article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
multiwavelength follow - up of gamma - ray burst ( grb ) afterglows has revolutionized grb astronomy in recent years , yielding a wealth of information about the nature and origin of grbs ( cf .
klose 2000 ; kulkarni et al .
2000 ; galama 2000 ; castro - tirado et al .
1999 and references therein ) .
one of the most important clues to the origin of grbs is the total amount of energy released in the event .
optical observations have helped to derive this important quantity in a number of cases by setting the distance scale through measurement of redshift , and by determining the degree of collimation of the initial emission through detailed study of the afterglow light curve .
clear evidence of the steepening of the light curve , signature of an originally collimated outflow expanding laterally , has been found within a few days of the grb event in several afterglows ( castro - tirado et al 1999 ; kulkarni et al 1999 ; stanek et al .
1999 ; halpern et al . 2000c ; sagar et al .
2000b ) as a part of an international collaborative programme coordinated by one of us ( ajct ) , the uttar pradesh state observatory ( upso ) , naini tal has been involved in a regular programme of grb follow - up since january 1999 .
so far successful photometric observations of five afterglows have been carried out at upso .
the results of four of these have been reported earlier ( sagar et al .
1999 , 2000a , 2000b ) . in this paper
we report the studies made on the fifth , namely , grb 000926 .
grb 000926 was detected on 2000 september 26 at 23:49:33 ut by the inter - planetary network ( ipn ) group of spacecrafts ulysses , russian gamma - ray burst experiment ( konus ) and near earth asteroid rendezvous ( near ) ( hurley et al . 2000 ) .
the burst had a fluence of @xmath16 erg @xmath17 .
the total duration of the grb is @xmath18 s , putting it in the class of `` long - duration '' bursts , to which all grbs with detected afterglows so far belong ( cf .
lamb 2000 ) .
optical observations taken on 2000 september 27 by gorosabel et al .
( 2000 ) and dall et al .
( 2000 ) detected independently a new bright point like source at @xmath19 with a strong candidacy for the afterglow from grb 000926 .
at the same location , frail & berger ( 2000 ) detected a new radio source with vla at 8.46 ghz and 4.86 ghz .
coincident ( within errors ) with the location of optical and radio afterglows , piro & antonelli ( 2000 ) and garmire et al .
( 2000 ) detected the afterglow of grb 000926 .
the near - ir @xmath20 detection of the afterglow has been reported by di paola et al .
( 2000 ) , while observations in @xmath21 passband by kobayashi et al .
( 2000 ) clearly detect the optical transient ( ot ) . the high resolution optical spectrum of the ot of grb 000926 taken on 2000 september 29.26 ut ( cf .
castro et al .
2000 ) indicate a redshift value of @xmath22 .
this is a refinement of @xmath23 reported earlier by fynbo et al .
( 2000c ) using the low resolution optical spectrum taken on 2000 september 27.91 ut .
fynbo et al .
( 2000d , e ) report the detection of the host galaxy with an r magnitude of @xmath24 . in order to provide accurate photometric magnitudes of the ot , we also imaged the field of grb 000926 along with sa 110 region of landolt ( 1992 ) .
a total of 20 stars in the field have been calibrated and their standard @xmath3 magnitudes are given here .
we present the details of our optical observations in the next section , and discuss the light curves and other results in the remaining sections .
recently price et al ( 2000b ) have presented an independent series of observations in multiple optical / ir bands .
the results obtained by them are consistent with those presented here .
the optical observations of the grb 000926 afterglow were carried out on 2000 september 29 .
we used a 2048 @xmath25 2048 pixel@xmath14 ccd system attached at the f/13 cassegrain focus of the 104-cm sampurnanand telescope of upso , nainital .
one pixel of the ccd chip corresponds to 0.@xmath2638 , and the entire chip covers a field of @xmath27 on the sky .
the ccd @xmath3 observations of the grb 000926 field along with landolt ( 1992 ) sa 110 region have also been carried out on 27th october 2000 for calibration purposes during photometric sky conditions .
the log of ccd observations is given in table 1 .
in addition to these observations , several twilight flat field and bias frames were also observed .
* table 1 .
* log of ccd observations of grb 000926 and sa 110 landoldt ( 1992 ) standard fields . [
cols="^,<,^,<",options="header " , ]
we have used the data published in gcn circulars in combination with the present measurements to study the optical flux decay of grb 000926 afterglow . fig .
2 shows a plot of the photometric measurements as a function of time .
the x - axis is log ( @xmath28 ) where @xmath29 is the time of observation and @xmath30 (= 2000
september 26.9927 ut ) is the time of grb trigger .
all times are measured in unit of day .
[ light ] the emission from grb 000926 ot is fading in both @xmath4 and @xmath2 passbands .
the flux decay of most of the earlier grb afterglows , barring a few , is generally well characterized by a single power law @xmath31 , where @xmath32 is the flux of the afterglow at time @xmath29 and @xmath33 is the decay constant .
in contrast , the optical light curves of grb 000926 ( fig .
2 ) can not be fitted by a single power - law .
overall the ot flux decay seems to be described by a broken power - law as expected in grb afterglows having jet - like relativistic ejecta ( sari et al .
1999 ; rhoads 1999 ) .
broken power - law in these light curves can be empirically fitted by functions of the form ( see sagar et al .
2000b for details ) @xmath34^{1/s } + f_g$ ] , where @xmath35 is the constant flux from the underlying host galaxy , @xmath36 and @xmath37 are asymptotic power - law slopes at early and late times with @xmath38 .
the parameter @xmath39 controls the sharpness of the break , a larger @xmath40 implying a sharper break . with @xmath41 ,
this function becomes the same as that used by stanek et al .
( 1999 ) to fit the optical light curve of grb 990510 afterglow .
@xmath42 is the flux of afterglow at the cross - over time @xmath43 .
the function describes a light curve falling as @xmath44 at @xmath45 and @xmath46 at @xmath47 . in jet models , an achromatic break in the light curve
is expected when the jet makes the transition to sideways expansion after the relativistic lorentz factor drops below the inverse of the opening angle of the initial beam . slightly later
, the jet begins a lateral expansion which causes a further steepening of the light curve .
we use the observations in @xmath4 and @xmath2 to determine the parameters of the jet model using the above function .
we find that the minimum value of @xmath48 is achieved for @xmath49 .
this indicates that the observed break in the light curve is sharp , unlike the smooth break observed in the optical light curve of grb 990510 ( cf .
stanek et al .
1999 ; harrison et al . 1999 ) but similar to the sharp break observed in the optical and near - ir light curves of grb 000301c ( cf .
sagar et al .
in order to avoid a fairly wide range of model parameters for a comparable @xmath48 due to degeneracy between @xmath50 and @xmath51 ( magnitudes corresponding to @xmath42 and @xmath35 respectively ) , @xmath36 , @xmath37 and @xmath40 , we have used a fixed value of @xmath52 in our further analyses .
the least square fit values of the parameters @xmath43 , @xmath53 , @xmath36 , and @xmath37 then work out to be [email protected] d , [email protected] , [email protected] and [email protected] respectively in @xmath2 band , with a corresponding @xmath48 of 0.93 per degree of freedom ( @xmath54 ) .
the fit yields the host galaxy contribution to be @[email protected] in @xmath2 .
this is slightly fainter than the value @xmath12 23.9 estimated by fynbo et al .
( 2000e ) from deep ccd imaging , but brighter than @[email protected] derived from hst measurements ( price et al 2000b ) .
the main reason for the discrepancy with the hst estimate is possibly the restriction of the hst photometry to within a 1.5 arcsec aperture , while no such cutoff has been applied to the photometric data used here .
the derived values of the above parameters do not change significantly with larger adopted values of @xmath40 , although a marginal decrease in @xmath55 is noticed in doing so . in @xmath4 band
the available data is sparse , and a fit with fixed @xmath56 and @xmath57 d yields @[email protected] , @[email protected] and @[email protected] .
note the poor constraint on @xmath37 and b - magnitude of the host galaxy , because of the scarcity of measurements in the late light curve .
one can improve the constraints on the decay indices somewhat by assuming a colour index for the host galaxy , and thereby fixing its b magnitude from the r magnitude determined above .
the restricted aperture hst photometry reported by price et al ( 2000b ) indicate a @xmath59 colour index of [email protected] for the host galaxy .
fixing the value of @xmath58 at 25.5 using this mean colour index , the fit yields @[email protected] and @[email protected] . on the other hand , if the host galaxy has a flat spectral distribution similar to that for grb 990123 ( castro - tirado et al .
1999 ) , then @xmath59 colour will be @xmath60 ( which is allowed within the errors of the hst measurements and our light curve fits ) , and @xmath58 will be 24.9 .
if we use this value for @xmath58 , then the best fit values of the decay constants become @xmath36 = [email protected] and @xmath37 = [email protected] .
we can therefore say with confidence that within the range of host galaxy colours allowed by the light curves and the hst measurements , the decay indices in @xmath4 and @xmath2 bands are consistent with each other , and that the break in the light curve displays the achromaticity expected in the case of a lateral expansion of collimated ejecta .
the best fit light curves for the @xmath2 and the @xmath4 bands , with an assumed host galaxy colour index of @xmath59=1.04 , are displayed in fig . 2 .
the decay parameters derived here are consistent with those obtained by price et al ( 2000b ) from their measurements .
we have constructed the grb 000926 afterglow spectrum on 2000 september 29.25 ut and 29.76 ut using the available optical , near - ir , radio and observations .
these epochs were selected for the long wavelength coverage possible at the time of observations of flux by piro & antonelli ( 2000 ) and garmire et al .
optical fluxes at the wavelengths of @xmath61 and @xmath62 passbands have been derived using the slope of the fitted light curve shown in fig . 2 for the epochs under consideration .
the fluxes for the desired epoch at 8.46 ghz and @xmath20 wavelength are taken from frail & berger ( 2000 ) and di paola et al .
( 2000 ) respectively assuming that the flux decay is similar to optical one .
we used the reddening map provided by schlegel , finkbeiner & davis ( 1998 ) for estimating galactic interstellar extinction towards the burst and found a small value of @xmath63 mag .
we used the standard galactic extinction reddening curve given by mathis ( 1990 ) in converting apparent magnitudes into fluxes and used the effective wavelengths and normalisations by fukugita et al .
( 1995 ) for @xmath61 and @xmath62 and by bessell & brett ( 1988 ) for @xmath20 .
the fluxes thus derived are accurate to @xmath12 10% in optical and @xmath12 30% in @xmath20 .
3 shows the spectrum of grb 000926 afterglow from to radio region .
it is observed that as the frequency decreases the flux increases from to radio wavelengths . for a chosen frequency interval
, we describe the spectrum by a single power law : @xmath64 , where @xmath65 is the flux at frequency @xmath66 and @xmath67 is the spectral index . in the to optical region at @xmath68 = 2.26 day ,
the value of @xmath67 is @xmath690.02 , while in the to @xmath20 region , the value is @xmath700.04 . at @xmath71 = 2.77 day the value of @xmath67 is @xmath720.04 for to @xmath20 region .
these values agree very well with the spectral slopes derived for the spectrum of the grb 000926 afterglow by piro & antonelli ( 2000 ) and garmire et al .
( 2000 ) . during @xmath68
= 1 to 4 days , the values of @xmath73 and @xmath74 are almost the same indicating no change in the optical spectral index before and after the time @xmath43
. however , the values of @xmath67 in optical ( @xmath75 and @xmath76 derived at @xmath77 and 3.9 day respectively ) and in optical to @xmath20 region ( @xmath78 at @xmath79 day ) are much steeper than those derived for the broadband ( x - ray to optical ) spectrum
. this may indicate presence of intrinsic extinction in the grb 000926 afterglow .
the amount of intrinsic extinction is determined in the following way assuming that the ot spectrum follows a single spectral slope from to near - ir . the observed flux at and @xmath20 wavelengths are used to determine @xmath67 as the extinction affects them least .
this slope predicts the fluxes of grb 000926 ot at @xmath3 wavelengths .
the differences between these predicted and the corresponding galactic extinction corrected observed fluxes are used to evaluate the intrinsic extinction . from these values , intrinsic extinction at @xmath20 is determined assuming that it follows the galactic extinction law given by mathis ( 1990 ) . the extinction corrected flux at @xmath20 and the observed flux
are used to derive modified value of @xmath67 .
this procedure is iterated till self consistent values of intrinsic extinction at @xmath3 and @xmath20 wavelengths are obtained . this yields @xmath80 mag and @xmath81 as the values for the intrinsic extinction
these are in agreement with the values determined independently by price et al ( 2000b ) from their optical data .
the observed fluxes corrected for both galactic and intrinsic extinction are also shown in fig .
the values of @xmath67 derived from them are @xmath82 and @xmath83 at @xmath84 2.26 and 2.77 days respectively . in the light of above , we conclude that the value of @xmath67 is @xmath85 in to near - ir spectral region and it has not changed before and after @xmath43 during @xmath86 1 to 4 days .
3 indicates that the peak frequency appears to lie in the millimeter region .
this peak frequency is thus similar to that of grb 970508 ( cf .
galama et al .
1998 ) and grb 000301c ( cf .
sagar et al .
2000b ) but different from that of grb 990123 ( galama et al . 1999 ) where the peak is in radio region and that of grb 971214 for which the peak is in optical / near - ir waveband ( ramaprakash et al .
1998 ) . from this
, one may infer that the synchrotron peak frequency may span a large range in grb afterglows .
the afterglow of a grb is thought to be synchrotron radiation from a relativistic shock driven into the circumburst environment . in this model , the light curves and spectral energy distributions are generally well fit by the power - law @xmath87 for a range of frequencies and times that contain no spectral breaks .
the decay slope is dependent on the dynamics of the fireball , and the spectral slope on the importance of cooling , but within each regime @xmath33 and @xmath67 are functions only of @xmath88 , the power - law exponent of the electron lorentz factor .
this provides a verifiable relation between the two observable quantities @xmath33 and @xmath67 , and a way to infer the fireball dynamics .
there is recent evidence that the fireball in at least some grbs are not spherical but are collimated , like jets , into a small solid angle . in such cases , theoretical models predict a break and a marked steepening in the afterglow light curve ( mszros & rees 1999 ; rhoads 1999 ; sari et al .
1999 , huang 2000 ) .
the time of occurrence of this break in the afterglow light curve depends upon the opening angle of the collimated outflow .
observational evidence for a break was found first in the optical light curve of grb 990123 afterglow ( castro - tirado et al .
1999 ; kulkarni et al . 1999 ) and recently in that of grb 990510 ( stanek et al .
1999 ) , grb 990705 ( masetti et al .
2000 ) , grb 991216 ( halpern et al . 2000c ) and grb 000301c ( sagar et al .
2000b ) afterglows . at late times , when the evolution is dominated by the lateral spreading of the jet , the value of @xmath33 is expected to approach the electron energy distribution index while the value of @xmath67 is expected to be @xmath89 if the cooling frequency is below the observing frequency and @xmath90 otherwise ( sari et al .
the expected values of @xmath33 and @xmath67 at late times are thus in agreement with the corresponding observed values of @xmath91 and @xmath85 for grb 000926 indicating that the afterglow emission from grb 000926 is of jet type and not spherical , and that the cooling frequency is above the range of observed frequencies .
the slope of the light curve before the break is then expected to be @xmath92 ( sari et al .
1998 ) , which is also consistent with the value @xmath93 derived by us . however , a much steeper late decay of flux with @xmath941 and @xmath12 5 , while the spectral index is the same ( @xmath95 ) , as reported by piro & antonelli ( 2000 ) and garmire et al .
( 2000 ) respectively is not consistent with the simple jet model which predicts similar late flux decay in both optical and regions .
it is possible that the geometry of the region emitting x - rays after the lateral spreading starts differs from that emitting at longer wavelengths , thus giving rise to apparent difference in the behaviour of the light curve .
another area of possible disagreement with the standard fireball model is the sharpness of the break in the light curve . while the expected sharpness of the transition could depend on the density profile of the ambient medium ( kumar and panaitescu 2000 )
, the very sharp transitions seen in grb 000301c and grb 000926 would be difficult to quantitatively explain in the standard fireball model .
this is an area that deserves a detailed theoretical study .
the anisotropy of the initial ejection needs to be incorporated into the derived energetics of the burst .
the redshift @xmath96 determined for the grb 000926 afterglow by castro et al .
( 2000 ) yields a minimum luminosity distance of 16.6 gpc for standard friedmann cosmology with hubble constant @xmath97 = 65 km / s / mpc , cosmological density parameter @xmath98 = 0.2 and cosmological constant @xmath99 = 0 ( if @xmath100 then the inferred distance would increase ) .
the grb 000926 thus becomes the second farthest grb after grb 971214 ( kulkarni et al .
1998 ) amongst the grbs with known redshift measurements so far . if the original emission were isotropic , the observed fluence of @xmath101 erg/@xmath102 ( hurley et al . 2000 ) between 25100 kev , would yield the total @xmath103ray energy release to be at least @xmath104 erg ( @xmath105 ) .
however , the estimated @xmath43 of 1.74 day implies , using the expression in sari , piran and halpern ( 1999 ) , a jet opening angle of @xmath106 radian , where @xmath107 is the number density of the ambient medium .
this means that the actual energy released from the grb 000926 is reduced by a factor of @xmath108 relative to the isotropic value and becomes @xmath109 ergs . of the over dozen grbs with known redshifts ,
seven with inferred energy budget of @xmath110 erg ( assuming isotropic emission ) are grb 000926 ( discussed here ) ; grb 000301c ( sagar et al .
2000b ) ; grb 991216 and grb 991208 ( sagar et al .
2000a ) ; grb 990510 ( harrison et al . 1999 ) ; grb 990123 ( andersen et al . 1999
; galama et al .
1999 ) and grb 971214 ( kulkarni et al .
recent observations suggest that grbs are associated with stellar deaths , and not with quasars or the nuclei of galaxies as some grbs are found offset by a median value of 3.1 kpc from the centre of their host galaxy ( cf .
bloom et al .
2000 ) . however , release of energies as large as @xmath111 erg or more in radiation is extremely difficult to accommodate within the popular stellar death models ( coalescence of neutron stars or the death of massive stars ) .
however , evidence is now mounting in favour of non - isotropic emission , with grb 000926 the sixth known case after grb 990123 ( cf .
castro - tirado et al .
1999 ; kulkarni et al .
1999 ) , grb 990510 ( cf .
harrison et al . 1999 , stanek et al .
1999 ) , grb 990705 ( cf .
masetti et al .
2000 ) , grb 991216 ( cf .
halpern et al
. 2000c , sagar et al .
2000a ) and grb 000301c ( cf .
sagar et al .
2000b ) .
thus , it is possible that all cases where the inferred isotropic equivalent energy is large , the emission actually is confined within narrow jets , with total energy much below the isotropic equivalent .
the @xmath103ray energy released then becomes @xmath112 erg , a value within the reach of the currently popular models for the origin of grbs ( see piran 1999 and references therein ) . the intrinsic visual extinction ( @xmath113 mag ) derived for the optical afterglow also carries a possible clue to the class of progenitor objects .
extinction of comparable magnitude have been noticed also in grb 971214 ( @xmath114 mag , ramaprakash et al .
1998 and refernces therein ) ; grb 980329 ( @xmath115 mag , palazzi et al . 1998 ) ; grb 980703 ( @xmath116 mag , vreeswijk et al . 1999 ) and grb 000418 ( @xmath117 mag , klose et al . 2000 ) .
the presence of dust extinction in the host galaxies broadly supports the proposal that grbs could be associated with massive stars embedded in star - forming regions of the grb host galaxies ( paczyski 1998 ) .
we have determined @xmath3 magnitudes of 20 stars in the field of grb 000926 .
these magnitudes are used to calibrate present photometric magnitudes of the ot of grb 000926 as well as those given in gcn circulars using differential photometric techniques .
the light curves show a steepening which could be well understood in terms of a jet model .
the parameters of the jet are derived by performing a non - linear least square fit of a broken power - law model to the light curves .
the flux decay constants at early and late times are [email protected] and [email protected] respectively , and the break occurs at [email protected] day .
the break is sharp and similar to that in grb 000301c ( cf .
sagar et al .
2000b ) .
the quasi - simultaneous spectral energy distributions determined for various epochs indicate that the spectral index of the grb 000926 afterglow has not changed significantly during this period .
a steepening of the flux decay with no corresponding change in spectral index is expected of a jet evolution .
the determined redshift yields a minimum distance of 16.6 gpc to grb 000926 , and the observed fluence yields an isotropic equivalent energy release of @xmath118 erg .
however , the break in the light curve at 1.74 days translate to a collimated emission with an opening angle of @xmath119 radian , reducing the estimated energy release to @xmath120 erg .
the late time flux decays at and optical wavelengths seem to be significantly different for grb 000926 afterglow .
similarly , the flux decays of grb 991216 afterglow ( cf .
frail et al .
2000 ) are also different at , optical and radio wavelengths .
the multi - wavelength observations of recent grb afterglows have thus started revealing features which require explanations other than generally accepted so far indicating that there may be yet new surprises in grb afterglows .
* andersen m. i. et al . , 1999 , science , * 283 * , 2075 * bessell m.s . ,
brett j.m . , 1988 ,
pasp , * 100 * , 1134 * bloom j.s . ,
kulkarni s.r . ,
djorgovski s.g . , 2000 ,
astro - ph/0010176 * castro s.m .
2000 , gcn observational report no .
851 * castro - tirado a. j. et al . , 1999 , science , * 283 * , 2069 * dall t. et al . , 2000
, gcn observational report no .
804 * di paola a. et al . ,
2000 , gcn observational report no .
816 * frail d.a . , berger e. , 2000 , gcn observational report no . 805 * frail d.a .
et al . , 2000 ,
apj , * 538 * , l129/astro - ph/0003138 * fynbo j.p.u .
et al . , 2000a , gcn observational report no .
820 * fynbo j.p.u .
, 2000b , gcn observational report no .
825 * fynbo j.p.u .
et al . , 2000c , gcn observational report no .
807 * fynbo j.p.u .
et al . , 2000d , gcn observational report no .
840 * fynbo j.p.u .
et al . , 2000e , gcn observational report no .
871 * fukugita m. , shimasaku k. , ichikawa t. , 1995 , pasp , * 107 * , 945 * galama t.j . , 2000 , to appear in proc . of the 5th huntsville gamma - ray burst symposium / astro - ph/0001465 * galama t.j .
et al . , 1998 ,
apj , * 500 * , l97 * galama t.j .
et al . , 1999 ,
nature , * 398 * , 394 * garmire g. , garmire a. , piro l. , garcia m.r . ,
2000 , gcn observational report no .
836 * gorosabel j. et al . ,
2000 , gcn observational report no .
803 * halpern j.p .
et al . , 2000a , gcn observational report no .
824 * halpern j.p .
et al . , 2000b , gcn observational report no
. 829 * halpern j.p .
et al . , 2000c , apj , * 543 * , 697 * harisson f.a .
et al . , 1999 ,
apj , * 523 * , l121 * hjorth j. et al .
, 2000a , gcn observational report no .
809 * hjorth j. et al . , 2000b , gcn observational report no . 814 * huang y.f . , dai z.g . , lu t. , 2000 , mnras , 316 , 943 * hurley k. , mazets e. , golenetskii s. , 2000 , gcn observational report no .
801 , 802 * klose s. , 2000 , rev .
modern phys .
* 13 * , 129 * klose s. et al , 2000 , apj , * 545 * , 271 * kobayashi n. et al . ,
2000 , gcn observational report no .
821 * kulkarni s.r .
et al . , 1998 , nature , * 393 * , 35 * kulkarni s.r .
et al . , 1999 , nature , * 398 * , 389 * kulkarni s.r .
, 2000 , to appear in proc . of the 5th huntsville gamma - ray burst symposium / astro - ph/0002168
* kumar p. , panaitescu a. , 2000 , apj , * 541 * , l9 * lamb , d.q . , 2000 , astro - ph/0005028 * landolt , a.r . , 1992 ,
aj , * 104 * , 340 * masetti n. et al . , 2000 ,
a&a , 354 , 473 * mathis j.s . , 1990 , araa , * 28 , * 37 * mszros p. , rees m. j. , 1999 , mnras , * 306 * , l39 * monet d. , 1997 , the pmm usno - a1.0 catalogue * paczyski b. , 1998 , apj , * 494 * , l45 * palazzi e. et al
. , 1998 , a&a , * 336 * , l95 * piran t. , 1999 , physics reports * 314 * , 575 * piro l. , antonelli l.a . , 2000 ,
gcn observational report no .
832 , 833 * price p.a .
et al . , 2000a , gcn observational report no .
811 * price p.a .
et al . , 2000b , submitted to apj letters,/astro - ph/0012303 * ramaprakash , a.n .
et al . , 1998 ,
nature , * 393 * , 43 * rhoads j.e . , 1999 , apj , * 525 * , 737 * rol e. , vreeswijk p.m. , tanvir n. , 2000 , gcn observational report no .
850 * sagar r. , pandey a.k , mohan v. , yadav r.k.s . ,
nilakshi , bhattacharya d. , castro - tirado a.j . , 1999 , basi , * 27 * , 3/astro - ph/9902196 * sagar r. , mohan v. , pandey a.k . ,
pandey s.b . ,
castro - tirado a.j .
, 2000a , basi , * 28 * , 15/astro - ph/0003257 * sagar r. , mohan v. , pandey s.b . ,
pandey a.k . ,
stalin c.s . ,
castro - tirado a.j . , 2000b , basi , * 28 * , 499/astro - ph/0004223 * sari r. , piran t. , halpern j. p. , 1999
, apj , * 519 * , l17 * sari r. , piran t. , narayan r. , 1998 , apj , * 497 * , l17 * schlegel d.j .
, finkbeiner d.p . , davis m. , 1998 , apj , * 500 * , 525 * stanek k. z. et al . , 1999 , apj * 522 * , l39 * veillet c. , 2000 , gcn observational report no
. 831 * vrba f. , canzian b. , 2000 , gcn observational report no .
819 * vreeswijk p.m. et al . , 1999 , apj , * 523 * , 171 | the ccd johnson @xmath0 and cousins @xmath1 photometric magnitudes are determined for 20 stars in the field of grb 000926 .
they are used to calibrate the present @xmath2 as well as published @xmath3 photometric magnitudes of the grb 000926 afterglow .
optical light curves of the afterglow emission are obtained in @xmath4 and @xmath2 passbands .
they show a steepening of the flux decay as expected of an anisotropic fireball losing collimation with the fall of the bulk lorentz factor .
we derive the early and late time flux decay constants as @xmath5 and @xmath6 respectively . steepening in the flux decay seems to have started around 1.7 days after the burst .
negligible galactic but relatively large intrinsic extinction amounting @xmath7 0.03 and [email protected] mag respectively are derived in the direction of grb 000926 .
the value of the spectral index in the optical near - infrared region is @xmath9 .
the determination of the redshift @xmath10 indicates a cosmological origin of the burst at a luminosity distance of 16.6 gpc .
the observed fluence in the energy range 20100 kev indicates , if isotropic , the release of @xmath11 ergs of energy . attributing the observed break in the light curve at 1.7 days to the onset of sideways expansion of a jet - like ejecta
, we infer an initial jet opening angle of @xmath120.14 radian .
this indicates a large anisotropy in the original emission and the amount of released energy is reduced by a factor of @xmath12 100 relative to the isotropic value , which can be understood in terms of the currently popular stellar death models . *
grb 000926 and its optical afterglow : another evidence for non - isotropic emission * * r. sagar@xmath13 , s. b. pandey@xmath13 , v. mohan@xmath13 , d. bhattacharya@xmath14 and a. j. castro - tirado@xmath15 * + + _ @xmath14raman research institute , bangalore 560 080 ,
india _ + _ @xmath15iaa - csic , p.o .
box 03004 , e-18080 , granada , spain _ keywords : photometry grb afterglow
flux decay spectral index |
low - mass x - ray binaries ( lmxbs ) consist of an accreting compact star , either a black hole or a neutron star ( ns ) and a low - mass ( @xmath6 ) donor star , in which mass transfer proceeds via roche - lobe overflow ( rlof ) .
there are about 200 lmxbs in the galaxy @xcite .
the formation of black hole lmxbs is still a controversial topic ( * ? ? ?
