the he i d@xmath3 line appears like an absorption feature cospatial with plages ( landman 1981 ) , and it almost disappears when we look at the solar disk .
this feature is also seen in absorption in surges , eruptive prominences , and weaker flares , whereas in emission in more intense flares ( zirin 1988 ) .
the he i d@xmath3 line is formed at middle chromosphere , and its correlation with x - ray flux and ca ii h & k lines suggests that the fractional area of the stellar disk covered by plages may be a key factor in the formation of d@xmath3 ( danks & lambert 1985 ) .
moreover it has been observed a slight rotational modulation in @xmath32 cet ( lambert & obrien 1983 ) and @xmath33@xmath34 ori ( danks & lambert 1985 ) .
historically , there have been basically two models to explain the line formation of he i d@xmath3 : ( i ) zirin ( 1975 ) suggested that he i triplet levels were populated by over - photoionisation of the he i atoms by euv and x - ray radiation , and subsequent radiative recombinations and cascade .
( ii ) wolff et al .
( 1985 ) argued that collisional excitation and ionization in the chromosphere contributed also to the he i d@xmath3 formation , and not only the euv and x - ray radiation from the corona .
however , the most recent models ( andretta & giampapa 1995 , lanzafame & byrne 1995 ) seem to indicate that the primary mechanism in the formation of the he i triplets is the collisional excitation and ionization ( followed by recombination cascade ) by electron impact .
the he i d@xmath3 line usually appears , in stars , in absorption , but sometimes is in emission .
there are two possible reasons : ( i ) temperature and/or electronic density conditions are higher than ordinary , like may occur in flares ( zirin 1988 , andretta & giampapa 1995 , lanzafame & byrne 1995 ) .
( ii ) as it has been seen in he i @xmath3110830 , depending on the position of the emitting region in the disk or off the limb , the he i d@xmath3 line would appear in absorption or emission .
since the he i @xmath3110830 is formed in emitting regions located at some distance from the stellar photosphere , when the emitting region is seen in projection against the stellar disk , he i @xmath3110830 line appears in absorption , and when the emitting region is observed off the stellar limb , the line is in emission ( simon et al .
1982 , wolff & heasley 1984 ) .
these conclusions could extend to the case of he i d@xmath3 , since it is produced at the same region that he i @xmath3110830 .
the he i d@xmath3 line has been studied only in some chromospherically active binaries as ii peg ( huenemoerder & ramsey 1987 ; huenemoerder et al .
1990 ) , dm uma ( hatzes 1995 ) , er vul ( gunn & doyle 1996 ) and gk hya ( gunn et al .
the observation of emission in the he i d@xmath3 line supports the detection of flare like events as in the case of ii peg ( huenemoerder & ramsey 1987 ) and the weak - lined t tauri star v410 tau ( welty & ramsey 1996 ) . in our spectra
the he i d@xmath3 line has been found in emission only during the flares of ux ari and ii peg .
we wish to emphasize that the detection of he i d@xmath3 in emission in the rs cvn systems seems to occur at orbital phases near to the quadrature . in our observations
we have detected he i d@xmath3 in emission at orbital phase 0.74 in ux ari ( montes et al .
1996b ) and at 0.76 in ii peg .
this line has been also observed in emission at orbital phases 0.22 , 0.26 , 0.77 in ii peg by huenemoerder & ramsey ( 1987 ) and huenemoerder et al .
probably we are observing a flare off the limb , i.e. when the plage regions are near the limb ( the active regions are preferably in the opposite faces of the stars ) , which is the most favourable situation to see an off the limb flare .
but we can not distinguish whether the emission is only due to the existence of the flare , or it is favoured by the relative position on the star .
the application of the spectral subtraction to our sample reveals that the he i d@xmath3 line appears as an absorption feature more frequently in giants than in dwarfs .
three out of five giants observed show clear absorptions ( bd cet , v1149 ori and hk lac ) and two of them exhibit a possible absorption ( ay cet and xx tri ) , while among iv and v luminosity class stars there are only two plain absorptions .
various authors seem to point out a more frequent presence of he i @xmath3110830 and @xmath315876 triplets in giants and supergiants than in dwarfs ( simon et al .
1982 , zirin 1982 , wolff & heasley 1984 ) .
zirin ( 1982 ) observed he i @xmath3110830 usually in absorption , but sometimes it appears in emission , especially in giant and supergiants , with a p cygni form , and he attributes it to a mass - ejection phenomenon ( see also obrien & lambert 1986 ) .
simon et al . ( 1982 ) saw that none of the single red giants , in their sample , having strong @xmath3110830 absorption or emission has prominent transition region emission lines or soft x - ray emission , and they proposed a scattering process - like responsible for the @xmath3110830 line formation .
smith ( 1983 ) attributed a larger intensity in @xmath3110830 line for giants and supergiants to the most efficient ionization by euv and x - ray radiation in atmospheres of coronally active giants .
other authors say that @xmath3110830 line is sometimes produced by the propagation of acoustic shock waves , or that he i @xmath3110830 transition represents a wind diagnostic .
some of the above proposed mechanisms could also be applied to the he i d@xmath3 line .
the na i d@xmath1 and d@xmath2 lines are collisionally - controlled in the atmospheres of late - type stars and are formed in the lower chromosphere .
so , the detection of filled - in absorption in the d@xmath1 and d@xmath2 lines may provide information about chromospheric emission .
( see the recent models of these lines for m dwarfs stars by andretta et al .
1997 ) . in the sun , barrado et al .
( 1995 ) and barrado ( 1996 ) have found changes in the ew of na i lines in spectra taken at different regions over the solar surface , and a relation with the filled - in absorption h@xmath0 that might indicate that there is a non - radiative effect in the formation of these lines . in other stars the d@xmath1 and d@xmath2 lines
have been observed in emission or as a filled - in in very active red dwarf flare stars ( pettersen et al .
1984 ; pettersen 1989 ) .
however , no systematic study of these lines has been performed in stars with different levels of activity , and in chromospherically active binaries only the negative and uncertain detection of filled - in in the few active binaries er vul and gk hya , respectively , has been reported in the recent studies of gunn & doyle ( 1996 ) and gunn et al .
( 1996 ) .
the application of the spectral subtraction technique in these lines is more difficult that in the h@xmath0 line , because their wings are very sensitive to the effective temperature , mainly in latter spectral types .
therefore , small differences in spectral type , not appreciated in the h@xmath0 line , produce significant changes of the subtracted spectra in the wings of the na i lines .
moreover , in this spectral region there is a large number of telluric lines , and in the spectra of some stars interstellar na i could be present .
however , the distances of the majority of the stars is lower than 50 pc and the effect of the interstellar na i is negligible . in spite of this problems
, some conclusions can be drawn . in the chromospherically active binaries analysed here
, the spectral subtraction reveals that the core of the na i d@xmath1 and d@xmath2 lines are filled - in by chromospheric emission in the more active star of the sample ( the star with h@xmath0 emission above the continuum , and with larger excess h@xmath0 emission ew ) .
the stars with only a small or without excess h@xmath0 emission as bd cet , ay cet , v1149 ori and kt peg do not exhibit excess emission in the na i lines .
moreover , the excess d@xmath1 and d@xmath2 emissions obtained are larger in the systems with larger excess h@xmath0 emission , and also increase in the flares observed in ux ari and ii peg . in short , we can conclude that the filled - in of the core of the na i d@xmath1 and d@xmath2 lines could be used as a chromospheric activity indicator .
in this paper we have analysed , using the spectral subtraction technique , simultaneous h@xmath0 , na i d@xmath1 , d@xmath2 , and he i d@xmath3 spectroscopic observations of 18 chromospherically active binary systems .
we have found excess h@xmath0 in all the systems , except kt peg which have also small emission in the ca ii h & k lines .
the subtracted h@xmath0 profile of the more active stars of the sample ( h@xmath0 in emission above the continuum ) has very broad wings , and is well matched using a two - components gaussian fit ( narrow and broad ) .
the broad component is primarily responsible for the observed variations of the profile , and its contribution to the total ew increases with the degree of activity .
so , we have interpreted this broad component emission as arising from microflaring activity that take place in the chromosphere of this very active stars .
the h@xmath0 observation of the eclipsing binary system ar lac at orbital phase 0.939 , when a 0.33 fraction of the hot component is hidden by the cool component , allowed us to detect the presence of prominence - like extended material seen off the limb of the cool component through the detection of excess h@xmath0 absorption located at the wavelength position corresponding to the hot component . a small excess absorption in the he i d@xmath3 line is also present at this orbital phase .
the application of the spectral subtraction technique reveals that the core of the na i d@xmath1 and d@xmath2 lines are also filled - in by chromospheric emission in the more active star of the sample .
the stars with only a small excess h@xmath0 emission do not exhibit excess emission in the na i lines .
this work has been supported by the universidad complutense de madrid and the spanish direccin general de investigacin cientfica y tcnica ( dgicyt ) under grant pb94 - 0263 .
we would like to thank the referee s. catalano for suggesting several improvements and clarifications .
|
this is the first paper of a series aimed at studying the chromosphere of active binary systems using the information provided for several optical spectroscopic features .
simultaneous h@xmath0 , na i d@xmath1 , d@xmath2 , and he i d@xmath3 spectroscopic observations are reported here for 18 systems . the chromospheric contribution in these lines
have been determined using the spectral subtraction technique .
very broad wings have been found in the subtracted h@xmath0 profile of some of the more active stars .
these profiles are well matched using a two - components gaussian fit ( narrow and broad ) and the broad component could be interpreted as arising from microflaring .
prominence - like extended material have been detected in a near - eclipse h@xmath0 observation of the system ar lac .
the excess emission found in the na i d@xmath1 and d@xmath2 lines by application of the spectral subtraction technique and the behaviour of the h@xmath0 line in the corresponding simultaneous observations indicate that the filling - in of the core of these lines is a chromospheric activity indicator . for giant stars of the sample the he i
d@xmath3 line has been detected in absorption in the subtracted spectra .
an optical flare has been detected in ux ari and ii peg through the presence of the he i d@xmath3 in emission in coincidence with the enhancement of the h@xmath0 emission .
|
heegaard floer homology is a collection of invariants for closed oriented three - manifolds , introduced by peter ozsvth and zoltn szab @xcite .
there are four versions , denoted by @xmath0 and @xmath1 , which are graded abelian groups . the hat version @xmath2 is defined as the homology of a chain complex @xmath3 coming from a heegaard diagram of the three - manifold @xmath4 .
the differentials count the number of points in certain moduli spaces of holomorphic disks , which are hard to compute in general .
there is also a relative version of the theory corresponding to pairs @xmath5 , where @xmath6 is a knot in @xmath4 .
if @xmath6 is null - homologous , then a seifert surface @xmath7 of @xmath6 induces a filtration of the chain complex @xmath3 , and the chain homotopy type of the filtered chain complex is a knot invariant .
the homology groups @xmath8 of successive quotients of filtration levels are called knot floer homology groups ( @xcite ) .
a cobordism between two three - manifolds induces homomorphisms on the heegaard floer homology groups of the two three - manifolds .
in fact , the homomorphisms on @xmath1 and @xmath9 can be used to construct an invariant of smooth four - manifolds with @xmath10 ( @xcite ) , called the ozsvth - szab invariant .
conjecturally , the ozsvth - szab four - manifold invariant is equivalent to the gauge - theoretic seiberg - witten invariant .
heegaard floer homology turns out to be a fruitful and powerful theory in the study of three - dimensional and four - dimensional topology .
it gives an alternate proof of the donaldson diagonalization theorem and the thom conjecture for @xmath11 ( @xcite ) .
heegaard floer homology also detects the thurston norm of a three - manifold ( @xcite ) .
moreover , knot floer homology detects the genus ( @xcite ) and fiberedness ( @xcite ) of knots and links in the three - sphere .
there is an invariant @xmath12 coming from the knot filtration , whose absolute value gives a lower - bound of the slice genus for knots in the three - sphere ( @xcite ) . despite its success
, there was no general method to compute the invariants .
there were combinatorial descriptions in certain special cases , but the computation for an arbitrary three - manifold was an open problem ( @xcite ) . in this paper
, we give an algorithm to compute @xmath2 for a three - manifold @xmath4 , and also @xmath8 for a knot @xmath6 in any three - manifold .
all our computations will be done with coefficients in @xmath13 .
we show that one can always find heegaard diagrams satisfying certain properties ( definition [ defn : nicediagram ] ) .
using such heegaard diagrams , which we call nice , it will be easy to compute @xmath14 and @xmath15 .
our main results are summarized in the following theorems .
[ thm : holodisk ] given a nice heegaard diagram of a closed oriented three - manifold @xmath4 , @xmath2 can be computed combinatorially .
similarly , for a knot @xmath16 , @xmath8 can be computed combinatorially in a nice heegaard diagram .
[ thm : nicediagram ] every closed oriented three - manifold @xmath4 admits a nice heegaard diagram . for a null - homologous knot @xmath6 in a closed oriented three - manifold @xmath4 ,
the pair @xmath5 admits a compatible nice heegaard diagram .
in fact , there is an algorithm to convert any pointed heegaard diagram to a nice heegaard diagram via isotopies and handleslides
. it will be interesting to compare our result with the recent work of ciprian manolescu , peter ozsvth and the first author in @xcite , where they gave a combinatorial description of knot floer homology of knots in @xmath17 , in all versions .
we hope this method can be generalized to compute some of the other versions , notably @xmath18 and @xmath19 .
it would also be nice to have a proof of the invariance of the combinatorial description without using holomorphic disks .
the paper is organized as follows . in section [ section - prelims ] , we give an overview of certain concepts in heegaard floer theory . in section [ section - holo ] ,
we give a combinatorial characterization of index one holomorphic disks in nice heegaard diagrams . in section [ section - algorithm ] , we give an algorithm to get such heegaard diagrams . in section [ section - examples ] , we give examples to demonstrate our algorithm for three - manifolds and knots in the three - sphere .
the first author wishes to thank zoltn szab for introducing him to the subject of heegaard floer homology and having many helpful discussions at various points .
this work was done when the second author was an exchange graduate student in columbia university .
he is grateful to the columbia math department for its hospitality .
he also would like to thank rob kirby and peter ozsvth for their continuous guidance , support and encouragement .
we thank matthew hedden , robert lipshitz , ciprian manolescu , peter ozsvth , jacob rasmussen and dylan thurston for making comments and having helpful discussions during the development of this work .
we also thank the referees for their comments and suggestions .
in this section , we review the definition of heegaard floer homology .
see @xcite for details .
the heegaard floer homology of a closed oriented three - manifold @xmath4 is defined from a pointed heegaard diagram representing @xmath4 .
a heegaard splitting of @xmath4 is a decomposition of @xmath4 into two handlebodies glued along their boundaries .
we fix a self - indexing morse function @xmath20 on @xmath4 with @xmath21 index zero critical points and @xmath21 index three critical points .
( we usually choose @xmath22 . )
then @xmath20 gives a heegaard splitting of @xmath4 , where the two handlebodies are given by @xmath23 $ ] and @xmath24 .
if the number of index one critical points or the number of index two critical points of @xmath20 is @xmath25 , then @xmath26 is a genus @xmath27 surface .
we fix a gradient like flow on @xmath4 corresponding to @xmath20 .
we require @xmath20 to have the property that @xmath4 contains a disjoint union of @xmath21 flow lines , each flowing from an index zero critical point to an index three critical point .
we get a collection @xmath28 of @xmath29 circles on @xmath30 which flow down to the index one critical points , and another collection @xmath31 of @xmath32 circles on @xmath30 which flow up to the index two critical points .
note that both @xmath33 and @xmath34 have @xmath21 components .
we fix @xmath21 points @xmath35 ( called _ basepoints _ ) in the complement of the @xmath29 circles and the @xmath32 circles in @xmath30 , such that each component of @xmath36 contains exactly one @xmath37 and each component of @xmath34 contains exactly one @xmath38 .
this is equivalent to the condition that the trajectories of @xmath37 s under the gradient like flow , hit all the index zero and all the index three critical points .
we write @xmath39 .
the tuple @xmath40 is called a _ pointed heegaard diagram _ for @xmath4 . there are some moves on a heegaard diagram that do not change the underlying three - manifold .
an _ isotopy _ moves the @xmath29 curves and @xmath32 curves in two one - parameter families @xmath41 and @xmath42 in @xmath43 , moving by isotopy , such that the @xmath29 curves remain disjoint and the @xmath32 curves remain disjoint for each @xmath44 . in a _
handleslide _ of @xmath45 , we replace a pair of @xmath29 curves @xmath46 and @xmath47 with a pair @xmath46 and @xmath48 , such that the three curves @xmath46 , @xmath47 and @xmath48 bound a pair of pants in @xmath43 disjoint from all the other @xmath29 curves .
a handleslide of @xmath49 is defined similarly .
there is also a move called _ stabilization _ , but we will not be using it in the present paper .
these moves are called _
heegaard moves_. heegaard floer homology is a certain version of lagrangian floer homology .
the ambient symplectic manifold is the symmetric product @xmath50 .
the two half - dimensional totally real subspaces are the tori @xmath51 and @xmath52 .
the generators for the chain complex @xmath53 are the intersection points between these two tori , and the boundary maps are given by counting certain holomorphic disks . for more details see @xcite , and see @xcite for the issue of _ boundary degenerations _ when @xmath54 .
when @xmath22 , the homology of the chain complex @xmath53 is an invariant for the three - manifold @xmath4 , written as @xmath2 . for a general @xmath21 , and
@xmath4 a rational homology three - sphere , we have @xmath55 where @xmath56 is the singular homology of the @xmath57-dimensional torus with coefficients in @xmath58 . when we have a link @xmath59 in @xmath4 ,
we ensure that @xmath59 is a union of flow lines from index zero critical points to index three critical points .
we also ensure that @xmath59 contains all index zero and index three critical points and contains no index one or two critical points .
we orient @xmath59 and @xmath30 and define @xmath37 s as the positive intersection points between @xmath59 and @xmath30 .
we write the other @xmath21 intersections as @xmath60 . such a heegaard diagram , denoted by @xmath61 ,
is called a _ pointed heegaard diagram _ for the pair @xmath62 .
( for a link with @xmath63 components , we usually choose @xmath64 . )
the knot ( link ) floer homology @xmath65 is defined similarly , where the boundary maps count a more restricted class of holomorphic disks .
see @xcite for details . in the present paper , we will use the cylindrical reformulation of the heegaard floer homology by lipshitz .
see @xcite for details . given a pointed heegaard diagram @xmath66 , the generators of the chain complex @xmath67 are given by formal sums of @xmath25 distinct points in @xmath30 , @xmath68 , such that each @xmath29 circle contains some @xmath69 and each @xmath32 circle contains some @xmath70 .
a connected component of @xmath71 is called a _
region_. a formal sum of regions with integer coefficients is called a _ @xmath72-chain_. given two generators @xmath73 and @xmath74
, we define @xmath75 to be the collection of all @xmath72-chains @xmath76 such that @xmath77 .
such @xmath72-chains are called _
domains_. given a point @xmath78 , let @xmath79 be the coefficient of the region containing @xmath80 in @xmath76 .
a domain @xmath76 is _ positive _ if @xmath81 for all points @xmath82 .
we define @xmath83 .
a heegaard diagram is _ admissible _ , if , for every generator @xmath73 , any positive domain @xmath84 is trivial .
if the three - manifold @xmath4 has @xmath85 , we require the heegaard diagram to be admissible .
fix two generators @xmath73 , @xmath74 and a domain @xmath86 .
let @xmath7 be a surface with boundary , with @xmath87 marked points @xmath88 on @xmath89 , such that the @xmath90 points and the @xmath4 points alternate .
the @xmath87 arcs on @xmath89 in the complement of the marked points are divided into two groups @xmath91 and @xmath92 , each containing @xmath25 arcs , such that the @xmath91 arcs and the @xmath92 arcs alternate .
let @xmath93 and @xmath94 be the projection maps from @xmath95 onto its first and second factors .
look at maps @xmath96 such that the image of @xmath97 is @xmath76 ( as @xmath72-chains ) and the image of @xmath98 is @xmath99 ( in second homology ) .
we also want the @xmath90 points on @xmath89 to map injectively by @xmath97 to the @xmath69 s and to map to @xmath100 in the unit disk by @xmath98 .
similarly we want the @xmath4 points to map injectively to the @xmath101 s by @xmath97 and to @xmath102 in the unit disk by @xmath98 .
furthermore we also require the @xmath91 arcs in @xmath89 to map to @xmath29 arcs by @xmath97 and under @xmath98 to map to the arc @xmath103 in @xmath104 joining @xmath100 to @xmath102 in half - plane @xmath105 .
similarly we require the @xmath92 arcs to map to @xmath32 arcs by @xmath97 and to map to the arc @xmath106 in @xmath104 in the half - plane @xmath107 by @xmath98 .
now fix complex structures on @xmath30 and @xmath108 and take the product complex structure on @xmath95 . a generic perturbation
gives an almost complex structure which achieves transversality for the homology class @xmath76 . in our case
, we can achieve this by a generic perturbation of the @xmath29 curves and the @xmath32 curves ( ( * ? ? ?
* lemma 3.10 ) ) . the holomorphic embeddings @xmath109 which satisfy the above conditions and whose homology class is @xmath76 form a moduli space , which we denote by @xmath110 . the _ maslov index _
@xmath111 of @xmath76 gives the expected dimension of @xmath110 .
it can be computed combinatorially in terms of the euler measure and the point measures , which are defined as follows . for a generator @xmath112 and a domain @xmath76 , @xmath113
is defined to be the average of the coefficients of the four regions around @xmath69 in @xmath76 .
the _ point measure _ @xmath114 is defined as @xmath115 .
if we fix a metric on @xmath30 which makes all the @xmath29 and @xmath32 circles geodesic , intersecting each other with right angles , then the _ euler measure _ @xmath116 is defined to be @xmath117 of the integral of the curvature on @xmath76 .
the euler measure is clearly additive , and if @xmath118 is a @xmath119-gon region , then @xmath120 . for a domain @xmath121 ,
the maslov index is given by @xmath122 if @xmath76 is non - trivial , the moduli space @xmath110 admits a free @xmath123-action coming from the one - parameter family of holomorphic automorphisms of @xmath108 which preserve @xmath124 and the boundary arcs @xmath103 and @xmath106 .
in particular , if @xmath125 , the _ unparametrized moduli space _
@xmath126 is a zero - dimensional manifold , and then the _ count function _
@xmath127 is defined to be the number of points in @xmath126 , counted modulo @xmath72 . the boundary map in the chain complex @xmath67 is given by @xmath128 for a three - manifold @xmath4 , the homology of the chain complex @xmath129 is isomorphic to @xmath130 .
note that the only non - combinatorial part of the theory is the count function @xmath127 .
we will need the following proposition , which asserts that only positive domains can have holomorphic representatives .
let @xmath76 be a domain in @xmath131 .
if @xmath76 has a holomorphic representative , then @xmath76 is a positive domain . in particular ,
if @xmath132 , then @xmath76 is a positive domain .
if @xmath76 has a holomorphic representative , then there exists some holomorphic embedding @xmath109 of the type described above .
then for any point @xmath82 , @xmath79 is simply the intersection number of @xmath133 and @xmath134 .
since both of them are holomorphic objects in the product complex structure , they have positive intersection number and hence @xmath135 . here
we require the complex structure on @xmath95 to be standard near the basepoints .
see @xcite for a general discussion .
if a domain @xmath76 has a holomorphic representative , the number of branch points of @xmath98 is given by @xmath136 ( @xcite ) .
furthermore , in such a situation the maslov index can also be calculated as @xmath137 , where @xmath138 denotes the number of branch points of @xmath97 , and @xmath44 denotes the number of _ trivial disks _ ,
i.e. the components of @xmath7 which are mapped to a point by @xmath97 ( which correspond to coordinates @xmath69 of @xmath73 with @xmath139 ) .
in this section , we study index one holomorphic disks in nice heegaard diagrams .
[ defn : nicediagram ] let @xmath140 be a pointed heegaard diagram for a three - manifold @xmath4 .
@xmath141 is called * nice * if any region that does not contain any basepoint @xmath37 in @xmath142 is either a bigon or square .
let @xmath4 be a closed oriented three - manifold .
suppose @xmath4 has a nice admissible heegaard diagram @xmath140 .
we choose a product complex structure on @xmath95 .
[ defn : empty ] a domain @xmath143 with coefficients @xmath144 and @xmath145 is called an * empty embedded @xmath119-gon * , if it is topologically an embedded disk with @xmath119 vertices on its boundary , such that at each vertex @xmath146 , @xmath147 , and it does not contain any @xmath69 or @xmath101 in its interior .
the following two theorems show that , for a domain @xmath86 , the count function @xmath132 if and only if @xmath76 is an empty embedded bigon or an empty embedded square , and in that case @xmath148 .
thus @xmath127 can be computed combinatorially in a nice heegaard diagram .
[ thm : holo=>embedded ] let @xmath143 be a domain such that @xmath125 . if @xmath76 has a holomorphic representative , then @xmath76 is an empty embedded bigon or an empty embedded square .
we know that only positive domains can have holomorphic representatives .
we also know that bigons and squares have non - negative euler measure .
we will use these facts to limit the number of possible cases .
suppose @xmath149 , where @xmath150 s are regions containing no basepoints .
since @xmath76 has a holomorphic representative , we have @xmath151 . since each @xmath150 is a bigon or a square
, we have @xmath152 and hence @xmath153 .
so , by lipshitz formula @xmath154 , we get @xmath155
. now let @xmath156 and @xmath157 , with @xmath158 .
we say @xmath76 hits some @xmath29 circle if @xmath159 is non - zero on some part of that @xmath29 circle .
since @xmath160 , it has to hit at least one @xmath29 circle , say @xmath161 , and hence @xmath162 as @xmath163 .
also if @xmath76 does not hit @xmath46 , then @xmath164 and they must lie outside the domain @xmath76 , since otherwise we have @xmath165 and hence @xmath166 becomes too large .
we now note that @xmath116 can only take half - integral values , and thus only the following cases might occur . * * case 1 .
* @xmath76 hits @xmath161 and another @xmath29 circle , say @xmath167 , @xmath76 consists of squares , @xmath168 , and there are @xmath169 trivial disks . * * case 2 . *
@xmath76 hits @xmath161 , @xmath170 consists of squares and exactly one bigon , @xmath171 , and there are @xmath172 trivial disks . * * case 3 . *
@xmath76 hits @xmath161 , @xmath170 consists of squares , @xmath173 , and there are @xmath172 trivial disks . using the reformulation by lipshitz , in each of these cases , we will try to figure out the surface @xmath7 which maps to @xmath95 . recall that a trivial disk is a component of @xmath7 which maps to a point in @xmath30 after post - composing with the projection @xmath174 .
the first case corresponds to a map from @xmath7 to @xmath30 with @xmath175 , and @xmath7 has @xmath169 trivial disk components . if the rest of @xmath7 is @xmath176 , then @xmath176 is a double branched cover over @xmath108 with @xmath177 and @xmath145 branch point ( for holomorphic maps , the number of branch points is given by @xmath178 ) , i.e. @xmath176 is a disk with @xmath179 marked points on its boundary .
call the marked points corners , and call @xmath176 a square . in the other two cases , @xmath7 has @xmath172 trivial disk components , so if @xmath176 denotes the rest of @xmath7 , then @xmath176 is just a single cover over @xmath108 .
thus the number of branch points has to be @xmath144 .
but in the third case the number of branch points is @xmath145 , so the third case can not occur . in the second case
, @xmath176 is a disk with @xmath72 marked points on its boundary .
call the marked points corners , and call @xmath176 a bigon . thus in both the first and the second cases ,
@xmath76 is the image of @xmath176 and all the trivial disks map to the @xmath180-coordinates ( which are also the @xmath181-coordinates ) which do not lie in @xmath76 .
note that in both cases , the map from @xmath176 to @xmath76 has no branch point , so it is a local diffeomorphism , even at the boundary of @xmath176 . furthermore using the condition that @xmath182 ( or @xmath183 ) @xmath184 whenever it is non - zero
, we conclude that there is exactly one preimage for the image of each corner of @xmath176 .
all we need to show is that the map from @xmath176 to @xmath30 is an embedding , or in other words , the local diffeomorphism from @xmath176 to @xmath76 is actually a diffeomorphism .
we will prove this case by case .
in this case we have an immersion @xmath185 , where @xmath176 is a square ( with boundary ) .
look at the preimage of all the @xmath29 and @xmath32 circles in @xmath176 .
using the fact @xmath20 is a local diffeomorphism , we see that each of the preimages of @xmath29 and @xmath32 arcs are also @xmath145-manifolds , and by an abuse of notation , we will also call them @xmath29 or @xmath32 arcs . using the embedding condition near the @xmath179 corners
, we see that at each corner only one @xmath29 arc and only one @xmath32 arc can come in .
the different @xmath29 arcs can not intersect and the different @xmath32 arcs can not intersect , and all intersections between @xmath29 and @xmath32 arcs are transverse . note that since the preimage of each square region is a square , @xmath176 ( with all the @xmath29 and @xmath32 arcs ) is also tiled by squares .
thus the @xmath29 arcs in @xmath176 can not form a closed loop , for in that case @xmath186\{inside of loop } has negative euler measure and hence can not be tiled by squares .
similarly the @xmath32 arcs can not form a loop . also no @xmath29 arc can enter and leave @xmath176 through the same @xmath32 arc on the boundary , for again the outside will have negative euler measure .
thus the @xmath29 arcs slice up @xmath176 into vertical rectangles , and in each rectangle , no @xmath32 arc can enter and leave through the same @xmath29 arc .
this shows that the @xmath29 arcs and @xmath32 arcs make the standard co - ordinate chart on @xmath176 , as in figure [ fig : holo_square ] .
we call the intersection points between @xmath29 and @xmath32 arcs in @xmath176 vertices ( and we are still calling the four original vertices on the boundary of the square @xmath176 corners ) .
note that to show @xmath20 is an embedding , it is enough to show that no two different vertices map to the same point .
assume @xmath187 are distinct vertices with @xmath188 .
there could be two subcases .
* both @xmath80 and @xmath189 are in @xmath190 .
* at least one of @xmath80 and @xmath189 is in @xmath191 .
we will reduce the first subcase to the second . assume both @xmath80 and @xmath189 are in the interior of @xmath176 . choose a direction on the @xmath29 arc passing through @xmath188 in @xmath30 , and keep looking at successive points of intersection with @xmath32 arcs , and
locate their inverse images in @xmath176 .
for each point , we will get at least a pair of inverse images , one on the @xmath29 arc through @xmath80 , and one on the @xmath29 arc through @xmath189 , until one of the points falls on @xmath191 , and thus we have reduced it to the second subcase . in the second subcase , without loss of generality , we assume that @xmath80 lies on a @xmath32 arc on @xmath191 . then choose a direction on the @xmath32 arc in @xmath30 through @xmath188 and proceed as above , until one of the preimages hits an @xmath29 arc on @xmath191 . if that preimage is on the @xmath32 arc through @xmath189 , then reverse the direction and proceed again , and this time we can ensure that the preimage which hits @xmath29 arc on @xmath191 first is the one that was on the @xmath32 arc through @xmath80 . thus we get @xmath72 distinct vertices in @xmath176 mapping to the same point in @xmath30 , one of them being a corner .
this is a contradiction to the embedding assumption near the corners . in this case
we have an immersion @xmath185 with @xmath176 being a bigon .
again look at the preimage of @xmath29 and @xmath32 circles .
all intersections will be transverse ( call them vertices ) , and at each of the @xmath72 corners there can be only one @xmath29 arc and only one @xmath32 arc .
again there can not be any closed loops .
we get an induced tiling on @xmath176 with squares and @xmath145 bigon .
this time the @xmath29 arcs can ( in fact they have to ) enter and leave @xmath176 through the same @xmath32 arc , but they have to do it in a completely nested fashion , i.e. , there is only one bigon piece in @xmath192 , the innermost bigon " .
thus @xmath176 decomposes into two pieces , the innermost bigon and the rest . in case
there are no @xmath29 arcs in @xmath190 , the rest might be empty , but otherwise it is a square . from the arguments in the earlier case , the @xmath32 arcs must cut up the square piece in a standard way , and from the previous argument the @xmath32 arcs must enter and leave the bigon in a nested fashion , as in figure [ fig : holo_bigon ] . again to show @xmath20 is an embedding , it is enough to show that it is an embedding restricted to vertices .
take @xmath72 distinct vertices @xmath193 mapping to the same point , and follow them along @xmath29 arcs in some direction , until one of them hits a @xmath32 arc on @xmath191 .
then follow them along @xmath32 arcs , and there exists some direction such that one of them will actually hit a corner , giving the required contradiction .
so in either case , @xmath20 is an embedding .
[ thm : embedded=>holo ] if @xmath143 is an empty embedded bigon or an empty embedded square , then the product complex structure on @xmath95 achieves transversality for @xmath76 under a generic perturbation of the @xmath29 and the @xmath32 circles , and @xmath194 .
let @xmath76 be an empty embedded @xmath119-gon .
each of the corners of @xmath76 must be an @xmath180-coordinate or a @xmath181-coordinate , and at every other @xmath180 ( resp .
@xmath181 ) coordinate the point measure @xmath182 ( resp .
@xmath183 ) is zero .
therefore @xmath195 .
also @xmath76 is topologically a disk , so it has euler characteristic @xmath145 . since it has @xmath119 corners each with an angle of @xmath196 , the euler measure @xmath197 .
thus the maslov index @xmath125 . by (
* lemma 3.10 ) , we see that @xmath76 satisfies the boundary injective condition , and hence under a generic perturbation of the @xmath29 and the @xmath32 circles , the product complex structure achieves transversality for @xmath76 .
when @xmath76 is an empty embedded square , we can choose @xmath176 to be a disk with @xmath179 marked points on its boundary , which is mapped to @xmath76 diffeomorphically . given a complex structure on @xmath30 , the holomorphic structure on @xmath176 is determined by the cross - ratio of the four points on its boundary , and there is an one - parameter family of positions of the branch point in @xmath108 which gives that cross - ratio .
thus there is a holomorphic branched cover @xmath198 satisfying the boundary conditions , unique up to reparametrization .
hence @xmath76 has a holomorphic representative , and from the proof of theorem [ thm : holo=>embedded ] we see that this determines the topological type of @xmath176 , and hence it is the unique holomorphic representative . when @xmath76 is an empty embedded bigon , we can choose @xmath176 to be a disk with @xmath72 marked points on its boundary , which is mapped to @xmath76 diffeomorphically .
a complex structure on @xmath30 induces a complex structure on @xmath176 , and there is a unique holomorphic map from @xmath176 to the standard @xmath108 after reparametrization .
thus again @xmath76 has a holomorphic representative , and similarly it must be the unique one .
theorems [ thm : holo=>embedded ] and [ thm : embedded=>holo ] make the count function @xmath127 combinatorial in a nice heegaard diagram . for a domain @xmath143 with @xmath125
, we have @xmath148 if @xmath76 is an empty embedded bigon or an empty embedded square , and @xmath199 otherwise .
in this section , we prove theorem [ thm : nicediagram ] . we will demonstrate an algorithm which , starting with an admissible pointed heegaard diagram , gives an admissible nice heegaard diagram by doing isotopies and handleslides on the @xmath32 curves .
for a heegaard diagram , we call bigon and square regions _ good _ and all other regions _ bad_. we will first do some isotopies to ensure all the regions are disks .
we will then define a complexity for the heegaard diagram which attains its minimum only if all the regions not containing the basepoints are good .
we will do an isotopy or a handleslide which will decrease the complexity if the complexity is not the minimal one .
let @xmath200 be a pointed heegaard diagram with a single basepoint @xmath201 .
we consider heegaard diagrams with more basepoints in the last subsection .
we do finger moves on @xmath32 circles to create new intersections with @xmath29 circles . after doing this sufficiently many times , every region in @xmath141 becomes a disk .
we first ensure that every @xmath29 circle intersects some @xmath32 circle and every @xmath32 circle intersects some @xmath29 circle .
if @xmath46 does not intersect any @xmath32 circle , we can find an arc @xmath202 connecting @xmath46 to some @xmath203 avoiding the intersections of @xmath29 and @xmath32 circles , as indicated in figure [ fig : no_isolate_alpha](a ) .
we can select @xmath202 such that @xmath202 intersects @xmath204 just at the endpoint .
doing a finger move of @xmath203 along @xmath202 as in figure [ fig : no_isolate_alpha](b ) will make @xmath46 intersect some @xmath32 circle .
similarly , if @xmath205 does not intersect any @xmath29 circle , we find an arc @xmath202 connecting @xmath205 to some @xmath47 so that @xmath206 contains a single point as in figure [ fig : no_isolate_beta](a ) .
we then do the operation as depicted in figure [ fig : no_isolate_beta](b ) . repeating the above process
, we can make sure that every @xmath29 circle intersects some @xmath32 circle and every @xmath32 circle intersects some @xmath29 circle .
note that the complement of the @xmath29 curves is a punctured sphere .
thus every region is a planar surface .
a non - disk region @xmath118 has more than one boundary component .
every boundary component must contain both @xmath29 and @xmath32 arcs since every @xmath29 ( resp .
@xmath32 ) circle intersects some @xmath32 ( resp .
@xmath29 ) circle .
then we make a finger move on the @xmath32 curve to reduce the number of boundary components of @xmath118 without generating other non - disk regions .
see figure [ fig : kill_non_disk ] for this finger move operation . repeating this process as many times as necessary , we will kill all the non - disk regions .
we consider heegaard diagrams with only disk regions .
note that our algorithm will not generate non - disk regions .
let @xmath207 be the disk region containing the basepoint @xmath201 .
for any region @xmath118 , pick an interior point @xmath208 and define the _ distance _ of @xmath118 , denoted by @xmath209 , to be the smallest number of intersection points between the @xmath32 curves and an arc connecting @xmath201 and @xmath210 in the complement of the @xmath29 circles . for a @xmath119-gon disk region @xmath118 ,
define the _ badness _ of @xmath118 as @xmath211 . for a pointed heegaard diagram @xmath141 with only disk regions , define the _ distance _ @xmath212 of @xmath141 to be the largest distance of bad regions .
define the _ distance @xmath213 complexity _ of @xmath141 to be tuple @xmath214 where @xmath215 are all the distance @xmath213 bad regions , ordered so that @xmath216 .
we call the first term the _ total badness of distance @xmath213 _ of @xmath141 , and denote it by @xmath217 . if there are no distance @xmath213 bad regions , then @xmath218 .
we order the set of distance @xmath213 complexities lexicographically . for a distance
@xmath213 pointed heegaard diagram @xmath141 with only disk regions , if @xmath219 , we can modify @xmath141 by isotopies and handleslides to get a new heegaard diagram @xmath220 with only disk regions , satisfying @xmath221 and @xmath222 .
we order the bad regions of distance @xmath213 as in the definition of the distance @xmath213 complexity .
now we look at @xmath223 .
it is a @xmath224-gon with @xmath225 .
pick an adjacent region @xmath226 with distance @xmath227 having a common @xmath32 edge with @xmath223 .
let @xmath228 be ( one of ) their common @xmath32 edge(s ) .
we order the @xmath29 edges of @xmath223 counterclockwise , and denote them by @xmath229 starting at @xmath228 .
we try to make a finger move on @xmath228 into the @xmath223 and out of @xmath223 through @xmath230 , as indicated in figure [ fig : finger_move_start ] when @xmath223 is an octagon .
our finger will separate @xmath223 into two parts , @xmath231 and @xmath232 .
if we reach a square region of distance @xmath233 , we push up our finger outside the region via the opposite edge , as in figure [ fig : finger_move_square ] . note that doing a finger move through regions of distance @xmath233 does not change the distance of any of the bad regions , since they all have distance @xmath234 .
we continue to push up our finger as far as possible , until we reach one of the following : 1 . a bigon region .
2 . a region with distance @xmath235 .
3 . a bad region with distance @xmath213 other than @xmath223 , i.e. , @xmath150 with @xmath236 .
4 . @xmath223 .
we will prove our lemma case by case .
before we reach the bigon region , all regions in between are square regions with distance @xmath233 .
after our finger moves inside a bigon region , our finger separates the bigon into a square and a new bigon , as in figure [ fig : finger_move_bigon ] . denote the new heegaard diagram by @xmath220 .
we have @xmath237 .
since @xmath232 is a square and is good , we get @xmath238 .
note that we will not increase the distance of any bad region since we do not pass through any region of distance @xmath235 and all bad regions has distance @xmath234 .
hence @xmath221 and @xmath222 .
let @xmath239 be the region with distance @xmath240 we reached by our finger .
suppose @xmath241 .
let @xmath220 be the new heegaard diagram .
see figure [ fig : finger_move_smaller_dist_region ] .
note that @xmath239 might be a bigon , which could be covered in both case 1 and case 2 .
we have @xmath237 and @xmath232 is good .
our finger separates @xmath239 into a bigon region @xmath242 and the other part @xmath243 .
when @xmath239 is a square or a bad region , @xmath243 will be a bad region of distance @xmath244 .
we might have increased the distance @xmath245 complexity , but we have @xmath221 and @xmath222 . in this case , we reach some distance @xmath213 bad region @xmath150 with @xmath236
. see figure [ fig : finger_move_brother_region ] for an indication .
denote by @xmath246 and @xmath247 the two parts of @xmath150 separated by our finger .
then @xmath246 is good while @xmath247 is a bad region of distance @xmath213 .
we have @xmath248 and @xmath249 .
thus the total badness of distance @xmath213 remains the same .
but we are decreasing the distance @xmath213 complexity since we are moving the badness from a later bad region to an earlier bad region .
hence for the new heegaard diagram @xmath220 , we have @xmath250 and @xmath222 .
this is the worst case and we need to pay more attention .
we divide this case into two subcases , according to which edge the finger is coming back through .
this subcase is indicated in figure [ fig : finger_move_back_adjacent ] . without loss of generality
, we assume the finger comes back via @xmath251 . in this case
, we see the full copy of some @xmath32 curve , say @xmath205 , one the right side along our long finger .
suppose @xmath252 .
note that @xmath253 since otherwise @xmath254 and we will reach either @xmath223 or @xmath226 at an earlier time .
now instead of doing the finger move , we handleslide @xmath203 over @xmath255 .
this is indicated in figure [ fig : finger_move_back_adjacent_handleslide ] .
note that after the handle slides , we are not increasing the distance of any bad region .
we have increased the badness of @xmath226 , but it is a distance @xmath227 region .
@xmath232 is a bigon region and @xmath237 .
thus for the new heegaard diagram @xmath220 after the handleslide , the total badness of distance @xmath213 is decreased by 1 .
we have @xmath221 and @xmath222 .
if we return through @xmath256 with @xmath257 , then , instead of the finger move through @xmath230 , we do a finger move through @xmath258 ( starting from @xmath228 ) .
if we reach one of the first three cases , we are decreasing the distance @xmath213 complexity by similar arguments as before .
suppose instead that we come back to @xmath223 , say via @xmath259 .
we claim that @xmath260 .
certainly we can not come back via @xmath258 .
the finger can not come back via @xmath256 since the chain of squares from @xmath256 is connected to @xmath230 .
if @xmath261 or @xmath262 , we could close the cores the two fingers to get two simple closed curves @xmath263 and @xmath264 , as indicated in figure [ fig : no_crossing_finger ] . then @xmath263 and @xmath264 intersect transversely at exactly one point and they are in the complement of the @xmath32 curves .
the complement of the @xmath32 curves is a punctured sphere .
attach disks to get a sphere . then as homology classes , we get @xmath265\cdot[c_2]=1 $ ] .
but @xmath266 . this is a contradiction .
thus we must have @xmath260 .
( the argument of this claim was suggested by dylan thurston . )
now , instead of the finger move through @xmath258 , we do another finger move through @xmath267 . continuing the same arguments , we see that we either end up with a finger which does not come back , or we get some finger that starts at @xmath268 and comes back via @xmath269 .
if the finger does not come back , we reduce it to the previous cases and the lemma follows .
if there is a finger which starts at @xmath268 and comes back at @xmath269 , we see a full @xmath32 circle .
we do a handleslide similar to the one in subcase 4.1 .
we have @xmath270 and @xmath271 .
we also have @xmath272 .
thus for the new heegaard diagram @xmath220 after the handleslide , the total badness of distance @xmath213 decreases .
we have @xmath221 and @xmath222 .
thus we end the proof of our lemma .
repeat this process to make @xmath273 . repeating the whole process sufficiently many times
will eventually kill all the bad regions other than @xmath207 . in this subsection , we show that our algorithm will not change the admissibility , that is , if we start with an admissible heegaard diagram , then our algorithm ends with an admissible heegaard diagram .
there are two operations involved in our algorithm : isotopies and handleslides , and we will consider them one by one .
the isotopy is the operation in figure [ fig : admissible_isotopy ] .
let @xmath141 and @xmath220 be the heegaard diagrams before and after the isotopy .
suppose @xmath141 is admissible .
for a periodic domain in @xmath220 @xmath274 we have @xmath275 and @xmath276 .
hence @xmath277 and @xmath278 .
note that the regions are all the same except those in figure [ fig : admissible_isotopy ] .
therefore , @xmath279 is a periodic domain for @xmath141 . since @xmath141 is admissible , @xmath76 has both positive and negative coefficients , and so does @xmath280 .
hence @xmath220 is admissible .
our handleslide operation is indicated in figure [ fig : admissible_handleslide ] .
suppose @xmath141 is admissible .
for a periodic domain in @xmath220 @xmath281 we get @xmath282 .
suppose @xmath283 , then @xmath284 and @xmath285 .
now @xmath286 is a periodic domain for @xmath141 . since @xmath141 is admissible , @xmath76 has both positive and negative coefficients .
hence @xmath280 has both positive and negative coefficients , so @xmath220 is admissible
. * remark .
* in fact , it can be shown that nice heegaard diagrams are always ( weakly ) admissible ( ( * ? ? ?
* corollary 3.2 ) ) .
we have similar conclusions for heegaard diagrams with multiple basepoints .
our algorithm could be modified to get nice heegaard diagrams in that case .
note that every region is connected to exactly one region containing some @xmath201 point in the complement of the @xmath29 curves , so we can define the distance and hence the complexity in the same way , and thus our algorithm works as before .
starting with an admissible one - pointed heegaard diagram , our algorithm described in section 4.1 gives an admissible heegaard diagram with only one bad region , the one containing the basepoint @xmath201 .
the algorithm can be modified for multiple basepoints as described above .
in this section , we give two examples to demonstrate our algorithm .
one is on knot floer homology and the other is on the heegaard floer homology of three - manifolds .
|
in this paper , we give an algorithm to compute the hat version of heegaard floer homology of a closed oriented three - manifold .
this method also allows us to compute the filtration coming from a null - homologous link in a three - manifold .
|
modeling large - scale networks has been stirring much developments in various fields such as machine learning and system identification .
the assembly of numerous systems interacting with one another arises in fields such as biology , e.g with the brain neurons in @xcite , optics with the atmospheric turbulence , and many others . due to the large size of input - output data batches , identifying locally
the behavior of the network is a major challenge that has been mainly addressed by using prior knowledge on how the subsystems , or nodes , are connected to one another .
one common assumption is sparsity and relies on the fact that each node is connected to a limited number of other nodes with respect to the network s size .
other well - studied structures include interconnected one - dimensional strings of subsystems in @xcite , or clusters of different subsystems with known connection patterns , named as alpha - heterogeneous in @xcite .
however the links between the subsystems in the network might not be known beforehand . in the so - called sparse plus low rank networks
-@xcite- , a few latent variables relate most of the measured nodes from which few of them influence each other .
model identification remains computationally challenging to handle the combination of these two matrices structure . in @xcite
very general graphs are studied , whose weighted adjacency matrix is approximated with kronecker products of a so - called initiator matrix .
such an operation replicates the network structure associated to the initiator matrix to higher dimensions .
therefore this matrix embeds all the required information on the network to construct it at higher scales .
it is moreover shown in @xcite that any network is approximated with such a structure after having ordered the nodes in the most adequate way . in this paper
we address the identification of ( 2d ) spatial - temporal dynamical models of the vector - auto - regressive ( var ) form .
the coefficient - matrices of this model are parametrized as sums of kronecker products .
@xcite establishes the equivalence between expressing a matrix as a sum containing few kronecker products and a low - rank approximation of a reshuffled matrix .
the latter has been studied in @xcite for function - related matrices , in which more general existence theorems are derived . moreover ,
when the matrix exhibits multiple symmetries or is block - toeplitz with toeplitz - blocks , it is guaranteed that such a low rank approximation of the reshuffled matrix exists , see @xcite and @xcite .
more than only enjoying the storage of a reduced number of entries , such a structure enables fast computations thanks to the very pleasant algebra of the kronecker product , @xcite . for a matrix written as @xmath0 with @xmath1 ,
matrix - matrix multiplication and inversion both require @xmath2 instead of @xmath3 for the unstructured case . for a given labeling of the network ,
the coefficient - matrices in the varx model may be represented with as few terms as possible in the kronecker sum while still guaranteeing a given level of performance with respect to the standard least squares solution with unstructured coefficient matrices .
a major challenge in the estimation is the computational efficiency .
we address this problem by parametrizing only the factor matrices , e.g @xmath4 , which gives rise to a bilinear least squares problem .
the estimation problem belongs to the class of multi - convex optimization : fixing all variables but one yields a convex problem , @xcite .
such a formulation shares similarities with the identification of hammerstein systems , see e.g @xcite for which a two - stage algorithm is proposed .
the contribution of this work includes the formulation of a kronecker - based varx model for networks with unknown communication links .
the estimation problem is solved using both a non - iterative three - stage algorithm and an iterative alternating least squares , for which we have extended a convergence proof to our case .
thirdly , the missing sensor data can be retrieved by formulating a multi - linear least squares .
this paper is organized as follows .
in the second section the class of _ kronecker networks _ is defined .
the third section formulates the kronecker varx identification framework while the fourth section describes the non - iterative overparametrized algorithm to estimate the factor matrices with minimum computational complexity .
an alternative is proposed in section v within the framework of multi - linear least squares .
section vi describes the missing data case .
this covers the case of non - rectangular measurement grids .
we study how additional structure can be considered on the factor matrices in section vii .
last , section viii is dedicated to numerical experiments .
_ notations . _
the vectorization operator for a matrix @xmath5 is @xmath6 .
@xmath7 reshapes the vector @xmath8 into a matrix whose size will be clear from the context .
the kronecker product between two matrices @xmath9 is denoted by @xmath10 . , the 2-norm of a vector @xmath8 is written as @xmath11 .
the number of non - zero - elements in a vector @xmath8 is @xmath12 while the sum in absolute value is @xmath13 .
@xmath14 is the largest eigenvalue of the positive - semidefinite matrix @xmath5 .
the nuclear norm of @xmath5 , denoted with @xmath15 , represents the sum of the singular values of @xmath5 .
[ @xcite ] [ def : def1 ] let @xmath5 be a @xmath16 block matrix with blocks @xmath17 in @xmath18 , given as : @xmath19 \in \mathbb{r}^{m_1 m_2 \times n_1 n_2}\ ] ] then the re - shuffle operator @xmath20 is defined as : @xmath21 \in \mathbb{r}^{m_1 n_1 \times m_2 n_2}\ ] ] reshuffling @xmath22 to form @xmath5 back is defined with the operator @xmath23 , i.e @xmath24 .
[ @xcite ] [ lem : kron ] let @xmath5 be defined as in definition [ def : def1 ] , and let @xmath25 , with @xmath26 . then
: @xmath27 the operation in lemma [ lem : kron ] can also be reversed by the definition of the inverse vec operator @xmath28 .
[ @xcite ] [ lem : skron ] let @xmath5 be defined as in definition [ def : def1 ] , and let an svd of @xmath20 be given as : @xmath29 and let @xmath30 , @xmath31 : @xmath32 the integer @xmath33 is called the _ kronecker rank _ of @xmath5 w.r.t .
the chosen block partitioning of @xmath5 as given in definition [ def : def1 ] .
[ def : adecomp ] let @xmath34 be a @xmath35 pattern matrix .
define @xmath36 ( with @xmath37 ) and @xmath38}$ ] as an @xmath35 diagonal matrix which contains @xmath39 in the diagonal entries of indices from @xmath40 to @xmath41 ( included ) and @xmath42 elsewhere , then an @xmath43-decomposible matrix ( for a given @xmath43 ) is a matrix of the following kind : @xmath44 } \otimes m_a^{(i ) } ) + \sum_{i=1}^\alpha ( i_{[\beta_{i-1}+1:\beta_i ] } p \otimes m_b^{(i ) } ) \ ] ] the matrices @xmath45 are the diagonal blocks of @xmath46 , while the matrices @xmath47 constitute the off - diagonal blocks , according to the structure of @xmath48 . for @xmath49 ( and @xmath50 ) , these matrices are simply called _
decomposable _ matrices .
the class of @xmath43-decomposable matrices will be denoted by @xmath51 , with for @xmath52 just the symbol @xmath53 will be used . as a generalization of this class of structured matrices ,
we define next the class of sums of kronecker product matrices .
[ def : skron ] the class of sums of kronecker product matrices , denoted by @xmath54 , contains matrices of the following kind : @xmath55 with @xmath56 and @xmath57
the sensor readings at time instance @xmath58 are stored in the matrix @xmath59 as : @xmath60\ ] ] with @xmath61 . in this paper
we will consider that the ( temporal ) dynamics of this array of sensors is governed by the following var(x ) model : @xmath62 with @xmath63 zero - mean white noise with covariance matrix @xmath64 .
the coefficient matrices @xmath65 and @xmath66 in the varx ( we will restrict for simplicity to the ar - case ) model , in general are highly structured . here
we will consider these coefficient matrices to be in the matrix sets @xmath67 or @xmath54 .
we will consider the case they belong to the set @xmath54 , and for the moment only focus on the coefficient matrices @xmath65 . to address an identification problem
we will parametrize these coefficient matrices as : @xmath68 with the vectors @xmath69 and @xmath70 parametrizing the matrices @xmath71 and @xmath72 in an affine manner .
if we consider the term @xmath73 as a temporally white sequence @xmath74 , then the arx model can be written as , @xmath75 using the following kronecker rule , for matrices @xmath76 of compatible dimensions such that the product @xmath77 exists , @xmath78 we can write the arx model as , @xmath79 with @xmath80 .
this can also be written explicitly as , @xmath81 \bigl(i_{r_i } \otimes s(k - i)\bigr ) \left [ \begin{matrix } m(b_i^{(1 ) } ) \\ \vdots \\
m(b_i^{(p ) } ) \end{matrix } \right ] \nonumber \\ & & + v(k)\end{aligned}\ ] ] the ar(x ) models , or are called _ kronecker arx network models _ , or briefly _ kronecker arx models _ ( krarx ) ( pronounced as `` _ _ quarks _ _ '' ) .
given the model structure of krarx models , the problem of identifying these models from measurement sequences @xmath82 is fourfold : 1 .
the temporal order index @xmath83 .
the spatial order index @xmath84 for each coefficient matrix .
3 . the parametrization of the matrices @xmath71 and @xmath72 .
an example of a parametrization of the matrices @xmath71 and @xmath72 is ( block ) toeplitz .
the estimation of the parameter vectors @xmath69 , @xmath70 up to an ambiguity transformation .
this requires the specification of a cost function .
an example of such a cost function using the model is the following least squares cost function , @xmath85 for data batches with @xmath86 temporal samples . by the selection of the parameter @xmath83 and particular choices of the parametrizations in step 3 above , various special cases of restricting the coefficient matrices @xmath65 in to particular sets ( such as @xmath67 or @xmath54 ) can be considered .
further constraints to the ( least - squares ) cost function , such as in , might be introduced to look for sparsity in the parametrization vectors @xmath69 and @xmath70 .
an important challenge of the parameter estimation problem is the _
computational efficiency_. the covariance matrix estimation in high dimensional spaces has already been addressed in @xcite and is not considered further on in this document .
consider the krarx model then we can define the matrix @xmath87 with @xmath88 and write this model as : @xmath89 according to and the definition of the re - shuffling operator @xmath90 we have , @xmath91 therefore a way to find the spatial order index ( assuming it is the same for all @xmath87 ) is via the kronecker rank .
let this be denoted by @xmath33 , then we write , @xmath92 and with the definition of @xmath93 , @xmath94 , the coefficient matrix @xmath87 can be written as , @xmath95 not knowing the kronecker rank , a possible way to retrieve this parameter and the coefficient matrices @xmath87 for a given temporal order @xmath83 from the data is via the following multi - criteria cost function : @xmath96 let the estimated coefficient matrices be denoted by @xmath97 , then subsequently an svd of the matrices @xmath98 provides estimates of the terms @xmath99 , @xmath100 and @xmath101 in the kronecker products in .
it should be remarked that this way of formulating the identification problem does not require a parametrization of the matrices @xmath71 and @xmath72 as stipulated in the third step of the fourfold generic identification problem formulation outlined in section [ sec_idep ] .
since the rank operator in the cost function turns this cost function into a non - convex optimization problem , the nuclear norm can be used to convexify this problem .
this would then yield the following problem formulation , @xmath102 where an additional weighting parameter @xmath103 is introduced .
the nuclear norm regularization on @xmath83 matrices of size @xmath104 is prohibitive especially when handling large datasets .
it would indeed imply solving e.g a alternating direction method of multipliers algorithm ( admm ) with a singular - value decomposition on large matrices at each iteration or expensive matrices inversions scaling up to @xmath3 .
however we see in the following section that some more efficient computations can be performed by parallelizing the optimization problem . in this paragraph
we consider @xmath105 for the sake of clarity .
the least squares term in the cost function can be addressed row by row .
the @xmath106-th line of @xmath107 , where @xmath108 , and @xmath109 integers , is denoted with @xmath110 .
using `` standard '' matlab notation to select part of a matrix , the matrix @xmath111 and @xmath112 are related as : @xmath113 to further clarify this notation , we refer to figure 1 for a display of the different matrices in the above relation . and @xmath114.,scaledwidth=40.0% ]
we state the following lemma : if @xmath115 , then for all @xmath116 .
let the pca decomposition of @xmath114 be such that : @xmath117 where @xmath118 . from , @xmath119 .
the vectors @xmath120 are not necessarily full column rank , therefore @xmath121 .
the reverse implication is not true .
denote an upper - bound on the rank of @xmath114 by @xmath122 . by assuming that the pca decomposition of @xmath123 is given with @xmath124 and @xmath125 , a low - rank matrix @xmath114 is built using .
we describe the algorithm in the following lines .
let the measurement at position @xmath126 in the data matrix @xmath59 correspond to the @xmath106-th entry of the vector @xmath127 . then a rank - constrained least squares optimization
is formulated to estimate the matrix @xmath128 without over - fitting : @xmath129 which is solved using the relaxed problem : @xmath130 a pca of the low - rank matrix @xmath131 yields the following decomposition : @xmath132 where @xmath133 .
this decomposition is unique up to a ( non - singular ) ambiguity transformation @xmath134 : @xmath135 therefore such a pca can not be performed for @xmath136 independent well - chosen rows , as it would yield @xmath136 different ambiguity transformations .
there remains @xmath137 matrices of size @xmath138 to estimate the full factor matrices @xmath139 .
the latter have to be consistent with the estimation in and consider the same ambiguity transformation . therefore ,
if is solved e.g for @xmath140 , then for all @xmath141 $ ] , we solve the constrained least - squares optimization : @xmath142 and for all @xmath106 such that @xmath143 , where @xmath144 $ ] : @xmath145 these @xmath137 least squares can be performed in parallel , each of which corresponds to one sensor location as can be visualized in figure 2 .
choosing the sensor location in position @xmath146 in is not unique . which data to consider in order to estimate the coefficient - matrices with minimum computational complexity .
blue entry : minimization of type - .
green entry : minimization of type .
yellow entry : minimization of type .,scaledwidth=20.0% ] the three - step algorithm has been discussed for the case @xmath147 . in most cases ,
the kronecker rank is not known _ a priori _ and it has to be detected with cross - validation .
therefore we analyze how to deal with higher kronecker ranks , e.g @xmath148 . in the previous paragraph ,
the kronecker rank was limited by the size of the submatrices @xmath149 .
therefore a submatrix of size @xmath150 shall be selected such that the pca in is then carried out on a rank - deficient matrix .
for example , the output data associated with the set of indices @xmath151 conveys enough information to retrieve the factor matrices in this case .
figure 3 illustrates the sensor locations .
the regularized least squares in is then extended into : @xmath152 where the rank - deficient matrix is : @xmath153 similarly , the pca is performed on @xmath154 and the least squares in and are formulated for 4 neighboring points . and @xmath155 .
the color code is the same as in figure 2.,scaledwidth=20.0% ] the computational complexity is reduced exploiting the kronecker structure , and is attractive for being non - iterative and parallelizable to a large extent .
although it is possible to enforce some additional structure , e.g block - banded with banded - blocks , the framework is nonetheless less elegant than the alternating least squares minimization that we propose in the next paragraph to embed more structure on the factor matrices .
this approach is formulated based on the representation .
this forms has the advantage that constraints on the parametrizations of the matrices @xmath71 and @xmath72 can be more easily taken into consideration .
for example the parametrization may enforce the matrices to have a ( block ) toeplitz structure with entries in a particularly chosen set , e.g. @xmath156 .
denote this set for the parameter vectors @xmath69 , @xmath70 resp .
by @xmath157 and @xmath158 and let us assume @xmath83 and @xmath33 to be given .
denote moreover : @xmath159 \\
\tilde m_{b_i } & : = & \left [ \begin{matrix } m(b_i^{(1 ) } ) \\ \vdots \\
m(b_i^{(p ) } ) \end{matrix } \right ] \end{aligned}\ ] ] then we have the following bi - convex optimization problem @xmath160 here again we consider @xmath161 for the sake of clarity .
the krarx model then reads : @xmath162 such that @xmath163 .
the @xmath164-th column of @xmath165 , resp .
@xmath166 , is denoted with @xmath167 , resp .
@xmath84 . given an initial guess of the matrix @xmath165 denoted as @xmath168 ,
the following three steps are performed : 1 . _
step 1 : _ a minimization step over the columns of the matrix @xmath166 : @xmath169 2 .
_ step 2 : _ the estimates @xmath170 are normalized : @xmath171 3 .
_ step 3 : _ a minimization step over the columns of the matrix @xmath165 : @xmath172 in this section we study the convergence of the alternating least squares repeating the three steps in - .
the proof relies on the work in @xcite that establishes the result when @xmath173 are vectors .
therefore we review here the main theorems of the proof , and highlight the changes in appendix .
we first reformulate the equation . it can be shown that : @xmath174 where @xmath175 @xmath176 is defined similarly as @xmath177 .
@xcite analyse the convergence of the alternating least squares solution using the contraction mapping theorem , @xcite .
let @xmath178 denote the iteration counter .
let the initial estimate of @xmath179 be denoted by @xmath180 and denote the results of the subsequents three steps ( 1 to 3 ) be denoted resp .
@xmath181 then a _ functional _ representation of the three steps reads : @xmath182 these equations can be expressed using a single operator @xmath183 mapping the estimate @xmath180 to @xmath184 : @xmath185 _ [ the contraction mapping theorem , @xcite ] _ let @xmath186 be a non - empty complete metric space where @xmath187 is a metric on @xmath5 .
let @xmath188 be a contraction mapping on @xmath5 , i.e. , there is a nonnegative real number @xmath189 such that @xmath190 , for all @xmath191 .
then the map @xmath192 admits one and only one fixed point @xmath193 which means @xmath194 .
furthermore , this fixed point can be found from the convergence of an iterative sequence defined by @xmath195 for @xmath196 with an arbitrary starting point @xmath197 in @xmath5 .
we start by defining the following inner product .
[ def : x ] let @xmath198 and denote their columns with @xmath199 . for two matrices @xmath200 of conformable sizes , such that @xmath201 , similarly for @xmath202 , the inner product on @xmath203 is _ defined _ with : @xmath204 the matrices @xmath205 have the structure of the matrices @xmath200 in the above definition .
[ lem10 ] for the matrix @xmath206 and the defined inner product in definition [ def : x ] , the quantity @xmath207 is a norm on @xmath203 .
we define now two sets , associated to each of the variables @xmath166 and @xmath165 : @xmath208 @xmath209 are the estimates of resp .
@xmath210 .
if the following statements are true : * @xmath211 : the noise components in @xmath212 are independent identically distributed ( i.i.d ) with zero - mean and finite temporal and spatial variance .
* @xmath213 : the matrix @xmath214 is full column rank , which corresponds to the temporal persistency of excitation .
* @xmath215 : either @xmath216 or @xmath217 is known for all @xmath164 , and the first non - zero entry of @xmath218 is strictly positive . *
@xmath219 : the initial estimate @xmath220 is non - zero . then , the map @xmath221 is a contraction on @xmath222 and has a unique fixed point @xmath223 when @xmath224 , that corresponds to the true parameters .
this theorem proves that , whatever the ( non - zero ) initial conditions are , the alternating least squares in equations - converges to a global minimum when @xmath224 .
that global minimum corresponds to the true parameters of the krarx model .
the analysis focuses on one alternating step because the cost is identical for the other minimization step .
computing @xmath225 costs @xmath226 . at each iteration ,
the least - squares estimate of @xmath165 is computed as @xmath227 .
the matrix - matrix multiplication is done in parallel by sending to @xmath228 cores the vector @xmath229 , each of which processes @xmath230 with @xmath231 .
a cholesky factorization is then computed with @xmath232 , hence enabling faster inversion with gaussian pivoting on triangular matrices .
denote @xmath233 . the global cost for computing
the matrix - matrix multiplication @xmath234 is @xmath235 , which can be reduced by distributing the computations over different cores .
last , @xmath236 requires @xmath237 .
a total of @xmath238 operations are needed to solve one alternating minimization .
hence the als requires @xmath239 flops , where @xmath240 is the total number of iterations .
solving the unstructured least squares requires @xmath3 .
more importantly , part of the bottleneck resides into matrix - matrix multiplications for which graphical processing units ( gpu ) are efficient computing tools .
we extend here the framework considered in that collects the sensor data in a matrix .
let us assume that there are a few missing entries in the map @xmath59 , @xmath241 .
for example , the data might be collected on a circular array , which implies that the blocks in the matrices do nt share the same size .
consequently , the coefficient - matrices in the lifted varx model do not retain the kronecker structure .
therefore we embed the network in a rectangular enveloppe , with the added entries considered as unknowns . in this paragraph
we study the estimation of the factor matrices as well as the missing entries by working on the rectangular embedding .
denote the set of known , resp .
unknown , entries in @xmath242 with @xmath243 , resp . @xmath244 .
the global least squares boils down to : @xmath245 the estimation problem still belongs to the class of multi - convex problems , @xcite .
the alternating minimization algorithm consists of minimizing each set of variables while the others are fixed and iterate until convergence .
for example , we start by initializing @xmath246 and @xmath247 , for all @xmath248 , and optimize over @xmath167
. eventually this will provide with estimates of the unknown sensor measurements for all @xmath249 .
the parametrization of the factor matrices based on additional knowledge of the network may help either to reduce the computational complexity of the model identification step , or to cast the model into a structure useful for future use , e.g control .
the first category include banded , symmetric , toeplitz and circulant patterns whereas the second contains e.g sparse or sequentially semi - separable ( sss ) matrices .
exploring such structures on the factor matrices is very attractive numerically as the problem size reduces further .
the block - toeplitz toeplitz - blocks ( bttb ) structure arises e.g when modeling 2d homogeneous spatially - invariant phenomena on a rectangular grid .
the kronecker and bttb structures are related , but not equivalent .
an insight is given in the following lemma .
let @xmath250 .
* if @xmath5 is symmetric block - toeplitz , then @xmath5 has a kronecker rank of @xmath136 . *
if @xmath5 has a kronecker rank of 1 , it does nt _ in general _ imply neither that @xmath5 is block - toeplitz nor has toeplitz - blocks .
enforcing the coefficient matrices to be toeplitz fastens up the identification : at each alternating step , the unknown toeplitz is embedded into a circulant matrix which is then diagonalized using the discrete fast fourier transform .
the graphs corresponding to real - networks are sparse , @xcite .
we assume that a node is connected to a limited number of other nodes in the network relatively to the network s size , and therefore we seek to induce sparsity in the matrices @xmath71 and @xmath72 .
when the influence of neighboring nodes decay with the distance , a multi - banded structure is equivalent to a banded structure of each factor matrix .
when the graph does nt exhibit such regularity , one other possibility is to induce zero - elements in the estimated parameter vectors @xmath69 , @xmath70 .
let @xmath251 denote the zero - norm , then using again @xmath103 as regularization parameter , the following constrained optimization problem results , @xmath252 with @xmath253\ ] ] in order to preserve the biconvex nature the zero norm is replaced by the @xmath39-norm as follows , @xmath254 we can not guarantee theoretically the convergence of this alternating sparse least squares because the proof of theorem 6 in appendix relies on a closed - form expression of each update .
optimization of sparse regularized least squares as in has been widely studied in the literature , see e.g @xcite .
the so - called sss structure enables standard matrix computations ( @xmath255 ) to be done in linear computational complexity with respect to the matrix size .
for example , inverting a matrix @xmath256 written as @xmath257 in which both @xmath258 , @xmath259 have a sss structure requires @xmath260 operations instead of @xmath3 .
low - rank off - diagonal blocks of the factor matrices that are sought after can be enforced via nuclear norm regularization , see @xcite for more details .
the proposed _ quarks identification _ method is now illustrated with a real - life example of a network with unknown communication links . the atmospheric turbulence is a stochastic process that has been modeled via state - space and var models or simplified into diagonal var models , @xcite .
when a lightbeam with a flat wavefront passes through a turbulent medium , the wavefront gets distorted .
the spatial covariance of the wavefront is not sparse and hence there exists multiple connections from a subsystem to the other in the network . in this paper
the turbulence is modeled with a state - space model following the method in @xcite for one single layer .
the quarks identification is performed with the als algorithm , and the number of iterations is limited to 10 .
the variance accounted for ( vaf ) between two signals @xmath261 and @xmath262 is defined with : @xmath263 two signals with a vaf equal to @xmath264 are identical .
the performance criteria are both the vaf and relative root - mean - square - error ( rmse ) between the signals @xmath265 and @xmath266 .
an array of @xmath267 phase points is considered .
the identification set contains @xmath268 temporal measurements . in figure 4 and 5 the coefficient matrix @xmath269
is displayed when identifying a varx model with @xmath105 resp .
for the case of estimating the varx coefficient matrices via least - squares with @xmath270 sparsity regularization and with krarx model . in figure 6
is plotted the relative rmse in boxplots for different methods these different methods .
the patterns of coefficient - matrices identified differ from the method used , and although the sparsity constraint minimizes the number of non - zero entries , it remains detrimental to the overall performance quality with respect to the number of parameters needed to construct the matrix itself .
let us define a measure that we call _ model complexity _ as the number of non - zero entries needed to construct the @xmath83 coefficient - matrices .
for example , the complexity of a krarx model is at most @xmath271 -only the non - zero elements of the factor matrices- , while it reaches a total of @xmath272 for the full least squares estimation .
it is illustrated in figure 7 that displays the vaf with respect to the 0-norm of the entries needed to construct the full coefficient matrix .
the vaf obtained with the sparse identification decreases with increasing regularization parameter .
exact _ number of non - zero entry decreases only for high prior on sparsity , for which the vaf is already 0 . model with a kronecker structure , @xmath273 .
entries in log10 . ] .
sp2 corresponds to sparser model than sp1 . ]
this paragraph investigates how the kronecker rank of the coefficient - matrices evolves with increasing network size .
a total of 7500 points is considered in the identification batch .