* for a review ) , and will be discussed elsewhere . here
we focus on the formation and evolution of lmxbs with a ns .
the mainstream idea is that they either have an initially low - mass secondary , or descend from systems with initially intermediate - mass secondaries , i.e. , intermediate - mass x - ray binaries ( imxbs ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
considering the fact that nss must have formed from massive stars , and current lmxbs usually reside in relatively compact binary orbits ( with periods @xmath7days ; see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , the progenitors of these systems should generally have experienced common envelope ( ce ) evolution @xcite . during the ce phase and the subsequent supernova ( sn ) explosion ,
only a small part of the binaries can survive as i / lmxbs @xcite . for reviews on the formation and evolution of i
/ lmxbs , see @xcite and @xcite .
the stability of the mass transfer in i / lmxbs depends on the mass ratio , the orbital period , and the mass and angular momentum loss ( aml ; * ? ? ?
dynamically unstable mass transfer leads to a ( second ) ce evolution , with the ns spiraling into the envelope of the donor star . if the binary survives without merging , it becomes a compact binary with a helium ( he ) star around the ns
the subsequent mass transfer from the he star may lead to the formation of a partially recycled pulsar with a co white dwarf ( wd ) companion @xcite . for those binaries avoiding the ce phase , @xcite pointed out that there exists a bifurcation period ( @xmath8 ) for the initial orbital period , which separates i / lmxbs into converging or diverging systems .
the existence of this bifurcation period is due to the balance between the orbital expansion caused by nuclear evolution of the donor star and the orbital shrinking caused by aml .
binaries with @xmath9 continue to contract along the cataclysmic variable ( cv)-like or ultra - compact x - ray binary ( ucxb ) evolutionary tracks , while those with @xmath10 will become relatively wide binaries containing a recycled ns and a he or co wd @xcite .
@xcite made a systematic investigation of the evolution of i / lmxbs by using a henyey - type stellar evolution code to calculate 100 binary evolution sequences . in that work ,
the initial mass of the ns was fixed as @xmath11 , the initial mass of the donor ranged from @xmath12 , and the initial orbital period went from 4 hr to 100 days .
while the calculated results showed a variety of evolutionary channels which may explain the large diversity in observed lmxbs , they also indicated a number of remaining problems , in particular the low median x - ray luminosities as compared to those of observed lmxbs .
@xcite firstly combined the binary population synthesis ( bps ) method with detailed stellar evolutionary sequences of @xcite to follow the evolution of a population of i / lmxbs .
for early cases b and c mass transfer they adopted a fixed value of the critical ratio of the donor mass and the ns mass ( @xmath13 ) to determine whether the mass transfer is stable .
late cases b and c mass transfer was always assumed to be dynamically unstable followed by a ce phase . when dealing with ce evolution , they adopted constant values of the binding energy parameter @xmath14 , 0.3 and 0.5 , which parameterizes the structure of the donor s envelope .
their calculations demonstrated that incipient imxbs outnumber incipient lmxbs typically by a factor of @xmath15 .
it was also found that rather large values of @xmath16 are required to yield the i / lmxb birthrates that are consistent with the bmsp birthrates , but such values of @xmath17 would lead to an overproduction of luminous lmxbs in the galaxy .
there were still large discrepancies between the orbital period distributions of modeled and observed binary millisecond pulsars ( bmsps ) .
a possible solution to the above problems was to reduce the mean x - ray lifetime of lmxbs by a factor of @xmath18 . in this work ,
we revisit the formation and evolution of galactic i / lmxbs .
similar as @xcite , we perform a systematic evolutionary study of i / lmxbs , combining bps with detailed binary evolutionary sequences . to follow the formation and evolution of i / lmxbs , we take into account updated theoretical results and techniques on binary evolution achieved during the last ten years .
we use the rapid binary evolution algorithm developed by @xcite for bps calculation , which enables modeling of a large number of complex binary systems to obtain the incipient population of i / lmxbs , and an updated version of the @xcite code to calculate the detailed evolution of individual i / lmxbs . with them
we can follow the evolutionary tracks of the i / lmxb population from the incipient stage of a detached binary containing a newborn ns to the rlof phase , and finally to the remnant state with a bmsp .
the parameter distributions of i / lmxbs and bmsps can be clearly displayed , and compared with observations .
the most important and difficult to understand processes during binary evolution are unstable mass transfer and ce evolution ( * ? ? ?
* for a recent review ) . in the bps calculations , instead of using the empirical relations for the critical mass ratios
, we employ the numerically calculated results of @xcite to judge the stability of mass transfer in a primordial binary , and the parameter space obtained by ( * ? ? ?
* hereafter paper i ) to determine whether i / lmxbs can successfully evolve to be binary pulsars .
we also adopt the calculated values of @xmath19 given by @xcite to calculate the orbital change during the ce phase , which depend on the evolutionary state of the donor .
other improvement will be mentioned below .
although these updated parameters are still subject to large uncertainties , they provide physically more reasonable and realistic input as compared to previous bps studies .
this paper is organized as follows . in section 2
we describe the methods used in this paper .
we present the detailed results derived from our calculations and discuss their implications in section 3 .
we conclude in section 4 .
we use the bps code developed by @xcite and modified by @xcite to generate the incipient i / lmxb population .
this code , based on single star evolution with a series of analytic formulae @xcite , includes standard assumptions for the population of massive primordial binaries and analytic prescriptions to describe binary interactions , mass transfer and sn explosions .
we have updated the code in various aspects , in particular the formation and evolutionary processes of compact objects .
some key points are described below ( see * ? ? ? * for more details ) .
the stability of mass transfer is determined by the change of the donor s radius with respect to its rl radius .
usually a critical mass ratio @xmath20 (= donor mass / accretor mass ) is used as a criterion : if the mass ratio is larger than @xmath20 at the onset of rlof , the mass transfer is unstable and results in a ce evolution ; otherwise the mass transfer proceeds stably on a nuclear or thermal timescale . in previous studies , the values of @xmath21
were usually estimated by only comparing the mass - radius exponents for the donor stars , which describe the response of the star and the rl to mass loss @xcite .
however , it is well known that mass accretion can spin up the accretor @xcite and probably result in mass loss @xcite .
rapid mass accretion can drive the accreting star out of thermal equilibrium and cause it to expand , and this expansion may even cause the accretor to fill its own rl , leading to the formation of a contact binary @xcite . @xcite
numerically calculated the values of @xmath22 for a grid of binaries with different values of component masses and orbital periods , considering the responses of both the donor to mass loss and the accretor to mass accretion .
since the stellar expansion strongly depends on the accretion rate , they constructed three models with different assumed mass accretion histories . in model i , the mass accretion rate depends on the rotational velocity of the accretor .
since a small amount of transferred material is able to spin it up to be critically rotating @xcite , there is actually very small mass accreted in this case . in model ii
, it is assumed that half of the transferred mass is accreted and the other half is ejected out of binary system . in model iii ,
the mass transfer is assumed to be almost conservative .
the obtained values of @xmath20 vary with the orbital period , and can reach @xmath23 in model i , but are generally less than @xmath24 and @xmath25 in models ii and iii , respectively . for unstable mass transfer , we use the standard energy conservation equation @xcite to describe the subsequent ce evolution , @xmath26 where the indices 1 and 2 denote the primary and the secondary , and i and f the initial and final values , respectively ; @xmath27 is the mass of the primary@xmath28s envelope that is ejected out of the binary during the ce phase , @xmath29 is the rl radius of the primary at the onset of rlof , and @xmath30 is the ce efficiency with which the orbital energy is used to unbind the stellar envelope .
it has been pointed out that the internal energy within the envelope can also contribute to envelope ejection , and thus modify the values of @xmath19 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we employ the values of @xmath17 calculated by @xcite for a range of massive and intermediate - mass stars at various evolutionary stages , and take @xmath31 = 1.0 in the calculations .
we assume that nss are born from either core - collapse or electron - capture sne , depending on the masses of the progenitor stars .
we follow the criterion suggested by @xcite and @xcite to distinguish them .
the he core mass at the base of asymptotic giant branch is used to decide the formation of various co cores . ( 1 ) if the he core mass is less than @xmath32 , the star forms a degenerate co core and ends up forming a co wd .
( 2 ) if the core is more massive than @xmath33 , the star ultimately forms a feni core and collapses to a ns .
( 3 ) stars with core masses between @xmath34 and @xmath35 form partially degenerate one cores .
if in subsequent evolution the core mass increases to @xmath36 , then the core collapses to a ns due to electron capture ; otherwise the star leaves an one wd . under these prescriptions , for single stars , the stellar masses for core - collapse and electron - capture sne are in the range of @xmath37 , and @xmath38 , respectively @xcite .
the masse interval for electron - capture sne is in rough accordance with @xcite .
it depends on the assumptions for mass loss and the dredge - up process in stellar evolution ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and is blurred by mass transfer in binaries @xcite .
the newborn nss are assumed to receive a kick due to asymmetric sn explosions , and we utilize a maxwellian distribution for the kick velocity , @xmath39 assuming that the directions of the kicks are isotropically distributed .
we adopt @xmath40 @xcite and @xmath41 @xcite for nss formed from core - collapse and electron - capture sne , respectively .
the initial parameters in our bps calculation are taken as follows .
all the binary orbits are assumed to be circular due to tidal interactions @xcite .
the mass @xmath42 of the primary is in the range of @xmath43 , following the initial mass function of @xcite . ] .
the initial mass ratio @xmath44 has a uniform distribution between 0 and 1 , but only binaries with the secondary mass @xmath45 between @xmath46 and @xmath47 are chosen .
the logarithm of the orbital separation @xmath48 is uniformly distributed , lying between @xmath49 and @xmath50 .
we adopt solar metallicity , and assume a constant star formation rate of @xmath51 over the past 12 gyr @xcite .
@xcite obtained a smaller star formation rate of @xmath52 derived from _
detected pre - main - sequence stars .
if we adopt the latter value , the predicted number of galactic i / lmxbs will be reduced by several times , but the features of their distribution will not change . in fig . 1
, we plot the birthrates of incipient i / lmxbs as a function of the donor mass @xmath53 ( in the range of @xmath54 ) . in the left and right panels the nss
are assumed to form from core - collapse and electron - capture sne , respectively .
the green , red and black curves correspond to the results with the @xmath20 values from models i , ii and iii , respectively .
it is clearly seen that , in both core - collapse and electron - capture sn channels , imxbs dominate the whole population as expected .
although core - collapse sn - imxbs in model i have a birthrate higher than in other models by @xmath55 , the overall birthrate distributions do not show a remarkable difference .
thus we only take the calculated birthrates in model ii in the following study , not only because it is intermediate between the other two more extreme ones , but also because the corresponding results can better match the observed distribution of be / x - ray binaries @xcite . in fig . 2 we plot the distributions of incipient
i / lmxbs in the orbital period - donor mass plane ( left panel : the core - collapse sn channel ; right panel : the electron - capture sn channel ) .
after sne the binary orbits become eccentric due to mass loss and the kicks imparted on the nss . here
we assume that they are quickly circularized by tidal torques with the orbital am conserved .
the corresponding orbital separation is then set to be @xmath56 , where @xmath57 is the semimajor axis after the sne .
also shown are the histogram distributions of the birthrates , same as the red curves in fig . 1 .
in the core - collapse sn channel , the ns progenitors masses are @xmath58 , the resultant i / lmxbs have orbital periods @xmath59 days , and most of them have @xmath60 a few days .
the i / lmxbs birthrate roughly increases with the donor mass , because systems with lower - mass companions are more likely to merge during the ce phase ( see also * ? ? ?
in the electron - capture sn channel , the masses of the ns progenitors lie in the range @xmath61 .
the primordial binaries should have orbital periods longer than @xmath62 days , otherwise the mass transfer starts too early so that a wd rather a ns is produced .
the binding energies for such stars are relatively small , with the values of @xmath63 greater than unity when on asymptotic giant branch @xcite .
so the resultant i / lmxbs have orbital periods in a wide range @xmath64 days , but the majority are relatively long period ( @xmath65 days ) systems .
the i / lmxb birthrate slightly decreases with increasing donor mass .
the reason is that , for relatively massive companion stars , the orbital periods after ce are usually long because there are sufficient orbital kinetic energies .
the second mass transfer may not occur , so the primaries undergo core - collapse sne , which can readily disrupt the wide binaries .
finally there are no binaries with @xmath66 day and @xmath67 , because magnetic braking ( mb ) and gravitational radiation ( gr ) can not drive these systems into contact within hubble time .
based on the results in last subsection we then employ an updated version of the @xcite code to calculate the evolution of i / lmxbs . before doing this ,
we first select the binaries in which mass transfer is stable so that they can evolve to be binary pulsars .
super - edddinton accretion onto nss in imxbs may lead to delayed dynamical instability , and eventually ce evolution
. however , there are observational and theoretical hints for mass loss from the accreting nss that can stabilize the mass transfer ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
@xcite surveyed the allowed parameter ( @xmath68 ) space for stable mass transfer , and showed that the donor masses @xmath69 can be up to @xmath70 for a @xmath71 ns .
observations of high - mass x - ray binaries , in which the nss have accreted very small mass , show that the initial masses of nss range from @xmath72 for 4u 1538@xmath7352 to @xmath74 for vela x@xmath731 @xcite . in paper
i we extended the calculation by considering the ns masses @xmath75 in the range of @xmath76 .
the results are used to compare with the obtained incipient i / lmxb distribution in fig . 3
( binaries formed from both core - collapse and electron - capture sn channels are included ) .
the left and right panels are for nss with mass @xmath77 and 1.8@xmath78 , respectively .
the binaries distributed in the regions confined by the thick grey curves can stably evolve without entering the ce phase , and are selected for further investigation .
their corresponding birthrates are shown with the blue curves , totally ranging from @xmath79 to @xmath80 .
figure 4 demonstrates the distributions of the birthrates for these systems as a function of the initial donor masses and the orbital periods . here
the bin size of the donor mass is 0.2 @xmath81 and the orbital period ( in units of days ) changes logarithmically in steps of 0.2 .
we calculate the evolution of the selected i / lmxbs in a similar way as in paper i , taking into account aml caused by mb @xcite and gr @xcite .
the accretion rate of a ns is assumed to be limited by the eddington accretion rate , and we adopt the isotropic re - emission model , assuming that the excess material leaves the binary in the form of isotropic winds from the ns , carrying off the ns@xmath28s specific orbital am . figures 5 and 6 present two example evolutionary tracks . in fig .
5 the initial parameters are @xmath82 , @xmath83 and @xmath84 day .
the left , middle , and right panels depict the evolution of the orbital period @xmath85 with the age , the evolution of the donor radius @xmath86 and the mass transfer rate @xmath87 with the donor mass @xmath69 , respectively .
the donor starts to fill its rl and initiate ( case a ) mass transfer at @xmath88 gyr when @xmath89 day , mainly driven by mb .
the mass transfer rate first goes up to @xmath90yr@xmath4 , then decreases gradually as the orbital period reduces to @xmath91 day .
meanwhile , the donor mass decreases to be less than @xmath92 , and the donor becomes fully convective , so mb stops working .
the evolution afterwards is solely driven by gr , and the mass transfer rate drops down to @xmath93yr@xmath4 . at the end of the evolution ( the age of @xmath94 gyr ) , the donor mass and the orbital period
are reduced to @xmath95 and @xmath96 day respectively .
the duration of the mass transfer is about 3.77 gyr . in fig .
6 we show the evolutionary path with the same component masses but @xmath97 days .
the mass transfer starts at the age of 11.9 gyr , when the donor evolves to become a subgiant . during this case
b mass transfer , the orbital period increases all the time , up to 73 days .
the mass transfer rate varies in the range @xmath98yr@xmath4 ( the gap in the mass transfer rate is caused by the so - called bump - related detachment ; * ? ? ?
the binary finally leaves a ns with a he wd with mass of @xmath99 .
the mass transfer liftime is about 0.14 gyr .
the above examples suggest that , given the same donor star , the longer the initial orbital periods , the higher the averaged mass transfer rates , and the shorter the x - ray lifetimes @xcite .
this strongly influences the characteristics of observable lmxbs .
according to the calculated evolutionary sequences of the i / lmxbs , we plot their number distribution in the orbital period - mass transfer rate ( @xmath100 ) plane ( fig . 7 ) .
the left and right panels correspond to the results with a @xmath101 and @xmath102 ns , respectively .
each panel contains @xmath103 matrix elements , in which @xmath104 changes logarithmically from 0.0316 day to 1000 days with steps of 0.1 , and @xmath105 from @xmath106yr@xmath4 to @xmath107yr@xmath4 with steps of 0.125 .
the color reflects the number of i / lmxbs in the corresponding matrix element by accumulating the product of the birthrates of x - ray binaries passing through it with the time - span . a comparison of figs .
4 and 7 clearly demonstrates the change in the orbital periods during the evolution of i / lmxbs . as we have already seen from figs . 5 and 6 , in x - ray binaries with initially long orbital periods ,
the donors have evolved to be ( sub)giant stars at the onset of mass transfer , and the high mass transfer rates imply a short x - ray lifetime ; in short - period x - ray binaries , the donors are still on main - sequence during mass transfer , so the systems have a long - lasting mass transfer phase .
this difference in the evolutionary time leads to different number distribution in fig . 7 ,
which explains why most i / lmxbs tend to have relatively short orbital periods ( @xmath108 day ) .
another factor that influences the number distribution of lmxbs is their transient behavior .
the general idea is that , if the mass transfer rate is smaller than a critical mass transfer rate for a given orbital period , the accretion disk is likely subject to thermal and viscous instability , and the lmxb may appear as a transient x - ray source @xcite . in lmxbs irradiation from the ns
may help stabilize the disk to some extent by increasing the surface temperature @xcite .
the critical mass transfer rate for the disk instability is given by @xcite , @xmath109 and plotted with the red dashed line .
we assume that lmxbs under this line experience short outbursts separated by long quiescent intervals .
the majority of long - period lmxbs are transient sources spending most of their time in quiescence days .
this puzzle can be partly understood by the transient behavior of x - ray msps ( * ? ? ?
* and references therein ) .
these results are consistent with @xcite , who showed that more than @xmath110 persistent ns lmxbs have main - sequence donors and orbital periods between @xmath111 hours .
comparing the two panels of fig .
7 also shows that , lmxbs with a heavier ns have a larger coverage in the parameter space , since mass transfer in these binaries are more likely to be stable ( see fig .
3 ) . in fig .
7 we also plot galactic lmxbs with known orbital periods and accretion rates . here
the observational data are taken from ritter s lmxb catalogue @xcite , and from @xcite and @xcite . the filled circles and triangles represent persistent and transient lmxbs , respectively .
it is clearly seen that most lmxbs are compact systems except two sources : cyg x@xmath732 with a @xmath112 companion star in a 9.84 day orbit @xcite , and gx13@xmath1131 ( shown with the star symbol ) , which has the longest known orbital period ( of 601.7 hours ) for a galactic ns lmxb powered by rlof @xcite .
figure 8 depicts the distributions of the masses ( left panel ) and effective temperatures ( right panel ) of the donor stars .
the upper and lower panels are for persistent and transient lmxbs , and the black and red lines correspond to @xmath114 and @xmath115 nss , respectively .
it is seen that lmxbs dominate the population at the current epoch , because the initial phase of thermal - timescale mass transfer in imxbs , where a large fraction of the secondary mass is removed , is relatively short - lived .
the donor masses in lmxbs cluster around @xmath116 .
the donors temperatures are distributed in the range of @xmath117 k. figure 9 presents a more detailed comparison of the modeled orbital period ( left panel ) and accretion rate ( right panel ) distributions ( in red lines ) with observations ( in black lines ) .
the upper and lower panels are for @xmath118 and @xmath115 nss , respectively .
the numbers for observed sources are amplified by a factor of 500 .
although there seems broad agreement between the measured and predicted orbital periods , there exist a systematic discrepancy in the accretion rates , i.e. , the calculated average accretion rates ( @xmath93yr@xmath4 ) are lower than derived from observations by about an order of magnitude .
to quantitatively compare the univariate distributions of the calculated and observed quantities , we use kolmogorov - smirnov test to assist inference on their consistency @xcite .
the cumulative distribution functions ( cdfs ) for the calculated and measured orbital periods of the i / lmxbs are expressed as @xmath119 and @xmath120 , respectively .
we want to test the null hypothesis that @xmath121 for all @xmath104 , against the alternative that @xmath122 for some @xmath85 . by comparing the cdfs
, we can measure the maximum distance @xmath123 , where @xmath124 and @xmath125 are the numbers of x - ray binaries in the calculation and the observation , respectively . since @xmath126 , @xmath127 . at a given significance level @xmath128
, there exists a critical value @xmath129 , and @xmath130 allows rejection of the null hypothesis , indicating that the two distributions are significantly different .
figure 10 shows the cdfs for the orbital period ( left panel ) and the accretion rate ( right panel ) of the i / lmxbs .
the black , red and green curves represent the calculated cdfs with the ns mass of @xmath101 and @xmath102 , and the observed cdf , respectively . for the orbital period distribution , we obtain @xmath131 and 1.15 with a @xmath77 and @xmath132 ns , respectively ; while for the accretion rate distribution , @xmath133 and 3.4 , suggesting that the difference is significant .
we note that similar conclusions were also reached by @xcite and @xcite . here
one needs to be cautious that , although the numbers of lmxbs in the galactic disk with measured orbital periods and accretion rates have been substantially increased ( 51 and 45 respectively ) since then , this comparison is still severely subject to small number statistics as well as observational selection effects , e.g. , luminous sources are more likely to be observed . since in this work several key procedures involved in the formation processes of i / lmxbs have been significantly upgraded , the accretion rate discrepancy , if really exists , strongly suggests that something is missing or needs to be modified when modeling the evolution of lmxbs . a possible mechanism that may help solve
the problem is that there are extra aml mechanisms in the lmxb evolution that are not taken into account .
the calculated accretion rate distribution is peaked around @xmath134 yr@xmath4 , implying that the mass transfer is driven by gr .
there is similar situation in cvs , in which aml is thought to be dominated by mb and gr above and below the period gap , respectively .
however , modeling cv evolution indicates that the aml rate below the period gap is @xmath135 times the gr rate @xcite , suggesting the existence of some other aml mechanisms .
@xcite considered several kinds of aml mechanisms including isotropic wind from nss , outflows from the inner and outer langrangian points , and the formation of a circumbinary disk .
they found that only outflow from the outer langrangian point ( or a circumbinary disk ) can account for the extra aml rate , provided that @xmath136 of the transferred mass is lost from the binary .
these consequential aml mechanisms can not only substantially enhance the mass accretion rates , but also influence the value of the bifurcation periods ( e.g. , * ? ? ?
another possible origin of the accretion rate discrepancy is that the calculated secular mass transfer rate may not be identical to the accretion rate derived based on short time ( no more than @xmath137 years ) observations .
for example , x - ray irradiation of the donor star may driven mass transfer cycles in lmxbs @xcite . in this picture ,
once mass transfer begins , the accretion luminosity enhances the irradiation flux on the donor .
thermal expansion of the donor subsequently triggers an increasing mass transfer rate ( this stage is denoted as high state ) .
the donor evolves to a new thermal equilibrium state on kelvin - helmholtz timescale . when the donor nearly reaches the equilibrium state , its expansion rate decreases , then the mass transfer rate drops , and so does the irradiation flux .
the donor starts to contract , and its intrinsic luminosity increases .
this reduces the mass transfer to a quite low rate ( denoted as low state ) .
hence the lmxb undergoes cyclic mass transfer alternating between high and low states .
although spending most of their evolutionary time in low state , they are more likely to be detected in high state , in which the mass transfer rates can be significantly larger than the average ones . in both cases mentioned above the x - ray lifetime
is shortened , which can increase the birthrates of lmxbs derived from the observed number in the galactic disk .
however , the detailed processes related to either extra aml loss or cyclic mass transfer are rather uncertain and depend on poorly known input parameters .
numerical calculations show that irradiation in lmxbs may boost the mass transfer rates by tens of times @xcite .
furthermore , @xcite argued that the mass transfer cycles are not likely to work for transient sources with an unstable disk , because irradiation can not be sustained in a long quiescent stage . to illustrate the effect of irradiation in a simplified way , we arbitrarily increase the mass transfer rates of persistent lmxbs by a factor of @xmath138 , which is randomly distributed between 1 and 30 .
the corresponding evolutionary time is then reduced by the same factor .
the resultant distribution is plotted in fig .
11 , and we see that it seems to better match the observational data , especially for the luminous sources .
the values of @xmath139 decreases to be 3.8 and 2.9 correspondingly .
similar as in paper i , we follow the evolution of lmxbs to the formation of binary msps , and plot their orbital period distributions with the black lines in fig . 11 . in most cases
the mass transfer terminates when the donor star loses its hydrogen - rich envelope and becomes a he or co wd . for converging lmxbs
, we do not follow the evolution when the systems nearly reach the minimum periods .
they might evolve to be black widow systems ( e.g. , * ? ? ? * ; * ? ? ? * ) . at the end of each evolutionary sequence , we know the final orbital period and the secondary mass . by summing the calculated sequences , we obtain the orbital period distributions of bmsps .
we assume that the lifetimes of all bmsps are @xmath140 yr , so their numbers in each bin of the orbital period depends on their birthrates . in fig .
12 we also plot the observed orbital period distributions of binary pulsars in the red lines .
the upper and lower panels correspond to the initial ns mass of 1.0@xmath141 and 1.8@xmath142 , respectively .
depending on the amount of the accreted mass , the nss may be fully or partially recycled .
it is well known that accretion of @xmath143 material is sufficient to spin up a ns to millisecond period ( e.g. , * ? ? ?
* ) , so we distinguish bmsps from other partially recycled pulsars using this criterion . in the left panel
we show the theoretical and observational histograms for both fully and partially recycled nss , that is , those with any accreted mass @xmath144 , and those with measured spin periods @xmath145 1 s , respectively . in the right panel
we are confined with bmsps , that is , those with @xmath146 , and those with @xmath147 10 ms ( data from the atnf pulsar catalogue ; * ? ? ?
the orbital periods of galactic binary pulsars extend to @xmath148 1000 days , in agreement with our calculated results with @xmath149 , while those of bmsps tend to have systematically small values .
the known bmsps have orbital periods up to 200 days , longer than predicted if @xmath150 but likely compatible with that if @xmath151 .
however , our calculated distributions show a shortage of bmsps with @xmath152 day , which conflicts with observations ( see also * ? ? ? * ) .
figure 13 presents the cdfs for the orbital periods of binary pulsars with @xmath153 day .
similar as in fig .
12 , the left and right panels correspond to all binary pulsars and bmsps , respectively .
the line styles are same as in fig .
there are 115 and 77 binary pulsars in the galactic disk with @xmath154 s and @xmath155 ms in the atnf pulsar catalogue , respectively .
at a significance level @xmath156 , @xmath157 . in the left panel , @xmath158 and 3.9 for @xmath159 and @xmath160 , respectively .
in the right panel , @xmath161 and 2.1 for @xmath162 and @xmath160 , respectively . in both cases
the @xmath163 values are larger than @xmath164 at @xmath165 days .
the rarity of the predicted narrow bmsps is likely to be related to the accretion rate discrepancy discussed before , as the orbital period evolution largely depends on the mass transfer process .
it is caused by the fact that in our ( and almost all previous ) calculations rlof is mainly driven by nuclear evolution of the donor or mb , so that the bifurcation period is around @xmath166 day @xcite . for binaries with initial orbital periods around it
, mass transfer will finally lead to either wider or narrower orbits . only with the initial parameters in a very small parameter space
can lmxbs evolve into this orbital period range .
it seems that neither changing the mb law nor considering the feedback of x - ray irradiation can solve this problem in a satisfactory manner ( see also * ? ? ?
* for a relevant discussion ) .
similar problems also exist for the formation of ucxbs .
it was found that with the standard mb law , only a very small range of initial parameters are allowed for the binaries to evolve to ucxbs within hubble time @xcite ; if a saturated mb law is adopted , no ucxbs can be formed at all @xcite .
this makes it impossible to account for the relatively large number of observed ucxbs .
all these facts point to some missing ( or misunderstood ) aml mechanisms in the lmxb evolution , which will be investigated in a future work .
in this work , we incorporate detailed evolutionary calculations with bps to explore the formation and evolutionary tracks of i / lmxbs in the galaxy . by this way
, we can trace the evolution of the i / lmxbs from their incipient stage to the descendant binary pulsars . here
we summarize the results as follows .
the incipient i / lmxbs contain nss formed by both core - collapse and electron - capture sne .
their orbital periods are distributed in the range of @xmath167 days and @xmath168 days , respectively .
the total birthrate of galactic i / lmxb population is @xmath169 with a star formation rate of @xmath170yr@xmath4 . \2 .
since the observability of i / lmxbs depends on the x - ray lifetimes and the stability of the accretion disks , galactic lmxbs tend to be compact binaries with orbital periods @xmath1 1 day and donor masses @xmath171 .
these features are consistent with those of observed lmxbs . however , the average mass transfer rate @xmath172yr@xmath4 for persistent x - ray binaries is considerably lower than observed .
the orbital periods of bmsps are predicted to extend from less than 1 day to hundreds of days , covering the majority of observed bmsps .
however , the number of bmsps with @xmath165 day is severely underestimated .
this work was supported by the natural science foundation of china under grant numbers 11133001 , 11203009 and 11333004 , and the strategic priority research program of cas ( under grant number xdb09000000 )
. plane .
the red curves reflect their birthrates as a function of @xmath173 , same as in fig . 2 .
the grey curves confine the parameter space for stable mass transfer in x - ray binaries with an initial 1.0@xmath78 ( left panel ) and 1.8@xmath78 ( right panel ) ns , and the birthrates for binaries within it are shown with the blue curves .
[ figure],title="fig : " ] plane .
the red curves reflect their birthrates as a function of @xmath173 , same as in fig . 2 .
the grey curves confine the parameter space for stable mass transfer in x - ray binaries with an initial 1.0@xmath78 ( left panel ) and 1.8@xmath78 ( right panel ) ns , and the birthrates for binaries within it are shown with the blue curves .
[ figure],title="fig : " ] | we investigate the formation and evolutionary sequences of galactic intermediate- and low - mass x - ray binaries ( i / lmxbs ) by combining binary population synthesis ( bps ) and detailed stellar evolutionary calculations . using an updated bps code
we compute the evolution of massive binaries that leads to the formation of incipient i / lmxbs , and present their distribution in the initial donor mass vs. initial orbital period diagram .
we then follow the evolution of i / lmxbs until the formation of binary millisecond pulsars ( bmsps ) .
we find that the birthrate of the i / lmxb population is in the range of @xmath0 , compatible with that of bmsps which are thought to descend from i / lmxbs .
we show that during the evolution of i / lmxbs they are likely to be observed as relatively compact binaries with orbital periods @xmath1 1 day and donor masses @xmath2 .
the resultant bmsps have orbital periods ranging from less than 1 day to a few hundred days .
these features are consistent with observations of lmxbs and bmsps .
we also confirm the discrepancies between theoretical predications and observations mentioned in the literature , that is , the theoretical average mass transfer rates ( @xmath3yr@xmath4 ) of lmxbs are considerably lower than observed , and the number of bmsps with orbital periods @xmath5 day is severely underestimated .
these discrepancies imply that something is missing in the modeling of lmxbs , which is likely to be related to the mechanisms of the orbital angular momentum loss . |
cognitive radio is a subject of intense interest motivated by its potential for better usage of spectral resources . to explore the fundamental limits of such channels , and to make use of powerful techniques developed for capacity of channels with state known at transmitter , some information - theoretic cognitive models allow a cognitive node to possess non - causal knowledge of data originating elsewhere .
interestingly , in the recent years applications have emerged where knowledge of another nodes data prior to transmission is indeed practically viable .
examples include heterogeneous networks or coordinated networks , where some base stations can possess knowledge of the messages of other base stations by coordination .
other examples involve layered cell structures , where macro base stations can know the messages of pico base stations that are routed from the macro base station over backhaul links .
such heterogeneous or coordinated networks can be modeled by interference channels with cognitive transmitters @xcite .
contrary to the model in @xcite , when a cognitive transmitter does not have its intended receiver , it is called cognitive relay and helps other transmitters in a way of reducing the effective interference at the receivers . in this paper
, we consider the interference channel with a cognitive relay ( iccr ) where transmitter - receiver pairs constitute an interference channel and the cognitive relay helps the transmitters .
previous works in this area have generally focused on perfect and instantaneous channel state information at transmitter ( csit ) .
however , feedback delays are often present in real systems and make feedback information outdated .
fortunately , the usefulness of delayed csit has been explored in various channel models @xcite .
the iccr with delayed feedback , nevertheless , has not received much attention despite its importance .
the iccr was first considered in @xcite where an achievable rate region via a combination of dirty paper coding @xcite and beamforming was reported . in @xcite , a new achievable region
was presented by a combination of the han - kobayashi coding scheme @xcite and dirty paper coding , and an outer bound for the gaussian iccr was derived . for a discrete memoryless ( dm )
iccr , an outer bound was first derived in @xcite and then improved achievable rate regions and outer bounds were reported in @xcite .
the capacity region of dm - iccr is known in very strong and strong interference regime @xcite , but it still remains unknown under general channel conditions .
when capacity remains intractable , the degrees of freedom ( dof ) are often used to understand the asymptotic characteristics of the capacity .
the dof is defined as the ratio of the capacity of the channel of interest to a simple siso gaussian channel capacity , when transmit power goes infinity .
the dof of iccr has been studied in @xcite .
it was proved in @xcite that the two - user gaussian iccr has dof @xmath2 almost surely if perfect and instantaneous csit and csi at receiver ( csir ) are available , which implies that each receiver does not suffer from interference in an asymptotic sense . for the @xmath3 users with perfect csit and csir , achievable sum dof and outer bounds of the sum dof
were derived in @xcite .
although a conventional relay can not increase the dof @xcite , the authors of @xcite showed that a cognitive relay can improve dof unlike a conventional relay ; the optimal sum dof for @xmath3 users with perfect csit is @xmath4 if @xmath3 is odd while the sum dof for the @xmath3-user interference channel with perfect csit is @xmath5 by interference alignment @xcite .
the usefulness of delayed csit has been first demonstrated in @xcite for multiple - input single - output broadcast channel ( bc ) . in @xcite ,
the base station exploits the delayed csi to estimate the interference at each receiver in the previous transmission ( i.e. , the side information at the receivers ) and then retrospectively align the interfering signals with the help of the side information . for multiple - input multiple - output ( mimo ) bc with delayed csit ,
an outer bound of the dof region with @xmath3 users and the dof region with two users were derived in @xcite , and sum dof for a three - user case was obtained in @xcite .
new retrospective interference alignment schemes for an interference channel and an @xmath6 channel with delayed csit and delayed output feedback were proposed in @xcite , and the sum dof was derived .
the achievable dof reported in @xcite was improved in @xcite .
recently , @xcite proved the usefulness of ergodic interference alignment in a @xmath3-user interference channel when only delayed feedback is available and showed that the sum dof of 2 can be achievable as @xmath3 goes to infinity , which is the best dof result until now in a @xmath3-user interference channel with delayed csit .
derived in @xcite . in @xcite , the dof region with delayed csit was derived for general mimo interference channel with an arbitrary numbers of antennas .
the authors of @xcite showed that shannon feedback , which has both output feedback and delayed csit , strictly enlarges the dof region of the mimo interference channel compared to the case with delayed csit only @xcite . for delayed local csit , an achievable dof region of mimo interference channel
was derived in @xcite . in @xcite ,
the authors presented a hybrid csit model where one transmitter has perfect and instantaneous knowledge of channel matrices corresponding to one user while the other transmitter has only delayed csi corresponding to the other user , and derived the dof region of the mimo interference channel with hybrid csit .
moreover , the dof regions of mimo interference channel and broadcast channel without csit were derived in @xcite , and in addition to mimo interference channel and broadcast channel , the dof region of a cognitive radio channel without csit was reported in @xcite . in this paper
, we consider the interference channel with cognitive relay in the presence of various types of delayed feedback at the transmitter in independent and identically distributed ( i.i.d . ) fading channels . in all cases
perfect csir is assumed . unless explicitly mentioned otherwise ,
a two - user system is considered .
the types of delayed feedback information ( including no feedback ) are * delayed csit : transmitters and a cognitive relay know all channels after one sample delay . * delayed output feedback : transmitters and a cognitive relay know output of their intended receiver after one sample delay .
* delayed shannon feedback : transmitters and a cognitive relay know all channel gains and the output of intended receiver after one sample delay . * no feedback : both transmitters and a cognitive relay do not have any channel information .
for each type of delayed feedback , an outer bound for the dof region is derived .
focusing on the special case of the single - input single - output ( siso ) , a scheme is proposed that achieves the outer bound based on the retrospective interference alignment for each type of feedback . from the derived dof region , it is shown that the sum dof in the single - antenna network is @xmath7 for delayed csit , delayed output feedback , and delayed shannon feedback with the help of a cognitive relay , compared with the sum - dof of the interference channel which is only 1 regardless of availability of csit .
it is also shown that a cognitive relay does not extend the dof region in the absence of csit .
the proposed retrospective interference alignment scheme is extended to the mimo case .
it is shown that for the three types of delayed feedback information , the dof regions achieved by the proposed retrospective interference alignment scheme are similar , matching the dof outer bound for all antenna configurations . comparing with the dof region
when delayed feedback information is not available at both the transmitters and the cognitive relay , the delayed feedback information is useful for expanding the dof region when @xmath8 where @xmath9 , @xmath10 , and @xmath11 are the numbers of antennas at the transmitter , the relay , and the receiver , respectively . if delayed feedback is not available at the cognitive relay , the optimal dof region is shown to be achievable except when @xmath12 by the proposed retrospective interference alignment scheme .
our results quantitatively reveal the dof gain from the cognitive relay according to antenna configurations when only delayed feedback information is available .
finally , we compare the sum dof of iccr with those of broadcast channel and interference channel when only delayed csit is available . with the help of the cognitive relay
, iccr has an enlarged dof region compared to mimo interference channel .
furthermore , as a corollary of the above - mentioned results , lower and upper bounds are derived for the sum dof of a cognitive interference channel which is also known as a interference channel with a cognitive transmitter .
this paper is organized as follows .
section [ model ] describes the system model with various types of delayed feedback information . in section [ siso ] , focusing on the siso model as a special case , we propose a modified retrospective interference alignment scheme achieving the outer bound for siso model under various types of delayed feedback information .
section [ mimo_cr ] derives the dof regions for the multiple antenna scenarios .
section [ achievable_dof ] derives the achievable dof region when delayed feedback information is not available at the cognitive relay . in section [ outerbound ]
, we derive the dof outer bounds with and without delayed feedback information .
section [ discussion ] discusses a comparison with bc and ic under delayed csit and an extension to cognitive interference channel .
section [ conclusion ] concludes the paper .
this paper considers a mimo network consisting of two transmitters with @xmath9 antennas , two receivers with @xmath11 antennas , and a cognitive relay with @xmath10 antennas as shown in fig .
[ fig_ic_cr ] , where the desired and interference links are represented by solid and dashed lines , respectively .
the links experience i.i.d .
rayleigh fading .
transmitter @xmath13 has message @xmath14 intended for receiver @xmath13 , transmitter @xmath15 has message @xmath16 intended for receiver @xmath15 , and the cognitive relay has both @xmath14 and @xmath16 non - causally , where the messages @xmath14 and @xmath16 are independent .
channel outputs at time slot @xmath17 are @xmath18 @xmath19 where @xmath20,t},\cdots , y_{j[m_r],t}]^t \in\mathbb{c}^{m_r\times 1},\ j\in\{\a,\b\}$ ] , is the received signal at receiver @xmath21 , @xmath22,t}$ ] is the @xmath23-th element of @xmath24 , @xmath25 , is the transmitted signal from transmitter @xmath26 , @xmath27 is the transmitted signal from the cognitive relay , @xmath28 is the time varying channel matrix from transmitter @xmath26 to receiver @xmath21 , @xmath29 is time varying channel matrix from the cognitive relay to receiver @xmath21 , and @xmath30 is an i.i.d .
circular symmetric complex gaussian noise , @xmath31 , at receiver @xmath32 .
we assume that all channel coefficients are i.i.d .
circular symmetric complex gaussian random variables with zero mean and unit variance , @xmath33 . in this paper
, we assume perfect csir .
certain feedback information is available at the transmitters and cognitive relay with delay , represented by the following four cases , where @xmath34 and @xmath17 is the time index : 1 . delayed csit : @xmath35 , @xmath36 2 . delayed output feedback : @xmath37 , @xmath38 3 . delayed shannon feedback : @xmath39 , @xmath40 4 .
no feedback : @xmath41 @xmath42 each message @xmath43 is uniformly distributed , @xmath44 and @xmath45 are , respectively , encoding functions at transmitter @xmath26 and the cognitive relay for channel use @xmath17 and @xmath46 is the set of all channel matrices at time index @xmath17 , i.e. , @xmath47 @xmath48 and @xmath49 should satisfy the power constraint @xmath50\leq p$ ] and @xmath51\leq p$ ] , respectively where @xmath34 .
receiver @xmath26 decodes the message from the received signal with a decoding function @xmath52 such that @xmath53 . a rate pair @xmath54 is achievable if there exists a sequence of codes @xmath55 whose average probability of error goes to zero as @xmath56 .
the capacity region @xmath57 is defined as the set of all achievable rate pairs @xmath54 , and the dof region can be defined from the capacity region as @xmath58\big\}. \nonumber\end{aligned}\ ] ] the element @xmath23 of the received vector @xmath59 at time index @xmath17 is denoted as @xmath60,t}$ ] . in a similar manner ,
a subset of elements from this vector is denoted as follows : @xmath61,t}\triangleq\{y_{i[\ell_1],t},y_{i[\ell_1 + 1],t},\cdots , y_{i[\ell_2],t}\}\ ] ] in the same manner , we define a sequence of vectors over all ( causal ) time that select only a subset of the antennas : @xmath61}^t\triangleq\{y_{i[\ell_1]}^t , y_{i[\ell_1 + 1]}^t,\cdots , y_{i[\ell_2]}^t\}\ ] ] in the special case where only one antenna is selected across time we have @xmath60}^t=\{y_{i[\ell],1},y_{i[\ell],2},\cdots , y_{i[\ell],t}\}$ ] .
@xmath62 denotes that functions @xmath63 have the following tail characteristic : @xmath64 .
several specialized notations are shown in table [ table_notation ] that distinguish the dof regions under various conditions .
.dof notations for interference channel with cognitive relay [ cols="^,<",options="header " , ] table [ table2 ] reproduces from @xcite the sum - dof of the mimo interference channel , with @xmath65 transmit and @xmath11 receive antennas at respective nodes .
if the two transmitters partially cooperate , the channel becomes equivalent to the iccr with antenna configuration of @xmath66 .
therefore , the dof region of the interference channel with delayed csit is included by that of the iccr with delayed csit . if @xmath67 or @xmath68 , the sum dof is the same for both channels . if @xmath69 , however , the sum dof of the iccr is greater than that of the interference channel because the cognitive relay effectively produces partial cooperation between transmitters .
here we consider another extension to the cognitive interference channel consisting of one non - cognitive transmitter , one cognitive transmitter , and their intended receivers .
the cognitive transmitter has both messages intended for the two receivers as shown in fig .
[ cognitiveic ] .
the cognitive interference channel with perfect csit and csir has been studied in @xcite .
the inner and outer bounds of capacity region of the siso cognitive interference channel with perfect csit were given in @xcite .
for mimo cognitive interference channel , @xcite calculated the capacity region within a constant gap . in @xcite ,
the dof region of the cognitive interference channel with perfect csit was derived .
however , the dof region of the cognitive interference channel with delayed feedback has been unknown .
the dof of the iccr from the previous section can be used for a lower and an upper bound of the cognitive interference channel when feedback is delayed .
[ corollary : cognitiveic_lower ] the dof region of the cognitive interference channel with antenna configuration @xmath70 is lower bounded by that of the iccr with antenna configuration @xmath71 .
if cooperation between the cognitive relay and one transmitter is allowed in the iccr , the channel becomes equivalent to the cognitive interference channel where the non - cognitive transmitter has @xmath9 antennas , the cognitive transmitter has @xmath72 antennas , and each receiver has @xmath11 antennas .
therefore , the dof region of the cognitive interference channel with antenna configuration @xmath70 is lower bounded by that of the iccr with antenna configuration @xmath71 .
[ corollary : cognitiveic_upper ] the dof region of the cognitive interference channel with antenna configuration @xmath73 @xmath74 is upper bounded by the dof region of the iccr with antenna configuration @xmath75 . if only one transmitter exists in the iccr , the antenna configuration for this scenario is @xmath76 and it corresponds to the cognitive interference channel where the non - cognitive transmitter has @xmath9 antennas and the cognitive transmitter has @xmath10 antennas while each receiver has @xmath11 antennas .
hence , the upper bound of the dof region of the cognitive interference channel with antenna configuration of @xmath73 @xmath74 is that of the iccr with antenna configuration of @xmath77 . and @xmath78 .
] in this example the non - cognitive transmitter has two antennas , the cognitive transmitter has three antennas , and receivers have two antennas each .
the dof region of the @xmath79 iccr can serve as a lower bound , and the dof of @xmath78 iccr serves as upper bound .
the lower bound derived in section [ mimo_cr ] is @xmath80 where the maximum sum dof is @xmath81 .
the upper bound is obtained from the results of section [ mimo_cr ] as @xmath82 where the maximum sum dof is @xmath83 .
these lower and upper bounds are shown in fig .
[ dof_cognitive ] .
this paper studies the dof region of the two - user gaussian fading interference channel with cognitive relay ( iccr ) with delayed feedback .
three different types of delayed feedback are considered : delayed csit , delayed output feedback , and delayed shannon feedback . for the siso iccr
, the proposed retrospective interference alignment scheme using delayed feedback information achieves the dof region .
the sum dof of the siso iccr is @xmath1 with delayed feedback information , compared to the dof of 1 for siso interference channel in the absence of relay , regardless of csit . without feedback ,
the cognitive relay is not useful in the sense of dof . in the mimo case
, the optimal dof has been characterized under all antenna configurations if delayed feedback is provided to both the transmitters and cognitive relay .
dof benefits can be obtained over and above the open - loop system when @xmath84 or @xmath85 , and the sum dof is @xmath86 when @xmath87 and @xmath88 when @xmath89 . on the other hand , if @xmath90 or @xmath91 , then delayed feedback does not help in the sense of dof . in this scenario , the sum - dof ( for both open - loop and closed - loop ) is @xmath92 when @xmath93 and @xmath11 when @xmath91 .
if delayed feedback is unavailable at the cognitive relay , the proposed retrospective interference alignment scheme achieves the optimal dof except when @xmath12 , where existing upper and lower bounds do not meet .
delayed feedback is shown to extend the dof over and above the open - loop system when @xmath94 and @xmath95 .
in addition , in this paper upper and lower bounds are derived for the dof region of the two - user mimo cognitive interference channel under delayed feedback .
@xmath96,t}|y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) & = \frac{1}{c_1}\sum_{\ell=1}^{c_1}h(y_{\a[\ell],t}|y_{\a[1:\ell-1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \stackrel{(\mathrm{a})}{\geq } \frac{1}{c_1}\sum_{\ell=1}^{c_1}h(y_{\a[\ell],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\label{a1}\end{aligned}\ ] ] where ( a ) holds because conditioning reduces entropy . since @xmath97 if only delayed csit is available , we can write @xmath98,t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) & = h(y_{\a[\ell_1],t}|x_{\a , t } , y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) \nonumber\\ & = h(\widetilde{y}_{\a[\ell_1],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\label{entropy_y_1i}\\ h(y_{\b[\ell_2],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) & = h(y_{\b[\ell_2],t}|x_{\a , t } , y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) \nonumber\\ & = h(\widetilde{y}_{\b[\ell_2],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\label{entropy_y_2i } \end{aligned}\ ] ] where @xmath99,t}$ ] is defined as the output signal if we assume @xmath100 , i.e. , in the absence of @xmath101 .
we note that @xmath102,t}$ ] and @xmath103,t}$ ] have identical distributions since the respective channel gains and receiver noises have identical distributions .
thus , for all @xmath104 and @xmath105 , @xmath106,t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)=h(\widetilde{y}_{2[\ell_2],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) \label{equal1}\end{aligned}\ ] ] therefore , @xmath107,t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)= h(y_{\b[\ell_2],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n).\label{equal2}\end{aligned}\ ] ] this represents the statistical equivalence of channel outputs @xcite .
the result ( [ equal2 ] ) is also applicable even if @xmath97 is replaced by @xmath108 or @xmath109 for the delayed feedback or the shannon feedback information , since @xmath101 when either delayed feedback or shannon feedback information is available can be constructed from the given conditions for the delayed csit case .
thus , we can rewrite ( [ a1 ] ) as @xmath110,t}|y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)&\geq \frac{1}{c_1}\sum_{\ell=1}^{c_1}h(y_{\a[\ell],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) \nonumber\\ & \stackrel{(\mathrm{b})}{= } h(y_{\b[1],t}|y_{\a[1:c_1 - 1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n ) \nonumber\\ & \stackrel{(\mathrm{c})}{\geq } h(y_{\b[1],t}|y_{\a[1:c_1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \stackrel{(\mathrm{d})}{=}\frac{1}{c_2-c_1}\sum_{\ell=1}^{c_2-c_1}h(y_{\b[\ell],t}|y_{\a[1:c_1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \stackrel{(\mathrm{e})}{\geq } \frac{1}{c_2-c_1}\sum_{\ell=1}^{c_2-c_1}h(y_{\b[\ell],t}|y_{\b[1:\ell-1],t},y_{\a[1:c_1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & = \frac{1}{c_2-c_1}h(y_{\b[1:c_2-c_1],t}|y_{\a[1:c_1],t},y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\label{a2}\end{aligned}\ ] ] where ( b ) and ( d ) follow from the statistical equivalence of channel outputs , and ( c ) and ( e ) follow from the fact that conditioning reduces entropy .
thus , we have @xmath111,t}|y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)&\geq c_1h(y_{\a[1:c_1],t},y_{\b[1:c_2-c_1],t}|y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \stackrel{(\mathrm{f})}{\geq } c_1 h(y_{\a[1:c_1],t},y_{\b[1:c_2-c_1],t}|y_{\a[1:m_r]}^{t-1},y_{\b[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\end{aligned}\ ] ] where ( f ) follows from the fact that conditioning reduces entropy .
we use lemma [ lemma_mimo1 ] , as follows : @xmath112}^n|w_{\a},\mathcal{h}^n)&\geq \frac{1}{c_1}h(y_{\a[1:c_1]}^n|w_{\a},\mathcal{h}^n)\nonumber\\ & = \frac{1}{c_1 } \sum_{t=1}^{n}h(y_{\a[1:c_1],t}|y_{\a[1:c_1]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \stackrel{(\mathrm{g})}{\geq}\frac{1}{c_1 } \sum_{t=1}^{n}h(y_{\a[1:c_1],t}|y_{\a[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \stackrel{(\mathrm{h})}{\geq } \frac{1}{c_2 } \sum_{t=1}^{n}h(y_{\a[1:c_1],t},y_{\b[1:c_2-c_1],t}|y_{\a[1:m_r]}^{t-1},y_{\b[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & = \frac{1}{c_2}\sum_{t=1}^{n}\!\big[h(y_{\a[1:m_r],t},y_{\b[1:m_r],t}|y_{\a[1:m_r]}^{t-1},y_{\b[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & \ \ \ \ -h(y_{\a[c_1
+ 1:m_r],t},y_{\b[c_2-c_1 + 1:m_r],t}|y_{\a[1:c_1],t},y_{\b[1:c_2-c_1],t},y_{\a[1:m_r]}^{t-1},y_{\b[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\big]\nonumber\\ & \stackrel{(\mathrm{i})}{\geq } \frac{1}{c_2}h(y_{\a[1:m_r]}^n , y_{\b[1:m_r]}^n|w_{\a},\mathcal{h}^n)+n\cdot o(\mathrm{log}_2p)\end{aligned}\ ] ] where @xmath113 , @xmath114 , ( g ) follows from the fact that conditioning reduces entropy , and ( h ) follows from lemma [ lemma_mimo1 ] , ( i ) follows from the fact that for all @xmath115 , @xmath116,t},y_{\b[c_2-c_1 + 1:m_r],t}|y_{\a[1:c_1],t},y_{\b[1:c_2-c_1],t},y_{\a[1:m_r]}^{t-1},y_{\b[1:m_r]}^{t-1},w_{\a},\mathcal{h}^n)\nonumber\\ & = h(y_{\a[c_1 + 1:m_r],t},y_{\b[c_2-c_1 + 1:m_r],t}|y_{\a[1:c_1],t},y_{\b[1:c_2-c_1],t},y_{\a[1:m_r]}^{t-1},y_{\b[1:m_r]}^{t-1},x_{\a}^n , w_{\a},\mathcal{h}^n)\nonumber\\ & \leq o(\mathrm{log}_2p)\label{o_log}\end{aligned}\ ] ] because @xmath117,t}$ ] and @xmath118,t}$ ] do not affect the dof when the channel inputs @xmath119 and the channel outputs @xmath120,t}$ ] and @xmath121}$ ] are given . therefore , we have @xmath112}^n|w_{\a},\mathcal{h}^n)\geq \frac{1}{c_2}h(y_{\a[1:m_r]}^n , y_{\b[1:m_r]}^n|w_{\a},\mathcal{h}^n)+n\cdot o(\mathrm{log}_2p ) .
\nonumber\end{aligned}\ ] ] s. sridharan , s. vishwanath , s. a. jafar , and s. shamai , `` on the capacity of cognitive relay assisted gaussian interference channel , '' in _ proc .
toronto , ontario , canada , jul .
2008 .
s. rini , d. tuninetti , n. devroye , and a. goldsmith , `` the capacity of the interference channel with a cognitive relay in strong interference . '' in _ proc .
info . theory _ , st .
petersburg , russia , jul .