10 different experiments were carried out and the kronecker rank of a var1 model ranged from 1 to 4 .
figure 8 displays the relative rmse as a function of the total number of phase points , i.e @xmath228 .
the relative rmse decreases with increasing network size for all methods which is due to the fact that the frozen flow assumption shifts the turbulence phase from one column at every time sample and the shifted values -easily predictable- constitute the biggest part of the matrix . the ratio of new wavefront entries at each time step to the total number of phase points in the screen is equal to @xmath274 , which is the trend observed in figure 8 .
moreover , we observe that the gap between the least squares solution and the kronecker estimation does widen with increasing number of points but still keeps a very reasonable relative rmse .
figure 9 displays the number of entries requires to represent the model for both the standard least - squares solution and the kronecker - based representation .
in this paper is defined the class of kronecker networks for which the arx modeling part is investigated .
each coefficient - matrix of the var model is approximated with a sum of few kronecker matrices which offers high data compression for large networks .
estimating in least - squares sense the data matrices give rise to a bilinear problem which is addressed using two methods : first a three - stage non - iterative method is derived , then a iterative alternating least squares whose convergence was proved with a non - zero initial guess .
this second framework provides an elegant way of dealing with missing sensor data and adding further information on the factor matrices , e.g toeplitz , banded , sparse .
numerical examples on atmospheric turbulence data demonstrates the high compression capabilities of this model as well as its scalability for larger networks .
while the ordering of the nodes in the network is crucial for an efficient kronecker - representation of the coefficient matrix in the varx model , it may follow the intuition in some cases such as the one presented in the example .
the stability of the kvarx model identified has not been presented in this paper and is subject of current investigations .
the proof contains 4 points .
1 . positiveness : @xmath275 is positive because @xmath276 and the 2-norm for vectors @xmath277 are both positive .
if @xmath278 , and the matrix @xmath214 is full rank , then @xmath279 and @xmath8 is zero .
this implies that @xmath280 and therefore also the terms @xmath281 for all @xmath164 .
+ for the converse , we introduce a partitioning of @xmath214 such that @xmath282 .
@xmath283 each block - column @xmath65 is full column rank ( persistency of excitation ) . hence @xmath284 , and @xmath285 .
3 . let @xmath286 .
@xmath287 hence , @xmath288 4 . triangular inequality .
@xmath289 1 .
positiveness : @xmath275 is positive because @xmath276 and the 2-norm for vectors @xmath277 are both positive .
if @xmath278 , and the matrix @xmath214 is full rank , then @xmath279 and @xmath8 is zero .
this implies that @xmath280 and therefore also the terms @xmath281 for all @xmath164 .
+ for the converse , we introduce a partitioning of @xmath214 such that @xmath282 .
@xmath283 each block - column @xmath65 is full column rank ( persistency of excitation ) .
hence @xmath284 , and @xmath285 .
3 . let @xmath286 .
@xmath287 hence , @xmath288 4 . triangular inequality .
@xmath289 @xmath277 is therefore a norm .
the following lemma corresponds to lemma @xmath291 in @xcite .
it is used to prove that @xmath192 is a contraction mapping .
it is derived straightforwardly from the definition of the norm above .
let @xmath292 be defined with : @xmath293 . under assumption a2
, the magnitude of the directional derivative of @xmath294 along a direction vector @xmath295 attains its maximum when @xmath295 is in the same directions as @xmath296 .
it is shown that : @xmath297 key in this lemma is therefore to evaluate the term @xmath298 , which we get _ from the definition _ of the norm : @xmath299 the proof of the lemma is therefore unchanged compared with lemma @xmath291 of @xcite : the contour planes of @xmath294 are concentric spheres , and hence the gradient is in the radial direction .
the proof consists of first proving that @xmath192 is an operator mapping @xmath222 to @xmath222 as @xmath224 .
then the operator @xmath183 is a contraction mapping on @xmath222 , and last that the unique fixed point is equal to the true parameters @xmath300 .
note that three norms are used .
when @xmath8 is a vector , @xmath11 denotes the 2-norm .
when @xmath301 , @xmath302 is the norm related to the inner product from definition 4 .
else , the induced norm is used . as preliminary
, we recall the inequality @xmath303 .
we have : @xmath304 therefore : @xmath305 we start with the right - hand side term : @xmath306 @xmath176 has a finite spatial and temporal variance , hence @xmath307 is finite . since @xmath308 , @xmath309 therefore , @xmath310 @xmath311 is equal to @xmath39 if and only if @xmath312 .
else , use can be made of cauchy - schwarz inequality and the fact that @xmath166 is in @xmath313 to prove that it is inferior to 1 .
we conclude this first part with : @xmath314 therefore , @xmath183 goes from @xmath222 to @xmath222 .
let us now study the contraction mapping , and determine an upper bound on @xmath315 .
we have , for @xmath224 : @xmath316 multiplying by @xmath317 on both left sides gives : @xmath318 the right - hand side term reads : @xmath319 and hence , for all @xmath320 : @xmath321 @xmath322 being full column rank , we have : @xmath323 therefore , since @xmath324 , @xmath325 we have that @xmath326 , which then enables to prove that @xmath327 .
a similar reasoning holds to prove that @xmath328 .
we now introduce the quantity @xmath329 .
we have : @xmath330 where @xmath331 is the composition between the vectorized operator and @xmath183 . from @xmath332 ,
we decompose : @xmath333 let us start the analysis with @xmath334 , and by introducing a deviation from @xmath33 , denoted with @xmath335 : @xmath336 when @xmath337 , the difference is approximated with : @xmath338 therefore , @xmath339 where @xmath340 .
moreover , @xmath341 from the update rule @xmath342 . the relationship between the norms of @xmath335 and @xmath296 ( which are vectors in the same direction ) is @xmath343 .
hence : @xmath344 again using cauchy - schwarz and the fact @xmath345 .
it proves that @xmath192 is a contraction map on @xmath222 .
c. , jackson , k. , veran , j - p . ,
andersen , d. , lardire , o. and bradley , c. static and predictive tomographic reconstruction for wide - field multi - object adaptive optics systems .
, volume 31 , 2014 .
|
in this paper we propose a kronecker - based modeling of large networks with unknown interconnection links .
the class of kronecker networks is defined for which we formulate a vector autoregressive model .
its coefficient - matrices are decomposed into a sum of kronecker products . when the network is labeled such that the number of terms in the sum is small compared to the size of the matrix , exploiting this kronecker structure leads to high data compression .
two algorithms were designed for an efficient estimation of the coefficient - matrices , namely a non - iterative and overparametrized algorithm as well as an alternating least squares minimization .
we prove that the latter always converges to the true parameters for non - zero initial conditions .
this framework moreover allows for a convenient integration of more structure ( e.g sparse , banded , toeplitz ) on smaller - size matrices .
numerical examples on atmospheric turbulence data has shown comparable performances with the unstructured least - squares estimation while the number of parameters is growing _ only linearly _ w.r.t . the number of nodes instead of quadratically in the full unstructured matrix case .
network modeling , large - scale systems , kronecker product , low - rank approximation , multi - convex minimization .
|
as a new way to power mobile devices , there has been a growing interest in harvesting energy from the far - field radio - frequency ( rf ) signal transmissions . in particular , design of wireless - powered communication network ( wpcn ) has been studied as a significant application of rf energy harvesting , where remote user equipments ( ues ) utilize the energy harvested from wireless rf transfer of power for wireless communications .
one typical wpcn model has been proposed in @xcite , in which wireless energy transfer ( wet ) in the downlink ( dl ) and wireless information transmissions ( wit ) in the uplink ( ul ) are both coordinated by a hybrid access - point ( h - ap ) . in wpcns where the h - ap and ues all operate in half duplex ( hd ) , termed as hd - wpcns , a fundamental trade - off exists in allocations of resources to dl for wet and ul for wit @xcite , because allocating more resources to dl for wet increases transmit power of ues in the ul thanks to the increase in the amount of harvested energy , but also decreases resources allocated to ul for wit .
meanwhile , full duplex ( fd ) based wireless systems have a great deal of attention , in which the wireless nodes transmit and receive rf signals at the same time and in the same frequency band .
fd operation is expected to potentially double the spectral efficiency in wireless communications by cancelling the leakage of transmitted signal received by itself , i.e. , self - interference ( si ) . in particular , feasibility of the fd communication
has been verified by proof - of - concept ( poc ) , by showing that the power of residual si after self - interference cancellation ( sic ) is reduced close enough to the power of background noise @xcite , @xcite .
the simultaneous transmission and reception led by fd operation can also be utilized to improve throughput of wpcn . in @xcite ,
one wpcn model in which an h - ap with fd operation coordinates wet to and wit from a set of ues operating in half - duplex ( hd ) , denoted by fd - wpcn with hd ues ( fd - wpcn - hd ) , has been proposed .
it has been shown in @xcite that fd operation of the h - ap in the fd - wpcn - hd is able to increase throughput of wpcn when si is sufficiently cancelled at the h - ap . in @xcite ,
kang et al . have studied resource allocation in fd - wpcn - hd with energy causality constraint , i.e. , the constraint that signal transmission of a ue at a given time can utilize only the harvested energy which is harvested during the previous time . in addition , fd - wpcn - hd with orthogonal frequency - division multiplexing ( ofdm ) modulation has been studied in @xcite . in @xcite ,
finally , wireless powered relay network has been investigated , in which an fd relay that utilizes si as an energy source relays information of an hd source to an hd destination . in this paper , we study another type of wpcn in which not only the h - ap but also ues operate in fd , denoted by fd - wpcn with fd ues ( fd - wpcn - fd ) , as shown in fig .
[ fig_systemmodel ] .
thanks to the fd capability , ues can receive the signal sent by the h - ap to carry energy while transmitting their respective ul information .
we first propose transceiver structure for ues that enables simultaneous energy reception in the dl and information transmission in the ul . by extending our previous work in @xcite ,
we characterize energy usage in the proposed ue transceiver accounting for energy leakage from transmit chain to receive chain , and show that ues can also harvest energy from the leakage of their own ul transmissions ( i.e. , si ) as well as the received signal sent by the h - ap in the dl . based on a time - division - multiple - access ( tdma ) protocol for information transmissions in the ul ,
finally , the optimal time allocation to maximize the weighted sum - throughput in the network is studied , subject to a given total time constraint .
fig . [ fig_systemmodel ] shows a fd - wpcn - fd model considered in this paper .
this network consists of one h - ap and @xmath0 ues ( denoted by @xmath1 , @xmath2 ) , all of which are equipped with a single - antenna fd transceiver to transmit and receive signal over the same frequency band simultaneously . the h - ap transfers energy to ues in the dl by broadcasting a signal dedicated to carry energy .
at the same time , the h - ap also receives signals transmitted by ues in the ul , which contains information to decode . fd operation of the h - ap enables to simultaneously transmit / receive energy / information signals in the dl / ul , respectively . for decoding information received in the ul ,
an si canceller is deployed at the fd transceiver of the h - ap to eliminate si ( i.e. , the leakage of dl energy signal transmission received by itself ) , which interferes with reception of ul information signals . on the other hand , ues harvest energy from the received signals and charges their respective battery .
a part of charged energy is used to transmit information in the ul .
thanks to fd operation , ues are also capable of simultaneously transmitting / receiving information / energy signals in the ul / dl , respectively .
this is enabled by the transceiver shown in fig .
[ fig_ue_transceiver ] .
note that the energy signal broadcast by the h - ap in the dl does not carry any information to decode but energy only .
therefore , neither information decoder nor si canceller is deployed at the receiving end of a ue s transceiver . at the receiving end , instead
, an energy harvester ( e.g. rectifier ) is installed to harvest energy from the received signals .
it is worth noting that in this transceiver , a part of power amplifier ( pa ) output energy at the transmitting end leaks into the receiving end as si , while the remaining part of pa output energy is used to transmit information to the h - ap . at the @xmath3th user @xmath1 , as shown in fig .
[ fig_ue_transceiver ] , @xmath4 , @xmath5 , and @xmath6 represent the pa output energy at the transmitting end , energy that leaks into the receiving end , and energy used to transmit information , respectively , where @xmath7 models the near - field energy absorption by the receiving antenna . ] . for convenience , @xmath5 and @xmath6
are given , respectively , by @xmath8 in the network shown in fig .
[ fig_systemmodel ] , signal transmission from @xmath1 to @xmath9 , @xmath10 , is assumed to pass through a complex channel @xmath11 with channel power gain given by @xmath12 ( @xmath13 represents the h - ap later in this paper ) . when @xmath14 , in particular , @xmath15 represents the si channel at @xmath1 through which the leakage signal of @xmath1 is received by itself .
it is assumed that the channels remain constant during a block transmission time denoted by @xmath16 ( i.e. , follow quasi - static flat - fading ) , and @xmath17 with the channel reciprocity .
the h - ap is assumed to have perfect knowledge of @xmath18 , @xmath19 .
furthermore , it is also assumed that the h - ap knows the channel power gains @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 .
finally , @xmath25 , @xmath26 , is assumed to be known by the h - ap .
[ fig_framestructure ] shows a transmission protocol for fd - wpcn - fd . during whole block duration @xmath16 , the h - ap
transmits energy signal with a constant power @xmath27 in the dl , which is received by all ues . while broadcasting of energy by the h - ap
, ues transmit their own independent information to the h - ap in the ul , orthogonally over time by tdma .
the transmission block thus consists of @xmath0 slots each of which is allocated to @xmath1 for the ul information transmission .
the duration of the @xmath3th slot is given by @xmath28 , @xmath26 , where @xmath29 during the @xmath30th slot , the h - ap and user @xmath1 transmit dl signal to carry energy and ul signal to carry information , respectively .
therefore , the received signals at the h - ap and @xmath1 during the @xmath3th slot can be expressed , respectively , as @xmath31 @xmath32 with @xmath33 , @xmath34 , denoting the transmitted signal of the @xmath35 with @xmath36 = 1 $ ] .
furthermore , @xmath37 represents the transmit power of @xmath35 , @xmath34 .
finally , @xmath38 denotes the receiver noise at @xmath35 , @xmath34 .
it is assumed that @xmath39 , with @xmath40 representing a circularly symmetric complex gaussian random variable with mean @xmath41 and variance @xmath42 .
it is worth noting that the h - ap has to decode @xmath43 in ( [ eq_rxsignal_ap ] ) whereas @xmath1 does not decode information in the dl . at their respective receiving ends , therefore , the h - ap has to cancel the leakage of its own transmission acting as si , i.e. , @xmath44 in ( [ eq_rxsignal_ap ] ) , whereas it is not necessary for @xmath1 to cancel the leakage of its own transmission , i.e , @xmath45 in ( [ eq_rxsignal_ue_i ] ) .
in addition , the inactive ues during the @xmath3th slot , i.e. , @xmath9 s with @xmath24 and @xmath46 , also receives signals from the transmissions of the h - ap and @xmath1 . at an inactive ue @xmath9 , @xmath46 , therefore
, the received signal during the @xmath3th slot is given by @xmath47
in this section , the achievable throughput of fd - wpcn - fd presented in section [ sec : systemmodel ] is studied . in particular
, we aim to maximize the weighted sum - throughput in the ul transmissions . a ue @xmath1 can harvest energy from the received dl signals transmitted by the h - ap ( i.e. , @xmath48 in ( [ eq_rxsignal_ue_i ] ) ) during whole block duration .
in addition , ul transmissions of other ues received by @xmath1 ( i.e. , @xmath49 , @xmath46 ) can also be used for energy harvesting of @xmath1 .
furthermore , @xmath1 can also harvest energy from the leakage of its own transmission during the @xmath3th slot allocated to it for ul information transmission ( i.e. , @xmath45 in ( [ eq_rxsignal_ue_i ] ) ) . during a block duration @xmath16 , therefore , the harvested energy of @xmath1 , denoted by @xmath50 , can be obtained from ( [ eq_rxsignal_ue_i ] ) and ( [ eq_rxsignal_ue_j ] ) as @xmath51 where @xmath52 , @xmath26 , denotes the energy harvesting efficiency at @xmath1 . in ( [ eq_energy_ui ] ) , the harvested energy from the receiver noise is not considered assuming that the amount of harvested energy from the receiver noise is small enough to be neglected @xcite . in the fd ue transceiver
proposed in fig .
[ fig_ue_transceiver ] , it is worth noting that by the law of energy conservation , the energy harvested from the leakage of its own transmission , i.e. , @xmath53 in ( [ eq_energy_ui ] ) , is equivalent to the amount of energy that leaks from the transmitting end into the receiving end . from ( [ eq_definintion_energy ] ) , therefore , we have @xmath54 in each transmission block , a part of the harvested energy is used to generate the pa output signal at the transmitting end of the transceiver in fig .
[ fig_ue_transceiver ] .
the energy of pa output signal at @xmath1 is then given by @xmath55 with @xmath56 , @xmath26 , denoting the portion of the harvested energy used to generate pa output signal at @xmath1 in steady state.s are charged sufficiently at the beginning of each block and @xmath1 s use their respective harvested energy such that the energy charged in the battery at the end of each block reaches its initial state . ] note that @xmath57 in ( [ eq_definintion_energy ] ) can be represented as @xmath58 from ( [ eq_energy_ui])-([eq_e_i_t ] ) , we then have @xmath59 @xmath60 from which we have @xmath61 where @xmath62 .
it then follows that @xmath63 where @xmath64^t$ ] and @xmath65 with @xmath66 denoting a diagonal matrix of which the diagonal entries consist of the vector @xmath67 .
furthermore , @xmath68 in ( [ eq_transmissionenergy_matrix ] ) is a matrix with @xmath69 where @xmath70 denotes the element of matrix @xmath68 on the @xmath30th row and @xmath71th column .
finally , @xmath72 in ( [ eq_transmissionenergy_matrix ] ) is a vector given by @xmath73^t.\ ] ] from ( [ eq_transmissionenergy_matrix])-([eq_vector_beta ] ) , @xmath74 is then given by @xmath75 where @xmath76 , @xmath26 , is given by @xmath77 = { \bf{a}}^{-1 } { \boldsymbol{\beta}}.\ ] ] note that at the h - ap , the residual si after sic can be approximated as @xmath78 , where @xmath79 @xcite . given time allocation @xmath80 $ ] , achievable ul throughput of @xmath1 during the @xmath30th slot is then given from ( [ eq_rxsignal_ap ] ) and ( [ eq_pi ] ) by @xmath81 @xmath82 where @xmath83 is given by @xmath84 with @xmath76 , @xmath26 , given in ( [ eq_rho ] ) , and @xmath85 representing the gap of signal - to - interference - plus noise ratio ( sinr ) from the additive white gaussian noise ( awgn ) channel due to practical modulation and coding used .
as shown in ( [ eq_ul_rate ] ) , the throughput of the fd - wpcn - fd can be maximized by optimizing time allocated to each ue for ul information transmission , denoted by @xmath86 $ ] .
specifically , we maximize the weighted sum - throughput in this network by solving the problem formulated as follows : @xmath87 @xmath88 @xmath89 it can be easily shown that problem ( p1 ) is a convex optimization problem since @xmath90 and ( [ eq_prop_constraint1 ] ) are concave and affine functions of @xmath91 , respectively . the optimal time allocation solution for ( p1 ) , denoted by @xmath92 ,
is then obtained as follows .
[ proposition_opt_time_alloc ] the weighted sum - throughput of the fd - wpcn - fd is maximized by the optimal time allocation solution for ( p1 ) , given by @xmath93 $ ] with @xmath94 where @xmath95 denote the solution of @xmath96 , with @xmath97 defined as @xmath98 the proof is shown in appendix [ proof_proposition ] .
[ fig_rateregion ] compares achievable throughput region of fd - wpcn - fd with those of hd - wpcn and fd - wpcn - hd studied in @xcite , with @xmath99 , @xmath100db , @xmath101db , and @xmath102 .
it is assume that @xmath103 , @xmath104 , and @xmath105 .
for fd - wpcn - hd , in addition , sic is assumed to be perfect at the h - ap . for fd ues in fd - wpcn - fd , it is further assumed that @xmath106 with 15db isolation between transmission and reception .
it is observed from this figure that when @xmath107 with @xmath108 , the achievable throughput region of the fd - wpcn - fd is larger than those of hd - wpcn and fd - wpcn - hd with @xmath109 , where @xmath110 denotes the maximum peak transmit power at the h - ap in hd - wpcn and fd - wpcn - hd . when @xmath111 , in addition , the achievable throughput region of the fd - wpcn - fd becomes larger than those of hd - wpcn and fd - wpcn - hd with @xmath112 .
[ corollary_time_alloc_sumratemax ] the sum - throughput of the fd - wpcn - fd is maximized with a time allocation @xmath113 $ ] , with @xmath114 the proof is shown in appendix [ proof_corollary ] .
in this section , the sum - throughput of the hd - wpcn , fd - wpcn - hd , and fd - wpcn - fd are compared by simulations , with @xmath115 and @xmath116 = 0.5 , @xmath117 .
the bandwidth and noise spectral density are set as 1mhz and -160dbm / hz , respectively .
ues are assumed to be uniformly distributed within two concentric circles each of which has diameter of 5 m and 10 m , respectively , while the h - ap is located on the center of these concentric circles .
assuming that the average signal power attenuation at a reference distance of 1 m is 30db , the channel power gain @xmath20 is modeled as @xmath118 , @xmath119 , @xmath22 , with @xmath120 denoting the distance between @xmath1 and @xmath9 in meter , and the pathloss exponent @xmath121 .
we also assume that the short - term fading , denoted by @xmath122 , is rayleigh distributed , i.e. , @xmath122 is an exponentially distributed random variable with unit mean . for fd ues in fd - wpcn - fd , @xmath123 , @xmath117 , assuming 15db isolation between transmission and reception .
finally , we set @xmath85=9.8db assuming an uncoded quadrature amplitude modulation with the required bit - error rate given by @xmath124 @xcite . .
] in fig .
[ fig_sumrate_vs_p0 ] , the maximum sum - throughput of hd - wpcn , fd - wpcn - hd , and fd - wpcn - fd are compared with respect to @xmath27 values in dbm . for fd - wpcn - fd , @xmath111 is assumed with @xmath125 defined in the previous section , whereas perfect sic is assumed for fd - wpcn - hd .
we can observe in this figure that the average sum - throughput of fd - wpcn - fd is equivalent to those of hd - wpcn and fd - wpcn - hd with @xmath112 .
as compared to the fd - wpcn - hd with @xmath109 , the fd - wpcn - fd achieves sum - throughput gain of 4db with respect to @xmath27 .
furthermore , the sum - throughput of the fd - wpcn - fd is shown to increase faster with the increase of @xmath27 than that of the hd - wpcn with @xmath109 .
[ fig_sumrate_vs_sic ] compares the maximum sum - throughput of aforementioned wpcns for different values of sic gain , where sic gain is defined as @xmath126 .
the sum - throughput of the fd - wpcn - fd is shown to be larger than those of hd - wpcn and fd - wpcn - hd when sic gain at the h - ap is larger than 114db , i.e. , power of the residual si after sic is @xmath127 .
this implies that the use of fd ues proposed in fig .
[ fig_ue_transceiver ] efficiently improves the throughput of wpcn even with current sic technologies at the h - ap ( e.g. , @xcite and @xcite ) , which reduce the power of the residual si sufficiently close to @xmath128 after sic .
this paper studied an wpcn where both the h - ap and ues operate in fd .
thanks to the fd capability , ues can receive energy and transmit information simultaneously in the dl and ul , respectively .
we first proposed an fd transceiver structure for ues that enables simultaneous energy reception and information transmission , and then provided the energy usage model in the proposed ue transceiver .
furthermore , the optimal time allocation solution to maximize the weighted sum - throughput of the wpcn with fd h - ap and fd ues was obtained .
it was shown that the use of the proposed fd ues as well as fd h - ap with practical sic capability improves the throughput of wpcn significantly .
from ( [ eq_prop_primal])-([eq_prop_constraint2 ] ) , we have the lagrangian of @xmath129 as follows .
@xmath130 with @xmath131 representing the lagrange multiplier associated with ( [ eq_prop_constraint1 ] ) .
we thus have the dual function of ( @xmath132 ) as follows .
@xmath133 with @xmath134 denoting the feasible set of @xmath135 specified by ( [ eq_prop_constraint1 ] ) and ( [ eq_prop_constraint2 ] ) .
note that slater s condition holds for ( p1 ) because we can find a @xmath136 , @xmath137 , @xmath138 , with which @xmath139 .
therefore , @xmath129 is shown to be a convex optimization problem for which the strong duality holds , and the globally optimal solution for ( p1 ) thus satisfies the karush - kuhn - tucker ( kkt ) conditions @xcite , given by @xmath140 @xmath141 @xmath142 with @xmath143 s and @xmath144 denoting , respectively , the optimal primal and dual solutions of ( @xmath132 ) . with no loss of generality , we can assume @xmath145 from ( [ app_optimalitycondition2 ] ) because @xmath146 must hold for ( @xmath132 ) . from ( [ app_optimalitycondition3 ] )
, it then follows that @xmath147 where @xmath148 .
note that @xmath149 in ( [ eq_function_f ] ) is a monotonically increasing function of @xmath150 with @xmath151 and @xmath152 .
therefore , @xmath92 is uniquely obtained for given values of @xmath153 .
the proof of proposition [ proposition_opt_time_alloc ] is completed .
the sum - throughput of this network is maximized by solving ( p1 ) with @xmath154 . for the inequality in ( [ app_optimalitycondition4 ] ) to hold in this case
, we should have @xmath155 from ( [ app_optimalitycondition2 ] ) and ( [ app_timeportionratio ] ) , it follows that @xmath156 from ( [ eq_app_sumtime ] ) , the optimal time allocation solution in ( [ eq_corr_opt_tau ] ) is obtained .
corollary [ proposition_opt_time_alloc ] is now proved .
h. kim , h. lee , m. ahn , h. -b .
kong , and i. lee , `` joint subcarrier and power allocation methods in full duplex wireles powered communication networks for ofdm systems , '' _ ieee trans .
wireless commun .
15 , no . 7 , pp .
4745 - 4753 , july 2016 .
h. ju , k. chang , and m. -s .
lee , `` in - band full - duplex wireless powered communication networks , '' in _ proc .
ieee international conf . on advanced communi
( icact ) _ , pp .
23 - 27 , pyeongchang , korea , july 2015 .
b. p. day , a. r. margetts , d. w. bliss , and p. shcniter , `` full - duplex bidirectional mimo : achievable rates under limited dynamic range , '' _ ieee trans .
signal process .
60 , no . 7 , pp . 3702 - 3716 , july 2012 .
|
this paper studies a wireless powered communication network ( wpcn ) consisting of a hybrid access point ( h - ap ) and multiple user equipments ( ues ) , all of which operate in full - duplex ( fd ) .
we first propose a transceiver structure that enables fd operation of ues to receive energy in the downlink ( dl ) and transmit information in the uplink ( ul ) simultaneously .
we then provide an energy usage model in the proposed ue transceiver accounting for energy leakage from transmit chain to receive chain .
it is shown that throughput of fd wpcn with the proposed fd ues can be maximized by optimally allocating ul transmission time to ues by solving a convex optimization problem .
simulation results reveal that the use of the proposed fd ues efficiently improves the throughput of wpcn with practical self - interference cancellation capability at the h - ap .
[ section ] [ section ] [ section ] [ section ] [ section ] [ section ] [ section ]
|
the recent discovery @xcite of a new particle with properties consistent with a standard model ( sm ) higgs boson is clearly the most significant news from the large hadron collider ( lhc ) .
this discovery was supported by evidence for a higgs boson found by the cdf and d0 collaborations at the tevatron @xcite and completes our picture of the sm .
however , the sm leaves many fundamental questions open perhaps the most pressing issue being that the sm does not explain the value of the electroweak scale , the higgs mass , itself . clearly , a prime goal after the discovery is to thoroughly test the sm nature for this higgs - like signal .
the sm makes precise predictions for the production cross sections of the higgs boson @xmath12 ( via gluon - gluon fusion ( ggf ) , vector boson fusion ( vbf ) , associated production with an electroweak gauge boson @xmath1 ( vh ) and associated production with a @xmath13 pair ( ) ) , and its decay branching fractions into various final states ( @xmath14 , @xmath15 , @xmath16 , @xmath17 , and @xmath18 ) as a function of its unpredicted mass @xmath19 .
the observation of the higgs boson at the lhc is based primarily on the @xmath20 @xcite , @xmath15 @xcite and @xmath16 @xcite decay modes .
the higgs boson mass , @xmath19 , is quite precisely measured to be in the 125126 gev range using the high resolution @xmath14 and @xmath15 final states @xcite .
gev in the @xmath21 channel , the combined value from atlas is @xmath22 gev , in very good agreement with @xmath23 gev measured by cms .
we find it reasonable to assume that the lower @xmath19 value from the atlas @xmath24 measurement is due to a statistical fluctuation or unknown systematics .
] the evidence for the higgs boson at the tevatron is based principally on the @xmath17 decay mode @xcite , the observed enhancements being consistent with a large range of possible higgs masses . with the measurements in various channels , a comprehensive study of the properties of the higgs - like state becomes possible and has
the potential for revealing whether or not the higgs sector is as simple as envisioned in the sm . in particular
it is crucial to determine the higgs couplings to gauge bosons and to fermions as defined by the lagrangian @xmath25 h\ , , \label{ldef}\ ] ] where the @xmath26 are scaling factors for the couplings relative to their sm values , introduced to test possible deviations in the data from sm expectations . in principle
all the @xmath26 are independent , in particular the @xmath27 can be different for up- and down - type quarks and/or leptons . a significant deviation of any @xmath26 from unity would imply new physics beyond the sm . while fits to various combinations of @xmath26 s are performed by the experimental collaborations themselves @xcite , we find it important to develop our own scheme in order to bring all results from atlas , cms and the tevatron experiments together and test not only the sm but also specific models beyond .
such fits by theorists , using various parametrizations , were performed previously in @xcite . here , we go beyond these works by including all publicly available data as of the end of 2012 .
in particular we take into account the updates presented at the hadron collider physics symposium in nov .
2012 ( hcp2012 ) @xcite and at the open session of the cern council in dec .
2012 @xcite .
our parametrization is as follows .
we treat the couplings to up - type and down - type fermions , @xmath28 and @xmath29 , as independent parameters ( but we only consider the case @xmath30 , and we assume that the @xmath27 are family universal ) .
moreover , we assume a custodial symmetry in employing a single @xmath31 in eq .
( [ ldef ] ) .
the structure we are testing thus becomes @xmath32h\ , .
\label{ourldef}\ ] ] in general , the @xmath26 can take on negative as well as positive values ; there is one overall sign ambiguity which we fix by taking @xmath33 . even in this restricted context , various types of deviations of these three @xmath26 from unity are possible in extended theories such as two - higgs - doublet models ( 2hdms ) , models with singlet - doublet mixing , and supersymmetric models such as the minimal supersymmetric standard model ( mssm ) and the next - to - minimal supersymmetric standard model ( nmssm ) .
in addition to the tree - level couplings given above , the @xmath12 has couplings to @xmath2 and @xmath3 that are first induced at one loop and are completely computable in terms of @xmath28 , @xmath29 and @xmath34 if only loops containing sm particles are present .
we define @xmath35 and @xmath36 to be the ratio of these couplings so computed to the sm ( @xmath37 ) values .
however , in some of our fits we will also allow for additional loop contributions @xmath38 and @xmath39 from new particles ; in this case @xmath40 and @xmath41 .
the largest set of independent parameters in our fits is thus @xmath42 in this study , we focus on models in which the higgs decays only to sm particles , in particular not allowing for invisible ( e.g. @xmath43 , where @xmath44 is the lightest susy particle ) or undetected decays ( such as @xmath45 , where @xmath46 is a light cp - odd , perhaps singlet scalar ) .
this approach , when we allow in the most general case for the @xmath28 , @xmath29 , @xmath34 , @xmath47 and @xmath48 couplings to be fully independent , encompasses a very broad range of models , including in particular those in which the higgs sector consists of any number of doublets + singlets , the only proviso being the absence of decays of the observed @xmath0 state to non - sm final states .
( a fit for invisible higgs decays was performed early on in @xcite . )
this approach however does not cover models such as composite models and higgs - radion mixing models for which the @xmath49 coupling has a more complicated tensor structure than that given in .
our procedure will also be inadequate should the observed signal at @xmath0 actually arise from two or more degenerate higgs bosons ( see _ e.g. _
although the success of our fits implies that there is no need for such extra states , the explicit tests for degenerate states developed in @xcite should be kept in mind as a means to test directly for two or more higgs bosons contributing to the signal at 125126 gev .
this paper is organized as follows .
the experimental inputs and our fitting procedure are described in section 2 .
the results of three generic fits are presented in section 3 together with the results of a fit in two - higgs - doublet models .
section 4 contains our conclusions .
we perform fits employing all production / decay channels for which results are available from the atlas and cms collaborations at the lhc , as well as the tevatron cdf+d0 higgs results .
the experimental results are given in terms of signal strengths @xmath50 , the ratio of the observed rate for some process @xmath51 relative to the prediction for the sm higgs .
often it is the case that several production processes contribute to a given experimental channel .
for example , both vector boson fusion and gluon fusion can contribute to the vbf " channels ( or `` categories '' ) that are defined by a given set of experimental cuts .
in comparing theory to experiment it is thus important to incorporate the estimates from the experiments of the relative contributions of the theoretically distinct production / decay processes .
the values for the signal strengths in the various ( sub)channels as reported by the experiments and used in this analysis , together with the estimated decompositions into production channels are given in tables [ atlasresults][tevatronresults ] .
we adopt the simple technique of computing the @xmath52 associated with a given choice of the input parameters following the standard definition : ^2=_k ( _ k-_k)^2_k^2 , [ chisqdef ] where @xmath53 runs over all the experimentally defined production / decay channels employed , @xmath54 is the observed signal strength for channel @xmath53 , @xmath55 is the value predicted for that channel for a given choice of parameters and @xmath56 is the experimental error for that channel . the @xmath55 associated with each experimentally defined channel is further decomposed as _ k = t^i_k _ i [ murel ] where the @xmath57 give the amount of contribution to the experimental channel @xmath53 coming from the theoretically defined channel @xmath58 and @xmath59 is the prediction for that channel for a given choice of @xmath60 , @xmath61 , @xmath62 and ( for fits where treated as independent ) @xmath47 and @xmath48 . for the computation of the @xmath59 including nlo corrections we follow the procedure recommended by the lhc higgs cross section working group in @xcite .
in particular we include all the available qcd corrections for @xmath63 using ` higlu ` @xcite and for @xmath64 using ` hdecay ` @xcite , and we switch off the electroweak corrections .
the @xmath57 depend on the specific analysis and hence differ from experiment to experiment .
often , the @xmath57 are determined from simulations of a sm higgs signal . in some cases
, the experiments have done the unfolding of theoretical vs. experimental channels from the data and provide directly experimental results for the theoretically relevant @xmath65 s and the correlations between them .
+ @xmath19 ( gev ) & + & & & ggf & vbf & vh & tth + + @xmath66 & @xmath67 & 126.6 & 100% & & & + @xmath68 & @xmath69 & 126.6 & & 60% & 40% &
+ + inclusive & @xmath70 & 125 & 87% & 7% & 5% & 1% + + @xmath71 & @xmath72 & 125.5 & 95% & 3% & 2% & + +
vh tag & @xmath73 & 125.5 & & & 100% &
+ + @xmath74 & @xmath75 & 125 & 100% & & & + @xmath76 & @xmath77 & 125 & & 60% & 40% & + @xmath19 ( gev ) & + & & & ggf & vbf & vh & tth + + @xmath66 & @xmath78 & 125.8 & 100% & & & + @xmath79 & @xmath80 & 125.8 & & 60% & 40% & + + inclusive & @xmath81 & 125.8 & 87% & 7% & 5% & 1% + + 0/1 jet & @xmath82 & 125.8 & 97% & 3% & & + vbf tag & @xmath83 & 125.8 & 17% & 83% & & + vh tag & @xmath84 & 125.8 & & & 100% & + + vh tag & @xmath85 & 125.8 & & & 100% & + tth tag & @xmath86 & 125.8 & & & & 100% + + 0/1 jet & @xmath87 & 125.8 & 76% & 16% & 7% & 1% + vbf tag & @xmath88 & 125.8 & 19% & 81% & & + vh tag & @xmath89 & 125.8 & & & 100% & + @xmath19 ( gev ) & + & & & ggf & vbf & vh & tth + + combined & @xmath90 & 125 & 78% & 5% & 17% & + + combined & @xmath91 & 125 & 78% & 5% & 17% & + +
vh tag & @xmath92 & 125 & & & 100% &
+ with this framework programmed , our fitting procedure is as follows .
we first scan over a fine grid of the free parameters of the scenario considered , for example , @xmath60 , @xmath61 , @xmath62 with @xmath93 as computed from the sm - particle loops ( this will be fit * ii * below ) .
we obtain the value of @xmath94 associated with each point in the grid and thus determine the values of the parameters associated with the approximate minimum ( or minima ) . to get the true minimum @xmath94 , @xmath95 , and the associated best - fit " values and the one - standard deviation ( @xmath96 ) errors on them we employ minuit @xcite .
( the errors on parameters which are not input , @xmath97 and @xmath98 , are determined from the grid data . ) for plotting distributions of @xmath94 as a function of any one variable , we use the above grid data together with the best fit value , to profile the minimal @xmath94 value with respect to the remaining unconstrained parameters . the 68% , 95% and 99.7% confidence level ( cl )
intervals are then given by @xmath99 , @xmath100 and @xmath101 , respectively .
two - dimensional @xmath94 distributions are obtained analogously from a grid in the two parameters of interest , profiling over the other , unseen parameters ; in this case , we show contours of @xmath94 corresponding to the 68% ( @xmath102 ) , 95% ( @xmath103 ) and 99.7% ( @xmath104 ) confidence levels for 2 parameters treated jointly . + before presenting our results , a couple of comments are in order .
first of all , we stress that in models of new physics beyond - the - sm ( bsm ) , both the branching ratios and the production cross sections and distributions ( and indeed the number of higgs particles ) may differ from sm expectations . for any bsm interpretation of the higgs search
results it is absolutely crucial to have as precise and complete channel - by - channel information as possible @xcite .
unfortunately , not all the experimental analyses give all the necessary details .
below we comment on how we use the currently available information from the experiments .
the ideal case would of course be that the experiments consistently do the unfolding of theoretical vs. experimental channels from the data as mentioned above and always provide directly the experimental results for the theoretically relevant @xmath65 s ( see ) and the correlations between them .
* @xmath105 : we take our information from fig . 4 of @xcite .
this figure shows the results after unfolding to obtain the experimental results for the @xmath106 s as defined theoretically .
4 does make the approximation that vbf and vh can be lumped together ( have the same efficiencies after cuts ) and that ggf and tth can be similarly lumped together ( note that tth contributes less than @xmath107 ) .
we fit the 68% cl contour assuming that the @xmath108 follows a bivariate normal distribution . with this
, the correlation @xmath109 between the ggf and vbf+vh channels is automatically taken into account .
we also note that while fig .
4 of @xcite is for 126.6 gev , fig .
12 ( right ) in the same paper shows that there is a broad `` plateau '' as a function of the mass when the energy scale uncertainty is taken into account , implying that the results should not depend too much on the mass .
* @xmath110 : the signal strength in this channel reported by atlas @xcite is @xmath111 with a best fit mass of @xmath112 gev . at @xmath113 gev , the signal strength is @xmath114 , see fig . 10 in @xcite .
assuming that the discrepancy in the higgs mass determined from the @xmath3 and the 4 lepton final states is due to a statistical fluctuation ( rather than unknown systematics ) we use @xmath115 at @xmath113 gev , close to the combined best fit mass from atlas , in our fits .
+ alternatively , one could rescale the value of @xmath111 at @xmath116 gev for a higgs mass of 125 gev .
this would give @xmath117 at @xmath113 gev ( or @xmath118 at @xmath119 gev ) .
we checked that taking this alternative approach has only marginal influence on our results .
+ regarding the decomposition in production modes , no statement is made in the conference note or paper .
however , as it is an inclusive analysis , we take the relative ratios of production cross sections for a sm higgs as a reasonable approximation . to this end
, we use the ratios given by the lhc higgs cross section working group @xcite . *
@xmath120 : we adopt relative contributions of 95% and 5% vbf @xcite .
we do not include any result for 7 tev because the update presented at hcp is a combination of 7 and 8 tev .
* @xmath121 : atlas provides only an overall signal strength with no information on the decomposition with respect to production modes .
however , the conference note @xcite contains the results of unfolding to the theory - level @xmath106 s via a plot ( fig .
19 ) of the experimental results for @xmath122 versus @xmath123 at @xmath113 gev .
we fit the contours with the same procedure as for @xmath105 .
the correlation ( included in the fit ) is @xmath124 .
* @xmath105 : we follow the same procedure as for atlas @xmath125 , using figure 11 from @xcite .
the correlation ( included in the fit ) is @xmath126 .
* @xmath110 : no decomposition with respect to production modes is given in the conference note or paper . as it is a fully inclusive analysis , we use the relative ratios of production cross sections given by the lhc higgs cross section working group @xcite as a good approximation @xcite .
* @xmath127 : the information provided in the conference note and papers is incomplete ; our decomposition into production modes is based on @xcite .
our combination ( weighted mean ) agrees within 9% with that given by cms ( @xmath128 instead of @xmath129 ) . *
@xmath130 : as there is no information on possible contaminations by other production modes , we assume 100% vh or 100% tth production for the respective categories . *
@xmath131 : for the 0/1 jet and vbf tag categories we extract the decomposition into production modes from @xcite , assuming that there is no significant change in the efficiencies between @xmath132 gev and @xmath133 gev .
we use the efficiencies from the first three categories ( @xmath134 , @xmath135 and @xmath136 ) because they are the most sensitive ones ; they lead to very similar decompositions which we use in our analysis .
our combination ( weighted mean ) agrees within 6% with that given by cms ( @xmath137 instead of @xmath138 ) .
* @xmath105 and @xmath139 : no decomposition into production modes is given by the experiments .
we assume that the analyses are inclusive and we thus employ the ratios of the theoretical predictions for the ( sm ) higgs production cross sections .
* @xmath140 : we use the new results from hcp2012 @xcite assuming 100% vh .
for a first test of the sm nature of the observed higgs boson , we take @xmath37 ( quark , lepton and @xmath142 vector boson couplings to the higgs are required to be sm - like ) but we allow for additional new physics contributions to the @xmath3 and @xmath2 couplings , parameterized by @xmath38 and @xmath39 , coming from loops involving non - sm particles or from anomalies .
this fit , which we refer to as fit * i * , is designed to determine if the case where all tree - level higgs couplings are equal to their sm values can be consistent with the data .
for example , such a fit is relevant in the context of ued models where the tree - level couplings of the higgs are sm - like @xcite .
figure [ fit1 ] displays the results of this fit in the @xmath38 versus @xmath39 plane .
the best fit is obtained for @xmath143 , @xmath144 , and has @xmath145 for 19 degrees of freedom ( ) , giving a @xmath4-value of @xmath146 .
the results of this fit are summarized in table [ chisqmintable ] , together with the results of the other fits of this section . and
@xmath147 , assuming @xmath141 ( fit * i * ) .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point @xmath148 , @xmath149 .
[ fit1 ] ] we note that the sm ( @xmath141 , @xmath150 ) has @xmath151 and is hence more than @xmath6 away from the best fit in fig .
the number of degrees of freedom for the sm fit is 21 , implying a @xmath4-value of @xmath152 .
the largest @xmath52 contributions come from the @xmath153 channels from atlas ( @xmath154 ) , cms ( @xmath155 ) and tevatron ( @xmath156 ) , followed by the vbf result for @xmath139 from cms with @xmath157 .
next , we let @xmath60 , @xmath61 , @xmath62 vary , assuming there are no new particles contributing to the effective higgs couplings to gluons and photons , we take @xmath158 implying @xmath159 , @xmath160 as computed from the sm - particle loops .
the results for the one - dimensional and two - dimensional @xmath52 distributions are shown in figs .
[ fit2 - 1d ] and [ fit2 - 2d ] .
the value of @xmath34 is rather well determined to be close to unity .
it is intriguing that the best fit of @xmath34 is indeed just slightly below 1 , as any model with only higgs doublets or singlets requires @xmath161 .
the best fit values for @xmath29 and @xmath28 are sm - like in that they have magnitudes that are close to one . however , the best fit @xmath28 value is opposite in sign to the sm higgs case .
the preference for @xmath162 is at the level of @xmath163 see the first plot in fig .
[ fit2 - 1d ] .
this results from the fact that an enhanced @xmath3 rate ( as observed in the experimental data ) is obtained by changing the sign of the top - loop contribution so that it adds , rather than subtracts , from the @xmath164-loop .
in contrast , in the case of @xmath29 almost equally good minima are found with @xmath165 and @xmath166 .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane are given in table [ tab : fit2 ] .
note that , for the best fit point , the resulting @xmath98 and @xmath97 are in good agreement with the result of fit * i * above , for which @xmath167 and @xmath168 . here , however , the enhanced @xmath47 value derives from @xmath162 rather than from @xmath169 .
the best fit results are again tabulated in table [ chisqmintable ] .
distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
[ fit2 - 1d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
[ fit2 - 1d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
[ fit2 - 1d ] , title="fig : " ] distributions for the three parameter fit , fit
* ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
[ fit2 - 1d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane can be found in table [ tab : fit2 ] .
[ fit2 - 2d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane can be found in table [ tab : fit2 ] .
[ fit2 - 2d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane can be found in table [ tab : fit2 ] .
[ fit2 - 2d ] , title="fig : " ] + distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane can be found in table [ tab : fit2 ] .
[ fit2 - 2d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane can be found in table [ tab : fit2 ] .
[ fit2 - 2d ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , of @xmath60 , @xmath61 , @xmath62 with @xmath170 and @xmath171 as computed in terms of @xmath172 .
the red , orange and yellow ellipses show the 68% , 95% and 99.7% cl regions , respectively .
the white star marks the best - fit point .
details on the minima in different sectors of the ( @xmath28,@xmath29 ) plane can be found in table [ tab : fit2 ] .
[ fit2 - 2d ] , title="fig : " ] a negative sign of @xmath60while maintaining a positive sign of @xmath173is actually not easy to achieve .
( a sign change of both @xmath60 and @xmath173 would have no impact on the top quark induced loop amplitudes . )
it would require that @xmath173 is induced dominantly by the vev of a higgs boson which is _ not _ the higgs boson considered here .
hence , we have @xmath174 in most models , implying that it is important to study the impact of this constraint on our fits . the fit results when requiring @xmath175
are shown in the left two plots of fig .
[ fit2 - 1dpos ] and the top row of fig .
[ fit2 - 2dpos ] ; see also table [ tab : fit2 ] .
we observe that for this quadrant the results are consistent with sm expectations ( within @xmath176 ) .
interestingly the fit is not better than the sm itself : @xmath177 for @xmath178 , corresponding to @xmath179 .
distributions for the three parameter fit , fit * ii * , but imposing @xmath180 , @xmath181 ; the left two plots allow for @xmath182 ( @xmath183 ) , while in the right two plots @xmath184 ( @xmath185 ) .
[ fit2 - 1dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , but imposing @xmath180 , @xmath181 ; the left two plots allow for @xmath182 ( @xmath183 ) , while in the right two plots @xmath184 ( @xmath185 ) .
[ fit2 - 1dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , but imposing @xmath180 , @xmath181 ; the left two plots allow for @xmath182 ( @xmath183 ) , while in the right two plots @xmath184 ( @xmath185 ) .
[ fit2 - 1dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , but imposing @xmath180 , @xmath181 ; the left two plots allow for @xmath182 ( @xmath183 ) , while in the right two plots @xmath184 ( @xmath185 ) .
[ fit2 - 1dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , as in fig .
[ fit2 - 2d ] but with @xmath180 , @xmath181 , @xmath33 .
the upper row of plots allows for @xmath186 , while in the lower row of plots @xmath161 is imposed .
[ fit2 - 2dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , as in fig .
[ fit2 - 2d ] but with @xmath180 , @xmath181 , @xmath33 .
the upper row of plots allows for @xmath186 , while in the lower row of plots @xmath161 is imposed .
[ fit2 - 2dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , as in fig .
[ fit2 - 2d ] but with @xmath180 , @xmath181 , @xmath33 .
the upper row of plots allows for @xmath186 , while in the lower row of plots @xmath161 is imposed .
[ fit2 - 2dpos ] , title="fig : " ] + distributions for the three parameter fit , fit * ii * , as in fig .
[ fit2 - 2d ] but with @xmath180 , @xmath181 , @xmath33 .
the upper row of plots allows for @xmath186 , while in the lower row of plots @xmath161 is imposed .
[ fit2 - 2dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , as in fig .
[ fit2 - 2d ] but with @xmath180 , @xmath181 , @xmath33 .
the upper row of plots allows for @xmath186 , while in the lower row of plots @xmath161 is imposed .
[ fit2 - 2dpos ] , title="fig : " ] distributions for the three parameter fit , fit * ii * , as in fig .
[ fit2 - 2d ] but with @xmath180 , @xmath181 , @xmath33 .
the upper row of plots allows for @xmath186 , while in the lower row of plots @xmath161 is imposed .
[ fit2 - 2dpos ] , title="fig : " ] another possible model constraint is to require @xmath187 ( recall that @xmath188 by convention ) .
this constraint applies to any model containing only higgs doublets and singlets .
the 1d results for the combined requirement of @xmath189 and @xmath187 are shown in the right two plots of fig .
[ fit2 - 1dpos ] , and in the bottom - row plots of fig . [ fit2 - 2dpos ] .
we observe that the best fit values for @xmath28 and @xmath29 are only slightly shifted relative those found without constraining @xmath34 , and that accordingly the @xmath170 and @xmath171 at the best fit point are only slightly shifted .
however , the @xmath187 constraint does severely change the upper bound on @xmath47 , which for @xmath190 and @xmath191 mostly depends on the @xmath164-boson loop contribution .
the apparent sharpness of the boundary in the @xmath48 vs. @xmath47 plane is a result of the fact that these two quantities really only depend on @xmath28 for @xmath192 .
finally note that it has been shown in @xcite that single top production in association with a higgs is greatly enhanced when @xmath28,@xmath34 have opposite signs .
thus , the possibility of @xmath162 should be further scrutinized by precision measurements of the single top production cross section at the lhc .
.summary of results for fits * i**iii*. for fit * ii * , the tabulated results are from the best fit , cf . column 1 of table [ tab : fit2 ] .
[ cols="^,^,^,^,^",options="header " , ]
we are indebted to guillaume drieu la rochelle for help in extracting the signal strength information in the atlas @xmath20 and @xmath193 and the cms @xmath20 channels , and to albert de roeck for detailed discussions on the cms results .
moreover , we gratefully acknowledge discussions with ada farilla , jonathan hays , marumi kado , tom lecompte , koji nakamura and mayda velasco . this work was supported in part by us
doe grant de - fg03 - 91er40674 and by in2p3 under contract pics fr
usa no . 5872 .
ue acknowledges partial support from the french anr lfv - cpv - lhc , anr str - cosmo and the european union fp7 itn invisibles ( marie curie actions , pitn - ga-2011 - 289442 ) .
gb , ue , jfg and sk acknowledge the hospitality and the inspiring working atmosphere of the aspen center for physics which is supported by the national science foundation grant no .
gb , jfg and sk also thank the galileo galilei institute for theoretical physics for hospitality and the infn for partial support .
jfg acknowledges the generous hospitality while completing this work of the kavli institute for theoretical physics which is supported by the national science foundation under grant no .
nsf phy11 - 25915 .
99 g. aad _ et al . _ [ atlas collaboration ] , `` observation of a new particle in the search for the standard model higgs boson with the atlas detector at the lhc , '' phys
. lett .
b * 716 * ( 2012 ) 1 [ arxiv:1207.7214 [ hep - ex ] ]
. tevatron new physics higgs working group and cdf and d0 collaborations , `` updated combination of cdf and d0 searches for standard model higgs boson production with up to 10.0 fb@xmath194 of data , '' arxiv:1207.0449 [ hep - ex ] .
t. aaltonen _ et al . _
[ cdf and d0 collaborations ] , `` evidence for a particle produced in association with weak bosons and decaying to a bottom - antibottom quark pair in higgs boson searches at the tevatron , '' phys .
* 109 * ( 2012 ) 071804 [ arxiv:1207.6436 [ hep - ex ] ] .
lhc higgs cross section working group , a. david , a. denner , m. duehrssen , m. grazzini , c. grojean , g. passarino and m. schumacher _ et al .
_ , `` lhc hxswg interim recommendations to explore the coupling structure of a higgs - like particle , '' arxiv:1209.0040 [ hep - ph ] .
|
performing a fit to all publicly available data , we analyze the extent to which the latest results from the lhc and tevatron constrain the couplings of the higgs boson - like state at @xmath0 . to this end
we assume that only standard model ( sm ) particles appear in the higgs decays , but tree - level higgs couplings to the up - quarks , down - quarks and vector bosons , relative to the sm are free parameters .
we also assume that the leptonic couplings relative to the sm are the same as for the down - quark , and a custodial symmetry for the @xmath1 couplings . in the simplest approach ,
the effective higgs couplings to gluons and photons are computed in terms of the previous parameters .
this approach is also applied to two - higgs - doublet models of type i and type ii .
however , we also explore the possibility that the net higgs to @xmath2 and @xmath3 couplings have extra loop contributions coming from beyond - the - standard model physics .
we find that the sm @xmath4-value @xmath5 is more than @xmath6 away from fits in which : a ) there is some non - sm contribution to the @xmath3 coupling of the higgs ; or b ) the sign of the top quark coupling to the higgs is opposite that of the w coupling . in both these cases @xmath4-values @xmath7 can be achieved . since option b ) is difficult to realize in realistic models , it would seem that new physics contributions to the effective couplings of the higgs are preferred .
lapth-061/12 + lpsc12350 + lpt orsay 12 - 119 + nsf - kitp-12 - 236 g. blanger@xmath8 , b. dumont@xmath9 , u. ellwanger@xmath10 , j. f. gunion@xmath11 , s. kraml@xmath9
|
in order to improve , e. g. , the performance of nitride - based laser diodes , the density of threading dislocations penetrating into the corresponding layer systems need to be further reduced .
this is a difficult task , when growing such device structures on foreign substrates , e. g. , on sapphire , silicon carbide , or silicon due to a strain - related generation of extended defects .
therefore , the interest in the fabrication of bulk gan substrates as well as of free - standing gan layers ( fs - gan ) for a subsequent homo - epitaxial deposition of device structures is growing .
hydride vapor phase epitaxy ( hvpe ) has already proven to be well suited for the fabrication of fs - gan with threading dislocation densities below 10@xmath3 @xmath4 .
@xcite commonly , sapphire wafers are used as substrates for the hvpe process to grow fs - gan .
@xcite however , due to the low cost , available large wafer size , and well - established chemical etching techniques , silicon is very attractive to serve as a substrate material for the growth of fs - gan .
so far , two key problems , which depend on each other , have prevented a successful use of silicon as a substrate for the hvpe growth of fs - gan .
first , there is a high lattice mismatch ( 17@xmath5 ) and a large difference in the thermal expansion coefficient ( 56@xmath5 ) between gan and silicon giving rise to cracking of the gan during the growth as well as during the cooling of the gan substrate from the growth temperature down to room temperature .
second , silicon is not stable in the reactive ambiance and at the high temperatures applied during the hvpe process , where the reaction between silicon and the chlorine - based growth chemistry often results in a mixture of silicon , ga , and n including sin@xmath6 phases ( silicon - melt - back etching ) instead of the intended gan . @xcite aln buffer layers and a two - step growth mode with first a low - temperature ( low-@xmath1 ) step have been used to suppress the silicon - melt - back etching . also selective - area growth ( sag ) , where the templates for the hvpe consist of a few 100 @xmath0 m large growth windows to prevent cracking , has been applied .
@xcite however , the optimization of the protecting interlayer is very critical . for too thin aln buffer layers ,
the reaction with silicon hinders an epitaxial growth of gan . for too thick buffer layers ,
micro - cracks can appear during the growth , which again open a way for a reaction of si with ga or n. thus , although the growth of a few ten @xmath0 m thick gan layers on silicon by hvpe has already been reported about 10 years ago @xcite , a long - term hvpe growth of gan on silicon resulting in substrates with a thickness of several hundred @xmath0 m is still challenging .
furthermore , the use of sag or epitaxial lateral overgrowth techniques require micrometer masking , ex situ sio@xmath7 or sin@xmath6 deposition and multi - step photolithography , which increase the process complexity and costs .
hersee @xmath8@xcite proposed a nano - heteroepitaxy technique , taking advantage of three - dimensional stress relief mechanisms , which is expected to result in a further improvement of gan epilayers grown on foreign substrates . based on this method , several authors have discussed a new non - lithographic nano - heteroepitaxy approach relying on self - assembled nanostructures in order to reduce the process steps and costs .
@xcite kwon @xmath9 @xcite have used a 50 nm thick sputtered aln buffer on silicon , on which they have grown self - assembled nano - needles and -rods during a first low-@xmath1 hvpe step at 600 and 650 @xmath10c , respectively .
subsequently , 0.52
@xmath0 m thick gan layers have been overgrown in a second hvpe step at 1040 @xmath10c .
the optical near - band gap emission of these gan layers are red - shifted with regard to the expected spectral position , the origin of which has not been analyzed in detail . moreover
, longer high-@xmath1 hvpe runs to obtain thick gan layers have not been carried out . in this work
, we are reporting on the hvpe growth of thick gan layers on self - assembled gan nanostructures , the latter having been deposited on si(111 ) substrates with aln buffer layers using plasma - assisted molecular - beam epitaxy ( mbe ) .
the hvpe growth process is investigated by spatially and spectrally resolved cathodoluminescence ( cl ) spectroscopy in a scanning electron microscope ( sem ) at cross - sections of the layers . by adding gan nano - structures as nucleation sites and applying a two - step hvpe growth process , silicon incorporation into several ten @xmath0 m thick gan layers could be prevented .
three templates have been prepared by mbe for the hvpe overgrowth to obtain several ten to several hundred @xmath0 m thick gan layers on silicon . for sample s1 ,
a commercial kyma - si / aln template consisting of an n - type si(111 ) wafer and a 200 nm thick al - face aln buffer layer have been used . on top of this template ,
gan has been deposited by mbe at a temperature of 800 @xmath10c setting the iii / v flux ratio to 0.5 . as a result , we obtained gan nanostructures consisting of highly dense 0.6 @xmath0 m high nano - walls and of taller nano - wires ( nw ) with a low density as shown in the sem image of s1 in figure [ fig1 ] ( a ) and indicated by white arrows in the top view and cross - sectional image of figure [ fig1 ] ( a ) .
the very thin gan nws visible in figure [ fig1 ] ( a ) , which emerge from the gan matrix , resemble a lot the ones observed by fernndez garrido @xmath9 @xcite on high quality al - polar aln layers grown on sic ( 0001 ) .
for this reason we believe that the results obtained by fernndez garrido @xmath9 are also valid for sample s1 , that is the very thin gan nws are n - polar , whereas the gan matrix is ga - polar .
the template of samples s2 and s4 consisted of an n - type si(111 ) wafer and a 25 nm thick aln buffer layer grown by mbe at a temperature of 680 @xmath10c with a iii / v flux ratio close to unity .
this si / aln template was transferred to another mbe chamber to add gan nanostructures , which have been grown at 780 @xmath10c using a iii / v flux ratio of 0.31 .
the resulting nanostructures consist of a rough parasitic layer , straight nws with low density , and a large density of tilted nano - leafs [ cf .
sem image of s2 and s4 in figure [ fig1 ] ( b ) ] .
the template of the reference sample s3 consists of a si(111 ) wafer and a 15 nm thick aln layer grown by mbe at a temperature of 680 @xmath10c setting the iii / v flux ratio close to unity .
as shown in the atomic force microscopy ( afm ) image of figure [ fig1 ] ( c ) , the aln layer of s3 is almost contineous with scattered pits .
the layer is characterized by a dense distribution of small islands and a root - mean - square value of the roughness of about 0.8 nm . further details about the growth parameters and microstructure of comparable aln layers can be found elsewhere .
@xcite [ 4 cm ] the described templates showing very different surface structures were overgrown by thick gan layers using a home - made horizontal reactor hvpe system . during the gan growth , n@xmath11 and h@xmath11 were used as a carrier gas with a ratio of 1:1
the temperature of the ga boat was set to about 850 @xmath10c .
the samples s1 to s3 were overgrown separately , but using the same conditions .
we applied a two - step growth mode approach , which consisted of a low-@xmath1 ( 800 @xmath10c ) and a subsequent high-@xmath1 ( 950 @xmath10c ) step . during the first hvpe step , a lower temperature together with low hcl and nh@xmath12 fluxes have been chosen in order to prevent the evolution of silicon - melt - back etching .
the hcl and nh@xmath12 fluxes amounted to 0.01 and 0.6 slm during the first step and to 0.04 and 1.8 slm during the second step .
for s1 and s2s4 , the growth time of the low-@xmath1 step amounted to 20 and 50 minutes , respectively .
the high-@xmath1 step lasted for 30 minutes for all samples . in the case of sample s4 ,
an additional 360 minutes long high-@xmath1 step has been performed at 1063 @xmath10c in order to obtain a gan substrate of several hundred @xmath0 m thickness . in order to provide a more clear overview of the layer parameters and growth conditions ,
we summarize the thickness values of the used aln buffers and of the hvpe gan layers as well as the hvpe growth temperatures and corresponding growth times of the samples s1 to s4 in table [ tab1 ] .
.thickness of the aln buffer layer ( @xmath13 grown by mbe , growth temperature and time of the low-@xmath1 hvpe step ( @xmath14 and @xmath15 ) , growth temperature and time of the high-@xmath1 hvpe steps ( @xmath16 and @xmath17 , @xmath18 and @xmath19 ) , thickness of the gan layer after the low-@xmath1 hvpe step ( @xmath20 ) , and total thickness of the hvpe gan ( @xmath21 ) for the samples s1 to s4 . [ cols="<,^,^,^,^",options="header " , ] the morphology of all of these samples was imaged by means of a field - emission sem or by means of an afm .
the spatial distribution of the luminescence intensity as well as luminescence spectra from the cross - sections of the overgrown gan layers were measured using a cl system attached to the sem .
the cross - sections were prepared by using an ar - ion cutting machine .
sem and monochromatic cl images were acquired simultaneously for an accurate assignment of the local origin of the cl . a he - cooling stage allowed to vary the sample temperatures between 6 and 300 k. the cl system was operated with a photomultiplier tube for both monochromatic imaging and to record cl spectra . for the acquisition of both cl spectra and cl images , the acceleration voltage and the current of the electron beam were chosen to be 5 kv and about 1 na , respectively .
after the hvpe growth , the wafers were slowly cooled down from the growth temperature to room temperature over a period of 4 hours .
nevertheless , cracks appeared within the gan layers of s1 to s3 caused by the large tensile strain due to the large difference of the thermal expansion coefficients between silicon and gan .
the typical crack density amounts to several cracks per @xmath22 .
for s1 to s3 , the thickness of the hvpe gan layer amounts to 25 to 35 @xmath0 m .
as confirmed by an energy dispersive x - ray ( edx ) analysis at the cross - sections of these hvpe gan layers ( not shown ) , silicon - melt - back etching was successfully suppressed during the hvpe process .
figure [ fig2 ] shows the sem images of the surface of the hvpe gan layers . from sample to sample , the sem images exhibit a varying density and size of v - pits .
the v - pit density amounts to 6.5@xmath2310@xmath24 @xmath4 for s1 , 3@xmath2310@xmath24 @xmath4 for s2 and 1.4 @xmath2310@xmath25 @xmath4 for s3 .
the origin of these v - pits is discussed below .
using photoelectron diffraction to determine the polarity of the hvpe - gan layers @xcite , it has been confirmed that the layers are ga - polar ( not shown here ) . [ 4 cm ] in order to investigate the spatial distribution of the luminescence intensity within the hvpe layers , in particular , along the growth direction , we performed cl measurements at the cross - sections prepared by ion cutting .
figures [ fig3](a ) and ( b ) show normalized cl spectra at 6 k , which have been excited within regions located at the upper part of the cross - sections and within regions close to the aln interlayer of s1s3 , respectively , as indicated by the solid rectangles in the inset of figure[fig3](a ) . , i@xmath26 , and dx denote the spectral positions of the excitons bound to extrinsic , i@xmath7- as well as i@xmath26-type basal plane stacking faults , and the donor - bound exciton in gan , respectively.,width=377 ] luminescence spectra indicate sensitively the presence of extended defects such as inversion domain boundaries ( idb ) and stacking faults . according to the recent findings of auzelle @xmath9 @xcite , the presence of idbs in the hvpe - gan layers should result in relatively intense luminescence at a spectral position of about 3.45 ev .
however , such spectral line does not seem to be present in the cl spectra of samples s1s3 , as shown in figure [ fig3 ] , or at least its intensity is very weak .
the absence of the cl line at 3.45 ev indicates that the hvpe - gan exhibits a uniform ga - polarity .
furthermore , the cl spectra of the upper part of all hvpe gan layers are dominated by the donor - bound exciton ( dx ) recombination centered at about 3.475 ev .
however , the spectrum of s1 is significantly broadened toward the low - energy side . for the lower part
close to the aln / gan interface the cl spectra of s1 and s2 are dominated by luminescence lines centered at energy values below 3.44 ev .
we assign these lines to excitons bound to basal plane stacking faults ( i@xmath26 sf , i@xmath7 sf , and e sf ) .
@xcite for sample s3 , the cl spectrum of the lower part of the cross - section shows also significant contributions of sf - related luminescence , which are not as dominant as in s1 and s2 .
in contrast to s1 and s2 , the cl spectrum of the lower part of s3 clearly shows a shoulder at an energy value exceeding the one of the gan band edge [ cf .
dashed arrow in figure [ fig3](b ) ] . as will be discussed below
, the origin of this high - energy shoulder could be a high unintentional doping of the gan layer close to the si / gan interface due to silicon incorporation .
the spatial origin of the various parts of the cl spectra can be precisely determined by acquiring cl maps for the corresponding cl detection energies @xmath27 .
the ion - milled edges of the samples show smooth surfaces [ cf .
inset of figure [ fig3](a ) ] allowing for a mapping of the cl intensity without artifacts caused by a rough topography due to a typically non - uniform cleavage of gan .
figure [ fig4](a ) shows the cl map of the cross - section of sample s1 recorded for @xmath28 ev , which reveals that the donor - bound exciton emission originates from an about 20 @xmath0 m thick upper region of the hvpe gan , while the about 5 @xmath0 m thick bottom layer does not show any significant donor - bound exciton cl .
however , this region at the bottom of the hvpe gan shows significant emission for values of @xmath27 lower than 3.45 ev as shown in figures [ fig4](b)4(d ) , which we assign to sf - related luminescence .
thus , the cl maps clearly indicate two sharply separated regions along the growth direction , where the sf - related luminescence is found only within a 5 @xmath0 m thick bottom layer , while near - band gap luminescence originates mainly from the upper , thick part of the hvpe gan layer .
the near - band gap luminescence of the upper region shows an inhomogeneous distribution with dark vertical lines and bands resulting from non - radiative recombination at threading dislocations and/or grain boundaries , the density of which decreases along the growth direction . .
@xmath28 ev corresponds to the donor - bound exciton in gan .
the dashed lines mark the gan / si interface.,width=377 ] since the formation of sfs in gan is known to be promoted at low growth temperatures @xcite as well as for growth along directions different from the @xmath2-direction @xcite , we conclude that the 5 @xmath0 m thick sf - rich bottom region of the gan layer is a result of the epitaxy during the first , low-@xmath1 hvpe step . due to the typically high growth rate of the hvpe process ,
a temperature of 800 @xmath10c is not sufficiently high to prevent instabilities of the crystallographic phase leading to the evolution of sfs or even of inclusions of the cubic phase of gan .
moreover , at the beginning of the low-@xmath1 growth step , the nucleation of the hvpe gan on the nanostructures is followed by lateral growth , until coalescence is completed , with the latter also being known to lead to the formation of sfs .
@xcite note that both the cl spectrum of figure [ fig3](a ) and the cl map of figure [ fig4](b ) of sample s1 show significant emission for @xmath29 ev also within the upper thick part of the hvpe gan .
however , the spatial distribution of the emission at this energy differs significantly from the one of the region at the bottom of the hvpe gan , and , in contrast to the sf emission , comes from an area also showing dx emission , thus indicating a different origin of this emission .
we clearly observe large v - shaped and small column - like contrasts of the cl mapping throughout the upper and lower regions of the layer , respectively .
figure [ fig5 ] displays a magnified cl image of the lower part of the cross - section of sample s1 for a detection energy @xmath29 ev . at the bottom
, the low-@xmath1 hvpe gan shows cl features , which are arranged in columns and are elongated along the basal plane .
the latter observation supports the assignment of this emission to sf - related luminescence . directly from the interface to the low-@xmath1 hvpe gan , the cl map of the high-@xmath1 hvpe gan is characterized by v - shaped features as indicated by dashed lines .
specifically , the apex of these features originates at the boundary between two of the columns situated beneath .
as described recently by lee @xmath9 , @xcite v - shaped intensity distributions in cl maps of the cross - section of thick hvpe gan layers are caused by the evolution of @xmath30 facets initiated by threading dislocations .
these facets , which form the observed v - pits , are nitrogen - terminated and promote the incorporation of oxygen atoms during growth .
@xcite since oxygen atoms give rise to a donor level with a donor - to - valence band transition at about 3.42 ev @xcite , the near - band edge cl line of the upper region of s1 is broadened toward the low - energy side , and the motion of the growth front of the @xmath30 facets can be clearly imaged for @xmath29
ev . according to lee @xmath31 @xcite the v - pits can get filled during the growth by the evolution of faster growing facets at the bottom of the @xmath30-facetted v - pits giving rise to the observation of v- and u - shaped forms with varying opening angles . in figure
[ fig5 ] , the marked angles 1 and 2 amount to 62 and 55@xmath10 with respect to the basal plane , respectively . ev . at the bottom
, the firstly grown gan shows sf - related cl features , which are arranged as columns . just above this layer ,
the cl map is characterized by v - shaped features as indicated by the dashed lines .
the marked angles 1 and 2 amount to 62 and 55@xmath10 , respectively.,width=264 ] for the hvpe overgrowth of s1 , we summarize the experimental results as follows . during the low-@xmath1 hvpe step ,
a columnar growth occurs with a mean column diameter of about 0.3 @xmath0 m . at the very
beginning of the subsequent high-@xmath1 hvpe step , v - pits evolve , the growth dynamics of which are governed by facet growth leading to an increase of the concentration of incorporated impurity atoms such as oxygen .
figures [ fig6](a ) and 6(b ) show cl images of identical regions from the cross - section of sample s2 for cl detection energies of @xmath28 and 3.425 ev , respectively .
again , we observe two distinct parts of the hvpe gan representing the hvpe grown at low @xmath1 and high @xmath1 , where the optical emission of the former is related to basal plane sfs , while the emission of the latter is related to the donor - bound exciton in gan . moreover , the cl distribution mapped at 3.425 ev is very weak throughout the whole upper part of the hvpe gan and does not show any significant features such as v - shaped patterns . an enlarged section of the superposition of the cl maps of figures [ fig6](a ) and 6(b ) is shown in figure [ fig6](c ) for a direct correlation . in the lower part of figure [ fig6](c ) , the cl intensity forms stripe segments oriented along the basal plane confirming again their basal plane sf - related origin .
these stripe segments are arranged within columns with an average diameter of 2 @xmath0 m , the borders of which are continued within the upper part of figure [ fig6](c ) ( cf . dashed arrows ) indicating a columnar growth of the gan along the @xmath2-direction during the high-@xmath1 hvpe step .
some of the dark border lines are disappearing along the growth direction due to coalescence of neighboring columns .
we do not observe any indication of the evolution of @xmath30 or other slowly growing facets , i. e. , of a v - pit - like growth at this stage of the epitaxy .
nevertheless , v - pits are also present at the surface of s2 as visible in figure [ fig6](a ) and shown in figure [ fig2](b ) .
these v - pits were , however , formed within a region close to the surface , where either a bundle of or single dark - line defects are found at their apex as shown in the cl image of figure [ fig7 ] .
the indicated angle of 62@xmath10 corresponds to @xmath30 facets . and
( b ) 3.425 ev , which correspond to the donor - bound exciton and exciton bound to the i@xmath26-type sf in gan , respectively .
the dashed lines mark the gan / si interface .