2011 .
h. charmchi , g. a. hodtani , and m. nasiri - kenari , `` a new outer bound for a class of interference channels with a cognitive relay and a certain capacity result , '' _ ieee communications letters _ ,
2 , pp . 241 - 244 , feb .
a. dytso , s. rini , n. devroye , and d. tuninetti , `` on the capacity region of the two - user interference channel with a cognitive relay , '' _ ieee trans .
wireless commun .
_ , vol . 13 , no . 12 , pp . 6824 - 6838 , dec .
2014 .
m. j. abdoli , a. ghasemi , and a. k. khandani , `` on the degrees of freedom of three - user mimo broadcast channel with delayed csit , '' in _ proc .
info . theory _ , st .
petersburg , russia , jul .
2011 . c. s. vaze and m. k. varanasi , `` the degrees of freedom region and interference alignment for the mimo interference channel with delayed csit , '' _ ieee trans .
info . theory _
58 , no . 7 , pp . 4396 - 4417 , jul .
r. tandon , s. mohajer , h. v. poor , and s.shamai , `` degrees of freedom region of the mimo interference channel with output feedback and delayed csit , '' _ ieee trans .
info . theory _ ,
1444 - 1457 , mar .
2013 .
k. mohanty , c. s. vaze , and m. k. varanasi , `` the degrees of freedom region for the mimo interference channel with hybrid csit , '' _ ieee trans .
info . theory _ , to be publised .
c. huang , s. a. jafar , s. shamai , and s. vishwanath , `` on degrees of freedom region of mimo networks without channel state information at transmitters , '' _ ieee trans .
info . theory _ , vol .
2 , pp . 849 - 857 , feb . 2012 . c. s. vaze and m. k. varanasi , `` the degree - of - freedom regions of mimo broadcast , interference , and cognitive radio channels with no csit , '' _ ieee trans .
info . theory _
58 , no . 8 , pp . 5354 - 5374 , aug . 2012 .
n. devroye , p. mitran and v. tarokh , achievable rates in conitive radio channels , " _ ieee trans .
info . theory _
1813 - 1827 , may . 2006 .
i. maric , r. d. yates and g. kramer , capacity of interference channels with partial transmitter cooperation , " _ ieee trans .
info . theory _ ,
3536 - 3548 , oct . | this paper studies the interference channel with a cognitive relay ( iccr ) under delayed feedback .
three types of delayed feedback are studied : delayed channel state information at the transmitter ( csit ) , delayed output feedback , and delayed shannon feedback .
outer bounds are derived for the dof region of the two - user multiple - input multiple - output ( mimo ) iccr with delayed feedback as well as without feedback . for the single - input single - output ( siso ) scenario ,
optimal schemes are proposed based on retrospective interference alignment .
it is shown that while a cognitive relay _ without _ feedback can not extend the sum - dof beyond @xmath0 in the two - user siso interference channel , delayed feedback in the same scenario can extend the sum - dof to @xmath1 .
for the mimo case , achievable schemes are obtained via extensions of retrospective interference alignment , leading to dof regions that meet the respective upper bounds . |
this work was partially supported by the massey postdoctoral foundation , new zealand .
99 p. amaudraz et al . , ; m. arneodo et al . , ; m. arneodo et al . , .
a. baldit et al . , na51 collaboration , .
e. a. hawker et al . ,
e866/nusea collaboration , .
k. ackerstaff et al . , hermes collaboration , .
w. g. seligman et al .
, ccfr - collaboration , ; a. o. bazarko et al .
, ccfr - collaboration , ; s. a. rabinowitz et al . , ccfr - collaboration , .
m. arneodo et al . , nmc - collaboration , .
a. w. thomas , .
h. holtmann , a. szczurek , and j. speth , .
j. speth and a. w. thomos , .
s. kumano , . c. boros , j. t. londergan , and a. w. thomas , .
s. j. brodsky and b. q. ma , .
f. g. cao and a. i. signal , .
m. alberg , e. m. henley , x. ji , and a. w. thomas , .
m. alberg , t. falter , and e. m. henley , . c. boros and a. w. thomas , .
b. q. ma , i. schmidt , and j. j. yang , hep - ph/9906424 ( 1999 ) .
j. schechter , a. subbaramann , and h. weigel , .
t. feldmann , p. kroll , and b. stech , .
f. g. cao and a. i. signal , .
j. f. donoghue , b. r. holstein , and y. c. lin , .
l. burakovsky and t. goldman , .
m. glck , e. reya , and m. stratmann , . m. glck , e. reya , and a. vogt , . m. miyama and s. kumano , . f. g. cao and a. i. signal , in preparation .
parameters in eq .
( [ qfit ] ) at @xmath63 gev@xmath57 . 0.5 cm [ cols="^,^,^,^,^,^",options="header " , ]
fig . 1 . @xmath72 $ ] with @xmath49 gev .
the dashed , dotted , and dashed - dotted curves are the contributions from @xmath105 , @xmath106 and @xmath107 components respectively . the solid curve is the sum of above three contributions .
the thinner and thicker curves correspond to the scales @xmath81 gev@xmath57 and @xmath63 gev@xmath57 respectively . | the sea quark content of the @xmath0 baryon is investigated using light - cone baryon - meson fluctuation model suggested by brodsky and ma .
it is found that the @xmath0 sea is flavour asymmetric ( @xmath1 ) and quark - antiquark asymmetric ( @xmath2 ) .
our prediction for the flavour asymmetry , @xmath1 , is significantly different from the @xmath3 prediction ( @xmath4 ) , while our prediction for the @xmath5-@xmath6 asymmetry is consistent with the @xmath3 prediction .
pacs numbers : 11.30.hv ; 14.20.-c keywords : flavour asymmetry ; baryon sea ; baryon - meson fluctuation ; @xmath3 16.5 cm 23.5 cm -1.0 cm -0.2 cm -0.2 cm # 1#2#3#4#1 # 2 ( # 4 ) # 3 * the flavour asymmetry and quark - antiquark asymmetry in the @xmath0-sea * institute of fundamental sciences , massey university , private bag 11 222 , palmerston north , new zealand the nucleon sea exhibits two interesting properties : flavour asymmetry @xcite and quark - antiquark asymmetry @xcite . while there have been many studies of the nucleon sea from both experiment ( see e. g. @xcite ) and theory ( see e. g. @xcite and references therein ) , the studies of the sea distributions of the other baryons in the baryon octet predicted by the @xmath3 quark model are very few .
it is of interest to know whether the sea of the other members of the baryon octet has the same properties ( flavour asymmetry and quark - antiquark asymmetry ) as the nucleon sea .
also , through the study of the quark sea of the other members of the baryon octet , we can improve our understanding of the structure of the baryons and the non - perturbative properties of qcd .
alberg _
et al . _
@xcite pointed out that the valence and sea quark distributions of the @xmath7 may exhibit large deviations from the @xmath3 predictions , and these parton distributions could be obtained from drell - yan experiments using charged hyperon beams on proton and deuteron targets .
alberg , falter , and henley @xcite studied the flavour asymmetry in the @xmath0 sea employing the meson cloud model and effective lagrangian for the baryon - meson - baryon interaction , and found large deviations from @xmath3 .
boros and thomas @xcite calculated the quark distributions of @xmath8 and @xmath7 employing the mit bag model .
it was found that the valence quark distributions are quite different from the @xmath3 predictions and that the quark sea is flavour asymmetric .
more recently , ma , schmidt and yang @xcite showed that there are significant differences between the predictions of perturbative qcd and @xmath9 quark - diquark model for the flavor and spin structure of the @xmath8 baryon s quark distributions near @xmath10 . in this letter we shall investigate the flavour asymmetry and quark - antiquark asymmetry of the @xmath0 sea using the light - cone baryon - meson fluctuation model ( lcm ) suggested by brodsky and ma @xcite . the baryon - meson fluctuation
( meson cloud ) mechanism is very successful in understanding on the flavour asymmetry and quark - antiquark asymmetry of nucleon sea .
the various fluctuations can be described via corresponding baryon - meson - nucleon lagrangians @xcite . recently , brodsky and ma @xcite proposed that the baryon - meson fluctuation could be described by using a light - cone two body wave function which is a function of the invariant mass squared of the baryon - meson fock state .
compared to the commonly used effective lagrangian method ( elm ) @xcite for the description of baryon - meson fluctuations , the lcm is relatively simple .
furthermore our study @xcite showed that the lcm can produce very similar results to the effective lagrangian method for a suitable choice of parameter .
the basic idea of the meson cloud model ( for recent reviews see refs .
@xcite ) is that the nucleon can be viewed as a bare nucleon surrounded by a mesonic cloud .
the nucleon wave function can be expressed in terms of bare nucleon and virtual baryon - meson fock states .
although this model was developed mainly in the study of nucleon sea , applying this model to the other baryons is straightforward . for the @xmath0
, the wave function can be written as |^+_physical = z |^+_bare + _ b m dy d^2 * k * _ _ b m(y , k_^2 ) |b(y , * k * _ ) ; m(1-y,-*k * _ ) , [ nmcm ] where @xmath11 is the wave function renormalization constant , @xmath12 is the wave function of fock state containing a baryon ( @xmath13 , @xmath14 , @xmath0 , @xmath15 ) with longitudinal momentum fraction @xmath16 , transverse momentum @xmath17 , and a meson ( @xmath18 , @xmath19 , @xmath20 ) with momentum fraction @xmath21 , transverse momentum @xmath22 . here
we consider the most energetically - favoured fluctuations in the baryon octet and meson octet .
the fluctuation @xmath23 is neglected due to the higher mass of @xmath24 ( @xmath25 gev while @xmath26 gev , @xmath27 gev ) .
it would seem that the fluctuation @xmath28 is also important in the calculations of @xmath29 and @xmath30 .
however ,
applying the common @xmath31-@xmath32 mixing scheme = cos ( u + d -2 s ) -sin ( u + d + s ) and assuming @xmath3 symmetry for the quark distributions in the @xmath31 and @xmath32 , we find that compared to the fluctuation @xmath33 , the contributions to the @xmath34 and @xmath35 from the fluctuation @xmath28 are suppressed by a factor of @xmath36 which is in the range of @xmath37 for mixing angle in the theoretically accepted range @xmath38 @xcite . the higher mass of the @xmath39 ( @xmath40 gev , @xmath41 gev ) also suppresses the contribution from this fluctuation .
thus we neglect this fluctuation in our calculation .
provided that the lifetime of a virtual baryon - meson fock state is much longer than the strong interaction time in the drell - yan process , the contribution from the virtual baryon - meson fock states to the quark and anti - quark sea of @xmath0 can be written as convolutions q(x)&=&_bm , [ qbm ] + ( x)&=&_bm ^1_x f_mb(1-y ) ^m ( ) , [ qbarbm ] where @xmath42 is fluctuation function which gives the probability for the @xmath0 to fluctuate into a virtual @xmath43 state f_bm(y)=^_0 d k_^2 | _ b m(y , k_^2)|^2 .
[ fbm ] a common practice in the evaluation of the wave function @xmath44 is to employ time - ordered perturbative theory in the infinite momentum frame and the effective meson - baryon - nucleon interaction lagrangian @xcite . on the other hand , brodsky and ma
@xcite suggested that this wave function can also be described by using light - cone two - body wave function which is a function of the the invariant mass squared of the baryon - meson fock state _ bm(y,*k*_)=a exp , [ phi ] where @xmath45 is a phenomenological parameter which determines the shape of the fluctuation function .
compared to the effective lagrangian method , eq .
( [ phi ] ) is quite simple .
furthermore , our study on the @xmath46-@xmath47 asymmetry in the nucleon sea @xcite showed that eq .
( [ phi ] ) can provide similar results to the effective lagrangian method for @xmath48 gev . because the spin structure of the baryon - meson - baryon vertex is the same for all members of the baryon octet ( ignoring fluctuations to decuplet baryons ) , we might expect that the value of @xmath45 should be similar for all the members of the baryon octet
. we will use @xmath49 gev and @xmath50 gev in our calculation as there is little constraint from experimental data or theoretical studies on the @xmath51 sea .
the normalization @xmath52 in eq .
( [ phi ] ) can be determined by the probability for the corresponding fluctuation .
we adopt the result given in ref .
@xcite for the probabilities of the various fluctuations and the fluctuation functions given in ref .
@xcite for a given type of fluctuation are defined as the sum of all iospin states : @xmath53 . ] : p_^+=3.2% , p_p |k^0=0.4% , + p_^0 ^+=p_^+ ^0=p_=1.85% .
[ p ] in the baryon - meson fluctuation model , the non - perturbative contributions to the quark and the anti - quark distributions in the @xmath0 sea come from the quarks and anti - quarks of the baryons ( @xmath54 and @xmath15 ) and mesons ( @xmath55 and @xmath20 ) in the virtual baryon - meson fock states .
so we need the parton distributions of the involved baryons and mesons as input . for the parton distribution in the pion , we employ the parameterization given by glck , reya , and stratmann ( grs98 ) @xcite and we neglect the sea content in the meson , that is , ^^+=u^^+&=&^^-=d^^-= v^ , + ^^0=u^^0&=&^^0=d^^0= v^ , + v^(x,_nlo^2 ) & = & 1.052 x^-0.495 ( 1 + 0.357 ) ( 1-x)^0.365 , [ vpion ] at scale @xmath56 gev@xmath57 . for the @xmath6 distribution in the @xmath20 we relate it to the @xmath58 distribution in the @xmath59 which are given in the grs98 parameterization @xcite also ^|k^0(x,_nlo^2)=u^k^+(x,_nlo^2 )
= 0.540 ( 1-x)^0.17 v^(x,_nlo^2 ) , [ vk0bar ] at scale @xmath56 gev@xmath57 .
in order to investigate the quark - antiquark asymmetry via @xmath60 in the @xmath0 sea , we also need to know the @xmath5-quark distribution in the @xmath8 , @xmath0 and @xmath15 , for which we use the parameterization for the @xmath5 quark distribution in the proton given by glck , reya , and vogt ( grv98 ) @xcite , d^p(x,_nlo^2)=0.400 x^-0.57 ( 1-x)^4.09 ( 1 + 18.2 x ) , at scale @xmath61 gev@xmath57 .
we evolve the distributions to the scale @xmath62 gev@xmath57 using the program of miyama and kumano @xcite in which the evolution equation is solved numerically in a brute - force method .
we found that at @xmath63 gev@xmath57 all parton distributions @xmath64 , @xmath65 and @xmath66 can be parameterized using the following form q(x , q^2)=a x^b ( 1-x)^c ( 1+d + e
x ) [ qfit ] with the parameters given in table 1 .
we estimate the uncertainty in solving the evolution equations numerically and parameterizating the parton distribution in the form of eq .
( [ qfit ] ) to be about @xmath67 in the @xmath68-region which we are interested in _ i.e. _ @xmath69 .
the effect of evolution from a lower scale to a higher scale is to make the parton distribution more concentrated in the small @xmath68 region .
thus we may expect that the @xmath68 position at which an asymmetry exhibits a maximum will move to smaller @xmath68 as we evolve to higher values of @xmath70 .
however , we do not expect the asymmetry to evolve away " at a higher @xmath70 scale if it exists at a lower scale such as @xmath71 .
we investigate the flavour asymmetry in the @xmath0 sea through calculating the differences between the antiquark distributions : @xmath72 $ ] , @xmath73 $ ] and @xmath74 $ ] which are given by x&= & x + & = & ^1_x dy , [ xdmu ] x&= & x + & = & ^1_x dy , [ xdms ] x&= & x _
^+ ^0(x ) + & = & ^1_x dy f_^+ ^0(1-y ) ^^0 ( ) .
[ xums ] @xmath75 $ ] comes from only fluctuation @xmath76 , while @xmath77 $ ] and @xmath78 $ ] come from also @xmath33 , @xmath79 as well as @xmath80 . in fig . 1 we present our results for @xmath77 $ ] at the scales @xmath81 gev@xmath57 and @xmath63 gev@xmath57 with @xmath49 gev .
it can be found that the contribution from the fluctuation @xmath33 is about twice as large as that from @xmath79 , and both are much larger than the contribution from @xmath80 . under evolution
the distributions move from larger @xmath68 to smaller @xmath68 the @xmath68 position at which @xmath72 $ ] exhibits a maximum shifts from about @xmath82 to @xmath83 and the maximum decreases about @xmath84 , which coincides with our naive expection .
the numerical results for @xmath72 $ ] , @xmath73 $ ] and @xmath74 $ ] at @xmath63 gev@xmath57 are given in figs . 2 and 3 for @xmath49 gev and @xmath50 gev respectively .
we can see that @xmath85 , that is the anti - quark distribution in the @xmath0 sea is flavour asymmetric . as is well known , the nucleon sea is also asymmetric and for the proton sea @xmath86 @xcite .
the main difference between the proton@xmath87 and @xmath88 is the replacement of a valance @xmath5 quark by a valance @xmath46 quark .
thus one may expect from @xmath3 symmetry that the @xmath88 sea to be @xmath89 .
this prediction is opposite to our above conclusion ( @xmath86 ) from the light - cone baryon - meson fluctuation model .
if the @xmath3 symmetry breaking in the @xmath90 , @xmath6 and @xmath47 distributions in the @xmath0 sea has the same source as that for the @xmath58 , @xmath5 , and @xmath46 quark masses , we may expect that @xmath72 < x[\ubar(x)-\sbar(x)]$ ] since the mass difference between the @xmath58 and @xmath5 quarks is far smaller than that between the @xmath58 and @xmath46 quarks .
however , our calculations ( see figs . 2 and 3 ) show that @xmath73 > x[\dbar(x)-\ubar(x ) ] > x[\ubar(x)-\sbar(x)]$ ] .
the relation @xmath72
> x[\ubar(x)-\sbar(x)]$ ] is opposite to our above argument , which implies that the dynamics responsible for the @xmath3 symmetry breaking in the quark distributions of the @xmath0 sea , as calculated in our model , are different from that responsible for the mass differences among the @xmath58 , @xmath5 and @xmath46 quarks .
another interesting question concerning the @xmath0 sea is the quark - antiquark asymmetry .
although the perturbative sea created from gluon - splitting is symmetric @xmath91 ( in the leading twist approximation in perturbative calculation ) , the non - perturbative sea , which may exist over a long time and has a strong connection with the bare " @xmath0 , may be asymmetric @xmath92 .
because of the existence of valance @xmath58 and @xmath46 quarks in the @xmath0 , it is difficult to measure the differences @xmath93 and @xmath94 in the @xmath0 sea .
the most likely experiment is to measure the difference between @xmath5 and @xmath6 . from the baryon - meson fluctuation model
the @xmath60 turns out to be : d(x)-(x)&= & d_^+(x)-_^+(x ) + d_^0 ^+(x)-_^0 ^+(x ) + d_p |k^0(x)-_p & = & ^1_x \ { d^p ( ) .
+ & & - ^^+ ( ) + & & . - f_p |k^0(1-y ) ^|k^0 ( ) } .
[ dmd ] the numerical results at scales @xmath71 and @xmath63 gev@xmath57 are presented in fig . 4 . once again one can find that evolution
pushes " the distributions to the small @xmath68 region .
it can be seen that @xmath95 in the @xmath0 sea
. however , the prediction for the behavior of @xmath60 depends strongly on the value of @xmath45 for @xmath49 gev @xmath96 in the smaller @xmath68 region and @xmath97 in the larger @xmath68 region , while for the @xmath48 gev @xmath97 in the smaller @xmath68 region and @xmath98 in the larger @xmath68 region .
this result is similar to our earlier finding on the @xmath99 in the nucleon sea @xcite employing the same light - cone baryon - meson fluctuation model , which suggests that @xmath3 symmetry in the sea holds in this case .
we turn to the discussion about @xmath45-dependence in our calculation . comparing figs
. 2 and 3 one can find that for the @xmath72 $ ] , @xmath73 $ ] and @xmath74 $ ] the shape and maximum of asymmetries are very similar for different @xmath45 , while the @xmath68 position at which the asymmetries exhibit maxima shifts slightly . the calculations with @xmath49 gev peak at about @xmath100 while the calculations with @xmath48 gev peak at about @xmath101 .
thus the calculations for the flavour asymmetry are not very sensitive to the value of @xmath45 , and @xmath68 being about @xmath102 is a good region to study the flavor asymmetry in the @xmath0 sea .
this observation is consistent with the prediction given in ref .
@xcite the region @xmath103 should be a good one to measure the flavour asymmetry in the @xmath51 sea .
the calculation for the @xmath60 ( see fig . 4 ) is much more sensitive to the value of @xmath45 than that for the flavour asymmetry the calculations with @xmath49 gev and @xmath50 gev even give opposite predictions for the @xmath68-dependence of @xmath60 .
we may expect that further calculation @xcite on the nucleon sea employing the light - cone baryon - meson fluctuation model may provide useful constraints on the value of @xmath45 , and thereby give more definite predictions on the sea quark content in the @xmath0 baryon . in summary
, besides the nucleon sea the studies on the sea quark content of the other members of the baryon octet are interesting and important since it is helpful to our understanding of both the structure of the octet baryons and non - perturbative qcd effects such as @xmath3 symmetry breaking and flavour asymmetry .
we calculated the sea quark content of the @xmath0 baryon employing the light - cone baryon - meson fluctuation model .
it was found that the @xmath0 sea is flavour asymmetric ( @xmath104 ) and quark - antiquark asymmetric ( @xmath2 ) .
our prediction for the flavor asymmetry , @xmath1 , is significantly different from the @xmath3 prediction ( @xmath4 ) , while our prediction for the @xmath5-@xmath6 asymmetry is consistent with the @xmath3 prediction . |
graphs considered in the paper are finite , without loops or multiple edges . in a graph @xmath13 ,
@xmath14 and @xmath15 ( @xmath16 and @xmath17 ) denote the vertex set and the edge set of @xmath7 , respectively . for undefined concepts and notation
we refer the reader to @xcite . for two vertices @xmath18 and @xmath19 in a graph @xmath7 , a @xmath20-geodesic is a shortest path between @xmath18 and @xmath19 .
a set @xmath5 of vertices of @xmath7 is convex if the vertices of every @xmath20-geodesic is contained in @xmath5 for every @xmath21 . according to duchet ,
convexity in graphs has been studied since the early seventies , when abstract convexity was studied in different contexts ( @xcite is an outdated , but very nice , survey on the subject ) .
convexity in graphs has taken many different directions , and different related parameters have been defined and widely studied , e.g. , the hull number @xcite , the geodetic number @xcite , and the convexity number @xcite of a graph .
recent papers on this subjects include @xcite , where the decision problem associated with these three parameters are shown to be @xmath10-complete , even when restricted to bipartite graphs , and in the case of the geodetic number , even when restricted to bipartite chordal graphs .
chartrand , fink and zhang generalized the concept of convexity to oriented graphs , and defined the convexity number for an oriented graph ; oriented analogues of the hull number and geodetic number are defined in @xcite .
we focus on the convexity number of oriented graphs ; although this generalization was introduced in 2002 , and the proof given by gimbel in @xcite on the @xmath10-completeness of determining the convexity number of an arbitrary graph is one of the shortest and neatest @xmath10-completeness proofs ever done , the problem of determining the convexity number of an oriented graph was not known to be @xmath10-complete until now .
we prove that determining the convexity number of an oriented graph is @xmath10-complete even when restricted to bipartite graphs of girth @xmath22 , with @xmath23 . an _ oriented graph _ is an orientation of some graph . in an oriented graph @xmath24 ,
@xmath14 and @xmath15 ( @xmath25 and @xmath26 ) denote the vertex set and the edge set of @xmath0 , respectively .
an _ oriented subgraph _
@xmath27 of an oriented graph @xmath24 is an oriented graph with @xmath28 and @xmath29 .
an oriented graph is _ connected _ if its underlying graph is connected . a _ directed path _ is a sequence @xmath30 of vertices of an oriented graph @xmath0 such that @xmath31 are distinct and @xmath32 for @xmath33 .
an oriented graph is _ strongly connected _ ( or _ strong _ ) if for every pair of distinct vertices @xmath18 and @xmath19 , there exists a directed path from @xmath18 to @xmath19 .
the _ girth _ of an oriented graph is the length of a shortest directed cycle .
a @xmath20-_geodesic _ in a digraph @xmath0 is a shortest @xmath20-directed path and its length is @xmath34 .
a nonempty subset , @xmath5 , of the vertex set of a digraph , @xmath0 , is called a _
convex set _ of @xmath0 if , for every @xmath21 , every vertex lying on a @xmath20- or @xmath35-geodesic belongs to @xmath5 . for a nonempty subset , @xmath36 , of @xmath25 , the _ convex hull
_ , @xmath37 $ ] , is the minimal convex set containing @xmath36 . thus @xmath38
= s$ ] if and only if @xmath5 is convex in @xmath0 .
the _ convexity number _ , @xmath6 , of a digraph @xmath0 is the maximum cardinality of a proper convex set of @xmath0 .
a _ maximum convex set _ @xmath5 , of a digraph @xmath0 , is a convex set with cardinality @xmath6 . since every singleton vertex set is convex in a connected oriented graph @xmath0 , @xmath39 .
the _ degree _ , @xmath40 , of a vertex @xmath19 in an oriented graph is the sum of its in - degree and out - degree ; this is , @xmath41 .
a vertex , @xmath19 , is an _ end - vertex _ if @xmath42 .
a _ source _ is a vertex having positive out - degree and in - degree @xmath43 , while a _ sink _ is a vertex having positive in - degree and out - degree @xmath43 . for a vertex @xmath19 of @xmath0 ,
the in - neighborhood of @xmath19 , @xmath44 , is the set @xmath45 and the out - neighborhood of @xmath19 , @xmath46 , is the set @xmath47 .
a vertex @xmath19 of @xmath0 is a _ transitive vertex _ if @xmath48 , @xmath49 and , for every @xmath50 and @xmath51 , @xmath52 . for graphs @xmath7 and @xmath53 , their _ cartesian product _ , @xmath54 , is the graph with vertex set @xmath55 , and such that two vertices @xmath56 and @xmath57 are adjacent in @xmath54 if either @xmath58 and @xmath59 is an edge in @xmath53 , or @xmath60 and @xmath61 is an edge in @xmath7 . for a vertex @xmath22 of @xmath7 , the subgraph of @xmath54 induced by the set @xmath62
is called an @xmath53-fiber and is denoted by @xmath63 . similarly , for @xmath64 ,
the @xmath7-fiber , @xmath65 , is the subgraph induced by @xmath66 .
we will have occasion to use the fiber notation @xmath65 and @xmath63 to refer instead to the set of vertices in these subgraphs ; the meaning will be clear from the context .
it is clear that all @xmath7-fibers are isomorphic to @xmath7 and all @xmath53-fibers are isomorphic to @xmath53 .
as mentioned , the concept of convexity number of an oriented graph was first introduced by chartrand , fink and zhang in @xcite , where they proved the following pair of theorems .
let @xmath0 be a connected oriented graph of order @xmath67 .
then @xmath68 if and only if @xmath0 contains a source , a sink or a transitive vertex .