( c ) magnified section of a superposition of ( a ) and ( b ) . in ( c ) , the dashed upwards showing arrows mark the boundaries of the column - arranged sf - features at the bottom of the gan layer , which continue into the upper part of the gan ( downwards showing arrows ) , thus indicating a columnar growth along the @xmath2-direction.,width=377 ] a reason for the appearance of the v - pits at the surface could be a non - uniform decomposition of the gan during cooling down to room temperature after the hvpe has been terminated . since the cooling rate is usually low in order to minimize cracking of the layer , the sample remains for a long time at high temperatures .
if the decomposition rate is higher at positions , where threading dislocations penetrate the surface , compared with the surrounding surface regions , v - pits can be formed at dislocation positions as shown in figure [ fig7 ] .
ev showing v - pits at the surface .
the indicated angle corresponds to the @xmath30 facet of wurtzite gan.,width=264 ] comparing the above discussed cl data of s1 and s2 , we assume that the different morphology of the underlying mbe - grown nanostructure ( cf .
figure [ fig1 ] ) , where the density of the nanowalls of s1 is much higher than the density of nws or nanoleafs of s2 , resulted in a columnar growth with very different mean column diameters ( 0.3 @xmath0 m for s1 and 2 @xmath0 m for s2 ) during the low-@xmath1 hvpe step . according to kwon @xmath32 @xmath33 @xcite , a larger distance of the nanostructures serving as the template for a hvpe overgrowth promotes a lateral epitaxy and therefore results in larger grains during the first hvpe step .
the very different mean column diameters formed during the low-@xmath1 hvpe step resulted in a different growth dynamics for s1 and s2 during the subsequent high-@xmath1 hvpe step . while in the case of s1 , a @xmath30 faceted growth dominates , we observe a columnar growth along the @xmath2-direction throughout the whole gan layer for s2 . for the reference sample s3 ,
the cl distribution of the cross - section differs significantly from the ones of s1 and s2 . in figure [ fig8](a ) , the cross - section cl image of s3 obtained for @xmath28 ev shows near - band gap optical emission throughout the whole thickness without a clear distinction between the low-@xmath1 and high-@xmath1 hvpe growth regions .
a few dark horizontal lines appear close to the bottom of the gan layer .
figure [ fig8](b ) depicts low-@xmath1 cl spectra measured within the upper part of the cross - section of s3 and directly at the aln / gan interface .
the latter is significantly broadened and blue - shifted compared with the spectrum obtained far away from the interface .
correspondingly , the cl maps of figures . [ fig8](c ) to 8(e ) show bright luminescence features within an 1 to 2 @xmath0 m thick interface region for the whole considered spectral range even for @xmath34 ev , which clearly exceeds the band edge energy of gan . above this interface region of the gan layer , we observe some bright , sf - related stripe segments in the cl map for @xmath35 ev of figure [ fig8](e ) .
the broadening and blue - shift of the cl spectrum obtained close to the interface indicate a strong doping of the near - interface region most probably by silicon atoms from the substrate .
( b ) low-@xmath1 cl spectra excited at the upper part of the cross - section ( solid line ) and at the aln / gan interface ( dashed line ) of s3 .
( c)(e ) : magnified cl images of a cross - section region close to the aln / gan interface of sample s3 differing in the cl detection energy as indicated . the energy value @xmath34 ev exceeds the band edge energy of gan .
@xmath28 ev and @xmath35
ev correspond to the donor - bound exciton and exciton bound to the i@xmath36-type stacking fault , respectively .
the dashed line marks the interface to the silicon substrate.,width=453 ] since the aln buffer layer exhibits a certain roughness , its real thickness can be locally smaller than the nominally expected 15 nm . moreover , it is known from transmission electron microscope investigations that our mbe grown aln buffers are closed , but not single crystalline layers , and therefore exhibit grain boundaries .
the latter could serve as fast diffusion paths for si @xcite leading to a local incorporation of silicon into the gan layer on a strong dopant level .
a silicon concentration on a dopant level can not yet be detected by edx and is probably not a result of a silicon - melt - back etching , which usually leads to a much higher silicon content .
we suppose that due to the hvpe growth directly on top of the uncovered aln buffer layer , a locally high silicon doping of the gan takes place leading to a broad and blue - shifted optical emission spectrum , which we have also observed for aln buffer layers as thick as 200 nm for a 6 hours lasting hvpe process at 800 @xmath10c .
the appearance of sf - related cl features just above this interface region could also be a signature of a strong silicon doping .
@xcite in figure [ fig9 ] , we compare the cl spectra excited at the upper part of the cross - sections of s1 to s3 at room temperature .
s1 shows the highest cl intensity , but also the largest linewidth ( fwhm@xmath37 mev ) .
both the higher intensity and the larger fwhm are probably a result of doping due to a growth controlled by @xmath30 or other slowly growing facets .
the maximum cl intensity of s2 is about 10@xmath5 lower than the one of s1 , whereas the linewidth of the spectrum is significantly smaller ( fwhm@xmath38 mev ) and is close to the linewidth of a corresponding spectrum obtained from a fs - gan substrate , which exhibits a dislocation density as low as @xmath39 @xmath4 .
the linewidth of the cl spectrum of s3 is equal to the one of s2 , but the cl intensity is about 60@xmath5 lower than that of s2 , which is probably due to a higher density of dark - line defects in s3 as compared with s2 .
this difference in the dark - line density can hardly be quantified by means of the cl maps shown in figures .
[ fig6 ] and [ fig8 ] , but can be indirectly concluded by comparing the v - pit density at the surface , which amounts to @xmath40 @xmath4 for s2 and @xmath41 @xmath4 for s3 ( cf .
figure [ fig2 ] ) . from the comparison of the cl results of s1 and
s2 with the ones of s3 , we can conclude that the hvpe of gan on silicon covered by gan nanostructures in addition to an aln buffer layer can prevent the incorporation of silicon and results in a higher luminescence efficiency of the hvpe gan as compared with the growth directly on top of an aln buffer layer .
moreover , the template of s2 is more suitable for an undisturbed growth along the @xmath2-direction , which in turn prevents an undesired incorporation of impurity atoms at the growth front . the conclusion drawn above encouraged us to try a deposition of several 100 @xmath0 m gan using the same template and growth parameters as for s2 , but adding a 6 hours long high-@xmath1 hvpe step at 1063 @xmath10c . as a result
, we obtained sample s4 showing a mean thickness of the deposited gan of 600 @xmath0 m .
figure [ fig10](a ) shows an sem image of the cleaved edge of s4 . after the hvpe growth and cooling down process
, the silicon substrate was self - separated from the hvpe layer .
however , the surface of s4 is apparently rough , and the lower region of the hvpe layer shows a dark contrast in the sem image . as checked by edx , this dark part of the hvpe layer is not any more gan , but a mixture of ga , n , and silicon .
moreover , the cl maps of figures [ fig10](b ) and 10(c ) obtained from an ion - polished region of the cross - section of s4 indicate that the upper , several 100 @xmath0 m thick
gan does not exhibit a single , but a polycrystalline structure consisting of several 10 to 100 @xmath0 m large crystals .
the bright stripes visible in the cl map of figure [ fig10](c ) , which are complementary to the luminescence in figure [ fig10](b ) , correspond to luminescence of excitons bound to i@xmath26-type basal plane sfs and are therefore oriented perpendicular to the @xmath2-direction .
consequently , the @xmath2-axies of the grains , which can be recognized in the cl maps , are tilted to each other by several degrees .
the inset of figure [ fig11 ] shows a large polished piece of s4 , whereas the main graph depicts the room temperature cl spectra of this polished piece and the one of the mentioned fs - gan sample .
both the intensity and the linewidth of the near - band edge cl of s4 are similar to the ones of the fs - gan sample , where the fwhm amounts to 75 mev . the cl intensity at the spectral range of the `` yellow luminescence '' ( centered at about 2.3 ev ) is more than 3 orders of magnitude lower than the near - band edge cl , and therefore , as low as for the fs - gan .
the only significant difference between the two spectra is the appearance of a weak , broad band in the spectrum of s4 centered at about 2.9 ev .
the origin of this luminescence band is most probably associated with the presence of a high density of sfs and possibly cubic segments as visible in the low - temperature cl map of the cross - section of s4 in figure [ fig10](c ) .
using a two - step hvpe process , we have grown several ten @xmath0 m thick gan layers on si(111 ) substrates covered by an aln buffer as protection layer , both without and with additional gan nanostructures as nucleation sites .
a 15 nm thick aln buffer layer in combination with a low-@xmath1 hvpe step at 800 @xmath10c can successfully suppress a silicon - melt - back etching .
the gan layer deposited directly on the aln buffer by hvpe without mbe - grown gan nanostructures shows a significant broadening and blue - shift of the luminescence spectrum , when excited close to the aln / gan interface , which indicates a strong doping of the interface region due to silicon incorporation via fast diffusion paths provided by grain boundaries of the aln buffer . when additional gan nanostructures are deposited by mbe on the aln buffer , the nucleation and growth during the subsequent low-@xmath1 hvpe step results in a columnar growth of the gan , where the column diameter is controlled by the morphology of the nanostructure .
due to the relatively low growth temperature as well as to the lateral growth after nucleation , the gan layer region of the low-@xmath1 hvpe step contains a high density of stacking faults , which are no longer present within the region of the high-@xmath1 hvpe growth . when the diameters of the columns formed during the low-@xmath1 hvpe step amount to a few 100 nm , the growth dynamics of the subsequent high-@xmath1 hvpe step is characterized by the formation of v - pits at a very early stage , which in turn is associated with an incorporation of impurity atoms . for diameters of the columns formed during the low-@xmath1 hvpe step on the order of a few @xmath0 m ,
the growth dynamics of the subsequent high-@xmath1 hvpe step is characterized by straight columnar growth along the @xmath2-axis throughout the whole gan layer .
room temperature luminescence spectra of such gan layers are very similar to spectra obtained from fs - gan with respect to both intensity and fwhm .
a several hours long high-@xmath1 hvpe step resulted in several hundred @xmath0 m thick gan layers , which show , however , a silicon contamination at their bottom part , a high density of sfs throughout the whole layer , and a polycrystalline structure .
our future research will be focused on the origin of the silicon incorporation during a long - time , high-@xmath1 hvpe process and the necessary template structure as well as growth parameters to prevent the reaction of silicon with the hvpe gan .
we thank oliver brandt for fruitful discussions and raffaella calarco as well as holger t. grahn for a careful reading of the manuscript .
moreover , we thank o. romanyuk for the investigation of the polarity of the hvpe gan by photoelectron diffraction and we gratefully acknowledge financial support of this work by the sino - german center for science promotion through the project gz 736 .
|
several ten @xmath0 m gan have been deposited on a silicon substrate using a two - step hydride vapor phase epitaxy ( hvpe ) process .
the substrates have been covered by aln layers and gan nanostructures grown by plasma - assisted molecular - beam epitaxy . during the first low - temperature ( low-@xmath1 )
hvpe step , stacking faults ( sf ) form , which show distinct luminescence lines and stripe - like features in cathodoluminescence images of the cross - section of the layers .
these cathodoluminescence features allow for an insight into the growth process . during a second high - temperature ( high-@xmath1 ) step ,
the sfs disappear , and the luminescence of this part of the gan layer is dominated by the donor - bound exciton . for templates consisting of both a thin aln buffer and gan nanostructures , a silicon incorporation into the gan grown by hvpe is not observed .
moreover , the growth mode of the ( high-@xmath1 ) hvpe step depends on the specific structure of the aln / gan template , where in a first case , the epitaxy is dominated by the formation of slowly growing facets , while in a second case , the epitaxy proceeds directly along the @xmath2-axis . for templates without gan nanostructures , cathodoluminescence spectra excited close to the si / gan interface
show a broadening toward higher energies indicating a silicon incorporation on a high dopant level .
|
ngc 2146 is a starburst galaxy located at a distance of 14.5 mpc ( 1@xmath9 70 pc ) .
its infrared luminosity measured with iras at 60 @xmath10 and 100 @xmath10 is @xmath11 @xmath12 .
this value and the large 25 @xmath10 to 60 @xmath10 flux ratio place the object in the lower part of the ir luminosity / spectral index plane populated by superluminous galaxies without active nuclei ( hutchings et al .
ngc 2146 has a central molecular ring ( jackson & ho @xcite , young et al .
@xcite ) , and an outflow of hot gas along the minor axis driven by supernova explosions and stellar winds in the starburst region ( armus et al .
@xcite , della ceca et al .
these characteristics reveal a strong similarity to the prototype starburst galaxy m82 ( at a distance of 3.2 mpc and with an infrared luminosity of @xmath13 @xmath12 ) , although ngc 2146 does not have a companion that may have triggered the starburst ( fisher @xmath0 tully @xcite ) .
it has been suggested that the starburst in ngc 2146 is the result of a far evolved merger ( condon et al .
@xcite ; young et al .
@xcite , hutchings et al .
@xcite ) , but a fully convincing kinematic and material trace of the merger has not yet been found .
+ the nuclear region is partly obscured by a strongly absorbing dust lane ( benvenuti et al .
@xcite ) . the high extinction ( a@xmath14 @xmath5 5 mag ; young et al . @xcite ; hutchings et al .
@xcite ; smith et al .
@xcite ) makes the optical observation of the inner star - forming regions impossible .
+ strong non thermal radio emission from the centre of ngc 2146 has been detected by kronberg & biermann ( @xcite ; hereafter kb ) , by condon et al .
( @xcite ) , and by lisenfeld et al .
( @xcite , @xcite ) .
about 20 point sources were detected at 5 ghz in the central 800 pc of ngc 2146 by glendenning & kronberg ( @xcite ) using the vla , and these were interpreted as radio supernovae ( rsn ) or supernova remnants ( snrs ) . in this paper we report the detailed spatial distribution of these point sources , together with their number and fluxes derived from radio continuum observations at 5 ghz obtained with merlin and the vla . from additional 1.6 ghz merlin observations we obtained the spectral index of 9 of these sources , which allow to identify their nature .
our results open the possibility of a direct comparison with similar observations of m82 ( kronberg et al .
@xcite ; muxlow et al .
@xcite ) and other strong and nearby starburst galaxies , such as ngc1808 , ngc4736 and ngc5253 ( saikia et al .
@xcite ; duric & dittmar @xcite ; turner et al .
cccccc + map & frequency & observation & weighting & beam & noise + code & ( ghz ) & & & hpbw(@xmath15 ) & ( @xmath16 ) + + + a & 5 & vla & purely uniform@xmath17 & @xmath18 & @xmath19 + b & 5 & merlin & purely uniform@xmath17 & @xmath20 & @xmath21 + c & 5 & merlin+ & uniform & @xmath22 & @xmath23 + & & vla a - array & & & + d & 1.6 & merlin & purely natural@xmath24 & @xmath25 & @xmath23 + + + + @xmath17 : using the _ robust _ parameter = -5 ; @xmath24 : using the _ robust _ parameter = + 5 [ aips task imagr ] .
ngc 2146 was observed for 23 h with merlin ( 6 antennas ) in november 1996 ( one day ) and february 1997 ( two days ) .
the observing frequency was 4.994 ghz ( @xmath26 = 6 cm ) , with a bandwidth of 15 mhz in both circular polarizations ; the data were taken in spectral - line mode ( 15 @xmath27 1mhz channels ) .
oq208 ( 2.39 jy ) and 0602 + 780 ( 0.126 jy ) were used as flux and phase calibration sources , respectively .
+ ngc 2146 was observed in the l - band with merlin ( 6 antennas , including the lovell telescope ) in may 1997 , for a period of 15 h. the observing frequency was 1.658 ghz ( @xmath26 = 18 cm ) , with a bandwidth of 15 mhz in both circular polarizations .
the passband and relative gains of the antennas were determined using the point source calibrator 0552 + 398 , with a flux density of 2.31 jy ; the compact source 0602 + 780 ( 0.109 jy ) was used to determine the telescope phases .
+ moreover , we used observations of ngc 2146 taken in september 1985 with the vla ( 27 antennas ) in the a , b and c array configurations ; the observing frequency was 4.885 ghz , with an integration time of 18 h. + the 5 ghz data from merlin and the vla a - array were combined to increase the brightness sensitivity and @xmath28 coverage , in order to improve the information about the weaker sources already marginally detected in the 5 ghz merlin observation . to combine the two data sets it was necessary to shift the merlin phase reference position slightly .
a re - weighting was also necessary to compensate for different origins of the data sets .
several images were produced using the aips task imagr , and deconvolved with the clean algorithm ( hgbom @xcite ) .
the details of these maps such as restoring beam , the applied weighting , and the rms noise , are summarized in table [ maps ] .
the rms noise in source - free areas of the images is consistent with the expected thermal noise levels .
+ since the wardle telescope was not present in the observations , the length of the shortest baseline of the merlin array at 1.6 ghz is 9.2 km ( darnhall lovell telescopes ) . at this frequency
the interferometer is then not sensitive to structures larger than @xmath29 ( 280 pc ) . at 5 ghz
the shortest baseline of the merlin array is 9.5 km ( darnhall mark2 telescopes ) , so that @xmath30 = @xmath31 ( 90 pc ) .
the combination of the 5 ghz merlin data with the vla data gives a shortest baseline of 0.68 km , so that @xmath32 = @xmath33 ( 1250 pc ) .
the short - spacing effect does not influence the results at either frequency since the size of the sources we are searching for is smaller than @xmath34 ( 100 pc ) .
furthermore , most of the sources are either unresolved or have extensions of a few beam widths and hence the flux measurement is not influenced by the lack of short spacings .
figure [ vla-5 ] shows the uniformly weighted abc - array vla map ( a ) of ngc 2146 . due to the higher resolution ( 037 @xmath27 032 instead of 12 @xmath27 12 )
the comparison of this map with that of kb reveals a larger number of compact radio sources ( @xmath35 18 instead of 5 ) .
any feature above 0.1 mjy ( 3@xmath36 level ) in the 5 ghz merlin+vla map ( c ) is taken to be a source .
our map lacks the long tongue of emission to the se , seen in the map of kb .
the global features found by kb , such as the s - shape of the central emission region ( by kb assumed to be a triple source , but now resolved into a larger number of sources ) , its position angle ( @xmath35 105@xmath37 ) , and the main bar of radio emission ( @xmath35 128@xmath37 ) , are consistent with those in our map .
convolution of our vla map to the 12 resolution of the kb map shows that the peak brightness is of the order of 6 mjy / beam as measured by kb , and that the se tongue of emission is still present , though less extended than in the earlier maps . from the map shown in fig .
[ vla-5 ] it is evident that the point sources present in the se part of the galaxy are more numerous than in the nw part ( 13 compared to 4 , assuming that the centre of symmetry of ngc 2146 is near the strong central source 37.6 + 24.2 ) .
the extended emission of the se region is also more pronounced , giving the impression of an asymmetry in ngc 2146 , in favour of a stronger starburst activity along the se arm .
@xmath38 interferometer observations ( young et al .
1988 ) show a higher concentration of molecular gas in this part of the galactic disk . + table [ observed ] contains the positions and the fluxes of the eighteen compact sources derived from our observation , listed by coordinate name with increasing r.a .
, following the convention of kronberg & wilkinson ( @xcite ) .
the reference position is at @xmath39 and @xmath40 ( j2000 ) .
the coordinates of the emission peaks were determined from a gaussian fit of each component using the aips task jmfit ( which also gives the absolute positional uncertainties ) in the uniformly weighted 5 ghz merlin+vla map ( c ) . from this map
we also derived the peak and integrated flux densities at 5 ghz ( s@xmath41 ) .
the same quantities were derived at 1.6 ghz ( s@xmath42 ) from the naturally weighted merlin map ( d ) .
the numbers in parentheses are the errors of the derived quantities .
only nine sources are detected at 1.6 ghz , due to a lack of sensitivity in the merlin observations , and possibly also to the steep positive spectral indices which may place the remaining sources below the 3 @xmath36 level .
the spectral indices ( @xmath1 ) were obtained from the integrated flux densities at both frequencies , where we adopt the convention @xmath2 .
the flux densities have been computed using the aips task jmfit , when the sources were either unresolved or shell - like , and using the task blsum for sources with extended structures impossible to fit by a single gaussian .
the integrated flux densities were calculated up to a 3@xmath36 level to avoid confusion with the extended background emission . + in table [ computed ] we report the deduced quantities for the eighteen sources : the total emitted power at 5 ghz ( l@xmath43 ) and 1.6 ghz ( l@xmath44 ) , for a distance of 14.5 mpc , based on the peak flux values of table [ observed ] , and the brightness temperature calculated from the integrated flux intensities at 5 ghz ( table [ observed ] ) .
the diameter ( d ) was measured as described in appendix 1 of burns et al .
( @xcite ) , where @xmath45 , + @xmath46 , @xmath47 the radius taken from the map , and @xmath48 the half width at half maximum of the gaussian beam .
the diameter for the unresolved sources is 8.6 pc ( table [ computed ] ) and their brightness temperatures are obviously lower limits because their true dimensions are not known .
l l l l l l l l + & r.a . & dec .
& peak flux & integrated flux & peak flux & integrated flux & spectral & + & ( j2000 ) & ( j2000 ) & s@xmath41 & s@xmath41 & s@xmath42 & s@xmath42 & index & + source name & 06@xmath4918@xmath50 & 7821 & ( mjy / beam ) & ( mjy ) & ( mjy / beam ) & ( mjy ) & @xmath1 & notes@xmath51 + + + 35.0 + 28.0 & 3496 & 280 & 0.27 ( 0.04 ) & 0.27 ( 0.06 ) & & & & u + 35.3 + 30.9 & 3532 & 309 & 0.32 ( 0.03 ) & 0.75 ( 0.11 ) & & & & r + 36.6 + 27.5 & 3662 & 275 & 0.56 ( 0.04 ) & 0.61 ( 0.06 ) & 0.15 ( 0.04 ) & 0.15 ( 0.07 ) & + 1.3 & u + 36.8 + 24.0 & 3678 & 240 & 0.50 ( 0.03 ) & 2.39 ( 0.20 ) & & & & r + 37.6 + 24.2 & 3759 & 242 & 1.56 ( 0.04 ) & 1.56 ( 0.06 ) & 3.90 ( 0.04 ) & 5.30 ( 0.08 ) & -1.1 & u + 37.7 + 23.7 & 3769 & 237 & 0.31 ( 0.03 ) & 0.45 ( 0.08 ) & 0.26 ( 0.04 ) & 0.26 ( 0.07 ) & + 0.5 & u + 37.9 + 23.7 & 3791 & 237 & 0.35 ( 0.06 ) & 1.42 ( 0.09 ) & 0.30 ( 0.04 ) & 0.30 ( 0.07 ) & + 1.3 & r + 38.1 + 26.1 & 3812 & 261 & 0.15 ( 0.04 ) & 0.15 ( 0.06 ) & & & & u + 38.2 + 24.3 & 3821 & 243 & 0.22 ( 0.03 ) & 0.29 ( 0.07 ) & & & & u + 38.4 + 23.8 & 3844 & 238 & 0.20 ( 0.03 ) & 1.36 ( 0.25 ) & & & & r + 38.6 + 24.0 & 3859 & 240 & 0.28 ( 0.03 ) & 0.61 ( 0.11 ) & & & & r + 38.9 + 22.5 & 3890 & 225 & 0.30 ( 0.03 ) & 1.15 ( 0.20 ) & 0.29 ( 0.03 ) & 4.10 ( 0.46 ) & -1.1 & r@xmath17 + 39.0 + 17.9 & 3905 & 179 & 0.28 ( 0.04 ) & 0.28 ( 0.06 ) & 0.16 ( 0.04 ) & 0.16 ( 0.07 ) & + 0.5 & u + 39.2 + 20.7 & 3916 & 207 & 0.26 ( 0.04 ) & 0.26 ( 0.06 ) & & & & u + 39.5 + 17.7 & 3948 & 177 & 0.28 ( 0.03 ) & 0.71 ( 0.11 ) & 0.27 ( 0.04 ) & 0.36 ( 0.08 ) & + 0.6 & r + 40.2 + 18.1 & 4016 & 181 & 0.23 ( 0.03 ) & 0.37 ( 0.08 ) & 0.18 ( 0.04 ) & 0.18 ( 0.07 ) & + 0.6 & u + 41.2 + 17.3 & 4116 & 173 & 0.17 ( 0.04 ) & 0.17 ( 0.06 ) & & & & u + 41.4 + 15.0 & 4136 & 150 & 0.25 ( 0.04 ) & 0.25 ( 0.06 ) & 0.50 ( 0.04 ) & 0.50 ( 0.07 ) & -0.6 & u + + + + @xmath51 r : resolved only at 5 ghz ; r@xmath17 : resolved at both frequencies ; u : unresolved c c c r c + & & l@xmath43 & brightness & l@xmath44 + & d & from peak flux & temperature@xmath52 & from peak flux + source name & ( pc ) & @xmath53 & ( k ) & @xmath54 + + + 35.0 + 28.0 & 8.6 & 6.78 & @xmath551260 & + 35.3 + 30.9 & 9.4 & 8.04 & 3010 & + 36.6 + 27.5 & 8.6 & 14.07 & @xmath552840 & 3.77 + 36.8 + 24.0 & 18.4 & 12.56 & 2455 & + 37.6 + 24.2 & 8.6 & 39.20 & @xmath557265 & 98.00 + 37.7 + 23.7 & 8.6 & 7.79 & 2095 & 6.53 + 37.9 + 23.7 & 24.8 & 8.80 & 820 & 7.54 + 38.1 + 26.1 & 8.6 & 3.77 & @xmath55 700 & + 38.2 + 24.3 & 8.6 & 5.53 & @xmath551350 & + 38.4 + 23.8 & 22.6 & 5.03 & 945 & + 38.6 + 24.0 & 9.1 & 7.04 & 2610 & + 38.9 + 22.5 & 28.5 & 7.54 & 500 & 7.29 + 39.0 + 17.9 & 8.6 & 7.04 & @xmath551305 & 4.02 + 39.2 + 20.7 & 8.6 & 6.53 & @xmath551210 & + 39.5 + 17.7 & 9.4 & 7.04 & 2850 & 6.78 + 40.2 + 18.1 & 8.6 & 5.78 & @xmath551725 & 4.52 + 41.2 + 17.3 & 8.6 & 4.27 & @xmath55 790 & + 41.4 + 15.0 & 8.6 & 6.28 & @xmath551165 & 12.56 + + + + @xmath52 rayleigh - jeans 37.6 + 24.2 has a large negative spectral index of @xmath1 = 1.1 , consistent with optically thin synchrotron emission .
the source has a position offset of @xmath56 north - east of the dynamic center of ngc 2146 ( r.a .
= 06@xmath4918@xmath50374 @xmath57 04 , dec .
= 7821232 @xmath57 2 ) deduced from the @xmath58 rotation curve measured with the plateau de bure interferometer , adopting a systemic velocity of 850 @xmath59 .
the source is not resolved in the 015 resolution 5-ghz map ( c ) , and its brightness temperature is too high to be due to free - free emission from ionized gas .
the strength of the source allowed us to map it with a resolution of 004 using only the merlin observations at 5 ghz ( b ) . even in this case the source is not resolved , indicating that it is a very strong and compact object ( less than 3 pc diameter ) .
this source was considered earlier to be the true nucleus of ngc 2146 ( kb ) .
however , the properties of the source are consistent with those found for rsn and snrs , and because of its steep spectrum and small size , we believe that this source is a rsn ( weiler et al .
@xcite ) .
38.9 + 22.5 has also a large negative spectral index @xmath1 = -1.1 .
the object is resolved both at 5 ghz and 1.6 ghz , and has the structure of a partially filled shell of @xmath35 30 pc diameter ( fig .
[ 2shells ] ) . as a consequence of its morphology ,
the size obtained from a gaussian fit corresponds to its largest angular extent , which may explain the relatively low brightness temperature at 5 ghz .
similar to the previous source we believe that also in this case we are dealing with either a snr or a rsn . while the steep spectrum indicates it to be a rsn , its large size and the shell - like morphology agree better with that of a snr .
this source has a negative spectral index of @xmath1 = - 0.6 , again consistent with synchrotron emission .
it is not resolved and the derived brightness temperature of 1200 k is a lower limit .
the spectral index suggests that this source is a snr .
there are 6 sources which have a positive spectral index ( @xmath60 ) .
four of them ( 36.6 + 27.5 , 39.0 + 17.9 , 40.2 + 18.1 and 37.7 + 23.7 ) are not resolved , so that their brightness temperatures are only lower limits . the other two sources ( 37.9 + 23.7 and 39.5 + 17.7 ) are resolved into amorphous structures . the brightness temperatures , though relatively high , are consistent with free - free emission .
in the following sects . ( 4.1 and 4.2 ) we try to give an explanation of the physical nature of the thermal and non - thermal sources detected in our observations . a positive spectral index with a slope higher than 0.5
is expected to arise from free - free emission either in the turnover region , where @xmath61 , or in the optically thick region , where @xmath62 .
the free - free optical depth in the radio domain is @xmath63 ( mezger & henderson @xcite , carlstrom & kronberg @xcite , wills et al .
@xcite ) , where @xmath64 is the observing frequency , @xmath65 the electron temperature , and em the emission measure , defined as @xmath66 where @xmath67 is the electron density and l the pathlength . using in eq.(1 ) the electron temperature t@xmath68 = 8000 k derived from radio recombination line observations of ngc 2146 ( zhao et al .
@xcite ) , we obtain for @xmath69 = 1 the emission measure @xmath70 @xmath71 at 1.6 ghz and @xmath72 @xmath71 at 5 ghz
. for sources with a very steep positive spectral index these values is only a lower limit since for these sources @xmath73 . if the thermal gas is optically thin ( @xmath74 ) and has an electron temperature of 8000 k , the applied source detection limit at 5 ghz of 0.1 mjy corresponds for an unresolved source to an em of @xmath75 . using the value em(@xmath76=1 ) , and for the pathlength l the diameter d of the sources ( table [ computed ] ) , we derive the electron densities given in table [ positive ] .
+ if we assume that the free free emission is thermal + bremsstrahlung from regions photoionized by hot , luminous stars , it is possible to estimate the lyman continuum photon rate @xmath77}$ ] from the measured radio flux , and from @xmath78 it is possible to estimate the number of exciting stars . for optically thin emission , hjellming & newell ( @xcite )
give @xmath79 where d is the distance of the galaxy ; for optically thick emission , turner et al .
( @xcite ) give @xmath80 where @xmath81 is the coefficient for recombinations to all but the ground state of hydrogen ( = @xmath82 ; tabulated for different electron temperatures and densities in hummer & storey @xcite ) .
having thus computed the lyman continuum photon rate @xmath78 , we use the compilation of stellar properties by panagia ( @xcite ) to derive the number of stars embedded in the individual objects observed by us .
the number of equivalent o6 stars by @xmath57 50 % , respectively .
] is given in table [ positive ] .
+ are there objects in ngc 2146 which have emission measures of the order of em @xmath83 @xmath71 , and if so , can there be stellar clusters of 1000 to 2000 o6 stars associated with them ?
+ classical galactic regions have ems smaller than @xmath84 @xmath71 . applying the definition of the emission measure , carlstrom & kronberg ( @xcite ) derived a hypothetical upper limit of @xmath85 @xmath71 for the ems of the ionized gas in m82 , assuming that the gas is uniformly distributed throughout a 500 pc disk , with an average density less than 300 @xmath86 ( duffy et al .
@xcite ) . from radio and millimeter observations ( seaquist et al
. @xcite ; carlstrom & kronberg @xcite ; wills et al .
@xcite ) , values of up to @xmath87 @xmath71 have been derived for m82 .
+ the ems required to explain the observations of ngc 2146 are higher and suggest that the sources we are dealing with must have outstanding properties .
large ems of up to @xmath88 @xmath71 have been found in our galaxy in compact and ultracompact regions ( see churchwell @xcite ) with typical sizes of 0.01 pc .
high ems ( @xmath89-@xmath88 @xmath71 ) have been recently also found in two galaxies , the starburst dwarf galaxy ngc5253 ( turner et al .
@xcite ) and the blue compact galaxy henize 2 - 10 ( kobulnicky & johnson @xcite ) , in less compact though ultra - dense regions with size between 0.1 and 8 pc .
the parameters of our unresolved thermal sources ( turnover frequency , size , em ) , together with their estimated equivalent number of o6 stars , are consistent with these sources .
as far as the resolved thermal sources ( 37.9 + 23.7 ; 39.5 + 17.7 ) are concerned , the equivalent number of o6 stars seems clearly unrealistic , especially for 37.9 + 23.7 , which is surely seen in the optically thick regime . the star cluster associated with 37.9 + 23.7
is not seen at optical wavelengths , nor is any other super star cluster .
+ we thus propose that the objects with positive spectral index are in reality a mixture of thermal and non - thermal sources , too compact to be fully resolved in our map . however , at the single frequency of 5 ghz it is difficult to separate between thermal and non - thermal emission
. furthermore free - free absorption by ionized gas could play a role in obscuring non - thermal sources , such as snrs and/or rsn , embedded in or located behind optically thick regions ( with different sizes and densities ) .
of course such an explanation presents difficulties in being physically realistic .
when we assume that the turnover frequency is equal to or larger than 5 ghz , the em of the thermal sources should be larger than @xmath90 @xmath71 , implying regions that surround the non - thermal objects , with very small sizes ( 0.01 - 0.1 pc ) , but containing hundreds of massive stars .
if the thermal sources are instead located in front of , and are not associated with the snrs or rsn , then the number of stars should be even larger .
however , the statistical probability of a line - of - sight coincidence of snr / rsn and non - associated compact regions should be very small for the six observed sources .
+ fortunately , the discovery of very dense thermal sources by kobulnicky & johnson ( @xcite ) , with sizes up to 8 pc and containing @xmath35 750 o7v equivalent stars , makes the possibility less extreme of actually dealing with thermal sources of parsec scales with a very high stellar density . in addition to this , we can not exclude that the free - free absorption is affecting only the 1.6 ghz observations . in this case
the required em would be @xmath91 @xmath71 , and as a consequence the sizes of the regions could be larger , and the stellar density smaller .
better constraints on the exact turnover frequency , the mixture of thermal and non - thermal emission , the lyman continuum rate from the free - free emission and the equivalent number of o6 stars present both , in the resolved and unresoved sources , can be obtained from high - resolution observations at still higher frequencies ( for instance 8 and 15 ghz ) . at these frequencies
the free - free radio emission will originate either from optically thin regions ( in this case the spectral index will change to a slope of @xmath35 -0.1 ) , or from non - thermal sources no longer affected by free - free absorption ( in this case the slope will be typical for synchrotron emission ) . + the presence of a number of compact and ultra - compact regions in ngc 2146 has also been claimed by puxley et al .
( @xcite ) and zhao et al .
( @xcite ) using observations of h53@xmath92 ( 43 ghz ) and h92@xmath92 ( 8 ghz ) recombination lines . to explain both ,
the strength of the 8 ghz continuum and the recombination lines , they require a two component model of regions with high electron densities ( @xmath93 @xmath94 ) and sizes of 0.2 pc , and lower electron density regions ( @xmath95 @xmath94 ) with parsec - scale sizes . + the reason why we do not see such ultra compact / ultra dense sources in the starburst galaxy m82 , which in many respects is similar to ngc 2146 , is unknown .
a sequence for the starburst development has been proposed for some starburst galaxies , including m82 , by rieke et al .
( @xcite ) and this picture may also apply to ngc 2146 . according to this sequence , ngc 5253 for instance , which has a flat radio spectrum and similar compact sources as ngc 2146 ,
is classified to be in a very young phase of a starburst , while m82 is showing an older phase in which the massive stars have already evolved off the main sequence , with many snrs clearly visible .
ngc 2146 is probably in between these two phases because it still has optically thick regions like those in ngc 5253 , and a smaller number of snrs than m82 , but nevertheless , more similar to m82 , exhibits a global radio spectral index with a non - thermal slope of -0.74 .
c + source & electron density & equivalent number & equivalent number + name & @xmath86 & of o6 stars@xmath96 & of o6 stars@xmath97 + + + 36.6 + 27.5 & @xmath5 2800 & 2060 & 1310 + 37.7 + 23.7 & @xmath5 2800 & 2060 & 960 + 37.9 + 23.7 & @xmath5 1600 & 16750 & 3040 + 39.0 + 17.9 & @xmath98 2800 & 2060 & 600 + 39.5 + 17.7 & @xmath98 2700 & 2410 & 1520 + 40.2 + 18.1 & @xmath98 2800 & 2060 & 790 + + + + @xmath52 for optically thick emission ( eq . [ eq : thicklyman ] ) + @xmath99 for optically thin emission ( eq . [ eq : thinlyman ] ) earlier radio frequency observations of ngc 2146 ( mccutcheon @xcite ; kb ) show that the galaxy is filled with non - thermal synchrotron emission of spectral index @xmath100 .
39 mjy obtained from new single - dish 5 ghz observations of ngc 2146 , obtained with the 100 m effelsberg radio telescope of the mpifr , and the total flux at 20 cm of 1228 mjy taken from the green bank 1.4 ghz northern sky survey ( white & becker @xcite ) .
] this situation resembles that of m82 , where the spectral index is @xmath101 ( white & becker @xcite ; khr et al .
@xcite ) . following current theories , this synchrotron emission is due to sn , rsn , and snrs which preferentially occur in the starburst region , as supported by the observed x - ray emission and outflows of hot gas along the minor axis .
using minimum - energy arguments ( miley @xcite ) , muxlow et al .
( @xcite ) have computed the number of snrs ( @xmath35 2000 ) necessary to explain the observed flux density at 5 ghz of the extended radio component in the central part of m82 and consequently the time ( @xmath102 yr ) necessary to produce such a flux density , assuming a supernova rate of 0.05 @xmath103 . following this computation
, we use the relation given by longair ( @xcite ) to calculate the minimum energy requirements for synchrotron radiation : @xmath104}\label{eq : minen}\end{aligned}\ ] ] where @xmath105 is the energy ratio of relativistic protons to electrons , taken to be 1 , v is the volume of the considered region , @xmath64 is the frequency and @xmath106 the luminosity .
the 5-ghz flux of the central region of ngc 2146 ( assumed to be a disk with a radius of 750 pc and thickness of 100 pc ) is 0.25 jy , which corresponds to a minimum energy of @xmath107 . to produce this energy ,
@xmath108 evolved snrs similar to those in the lmc ( @xmath109 each ) are required . a supernova rate of 0.15 @xmath103 ( see sect .
4.2.2 ) could produce the extended non - thermal radio emission over a time of @xmath110 yr . the individual events are probably too old and faint still to be seen individually , but their number and/or strength must be higher than in m82 , making the small number of snrs detected surprising .
the strength of the overall magnetic field is not taken into account , though it must play an important role in increasing / decreasing the synchrotron emissivity and the global spectral index of a galaxy .
+ we also computed the minimum energies and the magnetic field strengths of the three snr candidates presented in the previous section .
the equation for the magnetic field ( longair @xcite ) is @xmath111}\label{eq : minb}\ ] ] with the same units as in eq .
[ eq : minen ] .
the results are compiled in table [ minb ] .
compared with the values obtained by duric et al .
( @xcite ) for non - thermal compact sources in ngc 4736 , the agreement of minimum energy requirements and the higher magnetic field strength again favour an identification with fairly young snrs .
the total emitted power at 5 ghz ( table [ computed ] ) is typically more than ten times higher than those found by duric et al .
( @xcite ) and than that of cas a. instead it falls in the upper end of the range of the m82 sources identified as young snrs or rsn .
+ c c c + source & @xmath112 & @xmath113 + name & ( @xmath114 ) & ( @xmath115 ) + + + 37.6 + 24.2 & 1.5 & 410 + 38.9 + 22.5 & 6.0 & 135 + 41.4 + 15.0 & 0.4 & 215 + + the star formation rate ( sfr ) of ngc 2146 , calculated using the extinction - corrected @xmath116 flux , is @xmath117and is three times higher than that of m82 ( young et al .
the same result is obtained from the fir flux ( moshir et al .
@xcite ) , using the relation derived by thronson & telesco ( @xcite ) and thronson et al .
( @xcite ) .
the values for the sfrs in ngc 2146 and m82 then are 15 @xmath118 and 5 @xmath118 respectively .
van buren & greenhouse ( @xcite ) use the fir flux to calculate the supernova rate in m82 to be 0.06 @xmath103 . applying this method to ngc 2146 we obtain a supernova rate of 0.15 @xmath103 , which is again higher than the one in m82 . from these results , and despite the fact that the volume of the starburst region in ngc 2146 is larger than that in m82 , we can infer that the bursts in both galaxies have different strengths . as a consequence
, we should have a higher number of snrs and rsn in ngc 2146 than in m82 .
+ however , when we place the @xmath119 snrs and rsn detected in m82 at 5 ghz ( kronberg et al . @xcite ; muxlow et al .
@xcite ) at the distance of ngc 2146 , given the sensitivity ( @xmath120 mjy at @xmath121 level ) of our merlin+vla map ( c ) we expect to find at least 8 sources .
the number of sources detected in our observations is only one third of that expected . for 9 sources , however , we do not have spectral indices , and their continuum spectra may be distorted by free - free or synchrotron self - absorption as explained above
. a larger number of snrs than those detected in ngc 2146 would also be expected if , as argued by armus et al .
( @xcite ) , snrs are the possible cause of the hard x - ray component in ngc 2146 we may therefore speculate that the starburst in ngc 2146 is young , or repetitive , so that we are presently observing a strong and young burst still lacking a large number of snrs and rsn , but instead emitting strong optically thick thermal emission from ionized gas regions , possibly associated with super starclusters .
a first , also strong , burst of star formation may have occurred in the past , of which we now observe the intense extended radio component , and which could account for the overall steep non - thermal radio spectrum of the galaxy . +
a realistic hypothesis is also that some snrs and rsn are hidden because of strong free - free absorption , viz . in the mixed regions mentioned in the previous subsection .
free - free absorption has in fact been also invoked for several compact sources in m82 ( wills et al .
@xcite ) , even though with turnovers at much lower frequencies , with the only exception of 44.01 + 59.6 that has a positive spectral index at frequencies higher than 1 ghz .
the combination of vla and merlin 5 ghz observations of the starburst galaxy ngc 2146 allowed us to resolve seven of the eighteen hitherto unresolved sources detected with the vla .
we have measured the spectral indices between 1.6 and 5 ghz for nine sources .
three sources ( 37.6 + 24.2 , 38.9 + 22.5 , 41.4 + 15.0 ) seem to be good candidates for snrs or rsn , with total emitted powers at 5 ghz comparable to the strongest snrs found in m82 .
four other sources ( 36.6 + 27.5 , 39.0 + 17.9 , 40.2 + 18.1 and 37.7 + 23.7 ) have instead been identified with compact or ultra - dense regions that have large ems and a high - frequency turnover .
the emitted thermal emission indicates a large equivalent number of o6 stars , exceeding the number of those found in similar objects in other galaxies .
the large number of o6 stars may , however , be consistent with stars in super starclusters .
the nature of the last two sources ( 37.9 + 23.7 and 39.5 + 17.7 ) for which we have spectral index information are still a matter of discussion .
it seems realistic that they emit a mixture of thermal and non - thermal radiation , and may manifest the most powerful regions ever observed .
however , the burst of star formation in ngc 2146 has an extraordinary power , stronger than that in m82 . despite this and the fact that the overall radio spectral index of the galaxy
reveals a non - thermal origin , the number of discovered snrs is much smaller than that in m82 . for the explanation of this fact we propose two scenarios : * ngc 2146 is experiencing a burst of star formation stronger than that in m82 , but which we are observing in a younger phase
this phase still lacks a large number of snrs , but instead gives rise to strong thermal emission from optically thick dense ionized gas regions , associated with massive or super starclusters .
the origin of the non - thermal extended component must be a trace of past sn explosions , individually too old and faint to be seen , and belonging to a previous burst in the galaxy . * strong free - free absorption by very dense regions and , for the most compact sources , possibly also + synchrotron - self absorption , may conceal a certain number of snrs or rsn , rendering the identification of their nature difficult if only based on measurements of the spectral indices around the turnover frequencies . in this case
the global number of snrs or rsn can be higher , and this would corroborate the similarity between ngc 2146 and m82 .
this interpretation raises , however , the question why we do not observe similar compact or ultra - dense regions in m82 as well .
we would like to acknoledge dr .
peter thomasson for his helpful support and dr .
peter biermann for useful discussions .
we are grateful to the referee dr .
jean turner for constructive comments and valuable suggestions .
this work was supported by the dfg through the graduiertenkolleg the magellanic system , galaxy interaction and the evolution of dwarf galaxies. 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 present a high - resolution 5-ghz radio continuum map of the starburst galaxy ngc 2146 made with merlin and the vla ( a - array ) , in a search of radio supernovae and supernova remnants expected to be already produced by the most massive stars in the starburst . at 5 ghz ,
about 20 point sources were detected earlier by glendenning @xmath0 kronberg ( 1986 ) in the central 800 pc of ngc 2146 .
our observations with higher sensitivity and resolution made with merlin and the vla confirms the detection of 18 sources , and resolves 7 of them .
additional 1.6-ghz merlin observations disclose 9 sources coincident in position with those detected at 5 ghz , which allows us to derive their spectral indices @xmath1 , ( @xmath2 ) .
only 3 sources have indices ( @xmath3 @xmath4 0 ) consistent with synchrotron emission from supernova remnants or radio supernovae , while the others have very steep inverted spectra ( @xmath3 @xmath5 0 ) .
+ we suggest that the sources with positive spectral index are optically thick ultra
compact and/or ultra dense hii regions with high electron densities and high emission measures ( em @xmath5 @xmath6 @xmath7 pc ) .
minimum energy requirements indicate that these regions may contain up to 1000 equivalent stars of type o6 , comparable to the number of stars found in super starclusters .
when compared with m82 , the galaxy ngc 2146 lacks however a large number of supernova remnants .
we suggest that ngc 2146 is experiencing a burst of star formation stronger than that in m82 , but being in a younger phase . in this phase , only few stars have already exploded , whereas the others cause strong thermal emission from compact , optically thick ionized gas regions , around the young super starclusters .
+ we may , however , not exclude an alternative scenario in which strong free - free absorption at 1.6 ghz in foreground ionized gas with very high emission measures ( em @xmath5 @xmath8 @xmath7 pc ) hides a certain number of supernova remnants , thus rendering for some sources the observed inverted spectra .
|
the unidentified 21 @xmath0 m feature was first discovered in carbon - rich proto - planetary nebulae ( ppn , * ? ? ?
* ) and 16 21-@xmath0 m sources have been found to date @xcite .
although 20 years have past and a variety of chemical materials have been suggested as a possible carrier , no firm identification has been made .
previously proposed candidates include hydrogenated amorphous carbon @xcite , hydrogenated fullerences @xcite , nanodiamonds @xcite , tic nanoclusters @xcite and o - substituted five - member carbon rings @xcite , sic @xcite , etc . since the observation of the 21-@xmath0 m feature
is confined to the ppn phase of stellar evolution , one may ask whether the formation or excitation of the feature is related to the dynamical evolution of post - agb evolution . for example , one may conclude that the physical conditions ( temperature and density ) governing the chemical reactions may be related to the morphology and kinematics of the envelope .
one of the ways to reveal the morpho - kinematic properties of ppne is through spectral imaging of molecular lines ( especially in the co lines ) of the circumstellar envelope .
iras 07134 + 1005 ( @xmath4hd 56126 ) , a prototypical 21 @xmath0 m source , has an angular size of @xmath57@xmath38@xmath3 @xcite and therefore is well fitted for millimeter - wave interferometric observations .
optical _ hubble space telescope _
images of iras 07134 + 1005 reveal a bright central star surrounded by a low surface brightness elliptical nebula @xcite .
a near - ir imaging polarimetry of iras 07134 + 1005 reveals a thin ( both geometrically and optically ) , limb - brightened shell with an equatorial density enhancement @xcite . in comparison ,
the mid - ir images @xcite show a hollow structure with a bright rim , with the emission region of the 21 @xmath0 m feature coinciding with the rim structure .
the central star is variable with a period of 36.8 days , indicating a mass of 0.6 m@xmath6 @xcite .
an abundance analysis by @xcite reveals a metal - poor star with less than solar fe abundance but with solar or greater than solar abundances of c , n , o and @xmath7-process elements , indicating the star experienced third dredge - up when it was on the asymptotic giant branch ( agb ) . the fact that it displays the family of aromatic infrared bands also points to its carbon - rich nature . in this paper , we report the results of a radio interferometric observation of iras 07134 + 1005 in the @xmath8co @xmath12 line using the submillimeter array ( sma ) .
we also present the results of model fitting using _
shape_. the outline of this paper is as follows . in section 2
we describe the details of the observation and data reduction . in section 3
we present the observational results . in section 4
we present the procedure and results of the morpho - kinematic modeling with _ shape_. in section 5 we discuss the obtained results , and finally the present research is summarized in section 6 .
the observations of iras 07134 + 1005 were made on february 28 , 2005 with sma under good atmospheric conditions with a zenith optical depth at 230 ghz of 0.030.05 .
the observations covered the frequency ranges of 344.47270346.45638 ghz ( usb ) and 334.47221336.45589 ghz ( lsb ) , which include the 345.796 ghz co @xmath12 line in the usb .
the array used 6 elements in the compact array configuration , with baselines ranging from 10 to 70 m. the diameter of each antenna was 6 m , and the field of view of a single antenna was 34@xmath3 .
the observation was interleaved every 20 minutes with nearby gain calibrators 0739 + 016 and 0750 + 125 to track the phase variations over time .
we calibrated the date using idl
mir , which is a data reduction software package developed by the sma project .
the absolute flux calibration was determined from observations of ganymede , and was accurate to within 15% .
the final map has an accumulated on - source observing time of about 3 hours .
the single - sideband system temperature ranged from 300 to 550 k , depending on the atmospheric conditions and the telescope elevation .
the sma correlator had a bandwidth of 2 ghz with a frequency resolution of 0.812 mhz corresponding to 0.7 km s@xmath2 in the velocity resolution at 345 ghz .
the phase center of the map was taken at r.a.@xmath9 , decl.@xmath10 ( j2000 ) .
image processesing of the data was performed with the miriad software package @xcite .
the robust weighting gave a 2.28@[email protected]@xmath3 clean beam with a position angle of @xmath12 .
we checked the detection of the continuum emission by integrating over a 2 ghz range in lsb , but did not detect it .
the upper limit of the continuum emission is @xmath13 jy beam@xmath2 .
in figure 1 we present the total intensity profile of the @xmath8co @xmath12 line .
the profile exhibits a parabolic shape usually expected from a simple spherical outflow ( e.g. , * ? ? ?
* ) , and is about the same as those of the other co lines ( see spectra in , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the peak intensity is 38.1 jy at @xmath14 km s@xmath2 .
the velocity integrated intensity is 492.8 jy km s@xmath2 .
the line - width at the zero intensity level is 20 km s@xmath2 ( corresponding to an expanding velocity of 10 km s@xmath2 ) .
since there is no single - dish co @xmath12 spectrum of iras 07134 + 1005 in the literature , we can not evaluate the extent of flux recovery by comparing with zero - spacing data .
however , at least 70% of the flux seems to be recovered in the present interferometry . since we know the size of the molecular envelope ( @xmath57@xmath38@xmath3 ) measured by the bima observation in the co @xmath150 line ( *
* the bima observation recovered 100% of the flux ) , here we assume a gaussian brightness distribution with a fwhm of 8@xmath3 . under this assumption 70% of the flux
is recovered with our _ uv_-coverage of the present observation .
since the emission region seems to include small - scale structure , which is not resolved out by interferometer , the real recovery rate of the flux would be more than 70% .
figure 2 shows the total intensity map of the co @xmath12 line superimposed on the 12.5 @xmath0 m image taken by the oscir instrument mounted on the gemini north telescope @xcite .
the map center is taken at the position of the central star ( the central star can be seen , for example , in the 10.3 @xmath0 m and 11.7 @xmath0 m images ; see figure 1 in * ? ? ?
the co structure is clearly larger than the mid - infrared structure , and we can see a resolved co structure coincide well with the 12.5 @xmath0 m structure at the central region of the map ( within @xmath51.5@xmath3 from the map center ) .
this central structure has already been hinted to as being hollow in the co @xmath150 bima map @xcite , but the intensity map of figuer 2 seems to coincide better with the mid - infrared structure than with the bima map .
the central co structure appears to be surrounded by a spherical envelope with an angular size of about 8@xmath3 .
the size of the spherical envelope found in the present co @xmath12 map is almost the same as that found in the co @xmath150 bima map .
if we assume the distance to the star is 2.4 kpc @xcite , 8@xmath3 is translated to a linear size of @xmath16 cm .
the spherical envelope seems to be slightly elongated in the north - south direction , but this elongation is due presumably to the beam pattern exhibiting elongation in this direction .
figure 3 shows velocity channel maps in the co @xmath12 line .
the feature seen in figure 3 clearly exhibits a systematic variation in velocity .
the feature increases in size as one approaches the systemic velocity ( @xmath573 km s@xmath2 ) , suggesting a spherically expanding envelope .
in addition , the central structure also exhibits a systematic variation in velocity . in the channels of 6769 km s@xmath2 , an arch - like feature ( convex upward )
is seen in the northern side of the map center , in the channels of 7073 km s@xmath2 two intensity peaks are seen at the east and west sides of the map center , and in the channels of 7478 km s@xmath2 another arch - like feature ( convex downward ) is seen at the southern side of the map center .
the @xmath17@xmath18 diagrams presented in figure 4 also shows this systematic variation of the co envelope . a reasonable interpretation of these variations could be an expanding torus surrounded by a spherically expanding envelope if the torus has a certain inclination angle .
we will further consider the morpho - kinematic properties of the envelope in the later sections by constructing models .
in order to acquire a better understanding of the morpho - kinematic properties of iras 07134 + 1005 , we have constructed a model using the _ shape _ software package ( see e.g. , * ? ? ? * ) .
_ shape _ is a morpho - kinematic modeling tool used to create 3d models of astronomical nebulae .
the modeling strategy behind shape is to build a 3d model of the object , create channel maps and @xmath17@xmath18 diagrams , compare with the observational data and then refine the model if needed .
the comparison of the generated channel maps and @xmath17@xmath18 diagrams with those observed is done by visual inspection .
_ shape _ does not calculate the numerical hydrodynamic evolution or radiation transfer equations .
since the co @xmath12 line seems to be optically thick , we are only concerned with the morphology and kinematics , and did not try to reproduce the intensity distribution and line profile .
we assumed an axial symmetric geometry in an effort to reduce the number of model parameters .
we also assumed , as outlined in section 3 , that the envelope consists of an expanding torus and expanding sphere ; both of which share the symmetry axis .
the modeled sphere has a fixed outer - radius of 4@xmath3 and a constant , radially expanding velocity of 9 km s@xmath2 . to model the torus , we adjusted 5 different parameters : inner and outer radii , thickness ( height ) , inclination angles of the symmetry axis and the expansion velocity .
the best - fit parameters were determined by ( educated ) trial and error until the reproduced maps closely matched the observed .
the parameters obtained are : inner radius of 1.2@xmath3 , outer radius of 3.0@xmath3 , thickness of 4.5@xmath3 , and a constant radial expansion velocity of 8 km s@xmath2 .
the inclination angle was determined to be 38@xmath19 to the line of sight , and rotated 8@xmath19 from the north in counterclockwise ( i.e. , projected position angle is 8@xmath19 ) . in figure 5
we present the polygon - mesh view of the model .
the volume of the sphere and torus are each sampled with 20000 particles , which creates a sufficiently smooth map for comparison with the observations . in figures 6 and 7 , we present the channel maps and @xmath17@xmath18 diagrams of the best - fit model described above . since _
shape _ can currently only use a circular beam for simulating image convolution , we used a circular beam with a diameter of 2@xmath3 instead of the actual elliptical beam of @xmath20 .
the use of a circular beam should not affect the main features of the model maps , but could affect the details .
the features seen in the model maps seem to roughly reproduce the observation , even though some noticeable differences are evident .
the major difference is in the intensity distribution which is especially clear in the @xmath17@xmath18 diagram . in the observed @xmath17@xmath18 diagram ( figure 4 ) , the intensity is clearly stronger in the red - shifted ( high velocity ) side , while the model @xmath17@xmath18 diagram exhibits a symmetric intensity distribution in velocity with respect to the systemic velocity ( @xmath573 km s@xmath2 ) .
this asymmetry in the intensity may be due to radiative - transfer effects .
the velocity of the near side ( approaching side ) of the torus is , more or less , similar to that of the approaching side of the sphere , and therefore the emission from the near side of the torus is selectively absorbed by the sphere . on the contrary ,
the emission from the far side ( receding side ) of the torus is not absorbed by the sphere , because the velocity of the far side is clearly different from that of the near side of the sphere .
apart from the asymmetry seen in the @xmath17@xmath18 diagrams , we see further differences in the fine structure of the channel maps . for example
, the model map shows a smooth ring - like feature in the 73 km s@xmath2 channel , while the observational map clearly shows two intensity peaks at the east and west sides of the map center . a possible explanation for this kind of difference in structure could be a distorted shape to the torus . to check this possibility , we constructed a model with a distorted torus and found that this distortion may indeed explain the observed map . of course
, using a distorted model increases the number of free parameters and therefore has a better chance of matching the data .
nevertheless , showing the results of the distorted model may be valuable in paving the way for improvements to the simple , symmetric torus model .
in figure 8 we present the distorted torus model and the resulting maps .
the top - right panel shows the geometry of the torus . a sphere with a radius of 4@xmath3 similar to that of the symmetric model has been omitted for clarity .
the torus is basically a sphere with a hollowed out spheroidal region .
this torus has a radially expanding velocity field with a hubble type law of @xmath21 .
the length of the major axis of the spheroid is 3.6@xmath3 , and the length of the minor axis is 1.2@xmath3 .
the center of the spheroid is shifted to the south ( along with the symmetry axis ) by 0.7@xmath3 with respect to the center of the expanding sphere . therefore the one side of the hollow region is open and the other side is closed .
the inclination angle of the symmetric axis was determined to be 55@xmath19 to the line of sight , and rotated 14@xmath19 from the north in counterclockwise ( i.e. , projected position angle ) .
this distorted torus model seems to reasonably reproduce the fine structure seen in the observed maps .
the first glimpse into the inner region of the molecular envelope of iras 07134 + 1005 was made by @xcite in their bima co @xmath150 observation , revealing a cavity structure at the central part of the envelope . in the present observation , with the higher excitation temperature and critical density of the co @xmath12 line , we have more clearly revealed the innermost structure , which is interpreted as an expanding torus .
the existence of a torus is relatively common in evolved stars and pne .
in fact , tori have been found in various kinds of objects such as agb stars and pne ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which have been considered a remnant of an equatorial enhanced mass loss developed in the last moment of the agb ( see , e.g. , * ? ? ?
therefore , the morpho - kinematic information of the torus found in iras 07134 + 1005 possibly provides us useful information for considering its mass - loss history in the late agb and early post - agb . here , we first try to estimate some physical parameters of the torus of iras 07134 + 1005 by making use of the present data , and then we discuss the meaning of the morpho - kinematic properties of iras 07134 + 1005 in the scheme of the post - agb evolution , and also discuss a possible relation between the morpho - kinematic properties of the envelope and the origin of the 21 @xmath0 m feature .
the inner radius of the torus is one of the most important parameters , because it directly constrains the time since the central star left the agb .
@xcite predicted the inner radius of the molecular envelope by fitting radiative transfer models to the spectral energy distribution ( sed ) , and gave the value of 1.2@xmath3 ( the original value given by volk et al .
is @xmath22 pc kpc@xmath2 ; here , we assume the distance of 2.4 kpc to translate the units ) .
@xcite gave a dust inner radius of [email protected]@xmath3 from their direct imaging of the dust component by mid - infrared imaging .
in the present research , the best - fit _ shape _ model ( simple , symmetric model ) gives an inner radius of 1.2@xmath3 .
this value is basically consistent with the previous values , but somewhat larger than the dust radius given by @xcite .
a reasonable interpretation for this might be that there are no molecules in the region within @xmath23 , even though this might be due alternately to excitation effects ( i.e. , co molecules exist in the region @xmath24 , but are not bright in the co @xmath12 line due to high temperature and/or high density ) .
therefore , further co observations in the higher transition lines will give us a deeper insight on this point .
if we assume that the expanding velocity of the torus is a constant value , 8 km s@xmath2 , which is determined by fitting _ shape _ model , the dynamical time scale of the inner edge of the torus ( @xmath25 ) is estimated to be 1710 yr ( also at the distance of 2.4 kpc ) . if we assume the inner radius of 0.8@xmath3 , which corresponds to the dust inner radius given by @xcite , the dynamical time scale is estimated to be 1140 yr . on the other hand
, the outer radius of the torus determined in our _ shape _ modeling is 3.0@xmath3 , giving a dynamical time scale of 4270 yr under the assumption that the expanding velocity is also 8 km s@xmath2 .
( since the outer edge of the torus seems to interact with the outer spherical envelope , the time scale of 4270 yr might include significant uncertainty ) .
a straightforward interpretation of these time scales is that the mass loss of the central star has gone into the equatorial enhanced mass - loss mode , forming a torus 4270 years ago .
the phase of mass loss then continued for about 25603130 years , and finally stopped 11401710 years ago .
this kinematic history of the envelope is consistent with the time - variation of a mass - loss rate assumed in previous radiative transfer calculations , which reasonably reproduce the co observations @xcite .
we roughly estimate the lower limit of the torus mass from the integrated intensity of the observed co @xmath12 line .
( since the co @xmath12 line is most likely optically thick , here we estimate just a lower limit of the mass ) . to calculate the lower limit of the mass
, we used the optically thin formula given by @xcite . to derive the flux of the torus , we integrated the flux over a circle region ( radius is 1.5@xmath3 ) placed at the map center ( the derived value is 265.6 jy km s@xmath2 ) .
for a distance of 2.4 kpc and a representative co / h@xmath26 abundance of @xmath27 for carbon - rich envelopes @xcite , we obtained the value @xmath28
m@xmath6 . if we assume that the duration of the equatorial outflow is 2800 years as derived above , the lower limit of the mass - loss rate in this phase is calculated to be @xmath29 m@xmath6 year@xmath2 .
this lower limit of the mass and mass - loss rate are 2 orders of magnitude smaller than values given by radiative transfer equation calculations ( see , e.g. , * ? ? ?
this presumably means that the co @xmath12 line is significantly optically thick .
a noteworthy characteristic in the morpho - kinematic properties of iras 07134 + 1005 would be the absence of a jet ( i.e. , a high - velocity bipolar flow ) .
a jet consisting of molecular or ionized gas is often found in evolved stars and pne together with a torus @xcite , but it is not found in iras 07134 + 1005 .
the jet might be resolved out by interferometry , but even in such a case the existence of a high - velocity jet should be hinted in the line profiles given by previous single - dish observations as a high - velocity wing . however , previous co observations have not found the evidence of such a high velocity wing . therefore , the jet seems to be really missing in the envelope of iras 07134 + 1005 , unless it is still not accelerated enough to form a wing component in the line profile , or fully ionized and slipped from the molecular line observations ( i.e. , too weak to be detected in molecular lines ) . otherwise , there is no jet at any time .
@xcite recently suggested , based on statistical analyses of evolved stars and pne , that a jet is formed after a torus is formed ( although the formation mechanism of jets and tori is still not clear ) , and that the time lag between the torus and jet formation ( hereafter , `` jet lag '' after the fashion of the huggins s paper ) is several hundred years ( the median is 300 years but with a large scatter ) .
the jet lags of the huggins s sample ranged from 130 to 1660 years .
our kinematic analyses suggests that iras 07134 + 1005 seems to have left the agb roughly 1400 years ago .
according to the range of jet lags in @xcite , it would not be unnatural if a jet is well developed in the molecular envelope of iras 07134 + 1005 .
therefore , the absence of a jet might mean that the jet lag of iras 07134 + 1005 is relatively long , if the jet is yet to be formed .
interestingly , another prototypical example of the 21-@xmath0 m source , iras 22272 + 5435 @xcite , also shows no sign of a high - velocity jet , while this object seems to have a developed torus .
single - dish line profiles of the co rotational lines exhibit an unexceptional parabolic shape , in which no hints of a high velocity wing are found ( see , e.g. , * ? ? ?
fitting of radiative transfer models to the single dish spectra suggested that a high mass - loss rate ( @xmath30 m@xmath6 ) is required in the last 2000 years , suggesting that a developed torus , which is formed by the superwind , exists in the inner region of the molecular envelope .
although radio co imaging did not resolve the torus structure due presumably to its small angular size @xcite , optical @xcite , near - ir @xcite and mid - ir @xcite imaging found a toroidal dust shell in the innermost part of the envelope .
the resemblance of iras 22272 + 5435 to iras 07134 + 1005 tells us that all 21-@xmath31 sources might exhibit a large jet lag , and also that all the 21-@xmath31 sources are at an intermediate phase between the torus and jet formation , even though there still could be other interpretations ( i.e. , the jet is / was very weak , or there is no jet formed at any time ) .
the origin of the jet lag is not clear .
however , the long jet lag potentially constrains the physical conditions of the innermost region of the envelope , which can not be resolved by any existing equipments , if we assume a certain model .
for example , one of the possible models , which explains the formation of jets and tori ( and jet lag ) , is an accretion disk on a primary / secondary star ( see , e.g. , * ? ? ?
* ; * ? ? ?
in the accretion disk model , the jet lag is tightly related to the viscosity of the disk : the longer the jet lag , the lower the viscosity [ the mathematical relation between the jet lag and viscosity is given by @xcite ; see , equation 2 in his paper ] .
if the viscosity is related to the chemical composition of the envelope , we might be able to constrain the carriers of the 21 @xmath0 m feature .
as stated in sect 4 , we have to be careful in the interpretation of the distorted torus , because the distorted shape of the torus increases model parameters , and also because observations in optically thick lines , such as co @xmath12 , might not be the best for comparison with the detailed _ shape _ model
. however , the distorted torus model suggested in this paper seems to include an important indication .
it is that the observation might be reproduced with the asymmetry of the torus in the axial direction , which presumably means that the two jet flows have not been formed simultaneously ( in other words , not synchronized ) , while another interpretation still might be that the density distribution of the envelope is not homogeneous in the axial direction ( i.e. , dense in the north , but thin in the south ) .
whatever the case , the torus morphology provides us useful information about the physical conditions of the envelope , and/or helps to constrain the formation mechanism of jets .
this paper has reported the results of an sma observation of iras 07134 + 1005 in the co @xmath12 line .
we also performed morpho - kinematic analyses with _ shape_. the main results of this research are summarized below : 1 .
the innermost structure of the molecular envelope was clearly resolved , and it is interpreted as a geometrically thick expanding torus with an expansion velocity of 8 km s@xmath2 .
the inner radius of the torus determined by the _ shape _ model is 1.2@xmath3 , suggesting that the central star has left the agb about 1400 years ago .
the derived inner radius of the torus is consistent with the previous values determined by radiative transfer calculations and mid - infrared imaging of the dust component .
the outer radius of the torus determined by the _ shape _ model is 3.0@xmath3 , suggesting that the central stars has gone into the superwind phase from 4270 years ago . in conjunction with the time scale of the inner edge of the torus
, the duration of the superwind is estimated to be 25603130 years .
there is no jet observed in iras 07143 + 1005 .
the jets could be weak or absent , or have a long delay times , and these may be characteristic of the 21 @xmath0 m sources .
the work was supported by the research grants council of the hong kong under grants hku7020/08p and hku7033/08p , and by the financial support from seed funding programme for basic research in hku ( 200802159006 ) .
barths , d. , lbre , a. , gillet , d. , & mauron , m. 2000 , , 359 , 168 bujarrabal , v. , alcolea , j. , & neri , r. 1998 , , 504 , 915 bujarrabal , v. , alcolea , j. , & planesas , p. 1992
, , 257 , 701 buss , r. h. , et al .
1990 , , 365 , l23 chiu , p .- j . ,
hoang , c .- t .
, dinh - v - trung , lim , j. , kwok , s. , hirano , n. , & muthu , c. 2006 , , 645 , 605 dayal , a. , hoffmann , w. f. , bieging , j. h. , hora , j. l. , deutsch , l. k. , & fazio , g. g. 1998 , , 492 , 603 fong , d. , meixner , m. , sutton , e. c. , zalucha , a. , & welch , w. j. 2006 , , 652 , 1626 gledhill , t. m. , chrysostomou , a. , hough , j. h. , & yates , j. a. 2001 , , 322 , 321 grishko , v. i. , tereszchuk , k. , duley , w. w. , & bernath , p. 2001
, , 558 , l129 hill , h. g. m. , jones , a. p. , & dhendecourt , l. b. 1998 , , 336 , l41 hirano , n. , et al .
2004 , , 616 , l51 hony , s. , tielens , a. g. g. m. , waters , l. b. f. m. , & de koter , a. 2003 , , 402 , 211 hrivnak , b. j. , & bieging , j. h. 2005 , , 624 , 331 hrivnak , b.j . ,
volk , k. , geballe , t.r . , & kwok , s. 2008 , in _ iau symp . 251 : organic matter in space _ , s. kwok & s. a. sandford , eds . , cup , p. 213
huggins , p. j. 2007 , , 663 , 342 huggins , p. j. , bachiller , r. , cox , p. , & forveille , t. 1996 , , 315 , 284 huggins , p. j. , forveille , t. , bachiller , r. , & cox , p. 2000 , , 544 , 889 huggins , p. j. , muthu , c. , bachiller , r. , forveille , t. , & cox , p. 2004
, , 414 , 581 jura , m. , chen , c. , & werner , m. 2000 , , 544 , l141 knapp , g. r. , crosas , m. , young , k. , & ivezic , z. 2000 , , 534 , 324 knapp , g. r. , young , k. , lee , e. , & jorissen , a. 1998 , , 117 , 209 kwok , s. , volk , k. , & hrivnak , b. j. 2002 , , 573 , 720 kwok , s. , volk , k. , & hrivnak , b. j. 1989 , , 345 , l51 meixner , m. , skinner , c. j. , graham , j. r. , keto , e. , jernigan , j. g. , & arens , j. f. 1997 , , 482 , 897 meixner , m. , zalucha , a. , ueta , t. , fong , d. , & justtanont , k. 2004 , , 614 , 371 morris , m. 1987 , , 99 , 1115 morris , m. , & alcock , c. 1977 , , 218 , 687 nakashima , j. 2005 , , 620 , 943 nakashima , j. , et al .
2007 , , 134 , 2035 papoular , r. 2000 , , 362 , l9 sault , r. j. , teuben , p. j. , & wright , m. c. h. 1995 , astronomical data analysis software and systems iv , asp conference series , eds , shaw , r. a. , payne , h. e. and hayes , j. j. e. , 77 , 433 speck , a.k . ,
& hofmeister , a.m. 2004 , , 600 , 986 soker , n. , & rappaport , s. 2000 , , 538 , 241 steffen , w. , & lpez , j. a. 2006 , revmexaa , 42 , 99 ueta , t. , meixner , m. , & bobrowsky , m. 2000 , , 528 , 861 ueta , t. , et al .
2001 , , 557 , 831 ueta , t. , murakawa , k. , & meixner , m. 2005 , , 129 , 1625 van winckel , h. , & reyniers , m. 2000 , , 354 , 135 volk , k. , kwok , s. , & hrivnak , b. j. 1999 , , 516 , l99 von helden , g. , tielens , a. g. g. m. , van heijnsbergen , d. , duncan , m. a. , hony , s. , waters , l. b. f. m. , & meijer , g. 2000 , science , 288 , 313 webster , a. 1995 , , 277 , 1555 zuckerman , b. , dyck , h. m. , & claussen , m. j. 1986 , , 304 , 401
|
we report the results of a submillimeter array ( sma ) interferometric observation of 21-@xmath0 m source iras 07134 + 1005 in the co @xmath12 line . in order to determine the morpho - kinematic properties of the molecular envelope of the object
, we constructed a model using the _ shape _ software to model the observed co map .
we find that the molecular gas component of the envelopes can be interpreted as a geometrically thick expanding torus with an expanding velocity of 8 km s@xmath2 .
the inner and outer radii of the torus determined by fitting _ shape _
models are 1.2@xmath3 and 3.0@xmath3 , respectively .
the inner radius is consistent with the previous values determined by radiative transfer modeling of the spectral energy distribution and mid - infrared imaging of the dust component .
the radii and expansion velocity of the torus suggest that the central star has left the asymptotic giant branch about 11401710 years ago , and that the duration of the equatorial enhanced mass loss is about 25603130 years . from the absence of an observed jet
, we suggest that the formation of a bipolar outflow may lack behind in time from the creation of the equatorial torus .
|
even though general relativity is a very well - tested theory , there is a strong motivation to modify general relativity at large scale . this is because to explain the accelerating cosmic expansion in general relativity , a cosmological constant has to be included @xcite .
even though the existence of such a cosmological constant is predicted from quantum field theories , quantum field theories predict a cosmological constant which is @xmath0 times larger than the observed value of the cosmological constant .
this has motivated the study of modified theories of gravity , and the scalar - tensor theories are such a modification of general relativity which could explain accelerating cosmic expansion @xcite .
however , these theories in general contain higher derivative terms in the action , and such higher derivative terms gives rise to ostrogradsky ghost degree of freedom .
these in turn cause instabilities in the theory called the ostrogradsky instabilities .
it is possible to avoid ostrogradsky ghosts by using a theory with galileon symmetry @xcite .
even though this galileon theory also contains higher derivative terms in the action , the galilean symmetry ensures that the field equations are only second - order differential equations .
thus , the galilean symmetry ensures that the ostrogradski ghost instabilities are avoided in these theories . in the decoupling limit
, this theory contains a higher - order derivative interaction known as the cubic galileon .
it is possible to construct a theory with quartic and quintic galileon @xcite .
it is necessary to add couplings of the scalar to curvature tensors , away from the decoupling limit . this way the horndeski theory
is obtained , and field equations for both the scalar field and the metric are again second order differential equations @xcite .
so , horndeski theory is also free from ostrogradsky ghosts , and it does not contain the ostrogradsky instabilities associated with these ghosts . this horndeski formalism is a general formalism and it can be used to analyse a non - minimal coupling of other fields to the metric .
thus , it is possible to analyse a horndeski coupling of the metric to a vector field @xcite . in this formalism , the vector field is again coupled to the metric using a non - minimal coupling .
furthermore , the field equations for this vector field are again second - order field equations .
thus , the ostrogradsky ghost terms are avoided even for the horndeski coupling of a vector field to the metric . as we will be interested in applying such a vector - tensor - horndeski theory to astrophysics
, we will assume such a vector field is a fundamental field in nature , which is obtained by increasing the field content of general relativity , and it is not the usual electromagnetic field .
this is important as astrophysical objects are neutral and do not have an electric charge .
so , the electromagnetic field can not have a direct effect on the physics of such astrophysical objects .
furthermore , it is known that the gravity couples to the electromagnetic field in the usual way , and there is no reason for it to couple to the physical electromagnetic field in a non - minimal way in any astrophysical object . however ,
if we assume that there exists an astrophysical vector field , which has negligible effect on small scale , then it is possible that such an astrophysical vector field can couple to the metric in a non - minimal way .
this coupling of such an astrophysical vector field can have non - trivial effects on the astrophysical phenomena .
so , in this paper , we will analyse the effect of such a non - minimal coupling of an astrophysical vector field on the physics of a compact star .
it may be noted that such fundamental vector fields have been proposed as a solution to various different physical problems , and have also been used to explain many interesting astrophysical phenomena . just like the scalar fields
, the vector field have also been used to explain cosmic expansion @xcite , and even the naturalness problem @xcite .
it may be noted that there the vector - tensor theories produce many non - trivial phenomenological effects which can not be produced in the scalar - tensor theories @xcite .
it has been observed that anomalies exist in the alignment of the low multipoles of the cmb @xcite and the hemispherical asymmetry @xcite .
such anomalies suggest that there might be a preferential direction in the universe , and this can be explained using such a vector field .
furthermore , such fundamental vector fields have also been used to study inflation @xcite .
the occurrence of higher derivative terms can produce ostrogradski ghost instability in such vector - tensor theories .
however , such ostrogradski ghost instability can be avoided by using the horndeski formalism .
thus , it is interesting to study a vector - tensor - horndeski theory .
it may be noted that there are other interesting motivation to introduce such vector fields .
this is because there exists a discrepancy between the predicted and observed dynamics of galaxies @xcite , and it has been proposed that the discrepancy can be resolved by increasing the field content of general relativity @xcite .
the field content of this modified theory of gravity contains such a vector field .
it may be noted that this theory agrees with the predictions made by the modified newtonian dynamics @xcite .
as the field content of general relativity contains a vector field , it has been observed that astrophysical black holes in this modification of gravity can mathematically resemble a reissner
nordstrom solution @xcite .
however , the charge of this vector field is generated from mass and not an electromagnetic source @xcite . as the astrophysical objects are not charged , and so any coupling of metric to a electromagnetic field can not be applied to study such astrophysical objects .
however , as this additional vector field is produced from mass and not electric charge @xcite , it is possible that such a field can have direct effect on astrophysical phenomena .
furthermore , in this theory , there is a certain amount of freedom to choose the action for vector field , and it has been demonstrated that coupling the metric to a non - linear vector field can turn a black hole into a gray hole @xcite . as it has been demonstrated that different form of the vector field action can produce different physical results , it is interesting to investigate other forms of coupling of vector field to metric . furthermore , as this vector field occur in the field content of general relativity , it will also be interesting to investigate the astrophysical application of such a theory .
so , in this paper , we will use the horndeski formalism to couple a vector field to the metric , and we use this vector - tensor - horndeski theory to analyse a compact star .
in this section , we will analyse the main features of a vector - tensor - horndeski theory .
it has been proposed that by increasing the field content of general relativity certain astrophysical phenomena can be explained @xcite .
furthermore , for compact stars this would deform astrophysical solutions , and these deformed solutions would resemble a charged black hole solution @xcite .
it is physically important to point out that the vector field introduced here is not the usual electromagnetic field , but an astrophysical vector field and it has negligible effect at small scale .
however , it is expected to change the astrophysical dynamics at larger scale . in this section
, we will use the horndeski formalism to non - minimally couple such a vector field to the metric .
thus , we first introduce as astrophysical vector field @xmath1 with the field tensor @xmath2 .
we denote the source for such a vector field as @xmath3 .
now the energy - momentum tensor for this astrophysical vector field can be written as @xmath4 , and we will also denote the energy - momentum tensor for other fields in the theory by @xmath5 . there is a conserved charge associated with this vector field , as we can write a divergence free current @xmath6 . now using horndeski formalism @xcite , we couple this astrophysical vector field to the metric as @xmath7 where @xmath8 is given by @xmath9
it may be noted that if @xmath10 , this theory reduced to the usual vector - tensor theory .
however , we would like to analyse this modified vector - tensor - horndeski theory .
the dual tensors can be defined as @xmath11 and @xmath12 , where @xmath13}_{[\gamma\delta]}/ 4!$ ] is total asymmetric levi - civita tensor .
thus , we can write the lagrangian for this vector - tensor - horndeski theory as @xmath14 this lagrangian has a non - trivial coupling between the astrophysical vector field and the metric , which was not present in other vector - tensor theories of gravity @xcite .
the static spherically symmetric solutions for a vector field has been studied using the horndeski formalism @xcite .
we will apply this solution to analyse a compact star in this theory , because even though physically , this vector field is an astrophysical vector field , mathematically , this solution will resemble the static spherically symmetric solutions for a horndeski vector field @xcite .
however , unlike the electromagnetic fields which can not have a direct effect on compact stars , this vector field can change the behavior of compact stars .
now for this horndeski astrophysical vector field @xmath15 , we have @xmath16 , and for isotropic matter fields , we can take a perfect fluid with @xmath17 .
the static spherically symmetric metric for a compact star can be written in the schwarzschild - droste coordinates @xmath18 .
furthermore , for this astrophysical vector field , we write @xmath19 . here
we choose the units , such that @xmath20 . now
the metric for this solution can be written as @xmath21 the equation of motion for this solution can be written as @xmath22 we can write the hydrostatic equation for matter sector as@xmath23 for @xmath24 , and thus we obtain @xmath25 now we can write eqs .
( [ 1]-[5 ] ) in terms of thermodynamic parameters , and this can be done by redefine the metric function @xmath26 in terms of a mass function @xmath27 as @xmath28 so , we can obtain the differential change in mass @xmath29 , which is the mass stored in a layer with thickness @xmath30 as @xmath31,\ ] ] where @xmath32 is the mass of the compact object . now make the equations dimensionless by expressed them in terms of @xmath33 , and by using @xmath34 , @xmath35 , @xmath36 , @xmath37 and @xmath38 . here
@xmath39 , @xmath40 is the mass of the sun , and @xmath41 thus , we obtain @xmath42 we will analyse a compact start by solving eqs .
( [ 11]-[44 ] ) .
furthermore , we will also use the equation of state @xmath43 and a specific form of @xmath44 to obtain such solutions .
the function @xmath45 is important in analyzing the geometric mass inside a sphere of radius @xmath46 .
this is because we can use eq .
( [ dm2 ] ) , and write @xmath47 ' . \end{aligned}\ ] ] furthermore , from eq .
( [ 1 ] ) , we obtain @xmath48\label{mr3}.\end{aligned}\ ] ] it may be noted in the absence of the horndeski field , @xmath49 , the eq .
( [ mr3 ] ) reduces to the usual form @xmath50 in general relativity .
now by integrating eq .
( [ mr3 ] ) , we obtain @xmath51\big ) dr\label{m - r}.\end{aligned}\ ] ] now using this result , we can obtain @xmath52 \nonumber \\ & & \big(r_g^2+\frac{2\kappa } { r^2}(1-\frac{3m}{r})\big)=8\pi p\end{aligned}\ ] ] now the equation of state for this system can be written as @xmath43 , it is possible to integrate this equation of state to obtain the behavior of this system in this vector - tensor - horndeski theory .
the boundary conditions used for analyzing this system in this vector - tensor - horndeski theory are similar to those used in the einstein gravity , @xmath53 and @xmath54 .
so , the radius of the compact star can be taken to be @xmath55 , such that the pressure vanishes .
now we can analyse compact stars using this formalism . in astrophysics ,
the term compact star is used to collectively refer to white dwarfs , neutron stars , and black holes .
they are described by tolman - oppenheimer - volkoff equations @xcite . in this section , we will analyse the effect of the horndeski astrophysical vector field on the physics of such compact stars .
the gravitational binding energy of a system is the minimum energy that must be added to that system for it to stop being a a gravitationally bound system .
it is important to analyze the gravitational binding energy of compact stars , such as the neutron star , and such an analysis has been done using general relativity @xcite .
so , in this section , we shall analyze the effect of the astrophysical horndeski field on the gravitational binding energy of a compact star .
thus , using the definition of the mass function ( [ m - r ] ) , the total mass of the matter distribution of this system can be represented by @xmath56 this mass is the arnowitt - deser - misner ( adm ) mass and in the horndeski - vector - tensor theory .
it may be noted that @xmath57 is the corrections to the usual adm mass produced by the the horndeski vector field , @xmath58\big ) dr.\end{aligned}\ ] ] now we can define the density inside a proper volume element @xmath59 as the proper mass , @xmath60 we interpret the difference between the proper ( [ proper ] ) and the total mass ( [ m - r ] ) as the gravitational binding energy @xmath61 in general relativity , due to the absence of vector or horndeski fields , @xmath49 and @xmath62 .
however , as we have a non - minimal coupling to the horndeski field , the adm mass is corrected by a finite @xmath63 . to analyze the effect of the astrophysical vector field on compact stars , we need to solve the modified tolman - oppenheimer - volkoff equation , eq .
( [ tov ] ) . however , to solve this equation , we need to use the equation of state for the interior structure of the star .
it is possible to write this equation of state using the central density , @xmath64 , as a free parameter .
thus , we can obtain the mass and radius of a star by fixing the central density .
this will correspond to choosing a single point in the mass - radius diagram for the star . as this system
is described by a single parameter , we can obtain the full mass - radius curve of the star by varying @xmath65 .
it is possible to study the inner structure of the compact stars , such as the neutron stars @xcite .
this is because it possible to use microscopic many - body simulations to numerically analyze equation of state for such compact stars .
the equation of state for compact stars can be obtained using a mean - field theoretical description of such a system @xcite .
in fact , it is possible to describe the equation of state for a neutron star using the nucleon - nucleon interaction .
the equation of state for such a star is is a polytrope equation of state @xcite , @xmath66 the value of @xmath67 can be taken to be @xmath68 when @xmath69 @xcite .
we can now use this polytrope equation of state to analyze the effect of horndeski field on a neutron star .
, width=302 ] , width=302 ] .
, width=302 ] .
, width=302 ] we have used an adaptive step - size runge - kutta method for analyzing a non - linear - integro - differential equation @xcite,@xcite .
this is because the modified tolman - oppenheimer - volkoff equation is a non - linear - integro - differential equation . in figure .
( [ kappa01 ] ) , blue line represents the vector - tensor - horndeski theory for @xmath70 .
it predicts the existence of a typical neutron star with mass around @xmath71 , and a radius around @xmath72 .
thus , this case is not physical .
in fact , this case can also be used to set a bound on the strength of the coupling of this astrophysical vector field to general relativity . in figure .
( [ kappa0 ] ) , the black line represents the standard general relativity , @xmath73 .
the mass - radius diagram corresponding to this case is given in fig .
( [ kappa05])-([kappa001 ] ) . in fig .
( [ kappa05 ] ) , the red line represents the vector - tensor - horndeski theory with @xmath74 .
it does not predict the existence of a typical neutron star with mass around @xmath71 , and with a radius around @xmath75 . it may be noted that for @xmath76 , the numerical evaluation does not converge .
furthermore , a stable stellar configurations does not exist for this case , as the hydrostatic equilibrium equations are unstable . the mass - radius profile for a weakly coupled vector - tensor - horndeski theory ( @xmath77 ) is plotted in fig .
( [ kappa001 ] ) .
we observed that this profile around the maximum with mass is similar to the profile obtained in the general relativity .
however , for large radius , equilibrium configurations in the vector - tensor - horndeski theory are more massive than in general relativity .
it is possible to analyze a compact star to be represented by a constant density .
it may be noted that the compact stars with constant density have been analysed in general relativity @xcite ( see @xcite for a comprehensive review ) . in general relativity
the tolman - oppenheimer - volkoff equation admits an analytical solution , and this is obtained by imposing @xmath78 . here
the central pressure @xmath79 predicted by general relativity becomes infinite for the critical mass @xmath80 @xcite .
so , it is possible for stars with @xmath81 to indicate a deviation from general relativity @xcite .
thus , it is interesting to analyze the effect of a horndeski vector field for such a system .
we can first analyse a uniform mass density called as top - hat density profile inside the star , @xmath82 here we shall now analyse such a system using modified vector - tensor - horndeski theory .
now for this model and for simplicity , we also choose @xmath83 .
so , the unknown functions for this system are @xmath84 , and the metric function is @xmath85 .
exact solution for the vector - tensor - horndeski theory , for this system can be written as @xmath86 now using ( [ m ] ) and eq .
( [ dm ] ) , we obtain @xmath87 the metric of this compact star can be written as @xmath88 now using this mass profile , we can integrate eq , ( [ 44 ] ) , and obtain @xmath89 here we have assumed that this astrophysical vector field satisfies the following initial condition on the surface of star @xmath90 , @xmath91 as the original function is corrected by a @xmath92 dependent terms , it can be argued that such corrections are produced by the horndeski vector field .
as such a coupling between gravity and horndeski vector field is constraint by experimental data , it is possible to use the astrophysical data to constraint such a coupling .
we analyzed a compact star with a constant density .
however , for real stars , we expect that the density to be a function of @xmath46 .
so , it is interesting to analyse various different functional dependence of @xmath45 , and analyse the effect of horndeski vector field on a system described by such functions .
this is important to demonstrate that the dependence of the system on horndeski vector field is not a special feature of a system with a constant density .
now we can take a simple function @xmath45 of @xmath46 , such as @xmath93 , to demonstrate that the horndeski vector field effects the systems in which the density is not a constant .
now for such a system , we obtain @xmath94 the mass density of this system can be obtained from eq .
( [ 11 ] ) , and it is given by @xmath95 where we have defined @xmath96 as @xmath97 the pressure for this system can also be obtained , and it is given by @xmath98 where we have defined @xmath99 as @xmath100 thus , the pressure of this system is effected by the horndeski vector field .
this is because the pressure of this system is corrected by terms proportional to @xmath92 , and so this system is effected by the horndeski vector field .
it may be noted that it is possible to take other functions describing @xmath45 , and analyse the pressure of the compact stars using those functions .
this procedure can be repeated for those functions .
it is expected that the pressure in such systems will also depend on the horndeski vector field .
so , for example , we can take another form of the function @xmath101 , and demonstrate that the system described by this function will also be effected by the horndeski vector field .
thus , using this function , we obtain @xmath102 where @xmath103 we observe the this system is again corrected by terms proportional to @xmath92 , and so the physics of compact starts is effected by the horndeski vector field .
this function can be used to obtain the mass density and pressure of the compact star . as the original function were corrected by the horndeski vector field ,
the mass density and pressure of the compact star in this theory would again depend on the horndeski vector field .
these results can be compared with experimental data , and bounds on the existence of such an astrophysical vector field can be thus obtained .
thus , we have analyzed compact stars in a vector - tensor - horndeski theory , and observed that the dynamics of this system is corrected by terms which are proportional to the coupling constant of the horndeski field .
it may be noted that as we are using an astrophysical vector field , it is possible that such a vector field will also contain monopoles .
furthermore , as this vector field will have negligible effect at small distance , we can not rule out the existence of such monopoles in this vector field .
these monopoles can effect the astrophysical phenomena , and have a direct effect on the physics of compact stars .
thus , if we assume a monopole with charge @xmath104 is located at @xmath105 , then we can write @xmath106 .
the exact solution for this system can now be written as @xmath107 as the constant density solutions have been studied in general relativity @xcite , it is interesting to analyse different limits of such solutions .
it is possible to take a constant density solution , @xmath108 with @xmath83 , and the analyse the effect of horndeski vector field using @xmath109 , where @xmath110 is a constant . now using this form of @xmath111 ,
we can obtain @xmath112 now using this mass profile we can integrate eq .
( [ tov ] ) , and obtain the pressure as @xmath113 where we have defined @xmath114 as @xmath115 so , the existence of astrophysical monopole from horndeski vector field can correct the pressure of a compact star .
thus , astrophysical monopoles can have interesting effects on the physics of compact stars in a horndeski - vector - tensor theory of gravity . as the monopoles have not been detected in the electromagnetic vector field ,
there are strong constrains of including the effects of such monopoles in physical systems .
however , the horndeski vector field is different from electromagnetic vector field , so the constraint on the electromagnetic vector field from the absence of electromagnetic monopoles do not apply to such horndeski vector fields , and hence it is possible that such vector fields can change the physics of compact stars .
in this paper , we have analysed a theory of modified gravity . in this theory , the field content of general relativity
was increased to include an astrophysical vector field .
we have used the horndeski formalism to non - minimally couple this astrophysical vector field to the metric .
as we have used the horndeski formalism , this theory did not contain any ostrogradsky ghost degree of freedom .
we will analysed a compact star using this vector - tensor - horndeski theory .
thus , we analyzed the effect of such a horndeski vector field on the gravitational binding energy .
we used a polytrope equation of state for analyzing a neutron star in horndeski - vector - tensor theory of gravity .
we analysed this system using an adaptive step - size runge - kutta method @xcite .
we also analysed various different cases for this system , and obtained the mass density and pressure for the compact stars corresponding to those cases .
it was demonstrated that the horndeski changed the physics of this system for these different cases .
it would be interesting to compare the results of this paper to experimental data , and thus obtain bounds on the vector - tensor - horndeski theory .
finally , we proposed that it is possible for a monopole to exist in such a modified theory of gravity .
as no monopole has been detected in the electromagnetic field , there are strong constrains on the existence of monopoles in such a theory .
however , the horndeski vector field was not constraint by the constraint on the electromagnetic vector field , and so we analyzed the effects of an astrophysical monopole on the compact stars .
it was demonstrated that such an astrophysical vector field would change the pressure of the compact star .
it may be noted that it is possible to analyse compact stars with an anisotropy .
in fact , multipole moments for compact stars with anisotropic pressure have been studied @xcite .
the compact stars with anisotropy have been studied using a modified tolman - oppenheimer - volkoff equation @xcite .
it has been observed that the pressure anisotropy can effect the surface tension of these stars .
this is because the anisotropy decreases the value of the surface tension
. it would be interesting to introduce such an anisotropy using a vector field .
furthermore , this vector field can be coupled non - minimally to the metric using horndeski formalism .
it would be interesting to analyse the phenomenological effects of such a model .
the vector fields have been used to modify general relativity , and it has been possible to use this modified theory of gravity to obtain the correct dynamics of galaxies without the need for dark matter @xcite .
it would be interesting to analyse the dynamics of galaxies using a horndeski vector field
. it might be possible to obtain the correct dynamics of galaxies by suitable modifying such a theory , and by possible adding other fields to it .
it would also be interesting to analyse the effects of such a theory on inflation .
this is because vector fields have been used to study inflation @xcite , and it would be interesting to repeat this calculation using the horndeski formalism .
it is possible to make observations on a neutron star , and these observations can be compared with the predictions of the vector - tensor - horndeski theory .
this can be used to both verify the existence of a horndeski vector field , if such effects are detected . however , if such effects are not detected then it can be used to set bounds on the strength of the coupling of such a astrophysical vector field to general relativity .
it may be noted that as the @xmath74 does not predict the existence of a neutron star with a typical mass and radius , so this case does not fit the experimental data .
thus , this case is not physical .
in fact , this case also establish a bound on the strength of coupling parameter of the horndeski vector field to general relativity . it may be noted that this coupling does change the behavior of mass - radius diagram .
such a mass - radius diagram of a neutron star can be observed , and the observations can be compared with this analysis .
this can be used to test the existence of such a horndeski astrophysical vector field , and also establish a bound on the strength of coupling of such a field .
it may be noted that it is possible to obtain such experimental data for neutron stars using gravitational lensing @xcite .
however , it is also important to analyze the effect of horndeski vector field on the gravitational lensing for such an analysis .
it is possible to use the burst oscillations in the x - ray flux to obtain the observational behavior of mass - radius of a neutron star @xcite .
it is also possible to use quescent neutron stars to make such observations on a neutron star @xcite
. it would be interesting to analyze the effect of the horndeski vector field on neutron stars using such data
. it would be interesting to analyze other effects of this horndeski vector field which can be observed using compact stars .
the accretion in the reissner
nordstrom spacetime has already been studied @xcite .
it was observed that the electromagnetic field can have interesting effect on such an accretion around a compact star .
in fact , the effect of a monopole on the accretion has also been studied @xcite .
as the astrophysical objects are not changed , such systems can not be physically realized .
however , in this paper , we have proposed that a horndeski astrophysical vector field can couple to general relativity , and the bounds on the strength of such an coupling can be obtained from observations . such an astrophysical vector field will also change the accretion around compact stars .
it would be interesting to analyze the accretion in this vector - tensor - horndeski theory
. it would also be interesting to compare the result thus obtained to observations , and test the vector - tensor - horndeski theory .
a. g. riess et al .
j. 116 , 1009 ( 1998 ) s. perlmutter et al . ,
nature 391 , 51 ( 1998 ) a. g. riess et al . , astron .
j. 118 , 2668 ( 1999 ) s. perlmutter et al . , astrophys .
j. 517 , 565 ( 1999 ) a. g. riess et al .
, astrophys .
j. 560 , 49 ( 2001 ) j. l. tonry et al .
, astrophys .
j. 594 , 1 ( 2003 ) t. clifton , p. g. ferreira , a. padilla and c. skordis , phys .
513 , 1 ( 2012 ) v. acquaviva , c. baccigalupi , f. perrotta , phys .
d70 , 023515 ( 2004 ) r. klein , m. ozkan and d. roest , phys . rev .
d 93 , 044053 ( 2016 ) e. babichev and c. charmousis , jhep 1408 , 106 ( 2014 ) m. jamil , d. momeni and r. myrzakulov , eur .
j. c * 73 * , no . 3 , 2347 ( 2013 ) doi:10.1140/epjc / s10052 - 013 - 2347 - 4 d. momeni , m. j. s. houndjo , e. gdekli , m. e. rodrigues , f. g. alvarenga and r. myrzakulov , int .
j. theor .
* 55 * , no . 2 , 1211 ( 2016 ) doi:10.1007/s10773 - 015 - 2762 - 4 [ arxiv:1412.4672 [ gr - qc ] ] .
a. nicolis , r. rattazzi and e. trincherini , phys . rev .
d 79 064036 ( 2009 ) g. w. horndeski , int . j. theor .
10 , 363 ( 1974 ) c. deffayet , x. gao , d. a. steer and g. zahariade , phys . rev .
d 84 , 064039 ( 2011 ) k. land and j. magueijo , phys .
95 , 071301 ( 2005 ) k. land and j. magueijo , mnras 378 , 153 ( 2007 ) h. k. eriksen , f. k. hansen , a. j. banday , k. m. gorski and p. b. lilje , astrophys .
j. 605 , 14 ( 2004 ) l. h. ford , phys .
d 40 , 967 ( 1989 ) a. golovnev , v. mukhanov and v. vanchurin , jcap 0806 , 009 ( 2008 ) c. g. boehmer and t. harko , eur .
j. c 50 , 423 ( 2007 ) t. koivisto and d. f. mota , jcap 0808 , 021 ( 2008 ) j. beltran jimenez and a. l. maroto , phys .
d 78 , 063005 ( 2008 ) j. beltran jimenez , r. lazkoz and a. l. maroto , phys .
d 80 , 023004 ( 2009 ) m. -a .
watanabe , s. kanno and j. soda , phys .
102 , 191302 ( 2009 ) j. soda , class .
29 , 083001 ( 2012 ) a. maleknejad , m. m. sheikh - jabbari and j. soda , phys .
528 , 161 ( 2013 ) m. thorsrud , d. f. mota and s. hervik , jhep 1210 , 066 ( 2012 ) m. prakash , equation of state , in the nuclear equation of state , ed . by a. ausari , l. satpathy ( world scientific , singapore , 1996 ) . w. h. press , s. a. teukolsky , w. t. vetterling and b. p. flannery , numerical recipes , the art of scientific computing , ( cambridge university press , 2007 ) a. m. oliveira , h. e. s. velten , j. c. fabris and l. casarini , phys .
d * 92 * , no .
4 , 044020 ( 2015 ) doi:10.1103/physrevd.92.044020 [ arxiv:1506.00567 [ gr - qc ] ] .
buchdahl , phys .
116 , 1027 ( 1959 ) a. fuzfa , j. m. gerard and d. lambert , gen .
34 , 1411 ( 2002 ) p. s. joshi and d. malafarina , eur .
j. c 75 , 596 ( 2015 ) s. schlgel , arxiv:1610.03622 [ gr - qc ] . k. yagi and n. yunes , phys .
d 91 , 103003 ( 2015 ) r. sharma ands .
d. maharaj , j. astrophys .
28 , 133 ( 2007 ) a. f. boden , m. shao and d. van buren , astrophys .
j. 502 , 538 ( 1997 ) j.miralda-escude , astrophys . j. lett .
470 , l113 ( 1996 ) t. s. olson , phys .
c 63 , 015802 ( 2001 ) e. f. brown , l. bildsten and r. e. rutledge , astrophys .
j. 504 , l95 ( 1998 ) f. ficek , class . quant .
32 , 235008 ( 2015 ) a. k. ahmed , u. camci and m. jamil , class .
33 , 215012 ( 2016 )
|
in this paper , we will analyse a theory of modified gravity , in which the field content of general relativity will be increased to include a vector field .
we will use the horndeski formalism to non - minimally couple this vector field to the metric . as we will be using the horndeski formalism
, this theory will not contain ostrogradsky ghost degree of freedom .
we will analyse compact stars using this vector - tensor - horndeski theory .
|
recent three dimensional ( 3d ) radiation hydrodynamical simulations of the sun and main - sequence stars have led to significant improvements in our knowledge of stellar atmospheres and spectroscopic abundance measurements ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the major differences between the 1d and 3d model atmospheres result from the insufficient description of convection by the mixing length theory ( * ? ? ?
* mlt ) that is used in the 1d models .
convection becomes important in hydrogen - atmosphere ( da ) white dwarfs ( wds ) below about 15,000 k. spectroscopic analyzes of large samples of wds with 1d model atmospheres show the so - called `` high @xmath1 problem '' @xcite , where the spectroscopically determined masses of wds cooler than 13,000 k are systematically higher than their hotter counterparts by as much as 20% .
there is no known evolution effect or observational bias that should lead to such a change in mass for the cooler wds , and a larger average mass is not observed for cool wds with parallax measurements . in a series of papers , @xcite presented the first 3d model atmosphere calculations for da wds using the co@xmath4bold radiation - hydrodynamics code @xcite .
the predicted balmer line profiles are significantly different in the 3d models , and the `` high @xmath1 problem '' essentially disappears in the spectroscopic analyzes based on the 3d models .
the previously published model grid covers the effective temperature range of 6000 - 15,000 k and the surface gravity range of @xmath1 = 79 . to study the effects of 3d models on the physical parameters of the extremely low - mass ( elm ) wds with @xmath1 @xmath5 , here
we extend our 3d model atmosphere grid down to @xmath1 = 5 .
elm wds are helium - core remnants and almost always found in short period binary systems . the elm survey ( * ? ? ?
* ; * ? ? ?
* and references therein ) has discovered binaries with periods as short as 1220 min @xcite . improving the constraints on the elm wd physical parameters , including the spectroscopic mass and gravitational wave strain ,
is important for future gravitational wave missions in the milli - hertz frequency range .
in addition , all of the known eclipsing and/or tidally distorted double wd systems involve low - mass wds .
studying the mass - radius relation for these systems , @xcite and @xcite find that the 1d spectroscopic analysis underestimates the radius ( or overestimates the surface gravity ) for elm wds cooler than about 10,000 k. nltt 11748 is an excellent example , where the eclipses constrain the radius and surface gravity to @xmath6 = 0.04230.0433 r@xmath7 and @xmath1 = 6.326.38 @xcite , while the 1d spectroscopic analysis predicts @xmath1 = 6.83 @xcite .
therefore , the `` high @xmath1 problem '' seems to impact all convective da wds , including elm wds .
we outline the details of our 3d model calculations in section 2 , and present the 1d to 3d atmospheric parameter correction functions for the extended range of @xmath8 in section 3 .
we discuss the astrophysical implications , including the revised parameters for the elm survey sample , radius constraints from the eclipsing and tidally distorted systems as well as elm wd + milli - second pulsar binaries , and the zz ceti instability strip in section 4 .
we conclude in section 5 .
we have carried out 30 simulations of pure - hydrogen elm wd atmospheres with co@xmath4bold @xcite .
these 3d simulations in the range @xmath9 provide a direct extension to the grid presented in @xcite for @xmath10 .
the calculations also complement the cifist grid of co@xmath4bold simulations for stars and giants @xcite with @xmath11 .
figure [ fg : fhr ] illustrates the new elm wd 3d simulations in a hr - type diagram .
properties of the individual models , such as @xmath0 ( derived from the temporally and spatially averaged emergent stellar flux ) , @xmath1 , box dimensions , and computation time , can be found in table [ tab : grid ] .
@rccccccr@l & @xmath1 & @xmath12 & @xmath13 & @xmath14 & @xmath15 & @xmath16 & + & ( cm s@xmath3 ) & ( cm ) & ( cm ) & & & ( s ) & + 5977 & 5.00 & 7.56 & 8.38 & @xmath176.58 & 2.96 & 4.00 & 1&.17 + 6958 & 5.00 & 7.71 & 8.46 & @xmath176.08 & 2.99 & 4.00 & 1&.03 + 7513 & 5.00 & 7.62 & 8.33 & @xmath175.64 & 2.99 & 4.00 & 0&.92 + 8501 & 5.00 & 8.36 & 8.64 & @xmath177.77 & 3.06 & 3.30 & 0&.94 + 9001 & 5.00 & 8.38 & 8.64 & @xmath175.46 & 3.00 & 3.30 & 1&.19 + 9499 & 5.00 & 8.49 & 8.68 & @xmath178.29 & 3.01 & 3.30 & 1&.54 + 6044 & 5.50 & 7.04 & 7.66 & @xmath176.98 & 2.97 & 3.50 & 0&.75 + 6915 & 5.50 & 7.20 & 7.73 & @xmath176.58 & 3.01 & 3.50 & 0&.62 + 7967 & 5.50 & 7.37 & 7.98 & @xmath176.57 & 3.02 & 3.50 & 0&.47 + 8996 & 5.50 & 7.93 & 8.18 & @xmath177.47 & 3.02 & 2.90 & 0&.46 + 9499 & 5.50 & 7.93 & 8.18 & @xmath176.12 & 3.00 & 2.90 & 0&.67 + 9998 & 5.50 & 8.00 & 8.22 & @xmath176.75 & 3.01 & 2.90 & 0&.91 + 10500 & 5.50 & 8.06 & 8.29 & @xmath177.23 & 3.01 & 2.90 & 1&.40 + 5958 & 6.00 & 6.66 & 7.20 & @xmath177.21 & 2.99 & 3.00 & 0&.38 + 6971 & 6.00 & 6.78 & 7.28 & @xmath176.99 & 3.02 & 3.00 & 0&.21 + 7962 & 6.00 & 6.89 & 7.34 & @xmath176.82 & 2.99 & 3.00 & 0&.09 + 8963 & 6.00 & 7.00 & 7.42 & @xmath175.76 & 2.96 & 3.00 & @xmath170&.08 + 9491 & 6.00 & 7.32 & 7.65 & @xmath175.07 & 3.00 & 2.40 & @xmath170&.03 + 9999 & 6.00 & 7.46 & 7.71 & @xmath177.34 & 3.00 & 2.40 & 0&.21 + 10499 & 6.00 & 7.52 & 7.75 & @xmath175.75 & 3.01 & 2.40 & 0&.40 + 11001 & 6.00 & 7.49 & 7.75 & @xmath175.00 & 2.96 & 2.40 & 0&.77 + 5986 & 6.50 & 6.10 & 6.57 & @xmath177.01 & 2.99 & 2.50 & @xmath170&.03 + 6922 & 6.50 & 6.20 & 6.65 & @xmath177.48 & 3.00 & 2.50 & @xmath170&.19 + 7990 & 6.50 & 6.35 & 6.84 & @xmath176.76 & 2.99 & 2.50 & @xmath170&.31 + 8907 & 6.50 & 6.51 & 7.03 & @xmath176.66 & 3.00 & 2.50 & @xmath170&.42 + 9412 & 6.50 & 6.54 & 6.95 & @xmath175.29 & 3.02 & 2.50 & @xmath170&.49 + 9989 & 6.50 & 6.95 & 7.24 & @xmath176.75 & 3.02 & 2.50 & @xmath170&.49 + 10500 & 6.50 & 6.99 & 7.24 & @xmath175.80 & 3.01 & 2.05 & @xmath170&.32 + 11000 & 6.50 & 7.05 & 7.28 & @xmath175.77 & 3.01 & 2.05 & @xmath170&.13 + 11499 & 6.50 & 7.09 & 7.33 & @xmath175.83 & 3.01 & 2.05 & 0&.12 + 6011@xmath18 & 5.00 & 8.04 & 8.08 & @xmath177.93 & 8.38 & 5.00 & 1&.08 + [ tab : grid ] @rcccrrc@ & @xmath1 & @xmath19 char .
size & @xmath20 & & & @xmath19 mach + & ( cm s@xmath3 ) & ( cm ) & ( cm ) & & & + 5977 & 5.00 & 7.50 & 6.70 & 1.53 & 16.91 & @xmath170.42 + 6958 & 5.00 & 7.54 & 6.74 & 1.62 & 23.95 & @xmath170.26 + 7513 & 5.00 & 7.57 & 6.77 & 1.65 & 23.17 & @xmath170.12 + 8501 & 5.00 & 7.83 & 6.86 & 1.51 & 18.34 & @xmath170.07 + 9001 & 5.00 & 7.69 & 6.92 & 1.45 & 11.17 & @xmath170.26 + 9499 & 5.00 & 7.77 & 6.99 & 1.49 & 3.80 & @xmath170.58 + 6044 & 5.50 & 6.78 & 6.21 & 1.12 & 13.00 & @xmath170.50 + 6915 & 5.50 & 6.91 & 6.25 & 1.15 & 21.22 & @xmath170.34 + 7967 & 5.50 & 7.08 & 6.31 & 1.17 & 23.14 & @xmath170.14 + 8996 & 5.50 & 7.42 & 6.41 & 1.19 & 18.83 & @xmath170.06 + 9499 & 5.50 & 7.27 & 6.47 & 0.98 & 13.01 & @xmath170.22 + 9998 & 5.50 & 7.32 & 6.53 & 0.79 & 6.93 & @xmath170.42 + 10500 & 5.50 & 7.46 & 6.62 & 0.80 & 1.84 & @xmath170.86 + 5958 & 6.00 & 6.23 & 5.71 & 0.86 & 8.02 & @xmath170.62 + 6971 & 6.00 & 6.40 & 5.77 & 0.80 & 17.55 & @xmath170.43 + 7962 & 6.00 & 6.53 & 5.82 & 0.80 & 22.37 & @xmath170.27 + 8963 & 6.00 & 6.63 & 5.88 & 0.80 & 21.21 & @xmath170.04 + 9491 & 6.00 & 6.81 & 5.94 & 0.80 & 19.55 & @xmath170.05 + 9999 & 6.00 & 6.82 & 6.01 & 0.58 & 13.76 & @xmath170.24 + 10499 & 6.00 & 6.83 & 6.07 & 0.37 & 7.65 & @xmath170.39 + 11001 & 6.00 & 6.81 & 6.15 & 0.64 & 2.13 & @xmath170.73 + 5986 & 6.50 & 5.73 & 5.21 & 0.43 & 5.13 & @xmath170.71 + 6922 & 6.50 & 5.89 & 5.28 & 0.37 & 12.49 &
@xmath170.52 + 7990 & 6.50 & 6.00 & 5.33 & 0.33 & 19.63 & @xmath170.36 + 8907 & 6.50 & 6.13 & 5.38 & 0.37 & 21.18 & @xmath170.21 + 9412 & 6.50 & 6.31 & 5.42 & 0.38 & 20.45 & @xmath170.10 + 9989 & 6.50 & 6.47 & 5.48 & 0.30 & 19.67 & @xmath170.06 + 10500 & 6.50 & 6.40 & 5.55 & 0.17 & 16.19 & @xmath170.19 + 11000 & 6.50 & 6.35 & 5.60 & 0.00 & 9.80 & @xmath170.35 + 11499 & 6.50 & 6.36 & 5.68 & @xmath170.20 & 4.90 & @xmath170.56 + 6011@xmath18 & 5.00 & 7.34 & 6.60 & 1.76 & 17.79 & @xmath170.38 + [ tab : gran ] the numerical setup for the elm wd simulations is the same as that described in @xcite for a grid of @xmath21 points .
the properties of the boundary conditions are reported in ( * ? ? ?
* see sect .
the warmest simulations , represented by filled symbols in figure [ fg : fhr ] , have a convective zone that is thinner than the vertical dimension of the atmosphere . in those cases
, we use a bottom layer that is closed to convective flows with imposed zero vertical velocities . in cooler models
, the convective flux is transported in and out of the domain through an open boundary . in all cases
the lateral boundaries are periodic and the top boundary is open to material flows and radiation .
we use the same equation - of - state ( eos ) and opacity tables as those described in @xcite .
in particular , we employ the @xcite eos , the stark broadening profiles of @xcite , and the quasi - molecular line opacity of @xcite .
the wavelength - dependent opacities are sorted in the 8 bin configuration discussed in @xcite .
the elm wd 3d simulations show physical properties that are very similar to those of the 3d models of higher mass wds presented in earlier studies .
@xcite found that 3d effects on the thermal structure are well correlated with the density at rosseland optical depth @xmath22 .
since the atmospheric density can be kept roughly constant by decreasing both @xmath1 and @xmath0 , the elm simulations have properties similar to those of c / o - core wds but with a shift towards smaller @xmath0 values .
figure [ fg : f3da ] compares our simulation at @xmath237500 k and @xmath1 = 5.0 with the simulation at @xmath239500 k and @xmath1 = 7.0 from @xcite .
both simulations have a similar relative intensity contrast and granulation is visually comparable .
table [ tab : gran ] demonstrates that elm wds have a slightly larger maximum relative intensity contrast compared to higher mass degenerates .
elm wds are very similar to main - sequence simulations in that respect ( see figure 10 of * ? ? ?
furthermore , table [ tab : gran ] describes the characteristic size of granulation and decay time .
these quantities follow the same trends as those observed for higher mass wds @xcite , in particular the relation between the characteristic size of granulation to pressure scale height ratio and photospheric mach number .
our low gravity wd simulations reach photospheric plasma conditions and characteristic geometrical scales that are similar to main - sequence stars , although the radii are quite different . in order to better understand this link ,
we have computed a stellar model at @xmath24 6000 k , @xmath25 , @xmath26 = 8.51% , and a metallicity of [ m / h ] = @xmath174.0 based on solar elemental abundances from @xcite with updated cno values @xcite .
the model is an extension of the cifist grid for main - sequence stars where @xmath27 .
the simulation is representative of the extremely metal - poor f - dwarfs with the currently highest derived surface gravities @xcite .
the properties of the simulation are described in tables 1 and 2 , while figure [ fg : f3db ] shows the comparison with the pure - hydrogen elm wd at @xmath0 @xmath23 6000
k and @xmath1 = 5.0 .
the relative intensity contrast and size of the granules are similar , although the stellar atmosphere is denser by 58% at @xmath22 compared to the more opaque pure - hydrogen wd atmosphere .
it implies that their properties are only expected to be qualitatively similar .
nevertheless , it demonstrates that elm wds will provide an important connection between wds and main - sequence stars in future studies of the 3d effects on the determination of the atmospheric parameters from balmer line profiles .
we have computed mean 3d structures , hereafter @xmath28 structures , which are spatial and temporal averages of @xmath29 and @xmath30 performed over surfaces of constant @xmath31 and for 12 random snapshots .
we have employed those snapshots to define the @xmath0 values of our simulations ( table 1 ) . in figure
[ fg : ftstruct ] , we compare the temperature structures of @xmath32 simulations and 1d models at @xmath0 @xmath23 7000 k in the range @xmath33 .
convection is still fairly adiabatic in the photosphere for @xmath34 in this @xmath0 range , resulting in small differences between the @xmath32 and 1d cases except for the optically thin layers where 3d overshoot causes a significant dynamic cooling .
density decreases for lower surface gravities and convection becomes less efficient in the photosphere .
figure [ fg : ftstruct ] demonstrates that differences between @xmath32 and 1d structures become larger for elm wds at @xmath24 7000 k. the observed pattern is very similar to the one at constant @xmath1 and variable @xmath0 in figure 7 of @xcite .
we have computed spectra from the 3d simulations in order to study the 3d effects on the determination of atmospheric parameters .
first of all , we rely on mean spectra from @xmath32 structures , hereafter @xmath32 spectra . for wd simulations with @xmath34 ,
the normalized balmer line profiles computed from @xmath32 structures are nearly identical to the results of a full 3d spectral synthesis @xcite .
however , we have found that the @xmath32 spectra approximation becomes questionable at lower gravities , in line with main - sequence simulations where 3d inhomogeneities have a significant direct impact on predicted balmer line profiles @xcite .
we review first the differences between @xmath32 and 1d spectra and then consider the additional effects from 3d spectral synthesis in sect . 3.2 . figure [ fg : fprofiles ] shows the comparison of predicted @xmath35 and 1d balmer line profiles for three simulations at @xmath36 , which is an illustrative case for elm wds .
the 1d spectra adopt the ml2/@xmath37 = 0.8 parameterization of the mlt @xcite and the same microphysics as the 3d calculations .
the elm results can be compared to the case at @xmath38 in figure 16 of @xcite .
one major difference is that up to ten balmer lines are typically fitted simultaneously for elm wds , since the non - ideal effects and the sensitivity to surface gravity are shifted to higher balmer lines @xcite . for the simulation at @xmath39
k and @xmath40 in the middle panel of figure [ fg : fprofiles ] , it is only for h10 and higher lines in the series that @xmath32 profiles become shallower than 1d profiles due to differences in the predicted non - ideal effects .
the latter are very sensitive to the density stratification in the atmosphere @xcite .
the shape of the lower balmer lines is also significantly different between @xmath32 and 1d spectra at intermediate @xmath0 , which is mostly responsible for @xmath0 shifts .
the @xmath32 and 1d balmer lines become identical at high @xmath0 values ( see the right panel of figure [ fg : fprofiles ] ) . for elm wds , this transition appears at relatively low temperatures , and 3d effects are largely negligible on the thermal structures for @xmath41 k. similarly to c / o - core wds , convection becomes very inefficient in this regime and the thermal structure is essentially fixed by the radiation field , even though convective velocities are still large . for cool @xmath0 values , the left panel of figure [ fg : fprofiles ] suggests that 3d effects are small , but section 3.2 demonstrates that the @xmath32 approximation is not always good in this regime .
figure [ fg : fshifts ] presents the @xmath32 atmospheric parameter corrections found by fitting the @xmath28 spectra with our standard grid of 1d spectra .
the corrections were derived simultaneously for the nine balmer lines from h@xmath42 to h12 in the same way we fit observations .
the @xmath28 structures deviate significantly from their 1d counterparts in the upper layers ( @xmath43 ) due to the cooling effect of convective overshoot .
this results in deep @xmath35 cores for h@xmath37 and h@xmath42 , as observed in the middle panel of figure [ fg : fprofiles ] .
while we have no suggestion that the 3d simulations are inaccurate , there is no observational evidence for an adiabatic surface cooling in wds and it appears premature to account for these effects . as a consequence , the line cores were partially removed from the fits , as well as the entire h@xmath37 line , as discussed in sect .
3.1 of @xcite .
the @xmath32 atmospheric parameter corrections for elm wds in figure [ fg : fshifts ] follow the trend observed for higher gravity objects , despite the fact that more lines are included in the fitting procedure in the former case .
the @xmath35 @xmath1 corrections have a similar strength for elm and c / o - core wds , although they are restricted to a narrower range of @xmath0 at low surface gravity .
we have employed the linfor3d three - dimensional spectral synthesis code @xcite to compute 3d h@xmath42 profiles for elm white dwarfs . unlike our previous experiments for @xmath34 ,
the 3d spectra and their @xmath32 counterparts from properly averaged @xmath32 structures differ by a few percent , with the largest effects at lower gravities .
it is currently out of the scope of this work to proceed with the time - consuming computation of a full grid of 3d synthetic spectra .
however , to further constrain the precision of the 3d corrections , we have recomputed our grid of elm spectra with the so - called 1.5d approximation @xcite . under this formalism , we assume that the physical conditions do not vary in the horizontal direction for each point on the top of the computational box , i.e. each column is a plane - parallel atmosphere .
the 1.5d spectrum is then the average over all @xmath44 columns , for which the emergent 1d spectra can be easily calculated from our regular spectral synthesis code .
the full 3d spectral synthesis has the effect of coupling nearby grid points , hence it is expected to lie somewhere between the extreme cases of the 1.5d and @xmath32 approximations . in figure [ fg : fshifts ] , we compare the 1.5d ( blue ) and @xmath35 ( black ) atmospheric parameter corrections .
we notice that while the @xmath32 approximation is relatively good at @xmath45 , especially in terms of @xmath1 corrections , it becomes questionable at lower gravities .
we remind the reader that the 1.5d approximation overestimates 3d effects , hence suggests an upper limit to the 3d corrections . since both 1.5d and @xmath32 corrections have the same signs
, we suggest that applying the @xmath32 corrections provides a minimum 3d effect and a reasonable estimate of the atmospheric parameters of elm wds .
the deficiency of the @xmath32 approximation for elm wds is likely explained by multiple factors .
the temperature fluctuations are slightly larger at lower gravities although this could not alone explain the variations . on the other hand , we find that the line to continuum emergent intensity ratio has a significantly larger spatial variation for lower surface gravities .
it implies that fluctuations are less likely to cancel in the 3d spectral synthesis .
this could be related to the fact that broadening due to collisions with neutral atoms becomes increasingly important at the low @xmath0 values where strong 3d effects are observed in elm wds .
finally , we have verified that the @xmath32 approximation is not the cause for the slight residual bump in the 3d wd mass distribution at @xmath46 k @xcite or the presence of deep cores for the lower balmer lines compared to 1d spectra .
these features are also present when using 1.5d spectra .
@rrrrr@ & @xmath1 = 5.0 & @xmath1 = 5.5 & @xmath1 = 6.0 & @xmath1 = 6.5 + & ( cm s@xmath3 ) & ( cm s@xmath3 ) & ( cm s@xmath3 ) & ( cm s@xmath3 ) + 6000 & 6 & 12 & 13 & 24 + 6500 & @xmath1733 & @xmath1719 & @xmath1711 & 7 + 7000 & @xmath1764 & @xmath176 & 5 & 21 + 7500 & @xmath17203 & @xmath1750 & @xmath1719 & 16 + 8000 & @xmath17460 & @xmath17202 & @xmath17105 & @xmath1766 + 8500 & @xmath17555 & @xmath17425 & @xmath17197 & @xmath17152 + 9000 & @xmath17288 & @xmath17564 & @xmath17338 & @xmath17154 + 9500 & @xmath1799 & @xmath17243 & @xmath17557 & @xmath17151 + 10000 & 0 & @xmath17155 & @xmath17237 & @xmath17466 + 10500 & 0 & 78 & @xmath17116 & @xmath17211 + 11000 & 0 & 0 & 136 & 13 + 11500 & 0 & 0 & 0 & 123 + [ tab : tcorr ] we have derived functions to convert 1d atmospheric parameters of elm wds ( @xmath47 7.0 ) to their 3d counterparts . we have ensured that the transition is smooth at @xmath48 7.0 with the functions already presented in @xcite , which are more precise in the range 7.0 @xmath49 9.0 .
fortran 77 functions are available in the online appendix a. one can also rely on the data from tables [ tab : tcorr ] and [ tab : gcorr ] to derive alternative functions or request @xmath32 spectra to the authors .
our independent variables are defined as @xmath50 } - 8.0 ~,\ ] ] @xmath51 } - 10,000}{1000 } ~,\ ] ] and the fitting functions for the ml2/@xmath37 = 0.8 parameterization of the mlt are given below with the numerical coefficients identified in table [ tab : coeff ] .
@xmath52 } = ~&a_{13}+\big(a_{1}+a_{4}\exp\big[(a_{5}+a_{7}\exp[a_{8}t_{\rm x } \\ & + a_{9}g_{\rm x}])t_{\rm x}+a_{6}g_{\rm x}\big]\big)\exp\big((a_{2}+ \\ & a_{10}\exp[a_{11}t_{\rm x}+a_{12}g_{\rm x}])t_{\rm x}+a_{3}g_{\rm x}\big ) , \end{split}\ ] ] @xmath53 } = ~&b_{13}+(b_{1}+b_{10}\exp[b_{11}t_{\rm x}+b_{12}g_{\rm x } ] ) \\ & \exp\big([b_{2}+b_{4}\exp(b_{5}t_{\rm x}+b_{6}g_{\rm x})]t_{\rm x } \\ & + [ b_{3}+b_{7}\exp(b_{8}t_{\rm x}+b_{9}g_{\rm x})]g_{\rm x}\big ) . \end{split}\ ] ] @rrrrr@ & @xmath1 = 5.0 & @xmath1 = 5.5 & @xmath1 = 6.0 & @xmath1 = 6.5 + & ( cm s@xmath3 ) & ( cm s@xmath3 ) & ( cm s@xmath3 ) & ( cm s@xmath3 ) + 6000 & 0.010 & @xmath170.019 & @xmath170.017 & @xmath170.018 + 6500 & @xmath170.023 & @xmath170.037 & @xmath170.040 & @xmath170.024 + 7000 & @xmath170.128 & @xmath170.117 & @xmath170.122 & @xmath170.134 + 7500 & @xmath170.239 & @xmath170.225 & @xmath170.215 & @xmath170.239 + 8000 & @xmath170.218 & @xmath170.316 & @xmath170.315 & @xmath170.350 + 8500 & @xmath170.064 & @xmath170.181 & @xmath170.302 & @xmath170.340 + 9000 & @xmath170.035 & 0.054 & @xmath170.199 & @xmath170.283 + 9500 & 0.001 & 0.006 & 0.056 & @xmath170.150 + 10000 & 0.000 & 0.044 & 0.021 & 0.004 + 10500 & 0.000 & 0.037 & 0.027 & 0.002 + 11000 & 0.000 & 0.000 & 0.034 & 0.007 + 11500 & 0.000 & 0.000 & 0.000 & 0.024 + [ tab : gcorr ] @crcr@ coefficient & & coefficient & + ( @xmath0 ) & & ( @xmath1 ) & + a@xmath54 & @xmath177.3801851e+00 & b@xmath54 & 5.1729169e@xmath1702 + a@xmath55 & 6.0978389e@xmath1701 & b@xmath55 & @xmath172.8514123e+00 + a@xmath56 & @xmath178.5916775e@xmath1701 & b@xmath56 & @xmath172.8616686e+00 + a@xmath57 & 7.2038240e+00 & b@xmath57 & 3.2487645e+00 + a@xmath58 & @xmath171.1195542e@xmath1701 & b@xmath58 & @xmath177.3122419e@xmath1702 + a@xmath59 & @xmath178.2204863e@xmath1703 & b@xmath59 & @xmath173.5157643e@xmath1702 + a@xmath60 & 1.0550418e@xmath1701 & b@xmath60 & 2.0585494e+00 + a@xmath61 & 2.2982103e@xmath1703 & b@xmath61 & 2.9195288e@xmath1701 + a@xmath62 & @xmath172.3313912e@xmath1702 & b@xmath62 & @xmath179.2196599e@xmath1702 + a@xmath63 & @xmath171.5513621e@xmath1701 & b@xmath63 & @xmath173.2042870e@xmath1701 + a@xmath64 & 6.5429395e@xmath1701 & b@xmath64 & @xmath176.3154984e@xmath1701 + a@xmath65 & @xmath171.1711317e+00 & b@xmath65 & 8.3885527e@xmath1701 + a@xmath66 & 2.8245633e@xmath1703 & b@xmath66 & @xmath171.0552187e@xmath1704 + [ tab : coeff ]
@xmath67 @lr@ @xmath67 @lcr@ @xmath67 @rr@ @xmath67 @lcr@ @xmath67 @lr@ @xmath67 @lcc@ object & & & @xmath68 & & & @xmath69 & & & @xmath70 & @xmath19 h + & & & ( m@xmath7 ) & & & ( gyr ) & & & ( gyr ) & + j0112 + 1835 & 9740 & 140 & 5.77 & 0.05 & 0.160 & 662 & 98 & 8.63 & 1.07 & 1.8 & 0.62 & 0.03 & 0.85 & 0.04 & 2.68 & @xmath1722.40 + j0152 + 0749 & 10800 & 190 & 5.93 & 0.05 & 0.168 & 980 & 131 & 7.36 & 0.83 & 1.4 & 0.57 & 0.03 & 0.78 & 0.04 & & @xmath1722.80 + j0345 + 1748 & 8560 & 120 & 6.51 & 0.04 & 0.181 & 175 & 34 & 3.91 & 0.41 & 3.9 & 0.76 & 0.03 & & 7.18 & @xmath1722.16 + j0745 + 1949 & 8230 & 130 & 5.91 & 0.07 & 0.156 & 300 & 56 & 7.28 & 1.07 & 3.6 & 0.10 & 0.01 & 0.12 & 0.02 & 5.80 & @xmath1722.65 + j0818 + 3536 & 10190 & 190 & 5.87 & 0.09 & 0.164 & 3388 & 640 & 7.76 & 1.26 & 1.7 & 0.25 & 0.04 & 0.33 & 0.06 & 9.30 & @xmath1723.47 + j0822 + 2753 & 8800 & 130 & 6.36 & 0.05 & 0.175 & 496 & 72 & 4.57 & 0.51 & 4.2 & 0.76 & 0.08 & 1.05 & 0.11 & 8.18 & @xmath1722.31 + j0849 + 0445 & 10020 & 150 & 6.27 & 0.06 & 0.176 & 1101 & 158 & 5.08 & 0.61 & 1.9 & 0.65 & 0.04 & 0.89 & 0.05 & 0.45 & @xmath1722.38 + j0900 + 0234 & 8330 & 130 & 5.96 & 0.07 & 0.155 & 618 & 113 & 6.83 & 0.99 & 3.8 & & & &
+ j1005 + 3550 & 9760 & 140 & 6.00 & 0.05 & 0.166 & 1231 & 392 & 6.72 & 0.79 & 2.1 & 0.19 & 0.02 & 0.24 & 0.02 & 10.63 & @xmath1723.12 + j1112 + 1117 & 9240 & 140 & 6.17 & 0.06 & 0.169 & 266 & 42 & 5.57 & 0.70 & 2.7 & 0.14 & 0.01 & 0.17 & 0.02 & 12.73 & @xmath1722.56 + j1443 + 1509 & 8970 & 130 & 6.44 & 0.06 & 0.181 & 557 & 85 & 4.24 & 0.51 & 3.5 & 0.84 & 0.12 & 1.17 & 0.17 & 3.82 & @xmath1722.25 + j1512 + 2615 & 11250 & 180 & 6.93 & 0.06 & 0.250 & 766 & 96 & 2.85 & 0.31 & 0.4 & 0.31 & 0.04 & 0.39 & 0.05 & & @xmath1722.94 + j1518 + 0658 & 9650 & 140 & 6.68 & 0.05 & 0.197 & 309 & 41 & 3.38 & 0.35 & 0.9 & 0.58 & 0.03 & 0.78 & 0.04 & & @xmath1722.42 + j1538 + 0252 & 10030 & 150 & 5.98 & 0.06 & 0.167 & 1154 & 180 & 6.90 & 0.92 & 1.9 & 0.76 & 0.06 & 1.06 & 0.09 & & @xmath1722.86 + j1557 + 2823 & 12470 & 200 & 7.76 & 0.05 & 0.448 & 191 & 18 & 1.45 & 0.11 & 0.4 & 0.41 & 0.03 & 0.51 & 0.04 & & @xmath1721.92 + j1614 + 1912 & 8700 & 170 & 6.32 & 0.13 & 0.172 & 207 & 54 & 4.73 & 1.00 & 4.1 & & & & + j1630 + 2712 & 10200 & 170 & 6.03 & 0.08 & 0.169 & 2161 & 374 & 6.56 & 0.99 & 1.8 & 0.52 & 0.04 & 0.70 & 0.06 & & @xmath1723.13 + j1741 + 6526 & 10410 & 170 & 6.02 & 0.06 & 0.170 & 998 & 150 & 6.71 & 0.87 & 1.7 & 1.11 & 0.05 & 1.57 & 0.06 & 0.16 & @xmath1722.11 + j1840 + 6423 & 9120 & 140 & 6.34 & 0.05 & 0.177 & 740 & 109 & 4.70 & 0.54 & 3.3 & 0.65 & 0.03 & 0.88 & 0.04 & 4.78 & @xmath1722.47 + j2103@xmath170027 & 9900 & 150 & 5.79 & 0.05 & 0.161 & 1253 & 177 & 8.47 & 1.01 & 1.8 & 0.71 & 0.04 & 0.98 & 0.05 & 5.73 & @xmath1722.72 + j2119@xmath170018 & 9690 & 150 & 5.72 & 0.08 & 0.158 & 2835 & 507 & 9.09 & 1.38 & 1.7 & 0.74 & 0.04 & 1.03 & 0.05 & 0.58 & @xmath1722.82 + j2228 + 3623 & 7890 & 120 & 5.78 & 0.08 & 0.175 & 420 & 79 & 8.96 & 1.34 & 7.7 & & & & + j2236 + 2232 & 11310 & 170 & 6.54 & 0.04 & 0.182 & 355 & 43 & 3.81 & 0.40 & 1.1 & 0.39 & 0.03 & 0.51 & 0.04 & & @xmath1722.79 + [ tab : elm ] @xcite present a 1d spectroscopic analysis of 61 low - mass wds identified in the elm survey .
all but five of these objects are in confirmed binary systems with periods ranging from 12 min to about a day .
the majority of these targets are warmer than 12,000 k , where the 3d corrections are negligible . however , there are 23 convective wds that are in the right @xmath0 and @xmath71 range for 3d corrections to be significant .
table [ tab : elm ] presents the updated parameters for these 23 systems , including temperature , surface gravity , mass , radius , cooling age , merger time , and their gravitational wave strain .
the corrections are largest at around 8000 k for low - mass wds .
hence , the parameters for relatively cool objects like j0345 + 1748 ( nltt 11748 ) have changed significantly . since the 3d corrections lower the estimated @xmath71
, they also imply lower masses and gravitational wave strains for these systems .
1d spectroscopic analysis of @xcite find @xmath72 k and @xmath73 for nltt 11748 .
however , this was based on older wd models without the improved stark broadening profiles of @xcite . ] .
the 1d to 3d corrections from our models revise these to @xmath74 k and @xmath75 , a change of 0.3 dex in surface gravity . using the @xcite evolutionary models , the revised mass , distance , and radius are @xmath76 , @xmath77 pc , and @xmath78 , respectively .
our distance estimate is in excellent agreement with the parallax measured at the usno , @xmath79 mas .
@xcite use high - quality photometry of nltt 11748 to constrain the same parameters from the observed eclipses .
depending on the thickness of the surface hydrogen layer of the secondary ( more massive ) wd , they derive @xmath80 and @xmath81 .
our mass and radius measurements , including the 3d correction factors , are consistent with their results within the errors .
there are several tidally distorted and/or eclipsing elm wd systems that provide model - independent constraints on the radii .
nltt 11748 is discussed above ; it is an eclipsing system where the spectroscopically inferred radius from the 1d models were significantly smaller than the eclipse modelling suggested @xcite .
studying the radius constraints on the elm wd sample , @xcite find that the agreement between the spectroscopically inferred radius and the model - independent values is quite good for @xmath41 k. however , there are four relatively cool objects , lp 400 - 22 , nltt 11748 , j0745 , and j0751 , where the radius is underestimated or the mass/@xmath71 is overestimated .
figure [ fg : radius ] shows the comparison of the updated radius measurements , including the 3d corrections , for the same 10 stars from @xcite .
table [ tab : mr ] provides the references for the different radius measurements .
the agreement between the spectroscopic and photometric radii is now excellent over the temperature range 8,00017,000 k. there are only two deviant points , lp 400 - 22 and j0745 , and we refer the reader to @xcite and @xcite , respectively , for further discussion of these objects . ignoring lp 400 - 22 and j0745 , the excellent agreement between the radius measurements from these two independent methods suggests that the 3d model correction factors presented in this paper have the right amplitude to account for the differences seen in the earlier analysis relying on 1d models .
finally , we note that this independent verification of 3d results is unique to elm wds and fairly different to the one presented in @xcite for single c / o - core wds with trigonometric parallax measurements .
@crrrc@ object & & & & ref .
+ & & & & + j0056@xmath170611 & 12230 @xmath67 180 & 5.65 @xmath67 0.61 & 5.6 @xmath67 0.6 & 1 + j0106@xmath171000 & 16970 @xmath67 260 & 6.42 @xmath67 0.68 & 6.3 @xmath67 0.8 & 1 + j0112@xmath821835 & 9740 @xmath67 140 & 8.64 @xmath67 1.07 & 8.8 @xmath67 0.9 & 1 + j0345@xmath821748 & 8560 @xmath67 120 & 3.91 @xmath67 0.41 & 4.23 @xmath67 0.10 & 2 + j0651@xmath822844 & 16340 @xmath67 260 & 3.25 @xmath67 0.31 & 4.0 @xmath67 0.2 & 1 + j0745@xmath821949 & 8320 @xmath67 130 & 7.28
@xmath67 1.07 & 17.6@xmath83 & 3 + j0751@xmath170141 & 15740 @xmath67 250 & 13.15 @xmath67 1.04 & 13.8@xmath84 & 1 + j1741@xmath826526 & 10410 @xmath67 170 & 6.71 @xmath67 0.87 & 7.6 @xmath67 0.6 & 1 + j2119@xmath170018 & 9690 @xmath67 150 & 9.09
@xmath67 1.38 & 10.3 @xmath67 1.6 & 1 + j2236@xmath822232 & 11310 @xmath67 170 & 3.81 @xmath67 0.40 & @xmath859.9 & 4 + [ tab : mr ] there are several wd companions to milli - second pulsars with optical spectroscopy available in the literature .
these wds provide an independent method to estimate the mass of the neutron star through constraints on the mass of the wd and the semi - amplitude of the radial velocity variations .
hence , accurate mass ( and @xmath71 ) measurements for the elm wds in these systems are crucial .
many of these wds have temperatures in the range 8000 - 10,000 k , where 3d effects are significant .
psr j0218 + 4232 , j1012 + 5307 , and j1909@xmath173744 have companions with @xmath86 k and @xmath87 , based on 1d spectroscopic analyses @xcite .
the corrections in @xmath71 are on the order of 0.3 dex at these temperatures . for example
, @xcite derive @xmath88 k and @xmath89 for the elm wd companion to psr j1012 + 5307 .
taking the 3d corrections into account brings down these values to @xmath90 k and @xmath91 .
psr 1911@xmath175958a has a wd companion with @xmath92 k and @xmath93 @xcite .
the amplitude of the 3d corrections is lower at these temperatures , bringing down the temperature and surface gravity to 9780 k and @xmath94 , respectively .
psr j1738 + 0333 is an excellent system where the mass , radius , and the surface gravity of the wd companion can be estimated from both optical spectroscopy / photometry and radio observations . @xcite
obtain @xmath95 k , @xmath96 , @xmath97 , and @xmath98 from optical spectroscopy .
interestingly , they also derive @xmath99 ( from the orbital period decay rate of the binary , mass ratio , parallax , and photometry ) , @xmath100 ( from the mass ratio and orbital decay rate of the binary ) , and @xmath101 ( from photometry ) .
the latter measurements are independent of the 1d spectroscopic analysis and imply a lower surface gravity and a larger radius for this elm wd .
these provide additional evidence for the `` high @xmath71 problem '' in 1d spectroscopic analysis of elm wds , though ( admittedly ) the errors in surface gravity and radius are relatively large . based on the 3d corrections factors that we estimate ,
we revise the spectroscopic parameters of this companion to @xmath102 k and @xmath103 , which are consistent ( within the errors ) with @xmath99 estimated using the independent method .
we note that this wd companion also happens to be a pulsating wd @xcite .
figure [ fg : zz ] shows the instability strip for the pulsating dav wds based on 1d and 3d spectroscopic analyses .
here we only show the parameters for the davs derived from a uniform analysis by @xcite .
we also show the five elm pulsators currently known ( excluding the companion to psr j1738 + 0333 since the 1d model spectroscopic analysis of that star was not performed using the same models ) , and a large number of photometrically constant stars that help define the boundaries of the instability strip .
the solid lines show the boundaries of the instability strip from gianninas et al .
( 2014 ) .
there is a clear trend of increasing @xmath71 with decreasing temperature in the 1d spectroscopic parameters of these stars .
this trend disappears in the 3d version of the plot , where the 3d corrections lower the implied surface gravity .
the blue edge of the instability strip is well defined in the 1d analysis .
the 3d corrections move the stars systematically to slightly cooler temperatures and lower surface gravities .
the dashed lines show the revised boundaries based on the 3d corrected parameters . for the blue edge ,
the boundary is simply shifted to the right by 100 k. the blue edge of the instability strip is still well defined for both normal mass and elm wds in the 3d analysis . for the red edge , as in the 1d case , we use gd 518 and sdss j222859.93 + 362359.6 as the two anchor points at either extremity ( the highest and lowest surface gravity ) to define the 3d red edge , which now has a higher slope .
the equations of the revised blue and red edges are defined by equations [ eq : blue ] and [ eq : red ] , respectively .
given the small number of elm pulsators currently known , it is too early to definitively constrain the red edge of the instability strip for those stars .
the new zz ceti instability strip from the 3d analysis does not reveal any new elm pulsator candidates in the published list of elm wds , but the revised boundaries should be used in the future to look for new pulsating da wds .
@xmath104 @xmath105
we provide mean 3d model spectra for wds with @xmath106 , extending the previously published 3d model grid to very low - mass wds .
our 3d models now cover the entire temperature and surface gravity range for the observed population of convective da wds .
we list 1d to 3d correction factors , which can be as large as 0.35 dex in surface gravity .
these corrections largely resolve the discrepancies seen in the @xmath71 and radius estimates for relatively cool elm wds that are observed in double wd or wd + milli - second pulsar systems .
we also update the boundaries of the zz ceti instability strip for wds , including the elm pulsators as well as the normal ( @xmath107 ) wds .
support for this work was provided by nasa through hubble fellowship grant # hf-51329.01 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
mk and ag gratefully acknowledge the support of the nsf and nasa under grants ast-1312678 and nnx14af65 g .
ag acknowledges support provided by nasa through grant number hst - go-13319.01 from the space telescope science institute , which is operated by aura , inc .
, under nasa contract nas 5 - 26555 .
the computing for this project was performed at the ou supercomputing center for education & research ( oscer ) at the university of oklahoma ( ou ) .
acknowledges financial support by the sonderforschungsbereich sfb 881 `` the milky way system '' ( subproject a4 ) of the german research foundation ( dfg ) .
j.j.h . acknowledges funding from the european research council under the european union s seventh framework programme ( fp/2007 - 2013 ) / erc grant agreement n. 320964 ( wd - tracer ) .
allard , n. f. , kielkopf , j. f. , & loeillet , b. 2004 , , 424 , 347 althaus , l. g. , miller bertolami , m. m. , & crsico , a. h. 2013 , , 557 , aa19 antoniadis , j. , van kerkwijk , m. h. , koester , d. , et al .
2012 , , 423 , 3316 asplund , m. 2005 , , 43 , 481 asplund , m. , grevesse , n. , sauval , a. j. , & scott , p. 2009
, , 47 , 481 bassa , c. g. , van kerkwijk , m. h. , & kulkarni , s. r. 2003 , , 403 , 1067 bassa , c. g. , van kerkwijk , m. h. , koester , d. , & verbunt , f. 2006 , , 456 , 295 bergeron , p. , wesemael , f. , fontaine , g. , & liebert , j. 1990 , , 351 , l21 bhm - vitense , e. 1958 , , 46 , 108 brown , w. r. , kilic , m. , allende prieto , c. , & kenyon , s. j. 2010 , , 723 , 1072 brown , w. r. , kilic , m. , hermes , j. j. , et al .
2011 , , 737 , ll23 caffau , e. , ludwig , h .-
g . , steffen , m. , freytag , b. , & bonifacio , p. 2011
, , 268 , 255 callanan , p. j. , garnavich , p. m. , & koester , d. 1998 , , 298 , 207 eisenstein , d. j. , liebert , j. , harris , h. c. , et al .