[ no2 ] there is no connected graph of order at least @xmath69 with convexity number @xmath70 . taking an interesting direction for the subject of convexity in oriented graphs , in @xcite , tong ,
yen and farrugia introduced the concepts of convexity spectrum and strong convexity spectrum of a graph .
for a nontrivial connected graph @xmath7 , we define the _ convexity spectrum _
, @xmath71 , of a graph @xmath7 , as the set of convexity numbers of all orientations of @xmath7 , and the _ strong convexity spectrum _ , @xmath72 , of a graph @xmath7 as the set of convexity numbers of all strongly connected orientations of @xmath7 .
if @xmath7 has no strongly connected orientation , then @xmath72 is empty . the _ lower orientable convexity number _ , @xmath73 , of @xmath7 is defined to be @xmath74 and the _ upper orientable convexity number _ , @xmath75 ,
is defined to be @xmath76 .
hence , for every nontrivial connected graph @xmath7 of order @xmath77 , @xmath78 .
tong , yen and farrugia calculated the convexity and strong convexity spectra of complete graphs and also constructed , for every @xmath79 , a graph @xmath7 with convexity spectrum @xmath80 .
it is not very surprising that the strong convexity spectra of @xmath81 for @xmath82 is a `` large '' set , @xmath83 , missing only @xmath84 and @xmath85 .
nonetheless , we find very surprising that , in one hand , tong and yen proved in @xcite that @xmath86 for every pair of integers @xmath87 , and , in the other hand , we prove that the strong convexity spectrum of an @xmath88 grid , for any pair of integers @xmath89 , only lacks the set of integers @xmath90 .
so , an interesting question arises from the previous observation : what property in a graph determines a large strong convexity spectrum ?
we can discard regularity and high degrees ; grids are not regular graphs , and have both small maximum and minimum degree .
although we did not find what is so special about grids in terms of convexity , we managed to calculate the strong convexity spectra of all grids .
the rest of this paper is ordered as follows .
section [ snp ] is devoted to prove the @xmath10-completeness of the problem of determining the convexity number of a given oriented graph ; the problem remains @xmath10-complete even when restricted to bipartite oriented graphs of arbitrarily large girth . in section [ sgrids ] , we prove some basic results on the convexity number of general oriented graphs , and also introduce a concept of main importance to this work : the whirlpool orientation of a grid . using this concept ,
we prove that @xmath91 for any grid @xmath7 .
we finish the section with a result excluding some values from the strong convexity spectra of certain grids .
section [ scssg ] is devoted to calculate the strong convexity spectra of @xmath92 and @xmath93 grids for every integer @xmath67 . in section [ smain
] , the strong convexity spectra of @xmath88 grids for every pair of integers @xmath94 is calculated .
we define the problem oriented convexity number as follows . given an ordered pair @xmath95 , consisting of an oriented graph @xmath0 and a positive integer @xmath96 , determine whether @xmath0 has a convex set of size at least @xmath96 .
this first section is devoted to prove the @xmath10-completeness of oriented convexity number .
[ redu ] oriented convexity number restricted to bipartite oriented graphs of girth @xmath97 is @xmath10-complete .
let @xmath0 be an oriented graph .
given a subset @xmath98 , it can be verified in polynomial time whether @xmath99 is a convex set .
hence , oriented convexity number is in @xmath10 . in order to prove @xmath10-hardness ( and hence @xmath10-completness ) , we reduce an instance @xmath100 of the well - known @xmath10-complete problem clique to an instance @xmath101 of oriented convexity number such that @xmath102 if and only if @xmath103 , the encoding length of @xmath101 is polinomially bounded in terms of the encoding length of @xmath100 , and @xmath0 is bipartite .
let @xmath100 be an instance of clique .
let us assume that @xmath104 and @xmath7 is connected ; we construct @xmath0 as follows .
for every vertex @xmath18 of @xmath7 we create a directed hexagon , @xmath105 , with two antipodal distinguished vertices @xmath106 and @xmath107 . for every edge @xmath108
we add the arcs @xmath109 and @xmath110 to @xmath0 .
we create four additional vertices @xmath111 with arcs @xmath112 for @xmath113 and arcs @xmath114 for every @xmath115 . of theorem [ redu ] when @xmath7 is @xmath116 .
] clearly , @xmath117 and @xmath118 .
it is direct to observe that the digraph @xmath0 is bipartite and strongly connected . in figure [ redufig ]
the classes of the bipartition are given by the vertices of the same color ( black and white ) .
we also claim the following statements to hold .
[ c1 ] let @xmath18 be a vertex in @xmath7 and let @xmath99 be a convex set of @xmath0 with @xmath119 . if @xmath120 , then @xmath121 .
[ c2 ] let @xmath99 be a convex set of @xmath0 with @xmath119 .
if @xmath122 for @xmath123 , then @xmath124 .
[ c3 ] let @xmath125 be such that @xmath126 . if @xmath99 is a convex set of @xmath0 such that @xmath127 , then @xmath128 .
[ c4 ] if @xmath5 is a clique of @xmath7 , then @xmath129 is a convex set of @xmath0 .
it follows from claims [ c1]-[c4 ] that @xmath99 is a proper convex set of @xmath0 with @xmath119 , if and only if there exists a clique @xmath5 in @xmath7 such that @xmath129 . considering a maximum clique of @xmath7 , and a maximum proper convex set of @xmath0
, we obtain @xmath130 .
hence , if @xmath131 , then @xmath7 contains a clique of size at least @xmath96 if and only if @xmath0 contains a convex set of size at least @xmath132 .
since the encoding length of @xmath101 is linearly bounded in terms of the encoding length of @xmath100 , oriented convexity number is @xmath10-complete .
we will consider two cases .
first , assume that @xmath133 .
let @xmath134 and @xmath135 be vertices in @xmath136 .
clearly , @xmath137 is a @xmath138-geodesic in @xmath0 and @xmath139 is a @xmath140-geodesic in @xmath0 .
hence , @xmath121 .
for the second case suppose that @xmath141 and @xmath142 .
therefore , every @xmath138-directed path in @xmath0 uses the vertex @xmath106 , and every @xmath140-directed path in @xmath0 uses the vertex @xmath107 .
thus , @xmath133 and we are back to the first case .
suppose first that @xmath143 with @xmath144 .
assume without loss of generality that @xmath145 .
every @xmath146-directed path in @xmath0 uses the vertices @xmath147 and @xmath148 , thus @xmath149 . for every @xmath115
, @xmath150 is a @xmath151-geodesic in @xmath0 .
we conclude from claim [ c1 ] that @xmath121 for every @xmath115 , and it follows that @xmath152 .
if @xmath122 for @xmath123 and @xmath153 for some @xmath154 , then every @xmath155-directed path in @xmath0 uses the vertex @xmath147 and every @xmath156-directed path in @xmath0 uses the vertex @xmath148 .
thus , @xmath149 and we are done .
it follows from claim [ c1 ] that @xmath157 . since @xmath158
, it is easy to observe that @xmath159 is an @xmath160-geodesic in @xmath0 . hence , @xmath161 and claim [ c2 ] guarantees @xmath128 .
it is direct to verify that @xmath162 is a convex set of @xmath0 for every @xmath115 .
let @xmath163 and @xmath164 .
if @xmath141 and @xmath165 , then every @xmath166-directed path in @xmath0 uses the vertices @xmath106 and @xmath167 ; moreover , it must contain an @xmath168-directed path .
our claim follows from noting that @xmath169 .
it is easy to observe that the directed @xmath97-cycles in the previous construction can be replaced by directed @xmath170-cycles , and the directed path @xmath171 can be replaced by a directed path of length @xmath77 , in order to get the result for an oriented bipartite graph of arbitrary large girth .
we begin this section with two straightforward lemmas regarding convex sets in strong oriented graphs .
the first lemma is a direct observation , so the proof will be omitted .
[ girthcon ] let @xmath0 be a strongly connected oriented graph . if @xmath98 is a convex set such that @xmath119 , then @xmath99 induces a strong subdigraph of @xmath0
therefore , @xmath172 , where @xmath173 stands for the girth of @xmath7 .
[ comcon ] let @xmath0 be an oriented graph . if @xmath98 is a maximal convex set , then @xmath174 is a connected subdigraph of @xmath0 .
otherwise , let @xmath175 be the connected components of @xmath176 .
since @xmath99 is a convex set of @xmath0 , @xmath177 is a convex set of @xmath0 properly containing @xmath99 , a contradiction .
the following observation is simple , but also very useful while searching for adequate orientations to realize specific convex numbers .
the proof is straightforward and thus will be omitted .
[ ai - ao ] let @xmath7 be a triangle - free graph and let @xmath0 be an orientation of @xmath7 with a convex set @xmath99 .
let @xmath178 be adjacent to a vertex @xmath179 .
* if @xmath180 , then @xmath181 . *
if @xmath182 , then @xmath183 .
let @xmath77 be an integer .
we denote by @xmath184 the path on @xmath77 vertices , and we will assume without loss of generality that @xmath185 .
the @xmath186-_grid _ is the cartesian product @xmath187 .
hence , if @xmath188 , then @xmath189 .
although it is not standard , and it can be impractical in a different context , for the sake of simplicity we will denote the ordered pair @xmath190 as @xmath191 .
also , we will use the canonical embedding of the grid @xmath188 in the plane to define orientations of @xmath7 , and directed paths in an orientation @xmath0 of @xmath7 . to achieve this goal
, we will use directions to `` move '' on the grid , denoted as a sequence of movements using the symbols @xmath192 , which stand for up , down , left and right . as an example , consider a directed cycle denoted in the usual way , _
i.e. _ , @xmath193 ; with our notation we have the sequence @xmath194 , starting at vertex @xmath191 . in the following paragraph
, there is another example of the orientations that can be defined in this way .
let @xmath195 be integers and @xmath7 be the canonical plane embedding of the grid @xmath187 .
let @xmath196 be a connected subgraph of the interior dual of @xmath7 and let @xmath53 be the subgraph of @xmath7 induced by the faces in @xmath197 ( hence , the interior dual of @xmath53 is @xmath196 ) .
we define a _ whirlpool _ to be an oriented graph obtained from @xmath53 by the following orientation of its edges .
@xmath198 an example of a whirlpool is depicted in figure [ whirlpool ] , where the graph @xmath196 has the gray faces of the grid as its vertex set . in the rest of the figures
, the gray squares will always correspond to directed cycles . as the following proposition shows ,
whirlpools have a very important property related to convexity . [ lcon1grids ] if @xmath0 is a whirlpool , then @xmath199 .
let @xmath0 be a whirlpool .
hence , there exist a pair of integers @xmath195 and a subgraph @xmath53 of the grid @xmath187 such that @xmath0 is obtained from @xmath53 by the aforementioned orientation of its edges .
we affirm that every @xmath69-cycle in @xmath53 is an oriented @xmath69-cycle in @xmath0 . if @xmath99 is a @xmath69-cycle in @xmath53 , then @xmath200 for some @xmath201 .
it is not difficult to observe that if @xmath202 , then @xmath99 is an oriented cycle in @xmath0 .
else , @xmath203 ( the cycle @xmath99 in reverse order ) is an oriented cycle in @xmath0 .
hence , @xmath0 is strongly connected .
since every edge of @xmath53 belongs to a @xmath69-cycle , it follows that every arc of @xmath0 belongs to a directed @xmath69-cycle . recalling that @xmath0 is triangle free ,
it is clear that @xmath204 implies @xmath205 .
hence , if @xmath204 , then the vertex set of every @xmath69-cycle containing the arc @xmath206 is contained in the convex hull of @xmath207 .
but @xmath0 is strongly connected , so using the fact that the interior dual of @xmath53 is connected , it can be shown inductively that for every pair @xmath208 of vertices of @xmath0 , the convex hull of @xmath209 is @xmath25 .
therefore , @xmath199 .
[ con=1 ] for every grid @xmath7 , @xmath210 .
observe that if an oriented grid @xmath0 contains a whirlpool @xmath211 as a subdigraph , then every convex set containing at least two adjacent vertices of @xmath211 must contain @xmath212 .
also , if @xmath211 is a whirlpool , then the digraph @xmath213 obtained by the reversal of every arc of @xmath211 has the same properties as @xmath211 ; we will call such a digraph an _ anti - whirlpool_. for a given integer @xmath96 and an oriented graph @xmath0 , it is easier to prove @xmath214 than proving @xmath215 .
the following result excludes some values from the strong convexity spectra of grids . although simple , the complete proof of the lemma is to long to be included here .
the proof of the first item of the lemma is complete , as well as the cases @xmath216 and @xmath217 of the second item .
the proofs for the cases @xmath218 and @xmath219 can be obtained with similar arguments .
[ forbid ] let @xmath195 be integers . if @xmath188 , then : * @xmath220 .
* for every @xmath221 , if @xmath222 , then @xmath223 .
by theorem [ no2 ] we have @xmath224 . observing that every connected subdigraph of @xmath7 with @xmath225 or @xmath226 vertices has at least one vertex of degree @xmath227 , and thus does not admit a strong orientation , it follows from lemma [ girthcon ] that @xmath228 .
any strong orientation of @xmath7 has neither sinks nor sources . also , @xmath229 and hence the orientations of @xmath7 can not have transitive vertices .
it follows that @xmath230 .
let @xmath231 be integers such
that @xmath232 and suppose that a strong orientation @xmath0 of @xmath7 has a convex set @xmath99 of cardinality @xmath233 .
let @xmath234 be the set @xmath235 and assume without loss of generality that @xmath180 . since @xmath0 is strong , @xmath236 and @xmath237 .
it follows from observation [ ai - ao ] that @xmath238 and @xmath239 .
let us denote by @xmath240 the vertices above , below , to the left and to the right of @xmath241 , respectively , in g. since @xmath232 , we can assume without loss of generality that either @xmath242 or @xmath243 is a directed path in @xmath0 , and hence a geodesic .
but this contradicts that @xmath99 is a convex set .
hence , a convex set of cardinality @xmath233 can not exist .
let @xmath231 be integers such that @xmath244 and suppose that a strong orientation @xmath0 of @xmath7 has a convex set @xmath99 of cardinality @xmath245 .
let @xmath246 be the set @xmath235 .
let us assume without loss of generality that @xmath180 and @xmath247 .
we have two cases .
first consider @xmath248 .
again , we have two cases .
our first subcase is @xmath249 . as in the previous argument
, either @xmath243 or @xmath250 is a geodesic in @xmath0 , contradicting that @xmath99 is a convex set .
our second subcase is @xmath251 .
we will assume without loss of generality that @xmath252 .
hence , either @xmath253 or @xmath254 is a geodesic in @xmath0 , contradicting that @xmath99 is a convex set .
as a second case , consider @xmath255 , with two subcases .
our first subcase is @xmath249 .
either @xmath256 or @xmath257 is a geodesic in @xmath0 , a contradiction .
the second subcase is @xmath251 .
hence , at least one of the following directed paths is present in @xmath0 , and it is a geodesic : @xmath258 . but every case results in a contradiction .
hence , a convex set of cardnality @xmath245 can not exist .
the following pair of results deal with the convexity spectra of @xmath92 grids .
[ 2*n - part1 ] let @xmath67 be an integer and let @xmath7 be the grid @xmath259 .
if @xmath260 is an integer such that @xmath261 and @xmath262 , then @xmath263 .
let @xmath264 be the orientation of @xmath265 as a whirlpool .
we will consider two cases .
first , suppose that @xmath266 .
let @xmath267 be the orientation of @xmath268 as a whirlpool , if @xmath260 is odd , or as anti - whirlpool if @xmath260 is even .
in either case , the orientation of the edges @xmath269 and @xmath270 result in parallel arcs , _ i.e. _ , we have either the arcs @xmath271 and @xmath272 or the arcs @xmath273 and @xmath274 .
we will assume that @xmath260 is odd , the remaining case can be dealt similarly . if @xmath0 is the digraph obtained by orienting the two remaining edges as @xmath275 and @xmath276 , then it is clear that @xmath277 and @xmath278 . but
@xmath279 , and hence , @xmath280 is a convex set of cardinality @xmath281 . if @xmath99 is a convex set of @xmath259 such that @xmath282 , then @xmath283 . from the previous observations about @xmath284 and @xmath285 , we conclude that @xmath286 .
since @xmath264 and @xmath267 are whirlpools , we obtain @xmath287 .
therefore , @xmath280 is a maximum convex set of @xmath0 .
if @xmath288 , then we will assume that @xmath77 is odd , otherwise we are back in the previous case .
let @xmath267 be the orientation of @xmath289 $ ] as a whirlpool , if @xmath260 is odd , or as anti - whirlpool if @xmath260 is even .
again , we will assume that @xmath260 is odd .
orient the remaining edges in the following way : @xmath290 and @xmath291 . an argument similar to the one used in the previous case shows that @xmath292 and @xmath293 are convex sets of @xmath0 , and clearly @xmath294 . if @xmath99 is a convex set of @xmath259 such that @xmath282 , then @xmath295 . since @xmath296 and @xmath297 , then @xmath298 . but
@xmath299 and @xmath300 are @xmath301-geodesics in @xmath0 .
since @xmath302 and @xmath303 are whirlpools in @xmath0 , we obtain @xmath124 .
hence , @xmath292 is a maximum convex set of @xmath0 .
if @xmath67 is an integer and @xmath7 is the grid @xmath259 , then @xmath304 let @xmath0 be a strong orientation of @xmath7 and @xmath99 a maximum convex set of @xmath0 .
it follows from lemma [ comcon ] and the fact that @xmath305 $ ] is strong that , for some @xmath306 , either @xmath307 , or @xmath308 . hence , there are not odd integers greater than @xmath227 in @xmath8 .
assume without loss of generality that @xmath309 for some @xmath306 .
also assume without loss of generality that @xmath310 .
we will consider two cases .
consider for the first case @xmath311 . since @xmath0 is strong , either @xmath312 or @xmath313 . in the former case
@xmath314 is a @xmath315-geodesic in @xmath0 , a contradiction . in the latter case , it is direct to verify that @xmath316 is also a convex set ; since @xmath99 is maximum , we get @xmath266 . for the second case consider @xmath317 .
again , since @xmath0 is strong , either @xmath312 or @xmath313 . in the former case it is easy to check that @xmath316 is a convex set of @xmath0 , hence , @xmath266 . in the latter case ,
if @xmath318 , then @xmath319 is a convex set of @xmath0 , a contradiction .
hence @xmath320 and we have two further cases . if @xmath321 , then @xmath316 is a convex set of @xmath0 , thus @xmath266 . if @xmath322 , then @xmath323 is a convex set of @xmath0 , and hence @xmath324 . although a bit more complex , similar arguments can be used for the @xmath93 grids .
[ 3*n - c=3k ] let @xmath67 be an integer and let @xmath7 be the grid @xmath325 .
if @xmath260 is an integer such that @xmath306 , then @xmath326 . with @xmath327 and @xmath328 .
] we consider two cases .
if @xmath329 , then let @xmath264 be the orientation of @xmath330 as a whirlpool .
suppose that @xmath77 is even , the remaining case is analogous .
orient the remaining edges as @xmath331 .
it is straightforward to verify that this orientation @xmath0 of @xmath7 is strong and @xmath280 is a maximum convex set of @xmath0 . for @xmath332 ,
let @xmath264 be the orientation of @xmath333 as a whirlpool ( note that @xmath264 is empty for @xmath334 ) .
let @xmath267 be the orientation of @xmath335 as a whirlpool .
orient @xmath336 as a directed cycle .
also , orient @xmath337 as @xmath338 if @xmath260 is even , and reverse each of these arcs if @xmath260 is odd .
finally , orient the remaining arcs as @xmath339 .
let @xmath0 be the resulting orientation of @xmath7 , and let @xmath340 be the induced subdigraph @xmath341 $ ] of @xmath0 .
we will assume that @xmath260 is even , the remaining case can be dealt similarly .
observe that @xmath342 , and @xmath343 . since @xmath344 , @xmath345 , @xmath346 , and @xmath347 , it is clear that @xmath348 is a convex set of @xmath0 with @xmath349 .
let @xmath99 be a convex set of @xmath0 such that @xmath350 .
since every convex set in a strong digraph induces a strong subdigraph , we observe that @xmath351 .
but @xmath267 is a whirlpool , and hence @xmath352 .
note that @xmath353 and also that @xmath354 is a directed path in @xmath0 .
thus , @xmath355 . recall that @xmath346 and consider the directed path @xmath356 to conclude @xmath357 . since @xmath260 is even , @xmath358 is a directed cycle in @xmath0 .
this implies @xmath359 , but @xmath264 is a whirlpool , and hence @xmath360 and @xmath128 . therefore @xmath361 .
[ 3*n - c=3k+2 ] let @xmath67 be an integer and let @xmath7 be the grid @xmath325 .
if @xmath260 is an integer such that @xmath332 , then @xmath362 . with @xmath327 and @xmath363 .
] let @xmath264 be the orientation of @xmath364 as a whirlpool .
we will consider two cases .
for the first case consider @xmath365 , and let @xmath267 be the orientation of @xmath366 as a whirlpool , if @xmath260 is even , or as an anti - whirlpool , if @xmath260 is odd
. if @xmath260 is even , orient the remaining edges as @xmath367 to obtain the orientation @xmath0 of @xmath7 .
it is direct to verify that @xmath0 is strong . observing that @xmath368 and @xmath369
it is easy to verify that @xmath280 is a convex set of @xmath0 . if @xmath99 is a convex set of @xmath0 such that @xmath370 , then @xmath371 . but
@xmath372 , and hence @xmath351 .
recalling that @xmath267 is a whirlpool and observing that @xmath373 is a @xmath374-geodesic in @xmath0 , we conclude that @xmath124 .
hence @xmath375 . if @xmath260 is odd , orient the remaining edges as @xmath376 to obtain the orientation @xmath0 of @xmath7 .
this orientation is , locally , the dual orientation of the case when @xmath260 is even , so analogous arguments show that @xmath280 is a convex set and @xmath375 . as a second case ,
assume that @xmath377 or @xmath378 .
when @xmath260 is odd , orient @xmath379 as a directed path , and orient the arcs @xmath380 .
if @xmath381 , let @xmath267 be the orientation of @xmath382 as an anti - whirlpool and orient the remaining edges of @xmath7 as @xmath383 to obtain @xmath0 .
clearly @xmath0 is strong . also , it is direct to verify that @xmath280 is a convex set of @xmath0 with @xmath384 vertices .
let @xmath99 be a convex set of @xmath0 such that @xmath370 . if @xmath351 , then @xmath352 .
but @xmath385 is a @xmath386-geodesic in @xmath0 , and thus , @xmath387 for @xmath113 .
also , @xmath388 is a @xmath389-geodesic in @xmath0 .
this implies @xmath128 , because @xmath359 . otherwise , and because @xmath390 , @xmath360 and @xmath391 for some @xmath392 .
in any case , @xmath379 is the union of a @xmath393-geodesic and a @xmath394-geodesic in @xmath0 ( and the case @xmath395 is finished ) .
since @xmath396 is a @xmath397-geodesic in @xmath0 , we have @xmath351 and @xmath128 .
when @xmath260 is even , as in the previous case , we can orient the remaining edges of @xmath7 to obtain , locally , an orientation that is dual to the orientation when @xmath260 is odd .
hence , analogous arguments can be followed to prove that @xmath280 is a maximum convex set of @xmath0
. therefore , @xmath398 .
[ 3*n - c=3k+1 ] let @xmath67 be an integer and let @xmath7 be the grid @xmath325 .
if @xmath260 is an integer such that @xmath399 , then @xmath400 . for @xmath327 and @xmath401
] we will assume that @xmath260 is odd , the remaining case can be dealt similarly .
orient @xmath402 as a whirlpool to obtain @xmath403 .
we will consider two cases .
for the first case , suppose that @xmath378 .
let @xmath264 be the digraph obtained from @xmath404 by orienting @xmath405 as @xmath403 , @xmath406 as a directed path , and the remaining edge of @xmath407 $ ] as @xmath408 .
also , orient @xmath409 , and @xmath410 as directed paths .
orient the remaining edge as @xmath411 to obtain the digraph @xmath0 .
it is immediate to verify that @xmath0 is a strong digraph .
note that @xmath412 and @xmath413 . observing that @xmath414 , @xmath415 , @xmath416 , and @xmath417 ,
it is easy to conclude that @xmath280 is a convex set of @xmath0 with @xmath384 vertices .
let @xmath99 be a convex set of @xmath0 such that @xmath418 . since @xmath395 , then @xmath419 and hence @xmath420 . also , there is at least one vertex @xmath421 .
regardless of the choice of @xmath19 , the directed path starting at @xmath422 and defined by the sequence @xmath423 results from the union of a @xmath424-geodesic and a @xmath425-geodesic .
hence , @xmath426 and therefore @xmath427 as a second case , assume that @xmath428 .
let @xmath264 and @xmath267 be the digraphs obtained by orienting both @xmath404 and @xmath429 as whirlpools .
orient @xmath430 and @xmath431 as directed paths .
if @xmath432 , orient @xmath433 , and orient the same edge as @xmath434 otherwise . finally , orient the remaining edges as @xmath411 to obtain the digraph @xmath0 .
it is direct to verify that @xmath0 is strong .
observe that @xmath435 and @xmath436 .
noting that @xmath437 , @xmath438 and @xmath439 , it is not hard to verify that @xmath280 is a convex set of @xmath0 .
let @xmath99 be a convex set of @xmath0 such that @xmath418 . if @xmath440 , then @xmath441 , and @xmath360 .
if @xmath442 and @xmath351 , then @xmath352 . but
@xmath443 is a @xmath444-geodesic in @xmath0 , which implies @xmath445 .
the directed path with initial vertex @xmath446 and defined by the sequence @xmath447 is a @xmath448-geodesic in @xmath0 .
from here we observe that @xmath441 and thus @xmath360 . if @xmath442 and @xmath449 , then @xmath441 , and @xmath360 .
hence , in every case @xmath360 . since there is at least one vertex from @xmath450 in @xmath99 , necessarily @xmath451 .
the directed path starting at @xmath452 and defined by the sequence @xmath453 is a @xmath454-geodesic , and hence @xmath352 .
but if @xmath455 , it is easy to verify that @xmath124 .
hence , @xmath427 .
so far , we have every integer of the convexity spectrum of @xmath325 , except for @xmath69 .
our next theorem deals with the remaining case .