2006 , , 167 , 40 freytag , b. , steffen , m. , ludwig , h .-
, et al .
2012 , journal of computation al physics , 231 , 919 genest - beaulieu , c. , & bergeron , p. 2014
, , 796 , 128 gianninas , a. , bergeron , p. , & ruiz , m. t. 2011 , , 743 , 138 gianninas , a. , dufour , p. , kilic , m. , et al . 2014a , , 794 , 35 gianninas , a. , hermes , j. j. , brown , w. r. , et al .
2014b , , 781 , 104 grevesse , n. , & sauval , a. j. 1998 , , 85 , 161 hermes , j. j. , montgomery , m. h. , winget , d. e. , et al .
2013 , , 765 , 102 hermes , j. j. , brown , w. r. , kilic , m. , et al .
2014 , , 792 , 39 hummer , d. g. , & mihalas , d. 1988 , , 331 , 794 jacoby , b. a. , hotan , a. , bailes , m. , ord , s. , & kulkarni , s. r. 2005 , , 629 , l113 kaplan , d. l. , marsh , t. r. , walker , a. n. , et al .
2014 , , 780 , 167 kepler , s. o. , kleinman , s. j. , nitta , a. , et al .
2007 , , 375 , 1315 kilic , m. , brown , w. r. , allende prieto , c. , kenyon , s. j. , & panei , j. a. 2010a , , 716 , 122 kilic , m. , allende prieto , c. , brown , w. r. , et al . 2010b , , 721 , l158 kilic , m. , gianninas , a. , brown , w. r. , et al .
2013 , , 434 , 3582 kilic , m. , brown , w. r. , gianninas , a. , et al .
2014 , , 444 , l1 kilic , m. , hermes , j. j. , gianninas , a. , & brown , w. r. 2015 , , 446 , l26 kleinman , s. j. , kepler , s. o. , koester , d. , et al .
2013 , , 204 , 5 ludwig , h .-
, & steffen , m. 2008 , in precision spectroscopy in astrophysics , ed .
santos , l. pasquini , a.c.m .
correia , & m. romanielleo ( berlin : springer ) , 133 ludwig , h .- g . ,
behara , n. t. , steffen , m. , & bonifacio , p. 2009a , , 502 , l1 ludwig , h .-
, caffau , e. , steffen , m. , et al . 2009b , , 80 , 711 scott , p. , grevesse , n. , asplund , m. , et al .
2015a , , 573 , a25 scott , p. , asplund , m. , grevesse , n. , bergemann , m. , & sauval , a. j. 2015b , , 573 , a26 spite , m. , andrievsky , s. m. , spite , f. , et al .
2012 , , 541 , a143 steffen , m. , ludwig , h .-
& freytag , b. 1995 , , 300 , 473 tremblay , p .- e .
, & bergeron , p. 2009
, , 696 , 1755 tremblay , p .- e . , bergeron , p. , kalirai , j. s. & gianninas , a. 2010 , , 712 , 1345 tremblay , p .- e . , ludwig , h .-
g . , steffen , m. , bergeron , p. , & freytag , b. 2011 , , 531 , l19 tremblay , p .- e . , ludwig , h .-
g . , steffen , m. , & freytag , b. 2013a , , 552 , a13 tremblay , p .- e . , ludwig , h .-
g . , freytag , b. , steffen , m. , & caffau , e. 2013b , , 557 , a7 tremblay , p .- e . , ludwig , h .-
, steffen , m. , & freytag , b. 2013c , , 559 , a104 van kerkwijk , m. h. , bergeron , p. , & kulkarni , s. r. 1996 , , 467 , l89
.... function ml18_to_3d_dteff_elm(teff , logg ) real*8 ml18_to_3d_dteff_elm , a(13),teff , logg real*8 tx , gx , shift c ; c ; in : teff , effective temperature ( k ) c ; logg , surface gravity ( g in cm / s2 ) c ; c ; out : teff correction ( k ) 1d ml2/alpha = 0.8 to 3d c ; a(1)= -7.3801851e+00 a(2)= 6.0978389e-01 a(3)= -8.5916775e-01 a(4)= 7.2038240e+00 a(5)= -1.1195542e-01 a(6)= -8.2204863e-03 a(7)= 1.0550418e-01 a(8)= 2.2982103e-03 a(9)= -2.3313912e-02 a(10)=-1.5513621e-01 a(11)= 6.5429395e-01 a(12)=-1.1711317e+00 a(13)= 2.8245633e-03 tx=(teff-10000.0)/1000.00 gx=(logg-8.0 ) shift = a(13)+(a(1)+a(4)*exp((a(5)+a(7)*exp(a(8 ) * * tx+a(9)*gx))*tx+a(6)*gx))*exp((a(2 ) * + a(10)*exp(a(11)*tx+a(12)*gx))*tx+a(3)*gx ) .... function ml18_to_3d_dlogg_elm(teff , logg ) real*8 ml18_to_3d_dlogg_elm , b(13),teff , logg real*8 tx , gx , shift c ; c ; in : teff , effective temperature ( k ) c ; logg , surface gravity ( g in cm / s2 ) c ; c ; out : log g correction ( g in cm / s2 ) 1d ml2/alpha = 0.8 to 3d c ; b(1)= 5.1729169e-02 b(2)= -2.8514123e+00 b(3)= -2.8616686e+00 b(4)= 3.2487645e+00 b(5)= -7.3122419e-02 b(6)= -3.5157643e-02 b(7)= 2.0585494e+00 b(8)= 2.9195288e-01 b(9)= -9.2196599e-02 b(10)=-3.2042870e-01 b(11)=-6.3154984e-01 b(12)= 8.3885527e-01 b(13)=-1.0552187e-04 tx=(teff-10000.0)/1000.0 gx=(logg-8.0 ) shift = b(13)+(b(1)+b(10)*exp(b(11)*tx+b(12)*gx ) ) * * exp((b(2)+b(4)*exp(b(5)*tx+b(6)*gx ) ) * * tx+(b(3)+b(7)*exp(b(8)*tx+b(9)*gx))*gx )
|
we present an extended grid of mean three - dimensional ( 3d ) spectra for low - mass , pure - hydrogen atmosphere da white dwarfs ( wds ) .
we use co5bold radiation - hydrodynamics 3d simulations covering @xmath0 = 600011,500 k and @xmath1 = 56.5 ( @xmath2 in cm s@xmath3 ) to derive analytical functions to convert spectroscopically determined 1d temperatures and surface gravities to 3d atmospheric parameters . along with the previously published 3d models ,
the 1d to 3d corrections are now available for essentially all known convective da wds ( i.e. , @xmath1 = 59 ) . for low - mass wds , the correction in temperature is relatively small ( a few per cent at the most ) , but the surface gravities measured from the 3d models are lower by as much as 0.35 dex .
we revisit the spectroscopic analysis of the extremely low - mass ( elm ) wds , and demonstrate that the 3d models largely resolve the discrepancies seen in the radius and mass measurements for relatively cool elm wds in eclipsing double wd and wd + milli - second pulsar binary systems .
we also use the 3d corrections to revise the boundaries of the zz ceti instability strip , including the recently found elm pulsators .
|
potential energy functions play a central role in chemistry , and from these functions , the properties of a system can , in principle , be determined . often , the potential energy function is obtained from _ ab initio _ methods , but it can also be determined from experimental data using inverse techniques .
inverse problems theory to handle experimental data has been applied to refine potential energy functions from the second virial coefficient,@xcite differential cross - sections@xcite and phase shifts.@xcite the helium diatomic system at low temperatures has received considerable attention in recent years due to quantum and relativistic effects in their thermodynamics and transport properties.@xcite the theoretical interpretation of the results has been performed with the _ ab initio _
potential function.@xcite the helium dimer has a very small potential well depth , measuring approximately 11 k,@xcite consequently , highly accurate calculations are required to obtain the potential energy function.@xcite the present paper presents an alternative way to refine the helium potential energy function from viscosity coefficient data at low temperatures .
few studies have been conducted along this line and often at high temperatures,@xcite limit in which classical theory is valid , and in parametric form , usually for a lennard - jones potential energy curve.@xcite to gain insight into this problem , an investigation into the sensitivity@xcite of the viscosity coefficient to the potential energy function is conducted , and an adequate temperature range for experimental data was observed .
the present study shows a higher sensitivity for long range potential in the temperature range below 5 k. since it is necessary to use viscosity coefficient data at low temperatures in an inverse procedure , a quantum strategy to refine the potential energy curve is necessary and will be presented in this study .
the inverse problem was divided into two parts : in the first part , the cross - section is obtained from the viscosity coefficient , whereas in the second part , the potential energy function is determined from the cross - section .
the solution to the first part was determined using the tikhonov regularization under a laplace integral equation formulation .
in the second part , the relationship between the potential energy and the total cross - section , which is nonlinear , was linearized by a sensitivity analysis algorithm.@xcite the cross - section sensitivity matrix established has an approximate linear relation between the potential energy and total cross - section . in a previous study,@xcite the functional derivative of the quantum phase shift with respect to the potential energy function
was established and coupled with the variable phase equation .
for the first time , this set of coupled differential equations was used to establish the cross - section sensitivity matrix within exact quantum theory .
this problem is ill - posed , since sensitivity matrix elements possess values that are close to zero . in an attempt to circumvent this problem in a tractable way
, the tikhonov regularization procedure was used again . the result obtained by the inverse technique
was compared with recent _ ab initio _ potentials@xcite and exhibited excellent agreement .
the present work presents a general strategy to obtain an accurate inverted potential , comparable in quality with high level potential energy models and calculations .
chapman - enskog theory@xcite provides a well - established perturbation method to calculate the viscosity coefficient @xmath2 of a non - ideal gas from the collision integral @xmath3 in a first order approach , the viscosity coefficient is given by @xmath4 in which @xmath5 is the mass of the system , and @xmath6 is boltzmann s constant .
the collision integral of interest to determine the viscosity coefficient is set in the form @xmath7 with @xmath8 the temperature and @xmath9 the collision energy .
the transport cross - section @xmath10 in classical statistical mechanics is directly related to the scattering angle , but in quantum assumptions and for the bose - einstein statistic , the cross - section is associated with the phase shift,@xcite as follows : @xmath11 \label{eqq}\ ] ] with @xmath12 the wave number , @xmath13 planck s constant and @xmath14 the angular moment .
the phase shift @xmath15 is obtained from calogero s equation , @xmath16 \label{eqc}\ ] ] when @xmath17 in which @xmath18 and @xmath19 is the interatomic distance and @xmath20 is the system reduced mass .
more details about the calogero equation can be found in the published works of lemes et al.,@xcite viterbo2014 et al.,@xcite and braga and murrell.@xcite the viscosity coefficient as a function of the temperature is determined by solving these equations , obeying the sequence ( [ eqc ] ) , ( [ eqq ] ) , ( [ eqo ] ) and ( [ visco ] ) . in another approach ,
it is desired to obtain the potential energy function from the viscosity coefficient in the opposite direction of the direct problem , that is , in the sequence ( [ visco ] ) , ( [ eqo ] ) , ( [ eqq ] ) and ( [ eqc ] ) .
the first step consists of computing the collision integral @xmath21 from viscosity coefficient data using equation ( [ visco ] ) through simple algebraic manipulation .
the second step consists of the evaluation of the transport cross - section @xmath22 from the collision integral @xmath23 and these quantities are related by equation ( [ eqo ] ) .
the bold style was used to represent functions in matrix form . using the trapezoidal quadrature , equation ( [ eqo ] )
is transformed to an algebraic form @xmath24 the solution of which is determined by the tikhonov regularization method as @xmath25 in which @xmath26 is a first approximation for @xmath27 with @xmath22 found , it is possible to obtain the potential energy @xmath28 from the cross - section through equations ( [ eqq ] ) and ( [ eqc ] ) . the main difficulty of this strategy is to obtain the @xmath28 value from @xmath29 due to the nonlinear relationship between these two quantities .
this difficulty can be avoided by linearization of the problem as follows : @xmath30 in which the matrix @xmath31 is the sensitivity matrix , with elements @xmath32 nevertheless , it is more appropriate to use equation ( [ eqlinear ] ) as @xmath33 with elements of @xmath34 in a normalized form , such as @xmath35 here , the adequate solution can not be found by an orthodox inverse matrix algorithm , because the matrix @xmath34 is ill conditioned .
the inverse matrix will be given again by another tikhonov regularization @xmath36 in which @xmath37 is a parameter of regularization and is the identity matrix .
finally , the desired solution is given by @xmath38 in which @xmath39 is an initial guess for @xmath40 in this case , the regularization parameter is chosen by an l - curve , and provides an equilibrium between a residual norm and a solution norm .
the success of the present method relies on identifying a way to determine the sensitivity matrix . in a previous work,@xcite a differential equation for the functional derivative of the quantum phase shift with respect to the potential energy function , @xmath41 ,
is established and coupled to calogero s equation ( [ eqc ] ) , @xmath42 \end{array } \label{eqsp}\ ] ] in which @xmath43 is given by @xmath44 therefore , the value of @xmath45 at fixed values of @xmath46 and @xmath47 can be determined from coupled equations ( [ eqsp ] ) , at @xmath48 for different angular moments .
more details about this approach are given in the work of lemes et al.@xcite by deriving equation ( [ eqq ] ) with respect to the potential energy function at @xmath47 the sensitivity cross - section @xmath49 will be given by @xmath50\\ \times \left[s_{\delta_{l+2}}(\kappa , r^*)-s_{\delta_l}(\kappa , r^*)\right]\ ] ] where all of the necessary information is available from the previous step .
this method provides a new , simple and exact way to establish the sensitivity cross - section matrix @xmath34 english in equation ( [ tikep ] ) .
finally , the sensitivity matrix for the viscosity coefficient can be obtained by @xmath51 in which @xmath52 there is no need to determine matrix ( [ senseeta ] ) in our method , but this matrix suggests a new procedure to recover the potential energy function from viscosity coefficient data , such as using @xmath53 instead of equation ( [ eqlinear ] ) .
nevertheless , preliminary studies show that the condition number of matrix @xmath54 is larger than the condition number of matrix @xmath55 making this method more difficult than the procedure proposed in this study .
figure [ fig1](a ) shows the contour lines of the normalized sensitivity matrix to the transport cross - section data , @xmath56 this result provides important insight into the inverse procedure , such as the choice of the experimental data range to be used .
* figure 1 * from figure [ fig1](a ) , it can be observed that the transport cross - section with @xmath46 between 0.2 and 0.5 @xmath57 has a larger value of sensitivity at interatomic distances between 3 and 4 .
this region of @xmath46 provides experimental data that are more adequate to obtain the long range potential between 3 and 4 because small changes in the potential can cause large changes in the transport cross - section .
since the @xmath46 values of approximately 0.1 @xmath57 correspond to energy values of approximately 0.1 k , it is better to use viscosity experimental data obtained at low temperatures for the inverse procedure .
this conclusion can also be obtained from figure [ fig1](b ) , showing the sensitivity level curves for viscosity data .
only a small number of papers have reported experimental data on viscosity coefficients at low temperatures .
for the temperature range required to refine the potential curve between 3 and 4 , that is , below 5 k , the reference of becker et al.@xcite presents 11 values .
however , the quality of these data has been questioned@xcite due to the use of an old reference value for equipment calibration , resulting in a positive deviation of approximately 5 % as estimated by bich and vogel.@xcite to generate the necessary data of viscosity coefficients , the equation @xmath58\right\ } , \label{nachereq}\ ] ] was used.@xcite from this equation , 41 viscosity values , between 1 and 5 k , were determined and used to refine the potential curve .
the values adjusted by the equation ( [ nachereq ] ) , shown in figure [ fig2 ] , agreed with the calculated results with the most precise potential for description of the system,@xcite with an average error of 0.8735 % , an error lower than that reported for the experimental data ( 5 % ) .
therefore , henceforth we will refer to equation ( [ nachereq ] ) as experimental data .
* figure 2 * in our inversion procedure , an initial potential curve , @xmath59 is given as _ a priori _ information .
this approximate potential curve is given by @xmath60 } \right ] } \right\},{r_1 } < r < { r_2}}\\ { e_{\rm p}(r),r \ge { r_2 } } \end{array } } \right.\ ] ] as suggested in the reference of keil and danielson,@xcite in which @xmath61 @xmath62 and @xmath63 were used to modify the potential curve @xmath28 between 3 and 4 .
the function @xmath28 is the _ ab initio _ potential curve proposed in 2010,@xcite henceforth considered to be the reference potential .
the @xmath28 curve was previously tested to evaluate the second virial coefficient between 3 and 100 k with excellent agreement with the experimental data.@xcite when this potential is compared with the most accurate potential describing the system,@xcite the difference is less than @xmath64 between 2.4 and 4.5 . nevertheless , both curves , varandas@xcite and przybytek et al.@xcite , describe the viscosity coefficient with an error less than 5 % for temperatures between 1 and 5 k. therefore , both _
ab initio _ potential curves , namely , those of varandas@xcite and przybyteke et al.@xcite , are equally suitable to describe the viscosity coefficient between 1 and 5 k. the difference between the potential curves @xmath65 and @xmath66 is controlled by the parameter @xmath67 when @xmath68 is equal to 0.2 the average difference is approximately 10 % . in this case , if the potential curve @xmath65 is used to calculate the viscosity data the average error found was 12.46 % , which is greater than the experimental error of 5 % .
therefore , the calculated viscosity data from the potential curve @xmath65 are not within the experimental error , and it is not suitable for estimating the viscosisty coefficient .
the determination of the viscosity coefficient from a potential energy curve involves three steps : the determination of the phase shift using equation ( [ eqc ] ) ; followed by the determination of the transport cross - section using equation ( [ eqq ] ) ; and finally the determination of the collision integral using equation ( [ eqo ] ) from which the viscosity can be readily calculated .
numerical integration of calogero s equation ( [ eqc ] ) was performed using the euler algorithm , with a step of @xmath69 between 1.5 and 70000 .
next , in equation ( [ eqq ] ) , the sum is carried over even angular moments between 0 and 20 .
finally , the trapezoidal rule was used to numerically integrate equation ( [ eqo ] ) over 100 points between @xmath70 and 5.4 mev .
figure [ fig2 ] shows the experimental data@xcite together with the values calculated by equations ( [ eqc ] ) , ( [ eqq ] ) , ( [ eqo ] ) and ( [ visco ] ) , using the potential function @xmath71 the inverse problem consists of the determination of @xmath28 using known viscosity data , @xmath72 therefore , one has to solve the set of four equations in the reverse sequence : ( [ visco ] ) , ( [ eqo ] ) , ( [ eqq ] ) and ( [ eqc ] ) . the first step provides the values of the transport cross - section , which are shown in figure [ fig3 ] .
this step involves equation ( [ eqc ] ) , which is a linear ill - posed problem , whose solution is determined by the tikhonov regularization method , shown in equation ( [ tikq ] ) .
* figure 3 * the result shown in figure [ fig3 ] was obtained with @xmath73 and using a matrix * k * with dimensions of 41 ( temperature ) @xmath74 101 ( energy ) .
the value of @xmath26 is obtained from @xmath65 following the direct method . in the second part of the inverse procedure ,
the equations ( [ eqq ] ) and ( [ eqc ] ) are rewritten as ( [ lnq ] ) , which is another linear ill - posed problem , the solution of which is provided by equation ( [ tikep ] ) . before using equation ( [ tikep ] ) to refine the potential energy function , one needs to determine the matrix * s * from equations ( [ eqsp ] ) and ( [ senseq ] ) .
the coupled differential equations ( [ eqsp ] ) were solved using euler s method with initial conditions @xmath75 @xmath76 and @xmath77 . in the limit
@xmath17 @xmath45 is the desired sensitivity value at fixed @xmath78 @xmath79 and @xmath80 therefore , solving the coupled differential equations ( [ tikep ] ) , for different values of @xmath47 @xmath46 and @xmath14 , we obtain the @xmath81 values from which , using equation ( [ eqg ] ) , we calculated @xmath82 and consequently @xmath83 now , knowing @xmath84 equation ( [ tikep ] ) can be used in the following manner : first , an initial guess , @xmath85 is given for the unknown potential , and the value of @xmath26 is computed for this first potential estimation .
the difference , @xmath86 is then calculated between the @xmath87 obtained from the first part and @xmath88 a correction to the initial potential energy function can be evaluated interactively by using equation ( [ tikep ] ) . the refined potential @xmath89 minimizes the tikhonov criterion function , for which the balance between the residual norm and the solution norm is given by @xmath90 the potential curve @xmath91 obtained after three iterations , is shown in figure [ fig4 ] together with the reference potential @xmath66 and the initial choice @xmath71 the refined potential @xmath92 has an average error of 2.6 % between 2.4 and 4.5 compared with the _ ab initio _
potential curve.@xcite the potential depth is -11.1 k for the potential obtained in this work , while for the reference potential , the depth is -11.0 k ; therefore , an error of less than 1 % is observed .
the equilibrium distance is 2.96 for both potentials * figure 4 * the viscosity coefficient was determined using refined potential @xmath91 this result together with experimental data are shown in figure [ fig2 ] , for temperatures between 1 and 5 k. the improved potential is adequate to describe the experimental data with an average error of 1.6823 % against 12.4632 % when @xmath65 was used .
the average error found with the @xmath92 potential is less than the experimental error ( 5 % ) , therefore the inverted results provided a better potential than the initial curve @xmath71 the potential @xmath91 obtained directly from the experimental data , had equivalent results when compared with the _ ab initio _ potential presented in reference,@xcite in which the average error is 1.1173 % .
both the refined potential @xmath92 and the _ ab initio _
potential @xmath66@xcite are in agreement with the experimental value , with errors less than the experimental errors . therefore ,
both potential curves , @xmath92 and @xmath93 are equally suitable to describe the viscosity coefficient between 1 and 5 k.
the long range potential function for helium was determined in the present work from viscosity coefficient data at temperatures below 5 k. to the best of our knowledge , this approach has never been reported . in this case , quantum assumptions and the bose - einstein statistic are required to connect viscosity coefficient with interatomic potential .
this relationship is nonlinear , and a common method to solve this inverse problem can not be employed . to circumvent this difficulty , the sensitivity matrix was used to linearize this problem .
the initial problem was rewritten as a linear fredholm integral equation of the first order , whose solution is given by the tikhonov method .
the success of the proposed method depends on finding a cross - section sensitivity matrix by a consistent and robust procedure . in a previous work,@xcite a differential equation for the functional derivative of the quantum phase shift with respect to the potential energy
was established and coupled to calogero s equation . in the present work , this previous result was used to establish the cross - section sensitivity matrix within exact quantum theory , for the first time .
finally , the long range potential curve was obtained from experimental data and compared with recent _ ab initio _
potentials,@xcite showing an excellent agreement .
the average difference between both is less than 2.6 % between 3 and 4 , with an even smaller difference in the potential depth , -11.1 k against -11.0 k. the potential depth is important result , which shows that there is excellent agreement with a bound state prediction of the @xmath94 dimer.@xcite the refined potential describes the viscosity coefficient with an average error of 1.6823 % that is less than the experimental error ( 5 % ) , a result similar to that found for the _ ab initio _ potential ( 1.1173 % ) .
the present algorithm is general and can be used to determine the potential energy function whenever _ ab initio _ calculations can not be used .
the method could be applied without difficulty to an inverted second virial coefficient or other transport properties at low temperatures when quantum assumptions are needed .
we would like to thank cnpq and fapemig for their financial support .
h. e. cox , f. w. crawford , e. b. smith and a. r. tindell , _ mol .
phys . _ , 1980 , * 40 * , 705 - 712 . n. h. t. lemes , r. c. o. sebastio and j. p. braga , _ inverse probl .
_ , 2006 , * 14 * , 581 - 587 .
t. ho and h. rabitz , _ j. chem .
_ , 1989 , * 90 * , 1519 - 1525 . n. h. t. lemes , e. borges , r. v. souza and j. p. braga , _ int .
j. quantum chem .
_ , 2008 , * 108 * , 2623 - 2627 . n. h. t. lemes , j. p. braga , m. o. alves and .
costa , _ j. mol . model _ , 2014 , * 20 * , 2317 .
w. cencek , m. przybytek , j. komasa , j. b. mehl , b. jeziorski , k. szalewicz , _ j. chem
. phys . _ , 2012 , * 136 * , 224303 .
costa , n. h. t. lemes , m. o. alves , r. c. o. sebastio and j. p. braga , _ j. braz
_ , 2013 , * 24 * , 363 - 368 . a. j. c. varandas ,
a _ , 2010 , * 114 * , 8505 - 8516 .
m. przybytek , w. cencek , j. komasa , g. lach , b. jeziorski , and k. szalewicz , _ phys . rev
. lett . _ , 2010 , * 104 * , 183003 .
erratum , _ phys .
_ , 2012 , * 108 * , 129902 .
g. c. maitland , _ intermolecular forces : their origin and determination _ , oxford university press , london , 1981 .
s. u. kim and c. w. monroe , _ j. comput .
phys . _ , 2014 , * 273 * , 358 - 373 .
k. pachucki and j. komasa , _ j. chem .
_ , 2006 , * 124 * , 064308 . s. chapman and t. g. cowling , _ the mathematical theory of non - uniform gases : an account of the kinetic theory of viscosity , thermal conduction and diffusion in gases _ , cambridge university press , cambridge , 1953 .
j. o. hirschfelder , c. f. curtiss and r. b. bird , _ molecular theory of gases and liquids _ , wiley , new york , 1964 .
v. d. viterbo , n. h. t. lemes and j. p. braga , _ rev . bras .
ensino fis .
_ , 2014 , * 36 * , 1310
. j. p. braga , and j. n. murrell , _ mol .
phys . _ , 1984 ,
* 53 * , 295 - 299 .
e. w. becker , r. misenta and f. schmeissner , _ z. phys . _ , 1954 , * 137 * , 126 - 136 .
e. bich , r. h. and e. vogel , _ mol .
phys . _ , 2007 , * 105 * , 3035 - 3049 . p. j. nacher , _ j. chem .
phys . _ , 1994 , * 101 * , 6367 - 6368 .
m. keil and l. j. danielson , _ can .
_ , 1988 , * 66 * , 159 - 163 .
|
the long range potential of helium - helium interaction , which requires accurate _ ab initio _ calculation , due to the small value of the potential depth , approximately 11 k ( @xmath0 ev = 0.091 kj / mol ) at 2.96 , will be obtained in this study by an alternative technique .
this work presents a robust and consistent procedure that provides the long range potential directly from experimental data .
however , it is difficult to obtain experimental data containing information regarding such a small potential depth .
thereby , sensitivity analysis will be used to circumvent this difficulty , from which viscosity data at lower temperatures ( @xmath1 k ) were chosen as appropriate data to be used to retrieve the potential function between 3 and 4 . the linear relationship between the potential energy function and the viscosity coefficient
will be established under quantum assumptions and the bose - einstein statistic .
the use of quantum theory is essential , since the temperatures are below 5 k. the potential obtained in this study describes the viscosity with an average error of 1.68 % that is less than the experimental error ( 5 % ) , with the results being similar to those obtained for recent _ ab initio _ potentials .
* keywords : * variable phase method . viscosity coefficient . low temperature . sensitivity analysis .
|
in recent years , type - ii gasb / inas two dimensional ( 2d ) structures have attracted growing attention in both theoretical and applied research .
a distinguish feature of the gasb / inas quantum well ( qw ) structures and superlattices is that electrons and holes are mostly localized in the inas and gasb layers , respectively , and that the energy gap between the electron and hole subbands can be efficiently tuned by adjusting the widths of the layers , for reviews see e.g @xcite .
narrow gap type - ii gasb / inas superlattices , being characterized by a small effective mass , high mobility , excellent electron confinement , and the possibility of modulation doping , became an important system for mid - infrared radiation detection @xcite and were proposed as a novel candidate for terahertz ( thz ) radiation detectors @xcite .
furthermore , it has been shown most recently that type - ii gasb / inas qws can be tuned through the topological quantum phase transition by the variation of the layer thicknesses and/or the application of front- and back - gate voltage @xcite .
this property should enable the fabrication of high - quality materials with an interchanged band order , a crucial issue for nontrivial 2d topological insulators ( tis ) @xcite .
the inverted band order results in the emergence of topologically protected helical edge states , where carriers with opposite spin projections counter - propagate along each edge @xcite and supporting the spin quantum hall effect @xcite .
contemporary investigations of gasb / inas qws in the topologically non - trivial phase comprise theoretical works on their electronic properties @xcite and transport measurements @xcite . here , we report on the observation of a thz radiation induced linear and circular photogalvanic effects in type - ii gasb / inas qws in the inverted regime .
photogalvanic spectroscopy bridges electron transport and optics and has become a very efficient tool to study nonequilibrium processes in semiconductors and low dimensional structures , yielding information on their point - group symmetry , details of the band spin splitting , processes of momentum , energy , and spin relaxation etc .
, for reviews see @xcite .
most recently , the circular photogalvanic effect has been applied to study the rashba / dresselhaus spin - orbit coupling in non - inverted type - ii gasb / inas superlattices excited by polarized infrared radiation @xcite .
furthermore , the circular photogalvanic effect , resulting in a photon - helicity - dependent photocurrent , is a key ingredient for the realization for a all - electric scheme to detect the radiation stokes parameters @xcite .
our experiments demonstrate that photogalvanic currents can be excited efficiently in type - ii gasb / inas qws in the frequency range from about 1 to 3.5 thz .
linearly or circularly polarized terahertz radiation induces a @xmath0 electric current whose magnitude and direction depend on the radiation polarization state .
in particular , the photocurrent reverses its direction by switching the radiation polarization from left - handed to right - handed .
we present a kinetic theory of the observed circular and linear photogalvanic effect which describes well the experimental data .
we suggest that the photocurrents are of orbital origin and stem from the lack of an inversion symmetry in the structure .
experimental data on the photocurrent anisotropy reveal that the dominated inversion symmetry breaking mechanism responsible for the thz induced photocurrents is the structure inversion asymmetry .
the samples were fabricated on te - doped ( 100)-oriented gasb substrates by molecular beam epitaxy ( mbe ) .
the mbe chamber was equipped with evaporation cells for group iii elements ( al , ga , in ) and with cracking cells for group v elements ( as , sb ) .
the substrate temperature was controlled by pyrometer during growth .
the wafer was heated to 300@xmath1c pre - growth in the load lock chamber .
oxide desorption was realized at 580@xmath1c under sb stabilization flux . afterwards
the growth was started with a 100 nm thick gasb smoothing layer followed by the lower 200 nm lattice matched alas@xmath2sb@xmath3 barrier .
the explicit compound quantum well consists nominally of a 10 nm gasb and a 13 nm inas well . on top a 100 nm alas@xmath2sb@xmath3 barrier and a 5 nm gasb cap
finish the structure .
the composition of the qw widths aims at an inversion of the electron and hole subbands .
8x8 kp simulations suggest an inversion of about 70 mev at @xmath4 .
scanning electron microscopy ( sem ) images of the samples cross section , structure composition and thickness of all layers are shown in fig . [ figure_1_sample ] . to probe the photocurrent in different directions four pairs of ohmic contacts
were prepared ; two pairs in the corners of the @xmath5mm@xmath6 squared samples and two in the middle of the edges , see fig .
[ figure_2_phi ] .
the samples edges are oriented along the @xmath7 $ ] and @xmath8 $ ] directions . to characterize the samples ,
electrical measurements were done on hall - bar structured samples at helium temperature . in these measurements
the mobility is @xmath9 cm@xmath6/vs and the electron density is @xmath11 @xmath12 .
transport measurements show the coexistence of @xmath13- and @xmath14-type carriers and support the assumption of the sample being in the inverted regime .
to excite photocurrents we applied a high power line - tunable nh@xmath15 thz laser @xcite optically pumped by a pulsed transversely excited atmosphere ( tea ) co@xmath16 laser @xcite .
the laser operates at the frequencies @xmath17 = 1.07 , 2.03 , and 3.31thz .
the corresponding photon energies ( @xmath18 , 8.4 , and 13.7mev , respectively ) are smaller than the band gap as well as the size - quantized subband separation .
the radiation induces indirect optical transitions ( drude - like free - carrier absorption ) in the lowest conduction subband .
the laser generates single pulses with a duration of about 100ns and a repetition rate of 1hz .
the laser peak power , being of the order of @xmath19 kw , was controlled by the thz photon drag detector @xcite .
the laser radiation was focused onto a spot size of about 1.6 - 2 mm diameter depending on the radiation frequency .
the beam had an almost gaussian profile which was measured by a pyroelectric camera@xcite .
various polarization states of the radiation are achieved by transmitting the linearly polarized ( @xmath20 ) laser beam through @xmath21/4 or @xmath21/2 crystal quartz plates . by rotating a @xmath21/4 plate one
transfers the linear into elliptical polarization .
the polarization states are directly related to the angle @xmath22 between the initial linear polarization of the laser light and the optical axis of the plate . by that we obtain the degree of circular polarization given by @xmath23 and the bilinear combinations of the linear polarization vector components .
the latter following the stokes parameters @xcite are expressed via @xmath24 and @xmath25 @xcite . in some experiments we also used @xmath21/2 quartz plates providing a linearly polarized radiation with the polarization plane rotated by the azimuth angle @xmath26 from the initial linear polarization of the laser light along the @xmath27 axis .
the experimental geometry is sketched in the insets in panel ( c ) of fig .
[ figure_2_phi ] .
unbiased samples were excited at oblique incidence with the plane of incidence lying in the ( @xmath28 ) plane .
the angle of incidence @xmath29 was varied between @xmath30 to + 30@xmath1 , with respect to the layer normal @xmath31 $ ] . both photocurrent components perpendicular , @xmath32 , and parallel , @xmath33 , to the plane of incidence were investigated . in experiments aimed at the study of photocurrent anisotropy , we also used other orientations of the plane of incidence with respect to the @xmath34 and @xmath27 axis .
details of these geometries are discussed below . to avoid any signals stemming from the illumination of contacts or the sample edges ,
the samples were excited through a 2.7 mm diameter aperture in a metal mask covering the contacts and the edges .
the photocurrents were measured by the voltage drops across a 50@xmath35 resistor , fig .
[ figure_2_phi](c ) .
the peak voltage was recorded with a 1 ghz bandwidth storage oscilloscope .
several samples of the same batch were investigated at room temperature yielding the same results .
irradiating the middle of ( 001)-grown type - ii gasb / inas qw ( edges are not illuminated ) with polarized thz radiation at oblique incidence photocurrent we detect a signal in the absence of an external bias .
the signal follows the temporal structure of the laser pulse .
its response time is determined by the time resolution of our setup but it is at least 1 ns or shorter .
this fast response is typical for photogalvanic effects where the signal decay time is expected to be of the order of the momentum relaxation time @xcite , which according to transport data in our samples at room temperature is much shorter 1 ns .
the photocurrent is observed in the investigated frequency range from 1.1 and 3.3 thz .
photosignals are observed both in the direction perpendicular to the plane of radiation incidence [ transverse geometry , @xmath32 in figs .
[ figure_2_phi](a ) and ( b ) ] and in the direction lying in the incidence plane [ longitudinal geometry , @xmath33 in fig .
[ figure_2_phi](d ) ] .
figures [ figure_2_phi ] shows the dependences of the photocurrent components @xmath32 and @xmath33 on the angle @xmath22 for the angles of incidence @xmath36 .
we find that the polarization dependence of the transverse current is well fitted by @xmath37 and that of the longitudinal current by @xmath38 , where @xmath39 and @xmath40 are fitting parameters .
a distinguish feature of the photocurrent contribution @xmath41 is that it reverses the direction upon switching the radiation polarization from right - handed ( @xmath42 ) to left - handed ( @xmath43 ) circular polarization .
note that both contributions , @xmath44 and @xmath45 , are equal to zero for circularly polarized radiation .
these contributions are driven by the linear polarization component of elliptically polarized radiation . to prove this the photocurrent was measured with linearly polarized radiation .
the top right inset in fig .
[ figure_3_alpha ] shows an exemplary dependence of the transverse photocurrent @xmath32 on the azimuth angle @xmath26 obtained at the incidence angle @xmath46 .
the polarization dependencies of the transverse and longitudinal ( not shown ) components of the photocurrents are well described by @xmath47 and @xmath48 , respectively , using @xmath49 and @xmath40 obtained in the experiments with elliptically polarized radiation .
note that the offsets @xmath40 and @xmath50 have magnitudes substantially smaller than the amplitudes of the photocurrents @xmath51 and @xmath49 and are out of scope of our paper . as a function of the incidence angle
all photocurrent contributions change signs at @xmath52 .
this is demonstrated for photocurrent components @xmath51 and @xmath49 in fig .
[ figure_4_theta ] .
the similar behaviour is also observed for the transverse and longitudinal photocurrents induced by linearly polarized radiation , see fig .
[ figure_3_alpha ] showing the dependence of the transverse photocurrent on the angle of incidence .
figure [ figure_4_theta ] presents a set of the experimental data on the circular and linear photogalvanic currents obtained for thz radiation of different frequencies .
these data reveal that the increase of the radiation frequency leads to the decrease of the magnitudes of both photocurrent contributions .
such a trend corresponds to the well known spectral behavior of the drude absorption @xcite at @xmath53 , where @xmath54 is the momentum relaxation time of carriers .
below we demonstrate that the observed polarization dependences , the variation of the photocurrent with the angle of incidence , as well as the frequency behavior follows the microscopic theory of the circular and linear photogalvanic effects at the drude - like absorption in ( 001)-oriented qws . the observed circular photogalvanic effect ( the photocurrent proportional to @xmath51 ) allows us to analyse the relative strength and interplay of the structure inversion asymmetry ( sia ) and the bulk inversion asymmetry ( bia ) of the studied samples .
the technique relies on the phenomenological linear coupling between the polar vector of the photocurrent and the axial vector of the photon angular momentum , which is a prerequisite for the circular photogalvanic effect to occur and which can originate in qws from sia or bia , see also refs .
@xcite . to judge on the dominant mechanism of inversion symmetry breaking causing the photocurrent generation , we have studied the photocurrent anisotropy in the qw plane . to this end
we have measured the longitudinal and transverse circular photocurrents for four different orientations of the radiation incidence plane : ( @xmath28 ) , ( @xmath55 ) , and the planes ( @xmath56 ) and ( @xmath57 ) containing the cubic axes @xmath58 $ ] and @xmath59 $ ] .
our measurements reveal that in all these geometries the circular photocurrent is reliably detected in the direction perpendicular to the incidence plane only and being almost of the same magnitude .
these findings correspond to the photocurrent caused by sia .
the circular photogalvanic current caused by bia would flow in the plane of radiation incidence ( longitudinal photocurrent ) for the incidence planes ( @xmath56 ) and ( @xmath57 ) .
moreover , it would have an opposite sign of the transverse components for the incidence planes ( @xmath28 ) and ( @xmath55 ) .
therefore , together with the sia contribution we would observe substantial difference in the magnitude . both features are not detected .
thus , we conclude that the dominant mechanism of the photocurrent generation in the samples under study is related to sia .
finally , we note that illuminating the edges of the sample without a metal mask we have observed a strong circular photocurrent .
the photocurrent is excited at normal incidence .
this observation is in contrast to the data discussed above , where the @xmath0 current for @xmath60 vanishes .
as our samples are in the inverted regime and the photon energy of the thz radiation used is rather small , this is probably an evidence that the topological helical edge states are involved in the formation of such a photon helicity driven photocurrent .
however , this edge photocurrent requires a further careful study and is out of scope of the present paper .
in our experiments the photon energy of the thz radiation @xmath61 is much smaller than the electron fermi energy @xmath62 ( a few hundreds mev ) .
therefore , a relevant approach for the theoretical description of the photocurrent generation is quasi - classical . in this approach
@xcite the thz radiation is considered as an @xmath63 electric field @xmath64 where @xmath65 is the field amplitude , which acts upon the electrons .
the in - plane component @xmath66 of the field drives the electrons back and forth in the qw plane inducing an @xmath63 electric current at the field frequency @xmath67 .
the @xmath63 current does not follow the time dependence @xmath66 but is retarded with respect to the in - plane field by the phase @xmath68 , where @xmath54 is the momentum relaxation time .
the out - of - plane component @xmath69 pushes the electron density in the qw normal to the @xmath70 or @xmath71 directions depending on the field polarity .
this shift of the electron density along the growth direction in strongly asymmetric systems such as gasb / inas heterostructures results in a change of the electron mobility .
the modulation of the electron mobility with the same frequency @xmath67 results in a rectification of the @xmath63 electric current , which causes the emergence of a @xmath0 component of the current . for clockwise and counter clockwise rotating electric fields , corresponding to the @xmath72 and @xmath73 circular polarized radiations , respectively , the @xmath0 current flows in the opposite directions .
the rate of electron scattering , which determines the mobility , can be presented in general in an asymmetric heterostructure up to the first order in @xmath69 in the form @xmath74 where @xmath75 is the scattering rate at zero field , @xmath76 is the field - induced correction related to sia and @xmath77 is the electron charge .
for such a scattering rate , the kinetic theory developed in refs .
@xcite yields the following expression for the @xmath0 electric current density @xmath78 where @xmath79 is the electron density , @xmath80 is the effective mass at the fermi level , @xmath81 and @xmath82 are the fermi momentum and fermi velocity , respectively , @xmath54 is the momentum relaxation time , @xmath83 \:,\ ] ] @xmath84 is the parameter of the scattering asymmetry , @xmath85 \:,\ ] ] and @xmath86 and @xmath87 are the polar angles of the vectors @xmath88 and @xmath89 before and after the scattering event , respectively .
the current ( [ j ] ) is proportional to the square of the electric field amplitude , i.e. to the radiation intensity , and comprises both linear and circular photocurrents .
note that while in the theoretical consideration the current density @xmath90 is used , in the experiments the electric current @xmath91 is measured , which is proportional to the current density @xmath90 .
for the experimental geometry sketched in fig .
[ figure_2_phi](c ) , the components of the photocurrent ( [ j ] ) can be presented in the form @xmath92 \ , , \nonumber \\
j_{y } = \frac{e^3 n_e } { m^ * } \frac { \zeta_1 \tau_p^2 e_0 ^ 2}{1+(\omega\tau_p)^2 } \ , t_p t_s \ , \sin \theta [ \sin 4\varphi /2 + \omega\tau_p \sin
2\varphi ] \,,\end{aligned}\ ] ] where @xmath93 and @xmath94 are the amplitude transmission coefficients for @xmath14- and @xmath95-polarized radiation , respectively , @xmath96 @xmath97 is the amplitude of the electric field of the incident radiation , @xmath98 is the angle of refraction related to the incidence angle @xmath29 by @xmath99 , and @xmath100 is the refractive index .
the component of the photocurrent in the incidence plane @xmath101 contains only a linear photogalvanic current while the perpendicular component @xmath102 contains both linear and circular photogalvanic currents .
curves in figs .
[ figure_2_phi ] ( a ) , ( b ) and ( d ) show the polarization dependences of @xmath33 and @xmath32 , respectively , calculated after eqs .
( [ j_final ] ) .
they correspond well to these experimental data apart a small polarization - independent offset which is detected for some frequencies .
the theory also describes well the polarization behavior of the photocurrent excited by linearly polarized radiation : in this case , the terms in the square brackets should be replaced by @xmath103 and @xmath104 for @xmath33 and @xmath32 , respectively .
this transformation just reflects a change of the stokes parameters of radiation transmitted through @xmath105 plate as compared to those transmitted through @xmath106 plate .
the calculated dependence of the photocurrent on the azimuth angle @xmath26 is shown in fig .
[ figure_3_alpha ] for the transverse photocurrent excited by radiation with the frequency @xmath107thz .
we note that , in accordance with the eqs .
( [ j_final ] ) , for all fits in figs .
[ figure_2_phi ] and [ figure_3_alpha ] we used the same prefactor for the linear photogalvanic effects excited at a given frequency .
equations ( [ j_final ] ) also reveal that the photocurrent emerges at oblique incidence only and flows in the opposite directions for positive and negative angles of incidence @xmath29 .
moreover , the equations explain the overall variation of the photocurrent with the angle of incidence .
the corresponding calculated dependencies are shown by solid lines in figs .
[ figure_4_theta ] ( a ) and ( b ) for the circular and linear photogalvanic currents , respectively .
to summarize , our experiments show the photon helicity driven photogalvanic currents and photocurrents induced by the linearly polarized radiation can be effectively generated in type - ii gasb / inas qws at room temperature .
a microscopic theory developed in the framework of the boltzmann equation of motion suggests that the photocurrents are of orbital origin and stem from the lack of an inversion symmetry in the structure .
the performed experimental analysis of the circular photogalvanic effect anisotropy indicates that , in the context of photocurrents , the structure inversion asymmetry outweighs the bulk inversion asymmetry in the studied systems . in addition to photocurrents excited in the bulk of the qws , we also observed a strong circular photogalvanic current stemming from the illumination of the sample edges .
the edge photocurrent may be related to helical edges states in the studied qw structures characterized by the inverted band order , which is subject of a future work .
the photocurrents determined by the radiation stokes parameters , in particular , the one proportional to the radiation helicity , are detected in the whole investigated frequency range .
this observation , together with a subnanosecond response time , suggests type - ii gasb / inas qws as a promising candidate for the fast room - temperature all - electric detectors of thz radiation polarization .
we thank v.v . belkov for fruitful discussions . the work was supported by the elite network of bavaria ( k - nw-2013 - 247 ) , the dfg priority program spp1666 , the volkswagen stiftung program , the state of bavaria and the german research foundation ( ka2318/4 - 1 ) .
s.a.t . acknowledges support from the rfbr ( projects 14 - 22 - 02102 and 16 - 02 - 00375 ) .
f. qu , a.j.a .
beukman , s. nadj - perge , m. wimmer , b .-
nguyen , w. yi , j.thorp , m. sokolich , a.a .
kiselev , m.j .
manfra , c.m .
marcus , and l.p .
kouwenhoven , phys .
lett * 115 * ( 2015 ) 036803 doi : 10.1103/physrevlett.115.036803 s.n .
danilov , b. wittmann , p. olbrich , w. eder , w. prettl , l.e .
golub , e.v .
beregulin , z.d .
kvon , n.n .
mikhailov , s.a .
dvoretsky , v.a .
shalygin , n.q .
vinh , a.f.g .
van der meer , b. murdin , and s.d .
ganichev , j. appl .
physics * 105 * ( 2009 ) 013106 doi : 10.1063/1.3056393 c. drexler , s.a .
tarasenko , p. olbrich , j. karch , m. hirmer , f. mller , m. gmitra , j. fabian , r. yakimova , s. lara - avila , s. kubatkin , and s.d .
ganichev , nature nanotech .
* 8 * ( 2013 ) 104 doi : 10.1038/nnano.2012.231 s.d .
ganichev , p. schneider , v.v .
belkov , e.l .
ivchenko , s.a .
tarasenko , w. wegscheider , d. weiss , d. schuh , b.n .
murdin , p.j .
phillips , c.r .
pidgeon , d.g .
clarke , m. merrick , p. murzyn , e.v .
beregulin , and w. prettl , phys .
b. r * 68 * ( 2003 ) 081302 doi : 10.1103/physrevb.68.081302 s. giglberger , l.e .
golub , v.v .
belkov , s.n .
danilov , d. schuh , ch .
gerl , f. rohlfing , j. stahl , w. wegscheider , d. weiss , w. prettl , and s.d .
ganichev , phys .
b * 75 * ( 2007 ) 035327 doi : 10.1103/physrevb.75.035327 m. kohda , v. lechner , y. kunihashi , t. dollinger , p. olbrich , c. schnhuber , i. caspers , v.v .
belkov , l.e .
golub , d. weiss , k. richter , j. nittaand , and s.d .
ganichev phys .
b rapid communic . * 86 * ( 2012 ) 081306 doi : 10.1103/physrevb.86.081306
|
we report on the observation of photogalvanic effects induced by terahertz radiation in type - ii gasb / inas quantum wells with inverted band order .
photocurrents are excited at oblique incidence of radiation and consists of several contributions varying differently with the change of the radiation polarization state ; the one driven by the helicity and the other one driven by the linearly polarization of radiation are of comparable magnitudes .
experimental and theoretical analyses reveal that the photocurrent is dominated by the circular and linear photogalvanic effects in a system with a dominant structure inversion asymmetry .
a microscopic theory developed in the framework of the boltzmann equation of motion considers both photogalvanic effects and describes well all the experimental findings .
|
due to continued nanoelectronics scaling , metal interconnects as well as transistors in integrated circuits are becoming ever smaller and are approaching atomic scale dimensions @xcite . as a consequence , understanding the effects of size - dependent phenomena on material properties is becoming ever more critical to enable efficient device performance .
present day transistor technologies use copper as an interconnect material , however it is unknown if at the small cross - sections required for future technologies copper will be able to fulfill this function effectively due to increased resistances ; measured line resistivity increases dramatically for cu nanowires ( nws ) compared to bulk materials .
@xcite it is required that the electrical conductivity of small cross section copper nanostructures be understood to maintain acceptable power consumption in future nanoelectronics generations .
in particular the contribution of the individual scattering sources , such as surfaces,@xcite grain boundaries,@xcite electron - phonon interaction @xcite and impurities , to overall resistivity needs to be assessed to aid development of interconnects which minimize line resistance .
a succinct overview of these issues can be found in the review of josell , brongersma and tkei.@xcite in general the dominant scattering mechanism in metal nanostructures is dependent on processing conditions and the geometry of the resulting nanostructures .
for example @xcite have reported electron scattering to vary with grain size in the case of gold thin films .
grain boundary scattering is found to dominate for grain sizes much smaller than the electron mean free path , while for much larger grain sizes it is the surface that is critical to resistivity .
for grain sizes comparable to the mean free path , both mechanisms are responsible for the increases resistances relative to bulk gold . meanwhile , in the case of copper thin films @xcite have reported that it is grain boundary scattering which dominates .
@xcite report that the temperature dependence of resistivity in cu nws is consistent with surface scattering as the dominant scattering source .
this report is corroborated by @xcite , who find that diffuse surface scattering and line edge roughness is consistent with the temperature dependence of resistivity for cu nws down to 25 nm diameter and that the role of grain boundary scattering is minimal .
similarly , @xcite report that resistivity is dominated by diffuse surface scattering for 20 - 100 nm cu nws .
thus though the situation for metal nanostructures can vary by material and process conditions , the literature to date as pertains to cu nws seems quite emphatic ; the surface is the dominant scattering source . in low dimension
cu films the surface environment has been reported to drastically influence conductivity . in a series of studies,@xcite chawla and co
- workers have investigated the scattering at the surfaces of cu thin films .
they report that scattering at a cu - vacuum surface is partially specular , whereas after tantalum deposition the surface scattering becomes diffuse .@xcite the effect of oxygen has also been reported , with diffuse scattering again reported after oxidation.@xcite the effect of a variety of coating metals on cu thin films has also been reported from first principles simulations.@xcite metals with a density of states ( dos ) comparable to the dos of surface cu atoms ( al and pd ) lower resistivity , while those which do not ( ta , ti and ru ) increase resistivity . against this backdrop ,
it is probable that surface scattering is of particular importance for small diameter cu nws and thus we study unterminated ( i.e. , in vacuum ) and oxidized 1 and 3 nm cu nws using _ ab initio _ and semi - empirical computational techniques .
a high degree of anisotropy of cu nw conductivity with crystal orientation has previously been reported using tight - binding computational methods,@xcite thus we also consider the effects of crystal orientation .
we calculate transmission spectra and transmission pathways @xcite in each case for geometry optimized structures .
we find a strong dependence of transmission on crystal orientation and termination . [ 110 ] nws consistently show a larger transmission than [ 100 ] and [ 111 ] nws .
surface oxidation reduces transmission substantially compared to unterminated nws .
additionally we find that transmission even for 1 nm unterminated nws is larger at the nw surface than the core due to the existence of surface states .
we provide a simple analysis of transmission in nanoscale cu wires , which accounts for the trends in transmission we obtain .
, [ 110 ] and [ 111 ] nws , ( d ) , ( e ) and ( f ) show the respective o - terminated nws .
cu is orange and o is red . ] , [ 110 ] and [ 111 ] nws , respectively .
cu is orange and o is red . ]
the nw structures studied in this work are based on the bulk fcc cu structure , with [ 100]- , [ 110]- and [ 111]-oriented nws chosen with approximately 1 and 3 nm diameters . due to the small diameter of the nws studied in this work , atoms at the nw surface
will have an especially large contribution to the nw properties e.g. , 57% , 59% and 49% of the atoms are at the surface for the [ 100 ] , [ 110 ] and [ 111 ] 1 nm nws , respectively . hence in these nanostructures
, electrical resistivity can be expected to be dominated by surface scattering .
while standard complimentary metal oxide semiconductor ( cmos ) technologies employ ta based materials as barrier layers to prevent cu diffusion , deposited ta has been shown to cause diffusive scattering at the surface of thin cu films.@xcite as scaling continues , the relative contribution of surface scattering to total scattering increases dramatically and along with the barrier thickness means that ta - based barrier layers may well become unsuitable for efficient interconnect design . to isolate barrier effects
we have chosen to restrict this work to consideration of unterminated ( i.e. , in vacuum ) and oxygen terminated ( i.e. , surface oxidized ) nws .
the relaxed ( i.e. , zero strain ) structures of the nws studied are shown in fig .
[ fig : small - structures ] and fig .
[ fig : large - structures ] . density functional theory ( dft ) as implemented in the openmx software package @xcite is used to study the electronic properties of the 1 nm diameter cu nws .
the pbe@xcite formulation of the generalized gradient approximation ( gga ) exchange and correlation functional is used in conjunction with norm - conserving pseudopotentials@xcite and a strictly localized pseudo - atomic orbital ( pao ) basis,@xcite which enables us to perform the decomposition of the transmission into localized pathways .
the basis sets used are 6.0h - s4p2d2 and 7.0-s2p2 for cu and o respectively .
the first part of the basis set notation gives the pao cutoff radius in bohr , while the second part indicates the orbitals used for the valence electrons , e.g. , o 7.0-s2p2 gives a cut - off radius of 7.0 bohr and 8 basis functions ( 2 s functions and 6 p functions ) .
this cu basis set agrees reasonably with experiment , giving an optimal fcc lattice parameter of 3.63 and a bulk modulus of 135 gpa compared to experimental values of 3.615 and 137 gpa , respectively . the supercell approach is used and because the nws are periodic only along their axis , a cell size incorporating a minimum of 1 nm of vacuum in the transverse directions is used such that the interaction between periodic images is minimized .
the first brillouin zone is sampled using 11 k - points generated according to the monkhorst - pack method @xcite and an energy cutoff of 200 ry is used to generate the grid by which real space quantities are discretized .
the atomic positions in the nws are relaxed until all forces are less than 3x10@xmath0 hartree / bohr ; total energy is also minimized with respect to the lattice cell parameter along the nw axis for each nw orientation .
optimized lattice parameters along the nw axis are given in table [ tab : cell_parameter ] .
the 3 nm o - terminated nws are also relaxed using the openmx software package using the same parameters . [ 100 ] and [ 110 ] nw structures are fully optimized whereas we take a hollow - core approach ( described in the supporting information ) for the [ 111 ] nw due to the number of atoms in the unit cell . due to
its reduced computational demand , density functional tight binding ( dftb ) as implemented by the quantumwise software package@xcite is used to calculate the electronic transport properties of the 3 nm nws , using a density mesh cutoff of 15 hartree and an 11 k - point monkhorst - pack grid .
dftb results were compared to dft in a carefully chosen set of cases in order to explicitly assess their validity in nw structures .
the electrical properties of these nws are modeled in the context of the landauer - bttiker formalism , @xcite which includes the concepts of wide _ reflectionless _ contacts and electrodes .
this formalism relates the electric current through a nw attached to two electron reservoirs ( contacts ) with the probability that an electron can be transmitted via the relation @xmath1 \ , de \ , , \end{aligned}\ ] ] where @xmath2 is the electron charge , @xmath3 planck s constant , @xmath4 is the transmission coefficient per spin channel @xmath5 at energy @xmath6 and applied bias @xmath7 , @xmath8 is the fermi - dirac distribution , @xmath9 ( @xmath10 ) and @xmath11 ( @xmath12 ) are the temperature and chemical potential of the left ( right ) electrode , and the applied bias is given by @xmath13 in the case of a nw directly attached to identical electron reservoirs ( i.e. , same chemical potential and temperature ) we may write the linear response conductance in the zero - temperature limit as @xmath14 given that the nw is modeled as a perfectly periodic structure , all bloch waves propagate with unit probability .
this leads to the expression for the nw contact resistance @xmath15_{t_i = 1 ; e = e_f}^{-1 } \nonumber \\ = & \left [ \frac{e^2}{h } m(e_f ) \right]^{-1 } \ , , \end{aligned}\ ] ] where the nw acts as a ballistic waveguides for the total number of modes at the fermi level @xmath16 , which includes all bloch waves @xmath17 .
this quantity provides information on the electrical resistivity intrinsic to the nw s electronic structure . throughout this work
the reported transmission then corresponds to a sum over the total number of modes @xmath18 for each structure as a measure of their maximum potential conductivity as affected by their reduced scale and in the absence of scattering sources such as structural defects , grain boundaries , and phonons . even though surface scattering is expected to be the most important contribution to resistance at these dimensions at zero temperature ,
the contribution of electron - phonon scattering at finite temperatures can not be neglected .
@xcite report a decomposition of the resistance of copper lines with widths ranging from 75 to 500 nm with electron - phonon scattering estimated to account for 60% of the resistivity at 300k .
the effects of electron - phonon interactions on resistivity in sub 5 nm copper nws are currently unknown and have been excluded from this study .
it is worth noting that for interconnect applications these nws function would be to provide electrical contact between devices ( e.g. , transistors ) across a short length and with small voltage drops - ideally zero .
hence the use of the linear response conductance applies and its validity can be extended to arbitrarily high temperatures as long as the transmission coefficient remains approximately constant over the energy range in which the transport takes place . with the aim of providing some insight into the influence of the local chemical environment on electronic transport , a decomposition of transmission into spatially resolved pathways
is computed according to the formalism described by @xcite , as implemented in the quantumwise software package , in which the transmission through a plane perpendicular to the transport direction that divides the system in two regions @xmath19 and @xmath20 is decomposed as @xmath21 where @xmath22 and @xmath23 are atoms on each side of said plane .
we note that the use of a green s function implementation to the landauer approach is required for decomposing the transmission into local contributions between pairs of atoms and hence this approach is taken for computing all electronic transport properties .
the optimized cell parameters along the periodic direction of the nws for a single repeat unit are given in table [ tab : cell_parameter ] . in the case of the unterminated nws ,
the optimized cell parameter is very similar to the calculated bulk cell parameter ; this agreement is reflected in the nw structures wherein the atoms move only slightly from their bulk positions after geometry optimization ( see fig .
[ fig : small - structures ] ) .
when oxygen atoms are placed on the nw surface the situation changes dramatically with significant surface reconstruction seen for both the 1 nm and 3 nm nws .
there is a strong tendency for o to pull surface cu atoms away from the nw bulk - like core until the bonded cu and o atoms lie in a single plane . in bulk
cu(i ) and cu(ii ) oxide , each cu bonds to 2 or 4 o atoms for cu@xmath24o and cuo respectively , conversely each o atom bonds to 4 or 2 cu atoms for cu@xmath24o and cuo , respectively .
the surface reconstruction seen in this work emulates the kinds of structural motifs seen in these oxides .
the ratio of copper to oxygen ( cu : o ) at the nw surface is 1.5 , 1.7 , and 1.9 for the 1 nm oxidized [ 100 ] , [ 110 ] , and [ 111 ] nws respectively , i.e. , it lies between bulk cuo and bulk cu@xmath24o . in all cases , each o atom at the surface bonds to 3 cu ,
thus the oxidation is intermediate to the situation seen for the bulk oxides .
the surface oxide in the nws consists of a very thin curved layer which locally is essentially two dimensional .
thus the nw surfaces are well oxidized as the cu : o ratio is comparable to bulk copper oxide . for the [ 100 ] and [ 110 ] nws , each cu at the surface bonds to 2 oxygen atoms .
thus while the cu : o ratio at the surfaces are different , in the case of [ 100 ] and [ 110 ] nws , each o bonds to the same number of cu atoms and each cu bonds to the same number of o atoms . in spite of the similarity in the surface environment the surface geometries differ . for [ 100 ]
the o - cu - o motif is linear at cu , while for [ 110 ] most o - cu - o are v - shaped , a minority are linear .
the [ 111 ] nw surface has 13 cu atoms bonded each to 2 o atoms and 8 cu atoms bonded each to 1 o atom , thus in line with the higher cu : o ratio , this surface appears to be less oxidized than the [ 100 ] and [ 110 ] surfaces .
the cu atoms at the [ 111 ] surface which bond to two o have a mix of linear and v - shaped o - cu - o structural arrangements .
l * 5c & & & + & bulk & 1 nm & 3 nm & 1 nm & 3 nm + 100 ] & 3.63 & 3.60 & 3.63 & 3.75 & 3.64 + [ 110 ] & 2.57 & 2.56 & 2.59 & 2.86 & 2.73 + [ 111 ] & 6.29 & 6.26 & - & 6.46 & - + the surface reconstruction does not strongly affect the positions of cu atoms below the surface layers , this is clearly evident in the 3 nm nw structures shown in fig . [
fig : large - structures ] , beyond the first two cu surface layers the cu atoms occupy approximately equivalent configurations with respect to their bulk positions .
the reconstruction at the surface is associated with a lengthening of the optimized cell parameter relative to the bulk and unterminated nw , which is more severe for the 1 nm nws than 3 nm nws .
this suggests that there is competition between the surface of the nw and core over the relaxed cell parameter along the axis .
the oxidized surface favors an increase in length along this axis , presumably to better incorporate the surface o atoms in a layer .
conversely the core of the nw consists only of cu atoms with their geometry similar to bulk fcc cu and therefore favors a bulk - like cell parameter .
thus while unterminated nw cell lengths are similar for 1 and 3 nm nws , the oxidized 3 nm nws have a smaller ( i.e. , closer to the bulk value ) cell parameter than the oxidized 1 nm nws . a concomitant increase in nw diameter
is not observed , therefore this change in cell parameter occurs exclusively along the nw axis .
the surface structural rearrangement seen here for the o - terminated [ 100 ] nw , shown in fig . [ fig : small - structures](d ) and fig .
[ fig : large - structures](b ) , is very similar to that seen in experiment for surface oxidized cu(100 ) slabs@xcite suggesting the structural rearrangement we see for the nws in our simulations is representative of experimental structures with highly prepared surface treatments .
nws having larger transmission than [ 100 ] and [ 111 ] nws .
the zero of energy is taken to be at the fermi level . ]
l * 5r & & & + & bulk & 1 nm & 3 nm & 1 nm & 3 nm + 100 ] & 28.0 & 22.8 & 19.4 & 7.3 & 8.8 + [ 110 ] & 29.7 & 26.7 & 24.3 & 15.3 & 15.1 + [ 111 ] & 27.9 & 15.3 & 11.8 & 3.7 & 4.0 + the transmission spectra calculated for the terminated and oxidized 1 nm and 3 nm cu nws are shown in fig .
[ fig : transmission ] for a narrow energy range of @xmath2550 mev centered about the fermi energy . in each case , we find that the oxidized nw has a substantially lower electron transmission than the corresponding unterminated nw .
this result is consistent with the increase in resistance after chemical oxidation previously been reported by @xcite for cu thin films and arises from the reduction of the dos at the surface upon oxidation .
additionally , we find a strong dependence of transmission on nw orientation . in all cases
we find that the [ 110 ] nw has a larger transmission than [ 100 ] , which in turn has a larger transmission than [ 111 ] , irrespective of termination . as expected the transmission of the nws increases going from 1 nm to 3 nm diameter due to the increasing cross - sectional area .
the cross - sectional areas of the nws are not identical as diameters deviate from exactly 1 nm and 3 nm due to the different atomic arrangements of the crystal orientations .
thus , we also calculate the transmission per unit area @xmath26 , presented in table [ tab : transmission_area ] , to ensure that this variation in transmission with orientation is not due to deviation in cross sectional area .
there is ambiguity in how to define the cross sectional area of the nws due to atomic structure .
the atoms in the nw have volume which extends beyond the nuclear positions and the smaller the nw the greater this contribution will be to the nw cross section . for simplicity we have taken the cross - section to be circular , elliptical , or octagonal as appropriate , with dimensions equal to the maximum inter - nuclear width of the nw plus twice the cu atomic radius ( 1.28 ) .
we find that the transmission per area @xmath26 follows the same trend as the total transmission plotted in fig .
[ fig : transmission ] ; [ 110 ] has the largest @xmath26 and [ 111 ] the smallest , regardless of diameter or termination .
this result is consistent with the anisotropy in cu conductivity with crystal orientation reported recently by @xcite .
in fact , this orientation dependence is of such significance that for the 3 nm nws the _
oxidized _ [ 110 ] nw has a transmission comparable to the _ unterminated _ [ 111 ] nw at the fermi energy ; a remarkable result consistent with the shape of the fermi surface of bulk fcc copper , in which the fermi surface exhibits a vanishing dos along the @xmath27 directions .
the extent of the orientation dependence on transmission is also clearly evident in table [ tab : transmission_area ] . in the case of unterminated nws ,
the transmission for a given cross sectional area for [ 110 ] is approximately twice that of [ 111 ] nws , after oxidation this margin further increases with the [ 110 ] transmission per unit area approximately 3.5 times larger than for the [ 111 ] nw .
it may be anticipated that a more oxidized nw would show greater suppression in transmission , however the 1 nm [ 111 ] nw is in fact less oxidized than the [ 110 ] nw ; each surface cu in the [ 110 ] nw bonds to 2 oxygen atoms while surface [ 111 ] atoms bond to either 1 or 2 oxygen atoms . meanwhile ,
unterminated [ 100 ] and [ 110 ] nws show similar transmission per unit area ; [ 100 ] is about 4 units smaller than [ 110 ] for both 1 nm and 3 nm nws .
after oxidation , the transmission per unit area for [ 100 ] nws is approximately 60% of [ 110 ] nws , i.e. , transmission suppression on oxidation is relatively larger for [ 100 ] than [ 110 ] .
thus both the surface chemical environment and the nw orientation play a critical role in determining the overall transmission of the copper nw .
the arrows show the direction of transmission paths .
the arrow thickness represents the magnitude of the transmission while the color represents the orientation of the path with respect to the nw axis .
forward transmission is shown in blue , radial transmission in green , and backscattering in red .
pathways with transmission below 10% of the maximum value are omitted for clarity .
the dashed black lines mark the positions of the planes shown fig .
[ fig:1nm - cross - section](e ) . ] , ( b ) o - terminated [ 100 ] , ( c ) unterminated [ 110 ] , ( d ) o - terminated [ 110 ] , ( e ) unterminated [ 111 ] and ( f ) o - terminated [ 111 ] .
atomic positions are shown as `` + '' ( cu ) or `` x '' ( o ) .
the dot size represents the absolute size of the transmission pathway crossing the plane , while the dot color indicates the size and direction of the transmission pathway .
blue dots correspond to forward transmission and red dots to backscattering .
note the difference in the color scales ranges . ]
the transmission results show that orientation and surface environment drastically impact on cu nw electron transport .
these effects are not easy to decouple , however , a greater understanding of the relationship between them can be established by examining transmission through the nw at an atomic scale . we have calculated transmission pathways in the nws@xcite at the fermi level , which are shown in fig . [ fig:1nm - pathways ] for the unterminated [ 111 ] nw , to particularly elucidate the effects of nw crystal orientation and surface oxidation on transmission .
transmission pathways provide a visual representation of transmission through the nw with the arrow width representing the size of the transmission between an atom pair . as described in the caption of fig .
[ fig:1nm - pathways ] , the arrows are colored according to their angle with respect to the nw axis ; blue arrows correspond to forward transmission , green arrows to radial transmission ( i.e. , atoms within the same transverse plane ) , and red arrows backscattering or reflection . as expressed in eq .
( [ eq : pathways ] ) , the transmission pathways crossing a given plane perpendicular to the transport direction always sum to give the total transmission coefficient .
a more convenient visual representation is portrayed in fig .
[ fig:1nm - cross - section ] in which the intersections of transmission pathways with planes located halfway between non - equivalent atomic planes give a clear image of the decomposition of the total transmission into pathways per cross - sectional area per nw .
such planes are indicated by dashed black lines in fig .
[ fig:1nm - pathways ] for the case of the unterminated [ 111 ] nw .
these projections provide a heuristic representation of localized transmission and allow quick identification of where the most important contributions lie within the nw cross - section .
[ fig:1nm - cross - section](a ) shows transmission pathways for the unterminated [ 100 ] nw . there are transmission pathways both parallel ( dots centered at atomic positions ) and angled to the nw axis ( centered between atomic positions ) .
angled pathways are larger in magnitude than parallel pathways as represented by their larger dot diameter and their color .
this difference in magnitude can be attributed to the distance between the cu atoms being larger if a direct route parallel to the axis is taken ( 3.60 versus 2.53 ) .
it is is readily seen that pathways along the surface dominate electron transmission for this nw as the dots in the surface are larger than below the surface , where the parallel pathways are smaller and obscured by the symbols indicating the atomic positions .
the unterminated [ 110 ] nw transmission pathways are shown in fig .
[ fig:1nm - cross - section](c ) .
transmission involving only surface atoms occurs along the nw axis and is large ( indicated by the darker blue dots centered about atomic positions ) while pathways at an angle to the axis exhibit much lower transmission . below the surface
the situation is similar , transmission along the axis is larger than off - axis . in this orientation
the cu atom separation is similar for both parallel and off axis transmission and thus the more direct parallel route is favored . as for the unterminated [ 100 ] nw , transmission is larger at the nw surface than below the surface with the largest dots corresponding to pathways located at the surface and parallel to the nw axis .
the transmission pathways for the unterminated [ 111 ] nw , shown in fig .
[ fig:1nm - cross - section](e ) . in this case
, there is no significant transmission directly parallel to the nw axis , this is presumably because the cu interatomic distance along the nw axis is simply too large compared to the off axis distances ( 6.26 versus @xmath282.5 reflecting that no nearest neighbors align along the nw axis ) .
as previously mentioned , this result is consistent with the vanishing density of states found along the @xmath27 directions in the fermi surface of fcc cu ; there are no transmission paths parallel to the [ 111 ] direction as shown by the absence of arrows parallel to the nw axis in fig .
[ fig:1nm - pathways ] and of dots centered about atomic positions in fig .
[ fig:1nm - cross - section](e ) . as shown in fig .
[ fig:1nm - pathways ] transmission pathways tend to flow along the nw surface until they reach a cu atom with a low coordination number ( such as the one at the top center of the figure ) , at which point the electrons are scattered into the nw core where they continue to propagate along angled paths until they reach another surface where this process repeats .
the lack of transmission pathways parallel to the [ 111 ] nw axis makes the surface critical as all paths inevitably lead to the surface .
while there are some transmission pathways inside the core of the nw , transmission _ through _ the nw occurs primarily at the nw surface ; and much like the unterminated [ 100 ] and [ 110 ] nws , transmission is larger at the surface .
however , in this case as well as dots representing forward transmission , the dots representing negative values ( i.e. , electron backscattering or reflection paths ) are much larger than for the unterminated [ 100 ] and [ 110 ] nws . the transmission pathways of the unterminated nws consistently show that transmission is larger at the nw surface than in the core and also that backscattering for these nws is relatively small , particularly for the [ 100 ] and [ 110 ] nws in which the total magnitude of all reflecting paths is 14% and 13% of that of forward paths , whereas for [ 111 ] nws this value increases to 25% .
this surface - dominated transmission can be explained in terms of the increased local density of states ( ldos ) of surface atoms at the fermi level as a result of their low coordination .
this allows electrons otherwise localized at bonds to become available for transmission at the surface atoms .
[ fig:1nm - ldos ] shows the average local density of states ( ldos ) of an atom situated at the surface and of an atom close to the center of the nws cross section ; the atoms at the surface show a larger density of states in the range around the fermi level studied in this work .
unterminated [ 111 ] nws exhibit the largest boost with the average ldos value of a surface atom being 2.55 times that of an atom near the center of the nw at the fermi level ; similarly , atoms at the surface of [ 100 ] nws have an average surface - to - center ldos ratio of 2.41 and a considerably lower ratio of 1.39 is found for [ 110 ] nws . as might be expected from the transmission spectra shown in fig .
[ fig : transmission ] and the previous analysis , oxidation of the nws changes the behavior at the nw surfaces considerably . in the
[ 100 ] nw , transmission at and just below the nw surface is reduced upon oxidation , in fact fig .
[ fig:1nm - cross - section](b ) shows that near the surface forward transmission is largely suppressed and backscattering pathways are enhanced . in the core of the nw
the transmission is similar to the unterminated nw as shown by similarly arranged , sized and colored dots. thus in the case of the [ 100 ] nw , oxidation appears to affect the surface and immediate subsurface almost exclusively . in the case of the [ 110 ] nw , more significant back scattering
is seen upon surface oxidation with the appearance of the large red dots indicating a relatively large magnitude in fig .
[ fig:1nm - cross - section](d ) . unlike the [ 100 ] nw ,
transmission is also reduced in the core of the [ 110 ] nw upon surface oxidation .
a suppression of the pathways parallel to the nw axis can be clearly seen all the way to the center when comparing to the unterminated nw ( [ fig:1nm - cross - section])(c ) .
thus it appears that the effect of surface oxidation is felt deeper in the nw for the [ 110 ] orientation compared to [ 100 ] .
however , transmission per unit area for the oxidized [ 100 ] nw actually drops more relative to the unterminated nw compared to the [ 110 ] orientation , as reported in table [ tab : transmission_area ] .
this is likely caused by a reduction of transmission at the surface and immediate subsurface in the [ 100 ] oxidized nw , in which the surface has a larger contribution relative to the core in the unterminated case as shown in fig .
[ fig:1nm - ldos ] and discussed above .
the surface of the oxidized [ 111 ] nw is more disordered than the oxidized [ 100 ] and [ 110 ] nws studied in this work .
this disorder along with the indirect transmission associated with transport along this crystal orientation leads to a complicated , disordered transmission pathways ( fig .
[ fig:1nm - pathways](f ) ) consistent with the low transmission calculated for this structure .
in contrast to the [ 100 ] and [ 110 ] oxidized nws , there is large forward transmission at the surface for the oxidized [ 111 ] nw , which can be seen clearly in the mapping of the transmission pathways onto a cross sectional area of the [ 111 ] nw shown in fig .
[ fig:1nm - cross - section](f ) . however , the large backscattering paths in the core of the nw result in a lower overall transmission ; all paths in the core lead either to the surface ( where oxidation has largely suppressed forward transmission ) or are backscattered . to understand the behavior of the total transmission per nw given in the previous sections , we provide an analysis for a simple model describing the electron transmission behavior of round nws which partitions the nws into regions of similar behavior based on the magnitudes of the transmission pathways within the them .
the generalization of the model to other geometries is straightforward .
the model provides insight into the observation that transmission per unit area decreases as area increases for the unterminated nws and that the transmission per unit area is relatively constant for oxidized nws for a given orientation .
is the radius of the nw , @xmath29 is the radius of the bulk region , @xmath30 is the width of the subsurface region , i.e. , the depth to which @xmath26 is affected by the surface , and @xmath31 is the width of the surface . ]
the total transmission @xmath32 of a round nw of radius @xmath33 is divided into contributions from three distinct regions of the nw ; the surface , the subsurface and the bulk regions ; as shown in fig .
[ fig : model ] for a cross section of the nw .
the surface region is described by a thickness @xmath31 , the subsurface region by a thickness @xmath30 and the bulk - like core is described by a radius @xmath29 ; therefore @xmath34 to each region is assigned a transmission per unit area @xmath26 .
the surface transmission per unit area @xmath35 is assumed constant , whereas the subsurface transmission per unit area @xmath36 can vary with depth into the nw as signified by the grading in fig .
[ fig : model ] .
however , it is assumed that an effective transmission per unit area can be assigned to the subsurface region .
the bulk region is defined as the core atoms for which the transmission per unit area approaches the bulk behavior @xmath37 .
the area of the surface region can be expressed as @xmath38 = ( 2 \pi r ) \delta - \pi \delta ^2,\ ] ] which for @xmath39 can be approximated as @xmath40 similarly the area for the subsurface region can be expressed as @xmath41 = 2\pi ( r-\delta)d - \pi d^2,\ ] ] which for @xmath42 is given as @xmath43 the core region displaying bulk - like behavior has an area given by @xmath44 which for @xmath42 becomes @xmath45 the fact that @xmath46 and @xmath47 scale linearly with @xmath33 whereas @xmath48 scales as @xmath49 for large values of @xmath33 simply describes the fact that the surface and subsurface play decreasing roles as the nw s radius is increased . an effective transmission per unit area @xmath26
is then defined for each region and the total transmission is expressed as @xmath50 to explore the role of the surface and subsurface transmission per unit area in small cross section nws we consider four cases .
_ case i : @xmath51_. in this limit , the surface dominates and transport is given purely by the surface and represents , for example , the case of an atomic chain . _ case ii : @xmath52_. this condition describes when the cross sectional area of the nw is of a scale that a subsurface region is present but the nw diameter is not large enough for the core of the nw to display bulk behavior for transmission ( it should be noted that the subsurface region is not sharply defined as represented by the grading of the gray region used to define it in fig .
[ fig : model ] ) . in the transmission plots of fig .
[ fig:1nm - cross - section ] for the unterminated cu nws , it can be seen that the subsurface region has a lower transmission per unit area compared to the nw surface regions .
hence by decreasing the nw diameter from 3 nm to 1 nm , a larger transmission per unit area is obtained consistent with the results listed in table [ tab : transmission_area ] .
_ case iii : @xmath53_. as the nw radius increases , eventually a core region is formed whereby local transmission pathways approach the values obtained in bulk copper .
when this core region is of a comparable dimension to the surface and subsurface layer , the transmission per unit area will reflect the contributions from the different regions .
as the subsurface region in unterminated cu nws has a lower transmission per unit area than the bulk , the overall transmission per unit area will increase relative to case ii as the region forms .
iv : @xmath54_. as the nw s radius is further increased , the core region s contribution will quickly dominate due to the area increasing as @xmath49 as opposed to the linear dependence on nw radius for the surface and subsurface regions .
hence the transmission per unit area for the nw will increase with increasing @xmath33 until asymptotically approaching the bulk value .
the following picture emerges in the case of unterminated nws . for extremely small diameter nws with diameters of approximately 1 nm ,
surface and subsurface transmissions combine to provide a transmission per unit area that is comparable to bulk copper for [ 100 ] and [ 110 ] and less than bulk copper for [ 111 ] . as the nw diameter increased , for example up to 3 nm , the lower subsurface transmission with respect to surface transmission results in a net lowering of overall transmission per unit area compared to smaller diameter nws .
as the nw diameter is increased further , the core region begins to behave like a bulk conductor and the transmission rises and asymptotically approaches the bulk transmission per unit area .
the picture needs to be slightly modified for the case of nws with surface oxidation . as seen in fig .
[ fig:1nm - cross - section ] , the surface transmission is reduced significantly due to surface oxidation . from table [
tab : transmission_area ] , it is seen that the transmission per unit area for the 1 nm and 3 nm diameter nws are comparable . once oxidized , the transmission at the surface is suppressed and the transmission of the subsurface region dominates .
we anticipate that the transmission per unit area of the subsurface region for a given orientation is similar for 1 nm and 3 nm nws and this combined with the low surface transmission ( see fig .
[ fig:1nm - cross - section ] ) leads to the transmission per unit area for each orientation being approximately constant going from 1 nm to 3 nm diameter , as reported in table [ tab : transmission_area ] .
however , the transmissions of both 1 nm and 3 nm oxidized nws are much lower than found for the bulk , and hence larger cross section nws are required before significant bulk - like behavior from the core contributes to the net transmission per unit area .
the electron transport properties of 1 nm and 3 nm diameter copper nws have been calculated in each case for unterminated and surface oxidized nws of [ 100 ] , [ 110 ] and [ 111 ] crystal orientations .
we find even 1 nm diameter copper nws to be metallic , however electron transmission is strongly dependent on both surface termination and crystal orientation .
surface oxidation suppresses electron transmission compared to unterminated copper nws , consistent with previous reports for copper thin films .
the [ 110 ] oriented nws consistently show a higher electron transmission than [ 100 ] oriented nws which in turn have a higher electron transmission than [ 111 ] nws .
transmission in unterminated copper nws is larger at the nw surface than below the surface , irrespective of crystal orientation .
a different picture emerges for surface oxidized nws wherein transmission is lower at the nw surface than in the subsurface for [ 100 ] and [ 110 ] oriented nws .
however while the oxidized [ 111 ] nw surface transmission is reduced compared to the unterminated nw , it nonetheless continues to show larger surface transmission than subsurface transmission .
transmission per unit area decreases with increasing diameter for unterminated nws but remains approximately constant for the surface oxidized nws studied .
a simple model of transmission in round nws which divides the nw into regions based on local transmission pathways explains the differing behaviors of unterminated and oxidized nws .
briefly , transmission per unit area is lower in the subsurface region than at the surface for the unterminated nws studied , thus when the diameter is increased from 1 nm to 3 nm the transmission per unit area of the nw decreases as the subsurface region becomes relatively more important . in oxidized nws ,
the overall transmission per unit area is comparable for 1 nm and 3 nm nws .
the suppression of the surface transmission in this case increases the contribution of the subsurface to overall transmission and a similar subsurface transmission per unit area may be expected for 1 nm and 3 nm oxidized nws for the same crystal orientation resulting in similar overall transmission per unit area .
overall , our results suggest that conductivity in sub 3 nm copper nws is highly sensitive to crystal orientation and surface termination , and that careful material preparation and processing will be essential in order to maximize conductivity .
this work was performed as part of the intel - tyndall research collaboration sponsored by intel components research .
ass was funded under an irish research council postgraduate scholarship .
we are grateful to quantumwise a / s for providing access to the quantumwise simulation software .
31ifxundefined [ 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 link:\doibase 10.1103/physrevb.66.075414 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2711385 [ * * , ( ) ] link:\doibase 10.1017/s0305004100019952 [ * * , ( ) ] link:\doibase 10.1103/physrevb.1.1382 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2355435 [ * * , ( ) ] \doibase http://dx.doi.org/10.1146/annurev-matsci-082908-145415 [ , * * , ( ) ] link:\doibase 10.1103/physrevb.82.113409 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.apsusc.2013.02.037 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.041402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.81.155454 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3292022 [ * * , ( ) ] \doibase
http://dx.doi.org/10.1016/j.tsf.2012.05.070 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3157271 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.3489357 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.235423 [ * * , ( ) ] \doibase http://dx.doi.org/10.1103/physrevb.81.045406 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.4868979 [ * * , ( ) ] \doibase http://dx.doi.org/10.1038/nchem.546 [ * * , ( ) ] @noop @noop * * , ( ) \doibase http://dx.doi.org/10.1103/physrevb.47.6728 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.13.5188 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevb.65.165401 [ * * , ( ) ] http://stacks.iop.org/0953-8984/14/i=11/a=302 [ * * , ( ) ] link:\doibase 10.1147/rd.13.0223 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1103/physrevb.87.155414 [ * * , ( ) ]
|
density functional theory and density functional tight - binding are applied to model electron transport in copper nanowires of approximately 1 nm and 3 nm diameters with varying crystal orientation and surface termination .
the copper nanowires studied are found to be metallic irrespective of diameter , crystal orientation and/or surface termination .
electron transmission is highly dependent on crystal orientation and surface termination .
nanowires oriented along the [ 110 ] crystallographic axis consistently exhibit the highest electron transmission while surface oxidized nanowires show significantly reduced electron transmission compared to unterminated nanowires .
transmission per unit area is calculated in each case , for a given crystal orientation we find that this value decreases with diameter for unterminated nanowires but is largely unaffected by diameter in surface oxidized nanowires for the size regime considered .
transmission pathway plots show that transmission is larger at the surface of unterminated nanowires than inside the nanowire and that transmission at the nanowire surface is significantly reduced by surface oxidation
. finally , we present a simple model which explains the transport per unit area dependence on diameter based on transmission pathways results .
|
linear polarization can be an important diagnostic of relativistic jets and accretion flows associated with black hole systems . in the case of the massive black hole in the galactic center , sgr a * , the properties of its mm - wavelength linear polarization probes the accretion environment on scales inaccessible with other techniques ( bower et al.1999a , b ; aitken et al.2001 ; bower et al.2001 ; bower et al.2003 ; marrone et al.2006 ) . the apparent absence of linear polarization at wavelengths exceeding 2.7 mm and sharp rise in polarization fraction at shorter wavelengths sets an upper limit to the rotation measure ( rm ) .
this limits the mass accretion rate to @xmath11 at distances of @xmath12 schwarzschild radii from the black hole , which eliminates certain classes of accretion flow ( quataert & gruzinov 2000 ; agol 2000 ) , but is consistent with cdaf and jet models ( e.g. falcke , mannheim and biermann 1993 ) .
the rm measures the accretion rate by serving as a proxy for the electron column density once coupled with assumptions about the magnetic field .
equipartition between kinetic , magnetic and gravitational energy is often assumed to relate the electron density and magnetic field ( e.g. bower et al .
1999a ; melia & falcke 2001 ; marrone et al .
2006 ) . the discovery of variations in both the polarization angle ( bower et al.2005 ) and fraction ( marrone et al.2006 ) suggests two potential sources of variability .
the variations may be intrinsic and , therefore , offer evidence on the nature and structure of the emission region on a scale of @xmath13 .
changes in the rm along the line of sight may also induce external polarization variability .
structure in the accretion region on all scales within the bondi radius can contribute to rm variations .
there is thus strong motivation to accurately characterize the rm , its fluctuations , and the intrinsic variability of the polarized source . in 2 we present the detection of linear polarization of sgr a * at 3.5 mm , showing that it varies on short timescales as well as relative to historical non - detections . combining this in 3 with other data , our detection yields the best constraints so far on the rm .
we discuss the nature of the accretion flow and on the intrinsic source properties in 4 .
sgr a * was observed with the bima array at 3.5 mm between 28 march and 01 april 2004 ut .
all five observations were obtained in identical lst time ranges that corresponded roughly to 11 to 16 ut .
observations were obtained in a polarimetric mode that produced stokes @xmath14 , @xmath15 , @xmath16 and @xmath17 images , with the two frequency bands centered on 82.76 and 86.31ghz , each with 800 mhz bandwidth .
data were calibrated assuming that the flux density of j1733 - 130 was steady at its mean value of 2.7jy .
j1733 - 130 was observed for 5 minutes every 30 minutes .
observations were obtained in the b configuration giving a resolution of @xmath18 in position angle ( p.a . )
the data were filtered to remove spacings shorter than 20k@xmath20 , similar to the processing of previous sgr a * polarization experiments .
this @xmath21@xmath22 distance cutoff gives a total flux estimate accurate to better than 10% from 8.4 to 86 ghz ( bower et al.2001 ) .
a priori amplitude calibration using system temperature and default gain information was applied to j1733 - 130 .
both sources were phase self - calibrated .
polarization leakage terms from observations of 3c279 on 11 november 2003 were applied .
numerous sources of error contribute to the accuracy with which we can measure the polarization fraction and p.a .. in addition to statistical errors that include noise from the sky and the telescope , the leakage of polarization from one polarization handedness to the other is an important effect .
this converts a fraction of the total intensity into a false polarization signal .
leakage calibration uses observations of a bright calibrator source and simultaneously solves for the polarization of the source and the leakage terms appropriate for individual antennas .
imperfect correction leaves a residual false polarization signal and introduces errors that exist in the calibrator observation .
bima leakage solutions at 1.3 mm were found to be stable over periods of months to years ; variability in the solutions were due to changes in receiver orientation ( bower , wright & forster , 2002 ) .
typical variations in antenna leakage terms were @xmath23 over 2 years , indicating a @xmath24 variation in the average leakage solution for the 9 bima antennas . solutions at 3.5 mm are expected to have the same characteristics although this has not been studied in depth .
in addition to variations in the leakage error , statistical errors in the calibrator data introduce errors .
a comparison of actual measured p.a.s of 3c 279 at 3.5 mm and 1.3 mm from bima with measured p.a.s at 1.3 cm and 0.7 cm from the vla indicate excursions of 10@xmath3 20@xmath3 ( bower et al.2002 ) .
these deviations exceed those expected from leakage corrections and statistical errors .
they are possibly real source effects but may also be uncorrected instrumental errors .
the average flux densities of sgr a * and linear polarization e - vectors are listed in table [ fluxtable ] . in the polarization fraction and p.a .
errors we include the effect of a 0.5% polarization leakage error .
this error propagation is accurate in the case of the high snr detections in the lsb but presents a lower limit to the errors in the low snr detections in the usb .
inter - day flux density variations are detected at a significant level , with a reduced-@xmath25 of 4.9 for the hypothesis of constant flux density for sgr a * for 4 degrees of freedom .
the mean flux density is @xmath26jy and the rms variability amplitude is @xmath27jy .
polarization was detected in all five epochs , making this the first detection of polarization at 3.5 mm , the longest wavelength at which polarization has been detected in sgr a*. the mean polarization fraction is @xmath1% and its mean p.a .
is @xmath28 . the polarization fraction and p.a
. vary from day to day .
the detection of polarization is more robust in the lsb ( 82.8 ghz ) than in the usb ( 86.3 ghz ) .
the origin of this difference is unclear
. it may be due to changes in the leakage corrections and/or poor phase stability at the higher frequency .
we have previously seen data from bima in which polarization results in one sideband were more reliable than in the other ( bower et al.2001 ) .
the bandwidth used in these observations is sufficiently small that the linear polarization is not depolarized by rotation of the polarization vector across the observing band , as discussed below .
cccccccc ut day ( 2004 ) & @xmath14(jy ) & @xmath29(jy ) & @xmath15(mjy ) & @xmath16(mjy ) & @xmath30(mjy ) & p ( mjy ) & @xmath31 ( @xmath3 ) + lsb & & & & & & & + 88.5 & 1.94 & 0.11 & 20 & 34 & 5 & @xmath32 & @xmath33 + 89.5 & 1.96 & 0.20 & 31 & 9 & 4 & @xmath34 & @xmath35 + 90.5 & 1.71 & 0.20 & 29 & 16 & 8 & @xmath36 & @xmath37 + 91.5 & 1.69 & 0.20 & 29 & 18 & 8 & @xmath38 & @xmath39 + 92.5 & 2.12 & 0.14 & 55 & 21 & 4 & @xmath40 & @xmath41 + usb & & & & & & & + 88.5 & 1.98 & 0.12 & -10 & 11 & 5 & @xmath42 & @xmath43 + 89.5 & 1.97 & 0.18 & 7 & -15 & 4 & @xmath44 & @xmath45 + 90.5 & 1.68 & 0.19 & 3 & 3 & 7 & @xmath46 & @xmath47 + 91.5 & 1.76 & 0.16 & 9 & 6 & 7 & @xmath48 & @xmath49 + 92.5 & 2.25 & 0.09 & 30 & 19 & 4 & @xmath50 & @xmath51 + cccccc @xmath52 ( ghz ) & @xmath53(% ) & @xmath54 ( % ) & @xmath31 ( @xmath3 ) & @xmath55 ( @xmath3 ) & measurement epochs + 82.8 & 2.1 & 0.4 & 15.2 & 8.2 & 5 + 86.3 & 0.8 & 0.5 & 18 & 35 & 5 + 150@xmath56 & 12 & @xmath57 & 83 & 3 & 1 + 222@xmath56 & 11 & @xmath58 & 88 & 3 & 1 + 216@xmath59 & 10 & 1 & 115 & 13 & 2 + 230@xmath59 & 9 & 3 & 117 & 24 & 7 + 340@xmath60 & 6.1 & 2.0 & 145 & 9 & 6 + 350@xmath56 & 13 & @xmath61 & 161 & 3 & 1 + 400@xmath56 & 22 & @xmath62 & 169 & 3 & 1 +
table [ polntable ] lists the mean polarization p.a.s for all previous detections of linear polarization in sgr a*. due to systematic errors from polarization leakage calibration , the poor quality of upper sideband data at 3.5 mm , and the relatively small spacing between the two sidebands , we are unable to compute a meaningful rm estimate from the 3.5 mm data alone . the combined average polarization data are consistent with a single rm of @xmath5radm@xmath6 and an intrinsic polarization p.a .
fig.1 plots the best rm fit to the mean polarization p.a . data .
the best fit , with a reduced-@xmath25 of 29 computed using the errors quoted in table [ polntable ] , is obtained if the 85ghz p.a .
is rotated by @xmath64 .
a fit to the unwrapped 85ghz p.a . yields a reduced-@xmath25 of 54 and @xmath65radm@xmath6 , while an extra @xmath64 wrap yields @xmath66radm@xmath6 with an associated reduced-@xmath25 of 64 .
the reduced-@xmath25 for the best fit is unacceptably high because of the unrealistically small errors associated with the aitken et al.(2001 ) observations .
these single - epoch low - resolution measurements predate the identification of polarization variability in sgr a*. we therefore fit only to interferometric observations spanning multiple epochs .
the resulting fit properties depend on whether one regards measurements from the two sidebands at @xmath67ghz , where available , as being mutually independent .
if so , the reduced-@xmath25 for a fit including the two -180@xmath3-wrapped 3.5 mm points is 0.9 , but is 3.6 for the unwrapped 3.5 mm polarization pas . if not , the two @xmath25 values are 0.8 and 2.1 respectively . in both cases ,
the solution involving the wrapped 3.5 mm points is preferred .
we regard these cases as bounding the true significance of the rm . for the case of independent sideband measurements , the associated best fit is @xmath68 and @xmath69radm@xmath6 .
we adopt the rm @xmath70radm@xmath6 derived from all of the multi - epoch data . given the variability of the p.a . and
the uncertain systematics of different measurements , we consider the use of all data to provide the most conservative estimate of the rm . .
diamonds denote measurements for which polarization variability is detected , and their error bars denote the standard error of the mean of the p.a
. variations .
triangles denote single epoch measurements only .
the dashed line shows the best fit to all data , the solid line to all points excluding the aitken et al .
( 2001 ) data , and the dot - dashed line to the unwrapped 85ghz points , denoted by stars . ]
m@xmath71 ( ghez et al.1998 ) .
two choices for @xmath72 are shown : one in which the outer scale is large ( effectively infinite ) and another in which it is only three times larger than the inner cutoff radius . ]
the accretion rate implied by this rm depends on density and magnetic profiles assumed . following the prescription outlined by marrone et al.(2006 ) , . ]
we write @xmath73 , for @xmath74 and assume equipartition between magnetic , kinetic and gravitational energy .
the radius @xmath75 designates the point at which the flow becomes relativistic , and so ceases to contribute to the rm because its contribution to faraday rotation is decreased by a factor @xmath76 .
the outer scale is a proxy for the coherence length of the magnetic field fluctuations .
the estimated accretion rate for various adaf ( @xmath77 ) and cdaf ( @xmath78 ) models is shown in figure [ figmdotlabel ] as a function of @xmath75 .
the accretion rate is in the range @xmath79m@xmath71yr@xmath8 for @xmath80 , @xmath81 and @xmath82 .
the fact that a single rm accounts for the frequency dependence of all mean polarization p.a.s implies that the faraday rotation occurs external to the polarized source .
internal rotation would cause the rm to vary as a function of frequency .
the emission is optically thick at all frequencies at which polarization is detected ( falcke et al.1998 ; zhao et al.2003 , yuan et al.2003 ) , so if any internal faraday rotation did occur , the diminuation of opacity effects with frequency , which increases the depth down to which one observes emission , would cause a corresponding increase in the faraday rotation path length .
an interpretation of the p.a .
jitter observed at 85ghz in terms of rm variability would imply rms deviations of @xmath83radm@xmath6 , far smaller than the @xmath84radm@xmath6 rms deviations implied by the 340ghz marrone et al.(2006 ) fluctuations interpreted similarly .
however , the absence of a clear frequency dependence in the inter - day jitter makes it difficult to ascribe to rm fluctuations , and hence variations in the accretion rate .
table 2 shows the rms p.a .
deviations for the multi - epoch observations at 85 , 216 , 230 and 340ghz .
these are inconsistent with the @xmath85 dependence expected of rm fluctuations from a magnetoionic medium external to the source , suggesting instead that the jitter reflects changes in the intrinsic source polarization p.a .. nonetheless , the fact that the fluctuations at these frequencies were not observed simultaneously , coupled with the short time span of the 85ghz observations compared to the @xmath86month albeit sporadic sampling of the 230 and 340ghz measurements , still admits the possibility that the p.a .
dispersion observed at 85ghz is unrepresentative of its long - term average .
this appears unlikely .
if the @xmath87 rms p.a .
variations observed at 216230ghz were associated with rm fluctuations we would expect @xmath88 fluctuations at 85ghz and would not expect to find @xmath89 over a set of disparate measurements .
moreover , the presence of intrinsic p.a .
changes is unsurprising given that the polarization fraction is also intrinsically variable , as discussed below .
the lack of p.a .
jitter attributable to rm fluctuations can be interpreted in terms of an upper limit in accretion rate fluctuations .
the absence of clearly identifiable inter - day rm fluctuations suggests it is uniform on inter - day timescales .
the consistency of the observations over a large number of disjoint epochs and frequencies with a single rm further suggests that the underlying accretion rate is constant on the timescale over which the observed polarization pas were averaged .
the uncertainty in our rm fit places an upper bound on the accretion rate variations using @xmath90 , valid under the assumption of equipartition between magnetic , kinetic and gravitational energy ( e.g. marrone et al.2006 ) .
the 7% uncertainty in the rm implies @xmath91 fluctuations less than @xmath10% .
the limit on @xmath91 variations is largely consistent with the limits imposed by source flux density variations . in the jet model
the flux density scales as @xmath92 ( falcke et al.1993 ) .
the standard deviation of the 3.5 mm fluxes is 10% , comparable to the errors on the individual measurements , which imposes an upper limit of 14% on @xmath91 fluctuations .
marrone et al.(2006 ) detect 10% rms intensity variations at 340ghz .
the five intensity measurements from bower et al.(2005 ) at 216230ghz exhibit 29% modulations , implying @xmath91 fluctuations of 40% .
variability in polarization fraction is detected by all multi - epoch observations ( fig.[figpolnlabel ] ) , at 85 , 230 and 340ghz , and presumably occurs at 112ghz , where it was not detected at the 1.8% ( 1-@xmath30 ) level by a previous search ( bower et al .
2001 ) . a previous limit of 1% linear polarization at 86ghz ( bower et al.1999b ) demonstrates that it also varies on long timescales .
since the faraday screen is external to the source the polarization amplitude fluctuations must be intrinsic .
instrumental depolarization effects are too low to explain the variability : bandwidth depolarization is only important for rms greater than @xmath93radm@xmath6 at 85ghz , while beam depolarization is similarly improbable given the sub - mas size of sgr a * at mm wavelengths ( bower et al . 2004 ; shen et al .
spatial variation in the rm across the transverse extent of the source could depolarize the emission , but at even 85ghz this requires rm fluctuations @xmath94radm@xmath6 ( quataert & gruzinov 2000 ) .
both the high variability of the mm and sub - mm emission ( zhao et al.2004 ; wright & backer 1993 ; tsuboi et al.1999 ) and the linearly polarized emission of sgr a * are possibly associated with its excess sub - mm emission ( serabyn et al.1997 ; falcke et al.1998 ; melia et al.2001 ) .
adaf models that fit the cm to - sub - mm spectrum include at least two distinct populations of radiating particles ( yuan et al.2003 ) , with the second important at @xmath95ghz in order to explain the sub - mm bump .
the model of yuan et al.(2003 ) , in which the sub - mm emission is dominated by thermal electrons , overpredicts the polarization fraction at 85ghz ( cf .
fig.[figpolnlabel ] ) . in this model
the degree of linear polarization ranges from 32% at 85ghz to 70% at 400ghz assuming a uniform magnetic field and that faraday depolarization intrinsic to the source is unimportant ( see their fig.3b ) . in the context of this model
the ratio of the predicted to observed polarization levels can only be attributed to magnetic field inhomogeneity intrinsic to the source .
this ratio is a factor of 3 higher at 85ghz relative to the fraction in the range @xmath96ghz , over which it is constant within the errors .
it is hard to account for such an increase in magnetic field inhomogeneity at 85ghz , particularly if the source size only scales @xmath97 , as expected if its emission is self - absorbed . on the other hand , including synchrotron self - absorption effects , goldston , quataert & igumenshchev ( 2005 ) show that the polarization fraction is expected to increase by a factor of three over the range @xmath98ghz . the quasi - spherical accretion polarization model of melia ,
liu & coker ( 2000 ) and the two - component model of agol ( 2000 ) predict a 90@xmath3 p.a .
flip at @xmath99ghz which is at variance with the measured p.a.s at 150 , 230 and 340ghz . .
previous 86 and 112ghz non - detections are marked with arrows .
the error bars in the aitken et al.(2001 ) measurements , marked with triangles , reflect uncertainty in the contribution from dust emission rather than variability associated with the source . ]
we have reported here the detection of linear polarization in sgr a * at 3.5 mm .
this enables us to calculate the rotation measure and set a limit on the accretion rate .
the lack of frequency dependence for position angle fluctuations indicates that they are intrinsic to the source .
our result favors riaf / cdaf accretion models , with a shallow density distribution , over adaf and bondi - hoyle accretion flows , which have a steep profile and are more likely to produce rapid rm variations .
future wide bandwidth simultaneous observations with carma and the sma will fully characterize intrinsic and extrinsic changes in the polarization properties of sgr a * and allow us to investigate the accretion environments of other nearby low luminosity agn , such as m81 * ( brunthaler , bower & falcke 2006 ) .
agol , e. 2000 , , 538 , l121 aitken , d.k . ,
greaves , j. , chrysostomou , a. , jenness , t. , holland , w. , hough , j.h .
, pierce - price , d. & richer , j. 2001 , , 534 , l176 bower , g.c . , backer , d.c . , zhao , j .- h . , goss , w.m . & falcke , h. 1999a , , 521 , 582 bower , g.c . ,
falcke , h. , herrnstein , r.m . ,
zhao , j .- h . , goss , w. m. & backer , d.c .
2004 , science , 304 , 704 bower , g.c . , falcke , h. , wright , m.c.h . & backer , d.c .
2005 , , 618 , l29 bower , g.c . ,
wright , m.c.h .
, backer , d.c . & falcke , h. 1999b , , 527 , 851 bower , g.c . , wright , m.c.h .
, falcke , h. & backer , d.c .
2001 , , 555 , l103 bower , g.c . ,
wright , m.c.h . ,
falcke , h. & backer , d.c .
2003 , , 588 , 331 bower , g.c . ,
wright , m.c.h . & forster , j.r .
2002 , `` polarization stability of the bima array at 1.3 mm '' , bima memos # 89 brunthaler , a. , bower , g.c . &
falcke , h. 2006 , , 451 , 845 falcke , h. , mannheim , k. & biermann , p.l .
1993 , , 278 , l1 falcke , h. , goss , w.m . ,
matsuo , h. , teuben , p. , zhao , j .- h . &
zylka , r. 1998 , , 499 , 731 ghez , a.m. , klein , b.l . , morris , m. & becklin , e.e .
1998 , , 509 , 678 goldston , j.e . , quataert , e. & igumenshchev , i.v . 2005 , , 621 , 785 marrone , d.p . ,
moran , j.m . ,
zhao , j .- h .
& rao , r. 2006 , , 640 , 308 melia , f. , falcke , h. 2001 , , 39 , 309 melia , f. , liu , s. & coker , r. 2000 , , 545 , l117 melia , f. , liu , s. & coker , r. 2001 , , 553 , 146 quataert , e. & gruzinov , a. 2000 , , 545 , 842 serabyn , e. , carlstrom , j. , lay , o. , lis , d.c . ,
hunter , t.r . ,
lacy , j.h . & hills , r.e .
1997 , , 490 , l77 shen , z .- q .
, lo , k.y . , liang , m .- c .
zhao , j .- h .
2005 , , 438 , 62 tsuboi , m. , miyazaki , a. & tsutsumi , t. 1999 , in the central parsecs of the galaxy , asp conference series , vol . 186 .
h. falcke , a. cotera , w.j .
duschl , f. melia & m.j .
rieke , 105 wright , m.c.h . & backer , d.c .
1993 , , 417 , 560 yuan , f. , quataert , e. & narayan , r. 2003 , , 598 , 301 zhao , j .- h . ,
young , k.h . ,
herrnstein , r.m .
, ho , p.t.p . , tsutsumi , t. , lo , k.y . ,
goss , w.m .
& bower , g.c .
2003 , , 586 , l29 zhao , j .- h .
, herrnstein .
r.m . , bower , g.c . ,
goss , w.m .
& lui , s.m .
2004 , , 603 , l85
|
we report the detection of variable linear polarization from sgr a * at a wavelength of @xmath0 mm , the longest wavelength yet at which a detection has been made .
the mean polarization is @xmath1% at a position angle of @xmath2 with rms scatters of 0.4% and 9@xmath3 over the five epochs .
we also detect polarization variability on a timescale of days .
combined with previous detections over the range 150400ghz ( 7502000@xmath4 m ) , the average polarization position angles are all found to be consistent with a rotation measure of @xmath5radm@xmath6 .
this implies that the faraday rotation occurs external to the polarized source at all wavelengths .
this implies an accretion rate @xmath7yr@xmath8 for the accretion density profiles expected of adaf , jet and cdaf models and assuming that the region at which electrons in the accretion flow become relativistic is within @xmath9 .
the inferred accretion rate is inconsistent with adaf / bondi accretion .
the stability of the mean polarization position angle between disparate polarization observations over the frequency range limits fluctuations in the accretion rate to less than @xmath10% .
the flat frequency dependence of the inter - day polarization position angle variations also makes them difficult to attribute to rotation measure fluctuations , and suggests that both the magnitude and position angle variations are intrinsic to the emission .
|
a giant elliptical galaxy m87 ( ngc 4486 ) , located at 18.4 mpc , hosts one of the nearest and best - studied agn .
a relativistic one - sided jet is ejected from the galactic center , where a supermassive black hole ( smbh ) with a mass of @xmath1 @xcite resides , and extends to hundreds of kilo parsecs .
it was reported in @xcite that the luminous center of m87 and the agn are displaced significantly .
the displacement was confirmed in @xcite and this fact suggests that the smbh is not located at the center of mass of the galaxy , although there is a large uncertainty in the projected distance ( @xmath2 ) .
several scenarios have been suggested to interpret this displacement : ( 1 ) accelaration by an asymmetric jet , ( 2 ) recoil due to gravitational - wave emission caused by a merger of two smbhs , ( 3 ) binary motion with respect to another smbh and ( 4 ) gravitational perturbations from multiple massive objects at the galactic center . as a possible method to distinguish between these scenarios , observation of the proper motion of the smbh ( agn ) was suggested .
here we focus on the latter two scenarios
. binary formation of two smbhs through galaxy mergers is a basic process of smbh evolution .
the observed displacement can be explained if the center of mass of a smbh binary is located at the galactic center and only one of the smbhs has enough gas accretion to make it an agn . on the other hand ,
there is a possibility that multiple intermediate mass bhs ( imbhs ) or relatively small smbhs reside at the galactic center through hierarchical galaxy merging @xcite .
the displacement can also be explained by gravitational perturbation of the main smbh from these multiple bhs . in this latter scenario ,
the main smbh can form a binary with one of the imbhs or smbhs , and the center of mass will not be necessarily located at the galactic center .
if the observed displacement is caused by these mechanisms , a substantial amount of gravitational waves ( gws ) is emitted by the smbh binary .
thus , observation of gws from m87 can probe the black hole population at the galactic center . in this paper , following these scenarios , we estimate the gws under the constraints on the binary orbital elements from observations and discuss the detectability of gws of pulsar timing arrays .
this paper is organized as follows . in section 2 ,
we summarize the basic formulae of keplerian motion . in section 3
, we describe the formalism to calculate the amplitude of gws from a binary and its time derivative . in section 4
, we present our main results . in section 5
, we discuss the detectability of the gws and give a summary and discussion in section 6 . for the rest of this paper we set @xmath3 , unless otherwise specified .
the distance between the main smbh and the barycenter of its host galaxy , @xmath4 , is related to the projected displacement , @xmath5 , as , @xmath6 where @xmath7 is the angle between the line of sight and the direction from the smbh to the barycenter . in the following ,
we discuss the orbital elements of the smbh binary .
the orbital motion is approximated to the newtonian two - body problem since the binary has a typical separation of @xmath8 so that gravity is sufficiently weak and the energy loss due to gravitational waves is negligible .
the equation of motion for the relative motion of the smbhs is given by @xmath9 where @xmath10 , @xmath11 , @xmath12 is the reduced mass , and @xmath13 and @xmath14 are the mass of the main and second smbh , respectively .
the orbital elements are described by @xmath15 where @xmath16 , @xmath17 is the eccentricity ( @xmath18 ) , @xmath19 is the semi - major axis , @xmath20 is the orbital phase and @xmath21 . fixing the eccentricity ,
the semi - major axis is maximum ( minimum ) when the smbh is located at the pericenter ( apocenter ) of the orbit .
therefore , the semi - major axis is constrained as @xmath22 on the other hand , the orbital period is written by @xmath23 in the current case , the typical orbital period is @xmath24 yrs for @xmath25 and @xmath26 .
gravitational waves are emitted from a binary and the waveforms of the plus and cross polarizations are given by , @xmath27 here , @xmath28 and @xmath29 are the unit vectors of the gravitational wave principal axes where @xmath30 is the inclination and @xmath31 is the transverse - traceless part of the amplitude given by @xcite @xmath32 where @xmath33 and @xmath34
( [ eq : hplus ] ) and ( [ eq : hcross ] ) can be rewritten as , @xmath35 + \biggl[\frac{m}{r } - r^2\dot{\phi}^2 -\dot{r}^2\biggr]{\rm{sin}}^2\iota\biggr\ } \\
h_{\times}(t ) & = & \frac{2\mu}{r}{\rm{cos}}\,\iota \nonumber \\ & & \times \left\ { \left ( \frac{m}{r}+r^2\dot{\phi}^2 - \dot{r}^2 \right ) \sin{2\phi } - 2 r \dot{r } \dot{\phi } \cos{2 \phi } \right\ } , \end{aligned}\ ] ] where @xmath36 is the distance to m87 . assuming the photo - center of the galaxy is on the orbital plane , the inclination is constrained as @xmath37 in the case where the main smbh is located in front of ( @xmath38 ) and behind ( @xmath39 ) the photo - center , respectively . in the current situation ,
the period of gravitational waves is much longer than the observation time span of pulsar timing arrays ( @xmath40 ) so that the amplitude varies linearly with time on this time - scale .
thus , we consider the change in the amplitude in the observational time span , rather than the amplitude itself . to do this , we need the time derivative of the waveforms : @xmath41 - \frac{m}{r^2}\dot{r}\ { \rm{sin}}^2\iota\biggr\ } \label{eq : dhdt_plus } \\
\frac{dh_{\times}(t)}{dt } & = & \frac{2 \mu}{r } \cos{\iota } \left\ { 4 \left(\frac{m}{r } - \dot{r}^2\right ) \dot{\phi } \cos{2\phi } + \frac{m}{r^2 } \dot{r } \sin{2 \phi } \right\ } .
\label{eq : dhdt_cross}\end{aligned}\ ] ] these time derivatives can be expressed in terms of @xmath4 , @xmath17 , @xmath42 and @xmath14 by substituting eqs .
( [ eq : r ] ) and ( [ eq : dphidt ] ) to eqs .
( [ eq : dhdt_plus ] ) and ( [ eq : dhdt_cross ] ) .
first , we show the results for the case where the center of mass of two smbhs is located at the photo - center of m87 . fig .
[ fig : contour ] represents the absolute value of the time derivatives of the plus - mode and cross - mode amplitudes in @xmath43 plane for @xmath44 and @xmath45 .
hereafter , the inclination is set to @xmath46 except for fig . [
fig : inclination ] .
the time derivative is of order @xmath47 irrespective of the values of @xmath42 and @xmath17 , although the ratio of the two polarization modes depends significantly on the phase of the orbital motion @xmath20 , which is uniquely determined by @xmath42 and @xmath17 .
[ fig : mass ] shows the dependence of the time derivatives on the mass of the second smbh for @xmath44 , @xmath48 and @xmath49 .
the dependence is strong ( @xmath50 ) for @xmath51 and becomes relatively weak ( @xmath52 ) for @xmath53 .
next , considering that the observed displacement has a large uncertainty , we show the dependence of the same quantity as fig .
[ fig : mass ] on the displacement in fig .
[ fig : displacement ] .
the mass of the second smbh is set to @xmath54 and @xmath55 . from the figure
, we see that the change of gw amplitude increases drastically as the displacement decreases .
the current uncertainty in the displacement leads to the uncertainty in the change of about 4 orders of magnitude . here
it should be noted that the correspondence between the displacement and the semi - major axis of the main smbh , which is shown in the upper axis , depends on the eccentricity , which is set to 0.8 in this figure .
[ fig : inclination ] shows the inclination dependence fixing all the other parameters .
the inclination does not affect the overall amplitude so much , although the derivative of the cross - mode amplitude becomes zero at @xmath56 .
finally , we consider the second scenario where multiple bhs reside in the central region of m87 and the main smbh is forming a binary with a smaller bh . as remarked in section 1 , the center of mass of the binary need not be located at the galactic center and the binary can have a much smaller semi - major axis compared with the previous cases . because there is no constraint on the orbital elements ,
we compute the time derivative of gw amplitude for a number of random sets of orbital elements ( semi - major axis , eccentricity , phase and inclination ) , fixing the mass of the smaller bh .
the results are shown as a function of the semi - major axis in fig .
[ fig : random ] for @xmath57 and @xmath58 .
aside from @xmath14 and semi - major axis , the time derivative of gw amplitude has a strong dependence on the eccentricity and increases significantly for @xmath59 .
in this section , let us discuss the detectability of the gws from a possible smbh binary at the center of m87 .
nano - hertz gws from smbh binaries are targets of so - called pulsar timing arrays @xcite .
because the arrival times of pulses are modulated by gws , the time variation of timing residuals , the difference between the actual arrival time and the expectation in the absence of gws , follows the wave form of the gws crossing the earth .
thus , precise measurements of timing residuals can detect gws with a long period comparable to the time span of observation , typically @xmath40 . in the current case with the linearly - changing gws ,
the situation is totally different as we see below .
the timing residual as a function of time , @xmath60 , induced by the gws is generally described as , @xmath61 where @xmath62 is the direction of the gw propagation , @xmath63 is the unit vector of the direction of the pulsar , and @xmath64 is the difference between the metric perturbation at the earth and pulsar generated by gws : @xmath65 where @xmath66 and @xmath67 is the distance to the pulsar .
below , we consider only the first term , so called earth term , and neglect the second term , pulsar term , following previous works . assuming that the gws are linearly changing ,
that is , @xmath68 are constant for an observation period , we have , @xmath69 where @xmath70 are the polarization tensors . then , eq . ( [ eq : resh1 ] ) is written as , @xmath71 where @xmath72 is the antenna beam pattern and given by @xmath73 .\ ] ] thus , linearly changing gws cause timing residuals quadratic in time .
however , this is exactly the same behavior as the one due to the spin down of the pulsar , @xmath74 where @xmath75 is the pulse period and @xmath76 is the spin - down rate .
therefore , the linearly changing gws are absorbed by the spin - down rate in the fitting of timing residual data with a timing model .
in other words , the value of @xmath76 is biased by @xmath77 , and the effects of gws are absorbed as @xmath78 .
it should be noted that @xmath79 can be positive or negative depending on the position of the pulsar and the gw polarization , while it is natural to consider an intrinsic @xmath76 should be positive .
however , some of the observed pulsars have a negative @xmath76 value due to various effects such as acceleration of the pulsar along the line of sight , the differential galactic rotation and the shklovskii effect @xcite .
thus , we need to investigate the statistical feature of @xmath76 of many pulsars , rather than that of individual pulsars , in order to probe the existence of linearly - changing gws . here
, we propose the distribution function of pulsar spin - down rate as a more robust method to detect linearly changing gws . as noted above , the value of the bias @xmath80 depends on the pulsar position and the gw polarization .
an example is shown in fig .
[ fig : hdot ] for @xmath81 and @xmath82 , which shows the antenna beam pattern in the sky for the cross mode . as can be seen
, the positive and negative areas appear as a quadrupole pattern .
thus , if we make histograms of @xmath76 of the pulsars in the positive and negative regions of the sky separately , a systematic difference between the two histograms would be seen for @xmath83 . practically , because we do not know the polarization of the gws from m87 , we need to compare two histograms assuming any possible polarization and dividing the sky accordingly .
the typical value of @xmath76 of the millisecond pulsars which have been found so far is about @xmath84 and @xmath85 can be as low as @xmath86 @xcite . therefore , linearly changing gws with @xmath87 could be , in principle , detected by the above method if we have a sufficiently large sample of millisecond pulsars .
according to our calculation above , this level of gws would not be generated in the first scenario ( see fig . [
fig : displacement ] ) , unless the second black - hole mass is extremely large ( @xmath88 ) . on the other hand
, there is a small chance for the second scenario ( see fig .
[ fig : random ] ) , if the second bh mass is as large as @xmath89 , the semi - major axis is relatively small @xmath90 and the eccentricity is large . in practice , due to the limited number of millisecond pulsars ,
the statistical errors in the histograms diminish the sensitivity of this method .
the number of known millisecond pulsars is approximately 200 ( the atnf pulsar catalogue ) and will reach 1,400 and 3,000 by the planned ska1 and ska2 surveys , respectively @xcite .
it is beyond the scope of the current paper to estimate the sensitivity quantitatively taking the statistical errors and other observational systematics into account .
let us comment on another aspect of this method which is significantly different from the conventional pulsar timing arrays .
pulsar timing arrays need a long - term observation of order @xmath40 to follow the wave form of nano - hertz gws . therefore , millisecond pulsars which have very stable periods through this time scale are necessary and only @xmath91 of millisecond pulsars satisfy this requirement . on the other hand , in the current case , we do not need such a long - term observation for two reasons .
one is that we do not need to follow the wave form of the gws and another is that the relative importance of linearly changing gws and pulsar spin down in the timing residuals does not depend on the observation period ( see eqs .
( [ eq : resh2 ] ) and ( [ eq : resp ] ) ) .
although the pulsars still need to be stable enough for a period of a few years in order to determine @xmath76 very precisely , the requirement on pulsar stability would be less strict .
thus , we expect that more millisecond pulsars could be used for this purpose compared with the conventional pulsar timing array , although it should be noted that only @xmath92 of millisecond pulsars have @xmath93 and the increase would not be drastic .
in this paper , we estimated gravitational waves from a hypothetical smbh binary at the center of m87 .
this follows the scenario that the observed displacement between the agn and the galactic center can be explained by the existence of another smbh , or multiple imbhs or small smbhs .
because the period of the binary and the resulting gws is much longer than observational time span , the amplitude changes linearly with time and we calculated the time derivatives of the gw amplitudes .
we investigated their dependence on the orbital elements and the second bh mass taking the observational constraints into account .
there would be no chance to detect gws from a smbh binary which explains the observed offset of the agn with either the conventional pulsar timing arrays or a new method proposed above . in the second scenario where multiple imbhs and small smbhs reside and one of them is forming a binary with the main smbh
, there is a small chance to detect the gws with the new method .
however , the detection of such gws is not a direct evidence of the existence of multiple bhs and , hence , does not give a decisive explanation of the agn offset .
nevertheless , linearly changing gws would still be a useful tool to probe the center of m87 .
if multiple bhs reside at the center of m87 , their gravitational perturbations will affect not only the main smbh but also the distribution and velocity field of stars @xcite .
this could be another probe of the m87 center and will be studied in future .
kt is supported by grand - in - aid from the ministry of education , culture , sports , and science and technology ( mext ) of japan , no .
24340048 , no . 26610048 and no .
the research of js has been supported at iap and at cea saclay by the erc project no .
267117 ( dark ) hosted by universit pierre et marie curie ( upmc ) - paris 6 and at jhu by nsf grant no . oia-1124403 .
|
in this paper , we study gravitational - wave ( gw ) emission from a hypothetical supermassive black - hole ( smbh ) binary at the center of m87 .
the existence of a smbh other than that usually identified with the central agn is a possible explanation for the observed displacement ( @xmath0 ) between the agn and the galactic centroid , and it is reasonable to assume considering the evolution of smbhs through galaxy mergers . because the period of the binary and the resulting gws is much longer than the observational time span
, we calculate the variation of the gw amplitude , rather than the amplitude itself .
we investigate the dependence on the orbital elements and the second bh mass taking the observational constraints into account .
the frequency of the gws is too low to be detected with the conventional pulsar timing array and we propose a new method to detect such low - frequency gws with the distribution function of pulsar spin - down rates .
although the gws from a smbh binary which explains the observed displacement is extremely hard to be detected even with the new method , gws are still a useful way to probe the m87 center .
|
owing to their unique properties , alkaline - earth - like atoms have become the subject of intense experimental and theoretical research in recent years .
ultracold gases of these two - electron atoms have been proposed for the realization and study of a variety of novel quantum systems including heavy fermion materials and the kondo insulator@xcite , su(n)-symmetric hamiltonians@xcite , quantum information processing@xcite and artificial gauge fields@xcite .
quantum degenerate gases have been reported by only a few groups for isotopes of ytterbium@xcite ( yb ) , strontium@xcite , and calcium@xcite .
however , all of these experiments rely on zeeman slowers as the initial cooling stage . from the early days of laser cooling@xcite
it is well known that zeeman@xcite or chirp@xcite slowing of a thermal beam enhances the loading rate of a magneto - optical trap ( mot ) by many orders of magnitude as compared to loading from a background vapor directly .
for alkali metals the concept of a 2d - mot@xcite has been developed as a powerful alternative which results in a similar flux of cold atoms , but in a much more compact and versatile setup . here ,
atoms are transversely cooled from a background vapor into a well - collimated beam of cold atoms . because longitudinally slow atoms are cooled preferentially , the resulting velocity distribution peaks at several tens of meters per second allowing them to be captured by the 3d - mot
. often , a `` pushing '' beam or moving molasses is used to enhance the yield of the cold atom source further . 2d - mots are extremely efficient , yet uniquely simple and highly tunable .
in particular , a single setup can readily be used to cool multiple elements , whereas multispecies zeeman slowers require sophisticated designs .
2d - mots are therefore inherently well suited for producing and studying ultracold mixtures.@xcite furthermore , high - precision spectroscopy experiments , which are prone to frequency shifts from black - body radiation , can benefit from the fact that there is no direct line of sight from the source of thermal atoms to the cold or ultracold sample .
the traditional concept of a 2d - mot is limited to elements with sufficiently large vapor pressures near room temperature .
alkaline - earth - like elements , except for mercury@xcite , have high melting temperatures @xmath2 ( e.g. @xmath3 for yb ) and low vapor pressures ( e.g. less than @xmath4 at @xmath5 for yb@xcite ) . at room temperature
this effectively prevents any 2d - mot from being loaded from a background gas .
an alternative concept has been demonstrated for lithium@xcite , where the 2d - mot is side - loaded from an oven instead of background vapor and high loading rates of up to @xmath6 have been reported . in this paper
we report on the first successful implementation of a 2d - mot loaded via a dispenser for yb and the subsequent generation of quantum degenerate gases of both bosonic @xmath7 and fermionic @xmath8 . based on numerical simulations
, we have designed a very compact `` glass - cell '' setup for a 2d- and 3d - mot that allows extraordinarily large optical access for future experiments based on optical lattices .
state is given for the fermionic isotope @xmath8 . ]
yb is a rare - earth element with as many as seven stable isotopes .
its five bosonic isotopes ( @xmath9 , @xmath10 , @xmath11 , @xmath7 , @xmath12 ) are spinless ( @xmath13 ) , while the remaining two fermionic isotopes have nuclear spins of @xmath14 ( @xmath15 ) and @xmath16 ( @xmath8 ) .
the electronic structure of yb is dominated by its two valence electrons and hence very similar to alkaline - earth elements . as shown in fig .
[ fig : levelscheme ] , its level scheme consists of states with singlet and triplet electronic spin.@xcite the intercombination transitions connecting singlet and triplet states are only allowed due to imperfect @xmath17-@xmath18-coupling and thus are very weak . in particular , the transition from the ground state @xmath19 to the triplet state @xmath20 is even forbidden as a single - photon transition for the bosonic isotopes , although small transition amplitudes may be created artificially by magnetically - induced state mixing@xcite or via multi - photon schemes@xcite . in fermionic isotopes ,
hyperfine interaction yields a non - zero natural linewidth of the transition on the order of @xmath21 . with a lifetime of several seconds@xcite ,
it is metastable on experimentally relevant timescales and gives rise to a series of unique and intriguing features.@xcite the ultranarrow @xmath22 transition is not only crucial for novel optical frequency standards@xcite , its fascinating properties also make it ideal for the spectroscopic analysis of quantum gases .
energy levels of atoms confined to a site of an optical lattice can be probed locally in the optical domain , and due to the coherent nature of the probe even controlled manipulation of quantum states is possible .
the wavelength of the lattice acts as a tool to tune the differential ac - stark shifts of the two states of the spectroscopy transition . in this way
, different scenarios can be realized , e.g. a `` magic '' lattice@xcite where the main order differential light - shift induced by the trapping potential on the transition is cancelled , thus enabling high - precision spectroscopy .
moreover , state - selective lattices at wavelengths where the polarizability of either @xmath19 or @xmath20 vanishes have been proposed@xcite and allow an unprecedented degree of control of the system . for laser cooling ,
two complementary transitions are of interest .
the broad principal transition @xmath0 allows an extremely strong radiation pressure . in bosonic isotopes
, the lack of substructure in the ground state @xmath19 prevents any sisyphus cooling on this transition.@xcite on the other hand , the intercombination transition @xmath1 is very narrow and has a _ doppler cooling _ limit of merely @xmath23 , which provides ideal starting conditions for evaporative cooling .
since the @xmath19 ground state does not allow magnetic trapping , evaporation and further experimental steps require optical dipole traps . in combination with optical lattices ,
a setup with wide optical access is therefore desirable for yb experiments .
3d - mot as a function of its capture velocity .
results of numerical simulations are shown for magnetic field gradients of @xmath24 ( @xmath25 ) , @xmath26 ( @xmath27 ) and @xmath28 ( @xmath29 ) at detunings of @xmath30 , @xmath31 and @xmath32 , respectively , yielding optimal maximum loading rates . ] with two nearly closed cooling transitions , yb is a straightforward candidate for a dual mot scheme .
the principal transition @xmath0 is used for cooling in the 2d - mot to achieve a strong cooling force , while the subsequent 3d - mot operates on the intercombination transition @xmath1 that allows for a final temperature in the microkelvin range .
the central concept of our 2d-/3d - mot setup to overcome the low vapor pressure of yb is to use a side - loaded 2d - mot@xcite , but use dispensers instead of an oven . as shown in fig .
[ fig:2d3dsketch ] , the 2d - mot is formed by a perpendicular pair of counterpropagating , circularly polarized laser beams and a 2d magnetic quadrupole field generated by rectangular coils . it is loaded transversely from a dispenser that has been installed within a glass cell . using the narrow intercombination transition ,
however , the velocity capture range of the 3d - mot typically is only a few meters per second . as detailed further in section [ section:2d3dmot ]
, we use laser broadening@xcite to enhance the loading rate substantially . to find the optimal 2d - mot parameters we have performed detailed numerical simulations . using the geometry as described above , except for the absence of a pushing beam , we numerically integrated the possible trajectories of yb atoms emitted from the dispenser for different velocities assuming the mean radiation pressure in the 2d - mot .
we assume that yb atoms emitted from the dispenser stick to the glass cell at their first contact , which is justified by the high desorption temperature .
thus , atoms perform a `` single pass '' through the 2d - mot , in strong contrast to 2d - mots loading from a thermal background vapor .
spontaneous emissions and collisions have been ignored in this calculation .
we sampled the velocity distribution in steps of @xmath33 per axis up to a maximum of @xmath34 and weighted each trajectory with the flux of the source in the associated velocity bin assuming a thermal maxwell - boltzmann distribution .
the 2d - mot beams have an elliptic cross - section of @xmath35 in diameter and a homogeneous intensity of @xmath36 where @xmath37 is the saturation intensity of the transition @xmath0 . from these results
we then calculated the rate of atoms entering the 3d - mot volume at a velocity smaller than the capture velocity .
examples of simulation results for a capture velocity @xmath38 are shown in fig .
[ fig:2dmot_sim_exp]a . evidently , the 2d - mot efficiently decelerates and cools a substantial amount of atoms emitted by the dispenser , and the resulting beam of atoms is indeed slow enough to subsequently be captured by the 3d - mot .
the loading rate exhibits a distinct maximum at a magnetic quadrupole field gradient of @xmath26 and a detuning @xmath39 .
as expected due to the large linewidth of the cooling transition , the optimal gradient is significantly higher than for typical alkali 2d - mots .
experimentally obtained loading rates of our 2d-/3d - mot system for the same range of magnetic field gradients and detunings are shown in fig .
[ fig:2dmot_sim_exp]b , and they are in excellent agreement with the simulation . in particular , the optimal set of parameters is reproduced exactly within the resolution of the scans .
[ fig : sim_capturevelocity ] shows the expected loading rates as a function of the capture velocity for several 2d - mot configurations .
it illustrates a strong dependence of the optimal 2d - mot parameters on the capture velocity around @xmath40 and a saturation of the loading rate at values beyond some @xmath41 .
the vacuum system of our experimental apparatus consists of two separate glass cells that are mounted to a single stainless steel vacuum chamber . fig . [
fig : vacuumsystem ] shows a cad section view of the apparatus .
both cells are made from synthetic fused silica glass , and broadband anti - reflective coatings have been applied to their outer surfaces . while the 2d - mot is produced in the upper cell , the 3d - mot is located in the `` science '' cell below , and all subsequent experimental steps take place in this cell as well . the central chamber connects the 2d - mot cell to the science cell via two differential pumping stages , equipped with graphite tubes , along its central axis .
these are necessary to maintain the low pressure required to achieve sufficiently long lifetimes in the science cell ( @xmath42 ) even when the 2d - mot is in operation ( @xmath43 ) .
each cell is pumped by a @xmath44 ion getter pump , and an additional titanium sublimation pump has been installed for the science cell .
the chamber itself has been manufactured from solid , non - ferromagnetic steel with internal tubes to connect the pumps and other equipment to the cells .
its compact designs reduces the distance between both mots to @xmath45 , which is essential for an efficient transfer of atoms from the 2d - mot into the 3d - mot .
the 2d - mot cell ( @xmath46 ) contains two alvasource@xcite as-4-yb-500-s yb dispenser sources that generate the atomic beams loading the 2d - mot when heated .
both dispensers are connected to a high - power feedthrough and have been mounted upright on an insulating ceramic ring with the emitter slit facing towards the center of the cell . to prevent the yb atoms from coating any critical surface of the 2d - mot cell ,
aluminium plates with rectangular apertures have been installed in front of each dispenser .
we typically operate the dispensers at currents of about @xmath47 , which yields a sufficiently large flux of yb . during operation , the pressure measured close to the 2d - mot cell rises from some @xmath48 to a few @xmath49 , likely due to outgassing of the dispensers , as emitted yb directly sticks to the walls .
two identical pairs of external rectangular coils generate the magnetic quadrupole field for the 2d - mot .
the coils are water - cooled and produce the necessary gradients of more than @xmath50 without radiating a significant amount of heat .
the science cell ( @xmath51 ) provides excellent optical access to the atomic samples .
a pair of helmholtz coils generates the magnetic quadrupole field for the 3d - mot , but it has been designed for large a variety of tasks , such as magnetic traps for another species or bias fields for feshbach resonances in mixtures .
it is capable of producing gradients of up to @xmath52 or homogeneous fields of up to @xmath53 and does not reduce the optical access significantly .
the 2d - mot is formed by two elliptical beams of about @xmath54 by @xmath55 in diameter that are retroreflected in a cat s eye configuration .
we use a commercial diode laser system with a tapered amplifier that is frequency - doubled in a bow - tie cavity and provides about @xmath56 of laser power ( @xmath57 ) to each beam at a wavelength of @xmath58 .
the laser is offset - locked to a master ecdl using an additional @xmath59 laser diode and can easily be scanned across a broad frequency range .
the master laser itself is stabilized by fluorescence spectroscopy of an atomic beam .
additional light for pushing and imaging beams is generated by an injection - locked slave laser diode of the same type .
the 3d - mot has been set up in a retroreflected three - beam - configuration ( fig .
[ fig:2d3dsketch ] ) .
a commercial fiber laser is frequency - doubled in a waveguide and provides @xmath60 ( @xmath61 ) and @xmath62 ( @xmath63 ) for the axial and diagonal beams with a diameter of @xmath64 .
this laser is stabilized via doppler - free fluorescence spectroscopy of the atomic beam mentioned above , and its inherent narrow linewidth of less than @xmath65 is particularly useful for operating a mot at small detunings .
measured as a function of the fwhm @xmath66 of the broadened laser spectrum .
the high frequency edge of the spectrum is kept at a constant detuning from the @xmath1 resonance frequency . ]
fluorescence in the 2d - mot cell indicates that atoms are captured in a cigar - shaped region along the axis of the magnetic field .
however , the emerging flux of cold atoms can not be captured efficiently in a single - frequency 3d - mot setup operated close to the @xmath1 intercombination transition , because the narrow linewidth addresses only a very small velocity range .
the capture velocity can be increased by actively broadening the frequency spectrum of the laser , as has already been reported for yb setups using zeeman slowers@xcite .
this temporarily introduces additional frequency components , preferably further detuned to the red , at the expense of temperature , as detailed in section [ section : motperformance ] . due to the low saturation intensity of the transition ,
there is typically enough laser power to cover a broad frequency range at a reasonable power spectral density .
we use the rf signal driving the aom that controls the power in the 3d - mot beams to modulate the laser frequency across several megahertz at a modulation frequency of typically @xmath67 . to keep the minimal red detuning roughly at @xmath68
, we also increase the center detuning accordingly .
[ fig : motbroadening ] shows the mot loading rate as a function of the fwhm @xmath66 of the broadened laser spectrum for the bosonic isotope @xmath7 ; we attribute the reduced loading rate above @xmath69 to insufficient spectral power density . for optimal loading from the 2d - mot
, we typically operate the 3d - mot at a center detuning of @xmath70 , a broadening of @xmath71 , and a magnetic field gradient of @xmath72 . as shown in fig .
[ fig:2dmot_sim_exp]b , the 2d - mot is operated at a magnetic field gradient of @xmath26 and a detuning of @xmath73 .
because of the large linewidth of the @xmath0 transition used for the 2d - mot , higher laser power may still increase attainable loading rates significantly , however our experimental setup is currently limited by available laser power and power dissipation in the coils ( ca .
@xmath74 ) . when reducing the laser power for the 2d - mot by a factor of @xmath75 , the loading rate of the 3d - mot is decreased by a factor of @xmath76 . measured as a function of power in the pushing beam .
loading rates are normalized to the value without a pushing beam . ]
the transfer of atoms from the 2d - mot into the 3d - mot can be enhanced significantly by a weak additional beam along the axis of the 2d - mot pushing atoms through the differential pumping stage .
we have implemented a single pushing beam as shown schematically in fig .
[ fig:2d3dsketch ] at a detuning of @xmath77 from the @xmath0 transition . in fig .
[ fig : pushingbeam ] , the loading rate of the 3d - mot is shown as a function of power in the pushing beam .
clearly , a power around @xmath78 yields an optimal and robust loading rate the optimal power depends on its size and the details of its alignment , however , and is given here for reference only . because of the large linewidth of the @xmath0 transition , the radiation pressure of the weak pushing beam exceeds that of the 3d - mot many times .
we therefore align the pushing beam such that it is blocked by the differential pumping tube and does not enter the science cell . for the parameters discussed above , we found a typical value for the 3d - mot steady - state particle number of @xmath79 for @xmath80 , and we typically load @xmath81 bosonic @xmath7 atoms within @xmath82 . at its optimal loading parameters ,
the 3d - mot has a temperature of @xmath83 . to reduce the temperature of the sample as close to the doppler limit of @xmath84 as possible , the laser - broadening amplitude
is gradually reduced to zero within several hundreds of milliseconds , while the center detuning is simultaneously shifted closer to resonance , and the power is reduced to @xmath85 and @xmath86 in the diagonal beams and beams along the symmetry axis of the coils , respectively .
furthermore , the mot is compressed by increasing the field gradient to about @xmath87 . with these parameters ,
the mot temperature is reduced to typically @xmath88 , comparable to temperatures published for other yb mots@xcite .
for evaporative cooling and trapping of ultracold samples , we have set up a crossed optical dipole trap ( odt ) formed by a deep horizontal trap and a vertical trap that creates additional confinement in the crossing region .
the horizontal trap is created by a beam of up to @xmath89 produced by a solid state laser at @xmath90 and focused down to vertical and horizontal @xmath91-waists of @xmath92 and @xmath93 , respectively . for the vertical trap ,
part of the laser power is split off by a partially reflecting mirror and typically up to @xmath94 are focused to waists of @xmath95 and @xmath96 .
the powers in each beam are monitored by logarithmically amplified photodiodes and regulated using quartz aoms that also induce a frequency difference of about @xmath97 between the two beams . to transfer atoms into the odt
once the mot has been loaded , we abruptly switch on the dipole traps and gradually increase the magnetic field gradient to @xmath87 to strongly compress the mot , while its temperature is reduced as discussed above .
both dipole traps are initially operated at their full powers of @xmath98 and @xmath99 , corresponding to potential depths of @xmath100 and @xmath101 for the horizontal and vertical traps , respectively .
@xmath7 atoms are trapped in the horizontal odt at the same temperature measured for the single - frequency mot .
the transfer efficiency strongly depends on the final mot parameters , in particular the magnetic field gradient .
evaporative cooling is performed in two stages by lowering the depths of the odt . in the first stage , atoms are predominantly evaporated from the horizontal trap , but concentrated into the crossed region of the dipole trap as the temperature decreases .
therefor we exponentially ramp the horizontal trap down to @xmath103 ( @xmath104 ) within @xmath105 , but keep the vertical trap at constant depth .
subsequently , atoms are evaporated from the crossed region in a second stage .
within @xmath106 we exponentially reduce the depth of the horizontal trap to nominally @xmath107 ( @xmath108 ) and relax confinement further by lowering the power of the vertical trap to some @xmath109 ( @xmath110 ) .
it is crucial that the confinement along the axis of the horizontal trap provided by the vertical beam is lowered during the final evaporation stage .
however , the vertical trap has virtually no effect on the overall trap depth , and its final power can be varied across a large range to control the resulting trap geometry . the absorption images in fig .
[ fig : bec]a , b show the emergence of a bose - einstein condensate of @xmath7 during the final stages of the evaporation ramp .
we routinely produce nearly pure condensates of @xmath111 @xmath7 atoms with no discernible thermal fraction left .
the crossed dipole trap at its final parameters has an effective depth of several tens of nanokelvins and trap frequencies @xmath112 determined from dipole and quadrupole oscillations of the condensate .
we have produced degenerate fermi gases of the spin-@xmath113 isotope @xmath8 using a similar sequence as for @xmath7 , because its @xmath114-wave scattering length@xcite is favorable for evaporative cooling in a spin mixture .
the 2d - mot uses the @xmath0 ( @xmath115 ) hyperfine transition of @xmath8 , but its operation is complicated by the proximity of the @xmath116 hyperfine transition as shown in fig .
[ fig : levelscheme ] .
this results in a strong reduction of the observed 3d - mot loading rate at detunings close to the hyperfine splitting , but efficient operation of a 2d - mot is still possible within a narrow window around the same optimal detuning as observed for bosonic @xmath7 .
even more than in the case of @xmath7 , the loading rate strongly depends on the intensity of the 2d - mot beams , however .
due to its lower abundance and reduced 2d - mot performance , the loading time of the @xmath8 3d - mot is significantly larger than for @xmath7 . with an earlier version ,
i.e. reduced laser powers in the 2d- and 3d - mot beams , we obtained the following results : a spin mixture of about @xmath117 @xmath8 atoms was transferred from a mot of @xmath118 atoms to the dipole trap .
thermalization is faster than in @xmath7 due to the enhanced scattering rate , and we used an evaporation ramp of similar shape as described above , but steeper than for the bosons .
after @xmath119 of evaporation @xmath120 atoms at a temperature of @xmath121 were left . a typical absorption image of a degenerate fermi gas is shown in fig .
[ fig : bec]c .
in conclusion , we have presented a novel scheme for the efficient creation of quantum gases of yb and similar alkaline - earth - like atoms based on transverse loading of a 2d - mot .
loading rates observed in the subsequent 3d - mot are in agreement with numerical simulations of the 2d - mot setup .
the compact 2d-/3d - mot apparatus facilitates laser cooling of additional species to study quantum - degenerate mixtures and provides excellent optical access to the combined 3d - mot and science cell .
therefore , it allows a wide range of experiments to be conducted .
we routinely produce bose - einstein condensates of @xmath7 and degenerate fermi gases of @xmath8 with particle numbers well suited for further experiments .
this work has been supported financially by the deutsche forschungsgesellschaft within the sfb 925 and grk 1355 and the european union within the fet - open scheme ( isense ) .
the authors thank thomas rtzel , hans kessler , and jan carstens for experimental support .