[ n*3-c=4 ] if @xmath456 is an integer and @xmath7 is the grid @xmath325 , then @xmath457 for @xmath458 . ]
consider the standard plane embedding of @xmath7 and color the interior faces gray and white with a checkerboard - like pattern , coloring the square on the bottom left corner with gray .
we will define an orientation of the arcs of @xmath7 using this coloring , an example can be seen on figure [ n*3-c=4fig ] .
there are two rows of squares .
enumerate the gray squares in each row from left to right .
orient the bottom left corner square as a whirlpool and , from here , orient all the gray squares in its row alternating whirlpool and anti - whirlpool orientations .
orient all the gray squares in the upper row following the same principle , but start orienting as an anti - whirlpool the first gray square . at this point
, every arc dividing two interior faces of @xmath7 has received an orientation .
every remaining unoriented edge @xmath459 of @xmath7 divides a white square from the exterior face of @xmath7 .
thus , the edge @xmath459 lies in exactly one square of @xmath7 , and has one parallel arc @xmath460 in the square .
it @xmath459 is not an edge of a corner square , orient it in the same direction as @xmath460
. there are four edges belonging to the white corner squares that remain unoriented .
orient the remaining edges as @xmath70-paths in such way that there are not white oriented squares .
let @xmath0 be the digraph obtained by this orientation . clearly , @xmath0 is strong and the vertices of each gray square conform a convex set .
let @xmath99 be a convex set of @xmath0 such that @xmath461 .
there must be two gray squares @xmath462 and @xmath463 such that @xmath464 and @xmath465 . since @xmath19 is an interior vertex , it belongs to two white squares .
let @xmath18 be a vertex in the opposite corner in one of these white squares @xmath466 .
assume without loss of generality that @xmath19 is the lower left corner of @xmath462 , the upper right corner of @xmath463 , and the lower right corner of @xmath466 ; the remaining cases can be dealt similarly . if @xmath18 is the middle vertex of a @xmath467-path or a @xmath468-path of length two , then @xmath469 , and hence @xmath470 .
otherwise , let @xmath241 and @xmath471 be the upper right and lower left corners of @xmath466 , respectively .
hence , @xmath472 and either the sequence @xmath473 starting from @xmath241 determines an @xmath3-directed path of length @xmath69 , or the sequence @xmath474 determines a @xmath4-directed path of length @xmath69 . in either case , @xmath470 .
it can be verified inductively that @xmath124 , and hence , @xmath475 .
if @xmath456 is an integer and @xmath7 is the grid @xmath325 , then @xmath476 \setminus \ { 2 , 3 , 5 , 7\}.\ ] ] by virtue of theorem [ no2 ] and lemmas [ forbid ] , [ 3*n - c=3k ] , [ 3*n - c=3k+2 ] , [ 3*n - c=3k+1 ] and [ n*3-c=4 ] , it remains to prove that @xmath477 .
let @xmath0 be a strong orientation of @xmath7 and @xmath99 a convex set of @xmath0 .
lemmas [ girthcon ] and [ comcon ] imply that @xmath99 induces a strong subdigraph of @xmath0 and that @xmath316 induces a connected subdigraph of @xmath0 , respectively .
but every connected subgraph of @xmath7 with @xmath478 vertices has either a vertex of degree @xmath227 , and thus does not admit a strong orientation ; or does not have a connected complement .
hence , @xmath477 .
the following lemma is the cornerstone of the vast majority of the arguments we will use in this section .
it is easy to check that the vertices on the boundary of each of the gray filled squares conform a convex set .
we affirm that any convex set in @xmath53 has at most @xmath69 vertices .
let @xmath99 be a maximum convex set in @xmath53 and suppose that @xmath461 .
observe that the intersection of @xmath99 with the vertices of each gray filled square is either empty , or it has one vertex , or it has four vertices .
hence , since @xmath480 , the vertices of at least two squares are contained in @xmath99 .
if the vertices of two gray filled squares different from the central one are contained in @xmath99 , then there are at least two vertices of the central square in @xmath99 ; thus , the vertices of the central square are contained in @xmath99 .
so , by symmetry , we need only to consider two cases .
the first case is when the square on the lower left corner and the central square are contained in @xmath99 .
assume that the vertex in the lower left corner of @xmath53 is @xmath481 .
it is easy to check that @xmath482 and @xmath483 are @xmath484- and @xmath485-geodesics , respectively . from here , @xmath486 , and @xmath487 is a @xmath488-geodesic in @xmath53 .
therefore , there are at least two vertices of each gray filled square in @xmath99 and we can conclude that @xmath489 , a contradiction . in the second case , we have the upper left corner and the central square contained in @xmath99 .
now , @xmath490 is a @xmath491-geodesic in @xmath53 .
hence , there are at least two vertices of the lower left corner square in @xmath99 .
so , the lower left corner square is contained in @xmath99 and we have the condition of the first case .
as the reader would expect , the main part of the argument in the next lemma s proof is the construction of the orientation .
the following lemmas will use similar orientations , so the descriptions will be very detailed in the first ones , and will loose detail as the lemmas progress . consider the standard plane embedding of @xmath7 and color the interior faces gray and white with a checkerboard - like pattern , assigning gray to the square on the bottom left corner ( like in figure [ 4 * 4-c=4 ] ) .
we will define an orientation of the arcs of @xmath7 using this coloring .
enumerate the rows of squares from bottom to top .
enumerate the gray squares in each row from left to right .
orient the bottom left corner square as a whirlpool and , from here , orient all the gray squares in the first column and first row alternating whirlpool and anti - whirlpool orientations .
now , the first square or every odd row is oriented , so we can orient all the gray squares in the odd rows alternating whirlpool and anti - whirlpool orientations .
a similar idea can be used to orient all the gray squares in even rows , but start orienting as an anti - whirlpool the first gray square on the second row . at this point
, every arc dividing two interior faces of @xmath7 has received an orientation .
we will consider two cases .
first , suppose that @xmath77 and @xmath496 are even integers , hence every corner square of @xmath7 is gray , and every remaining unoriented edge @xmath459 of @xmath7 divides a white square from the exterior face of @xmath7 .
thus , the edge @xmath459 lies in exactly one square of @xmath7 , and has one parallel arc @xmath460 in the square .
orient @xmath459 in the same direction as @xmath460 .
all the edges of @xmath7 are now oriented ; let @xmath0 be the resulting oriented graph .
figure [ 4 * 4-c=4 ] is an example of this orientation .
it is easy to verify that @xmath0 is strongly connected , and the vertices of every gray square conform a convex set of @xmath0 . if @xmath99 is a convex set of @xmath0 such that @xmath461 , then @xmath99 intersects the vertices on at least two different gray squares @xmath462 and @xmath463 in odd columns and rows . since @xmath99 induces a strong subdigraph of @xmath0 , we may assume without loss of generality that @xmath462 and @xmath463 are gray squares in the same row and adjacent odd columns .
if @xmath462 and @xmath463 are in row @xmath497 and columns @xmath260 and @xmath498 , then the vertices of @xmath462 and @xmath463 , together with the vertices of the squares @xmath466 and @xmath499 in row @xmath500 and columns @xmath260 and @xmath498 , induce a subdigraph of @xmath0 isomorphic to the digraph @xmath53 of figure [ 4 * 4-c=4 ] .
therefore @xmath501 .
we can repeat this argument using squares @xmath463 and @xmath499 and the squares in column @xmath502 and rows @xmath497 and @xmath500 .
iterating this process we obtain @xmath124 . hence , @xmath503 . for the second case , assume that @xmath77 or @xmath496 is an odd integer . by virtue of lemma [ n*3-c=4 ]
, we assume that @xmath94 .
observe that there are exactly two white corner squares in @xmath7 .
except for the edges in the white corner squares , orient the remaining edges of @xmath7 as in the previous case .
for each of the white corner squares we have the two cases depicted in figure [ n*m - c=4fig ] ( the squares we are interested in are the bottom right corners ) , and two isomorphic cases obtained by reversing all the arcs of the previous ones . we will consider two cases . if @xmath504 , complete the orientation @xmath0 of @xmath7 as in figure [ n*m - c=4fig ] .
again , it is direct to verify that @xmath0 is strong and the vertices of every gray square conform a convex set of @xmath0 .
let @xmath99 be a convex set of @xmath0 such that @xmath461 .
as in the previous case , it suffices to show that there are two consecutive gray squares in the same row or the same column that intersect @xmath99 . since @xmath461 , there @xmath99 intersects at least two different gray squares
hence , the desired condition is held , unless the gray squares are precisely those adjacent to one of the white corners .
we will assume without loss of generality that one white corner square is the one in the bottom right , as in figure [ n*m - c=4fig ] .
let @xmath462 and @xmath463 be gray squares to the left and above the white corner , respectively .
let @xmath505 be the vertex in the upper left corner of @xmath506 , @xmath507 .
consider first the situation depicted by the digraph on the left in figure [ n*m - c=4fig ] .
it is clear that @xmath508 , and also that , starting from @xmath509 , the sequence @xmath510 determines a @xmath511-directed path of length @xmath69 .
hence , @xmath99 intersects two consecutive gray squares in the same row .
for the situation depicted by the digraph on the right in figure [ n*m - c=4fig ] , it is clear that @xmath512 .
it is also clear that , starting from @xmath509 , the sequence @xmath513 determines a @xmath511-directed path of length @xmath97 .
since @xmath99 is convex , it intersects two consecutive gray squares in the same row . as a final case ,
assume without loss of generality that @xmath514 .
since @xmath77 is an odd integer , the two white corners are those on the right side of @xmath7 . in the situation depicted by the digraph on the left in figure [ n*m - c=4fig ] , use precisely that orientation and the same argument as in the previous case . in the remaining case , use the orientation of figure [ n*m - c=4fig ] for the bottom right corner , and orient the upper right corner also as a directed path of length @xmath70 ( assume that it goes up and left ) .
this orientation is strong , the vertices of each gray square conform a convex set , and the same argument as the previous case shows that we can find two consecutive gray squares in the same column that intersect @xmath99 , and hence @xmath124 .
the idea of this proof is to generalize the orientation used in lemma [ n*m - c=4 ] , but using grids of size @xmath520 instead of squares in the odd - numbered rows and columns .
an example of this orientation is depicted in figure [ n*m - c = abfig ] .
first , suppose that @xmath521 and @xmath522 .
enumerate the rows and columns of squares of @xmath7 from down to up and from left to right , respectively .
color with gray and white the squares of @xmath7 in a checkerboard - like pattern , but considering rectangles of squares instead of single squares , in the following way .
color the squares in the first @xmath523 columns and @xmath524 rows , and the square in the @xmath460-th column and @xmath525-th row wih gray .
color the squares in the @xmath460-th column and the first @xmath524 rows , and the squares in the @xmath525-th column and the first @xmath523 rows in white .
color the rest of @xmath7 with a tiling of this coloring .
enumerate the rows and columns of gray rectangles in the pattern from left to right and from down to up . clearly , the gray rectangles in the even numbered rows ( columns ) are squares .
we can assume that @xmath526 , hence , the gray rectangles in the odd numbered rows ( columns ) are proper rectangles .
orient the gray squares in the even rows as in the proof of lemma [ n*m - c=4 ] ( in particular , the first gray square of the second row is an anti - whirlpool ) .
orient the gray rectangles in the odd rows as whirlpools or anti - whirpools in such way that every gray square ( possibly except the last gray square in every row or every column ) in a even row , together with the four gray squares sharing a vertex with it , induce a digraph isomorphic to @xmath53 ( figure [ 4 * 4-c=4 ] ) .
of course , to achieve this end , we also need to orient four additional arcs dividing either two white squares of @xmath7 , or a white square and the exterior face of @xmath7 . now , every unoriented edge of @xmath7 divides two white squares , or a white square from the exterior face in an even row or column of gray squares .
orient every edge dividing two white squares in the same direction as the closest arc in a gray square in the same row or column .
there are unoriented edges parallel to the arcs oriented in the previous step ; orient those edges in the same direction as the arcs they are parallel to .
the remaining unoriented edges form paths on the exterior face of @xmath7 joining pairs of gray rectangles .
orient those paths as directed paths in such way that , if any , the corner vertices of @xmath7 in a white square have in - degree and out - degree equal to one .
figure [ n*m - c = abfig ] shows an example of this orientation .
first , suppose that @xmath539 .
orient @xmath540 as a whirlpool .
now , re - orient the corresponding arcs in the fiber @xmath541 to obtain a directed path in the same direction ( left or right ) as the arc between @xmath542 and @xmath543 . also , orient the fiber @xmath544 as a directed path in the same direction as the arc between @xmath542 and @xmath543 .
assume without loss of generality that @xmath544 is oriented right .
finally orient down every edge in @xmath545 except for @xmath546 , which is oriented up . if @xmath0 is the resulting oriented graph , then it is easy to verify that @xmath547 is a convex set of @xmath0 , and @xmath548 . if @xmath522 , then orient @xmath549 as a whirlpool and @xmath550 as a whirlpool or as an anti - whirlpool in such way that the fibers @xmath551 and @xmath552 are isomorphic . orient the remaining edges arbitrarily ,
as long as there is one arc going up and one arc going down , to obtain @xmath0 . we will show that @xmath292 is a convex set of @xmath0 .
let @xmath553 be a @xmath554-path in @xmath0 such that every intermediate vertex of @xmath553 belongs to @xmath293 .
then @xmath553 can be codified as @xmath555 .
let @xmath556 be the sequence such that @xmath557 from the way we oriented @xmath302 and @xmath303 we conclude that @xmath558 defines a @xmath554-path in @xmath0 , strictly shorter than @xmath553 and such that @xmath559 .
hence , @xmath292 is a convex set of @xmath0 . recalling that @xmath560 , and using the fact that every convex set with more than @xmath561 vertices has at least two vertices in @xmath292 and at least two vertices in @xmath293 , we conclude that @xmath548 .
consider a coloring of the squares of @xmath7 similar to the coloring used in the proof of lemma [ n*m - c = ab ] , with the following differences . in the lower left corner of the aforementioned orientation of @xmath7
we have a rectangle , @xmath569 , of @xmath570 gray squares .
consider the subgraph @xmath302 of @xmath7 obtained by deleting the first @xmath571 columns and the first @xmath572 rows of vertices of @xmath7 .
color @xmath302 as in the proof of lemma [ n*m - c=4 ] .
finally , complete the checkerboard - like pattern with gray rectangles of size @xmath573 squares in the first row , and with gray rectangles of size @xmath574 squares in the first column .
we want our largest convex set to be a segment of @xmath569 , the gray rectangle in the lower left corner .
color white the squares in the first @xmath96 columns from the top row of @xmath569 .
also color white the squares in the first @xmath580 rows of the rightmost column of @xmath569 to obtain the gray region @xmath581 .
now , orient @xmath7 as in lemma [ n*m - c = ab ] , orienting @xmath581 as a whirlpool .
orient all the remaining edges to the right and down to obtain @xmath0 .
an example of this orientation is depicted in figure [ n*m - c = ab - k - lfig ] .
now , observe that the top right corner of @xmath581 coincides with the top right corner of @xmath569 .
hence , this orientation has the same properties as the orientation of lemma [ n*m - c = ab ]
. we can assume without loss of generality that @xmath582 .
otherwise , assuming that @xmath583 , we have that @xmath584 for some pair of integerers @xmath585 such that @xmath586 , @xmath587 .
hence , @xmath588 . from here
, it is easy to observe that the only proper convex sets of @xmath0 are the vertices in each of the gray regions of @xmath0 . since the largest one is @xmath581 , which has @xmath589 vertices , we conclude @xmath590 .
first , consider the case @xmath594 and @xmath514 .
if @xmath595 , then the result follows from lemma [ n*m - c = ab - k - l ] .
the orientations for @xmath596 are shown in figure [ 5 * 4-c = nm - kfig ] .
it can be easily verified that the set of white vertices is a largest convex set for this orientation .
now suppose that @xmath597 .
the case @xmath598 follows from lemma [ n*m - c = ab - k - l ] .
the orientation for @xmath599 is depicted in figure [ 5 * 4-c = nm - kfig ] .
again , it is not hard to verify that the sets of white vertices are largest convex sets for each of the orientations .
finally , suppose that @xmath600 .
if @xmath601 , then the result follows from lemma [ n*m - c = ab - k - l ] . the orientations for @xmath602 are depicted in figure [ 5 * 4-c = nm - kfig ] . as in the previous cases , the sets of white vertices are largest convex sets for the corresponding orientation .
observe the orientations of the @xmath609 and @xmath610 grids in figure [ 5 * 4-c = nm - kfig ] for @xmath599 .
clearly , the orientation of the latter can be obtained from the orientation of the former by adding an additional row of squares at the bottom of the grid , and orienting this new row as a whirlpool .
naturally , three arcs of the former orientation should change its direction , in this case , these are the arcs @xmath611 and @xmath612 of the former oriented graph .
we have a similar situation with the orientation of the @xmath613 and @xmath610 grids for @xmath599 and @xmath614 , respectively .
again , the latter can be obtained from the former by adding one row and one column of squares , orienting the new row as an anti - whirlpool and the new column following certain pattern .
the idea of this proof is to observe that the three orientations of the @xmath610 grids , shown in figure [ 5 * 4-c = nm - kfig ] , can be naturally extended to obtain the desired orientations .
first , observe that , independently of the value of @xmath77 , we can always add a new row of squares at the bottom of the grid .
notice that we are extending the largest convex set of the grid , but the complement of such set remains unchanged .
hence , if we consider the aforementioned orientations of the @xmath610 grid , we can conclude that @xmath616 for every integer @xmath608 . to add a new column of squares we have two different behaviors .
first , consider the orientations for @xmath599 and @xmath598 . in this case , to add a new column , we need to change the direction of one arc , as in the @xmath609 and @xmath610 grids for the @xmath599 argument .
but , to add further columns we will not need to change the direction of any arc , just add a column of squares oriented as a whirlpool . following this procedure we will obtain an orientation of the grid @xmath187 for every pair of integers @xmath617 , for @xmath599 and @xmath598 , respectively . in the other hand ,
when @xmath614 , adding a new column will not preserve the complement of our largest convex set , the complement will grow larger .
if we add a new column , following the pattern as in the example in figure [ n*m - c = nm - kfig ] , we will obtain an orientation of the grid @xmath187 for every pair of integers @xmath89 , for @xmath618 .
we can generalize the orientations used for @xmath599 and @xmath598 , respectively , to obtain orientations of the grid @xmath187 for @xmath619 and @xmath620 .
the idea is to extend the pattern horizontally , we will explain how it is done for the case @xmath621 , the remaining case is analogous .
we want to preserve the @xmath622 directed path defined by the sequence @xmath623 , and the whirlpool ( or anti - whirlpool ) of @xmath624 squares in the upper right corner of the grid .
the square of the upper left corner of the grid is oriented as a whirlpool and the rest of the grid is oriented as a whirlpool or anti - whirlpool , except for some alternating squares adjacent to the aforementioned directed path .
the remaining unoriented edges are oriented as @xmath625 and , if it is still unoriented , @xmath626 . following a similar idea , orientations for every grid @xmath187 for any pair of integers @xmath627 and
any @xmath628 can be obtained .
we depict the orientations for @xmath629 for @xmath630 in figure [ n*m - c = nm - kfig ] .
since all these orientations follow a similar pattern , it can be easily proved that each one has convexity number @xmath631 , as desired .
now , consider the case @xmath514 . if @xmath634 ( mod 4 ) , then lemma [ n*m - c = nb ] gives the desired orientation .
else , we give basic orientations that can be enlarged adding columns of whirlpools , either to the largest convex set , or to its complement . the orientation for @xmath599 is the orientation used for the @xmath613 grid ( figure [ 5 * 4-c = nm - kfig ] ) .
clearly , additional rows of whirlpools can be added in the bottom of the grid .
we give the orientations for @xmath635 in figure [ n*m - c = nm - kfig2 ] .
again , the set of black vertices is the complement of the largest set of the oriented graph . since it is easy to observe that the given orientations have the desired properties , and
when we extend them the arguments remain the same , this concludes the proof ot the lemma . as a final remark
, the reader might have noticed that the proofs are lengthy and technical , which makes them hard to follow . it would be a good problem to find a short proof for the main theorem of this article .
p. duchet , convexity in combinatorial structures , proceedings of the 14th winter school on abstract analysis .
circolo matematico di palermo , palermo , 1987 .
rendiconti del circolo matematico di palermo , serie ii , supplemento no . | let @xmath0 be a connected oriented graph .
a set @xmath1 is convex in @xmath0 if , for every pair of vertices @xmath2 , the vertex set of every @xmath3-geodesic , ( @xmath3 shortest directed path ) and every @xmath4-geodesic in @xmath0 is contained in @xmath5 .
the convexity number , @xmath6 , of a non - trivial oriented graph , @xmath0 , is the maximum cardinality of a proper convex set of @xmath0 .
the strong convexity spectrum of the graph @xmath7 , @xmath8 , is the set @xmath9 . in this paper
we prove that the problem of determining the convexity number of an oriented graph is @xmath10-complete , even for bipartite oriented graphs of arbitrary large girth , extending previous known results for graphs .
we also determine @xmath11 , for every pair of integers @xmath12 .
convexity number , convex set , spectrum , oriented graph , grid 05c 12 , 05c20 |
photon - tagged jets in heavy ion collisions provide the ideal framework to study the interaction of energetic partons traversing a strongly interacting , hot and dense medium like the quark - gluon plasma @xcite .
annihilation ( @xmath2 ) or compton scattering ( @xmath3 ) results in @xmath4jet events with a prompt photon , which does not interact strongly , and a parton , which is subject to energy loss in the medium and subsequent fragmentation into a jet of hadrons @xcite .
as the photon escapes the medium without interactions , the analysis of these events will provide the best possible determination of the initial kinematics of the parton , and therefore all modifications of the jet can be attributed to the interaction of the parton or the emerging jet with the medium @xcite .
azimuthal correlations with prompt photons as trigger particles reflect these interactions and provide information about the energy loss mechanism during the propagation of the parton through the medium @xcite .
the data presented here has been taken by the star experiment during a long au+au run at top energy of @xmath5 at the rhic collider .
events triggered on a high tower ( @xmath6 ) in the barrel electromagnetic calorimeter ( bemc ) were selected for fast analysis . in the configuration used for this analysis ,
about 180 million events were offered to the trigger , corresponding to an integrated luminosity of @xmath7 .
r0.49 this analysis studies azimuthal correlations between neutral clusters in the bemc @xcite and charged tracks reconstructed with the time projection chamber ( tpc ) @xcite .
neutral trigger particles are reconstructed by adding the energy of the three neighboring towers with the highest energy deposition to a trigger tower .
for each event , only the trigger cluster with the largest transverse energy is used and only if @xmath8 .
the cluster position is defined as the center of the highest energy tower of the cluster .
clusters with a matching tpc track with @xmath9 within 0.03 in @xmath10 and @xmath11 are vetoed to remove charged particles with large energy deposition .
charged particles with high transverse momenta are then associated with these neutral trigger particles . due to large acceptance losses near the tpc edges ,
the acceptance has been limited to @xmath12 .
figure [ dphi_pc ] shows the distribution of azimuthal angles between trigger clusters with @xmath13 , which are a mixture of neutral pions and photons , and associated particles with @xmath14 . a correction for
the tracking efficiency of the associated particle has been applied .
no background has been subtracted , and the reduced near - side yield due to the charged track veto has not been corrected for .
note that the background is very low .
this greatly reduces systematic uncertainties on the extracted properties ( both shape and yield ) of the correlation signal compared to previous analyses at lower @xmath15 @xcite .
clear near- and away - side correlations are visible in peripheral as well as central events , in contrast to the disappearance of back - to - back correlations between charged particles at lower transverse momenta in central collisions @xcite .
the associated yield @xmath16 , @xmath17 per trigger particle is determined by integration over the near- ( @xmath18 ) or away - side peaks ( @xmath19 ) , subtraction of a flat background determined from the region between the peaks , and normalization with the number of trigger particles in the event sample .
figure [ icp_near ] shows the ratio of associated near - side yield per trigger particle in central and peripheral collisions as a function of centrality : @xmath20 .
the near - side yield for neutral bemc trigger clusters decreases by about 40% with centrality for two different @xmath21 bins ( solid symbols ) .
this feature is not visible in correlations between charged particles with similar cuts for trigger and associated particles ( open symbols ) @xcite , which are consistent with an unmodified near - side correlation .
the decrease of the associated yield in correlations between neutral and charged particles can be attributed to single particle high-@xmath22 suppression , which reduces the pion yield per binary collision , but does not affect prompt photon production @xcite .
the result is a lower fraction of @xmath23 triggers @xmath24 , and a proportionally lower associated near - side yield per unidentified trigger particle , if a constant yield per @xmath23 trigger is assumed .
this decrease of the @xmath23 fraction can be described by using the @xmath23 fraction from a pythia simulation as reference , scaling @xmath25 with @xmath26 from @xcite , and normalizing to the most peripheral data point to allow comparison with @xmath27 ( grey band ) .
the data is consistent with the expected decrease of the fraction of @xmath23 triggers from approximately 80% in peripheral to about half of all trigger particles in the most central collisions .
figure [ away_yield ] shows the away - side yield per neutral trigger particle with @xmath13 as a function of centrality .
the away - side correlations have both @xmath4jet and dijet contributions .
we try to describe the away - side yield by adding the expected @xmath4jet and dijet yields from a pythia simulation after applying two separate suppression factors @xmath28 and @xmath29 .
the observed suppression is consistent with both parameters being similar to the nuclear modification factor : @xmath30 .
a separation of @xmath4jet and dijet correlations is necessary to draw final conclusions .
the measurement of either suppression factor is sufficient to determine both , as the combined @xmath31 is already known .
the absence of a near - side correlation in @xmath4jet events can be used to tag dijet events by requiring a charged track with @xmath32 in a search cone around the cluster with a radius of 0.15 in the @xmath10-@xmath11-plane . the cone size and @xmath15 cut
have been chosen for high purity due to low random background , and high efficiency , which is however limited by the near - side yield per @xmath23 trigger .
figure [ htag ] shows the away - side correlation for these hadron - tagged and all , i.e. tagged as well as untagged , triggers , with a common normalization to the total number of neutral trigger clusters , for two centrality bins .
the tagged correlation function is not corrected for the tagging efficiency . for both centralities , about 20% of all correlated back - to - back pairs have a tagged trigger , indicating that dijets contribute substantially to away - side correlations not only in peripheral , but also in central events . a quantitative statement about the relative contributions from @xmath4jet and dijets in central collisions will require a precise determination of the tagging efficiency from simulations or peripheral events , where dijet correlations dominate .
the dataset from the long au+au run in 2004 allows the analysis of azimuthal correlations between photonic trigger particles with energies above 10gev and associated particles above 4gev .
a decrease of the near - side yield indicates a large fraction of triggers from @xmath4jet events ( close to 50% ) .
the centrality dependence of the away - side yield is consistent with similar values for the away - side suppression @xmath31 for dijet and @xmath4jet correlations and the nuclear modification factor @xmath26 .
the separation of dijet and @xmath4jet contributions requires further analyses with better identification of these two sources . | the measurement of @xmath0+jet events with a direct photon in coincidence with an energetic parton can provide unique insights into the propagation and fragmentation of the parton in the presence of the hot and dense medium created in heavy - ion collisions .
one way to explore these effects is the study of azimuthal correlations between the direct photon and hadrons produced in the fragmentation .
we present azimuthal correlations between photons with a transverse momentum of more than 10gev and charged hadrons with a transverse momentum of more than 2gev in au+au collisions at @xmath1 that have been measured with the star detector at rhic .
the separation of correlations with direct photons and photons from hadronic decays will be discussed . |
massive stars affect the structure and the evolution of the clouds around them substantially . globules and clouds with bright rims and cometary appearance
are found in the vicinity of many nearby ob associations e.g. vela ( gum nebula ; sridharan 1992 ) , orion ( bally et al . 1991 ) , cepheus ( indrani & sridharan 1994 ) and rosette ( patel , xie & goldsmith 1993 ) .
in addition to affecting the morphology , these stars also accelerate the globules and possibly induce star formation in them as well .
there are many clouds in the orion complex which show evidence of being affected by the nearby stars .
lynd s dark cloud l1616 with its cometary appearance is one among them .
it also harbors a bright reflection nebula ngc 1788 , also known as ced 040 , excited by a poor star cluster .
study of a cloud with an associated reflection nebula offers many advantages .
since reflection nebulae mark close spatial association of relatively dense interstellar clouds with luminous stars of spectral type b1 or later , the distance to the star and hence to the cloud can be estimated to a reasonable accuracy .
this allows one to determine sizes , masses and luminosities reliably .
a reflection nebula with newly born stars provides an oppurtunity to study the effects of recent formation of stars with intermediate masses .
in particular , one that harbors a new - born cluster is most suited to study cloud fragmentation .
stars , newly formed or otherwise , have significant effects on the thermal balance of a cloud .
hence , study of such a cloud where there is an identifiable dominant energy source ( but not as disruptive as o stars ) permits one to investigate the heating and cooling in them .
l1616 appears to be unique among the reflection nebulae .
as shown in the next section , l1616 has an estimated star formation efficiency of @xmath4 14% , much larger than other clouds with associated reflection nebulae .
the cometary appearance of the cloud , the peculiar location of the nebula near its edge facing the bright orion belt stars , and its suspected high star formation efficiency indicate that the cluster formation in the cloud is possibly induced .
furthermore , no iras source with a spectrum typical of young stellar objects is found within the cloud boundary .
this may suggest that the formation of the cluster has pre - empted any further star formation - a self regulatory process often invoked in the literature .
thus , l1616 seems to be an object of considerable interest .
we mapped this cloud in j = 1 @xmath5 0 transitions of both @xmath1co and @xmath2co with a view to determine its mass more reliably and to discern any mass motions that may be present . in this paper
we present the maps and argue that induced star formation is most likely to be the case and that the cloud is fragmenting . in the following section we estimate the star formation efficiencies ( sfe ) of a few clouds with associated reflection nebulae and
show that l1616 is unique among them .
in section 3 we present the details of the observations and the maps of l1616 .
the results are discussed in section 4 .
star formation efficiency ( sfe ) usually refers to the ratio of the total mass in the stars to the sum of this stellar mass and the mass of the cloud or cloud segment they are born in .
we estimate the masses in these two components for the bright reflection nebulae in the following way using their optical and infrared data from the literature . *
masses of the clouds :* : : the cloud mass can be expressed as 2@xmath6 , where @xmath7 is the molecular hydrogen column density , @xmath8 the atomic mass of hydrogen , @xmath9 the solid angle of the cloud and @xmath10 its distance from the sun . to estimate @xmath9 and @xmath7 reliably , the clouds need to be mapped both in @xmath1co and @xmath2co .
however , for most of the reflection nebulae such maps are not available .
hence , we use their optical obscuration solid angles listed in the catalogues of dark clouds ( lynds 1962 ; feitzinger & stuwe 1984 ; hartley et al .
we wish to note that only the mass in dense portions of the cloud , which the optical obscuration would trace well , may be relevant for the estimation of sfe .
molecular hydrogen column density seems to be nearly constant in clouds over a wide range of sizes ( larson 1981 ) and we take this to be @xmath11 @xmath12 , typical of the small molecular clouds .
the masses estimated in this way will be very approximate .
nevertheless , we believe that this will not be a severe handicap for the relative comparison of sfe attempted here .
* masses in stars :* : : many optically bright reflection nebulae are also bright in far infrared ( fir ) and have been detected by iras .
their fir emission results from reprocessing of the radiation from the exciting stars by the dust in the surrounding cloud . in most cases , the spectral type of the brightest member of the exciting stars is also known .
then , if one assumes all stars to be of the same spectral type , a lower limit to the mass in stars is given by @xmath13 . here
@xmath14 and @xmath15 , respectively , are the mass and luminosity of the brightest star in the cluster derived from its spectral type ( allen 1976 ) , @xmath16 is the fraction of the stellar radiation that is reprocessed which is taken to be 0.3 ( albedo = 0.7 , witt & schild 1986 ) and @xmath17 is the fir flux density derived in the following way using the measured flux densities in the four iras bands ( margulis , lada & young 1989 ) : + @xmath18 in this expression the flux densities are in jy and @xmath19 , the longest wavelength at which the flux density is measured is in @xmath20 m , and @xmath21 the corresponding flux density .
the last term is the estimated flux emitted longward of @xmath19 ( myers et al .
1987 ) .
this is obtained by assuming that @xmath21 is also the maximum over the entire spectrum , and that the spectrum is like that of a blackbody for wavelengths longer than @xmath19 .
however , it should be noted that if star formation is _ very _ recent , as the luminosities of protostars change considerably with time ( lada 1991 ) , the above estimate of mass in stars can be inaccurate .
since the masses of both the components depend similarly on the distance , their ratio is independent of it resulting in the following expression for the minimum sfe . @xmath22 \times \frac{1}{(2\;m_{h}\;n_{h_{2}}\;\omega_{c } ) } \nonumber \\ { } ~&=&~ 1.286 \times 10^{9}~\frac{(m / m_{\odot})}{(l / l_{\odot } ) } \ ; \frac{(s_{fir}/erg\;s^{-1}\;cm^{-2})}{(\omega_{c}/\mu~\rm steradian)}\end{aligned}\ ] ] thus , the only quantities entering the estimates are : the solid angle of the clouds , the iras flux densities of the associated nebulae and the mass and the luminosity of the exciting star with the earliest spectral type .
we use this approach to determine sfe of a few clouds with associated reflection nebulae . in table 1
we have listed the names of the nebulae ( row 1 ) , of the brightest stars exciting them ( row 2 ) , their spectral type ( row 3 ) , mass ( row 4 ) and luminosity ( row 5 ) , the names of the associated clouds ( row 6 ) , their solid angle ( row 7 ) and mass ( row 8) , names of the associated extended iras sources ( row 9 ) , their flux densities in the four bands ( row 10 - 13 ) , the fir fluxes @xmath17 ( row 14 ) , and their estimated sfe ( row 15 ) .
the spectral type of the stars exciting the nebulae have been taken from the catalogues of reflection nebulae ( van den bergh 1966 ; van den bergh & herbst 1975 ) .
table 1 shows that l1616 has the highest estimated sfe .
two other clouds , l1630 and sc064 , also have high estimated sfe but their masses are likely to be under - estimates owing to their large physical sizes .
this can be seen from the following .
one could take the volume density of molecular hydrogen , n@xmath23 , in the clouds to be the same instead of the column density . then ,
for spherical clouds , the new sfe@xmath24 is given by , @xmath25 a value of 1100 @xmath26 has been assumed for n@xmath23 such that sfe and sfe@xmath24 are the same for the cloud l1616 ( this choice does not affect the relative values of sfe@xmath24 derived ) .
row 16 in table 1 lists sfe@xmath24 calculated this way using distances ( last row ) taken from the literature ( racine 1968 ; herbst 1975 ) . table 1 shows that sfe@xmath24 for the two clouds , l1630 and sc064 , are much smaller than that of l1616 .
nevertheless , they also appear peculiar and merit further studies . before leaving this section
, we wish to emphasise that the sfes calculated here are very approximate .
the estimated stellar mass being a minimum makes the calculated sfe to be a lower limit .
but , this is made quite uncertain because of the error in the estimate of the cloud mass .
for example , the assumption that the density is the same for all the clouds may not be valid . in this sense , both the small clouds ( l1616 & sg033 ) having high sfes may be a bias effect resulting from them being denser .
thus , detailed co observations are needed to confirm these sfes .
we have carried out such observations for the cloud l1616 .
above estimates suggest that , among the clouds with associated reflection nebulae , the cloud l1616 has an abnormally high sfe . however , as mentioned already , it is necessary to make detailed co maps to determine the cloud mass and hence the sfe more reliably .
such maps would also help to ascertain if mass motions are present .
for this purpose , we mapped the cloud l1616 in @xmath27 transitions of @xmath1co and @xmath2co with a @xmath28 beam and a grid - point spacing of one beam .
the observations were carried out during the winters of 90 - 91 and 91 - 92 using the 10.4 m millimeterwave telescope at the raman research institute campus , bangalore ( for a brief description of the telescope see patel 1990 ; sridharan 1993 ) .
a filter - bank spectrometer with 250 khz ( 0.65 kms@xmath29 ) resolution covering a total bandwidth of 64 mhz was used . during the observations ,
pointing was checked by beam switched continuum scans on jupiter ( see patel 1990 for details ) and the rms pointing error was estimated to be @xmath30 .
an ambient temperature load was used for calibration . during the observations the dsb t@xmath31 ranged from 600@xmath32k to 1200@xmath32k .
frequency switching by 15.25@xmath32mhz was used for all the observations and the spectra obtained were appropriately combined .
third order polynomials were fitted to estimate and remove the curved baselines .
most of the final spectra had an rms noise of @xmath40.25 k and the estimated rms error on the velocities is @xmath4 0.3 kms@xmath29 . we estimate the forward spillover and scattering efficiency , @xmath33 , to be 0.57 and 0.63 at the @xmath1co and @xmath2co frequencies , respectively , by comparing the measured antenna temperatures on the calibration sources ( mostly ori a ) with the source brightness temperatures reported in the literature ( taken to be 73 k and 16 k at the two frequencies , respectively , for ori a ) .
we have also corrected the spectra for the elevation dependence of @xmath33 .
the kinetic temperature and the @xmath2co column density were obtained from the spectra using the relations given by dickman ( 1978 ) , assuming that @xmath2co is optically thin , and that both @xmath34 and @xmath2co have the same excitation temperatures .
these relations take into account the 2.7 k microwave background as well as the fact that the usual rayleigh - jeans approximation ( @xmath35 ) is not valid at millimeter wavelengths for the typical kinetic temperatures that obtain in the molecular clouds .
figures 1 to 4 show the distributions of the @xmath2co column density , the kinetic temperature , the equivalent width and the integrated line intensity ( @xmath36 ) in the cloud l1616 .
the @xmath1co spectra were used to obtain the last three quantities while the @xmath2co spectra yielded the first .
the ellipse in figs.@xmath321 to 4 outlines the i band image of the nebula ngc 1788 ( see plate 62 of witt & schild 1986 ) associated with l1616 .
its major and minor axes are @xmath37 and @xmath38 long , respectively .
it is centred at the position of the iras source with @xmath39 and @xmath40 ( hereafter , the _ centre _ ) and oriented @xmath41 with respect to the west .
the star , hd 293815 , of spectral type b9v is the brightest visible member of the cluster of stars illuminating the nebula ngc 1788 .
it lies @xmath42 away from the _ centre _ in the nw direction .
following are the notable features in the maps : 1 .
the @xmath2co column density peaks at @xmath43 west of the _ centre _ with a _ shoulder _ around the nebula and falls off gradually and more or less symmetrically from the peak .
2 . the temperature distribution is more or less centered on the nebula with its peak at @xmath44 east of the _ centre _ and hence not coincident with the column density peak .
this indicates that the radiation from the stars is the dominant energy source .
the spatial distribution of equivalent widths is more or less symmetrically distributed and has its peak coincident with the column density peak .
fig.3 shows that most parts of the cloud have equivalent line widths greater than @xmath45 .
the contour map of the derived _
integrated @xmath1co line intensity _ distribution in l1616 bears similarity to both @xmath2co column density ( fig.1 ) and kinetic temperature ( fig.2 ) distributions .
for example , the two peaks , one each to the east and west of the _ centre _ , correspond to the temperature and column density peaks , respectively .
this similarity implies that although the @xmath1co integrated line intensity map has column density information , it is substantially altered by the distribution of the temperature in the cloud .
hence , using integrated @xmath1co line intensities to trace the column density distributions of molecular hydrogen is not always reliable .
as stated earlier , one of the primary objectives of the observations was to determine the cloud mass more reliably .
the mass of the cloud estimated from the @xmath2co column density map presented in the previous section using a n@xmath46 to n@xmath47 conversion factor of @xmath48 ( dickman 1978 ) is @xmath49 .
one can also use the integrated @xmath1co line intensity map to determine the cloud mass .
the mass found this way using a value of @xmath50 for the i@xmath51/n@xmath46 ratio ( strong et al .
1988 ) is @xmath52 .
the two mass estimates are in good agreement with each other .
we take their mean , which is @xmath53 , to be the cloud mass .
this agrees well with our rough estimate presented in the beginning , although larger by a factor of @xmath541.15 .
the sfe calculated using this new and more reliably determined cloud mass is @xmath412% ( estimated total mass of the stars / stellar and cloud mass ) .
this is still large compared to the average sfe value of @xmath55 3% ( evans & lada 1991 ) found for the nearby molecular clouds . one can also obtain the mass from the cloud - averaged line width , provided the cloud is in virial equilibrium . for l1616 ,
since the line is wide over most parts of the cloud , the actual cloud - averaged line width turns out to be @xmath56kms@xmath29 .
the corresponding virial mass required to keep the cloud bound is @xmath57 , five times larger than the estimated mass .
conversely , if the cloud is bound , its detected mass of @xmath53 would imply a cloud - averaged line width of only @xmath58 .
thus , it appears that the energy input from the stars in the cluster is _ fragmenting _ the cloud .
the excess turbulent motions of @xmath59kms@xmath29 over and above that needed for virial equilibrium and the present cloud size of @xmath60pc suggest that the fragmentation is possibly happening for the past 1 to 2 myr .
the spatial distributions of emission in the _ linewings _ shown in fig.5 also indicate the disturbed state of the cloud .
the average emission from the material in the velocity ranges 5.4 to 6.7 kms@xmath29 ( thin ) and 8.7 to 10.4 kms@xmath29 ( thick ) , respectively , are shown superimposed one over the other .
the _ blueshifted _ and _ redshifted _ emissions have large spatial extents and distinct peaks , each a little away from the _
centre_. they clearly indicate that substantial mass motion is present , possibly caused by the activity of the stars in the cluster .
for example , the spherical _ redshifted _ emission , whose peak is coincident with the column density peak , is found to contain @xmath5410% of the cloud mass .
the conclusion that the mass motions in the cloud is substantial is further supported by the coincidence of the column density peak with the line width peak .
it is also consistent with the conclusion of de vries et al .
( 1984 ) who , from a systematic survey of reflection nebulae , found that 60% of the clouds with associated reflection nebulae have broad lines ; the enhanced emission in the wings and local line broadening arise from mass motions due to the dynamical interaction between the molecular cloud and the stars illuminating the nebulae .
incidentally , the energy contained in the mass motions is @xmath61 ergs which is less than @xmath540.1% of the energy radiated by the stars in the cluster over the past few million years .
having found that the cloud l1616 has abnormally high sfe and shows clear signs of disintegration , we set out to determine the possible cause .
its high sfe and cometary appearance suggest an external cause .
figure 6 presents a wide angle view of the portion of the sky around the cloud l1616 showing its cometary tail towards @xmath62 southwest and all nearby stars of spectral type earlier than b2 .
only such stars can affect the morphology and kinematics of nearby clouds significantly as well as cause implosion in them .
table@xmath632 provides the details of these stars taken from the literature ( hirshfeld & sinnott 1985 ) .
the projected tail is directed nearly perpendicular to the galactic plane and its extension points to stars 3 and 4 as possible causes . however ,
when their distances from the cloud and their spectral types are considered , it is clear that star 3 ( @xmath64 orionis , a blue supergiant ) is the most influential one .
its present surface temperature and luminosity estimated from its spectral type ( b0ia ) are @xmath4 28000@xmath32k and @xmath65 , respectively ( allen 1976 ) .
this implies that the spectral type of its main - sequence ( ms ) progenitor is o6 with a mass of @xmath4 @xmath66 and a lifetime of @xmath4 4.3 myr . such a star may have accelerated the cloud through the _ rocket effect _ ( oort & spitzer 1955 ) , as well as causing it to implode and form the star cluster .
earlier works ( e.g. bertoldi 1989 , bertoldi & mckee 1990 ) have shown that a cloud can implode as well as acquire a tail structure in such a process . a simple estimate of the time taken by the cloud to cover the current star - cloud separation of 50 - 70 pc using the typical induced velocities of 10 - 15 kms@xmath29 ( bertoldi and mckee 1990 ) yields a value of @xmath67myr .
this being in good agreement with the expected age of the star @xmath64 orionis suggests that its progenitor could have propelled the cloud l1616 to its present position from an initial location close to the star within the stellar life - time . in the following ,
we proceed to check this suggestion more quantitatively . using two dimensionless parameters viz .
the column density parameter , @xmath68 , and the initial shock velocity parameter , @xmath69 , bertoldi ( 1989 ) has classified clouds exposed to ionising radiation .
the cloud l1616 , given its radius of @xmath70pc and molecular hydrogen density of 1200@xmath71 , has a value of @xmath72 for @xmath68 .
the upper and lower bounds to its initial distance of 50@xmath32pc ( set by the present projected cloud separation ) and 1@xmath32pc , respectively , constrain @xmath69 to lie between -0.62 to 1.08 .
these values for the parameters place l1616 in region ii of his fig.1 . in this case , almost all of the ionising radiation will be absorbed by the recombining gas streaming off the ionisation front and the ionisation front can be considered thin relative to the cloud size . in this approximation ,
the temporal evolution of the parameters ( mass , radius , ion density , distance and speed ) of a cloud exposed to radiation from a massive star are described by the equations presented in appendix a. there are three unknowns in the problem : the escape velocity of the ionised plasma , @xmath73 , the fractional cloud velocity at present , @xmath74 and the ratio of the initial and the final values of the cloud distance , @xmath75 . however , the latter two are not independent ( as can be seen from equation a4 ) thus reducing the independent unknowns to two .
the only constraint is that the cloud moves from @xmath76 to @xmath77 within the interaction time - scale set by the age of the star .
this is bound by its ms life - time of @xmath544.3 myr and its maximum age of @xmath545 myr .
fig.7 shows a contour plot of the time needed to move to the present separation , taken to be 50@xmath32pc , as a function of @xmath76 and @xmath73 .
for the same value of @xmath73 two initial positions , one closer to the star and another closer to the cloud , result in the same time of travel .
the former solution is preferred as it results in a larger average density enhancement ahead of the shock front ( which goes as @xmath78 ) , thus increasing the chances of external triggering .
although , @xmath76 and @xmath73 can not be uniquely fixed , it is clear that , for reasonable values of these parameters , the progenitor of the star @xmath64 orionis could have caused the propulsion of the cloud to its present position within the stellar life - time .
for a temperature of 10000@xmath32k , the sound speed in the ionised plasma is @xmath79 and , observationally , the plasma in a bright condensation in m16 is found to be streaming with a speed of @xmath80 ( courtes , cruvellier & pottasch , 1962 ) .
hence , we take @xmath73 to be typically between 12 and 15 kms@xmath29 .
the projected and 3-d separation between the star and the cloud are 50@xmath32pc and 70@xmath32pc , respectively . if these are taken to be the lower and upper limits to the distance travelled in the maximum available time of @xmath81myr , along with a corresponding escape speed for the plasma of 12 and 15 kms@xmath29 , the needed initial separations turn out to be 5@xmath82pc and 4@xmath82pc , respectively .
fig.8 shows the variation of the cloud parameters with time for these two cases .
the corresponding initial mass , initial radius and the present speed of the cloud for the two cases are : @xmath83m@xmath84 , 1.29@xmath32pc , 13.1@xmath32kms@xmath29 and @xmath85m@xmath84 , 1.34@xmath32pc , 18.0@xmath32kms@xmath29 , respectively .
for the lower velocity case , the measured radial velocity of @xmath86kms@xmath29 , which includes a galactic differential rotation contribution of @xmath87kms@xmath29 , would imply a proper motion velocity of @xmath88kms@xmath29 .
this would require that its direction of motion makes an angle @xmath89 with respect to the line of sight ( los ) or , equivalently , the difference in the distance to the star and the cloud from the sun is not more than @xmath5415@xmath32pc .
this is within the errors ( @xmath4 20% ) of the estimated distances to @xmath64 orionis and l1616 of 370 pc and 420 pc , respectively .
incidentally , the density @xmath90 behind the shock front varies from 150 to 15 @xmath26 ( see fig.8 ) .
the density ahead of the front @xmath91 is approximately , @xmath92 where , @xmath93 s and @xmath94 s are the corresponding sound speeds and temperatures . taking @xmath95 and @xmath96 to be 40 k and 10000 k , respectively , @xmath91 is found to vary from @xmath97 to @xmath98 over the last 5@xmath32myr .
this substantial density enhancement in the frontside of the cloud over a long period could have caused it to implode and form the cluster .
the cloud fragmentation timescale estimated earlier is consistent with this scenario .
a quantity which could have substantiated this further is the tail - stretching timescale .
however , it could not be obtained as no velocity gradient along the tail was discernible .
this may be due to the insufficient velocity resolution used ( @xmath99 at 115 ghz ) .
also , it may be difficult to determine any possible gradient due to the confusion resulting from the mass motions in the cloud generated by the cluster . nevertheless , assuming that the tail makes an angle of @xmath100 with the los and the projected length of the tail is @xmath101pc , the upper limit to the velocity gradient can be translated to a lower limit to the tail stretching time - scale of @xmath102myr .
this agrees with the available travel time , but does not constrain it well . in conclusion ,
the above discussion is consistent with the hypothesis that the star @xmath64 orionis has been the external trigger .
a measurement of the proper motion velocity of hd293815 , the brightest star exciting the reflection nebula , will confirm this .
we shall now summarise our main findings .
our main objectives were to determine the mass of the cloud l1616 and to find out if large scale mass motions are present .
for this purpose , we mapped this cloud in the j = 1 @xmath5 0 transitions of both @xmath1co and @xmath2co .
we have presented the distributions of the @xmath2co column density , the kinetic temperature , the equivalent width and the integrated line intensity in the cloud .
they clearly indicate that the star cluster exciting the reflection nebula is the dominant energy source and that substantial mass motions are present in the cloud .
the masses determined from the maps of @xmath2co column density and @xmath1co integrated line intensity are in agreement with each other and suggest that the mass of the cloud l1616 is @xmath53 .
this confirms the high star formation efficiency suspected in this cloud if the stellar mass estimated from the fir fluxes is reasonably correct .
however , the virial mass estimated from the cloud - averaged line width of @xmath103 is @xmath57 .
this is much larger than the above estimates and suggests that the cloud may be disintegrating .
the evidence for mass motions supports this .
the cometary appearance , the location of the star cluster near the edge facing the bright orion stars , and the estimated high star formation efficiency suggest that the cluster formation may have been externally triggered . the star _
@xmath64 orionis _ located opposite to the tail direction is the most likely external trigger .
it is found that this star could have propelled the cloud from its initial distance to its present position in a time - scale of @xmath545 myr for reasonable values of the two parameters viz . the initial distance to the cloud from the star and the escape velocity of the ionised gas .
if the cluster - formation is externally triggered , the average density enhancement must have been high .
this requires that the cloud was initially close to the star which leads to a large space velocity of @xmath104kms@xmath29 at present .
then , the measured radial velocity requires the cloud to have a proper motion velocity of @xmath105 .
i am grateful to the faculty of the joint astronomy program , department of physics , indian institute of science , bangalore for providing me an opportunity to do research in astronomy .
my sincere thanks are due to the staff of the millimeterwave laboratory and the observatory at the raman research institute for their help and support during the observations .
i am also thankful to k.r.anantharamiah and t.k.sridharan for their critical comments on the manuscript and to the anonymous referee for the useful suggestions and critical comments .
i allen c.w . , 1976 , astrophysical quantities .
william clowes and sons , london , p. 209
i bally j. , langer w.d . ,
wilson r.w . , stark a.a . ,
pound m.w . , 1991 , in falgarone e. , boulanger f. , duvert g. , eds , proc .
147 , fragmentation of molecular clouds and star formation .
kluwer , dordrecht , p.11 i bertoldi f. , 1989 , apj , 346 , 735 i bertoldi f. , mckee c.f .
, 1990 , apj , 354 , 529 i courtes g. , cruvellier p. , pottasch s.r . , 1962 , ann .
, 25 , 214 i de vries c.p . , brand j. , israel f.p . ,
de graauw th .
, wouterloot j.g.a .
, van de stadt h. , habing h.j . , 1984 , a&as , 56 , 333 i dickman r.l . , 1978 ,
apjs , 37 , 407 i evans@xmath106ii n.j . , lada e.a . , 1991 , in falgarone e. , boulanger f. , duvert g. , eds , proc .
147 , fragmentation of molecular clouds and star formation .
kluwer , dordrecht , p.293 i feitzinger j.v .
, stuwe j.a
. , 1984 , a&as , 58 , 365 i hartley m. , manchester r.r . ,
smith r.m .
, trittou s.b .
, goss w.m . , 1986 , a&as , 63 , 27 i herbst w. , 1975 , aj , 80 , 212 i hirshfeld a. , sinnott r.w . , 1985 , sky catalogue 2000.0 vol.2 .
cambridge univ . press ,
cambridge i indrani c. , sridharan t.k . , 1994 , ja&a , 15 , 157 i lada c.j . , 1991 , in lada c.j . , kylafis n.d .
, eds , nato asi series c , 342 , the physics of star formation and early stellar evolution .
kluwer , dordrecht , p.329 i larson r.b .
, 1981 , mnras , 194 , 809 i lynds b.t . , 1962 , apjs , 7 , 1 i margulis m. , lada c.j . ,
young , e.t . , 1989 ,
apj , 345 , 906 i myers p.c .
, fuller g.a . ,
mathieu r.d . ,
beichman c.a . ,
benson p.j .
, schild r.e .
, emerson j.p . , 1987 , apj , 319 , 340 i oort j.h . , spitzer l. , 1955 , apj , 121 , 6 ipatel n.a . , 1990 ,
indian institute of science , bangalore i patel n.a . , xie t. , goldsmith p.f .
, 1993 , apj , 413 , 593 i racine r. , 1968 , aj , 73 , 223 i spitzer l. , 1978 , physical processes in the interstellar medium .
john wiley and sons , new york , p. 262
i sridharan t.k .
, 1992 , ja&a , 13 , 217 i sridharan t.k . , 1993 ,
bull.astron.soc.india , 21 , 339 i strong a.w . , blomen j.b.g.m . ,
, grenier i.a .
, hermsen w. , lebrun f. , nyman l.a . ,
pollock a.m.t . , thaddeus p. , 1988 , a&a , 207 , 1 i van den bergh s. , 1966 , aj , 71 , 990 i van den bergh s. , herbst w. , 1975 , aj , 80 , 208 i witt a.n .
, schild r.e . , 1986 ,
apjs , 62 , 839 an early type star , when first formed , is expected to pump uv photons into the surrounding ism rather abruptly .
many dense molecular condensations within its stromgren radius ( which is 12@xmath32pc for a star of spectral type o6v for an assumed value of 10@xmath32@xmath26 for the mean hydrogen density of the surrounding medium ) will experience a weak r - type ionisation front passing quickly around them , after which the stellar uv radiation will drive an ionisation front into the denser cloud .
the ionised hydrogen produced on the side of the cloud facing the star , being at a higher pressure than the gas outside owing to its higher density , would expand producing a recoil on the cloud and thus accelerate it away from the star .
the loss of this gas also results in the reduction of the cloud mass , and hence the cloud radius . in this appendix
we derive equations describing how the cloud s position @xmath107 , radius @xmath108 , mass @xmath109 and relative velocity with respect to the star @xmath93 change with time under certain assumptions .
let @xmath110 and @xmath111 , respectively , be the present velocity , mass and radius of the cloud and @xmath76 and @xmath77 , respectively , be its initial and present distance from the star @xmath64 orionis . integrating the equation of motion , taking @xmath93 to be zero initially , one gets the following result for rocket acceleration , @xmath112 here , @xmath113 is @xmath93 expressed as a fraction of @xmath73 , the velocity at which the ionised gas escapes .
the latter is nearly equal to the sound speed in the ionised plasma which is @xmath5411.4 kms@xmath29 . in obtaining the above result we have ignored the deceleration due to sweeping up of matter .
a cloud of radius @xmath541 pc ploughing through a medium with an average hydrogen density of 1 @xmath26 over a distance of @xmath5470 pc would sweep up only a mass of @xmath544 m@xmath84 . since this is much smaller compared to the present cloud mass , deceleration due to
it can be ignored .
this is particularly justified in the case of l1616 because the cloud is at an altitude of @xmath114190 pc while the star @xmath115orionis is at an altitude of @xmath114120 pc ( their galactic longitudes and latitudes are : [ 203.5 , @xmath11624.7 ] and [ 205.2 , @xmath11617.2 ] , respectively ) .
thus , most of its motion would have taken place at high altitudes , nearly perpendicular to the galactic plane , where the general ism density is likely to be quite low resulting in negligible swept - up mass .
the rate at which the cloud loses mass is given by @xmath117 , where @xmath118 is the mass of a proton and @xmath119 is the number density of hydrogen ions just behind the ionisation front ( spitzer 1978 ) . here ,
both @xmath120 and @xmath108 are taken to depend on @xmath113 .
as the cloud recedes from the star , the ionising photon density at the cloud surface will vary and hence cause @xmath120 to vary .
since the cloud loses mass , the cloud radius would also change .
if the mean density is conserved , then @xmath121 is given by , @xmath122 assuming that the gas expands spherically and that the recombinations occuring in it nearly completely _ shield _ the stellar uv photons , one obtains the following expression for @xmath119 ( spitzer 1978 ) , @xmath123 where , @xmath124 is the stellar output rate of ionising photons and @xmath125 is the effective recombination probability for hydrogen taken to be @xmath126 @xmath127s@xmath29 .
now , equating the mass loss rate to @xmath128 and solving for @xmath129 using the fact that @xmath130 one obtains , @xmath131 \right ) \\ \ni { \rm where},~~ \beta~&=&~\frac{4m_{1}}{m_{p}}~\sqrt{\frac{4\alpha}{3\pi~n_{c } ~r^{2}_{1}}}\end{aligned}\ ] ] it is worth noting that , in addition to their boundary values , the cloud variables described by the above equations depend only on @xmath113 and not on @xmath93 or @xmath73 themselves .
the fact that the infinitesimal time @xmath132 taken for the cloud to move by an infinitesimal distance @xmath133 depends on the velocity and acceleration during that period results in the following equation : @xmath134\end{aligned}\ ] ] since it was not possible to obtain a closed form solution , we have numerically integrated this equation to get the time as a function of z. in the graphs shown in fig.8 , the time axis was derived in this way .