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 link:\doibase 10.1103/physreva.81.051603 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.053624 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.053619 [ * * , ( ) ] \doibase http://dx.doi.org/10.1038/nphys1535 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.170504 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.110503 [ * * , ( ) ] http://stacks.iop.org/1367-2630/12/i=3/a=033007 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.1523 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.011611 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.91.040404 [ * * , ( ) ] link:\doibase 10.1103/physreva.76.051604 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.030401 [ * * , ( ) ] link:\doibase 10.1103/physreva.79.021601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.190401 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.011610 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.011606 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.200401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.200402 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.030402 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.011608 [ * * , ( ) ] link:\doibase 10.1103/physreva.82.041602 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.130401 [ * * , ( ) ] link:\doibase 10.1103/physreva.85.031603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.59.2631 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.48.596 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.54.996 [ * * , ( ) ] link:\doibase 10.1103/physreva.58.3891 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.020401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.183004 [ * * , ( ) ] link:\doibase 10.1063/1.1726358 [ * * , ( ) ] link:\doibase 10.1103/physreva.80.013409 [ * * , ( ) ] link:\doibase 10.1103/physreva.37.119 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.123202 [ * * , ( ) ] link:\doibase 10.1103/physreva.69.042506 [ * * , ( ) ] link:\doibase 10.1103/physreva.72.032506 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.083001 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.173002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.050801 [ * * , ( ) ] link:\doibase 10.1103/physreva.75.020501 [ * * , ( ) ] \doibase http://dx.doi.org/10.1038/nature03541 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.083002 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.050501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.103.063001 [ * * , ( ) ] in @noop _ _ , ( , ) pp .
link:\doibase 10.1103/physreva.60.r745 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.012719 [ * * , ( ) ]
|
we report on the first experimental setup based on a 2d-/3d - mot scheme to create both bose - einstein condensates and degenerate fermi gases of several ytterbium isotopes .
our setup does not require a zeeman slower and offers the flexibility to simultaneously produce ultracold samples of other atomic species .
furthermore , the extraordinary optical access favors future experiments in optical lattices .
a 2d - mot on the strong @xmath0 transition captures ytterbium directly from a dispenser of atoms and loads a 3d - mot on the narrow @xmath1 intercombination transition .
subsequently , atoms are transferred to a crossed optical dipole trap and cooled evaporatively to quantum degeneracy .
|
the most precise electric dipole moment ( edm ) measurement @xcite on the electron gives @xmath2 , a result consistent with zero .
a non - zero result at this level would necessarily mean the violation of time reversal symmetry ( t ) symmetry and the existence of new particle physics , beyond the standard model @xcite .
although the standard model does exhibit t violation through the ckm mechanism , the electron edm produced by this mechanism is roughly ten orders of magnitude below the present experimental sensitivity and is for all practical purposes zero .
by contrast , many modern extensions of particle theory lead quite naturally to a value in the current range of @xmath3 or a little below @xcite . since our experiment using cold ybf molecules aims to be more sensitive than this ,
it is a search for new physics . assuming the validity of cpt
, an electron edm at the @xmath4 level would also imply a new type of cp violation .
this would be of great interest for cosmology as it seems to be a necessary ingredient in understanding the observed matter - antimatter asymmetry of the universe @xcite .
the interaction between @xmath5 and an applied field @xmath6 can be expressed by the effective non - relativistic hamiltonian @xmath7 . for a free electron , @xmath8 and @xmath9
is a unit vector along the spin .
if the electron is part of an atom or molecule , @xmath9 lies along the spin of the system and @xmath10 is a factor that depends on the structure . some heavy atoms and molecules have the virtue that @xmath11 , and
then it is called the enhancement factor @xcite .
this coupling resembles the interaction @xmath12 of the magnetic moment @xmath13 with a magnetic field @xmath14 , where @xmath15 accounts for the atomic or molecular structure .
it is instructive to compare these two interactions in the case of a free electron with an edm of , say , @xmath16 , just below the present limit . in a @xmath17 field
the edm energy is so small that it equals the magnetic energy in a field of only @xmath18 .
controlling the stray magnetic field at that level seems close to impossible , especially when applying the electric field .
heavy atoms such as cs and tl alleviate this problem by their large enhancement factors .
in particular , @xmath19 for the thallium atom @xcite , which relaxes the necessary field control to the challenging , but achievable ft level . two magnetic effects are most troublesome .
( i ) stray magnetic fields vary both in space and time .
( ii ) atoms moving through the large electric field experience a motional magnetic field @xcite @xmath20 . in both cases
the unwanted field components are typically many orders of magnitude larger than 1 ft and heroic efforts were needed to reach the current precision @xcite .
heavy polar molecules offer substantial relief from these difficulties @xcite .
first , the enhancement factors are generically much larger @xcite because the electron edm interacts with the polarisation of the charge cloud close to the heavy nucleus . in an atom this polarisation follows from the mixing of higher electronic states by the applied electric field . in a polar molecule ,
these electronic states are already strongly mixed by the chemical bond and it is only rotational states that have to be mixed by the applied field .
since these are typically a thousand times closer in energy , the molecular enhancement factor is correspondingly larger . for the ybf molecule used in our experiment , the enhancement factor @xcite is @xmath21 at our operating field of @xmath22 , which relaxes the requirement on field control to the @xmath23 level .
figure [ polarisation ] shows the enhancement ( expressed as an effective electric field ) as a function of the laboratory field .
note that the polarisation , due to rotational state mixing , saturates at relatively modest fields . there is a second advantage to ybf .
being polar , this molecule has a strong tensor stark splitting between sublevels of different @xmath24 , the total angular momentum component along the electric field direction .
this strongly suppresses the zeeman shift due to perpendicular magnetic fields , including the motional field @xmath25 . for our typical operating parameters , the motion - induced false edm
is reduced by this mechanism to a level below @xmath26 @xcite , which is negligible .
our experiment uses a cold , pulsed , supersonic beam of ybf radicals @xcite in a magnetically shielded vertical vacuum chamber @xmath27 m high .
the electronic , rotational and vibrational ground state @xmath28 is a hyperfine doublet with states @xmath29 split by 170mhz .
we first deplete the @xmath30 state by laser excitation of the @xmath30 molecules on the @xmath31 transition .
this laser beam is called the pump . a radiofrequency magnetic field , which we call the first beam splitter
, then drives the @xmath32 molecules into a symmetric coherent superposition of the @xmath33 states , as described later in more detail .
next , parallel dc electric and magnetic fields are applied to introduce a phase shift @xmath34 between the two superposed states . here
@xmath35 and @xmath36 appear as functions of time because they are the fields in the molecular rest frame . at time
@xmath37 the molecules interact with a second oscillating field , the recombining beam - splitter , that couples the symmetric part of the @xmath30 coherence back to the @xmath32 state .
the resulting @xmath32 state population exhibits the usual @xmath38 fringes of an interferometer .
we detect the complementary @xmath30 population using fluorescence induced by a probe laser on the @xmath39 transition .
figure[fringes ] shows the interference fringes observed in this fluorescence when the magnetic field is scanned .
the beam splitter is an rf magnetic field perpendicular to @xmath6 and along the beam direction .
when a pulse of molecules arrives at the centre of the rf loop it is subject to a short pulse of resonant @xmath40mhz radiation , which induces hyperfine population transfer .
it is important that the molecules move as little as possible during this transition because unwanted phase shifts can occur if the fields @xmath41 or @xmath42 rotate during the splitting ( or recombining ) transition .
when combined with other imperfections of the apparatus , such a phase can produce a false edm and is therefore undesirable .
figure[rf_lineshape ] shows the lineshape measured using a @xmath43s - long rf pulse , together with a fit to the standard rabi formula@xcite showing good agreement . with high power amplifiers , we are now able work with pulses @xmath44s long , corresponding to a beam movement below 6 mm . the peak transition probability in figure[rf_lineshape ] , taken using the first loop ,
is approximately 0.8 .
this is due to the beam velocity spread , which gives the gas pulse a length of 5 cm at the first loop and 10 cm at the second .
since the rf field strength varies along the beam line , this spatial spread produces a small distribution of rabi frequencies , which could be improved by having a more monoenergetic beam .
the arrival of each ybf pulse at our detector is recorded with @xmath45 resolution .
this time - resolved data gives us a spatial resolution of @xmath46 at the second rf loop .
this is the basis of a useful diagnostic technique : varying the timing of the rf pulse allows us to map out the field distribution in the apparatus .
this works very well for electric fields because the stark shift of the hyperfine transition is large .
it can also be used to probe magnetic fields , albeit with slightly lower spatial resolution .
with the ybf interferometer working as described above , the electron edm measurement is straightforward in principle .
a small magnetic field is applied to bias the interferometer phase to the point of highest slope . the applied electric field
is then reversed and the change in the interferometer output constitutes the edm signal . in practice , we also reverse the direction of the magnetic field and modulate its amplitude .
the edm is correlated with the relative direction of e and b while the signals demodulated on the individual switching channels yield valuable information about the drift of the interferometer setup and some systematic effects .
additional modulations can be applied : the relative phase between the splitter and recombiner rf fields , the laser frequency , the rf amplitudes and the rf frequencies . with four switching channels ( e.g. e reversal , b reversal , relative rf phase and a small step @xmath47 )
there are sixteen possible states of the machine .
we cycle through all of them 256 times to form a `` block '' of data consisting of 4096 individual points .
the repetition rate is determined by the 25hz ablation laser which forms the ybf beam , thus each block of data requires about @xmath48minutes of real time to record .
slow drifts in the signal are cancelled by choosing appropriate modulation waveforms for the switching channels @xcite .
figure [ noise_a ] shows the power spectrum of the noise in our molecular beam signal , as measured by the photomultiplier recording laser - induced fluorescence from the @xmath30 state . above about 10hz
the spectrum is largely flat , as one would expect for shot noise due to the statistics of the photon arrivals , but at lower frequencies @xmath49 noise dominates .
for this reason , the noise we observe in our edm measurement usually exceeds the shot noise level .
when the ablation target is moved to expose a fresh spot to the yag laser the excess is typically a factor of 1.5 , increasing to 2 over the course of an hour or so , at which point we reset the target .
this non - statistical contribution to the noise in the edm measurement generates excess noise in the wings of the normalized distribution shown in figure [ noise_b ] . although it is a small effect ,
this must be taken into account if we are to assign reliable confidence limits to our edm measurement .
we use a novel bootstrap method .
the basic idea is that the experiment itself provides the best available estimate of the true underlying probability distribution @xcite .
each block of data provides a measurement of the edm . by randomly choosing values of the edm from this experimental data set
, we create an ensemble of additional synthetic edm data sets having the same distribution .
note that chosen points are not removed from the pool .
statistical averages are then performed using these empirical data sets .
this method works without needing any analytic model for the probability distribution ; in particular , we avoid making the usual assumption of a normal distribution .
figure [ bootstrap ] shows the bootstrapped sampling distribution for our 13kv / cm data set and the corresponding cumulative distribution .
one sees that the bootstrap method is putting excess probability into the wings of the distribution , as expected .
the integral of this probability distribution tells us that the true confidence limits are about 20% larger than one would derive from applying normal gaussian statistics to the data set .
the fluctuating magnetic field in the laboratory is a possible source of additional noise in the measured edm ( or worse , a systematic error if changes in the laboratory field are correlated with the switching of the electric field ) .
the ybf beam is protected from external fields by two layers of magnetic shielding ; a flux gate magnetometer between the shields monitors the residual field during data collection .
figure [ bnoise_a ] shows the correlation between the field measured by the molecular beam and by the magnetometer , both multiplied by @xmath50 to express them in units of edm .
the slope of the line shows that the shielding factor of the inner shield is approximately @xmath51 . as there is no correlation between the external b field and the electric field reversal , we are confident that this magnetic field noise does not generate a false edm at the present level of sensitivity .
we employ a veto which discards data when the external field noise exceeds a certain threshold .
this improves the uncertainty in the edm measurement , as illustrated in figure [ bnoise_b ] for a subset of our data .
this plots the 67% confidence limit on the edm vs. the magnetic field veto level .
the optimal cutoff level is approximately where the external magnetic field induces fluctuations comparable to the edm measurement error bar .
a veto at this level typically rejects about 15% of our data . during edm data acquisition runs , we periodically reverse the electric and magnetic connections to the apparatus manually .
this guards against a false edm generated by the external apparatus ( for example magnetic fields from the high voltage relays ) or the data acquisition electronics .
because the experiment has been carefully designed to minimize such effects , we do not find any signal correlated with these manual reversals .
the saturation of the enhancement shown in figure [ polarisation ] is another powerful discriminant against systematics .
the interferometer phase induced by a true edm would have to vary in this way , whereas a systematic error would be unlikely to do so . for example , the phase shift due to the magnetic field of a relay is constant , whilst the effect of leakage currents or electrical breakdown grows linearly or faster with the applied voltage .
we have recorded data at the field values indicated in the figure . from the data taken so far
, we can already see that we do not have any strong disagreement with the result of reference @xcite , however we have discovered a new systematic effect that needs to be addressed before we can confidently give a new result at a higher level of accuracy . as noted previously
, the tensor stark splitting causes the zeeman shift to depend only on the magnetic field component @xmath52 along the electric field direction . as a result ,
false edms due to the @xmath53 motional field and to geometric phases induced by rotating magnetic fields are utterly negligible for the current beam experiment .
this strong coupling of the molecular axis to the external electric field is a major blessing , however , it does also bring with it a potential problem . if the direction of the electric field does not reverse exactly when the voltages are switched , this change of z - axis can produce a change of @xmath52 and hence a zeeman shift that mimics an edm . in our apparatus ,
the electric field plates are split into three regions so that the splitter and recombiner transitions can take place at a lower field than the central field region of the interferometer . in the gaps between these regions ,
the field lines are curved , causing off - axis molecules to experience a rotation of the local z - axis as they fly through .
if the electric field were to reverse perfectly this would not cause a problem since the rotation angle would be the same for both electric polarities , but of course no reversal can be perfect .
consequently , a fixed magnetic field roughly perpendicular to the rotating electric field can induce a false edm .
we have demonstrated this by applying a strong perpendicular magnetic field and deliberately introducing a very large asymmetry when the electric field reverses .
figure [ systematic ] shows a large false edm induced in this way . under our normal operating conditions , we estimate that this effect should not be larger than @xmath54 .
because it is difficult to guarantee that we understand these fringe fields perfectly and because we are aiming to measure @xmath5 at the @xmath55 level of accuracy , we are now replacing the electrode structure with a single pair of field plates so that the molecules will experience the same magnitude and direction of electric field throughout the entire interferometer .
in the new high - field splitter and recombiner regions , the field must be homogeneous across the beam to within 0.1% in order to avoid excessive stark broadening of the rf transition .
this places severe limits on the parallelism of the plates but the required precision has been achieved with specialist machining techniques and the new plates are being installed at time of writing .
at present we anticipate making a measurement with accuracy in the @xmath56 range as soon as the new plates are working , which will improve on the result of ref .
@xcite . over the next few years
we expect to reach an accuracy of @xmath57 without additional major modifications . in order to be sure of systematics at this level
, it may also be necessary to perform a control experiment using caf molecules .
these are similar to ybf structurally and magnetically , but have @xmath58 times less sensitivity to the electron edm according to the expected @xmath59 scaling @xcite .
we have already made a cold supersonic caf beam in our apparatus and have seen the transitions that are needed .
it seems possible to go significantly further in measurement accuracy by decelerating the ybf molecules to increase the time that they spend in the interferometer . with this in mind
, we have demonstrated that our ybf beam can be decelerated @xcite and we have made substantial progress in understanding how to bring heavy polar molecules close to rest using an alternating gradient decelerator @xcite . with the use of an intense , slow ybf beam , there is no obvious obstacle to a measurement at the level of @xmath60 .
such high precision would provide a probe of new cp - violating elementary particle physics up to mass scales in the range of many tev , testing some models far beyond the reach of current accelerators including the lhc at cern .
we acknowledge support from pparc , epsrc , the royal society ( uk ) , and the cold molecules research training network of the european commission .
we are indebted to jon dyne for expert technical assistance .
kozlov and v.f .
ezhov , _ phys .
rev . _ * a49 * , 4502 ( 1994 ) ; m.g .
b _ * 30 * l607 ( 1997 ) ; a. v. titov , n. s. mosyagin , v. f. ezhov , _ phys . rev
. lett . _ * 77 * 5346 ( 1996 ) ; h.m .
quiney , h. skaane , i.p .
b _ * 31 * l85 ( 1998)(after correcting for the trivial factor of 2 between * s * and * @xmath61 * their result becomes 26 gv / cm ) ; f.a .
parpia , _ j. phys .
b _ * 31 * 1409 ( 1998 ) ; n. mosyagin , m. kozlov , a. titov , _ j. phys .
b _ * 31 * l763 ( 1998 ) .
|
we present progress towards a new measurement of the electron electric dipole moment using a cold supersonic beam of ybf molecules .
data are currently being taken with a sensitivity of @xmath0 .
we therefore expect to make an improvement over the tl experiment of commins group , which currently gives the most precise result .
we discuss the systematic and statistical errors and comment on the future prospect of making a measurement at the level of @xmath1 .
address = centre for cold matter , imperial college london , london , sw7 2bw , uk + e - mail : [email protected]
|
measurements of strange particles provide information on the properties of strongly interacting matter at early stages of heavy ion collisions . on the assumption that a deconfinement phase transition occurs in heavy ion collisions
, we have to look for qualitative changes ( onset phenomena ) when varying external parameters like system size and energy density .
therefore , we study @xmath3 , @xmath0 and @xmath1 production as a function of beam energy and @xmath5 production dependent on the collision centrality .
this contribution reports on the first results of the @xmath3 production at 30 @xmath2gev and @xmath0 and @xmath1 production at 40 @xmath2gev .
the na49 experiment @xcite is a large acceptance hadron spectrometer which consists of four tpcs .
the two vertex tpcs are located in a strong magnetic field .
they allow the measurement of charge and momentum of the detected particles .
the two main tpcs are outside the magnetic field .
they are used for particle identification via de / dx measurement . a veto calorimeter which is placed downstream the tpcs measures the energy of spectator nucleons from which the collision centrality can be inferred .
in fig . [ mt_spec ] results on @xmath6-spectra of hyperons at 40 and 158 @xmath2gev at midrapidity measured in central pb+pb collisions by na49 are summarized .
because of the good signal to background ratio at 40 @xmath2gev @xmath3 and @xmath0 cover almost in the full @xmath6-region .
the @xmath0 at 158 @xmath2gev have already been published @xcite , while the @xmath3-spectra at both energies are submitted for publication @xcite .
all other presented results are still preliminary . at low and high @xmath6
the spectra deviate from an exponential shape .
because of this all spectra are fitted with a simple exponential function only in a limited @xmath7-range ( 0.2 - 1.4 gev/@xmath8 ) .
from these fits the following inverse slope parameters at 40 @xmath2gev are derived : @xmath9 ( 210@xmath1011 ) mev , @xmath11 ( 218@xmath1039 ) mev . at 40 @xmath2gev the signals of @xmath12 and @xmath13 hyperons
are summed because of low statistics .
transverse mass spectra of @xmath3 ( triangles ) , @xmath0 ( circles ) , @xmath1 ( squares ) produced in central pb+pb collisions at 40 ( left ) and 158 @xmath2gev ( right ) .
the @xmath3 spectra are uncorrected for feeddown from @xmath0 and @xmath1 decays . ]
the na49 detector is able to measure particles in a broad rapidity range . in fig .
[ rap ] the rapidity distributions of @xmath3 , @xmath0 and @xmath1 produced in central pb+pb collisions at several beam energies are shown .
the @xmath3 rapidity spectra at 30 @xmath2gev , the @xmath5 rapidity spectra at 40 @xmath2gev and the @xmath1-spectra at 40 and 158 @xmath2gev are parametrised by gaussians as indicated by the full lines .
the @xmath3-spectrum at 80 and the @xmath0-spectrum at 158 @xmath2gev are parametrised by the sum of two gaussians displaced symmetrically with respect to midrapidity
. a clear evolution of the shape of the @xmath3-spectra is visible . for @xmath5 and @xmath1
we observe an increase of the width of the rapidity distributions with beam energy , but no indication for a change of the shape .
the measured rapidity spectra have been integrated to obtain the total multiplicities in the full phase space , for numerical values see table [ 4piyield ] .
rapidity spectra of @xmath3 ( triangles ) , @xmath5 ( circles ) , @xmath1 ( squares ) produced in central pb+pb collisions at 30 , 40 , 80 and 158 @xmath2gev .
filled symbols are measured , open ones are reflected at midrapidity . for clarity the @xmath0-spectra are scaled by a factor 5 and the @xmath1-spectra by a factor 20 .
the @xmath3 spectra are uncorrected for feeddown from @xmath0 and @xmath1 .
all spectra are preliminary except @xmath5 at 158 @xmath2gev . ]
@l*15@l beam energy & 30 @xmath2gev & 40 @xmath2gev & 80 @xmath2gev & 158 @xmath2gev + @xmath14 & 7.62 gev & 8.73 gev & 12.30 gev & 17.30 gev + @xmath15 & 7.2% & 7.2% & 7.2% & 10% ( @xmath1 : 20% ) + @xmath16 & [email protected]@xmath104.0 & [email protected]@xmath103.4 & [email protected]@xmath103.5 & [email protected]@xmath105.0 + @xmath17 & & [email protected]@xmath100.24 & & [email protected]@xmath100.62 + @xmath18 & & [email protected]@xmath100.04 & & [email protected] + @xmath19 & & & & [email protected] +
the total multiplicities of @xmath3 , @xmath0 and @xmath1 are plotted as a function of the centre of mass energy in fig .
[ energy_dep ] .
all yields are divided by the total multiplicities of the negatively charged pions .
the energy dependence of the @xmath3 hyperons shows a clear maximum at about 30 @xmath2gev ( @xmath207.62 gev ) . the @xmath21 ratio shows a weak maximum at 40 @xmath2gev ( @xmath208.73 gev ) .
the non - monotonic behaviour is not seen for the @xmath22 ratio .
the dashed lines show a prediction of a grand canonical hadron - gas model , which includes the variation of the baryo - chemical potential and the temperature with the beam energy @xcite .
the model describes the basic features of our data within 30% .
in fig . [ cen_dep ] the centrality dependence of @xmath5 production at 40 @xmath2gev at midrapidity is shown .
the yields are normalised by the mean number of wounded nucleons @xcite .
the @xmath5 production increases from peripheral to midcentral collisions and this behaviour is followed by a saturation toward central reactions .
20 afanasiev s v 1999 _ nucl . instrum .
methods _ a * 430 * 210 afanasiev s v 2002 _ phys .
_ b * 538 * 275 alt c in press ( nucl - ex/0311024 ) cleymans j and redlich k 1999 _ phys .
_ c * 60 * 054908 and private communication bialas a , bleszynski m and czyz w 1976 _ nucl .
_ b * 111 * 461
|
new results on the production of @xmath0 and @xmath1 hyperons in pb+pb interactions at 40 @xmath2gev and @xmath3 at 30 @xmath2gev are presented .
transverse mass spectra as well as rapidity spectra of these hyperons are shown and compared to previously measured data at different beam energies .
the energy dependence of hyperon production ( 4@xmath4 yields ) is discussed .
additionally , the centrality dependence of @xmath5 production at 40 @xmath2gev is presented .
|
acoustoelasticity is now a well established experimental technique used for the non - destructive measurement of third - order elasticity ( toe ) constants of solids , and its principles can be found in standard handbooks of physical acoustics , such as @xcite . the underlying theory , however , is quite intricate , and over the years several authors have produced different and irreconcilable expressions for the shift experienced by the wave speed when elastic wave propagation is coupled to a pre - strain .
a common mistake found in the literature consists of the implicit assumption that since a small pre - strain and a small - amplitude wave are described by linearized equations , they can be superposed linearly . with that point of view , the coupling between the two phenomena is of higher order and can be neglected in the first approximation .
the flaw in that reasoning is simply that these phenomena are _ successive _ , not linearly superposed .
experiencing an infinitesimal pre - strain from a stress - free configuration is a linear process ; propagating an infinitesimal wave in a stress - free configuration is another linear process .
but in acoustoelasticity , the wave travels in a pre - stressed , not a stress - free solid , and the laws of linear elastodynamics must be adapted to reflect this fact .
the final outcome of this analysis is that not only second - order , but also toe constants appear in the expression for the speed of an acoustoelastic wave .
for example , a longitudinal wave travelling in a solid subject to a hydrostatic stress @xmath1 will propagate with the speed @xmath2 given by @xmath3 where @xmath4 is the mass density _ in the unstressed configuration _ ,
@xmath5 and @xmath6 are the ( second - order ) lam coefficients , and @xmath7 , @xmath8 , @xmath9 are the ( third - order ) landau coefficients @xcite .
this equation was correctly established as early as 1925 by brillouin and confirmed in 1953 by hughes and kelly , although several erroneous expressions have appeared in between and since ( see , e.g. , @xcite ) . from our contemporary perspective , the easiest way to re - establish and further this expression is to rely on the modern theory of incremental ( also known as small - on - large ) elasticity , which presents compact expressions for the tensor of _ instantaneous elastic moduli _ , @xmath10 , given by its components @xmath11 & \mathcal{a}_{0ijij } = j^{-1}\frac{\lambda_i w_i - \lambda_j w_j}{\lambda_i^2 -\lambda_j ^2 } \lambda_i^2 , & i \ne j , \ ; \lambda_i \ne \lambda_j , \\[1ex ] & \mathcal{a}_{0ijji } = j^{-1}\frac{\lambda_j w_i - \lambda_i w_j}{\lambda_i^2 -\lambda_j ^2 } \lambda_i \lambda_j , & i \ne j , \ ; \lambda_i \ne \lambda_j , \\[1ex ] & \mathcal{a}_{0ijij } = j^{-1 } ( \lambda_i^2 w_{ii } - \lambda_i\lambda_jw_{ij } + \lambda_i w_i)/2 , & i \ne j , \ ; \lambda_i = \lambda_j,\\[1ex ] & \mathcal{a}_{0ijji } = j^{-1 } ( \lambda_i^2 w_{ii } - \lambda_i\lambda_jw_{ij } - \lambda_i w_i)/2 , & i \ne j , \ ; \lambda_i = \lambda_j , \end{aligned } \right .
\label{aoijkl}\ ] ] in the coordinate system aligned with the principal axes of pre - strain , which , since the material is isotropic , coincide with the principal axes of pre - stress . here
, @xmath12 is the strain - energy density per unit volume , which is a symmetric function of the principal stretches @xmath13 of the deformation , @xmath14 , @xmath15 and @xmath16 ; see , for example , @xcite for details .
the corresponding principal cauchy stresses are given by @xmath17 .
it is worth noting here for later reference that it follows from @xmath18 that @xmath19 a small - amplitude body wave may travel at speed @xmath20 in the direction of the unit vector @xmath21 with polarization in the direction of a unit vector @xmath22 provided the eigenvalue problem @xmath23 is solved with @xmath24 , where @xmath25 is the _
acoustic tensor _ , which has components @xmath26 note that sometimes the acoustic tensor is defined by _ without _ the factor @xmath27 , in which case the @xmath4 in would be replaced by the density in the deformed configuration , i.e. @xmath28 .
notice that in this theory the amplitude of the acoustic wave is infinitesimal , but that there is _ no restriction _ on the choice of @xmath12 or on the magnitude of the underlying pre - strain because we are in the context of _ finite _ nonlinear elasticity .
it follows that the expressions can be specialized in various ways , in particular to _ weakly _ nonlinear elasticity ( where @xmath12 is expanded in terms of some measure of strain ) and to special pre - strains . hence , to access the toe constants , we take @xmath12 as @xmath29 where @xmath30 , are invariants of the green lagrange strain tensor @xmath31 .
the components can then be expanded to the first order in terms of , for example , the volume change @xmath32 in the case of a hydrostatic pressure , or of the elongation @xmath33 in the case of uniaxial tension . to access fourth - order elasticity ( foe ) constants ,
we take @xmath12 as @xmath34 where @xmath35 , @xmath36 , @xmath37 , @xmath38 are the foe constants and we push the expansions of the wave speed up to the next order in the strain .
this is what we refer to as the _ large acoustoelastic effect_. in fact , the main purpose of this investigation is to provide a theoretical backdrop to the current drive to determine experimentally the foe constants of soft solids such as isotropic tissues and gels in order to improve acoustic imaging resolution ( see , e.g. , @xcite ) . because soft solids are often treated as _
incompressible _ , so that the constraint @xmath39 must be satisfied , we shall also consider the expressions for the toe and foe strain - energy densities in their incompressible specializations , specifically @xmath40 for toe incompressibility and @xmath41 for foe incompressibility .
hence , in the transition from compressible to incompressible elasticity , the number of second - order constants goes from two to one , of third - order constants from three to one , and of fourth - order constants from four to one .
this was established by hamilton et al . in 2004 , although it can be traced back to ogden in 1974 and even earlier , to bland in 1969 .
specifically , in that transition we note that the elasticity constants behave as @xcite @xmath42 for the second - order constants , where @xmath43 is the ( finite ) young s modulus , @xmath44 for the third - order constants , @xmath45 for the fourth - order constants , and @xmath46 for a combination of second- and third - order constants .
some constants have the explicit limiting behaviour @xmath47 where @xmath0 remains a finite quantity , of the same order of magnitude as @xmath6 ; see @xcite for analysis of the behaviour of @xmath7 , @xmath8 , , @xmath38 in the incompressible limit and the connection with @xmath0 .
note that the limiting behaviour for @xmath35 in terms of the initial poisson s ratio @xmath48 and young s modulus @xmath43 is thus @xmath49 .
this was shown in @xcite , but mistyped in equations ( 49 ) and ( 89 ) therein as @xmath50 . in this paper , we treat in turn the case of hydrostatic pre - stress ( section [ hydrostatic pressure ] ) and of uniaxial pre - stress ( section [ uniaxial tension ] ) , and we provide expansions of the body wave speeds up to the second order in the pre - strain , and also in the pre - stress , for compressible solids and in the relevant incompressible limits .
consider first a cuboidal sample of a compressible solid with sides of lengths @xmath51 in its ( unstressed ) reference configuration .
we define the reference geometry in terms of cartesian coordinates @xmath52 by @xmath53 .
the material is then subject to a pure homogeneous strain @xmath54 and deformed into the cuboid @xmath55 , where @xmath56 are the cartesian coordinates in the deformed configuration , and the constants @xmath13 are the principal stretches of the deformation .
we now specialize the deformation to a pure dilatation so that @xmath57 . since the material is isotropic the cauchy stress @xmath58 is spherical , @xmath59 say , with @xmath60 corresponding to hydrostatic tension ( pressure ) .
the pre - stress is computed as @xmath61 or , more conveniently , as @xmath62 , where @xmath63 . up to the second order in the volume change @xmath64
, we find @xmath65 where @xmath66 is the infinitesimal bulk modulus , and @xmath67 ( it is a simple matter to check that the expansion is equivalent to one first established by @xcite . )
conversely , the volume change is expressed in terms of the hydrostatic stress as @xmath68 here the pre - stress does not generate preferred directions and the solid remains isotropic .
it follows that two waves may propagate in any direction , one longitudinal , with speed @xmath69 , and one transverse , with speed @xmath70 , given by @xmath71 where these quantities are computed from the formulas .
expanding in terms of the volume change , we obtain @xmath72 where @xmath73 are the coefficients for the classical ( linear ) acoustoelastic effect , and @xmath74 b_{t } & = & \frac{1}{2 } \lambda + \frac{5}{9 } \mu+ \frac{1}{2 } a + \frac{13}{6 } b + c + e + f + \frac{2}{3 } g\end{aligned}\ ] ] are the coefficients of the large ( quadratic ) acoustoelastic effect . alternatively , we may use to express the wave speeds in terms of the pre - stress rather than the pre - strain , as @xmath75 thus recovering the classical formulas @xmath76 for the ( linear ) acoustoelastic effect @xcite , and establishing the formulas @xmath77 for the large ( quadratic ) acoustoelastic effect , where @xmath78 is given in . for an application of these results
, we turn to the classical data of @xcite .
they performed acoustoelastic experiments and plotted the stress - dependent shear and bulk moduli , defined as @xmath79 against the hydrostatic pressure @xmath80 for polysterene and for pyrex .
for the former , the variations are clearly linear , and nothing would be gained by including foe effects . for the latter ,
there is a marked departure from the linear acoustoelastic effect at high pressures , and we thus use to obtain a better fit to the data . by digitizing the data and performing a standard least square optimization , we found that for pyrex , @xmath81 k(p ) & = & 3.27 - 7.80 \times 10^{-5}p + 4.07 \times 10^{-9}p^2,\end{aligned}\ ] ] where @xmath82 and @xmath83 are expressed in units of @xmath84 bars and @xmath85 in bars .
the fit to the data is shown in fig .
[ fig : hydro ] . according to
, the quantities give access to @xmath5 and @xmath6 , to two linear combinations of the third - order constants , and to two linear combinations of the fourth - order constants .
specifically , we may solve to find the lam constants as @xmath86 similarly , we find two linearly independent combinations of the three third - order constants : @xmath87 - 4 m(0 ) , \quad a + 9b + 9c = - 9k(0 ) \left[1+k'(0 ) \right]/2.\ ] ] finally , we also find two combinations of the fourth - order constants : @xmath88 & \phantom { e + f + \dfrac{2}{3}g = \dfrac{7}{6}k(0 ) } + \dfrac{1}{2}k(0)\left [ k'(0 ) + \dfrac{13}{3}m'(0 ) + k'(0)m'(0 ) + k(0)m''(0 ) \right],\notag \\[4pt ] & e + 3f + g + 9h = \dfrac{17}{24}k(0 ) + \dfrac{3}{2}k(0)\left[k'(0 ) + \dfrac{1}{4}\left(k'(0 ) \right)^2 + \dfrac{1}{4}k(0)k''(0)\right].\end{aligned}\ ] ] hence , for the pyrex data : @xmath89 clearly , the data of @xcite are not sufficient to determine all the third- and fourth - order constants . an additional relation is needed to determine experimentally all the third - order constants , and a further two relations to determine the fourth - order constants .
these may be provided from , for example , wave speed measurements in uniaxially deformed samples , as we show in the following section .
we conclude this section by evoking the work of p.w .
bridgman on `` the physics of high pressures '' @xcite , which won him the 1946 nobel prize .
he carried out countless high pressure measurements on solids , and fitted his data with a quadratic equation for the volume change , in the form @xmath90 .
direct comparison of this experimental law with reveals the identifications @xmath91 and @xmath92 .
further , acoustoelastic measurements give direct access to these constants , because @xmath93 and @xmath94 $ ] . in bridgman
s experiments , very high pressures are required to obtain second - order deformations ; however , his coefficients @xmath95 and @xmath96 can also be determined by pressurizing the sample only linearly , and then measuring the speeds of transverse and longitudinal waves .
in fact , had bridgman been able to propagate such waves in his finitely deformed samples he would have had access to the third term in the expansion , @xmath97 say , because the continuation of is @xmath98p^3,\ ] ] where @xmath99 is found to be expressible as @xmath100 in which each term can be determined experimentally using the equations of large acoustoelasticity above .
we again consider the cuboidal sample as in section [ hydrostatic pressure ] , but now the cuboid is subject to a uniaxial tension in the @xmath101 direction so that it deforms homogeneously with elongation @xmath102 in the @xmath103-direction . by symmetry it contracts laterally and equibiaxially with elongation @xmath104 .
the nominal stress ( axial force per unit reference area ) is @xmath105 .
the condition that the lateral faces are free of traction is @xmath106 which determines the extent of the lateral contraction @xmath107 ( assuming that @xmath108 ) and can be written in terms of @xmath109 , up to second order , as @xmath110 where @xmath111 \frac{b}{2(\lambda+\mu ) } + \frac { \mu^2 c}{2(\lambda+\mu)^3 } , \label{alphabeta}\ ] ] and again @xmath112 is the infinitesimal bulk modulus . note that collapses to @xmath113 in the incompressible limits and , as expected from the expansion to second order of the connection @xmath114 between the stretch ratios of an incompressible solid in uniaxial tension . substituting with into the expression @xmath115 for the uniaxial stress
, we obtain the relation between the pre - stress and the pre - strain , up to the second order , as @xmath116 where @xmath117
a + \frac{3\mu\left(3\lambda^2 + 4 \lambda\mu + 2 \mu^2\right)}{2\left(\lambda + \mu\right)^3 } b + \frac{\mu^3}{\left(\lambda + \mu\right)^3 } c.\quad\end{aligned}\ ] ] ( note that although @xmath118 , the infinitesimal young s modulus , we shall not use @xmath43 hereon . )
conversely , @xmath119 in the incompressible limits and , the stress
strain and strain stress relations reduce to @xmath120 now we examine the possibility of a small - amplitude body wave travelling in the direction of the unit vector @xmath21 with polarization in the direction of a unit vector @xmath22 . for a general direction of propagation , the speeds of the different waves are found as the eigenvalues of the acoustical tensor @xmath25 in , i.e. as the roots of a cubic . in this paper , however , we focus on the wave speeds that are relatively simple to determine experimentally . the speeds of non - principal body waves are not easily accessible because they would require transducers to be placed at an angle to the faces of the cuboid , and not in full flat contact , which would lead to additional transmission problems .
this leaves the principal body waves , as explained by @xcite : first , waves in the direction @xmath103 of the tension , and , second , principal waves in the @xmath121 and @xmath122 directions , which are equivalent by symmetry . in the direction of tension there exists a longitudinal wave with speed @xmath123 , say , and two transverse waves , propagating with the same speed @xmath124 , say ( in fact , these latter two waves may be combined to form a transverse circularly - polarized wave ) .
hence , @xmath125 for the longitudinal wave , and @xmath126 , @xmath127 for the transverse wave , where @xmath128 and @xmath129 are unit basis vectors corresponding to the coordinates @xmath103 and @xmath121 , respectively . using and we then have @xmath130 expanding the elastic moduli in terms of the elongation @xmath109 , we obtain @xmath131 where @xmath132 a_{12 } & = & \dfrac{\mu}{\lambda+\mu}\left[4(\lambda+\mu ) + \dfrac{\lambda+2\mu}{4\mu } a + b \right],\end{aligned}\ ] ] for the classical acoustoelastic effect , and @xmath133 a+\frac{3 \left ( \lambda + 2 \mu\right ) \left(12 \lambda^2 + 21 \lambda \mu + 10 \mu^2
\right)}{2\left(\lambda + \mu \right)^3 } b\notag\\[1ex ] & + & \mu \left [ \frac{9}{\lambda + \mu}+ \frac{\mu^2}{\left ( \lambda + \mu \right)^3 } \right ] c - \frac{\lambda^2 a}{2\left(\lambda + \mu \right)^3 } \left ( b + c \right)-\frac{3\lambda^2 + 2\lambda\mu+2\mu^2}{(\lambda+\mu)^3}bc\notag\\[1ex ] & -&\frac{3 \lambda^2 + 2 \lambda \mu + 2 \mu^2}{\left(\lambda + \mu \right)^3}b^2-\frac{2\mu^2}{\left(\lambda + \mu \right)^3}c^2+\frac{6\left ( \lambda+2\mu \right)}{\lambda+ \mu}e + 3\left ( \frac{\lambda+2\mu } { \lambda+ \mu}\right)^2 f\notag\\[1ex ] & + & 2\left [ 6 + \frac{\lambda^2}{\left ( \lambda + \mu\right)^2}\right]g + \frac{12 \mu^2}{\left ( \lambda + \mu\right)^2}h,\end{aligned}\ ] ] @xmath134 & -&\frac{5\lambda^2 + 2\lambda\mu+2\mu^2}{8(\lambda+\mu)^3}ab-\frac{\mu^2}{4(\lambda+\mu)^3}(a+4b)c+\frac{3\mu \left(\lambda+2\mu \right)}{4\left(\lambda+\mu \right ) ^2}e+\frac{\mu^2}{\left(\lambda+\mu \right)^2}f\notag\\[1ex ] & + & \left [ 2 + \frac{\lambda^2}{\left(\lambda+\mu \right)^2}\right]g,\end{aligned}\ ] ] for the large acoustoelastic effect . for propagation perpendicular to the direction of tension , we take @xmath135 .
then there exists a longitudinal wave , with @xmath127 and speed @xmath136 , a transverse wave polarized in the direction of tension @xmath137 with speed @xmath138 , and a transverse wave polarized perpendicular to the direction of tension , with @xmath139 and speed @xmath140 .
these speeds are given by @xmath141 expanding the elastic moduli in terms of the elongation @xmath109 , we obtain @xmath142 where @xmath143 a_{21 } & = & \dfrac{(\lambda + 2\mu)(4\mu+a ) + 4\mu b}{4(\lambda+\mu)},\\[1ex ] a_{23 } & = & - \dfrac{\lambda(4\mu+a)-2\mu b}{2(\lambda+\mu)},\end{aligned}\ ] ] are the classical acoustoelastic coefficients , and @xmath144 & -&\frac{\lambda^2}{4(\lambda+\mu)^3}a^2-\frac{2(3\lambda^2 + 2\lambda\mu+2\mu^2)}{(\lambda+\mu)^3}b^2-\frac{2\mu^2}{(\lambda+\mu)^3}c^2 -\frac{5\lambda^2 + 2\lambda\mu+2\mu^2}{2(\lambda+\mu)^3}ab\notag\\[1ex ] & -&\frac{3\lambda^2 + 2\lambda\mu+6\mu^2}{(\lambda+\mu)^3}bc -\frac{\lambda^2 + 2\mu^2}{2(\lambda+\mu)^3}ac+\frac{3\lambda(\lambda-2\mu)}{2(\lambda+\mu)^2}e+\frac{3\lambda^2 + 4\mu^2}{(\lambda+\mu)^2}f\notag\\[1ex ] & + & 4\left[1+\frac{\lambda^2}{(\lambda+\mu)^2}\right]g+\frac{12\mu^2}{(\lambda+\mu)^2}h , \notag \\[4pt ] b_{21 } & = & b_{12 } - \gamma -\frac{3\kappa \mu^2}{\left ( \lambda + \mu \right)^2 } , \notag \\[4pt ] b_{23 } & = & -\frac{\lambda \mu^2 } { ( \lambda + \mu)^2 } + \frac{\lambda(3\lambda^2-\mu^2)}{4(\lambda+\mu)^3}a -\frac{3(3\lambda^2 + 2\lambda\mu+\mu^2)\mu}{2(\lambda+\mu)^3}b-\frac{2\mu^3}{(\lambda+\mu)^3}c-\frac{\lambda^2}{8(\lambda+\mu)^3}a^2\notag\\[1ex ] & -&\frac{3\lambda^2 + 2\lambda\mu+2\mu^2}{2(\lambda+\mu)^3}b^2 -\frac{2\lambda^2+\lambda\mu+\mu^2}{2(\lambda+\mu)^3}ab-\frac{\mu^2}{2(\lambda+\mu)^3}(a+2b)c-\frac{3\lambda\mu}{2(\lambda+\mu)^2}e\notag\\[1ex ] & + & \frac{\mu^2}{(\lambda+\mu)^2}f+\left[2+\frac{\lambda^2}{(\lambda+\mu)^2}\right]g , \end{aligned}\ ] ] are the ` large acoustoelasticity ' coefficients .
we may also find expressions for the acoustoelastic effect in terms of the pre - stress @xmath145 as @xmath146 where @xmath147 ( no sum on repeated @xmath148 ) , and @xmath149 ( @xmath150 ) .
using , we find that @xmath151 where @xmath152 .
in particular , we recover @xmath153 , \notag \\[1ex ] c_{12 } & = & \dfrac{1}{3\kappa}\left[4(\lambda+\mu ) + \dfrac{\lambda+2\mu}{4\mu } a + b \right ] , \notag \\[1ex ] c_{22 } & = & \dfrac{2}{3\kappa}\left[b + c - \dfrac{\lambda}{\mu}\left(\lambda+2\mu+\dfrac{a}{2}+b\right ) \right ] , \notag \\[1ex ] c_{21 } & = & \dfrac{1}{3\kappa}\left[\lambda + 2\mu + \dfrac{\lambda+2\mu}{4\mu}a + b \right ] , \notag\\[1ex ] c_{23 } & = & - \dfrac{1}{3\kappa}\left[2\lambda + \dfrac{\lambda}{2\mu}a - b \right],\end{aligned}\ ] ] for the classical acoustoelastic effect @xcite .
note that @xmath154 and that @xmath155 , and hence in the incompressible limit @xmath156 . using these relations
, we establish that @xmath157 recall now that @xmath4 is the mass density in the reference configuration .
when expressed in terms of the deformed density @xmath158 equation becomes @xmath159 this relation is in fact exact in accordance with and the expressions @xmath160 , @xmath161 specialized accordingly with @xmath162 . finally , we take the incompressible limits of the elastic constants using the limiting values listed in section [ introduction ] . there , we have @xmath163 and @xmath164 , unsurprisingly , because longitudinal homogeneous plane waves may not propagate in incompressible solids . for the transverse principal waves travelling in the direction of tension , @xmath165 in terms of the elongation @xcite , and @xmath166 \sigma_1 ^ 2,\ ] ] in terms of the pre - stress .
for the transverse principal waves travelling perpendicular to the direction of tension , @xmath167 in terms of the elongation , see @xcite , and , using , @xmath168 \sigma_1 ^ 2,\notag \\ & \rho v_{23}^2 = \mu - \left(1 + \dfrac{a}{6 \mu}\right ) \sigma_1 + \frac{1}{9\mu^2 } \left ( 8 \mu + 3a + \frac{a^2}{8\mu } + 3d \right ) \sigma_1 ^ 2,\end{aligned}\ ] ] in terms of the pre - stress . for an application
, we use data for a sample of silicone rubber that has been subjected to a standard tensile test .
figure [ fig : silicone1 ] displays the variation of the tensile cauchy stress component @xmath145 with the elongation up to a maximum stretch of about 250% , at which stage one end of the sample snapped out of its grip . over that range , the toe strain energy density is not able to capture the behaviour of the sample adequately , as shown in fig .
[ fig : silicone1 ] , and is thus discarded . on the other hand , the foe strain energy gives an excellent least - squares fit , with coefficient of correlation @xmath169 .
in fact , the fit is very good up to a much larger value of the stretch than could be expected of this fourth - order approximate theory .
we determined the following values for the constants of second- , third- , and fourth - order elasticity : @xmath170 note that all three are of the same order of magnitude , as expected from the theory @xcite .
we also remark that @xmath171 , indicating that the corresponding mooney
rivlin solid is materially stable @xcite . using these values ,
the theoretical variations of the squared wave speeds @xmath172 , @xmath173 , @xmath174 are plotted versus the elongation in fig .
[ fig : silicone2 ] .
these curves suggest that it is sufficient to elongate the sample by about 20% to reveal quadratic acoustoelastic effects , and thus to determine @xmath0 experimentally .
this study was set within the framework of the theory of small - amplitude waves propagating in a deformed elastic solid .
existing general expressions for the speeds of body waves in a finitely deformed isotropic elastic solid were recalled and then specialized to fourth - order elasticity . specifically , the squared wave speeds were expanded in terms of the pre - deformation ( and , equivalently , in terms of the pre - stress ) in order to reveal the so - called acoustoelastic effect at the considered order .
while the classical acoustoelastic effect is concerned with the linear variation of the squared wave speed with the pre - stress , the expansion was extended to the quadratic regime ( large acoustoelastic effect ) .
explicit expressions for the quadratic coefficients for compressible and incompressible materials were determined for the cases of hydrostatic and uniaxial pre - stress .
these coefficients give access to the fourth - order elastic constants . in the case of an incompressible sample under uniaxial tension , for example
, measurement of the variations of the speed of a single transverse wave with respect to the pre - stress ( or pre - strain ) is enough to determine the elastic constants .
the wave speed in the unstressed configuration gives the initial shear modulus @xmath6 of second - order elasticity , the linear variation of the squared wave speed with the pre - stress or pre - strain gives the third - order landau - coefficient @xmath7 , and the quadratic variation gives the fourth - order constant @xmath0 . for compressible materials
, @xcite showed that any three of the seven equations , , , with just the linear terms retained , or equivalently , , excluding either the expression for @xmath175 or for @xmath176 , suffice to determine the two lam coefficients @xmath5 and @xmath6 of second - order elasticity and the three landau coefficients of third - order elasticity @xmath7 , @xmath8 , @xmath9 .
here we have shown that any four of these seven equations , with the quadratic terms now included , give access also to the fourth - order constants @xmath35 , @xmath36 , @xmath37 , @xmath38 ( again , excluding either the expression for @xmath175 or for @xmath176 ) . although we have restricted attention largely to third- and fourth - order elasticity the expressions for the components of the tensor of instantaneous elastic moduli @xmath177 given in apply for an arbitrary finite deformation relative to an unstressed configuration of an isotropic elastic material , and the subsequent incremental response depends on the finite deformation and its accompanying pre - stress . if instead there is an initial stress in the reference configuration then the components of @xmath177 are considerably more complicated , as detailed in @xcite , but they do clarify how the elastic moduli depend in general on the initial stress .
in particular , the dependence of @xmath177 on an initial hydrostatic stress @xmath178 can be put in the simple form @xmath179 where @xmath180 and @xmath181 are now stress - dependent lam moduli , with @xmath182 corresponding to tension ( pressure ) , their precise functional dependence determined by the choice of strain - energy function .
we refer to a recent paper by @xcite and references therein for a discussion of the stress dependence of moduli within the context of an implicit theory of elasticity .
this work was supported by erasmus funding from the european commission and travel funding from the university of salento and from the national university of ireland galway .
it was also supported by an international joint project grant from the royal society of london .
we thank stephen kiernan and michael gilchrist at university college dublin for assistance with the tensile test of a silicone sample . , incremental statics and dynamics of pre - stressed elastic materials , in waves in nonlinear pre - stressed materials , m. destrade , g. saccomandi ( editors ) , cism lecture notes , * 495 * , 126 .
springer , new york ( 2007 ) . , measurement of shear elastic moduli in quasi - incompressible soft solids , 18th international symposium on nonlinear acoustics , july 07 - 10 , 2008 stockholm , sweden , nonlinear acoustics fundamentals and applications , book series : aip conference proceedings , * 1022 * , 303306 ( 2008 ) .
|
classical acoustoelasticity couples small - amplitude elastic wave propagation to an infinitesimal pre - deformation , in order to reveal and evaluate non - destructively third - order elasticity constants . here
, we see that acoustoelasticity can be also be used to determine fourth - order constants , simply by coupling a small - amplitude wave with a small - but - finite pre - deformation .
we present results for compressible weakly nonlinear elasticity , we make a link with the historical results of bridgman on the physics of high pressures , and we show how to determine @xmath0 " , the so - called fourth - order elasticity constant of soft ( incompressible , isotropic ) solids by using infinitesimal waves . _
keywords : _ acousto - elasticity , high pressures , large pre - tension , elastic constants
|
more than seventy years after zwicky s first accounts of dark matter in galaxy clusters @xcite , and thirty five years after rubin s measurements of rotational velocities of spirals @xcite , the case for non - baryonic dark matter remains convincing .
recent precision observations of the cosmic microwave background @xcite and of large scale structures @xcite confirm the picture in which more than 90% of the matter in the universe is revealed only by its gravitational interaction .
the nature of this matter is not known . a class of generic candidates are weakly interacting massive particles ( wimps ) which could have been thermally produced in the very early universe .
it is well known that if the mass and cross section of these particles is determined by the weak scale , the freeze - out relic density is around the observed value , @xmath0 0.1 .
the prototype wimp candidate is the neutralino , or the lightest supersymmetric particle , which is stable in supersymmetric models where r - parity is conserved .
another recently discussed candidate is the lightest kaluza - klein excitation ( lkp ) in theories with universal extra dimensions . if a new discrete symmetry , called kk - parity is conserved , and if the kk particle masses are related to the weak scale , the lkp is stable and makes an excellent dark matter candidate .
a vast experimental effort to detect wimps is underway .
for excellent recent reviews we refer to @xcite , @xcite and @xcite .
the good news is that cryogenic experiments are now for the first time probing the parameter space predicted by susy theories for neutralinos . on the more pessimistic side , predicted wimp quark cross sections span several orders of magnitude , and ton or multi - ton scale detectors may be required for a detection .
however , such experiments are now in the stage of development , with prototypes being tested and installed in underground labs .
the challenge is immense , but the rewards would be outstanding : revealing the major constituents of matter in the universe and their doubtless profound implications for fundamental physics at the weak scale .
wimps can be detected directly , via their scattering off nuclei in terrestrial targets @xcite , or indirectly , via their annihilation products in the sun , earth , galactic halo and galactic center with neutrino telescopes and space - based detectors . here
we will briefly discuss direct detection only .
the differential rate for wimp elastic scattering off nuclei is given by @xcite @xmath1 where @xmath2 represents the number of the target nuclei , @xmath3 is the wimp mass and @xmath4 the local wimp density in the galactic halo , @xmath5 and @xmath6 are the wimp velocity and velocity distribution function in the earth frame and @xmath7 is the wimp - nucleus differential cross section .
the nuclear recoil energy is given by @xmath8 , where @xmath9 is the scattering angle in the wimp - nucleus center - of - mass frame , @xmath10 is the nuclear mass and @xmath11 is the wimp - nucleus reduced mass .
the velocity @xmath12 is defined as @xmath13 , where @xmath14 is the energy threshold of the detector , and @xmath15 is the escape wimp velocity in the earth frame .
the simplest galactic model assumes a maxwell - boltzmann distribution for the wimp velocity in the galactic rest frame , with a velocity dispersion of v@xmath16 270 km s@xmath17 and an escape velocity of v@xmath18 650 km s@xmath17 .
the differential wimp - nucleus cross section can have two separate components : an effective scalar coupling between the wimp and the nucleus ( proportional to a@xmath19 , where a is the target atomic mass ) and an effective coupling between the spin of the wimp and the total spin of the nucleus .
in general the coherent part dominates the interaction ( depending however on the content of the neutralino ) and the cross section can be factorized as @xmath20 , where @xmath21 is the point - like scalar wimp - nucleus cross section and @xmath22 denotes the nuclear form factor , expressed as a function of the recoil energy .
the left side of equation [ eq1 ] is the measured spectrum in a detector , while the right side represents the theoretical prediction .
it includes wimp properties which are completely unknown , such as the wimp mass @xmath3 and elastic cross section @xmath21 , quantities accessible from astrophysics , such as the density of wimps in the halo , @xmath4 , the wimp velocity distribution and the escape velocity ( which are however prone to large uncertainties ) and detector specific parameters : mass of target nucleus , energy threshold and nuclear form factor .
the nuclear form factor becomes significant at large wimp and nucleus masses , and leads to a suppression of the differential scattering rate .
figure [ rates ] shows differential spectra for si , ar , ge and xe , calculated for a wimp mass of 100gev , a wimp - nucleon cross section of @xmath23 cm@xmath19 and using the standard halo parameters mentioned above .
the spectrum was calculated for illustrative nuclei such as si ( light solid ) , ar ( light dot - dashed ) , ge ( dark solid ) , xe ( dark dashed).,width=302 ] a wimp with a typical mass between a few gev and 1tev will deposit a recoil energy below 50kev in a terrestrial detector . as for the predicted event rates for neutralinos , scans of the mssm parameter space under additional assumptions ( gut , msugra , etc ) and accounting for accelerator and cosmological constraints , yield about 10@xmath25 to 10 events per kilogram detector material and day @xcite .
evidently , in order to observe a wimp spectrum , low energy threshold , low background and high mass detectors are essential .
in such a detector , the recoil energy of the scattered nucleus is transformed into a measurable signal , such as charge , light or phonons , and at least one of the above quantities is detected . observing two signals
simultaneously yields a powerful discrimination against background events , which are mostly interactions with electrons , as opposed to wimps and neutrons scattering off nuclei ( see section [ experiments ] for a more detailed discussion ) . even for experiments with good event by event discrimination ,
an absolute low background is still important
. it can be achieved in both passive and active ways .
passive methods range from high material selection of detector components to various specific shieldings against the natural radioactivity of the environment and against cosmic rays and secondary particles produced in their interactions . active background reduction implies an active shield , commonly a plastic or liquid scintillator surrounding the detector .
an additional advantage is provided by a highly granular detector ( or good timing and position resolution ) , since multiple scatters within the detector volume allow both background reduction and a direct measurement of the neutron background for experiments with an event - by - event discrimination against electron recoils . in order to convincingly detect a wimp signal , a specific signature from a particle populating
our galactic halo is important .
the earth s motion through the galaxy induces both a seasonal variation of the total event rate @xcite and a forward - backward asymmetry in a directional signal @xcite .
the annual modulation of the wimp signal arises because of the earth s motion in the galactic frame , which is a superposition of the earth s rotation around the sun and the sun s rotation around the galactic center : @xmath26 where @xmath27 + 12 km s@xmath17 ( @xmath28 @xmath29 220 km s@xmath17 ) , @xmath30 @xmath29 30 km s@xmath17 denotes the earth s orbital speed around the sun , the angle @xmath31 is the inclination of the earth s orbital plane with respect to the galactic plane and @xmath32 , @xmath33 june 2@xmath34 .
the expected time dependence of the count rate can be approximated by a cosine function with a period of t = 1year and a phase of @xmath35 june 2@xmath34 : @xmath36 where @xmath37 , @xmath38 are the constant and the modulated amplitude of the signal , respectively . in reality , an additional contribution to s(t ) from the background must be considered .
the background should be constant in time or at least not show the same time dependency as the predicted wimp signal .
the expected seasonal modulation effect is very small ( of the order of @xmath30/@xmath39 0.07 ) , requiring large masses and long counting times as well as an excellent long - term stability of the experiment .
a much stronger signature would be given by the ability to detect the axis and direction of the recoil nucleus .
in @xcite it has been shown , that the wimp interaction rate as a function of recoil energy and angle @xmath9 between the wimp velocity and recoil direction ( in the galactic frame ) is : @xmath40,\ ] ] where v@xmath41 = ( m@xmath42 + m@xmath43)@xmath19e@xmath44/2m@xmath42m@xmath45 and v@xmath46=3v@xmath47/2 .
the forward - backward asymmetry thus yields a large effect of the order of @xmath48(v@xmath49/v@xmath50)@xmath29 1 and fewer events are needed to discover a wimp signal than in the case of the seasonal modulation @xcite .
the challenge is to build massive detectors capable of detecting the direction of the incoming wimp . at present , only one such detector exists ( drift @xcite ) with a total active mass of @xmath51170 g of cs@xmath52 , with larger modules being under consideration .
first limits on wimp - nucleon cross sections were derived about twenty years ago , from at that time already existing germanium double beta decay experiments @xcite . with low intrinsic backgrounds and already operating in underground laboratories ,
these detectors were essential in ruling out first wimp candidates such as a heavy dirac neutrino @xcite .
present ge ionization experiments dedicated to dark matter searches such as hdms @xcite are limited in their sensitivity by irreducible electromagnetic backgrounds close to the crystals or from cosmogenic activations of the crystal themselves .
next generation projects based on high - purity germanium ( hpge ) ionization detectors , such as the proposed genius @xcite , gerda @xcite , and majorana @xcite experiments , aim at an absolute background reduction by more than three orders in magnitude , compensating for their inability to differentiate between electron- and nuclear recoils on an event - by - event basis . note that the primary scientific goal of these projects is to look for the neutrinoless double beta decay , at a q - value around 2039kev in @xmath53ge .
solid scintillators operated at room temperatures had soon caught up with hpge experiments , despite their higher radioactive backgrounds .
being intrinsically fast , these experiments can discern on a statistical basis between electron and nuclear recoils , by using the timing parameters of the pulse shape of a signal .
typical examples are nai experiments such as dama @xcite and naiad @xcite , with dama reporting first evidence for a positive wimp signal in 1997 @xcite .
the dama results have not been confirmed by three different mk cryogenic experiments ( cdms @xcite , cresst @xcite and edelweiss @xcite ) and one liquid xenon experiment ( zeplin @xcite ) , independent of the halo model assumed @xcite or whether the wimp - nucleon interaction is taken as purely spin - dependent @xcite .
the dama collaboration has installed a new , 250 kg nai experiment ( libra ) in the gran sasso laboratory , and began taking data in march 2003 . with lower backgrounds and increased statistics , libra should soon be able to confirm the annual modulation signal .
the zaragosa group plans to operate a 107 kg nai array ( anais ) at the canfranc underground laboratory ( 2450 mwe ) in spain @xcite , and deliver an independent check of the dama signal in nai .
cryogenic experiments operated at sub - kelvin temperatures are now leading the field with sensitivities of one order of magnitude above the best solid scintillator experiments . specifically , the cdms experiment can probe wimp - nucleon cross sections as low as 10@xmath54cm@xmath19 @xcite .
this class of experiments will be covered in more detail in section [ cryo_mk ] .
liquid noble element detectors are rapidly evolving , and seem a very promising avenue towards the goal of constructing ton - scale or even multi - ton wimp detectors .
these approaches will be discussed in section [ liquid_noble ] .
many other interesting wimp search techniques have been deployed , and it is not the scope of this paper to deliver a full overwiev . for a significantly more detailed accounting of existing and future detection techniques , as well as critical discussions of backgrounds and of the reported positive signal
, we refer to two recent reviews @xcite .
cryogenic calorimeters are meeting crucial characteristics of a successful wimp detector : low energy threshold ( @xmath5510kev ) , excellent energy resolution ( @xmath551% at 10kev ) and the ability to differentiate nuclear from electron recoils on an event - by - event basis
. their development was driven by the exciting possibility of doing a calorimetric energy measurement down to very low energies with unsurpassed energy resolution . because of the t@xmath56 dependence of the heat capacity of a dielectric crystal , at low temperatures a small energy deposition can significantly change the temperature of the absorber .
the change in temperature can be measured either after the phonons ( or lattice vibration quanta ) reach equilibrium , or thermalize , or when they are still out of equilibrium , or athermal , the latter providing additional information about the location of an event .
the astounding energy resolution comes from the fact that a much lower energy ( @xmath551mev ) is required to produce an elementary phonon excitation compared to semiconductor detectors ( @xmath511ev ) .
if a second signal , such as charge or scintillation light is detected , identification of background events induced by electron recoils is possible .
it is based on the measured difference in the ratio of charge or light to phonons for electron and nuclear recoils .
if a class of events ( such as , for example , interactions within tens of micrometers of the detector s surface ) results in incomplete charge or light collection , then such events can leak into the parameter region predicted for nuclear recoils .
as will be discussed later , no such events have been observed yet in phonon - light detectors . on the other hand ,
the timing properties of the fast phonon signal in certain phonon - charge detectors allow a discrimination between surface events and interactions in the bulk .
the cold dark matter search experiment operates low - temperature ge and si detectors at the soudan underground laboratory in minnesota ( at a depth of 2080 m.w.e . ) .
the high - purity ge and si crystals are 1 cm thick and 7.6 cm in diameter , and have a mass of 250 g and 100 g , respectively .
they are operated at a base temperature around 50mk .
superconducting transition edge sensors photolitographically patterned onto one of the crystal surfaces detect the athermal phonons from particle interactions .
the phonon sensors are divided into 4 different channels , allowing to reconstruct the x - y position of an event with a resolution of @xmath511 mm .
if an event occurs close to the detector s surface , the phonon signal is faster than for events far from the surface , because of phonon interactions in the thin metallic films .
the risetime of the phonon pulses , as well as the time difference between the charge and phonon signals allow to reject surface events caused by electron recoils .
figure [ starttimes ] shows phonon start times versus ionization yield ( charge energy divided by the total recoil energy ) for electron recoil events ( collected with a @xmath57ba source ) and nuclear recoil events ( collected with a @xmath58cf source ) . events below a yield around 0.75 typically occur within 0 - 30@xmath59 m of the surface , and can be effectively discriminated ( a typical rise time cut is shown by the horizontal grey line in the figure ) while preserving a large part of the nuclear recoil signal .
ba gamma calibration events ( diamonds ) and @xmath58cf neutron calibration events ( dots ) .
the grey lines indicate timing and ionization - yield cuts , resulting in a high - rate of nuclear recoil efficiency and a low rate of misidentified surface events.,width=302 ] charge electrodes are used for the ionization measurement .
they are divided into an inner disk , covering 85% of the surface , and an outer guard ring , which is used to reject events near the edges of the crystal , where background interactions are more likely to occur .
figure [ zips ] shows a picture of cdms detectors in their cu holders .
the discrimination against the electron recoil background is based on the fact that nuclear recoils ( caused by wimps or neutrons ) produce fewer charge pairs than electron recoils of the same energy .
the ionization yield , defined as the ratio of ionization to recoil energy , is about 0.3 in ge , and 0.25 in si for recoil energies above 20kev .
electron recoils with complete charge collection show an ionization yield of @xmath291 . for recoil energies above 10kev ,
bulk electron recoils are rejected with @xmath6099.9% efficiency , and surface events are rejected with @xmath6095% efficiency .
the two different materials are used to distinguish between wimp and neutron interactions by comparing the rate and the spectrum shape of nuclear recoil events .
while the interaction rate of neutrons is comparable in ge and si , the wimp rate is much higher in ge for spin - independent couplings .
a stack of six ge or si detectors together with the corresponding cold electronics is named a tower. five towers ( 30 detectors ) are currently installed in the cold volume at soudan , shielded by about 3 mm of cu , 22.5 cm of pb , 50 cm of polyethylene and by a 5 cm thick plastic scintillator detector which identifies interactions caused by cosmic rays penetrating the soudan rock .
results from the first tower of 4 ge and 2 si detectors @xcite , which took data at soudan from october 2003 to january 2004 , showed no nuclear recoil candidate in 53 raw live days ( see figure [ cdms_y_erecoil ] ) .
after cuts , the net exposure was 22kgdays for ge and 5kgdays for si .
the resulting upper limit on wimp - nucleon cross sections for spin - independent couplings and a standard halo is 4@xmath6110@xmath54cm@xmath19 at the 90% c.l . at a wimp mass of 60gev ( see figure [ limits ] ) , which is four times below the best previous limit reported by edelweiss @xcite .
the limits on spin - dependent wimp interactions are competitive with other experiments , in spite of the low abundance of @xmath62ge ( 7.8% ) in natural germanium .
in particular , in the case of a pure neutron coupling , cdms yields the most stringent limit obtained so far , thus strongly constraining interpretations of the dama signal region @xcite ( figure [ limits - cdms - sd ] ) . for details on the cdms analysis for spin - dependent wimp couplings we refer to jeff filippini s contribution to these proceedings @xcite .
two towers ( 6 ge and 6 si detectors ) were operated at soudan from march to august 2004 .
the data is being currently analyzed in a similar fashion to the first soudan run , in which all cuts are determined in a blind manner from calibration data , without access to the signal region .
calibration runs were interspersed with science runs , and taken with a @xmath58cf ( neutron and gamma ) and a @xmath57ba ( gamma ) source .
the high - statistics ba data also allows for a calibration to surface events , providing a population of so called ejectrons , or low energy electrons which are ejected from an adjacent detector of from material close to a detector s surface .
for an overview of the currently running experiment at soudan and expectations from the 5tower run in 2005 , we refer to walter ogburn s contribution to these proceedings @xcite .
the edelweiss experiment operates germanium bolometers at 17mk in the laboratoire souterrain de modane , at about 4800m.w.e .
the detectors are further shielded by 30 cm of paraffin , 15 cm of pb and 10 cm of cu . as in the case of cdms
, they simultaneously detect the phonon and the ionization signals , allowing a discrimination against bulk electron recoils of better than 99.9% above 15kev recoil energy .
the charge signal is measured by al electrodes sputtered on each side of the crystals , the phonon signal by a neutron transmutation doped ( ntd ) heat sensor glued onto one of the charge collection electrodes .
the ntd sensors read out the thermal phonon signal on a time scale of about 100ms .
figure [ edelw - det ] shows a picture of a 320 g germanium bolometer . between 2000 - 2003 , edelweiss performed four physics runs with five 320 g ge crystals , accumulating a total exposure of 62kgdays @xcite . above an analysis threshold of 20kev , a total of 23 events compatible with nuclear recoils
have been observed .
figure [ edelw - yplot ] shows the ionization yield versus recoil energy for one edelweiss detector for an exposure of 9.16kgdays .
the derived upper limit on spin - independent wimp - nucleon couplings under the hypothesis that all above events are caused by wimp interactions , and for a standard isothermal halo , is shown in figure [ limits ] .
the allowed region in the ( @xmath63,@xmath64 ) plane for spin - dependent wimp couplings is shown in figure [ limits - sd ] , for a wimp mass of 50gev ( @xmath65 are the effective wimp couplings to proton and neutrons ) . the edelweiss experiment has ceased running in march 2004 , in order to allow the upgrade to a second phase , with an aimed sensitivity of 10@xmath66cm@xmath19 .
the new 50 liter low - radioactivity cryostat will be able to house up to 120 detectors .
because of the inability of slow thermal detectors to distinguish between low - yield surface events and nuclear recoils and the inherent radioactivity of ntd sensors , the collaboration has been developing a new design based on nbsi thin - film sensors . these films , besides providing a lower mass and radioactivity per sensor ,
show a strong difference in the pulse shape , depending on the interaction depth of an event @xcite .
the edelweiss collaboration plans to operate twenty - one 320 g ge detectors equipped with ntd sensors , and seven 400 g ge detectors with nbsi thin - films in the new cryostat starting in 2005 . , @xmath64 ) plane for a wimp mass of 50gev .
@xmath65 are the effective wimp couplings to proton and neutrons ( see @xcite for details ) .
the allowed regions are those within the parallel straight lines or ellipses.,width=302 ] the cresst collaboration has developed cryogenic detectors based on cawo@xmath67 crystals , which show a higher light yield at low temperatures compared to other scintillating materials .
the detectors are also equipped with a separate , cryogenic light detector made of a 30@[email protected]@xmath56 silicon wafer , which is mounted close to a flat surface of the cawo@xmath67 crystal . the temperature rise in both cawo@xmath67 and light detector
is measured with tungsten superconducting phase transition thermometers , kept around 10mk , in the middle of their transition between the superconducting and normal conducting state . a schematic picture and a photograph of a cresst detector is shown in figure [ cresst - det ] .
a nuclear recoil in the 300 g cawo@xmath67 detector has a different scintillation light yield than an electron recoil of the same energy , allowing to discriminate between the two type of events when both the phonon and the light signals are observed .
the advantage of cawo@xmath67 detectors is their low energy threshold in the phonon signal , and the fact that no light yield degradation for surface events has been detected so far .
however , about 1% or less of the energy deposited in the cawo@xmath67 is seen as scintillation light @xcite .
only a few tens of photons are emitted per kev electron recoil , a number which is further diminished for nuclear recoils , because of the involved quenching factor .
the quenching factor of oxygen nuclear recoils for scintillation light is around 13.5% relative to electron recoils @xcite , leading to a rather high effective recoil energy threshold for the detection of the light signal . while neutrons will scatter predominantly on oxygen nuclei , it is expected that wimps will more likely scatter on the heavier calcium and tungsten .
the quenching factor of tungsten at room temperatures has been measured to be around 2.5% @xcite , making it difficult to observe the light signal of wimp recoils above the thermal noise .
the most recent cresst results were obtained by operating two 300 g cawo@xmath67 detectors at the gran sasso underground laboratory ( 3800 m.w.e ) for two months at the beginning of 2004 @xcite . the total exposure after cuts was 20.5kgdays .
the energy resolution in the phonon channel was 1.0kev ( fwhm ) at 46.5kev , the low - energy gamma line being provided by an external @xmath68pb contamination .
a total of 16 events were observed in the 12kev - 40kev recoil energy region , a number which seems consistent with the expected neutron background , since the experiment had no neutron shield at this stage .
no phonon - only events ( as expected for wimp recoils on tungsten ) were observed between 12kev - 40kev in the module with better resolution in the light channel , yielding a limit on coherent wimp interaction cross sections very similar to the one obtained by edelweiss .
cresst has stopped taking data in march 2004 , to upgrade with a neutron shield , an active muon veto , and a 66-channels squid read - out system .
it will allow to operate 33 cawo@xmath67 detector modules , providing a total of 10 kg of target material .
the upgrade will be completed in early 2005 and the expected sensitivity is around 10@xmath66cm@xmath19 .
liquid xenon ( lxe ) has excellent properties as a dark matter detector .
it has a high density ( 3g / cm@xmath56 ) and high atomic number ( z=54 , a=131.3 ) , allowing experiments to be compact .
the high mass of the xenon nucleus is favorable for wimp scalar interactions provided that a low energy threshold can be achieved ( see figure [ rates ] for a comparison with other target nuclei ) .
lxe is an intrinsic scintillator , having high scintillation ( @xmath70 = 178 nm ) and ionization yields because of its low ionization potential ( 12.13 ev ) .
there are no long - lived radioactive xenon isotopes , and other impurities ( such as @xmath71kr ) can be reduced to a very low level by centrifugation or with a distillation tower and a cold trap .
krypton contamination levels as low as 1ppb ( parts per billion ) have already been achieved @xcite .
scintillation in liquid xenon is produced by the formation of excimer states , which are bound states of ion - atom systems .
if a high electric field ( @xmath511 kv / cm ) is applied , ionization electrons can also be detected , either directly or through the secondary process of proportional scintillation .
the elastic scattering of a wimp produces a low - energy xenon recoil , which loses its energy through ionization and scintillation .
both signals are quenched when compared to an electron recoil of the same energy , but by different amounts , allowing to use the ratio for distinguishing between electron and nuclear recoils .
the quenching factor for both scintillation and ionization depends on the drift field and on the energy of the recoil .
the boulby dark matter collaboration has been operating a single - phase liquid xenon detector , zeplini , at the boulby mine ( @xmath513000 m.w.e . ) during 2001 - 2002 .
the zeplini detector had a fiducial mass of 3.2 kg of liquid xenon , viewed by 3 pmts through silica windows .
it was inclosed in a 0.93 ton active scintillator veto , which helped in reducing the background from the radioactivity of the pmts and surroundings .
a picture inner detector is shown in figure [ zeplin - det ] .
a total exposure of 293kgdays had been accumulated . with a light yield of 1.5 electrons / kev ,
the energy threshold was at 2kev electron recoil ( corresponding to 10kev nuclear recoil energy for a quenching factor of 20% ) .
a discrimination between electron and nuclear recoils was applied by using the difference in the mean time of the corresponding pulses . to establish this difference ,
the detector had been calibrated with gamma sources ( above ground and at the mine ) , and with neutron sources above ground . using this statistical discrimination , a preliminary limit on spin - independent wimp cross sections comparable to cresst and
edelweiss has been achieved ( see figure [ limits ] ) .
the collaboration has developed two concepts for dual - phase detectors , zeplinii and zepliniii .
zeplinii will have a 30 kg fiducial target mass , observed by 7 pmts .
zepliniii will operate a lower target mass ( 6 kg liquid xenon viewed by 31 pmts ) at a higher field ( @xmath60 5kv / cm ) .
it is expected that both zeplinii and zepliniii will be deployed at the boulby mine in 2005 .
the japanese xmass collaboration develops liquid xenon detectors for solar neutrinos , double beta decay and dark matter searches @xcite . since 2003 , they have been operating a 1 kg dual - phase detector at the kamioka mine ( 2700 mwe ) .
the active volume was viewed by two uv - sensitive pmts through mgf@xmath52 windows , the detector being operated in the low drift field regime ( 250v / cm ) @xcite . at such fields , the proportional light signal is large for electron recoils , but very small or zero for nuclear recoils .
events which yield a signal only in the light channel can thus mimic a potential wimp candidate . a second prototype with an active volume of 15 kg viewed by 14 pmts is currently under operation at the kamioka underground laboratory .
a 100 kg single - phase detector is also being operated at kamioka @xcite .
thirty liters of liquid xenon are contained in a ( 31cm)@xmath56 oxygen free copper vessel , and viewed by 54 low - background , 2 inch pmts .
so far , the background levels are consistent within a factor of two with expectations , although an internal @xmath72rn source has been detected @xcite .
the us xenon collaboration plans to deploy and operate a 10 kg dual - phase detector in the gran sasso underground laboratory ( 3500 m.w.e ) by 2005 - 2006 . at present , a 3 kg prototype is under operation above ground , at the columbia nevis laboratory @xcite .
the detector is operated at a drift field of 1kv / cm , and both primary and proportional light are detected by an array of seven 2inch pmts operating in the cold gas above the liquid .
the active volume is defined by ptfe walls and be - cu wire grids with high optical transmission .
the detector is insulated by a vacuum cryostat and cooled down to a stable temperature of ( [email protected])c with a pulse tube refrigerator ( see figure [ 3kg - schematic ] for a schematic drawing of the prototype and figure [ 3kg - picture ] for a picture of the cryostat . ) . the xenon gas is continuously circulated and purified using a high temperature saes getter @xcite .
the performance of the chamber was tested with gamma ( @xmath74co ) , alpha ( @xmath68pb ) and neutron ( @xmath75ambe ) sources .
the depth of an event is reconstructed by looking at the separation in time between the primary and proportional scintillation signal ( the electron drift velocity for the applied electric field is known ) .
the x - y position is inferred with a resolution of 1 cm from the center of gravity of the proportional light emitted close to the seven pmts .
the measured ratio of proportional light ( s2 ) to direct light ( s1 ) for alpha recoils is 0.03 if the corresponding ratio for gamma events ( electron recoils ) is normalized to 1 , providing a very clear separation between these type of events .
more interesting is the ratio s2/s1 for nuclear recoil events .
it was established using a @xmath75ambe neutron source , by selecting events which were tagged as neutron recoils in a separate neutron detector placed under a scattering angle of 130deg .
if the s2/s1 ratio for electron recoils ( provided by a @xmath76cs source ) is again normalized to 1 , then s2/s1 for nuclear recoils was measured to be around 0.1 , the leakage of electron recoils into the s2/s1 region for nuclear recoils being @xmath77 ( for details see @xcite ) .
figure [ xenon - nrecoils ] shows a histogram of s2/s1 for events taken with the @xmath75ambe source , compared to the corresponding distribution of events from the @xmath76cs gamma source .
these first measurements are very encouraging , and will be further improved by using a csi photocathode immersed in the liquid .
the downwards going primary photons are converted into electrons , which are then drifted into the gas phase , where electroluminiscence occurs and a tertiary signal is observed by the pmts .
first tests of a csi photocathode in the 3 kg prototype have confirmed the high quantum efficiency ( around 12% at 1kv / cm ) which had been previously measured @xcite .
the first xenon detector with a fiducial mass of 10 kg ( xenon10 ) to be operated in gran sasso is currently under construction .
its goal is to achieve a sensitivity of a factor of 20 below the current cdms results , thus probing wimp cross sections around 2@xmath6110@xmath66cm@xmath19 .
we live in suspenseful times for the field of direct detection : for the first time , a couple of experiments operating deep underground probe the most optimistic supersymmetric models .
the best limits on wimp - nucleon cross sections come from cryogenic experiments with ultra - low backgrounds and excellent event - by - event discrimination power ( cdms , edelweiss , cresst ) .
although these experiments had started with target masses around 1 kg , upgrades to several kilograms have already taken place or are foreseen for the near future , ensuring ( along with improved backgrounds ) an increase in sensitivity by a factor of 10 - 100 .
other techniques , using liquid noble elements such as xenon , may soon catch up and probe similar parameter spaces to low - temperature cryogenic detectors .
it is worth emphasizing here that given the importance of the endeavor and the challenge in unequivocally identifying and measuring the properties of a dark matter particle , it is essential that more than one technique will move forward . if supersymmetry is the answer to open questions in both cosmology and particle physics , then wimp - nucleon cross sections as low as 10@xmath78cm@xmath19 are likely @xcite .
likewise , in theories with universal extra dimensions , it is predicted that the lightest kaluza klein particle would have a scattering cross section with nucleons in the range of 10@xmath79 - 10@xmath78cm@xmath19 @xcite .
thus , to observe a signal of a few events per year , ton or even multi - ton experiments are inevitable .
there are several proposals to build larger and improved dark matter detectors ( see @xcite for an exhaustive list ) .
the selection presented below is likely biased , but based on technologies which seem the most promising to date .
the supercdms project is a three - phase proposal to utilize cdms - style detectors with target masses growing from 27 kg to 145 kg and up to 1100 kg , with the aim of reaching a final sensitivity of 3@xmath6110@xmath79cm@xmath19 by mid 2015 .
this goal will be realized by developing improved detectors ( for a more precise event reconstruction ) and analysis techniques , and at the same time by strongly reducing the intrinsic surface contamination of the crystals .
a possible site for supercdms is the recently approved sno - lab deep - site facility in canada ( at 6000 m.w.e . ) , where the neutron background would be reduced by more than two orders of magnitude compared to the soudan mine , thus ensuring the mandatory conditions to build a zero - background experiment . for details on the supercdms project
we refer to paul brink s contribution to these proceedings @xcite . in europe , a similar project to develop a 100kg-1ton cryogenic experiment , eureca ( european underground rare event search with calorimeter array ) @xcite is underway .
the xenon collaboration is designing a 100 kg scale dual - phase xenon detector ( xenon100 ) , towards a modular one tonne experiment .
the baseline goal is 99.5% background rejection efficiency above a 16kev nuclear recoil energy threshold , full 3-d localization of an event and a liquid xenon self shielding surrounding the detector .
the aim of xenon100 is to probe the parameter space down to cross sections of 10@xmath80cm@xmath19 .
the full scale , one tonne experiment ( xenon1 t ) , which will operate ten xenon100 modules , will increase this sensitivity by another order of magnitude @xcite .
zeplin max , a r&d project of the boulby dark matter collaboration , is a further proposal to build a ton scale liquid xenon experiment .
the design will be based on the experience and results with zeplin ii / iii at the boulby mine @xcite . in looking back over the fantastic progress made in the last couple of years , and extrapolating into the future
, it seems probable that these , and other proposed projects , will have a fair chance to discover a wimp signature within the present decade . in conjunction with indirect searches and accelerator production of new particles at the weak scale
, they will allow to reveal the detailed properties of wimps , such as their mass , spin and couplings to ordinary matter , and shed light on their velocity and spatial distribution in our galactic halo .
f. zwicky , helv .
acta 6 ( 1993 ) 110 .
v. rubin and w.k .
ford jr . , astrophysical journal 159 ( 1970 ) 379 .
bennet et al . , ap .
j. suppl . 148
( 2003 ) 1 .
f. prada et al .
, astrophys .
j. 598 ( 2003 ) 260 .
gaitskell , direct detection of dark matter , ann .
nucl . part .
54 ( 2004 ) 315 - 359 . g. chardin , dark matter direct detection , astro - ph/0411503 , 2004 .
g. bertone , d. hopper , j. silk , particle dark matter : evidence , candidates and constraints , hep - ph/0404175 .
m. w. goodman and e. witten , phys .
d 31 ( 1985 ) 3059 .
j. d. lewin and p. f. smith , astrop .
phys . 6 ( 1996 ) 87 .
kim et al . , j. high energy phys .
0212 ( 2002 ) 034 , e. a. baltz and p. gondolo , j. high energy phys . 0410 ( 2004 ) 052 ,
m. battaglia et al . , eur .
j. c33 ( 2004 ) 273 .
a. drukier , k. freese and d. spergel , phys .
d 33 ( 1986 ) 3495 . k. freese , j. frieman and a. gould , phys .
d 37 ( 1988 ) 3388 . d.n .
spergel , phys .
d 37 ( 1988 ) 1353 . c.j .
copi , j. heo and l.m .
krauss , phys .
b 461 ( 1999 ) 43 .
d. p. snowden - ifft et al . , nucl .
instrum . and meth . in phys .
res . a 498 ( 2003 ) 155 - 164 .
ahlen et al . , phys .
b 195 ( 1987 ) 603 ; d.o . caldwell et al . , phys .
61 ( 1988 ) 510 ; d. reusser et al . ,
b 255 ( 1991 ) 143 . m. beck et al .
( heidelberg - moscow collaboration ) , phys .
b 336 ( 1994 ) 141 . l. baudis et al . , nucl .
inst . meth .
a 385 ( 1997 ) 265 ; l. baudis , a. dietz , b. majorovits , f. schwamm , h. strecker , h. v. klapdor - kleingrothaus phys . rev .
d 63 , 022001 ( 2001 ) 022001 , h. v. klapdor - kleingrothaus et al .
, astropart.phys .
18 ( 2003 ) 525 - 530 .
klapdor - kleingrothaus , nucl .
b 110 ( 2002 ) 364 , l. baudis et al . , nucl . instr . meth . in phys .
res . a 426 ( 1999 )
425 - 435 .
r. gaitskell et al . , ( majorana collaboration ) , nucl - ex/0311013 .
r. bernabei et al .
nuovo cim .
26 ( 2003 ) 1 .
b. ahmed et al .
( the boulby dark matter collaboration ) , astropart.phys . 19 ( 2003 ) 691 - 702 .
r. bernabei et al .
b 424 ( 1998 ) 195 , phys . lett .
b 450 ( 1999 ) 448 .
d. akerib et al .
( cdms collaboration ) phys .
d 68 ( 2003 ) 082002 .
d. akerib et al .
( cdms collaboration ) , first results from the cryogenic dark matter search in the soudan underground lab , phys .
93 ( 2004 ) 211301 . c. angloher et al . ,
limits on wimp dark matter using scintillating cawo@xmath67 cryogenic detectors with active background suppression , astro - ph/0408006 , accepted for publication in astroparticle physics .
a. benoit et al .
( edelweiss collaboration ) , phys .
b 545 ( 2002 ) 43 .
v. sanglard et al .
( edelweiss collaboration ) , astro - ph/0503265 , submitted to prd .
v. a. kudryavtsev ( for the bdm collaboration ) , dark matter experiments at boulby mine , astro - ph/0406126 .
h. baer et al . , jcap 0309 ( 2003 ) 007 , e. baltz and p. gondolo , physical review d 67 ( 2003 ) 063503 , a. bottino et al . , phys.rev .
d69 ( 2004 ) 037302 , u. chattopadhyay , a. corsetti and p. nath , phys.atom.nucl .
67 ( 2004 ) 1188 - 1194 , m. battaglia et al .
, eur.phys.j .
c33 ( 2004 ) 273 - 296 y. g. kim et al .
, jhep 0212 ( 2002 ) 034 , a. b. lahanas and d. v. nanopoulos , phys.lett .
b568 ( 2003 ) 55 - 62 , a. masiero , s. profumo and p. ullio , hep - ph/0412058 .
p. ullio , m. kamionkowski , p. vogel , jhep 0107 ( 2001 ) 044 . c. savage , p. gondolo and k. freese , phys .
d70 ( 2004 ) 123513 . c. j. copi and l. m. krauss , phys.rev .
d67 ( 2003 ) 103507 .
s. cebrian et al . , nucl.phys.proc.suppl .
114 ( 2003 ) 111 - 115 .
d. akerib et al .
( cdms collaboration ) , in preparation .
p. brink , beyond the cdms - ii dark matter search : supercdms .
j. filippini , limits on spin - dependent wimp - nucleon interactions from the cryogenic dark matter search , these proceedings .
r. w. ogburn , the on - going cdms - ii dark matter search at the soudan mine , these proceedings .
v. sanglard , from edelweiss - i to edelweiss - ii , astro - ph/0411629 .
edelweiss collaboration , sensitivity of the edelweiss wimp search to spin - dependent interactions , astro - ph/0412061 , submitted to phys .
b. n. mirabolfati et al .
, aip conference proceedings , volume 605 ( 2001 ) 517 .
e. aprile ( xenon collaboration ) , the xenon dark matter experiment , astro - ph/0502279 , in proceedings of the fifth international heidelberg conference on dark matter in astro and particle physics , october 3 - 9 , 2004 .
e. aprile et al . , astro - ph/0207670 , in proceedings of the international workshop on techniques and applications of xenon detectors ( xenon01 ) , icrr , univ . of tokyo ,
kashiwa , japan , december 2001 ( world scientific ) .
e.aprile et al . , submitted to phys .
d ( 2005 ) .
y. takeuchi ( xmass collaboration ) , recent status of the xmass project , 32nd international conference on high energy physics , august 16 - 22 , 2004 , beijing , china .
m. yamashita , t. doke , j. kikuchi , s. suzuki , double phase ( liquid / gas ) xenon scintillation detector for wimps direct search , astroparticle physics 20 ( 2003 ) 79 - 84 .
saes pure gas , inc . :
http://www.saesgetters.com/. e. aprile et al . , nucl .
inst . meth . a 338 ( 1994 )
j. ellis , k. a. olive , y. santoso and v. c. spanos , update on the direct detection of supersymmetric dark matter , hep - ph/0502001 .
g. servant and t. m. tait , new j. phys . 4 ( 2002 ) 99 .
p. brink ( for the supercdms collaboration ) , beyond the cdms - ii dark matter search : supercdms , these proceedings .
|
we have strong evidence on all cosmic scales , from galaxies to the largest structures ever observed , that there is more matter in the universe than we can see .
galaxies and clusters would fly apart unless they would be held together by material which we call dark , because it does not shine in photons .
although the amount of dark matter and its distribution are fairly well established , we are clueless regarding its composition .
leading candidates are weakly interacting massive particles ( wimps ) , which are cold thermal relics of the big bang , ie moving non - relativistically at the time of structure formation .
these particles can be detected via their interaction with nuclei in deep - underground , low - background detectors .
experiments dedicated to observe wimp interactions for the first time reach sensitivities allowing to probe the parameter space predicted by supersymmetric theories of particle physics .
current results of high sensitivity direct detection experiments are discussed and the most promising projects of the future are presented . if a stable new particle exists at the weak scale , it seems likely to expect a discovery within this decade .
|
strange non - chaotic attractors ( snas ) are geometrical structures existing between periodic and chaotic attractors .
snas have been insensitive to initial conditions as evidenced by the negative values of their lyapunov exponents . however , they have the complicated geometrical structure of fractals .
snas are generic in quasiperiodically forced systems . after pioneering work of grebogi @xcite , a large number of quasiperiodic systems were reported for the existence of sna behavior .
different routes , such as the heagy - hammel or torus doubling , fractalization , intermittency , blow out bifurcation routes , to sna have been identified @xcite .
a good number of nonlinear systems and electronic circuits exhibiting snas in their dynamics have been studied numerically and experimentally @xcite while a few systems have been studied analytically @xcite .
the phenomenon of chaos synchronization finding potential applications in secure communication has been enchanting researchers after the _ master - slave _ concept introduced by pecora and carroll @xcite .
different types of synchronization phenomenon , such as complete , phase , lag , anti - phase and generalized synchronization , have been identified in coupled chaotic systems .
complete synchronization of identical chaotic systems have been confirmed through the negative values of the _ master stability function ( msf)_@xcite . the msfs for a few nonlinear systems and simple electronic circuits have been studied @xcite .
an explicit analytical solution explaining the complete synchronization of identical @xmath0 circuits has been presented @xcite .
synchronization of coupled driven - damped squids exhibiting snas in their dynamics have been synchronized using _ in - phase driving _
method @xcite .
however , the nonlinear systems exhibiting snas in their dynamics have not been studied analytically hitherto .
recently , the phenomenon of complete synchronization has been observed in coupled quasiperiodically - forced systems through unidirectional coupling of the systems@xcite . in this paper , we present an explicit analytical solution to the complete synchronization phenomenon observed in the above said unidirectionally - coupled system exhibiting snas in their dynamics .
+ this paper is divided into two sections . in section [ sec:1 ]
we present the normailized state equations of the circuit and summarize the analytical solutions of the individual circuit system through phase portraits and poincare sections .
explicit analytical solutions to the coupled state equations are presented in section [ sec:2 ] .
further , the analytical solutions obtained in this section have been used to study the synchroniation dynamics through phase portraits .
the stability of the synchronized state is confirmed through the msf .
the circuit under present study consists of a forced , series lcr circuit with a chua s diode connected parallel to the capacitor .
the circuit is subjected to sinusoidal external forcing by two voltage sources @xmath1 .
a schematic diagram of the circuit is as shown in fig .
[ fig:1 ] .
the state equations of the circuit is written as @xmath2 where @xmath3 and @xmath4 is the piecewise linear function given by @xmath5,\end{aligned}\ ] ] where @xmath6 and @xmath7 being the voltage across the capacitor @xmath8 and the current through the inductor @xmath9 , respectively . the value of the negative slopes of the inner , outer regions and the breakpoints in the @xmath10 characteristic curve of the nonlinear element are given as @xmath11 and @xmath12 respectively . with the rescaling parameters @xmath13 , @xmath14 , @xmath15 = @xmath8/@xmath16 , @xmath17 = @xmath18 , @xmath19 = @xmath20/@xmath21 , @xmath22 = @xmath23/@xmath21 , @xmath24 = @xmath25/@xmath26 , @xmath27 = @xmath28/@xmath26 , and @xmath29 = @xmath30/@xmath31
, the normalized state equations of the quasiperiodically - forced circuit can be written in the autonomous form as @xmath32 [ eqn:3 ] where , @xmath33 and @xmath34 with the circuit parameters taking the values @xmath35 and @xmath36 , the rescaled values of the normalized state variables takes the values , @xmath37 , a=-1.0944 , b=-0.5904 , @xmath38 and @xmath39 .
the amplitude of the second force @xmath40 has been taken as the control parameter .
the sna dynamics of the circuit exhibiting the _ heagy - hammel _ and _ fractalization _ routes has been studied analytically @xcite .
the analytical solutions obtained for each piecewise linear region is summarized as follows .
+ since the roots @xmath41 in the @xmath42 region are real and unequal , the fixed point @xmath43 corresponding to the @xmath44 region is a _ saddle _ or _ hyperbolic fixed point_. the state variables @xmath45 and @xmath46 are @xmath47 @xmath48 the roots @xmath49 in the @xmath50 region are a pair of complex conjugates with a negative real part .
hence the fixed points @xmath51 corresponding to the @xmath52 regions is a _ stable spiral fixed point _ .
the state variables @xmath45 and @xmath46 are @xmath53 @xmath54 where @xmath55 and @xmath56 , @xmath57 corresponds to @xmath58 and @xmath59 regions respectively .
the ratio of the frequencies , ( @xmath60 ) has been found to be an integral multiple of the golden ratio , i.e. @xmath61 .
hence , the system must exhibit sna dynamics for a proper choice of the amplitudes of the external forcing terms @xmath62 and @xmath40 .
the sna dynamics of the circuit exhibiting the hh route has been studied analytically through phase portraits and poincare maps .
[ fig:2 ] shows the phase portraits and their corresponding poincare maps for a 1-torus for @xmath63 , a 2-torus for @xmath64 and a sna followed by chaos in the range @xmath65 .
[ fig:2](a(i))-[fig:2](a(iv ) ) represent the phase portraits of the heagy - hammel route in the @xmath66 phase plane .
the correponding poincare maps of the hh route are shown in fig . [ fig:2](b ) . in the next section
, we present the normalized state equations of the response system and explain the the analytical solutions obtained for the coupled system .
if we consider the system given by eq .
( [ eqn:3 ] ) acting as the drive system drives a response system with state variables @xmath67 ) then the normalized state equations of the response system is given by @xmath68 [ eqn:10 ] where , @xmath69 the circuit parameters of the response system takes the same value as that of the drive system . the drive and the response systems given by eqs .
( [ eqn:3 ] ) and ( [ eqn:10 ] ) could be used to obtain explicit analytical solutions for studying the synchronization dynamics . from the state equations of the response system given by eqs .
( [ eqn:10 ] ) , we observe that the dynamics of the response is influenced by the drive through the coupling parameter . because the circuit equations are piecewise linear , each picewise linear region of the two sytems
could be coupled together to get a new set of equations which could be solved for each region .
the difference system obtained from eqs .
( [ eqn:3 ] ) and ( [ eqn:10 ] ) are @xmath70 [ eqn:12 ] where @xmath71=@xmath72 , @xmath73=@xmath74 and @xmath75 takes the values @xmath76 or @xmath77 depending upon the region of operation of the drive and response system . from the new set of state variables @xmath78
, the state variables of the response system @xmath79 could be written as @xmath80 [ eqn:13 ] one can easily establish that a unique equilibrium point @xmath81 exists for eq .
[ eqn:12 ] in each of the following three subsets @xmath82 form eq .
( [ eqn:14 ] ) it has been found that the origin @xmath43 is the fixed point for all the three piecewise linear regions of the difference system . in the first case , @xmath83 and @xmath84
take the values @xmath85 and @xmath86 respectively , which has been taken as the @xmath87 region of the difference system .
the stability determining eigenvalues are calculated from the stability matrix @xmath88 the eigenvalues in this region are real , negative and distinct for @xmath89 while they are a pair of complex conjugates with negative real parts for @xmath90 .
hence , the stability of the fixed point transforms from a _
stable node _ to a _ stable spiral _ for @xmath90 .
+ in the second case , @xmath83 and @xmath84 take the values @xmath91 and @xmath92 respectively , which has been taken as the @xmath93 regions of the difference system .
the stability determining eigen values are calculated from the stability matrix @xmath94 the eigenvalues in this region are found to be a pair of complex conjugates with a negative real part , for all the values of the coupling strength .
[ fig:3 ] shows the bifurcation of the real eigenvalues of the jacobian matrix given by eq .
( [ eqn:15 ] ) in the @xmath95 as a function of the coupling parameter .
the red and green lines show the two real roots while the blue line shows the real part of the complex conjugate roots .
the tranformation of the eigenvalues at the critical value of the coupling paramter @xmath96 indicates the mechanism of synchronization .
when the coupling paramter @xmath97 , the coupled systems become independent of each other .
hence the drive and the response systems given by eqs .
( [ eqn:3 ] ) and ( [ eqn:10 ] ) have the same solution for their state variables in all the three piecewise linear regions .
the systems are operated with different set of initial conditions given by ( @xmath98 ) and ( @xmath99 ) .
owing to a difference in the initial conditions , the snas of the drive and response become unsynchronized for @xmath100 .
the analytical solutions to the state variables in each of the three regions are as given in eqs .
( [ eqn:6])- ( [ eqn:8 ] ) .
+ now we present explicit analytical solutions for the dynamics of the response system for coupling strengths @xmath101 , leading to complete synchronization of the drive and the response .
this is achieved by finding a solution to the normalized state variables of the difference system given by eq .
( [ eqn:12 ] ) .
the solution of those equations are , @xmath102^t$ ] for which the initial conditions are written as @xmath103 @xmath104 . since eq .
( [ eqn:12 ] ) is piecewise linear , the solution to each of the three piecewise linear regions can be obtained explicitly . in the @xmath87 region @xmath83 and @xmath84
takes the values @xmath85 and @xmath86 respectively . hence the normalized equations obtained from eqs .
( [ eqn:12 ] ) are @xmath105 [ eqn:17 ] differentiating eq .
( [ eqn:17]b ) with respect to time and using eqs .
( [ eqn:17]a , [ eqn:17]b ) in the resultant equation , we obtain @xmath106 where , @xmath107 and @xmath108 .
the roots of eq .
( [ eqn:18 ] ) are @xmath109 . from stability analysis , for @xmath89 , @xmath110 , and hence the roots @xmath111 and @xmath112 are real , negative and distinct . using the method of undetermined coefficients , the general solution to eq .
( [ eqn:18 ] ) can be written as @xmath113 where @xmath114 and @xmath115 are integration constants and @xmath116 [ eqn:20 ] the constants @xmath114 and @xmath115 are given as , @xmath117 [ eqn:21 ] differentiating eq .
( [ eqn:18 ] ) and using it in eq .
( [ eqn:17]b ) we get @xmath118 from the results of @xmath119 obtained from eqs .
( [ eqn:19 ] ) , ( [ eqn:22 ] ) and @xmath120 obtained from eqs .
( [ eqn:6 ] ) , ( [ eqn:7 ] ) , @xmath121 and @xmath122 can be obtained from eqs .
( [ eqn:13 ] ) . for @xmath90 , @xmath123 and the roots @xmath111 and @xmath112 are a pair of complex conjugates with a negative real part given as @xmath124 , with @xmath125 and @xmath126 .
the general solution to eq.(13 ) can be written as , @xmath127 where the constants @xmath128 are as given in given in eq .
( [ eqn:20 ] ) .
differentiating eq .
( [ eqn:23 ] ) and using it in eq .
( [ eqn:17]b ) we get , @xmath129 the constants @xmath114 and @xmath115 are given as , @xmath130 [ eqn:25 ] from the results of @xmath119 obtained from eqs .
( [ eqn:23 ] ) , ( [ eqn:24 ] ) and @xmath120 obtained from eqs .
( [ eqn:6 ] ) , ( [ eqn:7 ] ) , @xmath121 and @xmath122 can be obtained from eqs .
( [ eqn:13 ] ) . in the @xmath93 region @xmath83 and @xmath84 takes the values @xmath131 . hence the normalized equations obtained from eqs .
( [ eqn:12 ] ) are @xmath132 [ eqn:26 ] differentiating eq .
( [ eqn:26]b ) with respect to time and using eqs .
( [ eqn:26]a , [ eqn:26]b ) in the resultant equation , we obtain @xmath133 where , @xmath134 and @xmath135 . from the stability analysis
we could confirm that the roots @xmath136 and @xmath137 are a pair of complex conjugates given as @xmath138 , with @xmath139 and @xmath140 , for all values of the coupling strength .
hence the state variables , @xmath141 can be written as @xmath142 @xmath143 the constants @xmath144 and @xmath145 are the same as the contants given in eq .
( [ eqn:20 ] ) and ( [ eqn:25 ] ) respectively , except that the constants @xmath146 are replaced with @xmath147 . from the results of @xmath119 obtained from eqs .
( [ eqn:28 ] ) , ( [ eqn:29 ] ) and @xmath120 obtained from eqs .
( [ eqn:8 ] ) , ( [ eqn:9 ] ) , @xmath121 and @xmath122 can be obtained from eq .
( [ eqn:13 ] ) .
now let us briefly explain how the solution can be generated in the @xmath148 phase space . with the _ time ( t ) _ being considered as the independent variable ,
the state variables evolves within each piecewise linear region depending upon its initial values .
if we start with the initial conditions @xmath149 in the @xmath87 region at time @xmath150 , the arbitrary constants @xmath114 and @xmath115 get fixed .
thus @xmath151 evolves as given by eq .
( [ eqn:22 ] ) up to either @xmath152 , when @xmath153 or @xmath154 when @xmath155 .
the next region of operation @xmath156 or @xmath157 thus depends upon the value of @xmath71 in the @xmath95 region at that instant of time . as the trajectory enters into the next region of interest , the arbitrary constant corresponding to that region
could be evaluated , with the initial conditions to that region being either ( @xmath158 ) or ( @xmath159 ) . during each region of operation , the state variables of the response system evolves as @xmath160 and @xmath161 , respectively .
the procedure can be continued for each successive crossing . in this way
, the explicit solutions can be obtained in each of the regions @xmath162 , @xmath163 of the response system .
the solution obtained in each region has been matched across the boundaries and used to generate the dynamics of the response system .
+ the analytical solutions obtained above for the response system can be used to explain the phenomena of complete synchronization through phase portraits and timeseries analysis .
the initial conditions of the drive and the response systems are so chosen such that the snas of the two system exist in different regions of phase space .
the initial conditions of the drive and response systems has been fixed as ( @xmath164=-0.5 , @xmath165=0.1 ) and ( @xmath166 = 0.5 , @xmath167=0.11 ) respectively .
[ fig:4](a ) , [ fig:4](b ) shows the phase portraits of the snas of drive and response systems .
the unsynchronized state of the coupled system for the coupling parameter @xmath97 , in the @xmath168 phase plane and their corresponding trajectory in the @xmath169 plane is shown in fig . [
fig:4](c ) and [ fig:4](d ) respectively . for
@xmath90 , the fixed point of the @xmath95 region transform into a _
stable spiral _ indicating the asymptotic convergence of trajectories towards the origin , within the synchronization manifold .
the phase portraits of the drive and the response systems for the value of the coupling parameter @xmath170 is shown in fig .
[ fig:5](a ) , [ fig:5](b ) , respectively .
[ fig:5](c ) shows the complete synchronization of the coupled systems in the @xmath168 phase plane and their corresponding trajectory in the @xmath169 plane as in fig .
[ fig:5](d ) . from the phase portraits and the time series plots obtained it could be inferred that for the coupling parameter taking the value @xmath171 , the response system which is operating with a different set of initial condition and for a different value of external periodic force , completely synchronizes with the drive .
the stability of the synchronized state for the x - coupled system is confirmed through the _ msf _ obtained from the normalized state equations [ eqn:3 ] and [ eqn:10 ] .
the msf for the coupled sna system is as shown in fig .
[ fig:6 ] .
the figure shows a broader stable synchronized region within @xmath172 indicated by the negative values of @xmath173 , confirming the synchronization of the coupled system .
we have presented in this paper an explicit analytical solution to the normalized state equations of coupled quasiperiodically - forced systems exhibiting snas in their dynamics .
the solutions thus obtained have been used to explain the phenomenon of complete synchronization observed in the coupled system through phase portraits .
analytical solutions for the synchronization of simple chaotic systems have been reported recently @xcite . to the best of our knowledge ,
this is the first time that synchronization of snas is studied analytically . c. grebogi , e. ott , s. pelikan , and j. a. yorke , physica d * 13 * , 261 ( 1984 ) .
j. f. heagy and s. m. hammel , physica d * 70 * , 140 ( 1994 ) .
a. venkatesan , m. lakshmanan , a. prasad and r. ramasamy , phys .
e * 61 * , 3641 ( 2000 ) .
a. prasad , s. s. negi and r. ramasamy , int .
chaos * 11 * , 291 ( 2001 ) .
m. agarwal , a. prasad and r. ramasamy , phys .
e * 81 * , 026202 ( 2010 ) .
k. srinivasan , d. v. senthil kumar , r. suresh , k. thamilmaran and m. lakshmanan , int .
chaos * 19 * , 4131 ( 2009 ) .
k. thamilmaran , d. v. senthil kumar , a. venkatesan and m. lakshmanan , phys .
e * 74 * , 036205 ( 2006 ) .
a. venkatesan , k. murali and m. lakshmanan , phys .
a * 259 * , 246 ( 1999 ) .
z. liu and z. zhu , int .
chaos * 6 * , 1383 ( 1996 ) .
a. arulgnanam , a. prasad , k. thamilmaran and m. daniel , chaos solitons fract . * 75 * , 96 ( 2015 ) .
a. arulgnanam , a. prasad , k. thamilmaran and m. daniel , int .
chaos * 25 * , 1530020 ( 2015 ) .
g. sivaganesh , j. korean phys .
soc . * 69 * , 124 ( 2016 ) .
g. sivaganesh , chin .
* 52 * , 1760 ( 2014 ) .
m. pecora and l. carroll , phys .
64 * , 821 ( 1990 ) .
m. pecora and l. carroll , ieee trans .
. sys . * 38 * , 453 ( 1991 ) .
m. pecora , l. carroll , a. johnson , j. mar and j. f. heagy , chaos * 7 * , 520 ( 1997 ) .
m. pecora and l. carroll , phys .
* 80 * , 2109 ( 1998 ) .
m. pecora and l. carroll , int .
chaos * 9 * , 2315 ( 1999 ) .
l. huang , q. chen , y. c. lai and m. pecora , phys .
e * 80 * , 036204 ( 2009 ) .
g. sivaganesh , j. korean phys . soc . * 68 * , 628 ( 2016 ) .
g. sivaganesh , chin .
lett . , * 32 * , 010503 ( 2015 ) .
p. r. venkatesh , a. venkatesan , m. lakshmanan , pramana * 86 * , 1195 ( 2016 ) .
r. ramasamy , phys .
e * 56 * , 7294 ( 1997 ) .
|
in this paper we present an analytical study on the synchronization dynamics observed in unidirectionally - coupled quasiperiodically - forced systems that exhibit _ strange non - chaotic attractors ( sna ) _ in their dynamics .
the sna dynamics observed in the uncoupled system is studied analytically through phase portraits and poincare maps .
a difference system is obtained by coupling the state equations of similar piecewise linear regions of the drive and response systems .
the mechanism of synchronization of the coupled system is realized through the bifurcation of the eigenvalues in one of the piecewise linear regions of the difference system .
the analytical solutions obtained for the normalized state equations in each piecewise linear region of the difference system has been used to explain the synchronization dynamics though phase portraits and timeseries analysis .
the stability of the synchronized state is confirmed through the _ master stability function_. an explicit analytical solution explaining the synchronization of snas is reported in the literature for the first time .
|
this paper is a theoretical study of multiplicity signatures in coherent production of charged - pairs of parabosons [ 1 - 3 ] of order @xmath0 in comparison with those arising in the case of ordinary bosons , @xmath1 .
the investigation is partially motivated by phenomenological coherence analyses , circa 1970 , of inelastic @xmath12 pair production from fixed targets with laboratory kinetic - energies up to 27 gev by c.p . wang [ 4 ] , and by horn and silver[5 ] .
the present analysis is possible because the conserved - charge boson coherent states of [ 5 , 6 ] are analogous to ones recently constructed for order @xmath0 parabosons [ 7 ] .
one physical consequence of order @xmath0 " is that two or less such bosons can occupy a totally anti - symmetric state . appendix 1 contains a brief review of @xmath0 paraboson statistics .
both the @xmath0 and @xmath1 models considered in this paper are notably `` statistical '' and crudely `` unrealistic '' in being free of kinematic and dynamical parameters associated with physically important production quantities such as the distributions of available energy / momentum , the kinematic size and other characteristics of the production region , resonances , masses and other conserved quantum numbers ( e.g. isotopic spin in the @xmath1 case [ 8 ] ) , because only the @xmath13-charge conservation constraint has been imposed . consequently , unlike what some readers would expect the situation to be from consideration of other particle - production models , in the figures the relative sizes of different aspects of the peaks , and of other structures , of the surfaces / curves can not be adjusted without additional assumptions and complications .
so despite the authors efforts , an interested viewer must sometimes force oneself to `` look '' in examination of some aspects of the figures . in the @xmath1 case ,
more realistic models / arguments were made [ 8 ] to investigate such issues as independent - emission and coherence in @xmath14 pair production but these analyses are by now , of course , extremely primitive versus the current very sophisticated computational treatments [ 9 , 10 ] of multi - particle production in high - energy collisions .
for instance , even in electron - positron collisions a wide variety of coherence phenomena such as stage - one and stage - two spin - correlations in fermion - antifermion pair production , bose - einstein interference effects , and color - reconnection effects are used in various computational frameworks to describe and investigate particle production at lep . despite the lack of inclusion of other physics that would be required in a realistic application
, we think that the focus in this paper on a comparison of these parameter - free statistical models is interesting and instructive with respect to @xmath0 versus @xmath1 parastatistics , and to the associated coherences involved in the production of conserved - charge paraboson pairs . from the then available data , in [ 4 ]
wang reported a regularity when , for various laboratory kinetic energies of the incoming primary particle , the relative frequency of events @xmath15 was plotted against the mean multiplicty @xmath10 for charged @xmath12 pairs for @xmath16 , @xmath17 , @xmath18 , @xmath19 and @xmath20 collisions .
the data set was a compilation of about 50 inelastic production experiments with incoming primary kinetic energies extending from below production threshold to @xmath21 .
wang tried to fit the regularity with two poisson - type distributions . in [ 5 ] , horn and silver constructed instead a simple parameter - free statistical model and found that , ignoring fluctuations in the data due to the specific structure of the various channels , the model did agree [ 5 , 7 ] with the universal trends of the experimental regularity reported by wang .
one construction of the horn - silver model is / was from the @xmath1 conserved - charge coherent states .
versus the inelastic reactions multiplicity data , in [ 5 ] horn and silver argued for a statistical treatment of the gross features because ( i ) momentum conservation should be a weak constraint since the emitted pions occupy a small part of the available phase space , ( ii ) total isospin conservation on the distribution of charged pions should also be weak since neutral pions are summed over , and so ( iii ) charge conservation remains as the important constraint .
they derived the @xmath15 distribution discussed below .
their @xmath22 distribution , i.e. the percentage of events with
m " @xmath14 pairs and no extra @xmath23 s , had been considered earlier by kastrup [ 11 ] . at the present time , the only way known to construct the analogous model in the @xmath0 case is to use conserved - charge coherent states . an alternative formulation with respect to either the statistical aspects or the coherence aspects is not currently available .
so we now briefly digress to discuss what are these conserved - charge coherent states and then compare them with their @xmath1 counterpart : a charged - paraboson pair in order @xmath0 consists of one a quanta of charge ` + 1 ' and one b quanta of charge ` -1 ' .
the hermitian charge operator is defined by @xmath24 where @xmath25 are the parabose number operators , see appendix 1 .
although @xmath26 does not commute with @xmath27 or @xmath28 , and although the paraboson pair operators @xmath29 , since @xmath30=0,\quad \lbrack q , ba]=0,\quad \lbrack ab , ba]=0\ ] ] the @xmath0 coherent state can be defined as simultaneous eigenstate of @xmath31 and @xmath3 : @xmath32 note that unlike in the @xmath1 case ( see below ) , here two complex numbers @xmath33 and @xmath34 arise because @xmath2 and @xmath3 are fundamentally distinct operators .
consequently , in the following multiplicity considerations , two non - negative parameters occur which are the moduli of these two complex numbers , @xmath35 and @xmath36 which involve the modified bessel functions @xmath37}(u)$ ] and @xmath38}(v)$ ] are given in [ 7 ] for @xmath39 and @xmath40 ; note @xmath41=$ ] `` integer part of @xmath42 '' . for @xmath43 ,
the @xmath44th - moment of the mean multiplicity for a @xmath0 charged - paraboson pair is @xmath45 . for @xmath9
fixed , the percentage of events with @xmath8 such pairs , @xmath15 , is the square of the moduli of the expansion coefficients of @xmath46 in terms of the two - mode parabose number fock states ( see [ 7 ] ) .
the explicit multiplicity formulas corresponding to the figures shown in this paper can be found below in section 2 .
the simpler @xmath1 conserved - charge coherent state @xmath47 satisfies @xmath48 and @xmath49 .
since @xmath2 = @xmath3 for @xmath1 , @xmath50 involves a single complex number @xmath51 , and so the horn - silver multiplicity analysis involved a single , non - negative real parameter latexmath:[$x\equiv because the @xmath0 construction involves two non - negative real parameters @xmath53 and @xmath54 , we consider a 3d plot of @xmath55 versus the quarter plane defined by @xmath56 and @xmath57 .
this is the obvious generalization of wang s plot and it is the focus of section 2 and its figures .
section 3 contains some additional multiplicity formulas which are similar in structure for @xmath58 .
section 4 contains a brief discussion of some of what has / hasnt been learned from this analysis of coherent production of charged - pairs of parabosons of order @xmath0 .
appendix 2 and its figures are focused on how different shaped @xmath59 charts , or coordinate patches , appear when mapped onto the two - dimensional surface .
the analysis in this section is straight - forward because there are simple explicit mathematical formulas involving modified bessel functions and because mathematica " [ 12 ] notebooks are convenient for displaying the parametric 1-parameter @xmath1 curve " and 2-parameter @xmath60 varying - width , folded ribbon . "
we find that a concise curve / ribbon " terminology is useful and appropriate for describing , for example , figs .
1a , b , which are for @xmath61 , and @xmath62 . on a 2d side of the 3d display ,
the @xmath60 ribbon consequently appears projected as a relatively narrow , two - dimensional region which will be denoted as a band " ; see fig .
1b which shows the front xz side of fig .
1a . in spite of their being from the more complex @xmath63 asymmetric case ,
these two lead figures have been chosen in order to _ ab initio _ discuss some details and caveats . if the reader initially finds figs .
1a , b and their description too complex , the reader might go directly to figs . 2 and 3a , b which are for the simpler @xmath64 symmetric case .
1a , b , and in the remaining figures in this paper , the solid line is the @xmath1 curve . in these two figures , this @xmath1 curve is also the @xmath65 line on the @xmath60 ribbon . near the peak ,
the fold " is the bottom edge of the ribbon .
as shown , the @xmath60 ribbon consists of open - circles for the non - folded @xmath66 region , and of solid - circles for the folded @xmath67 region .
the upper edge of the ribbon is the @xmath68 set of points .
slightly to the right of the peak , one can see from the solid - circles that each line of dots traveling leftward down the page , bends under ( or over " , whichever as the viewer prefers ) the fold to reach the @xmath65 line on the ribbon . for @xmath9 odd , we assume it is a complete fold " . the words
complete fold " mean that one part of the @xmath69 ribbon lies exactly on top of another part , and so there is no difference in going under " or over " such a fold .
due to the very limited analytic results , mental / visual care is generally advisable : ( i ) only for the cases of @xmath70 and for @xmath62 is it proven that the @xmath1 curve always lies on the @xmath0 two dimensional surface . in these two cases
the proof appears un - instructive for then the parametric - solution @xmath71 ( and for @xmath70 also @xmath72 ) avoids confronting different ratios of modified bessel functions . for an arbitrary @xmath9 value , from the asymptotic limits for the modified bessel functions
, it follows that as @xmath73 , the curve and as @xmath74 , the ribbon must begin at the same point : the origin for @xmath75 and @xmath76 for @xmath77 .
similarly , for large parameters for arbitrary @xmath9 , the curve and ribbon will also be at the same 3d point with @xmath78 , if @xmath79 .
( ii ) only for @xmath9 even is it known that there is a complete fold of the ribbon because then @xmath6 , @xmath10 , and @xmath11 are each @xmath64 symmetric . for @xmath9
even , the fold is the line @xmath80 , c.f . the simpler figs . 2 and 3a , b .
the line of dots traveling leftward down the page " discussed above for the peak region in figs .
1a , b are a set of @xmath81 values from a unit - negative - slope diagonal in the @xmath81 domain , see last figure of this paper and its discussion fig .
8a , b .
second , the reader should be aware that the dots " displayed to show the two - dimensional surface correspond to specific @xmath82 parameter values and that the associated @xmath83 , @xmath84 values are equally spaced , c.f .
5a below where some @xmath81 coordinate values are shown . in some figures
, @xmath85 is not equal to @xmath86 .
we find that a careful usage of a few un - connected dots " does not mislead in displaying these two - dimensional surfaces .
the dots " should not be confused with a random generation of data points , such as in a scatter plot .
note that the @xmath53 and @xmath54 parameters do not map into an orthogonal coordinate chart on the ribbon .
in general , at each point on the ribbon the curvature is non - zero .
third , in consideration of figs .
1a , b , as well as the other figures in this paper , it should be noted that , even near the peak , the width of the ribbon is narrow versus , for instance , the half - width of the @xmath1 peak .
this fact , besides the folding and the role of the @xmath53 and @xmath54 parameters , would be an important issue in attempting empirically to distinguish coherent production pairs of @xmath0 parabosons versus those of ordinary @xmath1 bosons .
on the other hand , the ribbon width is indeed non - zero and to the eye there is generally a systematic above , or below , displacement of the visual center of the @xmath0 band " from the @xmath1 curve as in figs .
1a , b . for the @xmath61 , and
@xmath62 case shown in figs .
1a , b the formulas for the @xmath1 curve are @xmath87 and for the @xmath0 ribbon are @xmath88 as in ( 4 - 8 ) , a superscript `` @xmath89 or @xmath90 '' , for @xmath1 or @xmath0 , can be respectively written on @xmath10 , @xmath91 , and @xmath15 when needed to avoid confusion .
usually one knows the @xmath92 order from the working context , and so these superscripts can often be suppressed . before discussing other specific @xmath8 and @xmath9 cases , we consider the parametric formulas for an arbitrary curve and ribbon
. the @xmath1 curve @xmath93 is parametrized by the real non - negative parameter @xmath94 . for @xmath9 @xmath95 , in terms of modified bessel functions , @xmath96 , @xmath97 equivalently , @xmath98 follows due to the recursion relation for @xmath99 s .
the probability for `` @xmath8 boson pairs + @xmath9 positive bosons '' is @xmath100 for example , in the analysis of `` ideal '' data for the production of purely multi - pion final states , @xmath15 would be the probability for the production of @xmath101 @xmath102 s and @xmath103 @xmath104 s . for @xmath1 ,
it is instructive to consider the statistically - fundamental @xmath105 with an @xmath8-independent normalization constant @xmath106 because in this way one sees that it is via normalization that the modified bessel function appears in development of the simple idea of charge - conservation - constrained poisson distributions for independent @xmath107 and @xmath108 production [ 5 , 8 ] . for multi - pion final states
wang [ 1 ] , and later horn and silver [ 2 ] in consideration of ( 12 ) as a model , plotted @xmath15 versus @xmath10 for small values of @xmath8 and @xmath9 .
this is the @xmath109 plane , or `` front '' projection of the 3d figures considered in this paper .
the @xmath0 two - dimensional surface @xmath110 is parametrized by the two real non - negative parameters @xmath53 and @xmath54 .
@xmath111as for the @xmath1 curve discussed above , only the @xmath112-coordinate is @xmath8 dependent .
( @xmath9 even ) : the case of @xmath9 even and @xmath9 @xmath95 is symmetric in @xmath53 @xmath113 @xmath54 with @xmath114 , see below , and the @xmath53 @xmath113 @xmath54 symmetric @xmath115 ( @xmath9 odd ) : the case of @xmath9 odd and @xmath9 @xmath95 is asymmetric in @xmath53 @xmath113 @xmath54 with @xmath116 , and the @xmath53 @xmath113 @xmath54 asymmetric @xmath117 for @xmath43 , the probability for `` @xmath8 paraboson pairs + @xmath9 positive parabosons '' is @xmath118}i_{[\frac{q}{2}]}(u)(\frac{v}{2})^{-[\frac{q+1}{2}]}i_{[\frac{q+1}{2}]}(v ) \\
\tilde{p}_{q , m;i } & = & \frac{u^{2r}v^{2s}}{2^{2m}[\frac{m+i}{2}]![\frac{q+m+i}{2}]![\frac{m+1-i}{2}]![\frac{q+m+1-i}{2}]!}\end{aligned}\ ] ] where in @xmath119 @xmath120 \\ s & \equiv & [ \frac{m+(-)^{q+m+i}i}{2}+\frac{1+(-)^{q}}{4}]\end{aligned}\ ] ] in paraboson statistics , it is sometimes convenient to use @xmath121=$ ] `` integer part of @xmath94 '' , and @xmath121!=[x][x-1]\cdots 1 $ ] , @xmath122!=1 $ ] . note that this @xmath121 $ ] symbol occurs in ( 18 - 21 ) .
the use of the symbol @xmath121 $ ] enables a compactification of several analytic expressions and its use is convenient in computational notebooks .
however , it is sometimes instructive , as in ( 13 - 16 ) , to write out the various cases by letting @xmath123 , or @xmath124 ; @xmath125 ; or similarly for @xmath8 or @xmath126 . throughout this paper
, the appearance of `` square braces '' always denotes such an integer truncation .
note that the summation over
i " in ( 17 ) is also due to the fact that @xmath127 [ 7 ] .
the compact expressions are @xmath128}(u)}{i_{[\frac{q}{2}]}(u)}+\frac{vi_{[\frac{q+3}{2}]}(v)}{i_{[\frac{q+1}{2}]}(v)}\right ) \\
< n^{2}>^{(2)}=<n>^{(2)}+\frac{1}{4}\left ( \frac{u^{2}i_{[\frac{q+4}{2}]}(u)}{i_{[\frac{q}{2}]}(u)}+2uv\frac{i_{[\frac{q+2}{2}]}(u)}{i_{[\frac{q}{2}]}(u)}\frac{i_{[\frac{q+3}{2}]}(v)}{i_{[\frac{q+1}{2}]}(v)}+\frac{v^{2}i_{[\frac{q+5}{2}]}(v)}{i_{[\frac{q+1}{2}]}(v)}\right)\end{aligned}\ ] ] since from ( 17 ) the @xmath0 case non - trivially involves two parameters , @xmath53 and @xmath54 , the figures in this paper are 3d ones with @xmath129 plotted versus the two - dimensional @xmath130 @xmath131 and @xmath132 @xmath133 plane .
only the first octant is used . for the case
@xmath77 , @xmath70 the parametric expressions are very simple : @xmath134 @xmath135 this is the simplest @xmath64 symmetric case and the right " @xmath136-side is shown in fig .
as previously mentioned , in the symmetric case , the fold is along @xmath80 which is shown by the dashed line in fig .
the two edges of the original ribbon , @xmath65 and @xmath68 are respectively mapped by @xmath137 , @xmath138 , into the @xmath1 curve .
only open - circle points from the @xmath139 region are shown . all @xmath77 curves / ribbons drop down from @xmath140 .
3a , b , respectively show the 3d plot and the right @xmath136-side for another symmetric case @xmath141 , @xmath70 . note that , unlike in the preceding fig .
2 , in this case the @xmath1 curve now lies on the top of the ribbon until there is an almost @xmath142 twist , after the peak where the @xmath1 curve is on bottom . in fig .
3b the @xmath136-side quite clearly shows this twist .
the parametric expressions are respectively @xmath143 , @xmath144 . in appendix 2
, there is a discussion of the @xmath145 coordinate charts associated with the asymmetric case @xmath141 , @xmath62 shown in figs .
the parametric equations are respectively @xmath146 , @xmath147 .
[ the omitted figures for the asymmetric case @xmath77 , @xmath62 similarly show the fold occurring in the @xmath148 region .
4a is the 3d display and fig .
4b is the projection onto the front @xmath109 plane .
4c is from a different 3d viewpoint .
it provides a view back into the origin , with a @xmath149 floor below and a @xmath149 ceiling above , in a left - to - right , front axes labeling " .
this figure shows how nearly vertical the ribbon is .
lastly , fig .
4d shows a close - up of the @xmath141 , @xmath62 peak .
this figure illustrates the fold . in it
the upper solid line is the @xmath1 curve and the lower solid line is the @xmath80 curve which separates the open - circles ( @xmath139 ) from the solid - circles ( @xmath148 ) .
the different shaped @xmath59 charts of appendix 2 can be used in investigating the folding which occurs in a transverse crossing of the ribbon .
similarly , the lines @xmath150 , also @xmath151 for @xmath9-odd , @xmath80 , and for @xmath152 also the independent @xmath1 curve , can be used .
both techniques are useful in studying what occurs in going out along the ribbon . in proceeding from the origin , or for @xmath77 from @xmath153 ,
there are twist(s ) and a peak ( absent for @xmath77 ) . if as in fig .
4c , the ribbon is viewed back in from the high @xmath53 and @xmath54 parametric coordinates , then in moving out from the origin but viewing back towards the origin , the almost @xmath142 twisting of the @xmath154 edge from the peak onward is counter - clockwise for @xmath155 . for @xmath156
there is also a twist near the origin . for the peak in @xmath157 , c.f .
4a , the @xmath112 coordinate values are : for the @xmath151 top edge , @xmath158 ; for the @xmath1 curve , @xmath159 ; for the @xmath80 curve @xmath160 ; and for the fold @xmath161 .
in the @xmath0 case , the @xmath44-th moment of the mean multiplicity @xmath162 where @xmath163 .
it follows that @xmath164 where , with @xmath165 a binomial coefficient , @xmath166}(u)i_{[\frac{q+1 + 2l}{2}]}(v ) \\ \bar{c } & = & 2^{[\frac{q}{2}]+[\frac{q+1}{2}]}u^{-[\frac{q}{2}]}v^{-[\frac{q+1}{2}]}\end{aligned}\ ] ] the function @xmath167 is generated by @xmath168 .
the expression for @xmath169 is ( 18 ) above .
note that as in the @xmath1 case , @xmath170 when @xmath70 , @xmath171 , and @xmath111for fixed @xmath9 , @xmath172 . for ( @xmath173 ,
there are the following relations to the @xmath174 expressions : @xmath175 in ( 27 - 32 ) and exchange @xmath176 , @xmath177 in ( 27 , 35 ) and exchange @xmath178 ; and @xmath179 .
the analogous formulas for the @xmath1 case are @xmath180 where @xmath181 .
so @xmath182 , @xmath183 where @xmath184 the function @xmath185 is generated by @xmath186 . the expression for @xmath187
is given after ( 12 ) above .
for ( @xmath173 for @xmath1 , @xmath188 , @xmath189 , and @xmath179 .
the 2d surface occurs in the 3d plots of the relative probabilities @xmath190 versus @xmath10 , @xmath91 as required by the coherences embodied in the equations in ( 3 ) for the @xmath0 conserved - charge coherent states . unlike the analogous @xmath1 case , where such a model was motivated from a regularity in @xmath12 data , it is necessary in the case of the production of the charged - pairs of parabosons of order @xmath0 to assume an analogous coherent production mechanism and to assume a reasonable , but definite , treatment of the @xmath2 and @xmath3 operators .
neither of these two assumptions might be true if paraboson pairs are found to be produced in nature in high energy physics collisions , e.g. in central - diffactive - exchange experiments [ 14 ] , or in some area of contemporary astrophysics / cosmology , e.g. in dark matter detectors [ 15 ] or by very high energy cosmic rays [ 16 ] . in the infra - red domains of qed and of qcd ,
coherent - state / degenerate - state coherence is a well - known phenomena .
on the other hand in qcd , partonic jets dominate the very high energy hadronic production processes . in the case of the lower energy @xmath14 production , as was emphasized by horn - silver , and others , the situation is complex and there is much more physics and phenomenological structure than that incorporated in the reference @xmath1 model considered in this paper . in the approximation in
which one neglects such additional physics , this @xmath1 model is parameter free because the @xmath94 parameter is effectively replaced by @xmath10 , the mean number of final charged pairs .
similarly , the analogous @xmath0 model is parameter free for the @xmath53 and @xmath54 parameters are effectively replaced by @xmath10 and @xmath11 . at the present time , the existence and relative importance of other conserved quantum numbers , of resonances or other phenomenological interaction effects , etc .
is completely unknown for a coherent production of pairs of parabosons of order @xmath191 nevertheless , with respect to experiments [ 14 - 16 ] , one conclusion from this paper is that from a parameter - free statistical model , one would expect a signature of relatively narrow bands , due to projection of varying - width folded ribbons , will be present in 2d analysis of pair multiplicities from coherent production of parabosons of order @xmath192 . in the case of @xmath0 parabosons , there will be a second kinematic variable [ 4 ] and multiplicity data will not scale in terms of a single - variable curve [ 11 ] .
for instance , if the variable for the @xmath1 curve is @xmath193 , then for @xmath0 parabosons there will also be a dependence on @xmath194 . in fig .
8a , with @xmath193 fixed , the sensitivity to the second variable @xmath194 is very striking and significantly greater than the naive width of the ribbon . in the region of the peak ,
1b and 4b show that this signature is considerably enhanced when there is an extra positive paraboson , versus the case of only charged - paraboson - pairs , fig . 3b . while an explicit dynamical production model is required to theoretically investigate the sensitivity with respect to specific kinematic variables , in independent - particle - emission models the total energy in the emitted particles is monotonically related to the intensity strength of the source .
this suggests that two simple and useful kinematic variables are the sum , @xmath195 , and difference @xmath196 , of the total emitted @xmath197 and @xmath198 energies .
the `` identical particle '' defining parabose tri - linear relations leave the two pair operators , @xmath2 and @xmath3 , fundamentally distinct which ( 3 ) and present analysis maintain by the use of the two distinct @xmath53 and @xmath54 parameters . in the @xmath1 case ,
the @xmath199 parameter can be interpreted as a product of the intensity strengths of the @xmath102 and @xmath104 sources . in the @xmath0 case ,
@xmath53 and @xmath54 can be interpreted as the intensity strengths of the @xmath2 " and @xmath3 " sources .
their difference is due to the two distinct orderings of the @xmath27 and @xmath200 operators .
the physical observables are not always @xmath53 @xmath113 @xmath54 symmetric . in particular
, the presence of a @xmath9 odd total charge produces an asymmetry in the folding of the 2d surface ; for instance , see eqs .
( 15,16 ) and ( 8) .
there is overall @xmath201 symmetry ( @xmath13 charge symmetry ): the probability for @xmath202 @xmath203 s and @xmath204 @xmath205 s " equals probability for @xmath206 @xmath203 s and latexmath:[$(m + the @xmath13 charge might be a `` hidden conserved - charge '' such as to yield pure pair production , i.e. only final state events with @xmath70 would occur such as in the production of strange particles via the strong interactions . in this case , or by a designed selection of only pair final state events , the @xmath0 versus @xmath1 results of this paper can still be used with @xmath210 .
this is the simpler symmetric case and because of @xmath70 there would be more analytic control in such an analysis .
one of us ( can ) thanks experimental and theoretical physicists for discussions .
we thank ted brewster , lorene evans , anthony poole , and james l. wolf for computer support .
this work was partially supported by u.s .
dept . of energy contract
. de - fg 02 - 86er40291 . in local , relativistic quantum field theory ,
identical particles obey either ( i ) parabose and parafermi statistics [ 1 - 3 ] for which the number of particles in an antisymmetric or a symmetric state , respectively , can not exceed a given integer @xmath92 , or ( ii ) for two space dimensions , infinite statistics based on the braid group [ 13 ] . in parabose statistics , instead of bilinear , there are fundamental trilinear commutation relations : @xmath211=2\delta _ { kl}a_m , \ ; [ a_k,\{a_l^{\dagger } , a_m^{\dagger } \}]=2\delta _ { kl}a_m^{\dagger } + 2\delta _ { km}a_l^{\dagger } , \\ \lbrack a_k,\{a_l , a_m\}]=0 , \hspace{2pc } ( k , l , m=1,2 ) \end{array}\ ] ] where @xmath212\equiv cd - dc$ ] , and @xmath213 . in the case of only two kinds of parabosons ,
there are some simple commutation relations between `` a '' paraboson operators and the `` b '' paraboson operators : letting @xmath214 @xmath215 , @xmath216=0 , \ ; [ b , a^2]=0 , \ ; [ a^{\dagger } , b^2]=0 , \ ; [ b^{\dagger } , a^2]=0\ ] ] plus the hermitian conjugate relations .
order @xmath0 is simpler because there is the self - contained set " of 3 relations [ 2 ] @xmath217 the parabose number operators for @xmath0 order are defined by @xmath218 in [ 7 ] , the state - vector space for two - mode parabosons and the order @xmath0 conserved - charged parabose coherent states were constructed . in a determination of the statistical uncertainties for an application of the parametric formulas in section 2 in comparison with actual experimental data , the structure(s ) of the folded , partially - twisted ribbons versus @xmath1 curves will enter and have to be correctly treated .
for this reason , and for possible more abstract theoretical / mathematical use by other readers , the following details about how different shaped @xmath145 coordinate patches appear when mapped onto the 2d surface should be of interest .
the lines @xmath150 , also @xmath151 for @xmath9-odd , and @xmath80 provide useful standard references in examining what is taking place on each ribbon .
in addition , for @xmath152 the @xmath1 curve is a different reference since it no longer is the @xmath150 line .
the @xmath141 , @xmath62 case which is @xmath64 asymmetric is used in this appendix .
this case is both generic and visually simple . figs .
4a-4d were earlier displays of this case .
the discussion begins with the non - folded @xmath154 region : figs .
5a , b show how the @xmath219 strip " from @xmath220 to @xmath221 is mapped onto the peak region of the ribbon .
in 3d fig .
5a , the solid line that lies on the bottom of the patch " on the ribbon is the @xmath80 set of points .
the open - circle points on the top of the patch are the @xmath151 set of points .
the points on the origin side in fig .
5a are for the vertical @xmath220 line in fig .
5b . the points on the patch farthest from the origin , which appear to have a nearly constant negative slope , are for the other vertical line in fig . 5b , i.e. for @xmath221 . here
there have been 5 sets of vertical points mapped from fig .
5b onto the 5a patch on the ribbon . figs .
6a , b similarly show the differences when the @xmath219 points arise from three more widely separated constant-@xmath53 vertical lines with respectively @xmath222 . to the left - side of the peak , @xmath223 ,
the spacing between these constant-@xmath53-value points is rather small , whereas on the right of the peak , where @xmath224 , the spacing between such points is significantly greater .
the points are equally spaced in @xmath54 so there are more points taken at @xmath224 than at @xmath223 .
horizontal strips " can similarly be used to study the folded @xmath225 region .
however , a ring " region or a unit - negative - slope strip " region can be used which includes both @xmath139 and @xmath148 domains .
7a , b show what occurs when points in a ring " of radii @xmath226 and @xmath227 are mapped onto the ribbon .
in 7a , the open - circle - dots are again from the simpler @xmath139 region , whereas the solid - dots are from the @xmath148 region where the fold occurs .
7c shows a close - up where the @xmath150 points ( @xmath1 curve ) are on the upper solid curve ; the @xmath80 points which divide the ring in half are on the lower solid curve .
[ if instead , the points are from pie - slice - like sectors , the associated lines on the ribbon then extend indefinitely on out the ribbon . ]
lastly , fig .
8a , b show what happens when points are taken from a somewhat similar , but sometimes more useful , unit - negative - slope strip in the @xmath228 quadrant . in fig . 8a
the solid line is the @xmath80 curve .
the line of dots traveling leftward down the page " discussed in the text for the peak region in figs .
1a , b are a set of @xmath81 values from such a unit - negative - slope diagonal . if instead , the points are taken from the orthogonal unit - positive - slope strips , the associated lines on the ribbon will start on a @xmath229 or @xmath230 line , depending on which side of @xmath231 the strip is from , and then the line will extend indefinitely on out the ribbon . h.s .
green , phys.rev . *
90 * , 270 ( 1953 ) ; d.v .
volkov , sov .
jetp , * 11 * , 375 ( 1960 ) ; and c.f .
w. pauli , handbuch der physik vol .
* 24 * , p. 113 of sec . 14
( 1934 ) ; p. 128 - 9 in
_ a scientific biography of w. pauli _ , c.p .
enz ( 2002)(oxford u. press ) .
o.w . greenberg and a.m.l .
messiah , phys.rev . * 136 * , b248 ( 1964 ) ; _ ibid .
_ , * 138 * , b1155 ( 1965 ) . y. ohnuki and s. kamefuchi , _ quantum field theory and parastatistics _
( 1982 ) ( univ . press of tokyo ) .
wang , phys.rev . * 180 * , 1463 ( 1969 ) .
wang plotted @xmath232 for @xmath70 versus the equivalent , mean number of charged prongs @xmath233 @xmath234 @xmath235 ( horn - silver ) .
figure 1a in wang s paper shows the laboratory kinetic energies of the data points .
d. horn and r. silver , phys.rev .
* d2 * , 2082 ( 1970 ) .
d. bhaumik , k. bhaumik , and b. dutta - roy , j. phys .
a : * 9 * , 1507 ( 1976 ) ; j.r . klauder and b
skagerstam , p. 43 - 48 , _ coherent states_(1985 ) ( world sci . pub .
co. , singapore ) ; and see g.s .
agarwal , j. opt .
. am . * b5 * , 1940 ( 1988 ) .
s. jing and c.a .
nelson , j. phys .
a : * 32 * , 4131 ( 1999 ) . see , and references therein , d. horn and r. silver , annals of phys .
( ny ) * 66 * , 509 ( 1971 ) ; d. horn , physics rep . * 4 * , 1 ( 1972 ) ; and d. horn and f. zachariasen , _ hadron physics at very high energies _ ( 1973 ) ( w.a .
benjamin , reading , ma ) .
t. sjostrand and m. bengtssson , comp .
. comm . * 43 * , 367 ( 1987 ) ; g. marchesini and b.r .
webber , _ ibid .
_ * 67 * , 465 ( 1992 ) ; s. jadach , z. was , r. decker , and j.h .
kuhn , _ ibid . _ * 76 * , 361 ( 1993 ) ; s. jadach , b.f.l .
ward , and z. was , _ ibid . _ * 124 * , 233 ( 2000 ) ; b.f.l . ward and s. jadach , acta phys .
b33 * , 1543 ( 2002 ) ; l. lonnblad , _ comp .
_ * 71 * , 15 ( 1992 ) ; and t. sjostrand , p. eden , c. friberg , l. lonnblad , g. miu , s. mrenna and e. norrbin , _ ibid . _ * 135 * , 238 ( 2001 ) .
see , and references therein , _ perturbative quantum chromodynamics _
mueller ( 1989 ) ( world sci . pub .
co. , singapore ) ; yu l. dokshitzer , v.a .
khoze , a.h .
mueller and s.i .
troyan , _ basics of perturbative qcd _ , ed .
j. tran thanh van ( 1991 ) ( gif - sur - yvette , editions frontieres ) ; j.c .
collins and d.e .
soper , ann .
nucl . part .
37 * , 383 ( 1987 ) ; r.k .
ellis , w.j .
stirling and b.r .
webber , _ qcd and collider physics _ ( 1996 ) ( cambridge uni . press ) ; b. andersson , g. gustafson , g. ingelman and t. sjostrand , phys .
rep . * 97 * , 31 ( 1983 ) ; b. andersson , _ the lund model _ ( 1997 ) ( cambridge u. press ) ; l.l .
frankfurt and m.i .
strikman , phys . rep . * 160 * , 235 ( 1988 ) ; l.n .
lipatov , _ ibid . _ * 286 * , 131 ( 1997 ) ; i. dremin and j.w .
gary , _ ibid . _
* 349 * , 301 ( 2001 ) ; s. catani , f. krauss , r. kuhn , and b.r .
webber , jhep * 0111 * , 63 ( 2001 ) ; c.f .
berger , t. kucs and g. sterman , phys . rev . *
d68 * , 014012 ( 2003 ) ; a. hebecker , phys . rep . * 331 * , 1 ( 2000 ) ; and e.a . de wolf , j. phys .
g : * 28 * , 1023 ( 2002 ) .
kastrup , nuc . phys .
* b1 * , 309 ( 1967 ) ; h. gemmel and h.a .
kastrup , z. phys .
* 229 * , 321(1969 ) ; nuc .
* b14 * , 566(1969 ) .
, version 4.2 , wolfram research , inc.(champaign , il ) . , ed .
f. wilczek ( 1990 ) ( world sci . pub .
co. , singapore ) ; and a. khare , _ fractional statistics and quantum theory _ ( 1997 ) ( world sci . pub .
co. , singapore ) .
albrow and a. rostovtsev , hep - ph/0009336 ; v. khoze , a.d .
martin , and m. ryskin , eur .
j. * c14 * , 525(2000 ) ; hep - ph/0207313 ; a. de roeck , v. khoze , a.d .
martin , et.al , eur .
j. * c25 * , 391(2002 ) ; cdf collab . ,
d. acosta , et.al .
, hep - ex/0311032 ; totem collab . , http://totem.web.cern.ch/totem .
dama collab . ,
r. bernabei , et.al .
* 26 * , 1 ( 2003 ) ; cdms collab . ,
r. abusaidi , et.al .
84 * , 5699(2000 ) ; edelweiss collab .
, a. benoit , et.al .
, phys . lett . *
b513 * , 15(2001 ) ; igex collab . ,
h. morales , et.al .
_ , * b489 * , 268(2000 ) ; hdms collab . , h.v .
klapdor - kleingrothaus , et.al .
. supp . * 110 * , 364(2002 ) ; zeplin collab . ,
j. dawson , et.al .
, * 110 * , 109(2002 ) ; simple collab . ,
collar , et.al .
, new j. phys .
* 2 * , 14(2000 ) ; picasso collab . ,
n. boukhira , et.al .
. supp . * 110 * , 103(2002 ) .
agasa collab .
, m. takeda , et.al .
, astropar . phys .
* 19 * , 447 ( 2003 ) ; hi res collab .
berman , et.al , nuc .
. supp . * 117 * , 106 ( 2003 ) ; auger collab . ,
http://www.auger.org/ ; and c.f .
v.a . karmanov and a.e .
kurdyavtsev , hep - ph/0207321 .
1a : the probability @xmath236 of 2 charge - paraboson - pairs plus 1 positive paraboson " versus @xmath10 , @xmath11 .
this is the more complex @xmath63 asymmetric case .
the solid line is the @xmath1 curve .
near the peak , the fold is the bottom edge of the ribbon .
the open - circles are for the non - folded @xmath139 region , and the solid - circles for the folded @xmath148 region , see discussion in sec .
2 . fig .
2 : the right - side projection , or @xmath136-side , for the probability @xmath237 of zero charge - paraboson - pairs plus zero additional single charged parabosons " versus @xmath11 .
since @xmath9 is even , this is the simpler @xmath63 symmetric case .
only open - circles from the @xmath139 region are shown .
the fold is the dashed line @xmath80 .
3a : the probability @xmath238 of 1 charge - paraboson - pair plus zero additional charged parabosons versus @xmath10 , @xmath11 .
this is also the simpler @xmath239 symmetric case so only open - circles from the @xmath240 region need to be shown .
the solid line is the @xmath1 curve .
the fold is the dashed line @xmath80 .
8a : for @xmath241 , the image of the points in the unit - negative - slope domain in fig . 8b .
the @xmath81 values shown are a subset from the unit - negative - slope line of points nearest the origin in fig .
note the location of the end points @xmath245 and @xmath246 .
|
a parameter - free statistical model is used to study multiplicity signatures for coherent production of charged - pairs of parabosons of order @xmath0 in comparison with those arising in the case of ordinary bosons , @xmath1 .
two non - negative real parameters arise because @xmath2 " and @xmath3 " are fundamentally distinct pair operators of charge ` @xmath4 ' , a - quanta and charge ` @xmath5 ' , b - quanta parabosons . in 3d plots of @xmath6@xmath7 the probability of @xmath8 paraboson charged - pairs + @xmath9 positive parabosons " versus @xmath10 and @xmath11 , the @xmath1 curve is found to lie on the relatively narrow 2d @xmath0 surface
. = -5 mm * coherent production of pairs of parabosons of order 2 * + nicholas frascino and charles a. nelson + _ department of physics , state university of new york at binghamton + binghamton , n.y .
13902 _
|
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.