18.0 cm [ t ] @ lcccccccccc nebula & vdb017 & vdb013 & vdb033 & vdb052 & vdb055 & vdb123 & vhe05 & vhe30 & vhe17 & vhe28 + star & + 30@xmath135549 & + 30@xmath135540 & 293815 & 37903 & 38023 & 170634 & & & & + spt & b8v & b8v & b9v & b1.5v & b4v & b7v & * & * & b1v & b5vp + m@xmath136 ( m@xmath84 ) & 4.1 & 4.1 & 3.6 & 12.7 & 7.7 & 4.7 & 3.2 & 3.2 & 14.2 & 6.5 + l@xmath136 ( l@xmath84 ) & 179 & 179 & 117 & 6887 & 1406 & 283 & 79 & 79 & 9727 & 794 + cloud & l1450 & l1452 & l1616 & l1630 & l1641 & l572 & sc025 & sc102 & sg033 & sc064 + @xmath9 ( @xmath137str . ) & 358 & 358 & 12 & 611 & 611 & 1322 & 171 & 764 & 8 & 212 + m@xmath138 ( m@xmath84 ) & 6739 & 2227 & 157 & 13862 & 6617 & 18935 & 22279 & 52259 & 547 & 3991 + iras & x0326 & x0322 & x0504 & x0539 & x0539 & x1827 & x0809 & x0916 & x0833 & x0857 + name & + 312 & + 307 & @xmath116034 & @xmath116019 & @xmath116081 & + 012 & @xmath116356 & @xmath116482 & @xmath116405 & @xmath116435 + 12@xmath137 m ( jy ) & 91 & 6 & 30 & 2090 & 52 & 32 & 3 & 14 & 117 & 984 + 25@xmath137 m ( jy ) & 30 & 6 & 68 & 177700 & 53 & 65 & 4 & 28 & 190 & 4410 + 60@xmath137 m ( jy ) & 1080 & 20 & 323 & 93400 & 304 & 375 & 14.5 & 196 & 1100 & 26400 + 100@xmath137 m ( jy ) & 1190 & 49 & 594 & 77400 & 787 & 750 & 29 & 392 & 2700 & 24300 + s@xmath139 ( 10@xmath140 ) & 1024 & 45 & 467 & 302246 & 553 & 536 & 27 & 266 & 1784 & 25530 + sfe ( % ) & 0.84 & 0.04 & 13.42 & 10.55 & 0.06 & 0.09 & 0.08 & 0.18 & 4.04 & 11.32 + sfe@xmath24 ( % ) & 0.13 & 0.01 & 13.42 & 1.23 & 0.01 & 0.01 & 0.01 & 0.01 & 2.20 & 2.47 + dist .
( pc ) & 501 & 288 & 417 & 550 & 380 & 437 & 1318 & 955 & 955 & 501 + 12.5 cm [ t ] @ rcclcc index & sao & dist .
& spt & @xmath141 & @xmath142 + no . &
& pc . & & h@xmath106m@xmath106s & @xmath143 + 1 & 132406 & 500 & o9.5v & 5 36 14 & @xmath1162 37 39 + 2 & 132387 & 540 & b1.5v & 5 35 25 & @xmath1164 50 32 + 3 & 132346 & 370 & b0ia & 5 33 41 & @xmath1161 13 56 + 4 & 132269 & 560 & b2v & 5 31 31 & @xmath1161 04 07 + 5 & 132210 & 550 & b2v & 5 28 55 & @xmath1166 44 41 + 6 & 132222 & 560 & b0v & 5 29 31 & @xmath1167 20 13 + 7 & 112830 & 560 & b1.5v & 5 27 19 & + 1 45 05 + 8 & 112861 & 470 & b1.5v & 5 28 37 & + 3 15 21 + 9 & 112734 & 490 & b1v & 5 22 09 & + 1 48 08 + 10 & 112697 & 470 & b1v & 5 20 12 & + 3 29 52 + 11 & 131451 & 460 & b2v & 4 41 41 & @xmath1168 35 44 + & l1616 & 420 & cloud & 5 04 30 & @xmath1163 25 00 + fig.1 : distribution of @xmath2co column density in l1616 .
the ellipse marks the position , the orientation and the extent of the i band image of ngc 1788 .
the peak contour level is 3.2@xmath14410@xmath145 @xmath12 and the spacing is 0.2@xmath14410@xmath145 @xmath12 .
fig.3 : distribution of equivalent width obtained from @xmath1co spectra .
the innermost contour level is 3.1 kms@xmath29 and the spacing is 0.2 kms@xmath29 .
the outer open contours are possibly artefacts arising from the uncertainty in the measured widths owing to poorer signal to noise ratios .
the portions of the cloud material lying between the open and closed contours have linewidths @xmath146kms@xmath29 , but are not shown for the sake of clarity .
the ellipse outlines the i band image of ngc 1788 .
fig.4 : distribution of integrated @xmath1co line intensity in l1616 .
the i band image of ngc 1788 is outlined by the ellipse .
the peak contour level is 33 kkms@xmath29 and the spacing is 2 kkms@xmath29 .
fig.5 : the _ thin _ and _ thick _ contours show the distribution of @xmath1co emission averaged over the velocity ranges 5.4 to 6.7 kms@xmath29 and 8.7 to 10.4 kms@xmath29 , respectively .
the elliptical i band image of ngc 1788 is also marked .
the peak contour level is 5 k with a spacing of 1 k. fig.6 : bright orion stars near the cloud l1616 .
its cometary tail is indicated .
the details of the stars are given in table 2 .
the tail when extended points to stars 3 and 4 .
star 3 is @xmath64 orionis , a blue supergiant .
owing to its nearness to the cloud and earlier spectral type , progenitor of @xmath64 orionis is more likely to be responsible for both the morphology and the efficient star formation of the cloud l1616 .
fig.7 : shows the contour plot of the time taken by the cloud l1616 to travel from its initial position to its present one , taken to be 50@xmath32pc , as a function of the initial distance , @xmath76 , and the escape velocity of the ionised gas , @xmath73 , when accelerated by the star @xmath115orionis through the rocket effect .
contours are labelled with the time taken in units of million years .
the parameter space between the 4.2@xmath32myr and 5.1@xmath32myr contours is favourable for the star being the external trigger .
fig.8 : shows the time evolution of the cloud parameters as it gets accelerated through the rocket effect . in each of the graphs , thin and thick curves
correspond to two cases with @xmath147 of 12@xmath32(15 ) kms@xmath29 , 5@xmath32(4 ) pc and 50@xmath32(70 ) pc , respectively .
distance from the star , ion density just behind the shock front , velocity and mass expressed as a fraction of its present value , 180 m@xmath84 , are the respective parameters plotted in the four graphs . | with its cometary appearance and a reflection nebula near its edge facing some bright orion stars , the lynd s cloud l1616 shows ample evidence for being affected by one or more of these massive stars . to estimate its mass and star formation efficiency as well as to determine if it is gravitationally bound , we mapped this cloud in j=1@xmath00 transitions of @xmath1co and @xmath2co .
it is found that the distribution of the emission in the line _ wings _ show clear evidence for substantial mass motions . also , the `` virial '' mass of the cloud is found to be five times the actual cloud mass determined from the @xmath2co column density map . it is argued that this cloud has abnormally high star formation efficiency and is possibly disintegrating .
the morphology and the location of the cloud indicate that it is being affected by the star @xmath3 orionis which is also possibly responsible for the cloud s unusual star formation efficiency . over a range of values of the relevant parameters ,
the star is found to quantitatively satisfy the requirements of being the cause of the observed characteristics of the cloud .
= 0.2 cm i keywords : stars : formation ; ism : clouds ; ism : individual : l1616 ; reflection nebulae .
running title : cloud l1616 proofs to : b.ramesh |
batygin & brown ( 2016 ) showed that a hypothetical planet ( `` planet nine '' ) in the outer solar system can explain several otherwise intriguing orbital properties of distant kbos .
recently , bailey et al .
( 2016 ) and gomes et al .
( 2016 ) suggested that planet nine , which has an inclined orbit relative to the orbital plane of the major planets , may also be responsible for generating the @xmath0 solar obliquity ( the misalignment angle between the sun s rotational equator and the solar system invariant plane ) .
these studies were based on somewhat formal treatments and involved numerical integrations . in this note
we present a pedestrian , but accurate calculation of the solar obliquity generated by planet nine .
this calculation yields a simple and transparent description of how the solar spin orientation depends on the property of planet nine .
batygin & brown ( 2016 ) showed that to explain the spatial clustering of the orbits of distant kbos , planet nine ( labeled `` @xmath1 '' ) must have mass @xmath2 , perihelion distance @xmath3 au , significant eccentricity ( @xmath4 ) and tens of degrees of orbital inclination with respect to the solar system invariant plane .
the angular momentum of planet nine is @xmath5 ( where @xmath6 is the unit vector ) , with l_p = 0.276l_j(m_p10 m _ ) ( a_p400au)^1/2(1-e_p^2)^1/4 , [ eq : lp]where @xmath7 is the orbital angular momentum of jupiter and we have defined the `` effective '' semi - major axis @xmath8 .
planet nine exerts a torque on each of the `` canonical '' planets ( labeled `` @xmath9 '' , from mercury to neptune ) ; this torque tends to induce a retrograde nodal precession of @xmath10 ( the orbital angular momentum unit vector of planet @xmath9 ) around @xmath6 at the characteristic rate _
jp=3m_p4m_(a_ja_p)^3 n_j , where @xmath11 are the semi - major axis and mean motion of planet j. note that @xmath12 depends on @xmath13 , so each planet has a different @xmath12 .
however , since the precession frequency due to mutual planet - planet interactions is much larger than the differential @xmath12 , all the canonical planets are strongly coupled , with their angular momentum axes aligned to each other , i.e. , @xmath14 [ see lai & pu ( 2016 ) for a precise calculation of the mutual inclinations induced by an inclined external perturber ] .
the orbital angular momentum unit vector @xmath15 of the canonical solar system planets then evolves according to the equation = _ l _ p ( _ p ) = jl_p_l_p ( ) , [ eq : dl]where @xmath16 is the inclination of planet nine ( @xmath17 ) , @xmath18 is the total angular momentum j = j = l+l_p = l+l_p_p , with @xmath19 ( note that the spin angular momentum of the sun , @xmath20 , is much smaller ) , and & & _
l=_j l_j_jpl = 2.74_jp + & & = 287.5gyr(m_p10 m _ ) ( a_p400au)^-3 .
the spin axis @xmath21 ( unit vector ) of the sun evolves due to the torques from all planets , = _ ps . _
sl ( _ ) , [ eq : ds ] where @xmath22 is the angle between @xmath21 and @xmath15 and the characteristic spin precession frequency is given by _
= _ j _ j= _ j3k_q2k _ ( m_jm _ ) ( r_a_j)^3_. here @xmath23 is the angular frequency of the sun , and @xmath24,@xmath25 are defined through the sun s moment of inertia and quadrupole moment : @xmath26 and @xmath27 , with @xmath28 .
normalizing to the values @xmath29 and @xmath30 ( corresponding to @xmath31 ; mecheri et al .
2004 ) , we find _ ps . = 2.88
_ j = 255.8gyr _ ( p_10day)^-1 , where @xmath32 . equations ( [ eq : dl ] ) and ( [ eq : ds ] ) completely determine the evolution of the spin axis of the sun . in eq .
( [ eq : dl ] ) we have neglected the torque from the solar spin on @xmath15 , and in eq .
( [ eq : ds ] ) we have neglected the torque from @xmath33 on @xmath21 ; both are excellent approximations . to solve @xmath34 analytically
, we note that eqs .
( [ eq : dl ] ) implies that @xmath15 precesses around the constant unit vector @xmath35 at the rate @xmath36 .
we transform eq .
( [ eq : ds ] ) into the frame corotating with @xmath15 , giving ( _ t)_rot = ( jl_p_l_p- _ ps . _ sl ) _ .
[ eq : dsrot]in this rotating frame , @xmath35 and @xmath15 are constant in time , and for @xmath37 and constant @xmath38 , also changes due to stellar evolution .
we ignore these complications and treat @xmath39 as a free parameter . ]
( [ eq : dsrot ] ) describes a uniform rotation of @xmath21 around a fixed axis ( see lai 2014 ) .
we set up a cartesian coordinate system where @xmath40 and @xmath41 ( so that the polar and azimuthal angles of @xmath6 is @xmath16 and @xmath42 ) . in this coordinate system , eq . ( [ eq : dsrot ] ) reduces to & & s_xt=-_y s_z + _ z s_y,[eq : sxy1 ] + & & s_yt=-_z s_x , [ eq : sxy2]where & & _ y = _
l_p_p , + & & _ z _ ps . -_l_p ( ll_p+_p ) . for @xmath43 ( consistent with @xmath37 ) , eqs .
( [ eq : sxy1])-([eq : sxy2 ] ) can be solved ( assuming that @xmath21 is aligned with @xmath15 at @xmath44 ) : s_x=-_y_z_z t , s_y=_y_z(1-_z t ) .
thus the polar and azimuthal angles of @xmath21 are given by & & _
sl(s_x^2+s_y^2)^1/2= & & _ = ^-1s_ys_x= -_z t2 .
[ eq : phistar ]
-1.4 cm and the relative longitude of ascending node @xmath45 ( between planet nine and the solar equator ) .
the effective semi - major axis of planet nine @xmath46 ( where @xmath47 is the semi - major axis ) is shown as a function of @xmath16 , the inclination of planet nine relative to the orbital plane of the canonical solar system planets .
the mass of planet nine is set to @xmath48 ( upper panel ) and @xmath49 ( lower panel ) and the solar rotation parameter is set to @xmath50 days .
the three solid lines depict eq .
( [ eq : apcon ] ) with three values of @xmath45 ( covering the allowed range , @xmath51 ) . the dashed and dotted lines depict eq .
( [ eq : lpcon ] ) for different @xmath45 and @xmath52 ( the eccentricity of planet nine ) . the intersect of a solid line and a corresponding dashed / dotted line of the same color marks the values of @xmath53 and @xmath16 ( with the corresponding @xmath33 and @xmath52 ) required to generate the observed @xmath54 and @xmath45 . ,
title="fig : " ] -0.4 cm -1.4 cm days ( black ; the current solar rotation period ) , 15 days ( blue ) and @xmath55 ( red ; implying that the spin axis of the sun is constant in time , the limit considered by gomes et al . 2016 ) , all for @xmath56 .
note that the solid line ( depicting eq .
[ eq : apcon ] ) does not depend on @xmath57 .
, title="fig : " ] -0.4
cm -0.3 cm is shown as a function of @xmath33 for several values of @xmath45 , @xmath16 and @xmath52 , as indicated .
, title="fig : " ] -0.4
cm for a given @xmath57 , the values of @xmath22 and @xmath58 at @xmath59 gyrs depend on the parameters of planet nine ( @xmath60 ) through the combination of two frequencies , @xmath61 and @xmath62 . batygin & brown ( 2016 ) suggested that the longitude of ascending node of planet nine ( relative that of the solar equator ) , @xmath63 , is about @xmath64 and ranges from @xmath65 to @xmath66 . to produce this @xmath45 and @xmath54 over time 4.5 gyrs , the parameters of planet nine must satisfy the following conditions , derived from eqs .
( [ eq : tht])-([eq : phistar ] ) : & & a_p462(m_p10m_2_p_sl f)^1/3au,[eq : apcon ] + & & ll_p-_p,[eq : lpcon ] where @xmath67 and @xmath68 is given by eq .
( [ eq : lp ] ) , and we have defined _ sl , g , f|/2-| . figures 1 - 3 illustrate the parameter space of planet nine required produce @xmath54 and @xmath69 .
figure 1 shows the effective semi - major axis @xmath70 as a function of @xmath16 for several values of planet mass and eccentricity , assuming an `` averaged '' @xmath57 or 20 days .
figure 2 illustrates how the result depends on the solar rotation parameter @xmath57 .
figure 3 shows @xmath70 as a function of @xmath33 for several values of @xmath16 and @xmath52 .
in general , a larger @xmath33 requires a smaller @xmath16 , with a modest change in @xmath70 .
there exists a minimum value of @xmath16 , as indicated by eq .
( [ eq : lpcon ] ) . in all cases
, @xmath70 lies in the range between 340 au and 480 au in order to produce the desired solar spin orientation .
this work has been supported in part by nasa grants nnx14ag94 g and nnx14ap31 g , and a simons fellowship from the simons foundation . | bailey et al . ( 2016 ) and gomes et al . ( 2016 ) recently suggested that the 6 degree misalignment between the sun s rotational equator and the orbital plane of the major planets may be produced by the forcing from the hypothetical planet nine on an inclined orbit .
here we present a simple but accurate calculation of the effect , which provides a clear description of how the sun s spin orientation depends on the property of planet nine in this scenario . |
for decades , ecologists have used entropy - like quantities as measures of biological diversity .
the basic premise is that given a biological community or ecosystem containing @xmath2 species in proportions @xmath3 , the entropy of the probability distribution @xmath4 indicates the extent to which the community is balanced or ` diverse ' .
shannon entropy itself is often used ; so too are many variants , as we shall see .
but almost all of them share the property that for a fixed number @xmath2 of species , the entropy is maximized by the uniform distribution @xmath5 .
however , there is a growing appreciation that this crude model of a biological community is too far from reality , in that it takes no notice of the varying similarities between species .
for instance , we would intuitively judge a meadow to be more diverse if it consisted of ten dramatically different plant species than if it consisted of ten species of grass .
this has led to the introduction of measures that do take into account inter - species similarities @xcite . in mathematical terms , making this refinement essentially means extending the classical notion of entropy from probability distributions on a finite set to probability distributions on a finite metric space .
the maximum entropy problem now becomes more interesting .
consider , for instance , a pond community consisting of two very similar species of frog and one species of newt .
we would not expect the maximum entropy ( or diversity ) to be achieved by the uniform distribution @xmath6 , since the community would then be @xmath7 frog and only @xmath8 newt .
we might expect the maximizing distribution to be closer to @xmath9 ; the exact answer should depend on the degrees of similarity of the species involved .
we return to this scenario in example [ eg : three ] . for the sake of concreteness , this paper is written in terms of an ecological scenario : a community of organisms classified into species .
however , nothing that we do is intrinsically ecological , or indeed connected to any specific branch of science .
our results apply equally to any collection of objects classified into types .
it is well understood that shannon entropy is just one point ( albeit a special one ) on a continuous spectrum of entropies , indexed by a parameter @xmath10 $ ] .
this spectrum has been presented in at least two ways : as the rnyi entropies @xmath11 @xcite and as the so - called tsallis entropies @xmath12 ( actually introduced as biodiversity measures by patil and taillie prior to tsallis s work in physics , and earlier still in information theory @xcite ) : @xmath13 both @xmath11 and @xmath12 converge to shannon entropy as @xmath14 .
moreover , @xmath11 and @xmath12 can be obtained from one another by an increasing invertible transformation , and in this sense are interchangeable .
when @xmath11 or @xmath12 is used as a diversity measure , @xmath0 controls the weight attached to rare species , with @xmath15 giving as much importance to rare species as common ones , and the limiting case @xmath16 reflecting only the prevalence of the most common species .
different values of @xmath0 produce genuinely different judgements on which of two distributions is the more diverse .
for instance , if over time a community loses some species but becomes more balanced , then the rnyi and tsallis entropies decrease for @xmath15 but increase for @xmath16 .
varying @xmath0 therefore allows us to incorporate a spectrum of viewpoints on the meaning of the word ` diversity ' .
here we use the diversity measures introduced by leinster and cobbold @xcite , which both ( i ) reflect this spectrum of viewpoints by including the variable parameter @xmath0 , and ( ii ) take into account the varying similarities between species .
we review these measures in sections [ sec : spectrum][sec : measures ] . in the extreme case
where different species are assumed to have nothing whatsoever in common , they reduce to the exponentials of the rnyi entropies , and in other special cases they reduce to other diversity measures used by ecologists . in practical terms , the measures of @xcite have been used to assess a variety of ecological systems , from communities of microbes @xcite and crustacean zooplankton @xcite to alpine plants @xcite and arctic predators @xcite , as well as being applied in non - biological contexts such as computer network security @xcite . mathematically , the set - up is as follows .
a biological community is modelled as a probability distribution @xmath17 ( representing the proportions of species ) together with an @xmath18 matrix @xmath19 ( whose @xmath20-entry represents the similarity between species @xmath21 and @xmath22 ) . from this data , a formula gives a real number @xmath23 for each @xmath10 $ ] , called the ` diversity of order @xmath0 ' of the community . as for the rnyi entropies , different values of @xmath0 make different judgements : for instance , it may be that for two distributions @xmath24 and @xmath25 , @xmath26 now consider the maximum diversity problem . fix a list of species whose similarities to one another are known ; that is , fix a matrix @xmath19 ( subject to hypotheses to be discussed ) .
the two basic questions are : * which distribution(s ) @xmath24 maximize the diversity @xmath23 of order @xmath0 ? * what is the value of the maximum diversity @xmath27 ?
this can be interpreted ecologically as follows : if we have a fixed list of species and complete control over their abundances within our community , how should we choose those abundances in order to maximize the diversity , and how large can we make that diversity ? in principle , both answers depend on @xmath0 .
after all , we have seen that if distributions are ranked by diversity then the ranking varies according to the value of @xmath0 chosen .
but our main theorem is that , in fact , both answers are _ independent _ of @xmath0 : * * theorem [ thm : main ] ( main theorem)**__there exists a probability distribution on @xmath28 that maximizes @xmath29 for all @xmath10 $ ] .
moreover , the maximum diversity @xmath27 is independent of @xmath10 $ ] . _ _
so , there is a ` best of all possible worlds ' : a distribution that maximizes diversity no matter what viewpoint one takes on the relative importance of rare and common species .
this theorem merely asserts the _ existence _ of a maximizing distribution .
however , a second theorem describes how to _ compute _ all maximizing distributions , and the maximum diversity , in a finite number of steps ( theorem [ thm : comp ] ) .
better still , if by some means we have found a distribution @xmath24 that maximizes the diversity of _ some _ order @xmath1 , then a further result asserts that @xmath24 maximizes diversity of _ all _ orders ( corollary [ cor : irrelevance ] ) .
for instance , it is often easiest to find a maximizing distribution for diversity of order @xmath30 ( as in example [ eg : naive ] and proposition [ propn : graph - main ] ) , and it is then automatic that this distribution maximizes diversity of all orders .
let us put these results into context .
first , they belong to the huge body of work on maximum entropy problems . for example
, the normal distribution has the maximum entropy among all probability distributions on @xmath31 with a given mean and variance , a property which is intimately connected with its appearance in the central limit theorem .
this alone would be enough motivation to seek maximum entropy distributions in other settings ( such as the one at hand ) , quite apart from the importance of maximum entropy in thermodynamics , machine learning , macroecology , and so on .
second , we will see that maximum diversity is very closely related to the emerging invariant known as magnitude .
this is defined in the extremely wide generality of enriched category theory ( * ? ? ?
* section 1 ) , and specializes in interesting ways in a variety of mathematical fields . for instance , it automatically produces a notion of the euler characteristic of an ( ordinary ) category , closely related to topological euler characteristic @xcite ; in the context of metric spaces , magnitude encodes geometric information such as volume and dimension @xcite ; in graph theory , magnitude is a new invariant that turns out to be related to a graded homology theory for graphs @xcite ; and in algebra , magnitude produces an invariant of associative algebras that can be understood as a homological euler characteristic @xcite .
this work is self - contained . to that end , we begin by explaining and defining the diversity measures in @xcite ( sections [ sec : spectrum][sec : measures ] ) .
then come the results : preparatory lemmas in section [ sec : invt ] , and the main results in sections [ sec : main ] and [ sec : comp ] .
examples are given in sections [ sec : examples - simple][sec : examples - pos - def ] , including results on special cases such as when the similarity matrix @xmath19 is either the adjacency matrix of a graph or positive definite .
perhaps counterintuitively , a distribution that maximizes diversity can eliminate some species entirely .
this is addressed in section [ sec : elim ] , where we derive necessary and sufficient conditions on @xmath19 for maximization to preserve all species . finally , we state some open questions ( section [ sec : questions ] )
. the main results of this paper previously appeared in a preprint of leinster @xcite , but the proofs we present here are substantially simpler . of the new results , lemma [ lemma : elim - main ] ( the key to our results on preservation of species by maximizing distributions ) borrows heavily from an argument of fremlin and talagrand @xcite . [
[ p : conventions ] ] conventions + + + + + + + + + + + a vector @xmath32 is * nonnegative * if @xmath33 for all @xmath21 , and * positive * if @xmath34 for all @xmath21 .
the * support * of @xmath35 is @xmath36 and @xmath37 has * full support * if @xmath38 .
a real symmetric @xmath18 matrix @xmath19 is * positive semidefinite * if @xmath39 for all @xmath40 , and * positive definite * if this inequality is strict .
ecologists began to propose quantitative definitions of biological diversity in the mid - twentieth century @xcite , setting in motion more than fifty years of heated debate , dozens of further proposed diversity measures , hundreds of scholarly papers , at least one book devoted to the subject @xcite , and consequently , for some , despair ( already expressed by 1971 in a famously - titled paper of hurlbert @xcite ) .
meanwhile , parallel discussions were taking place in disciplines such as genetics @xcite , economists were using the same formulas to measure wealth inequality and industrial concentration @xcite , and information theorists were developing the mathematical theory of such quantities under the name of entropy rather than diversity . obtaining accurate data about an ecosystem
is beset with practical and statistical problems , but that is not the reason for the prolonged debate . even assuming that complete information is available , there are genuine differences of opinion about what the word ` diversity ' should mean .
we focus here on one particular axis of disagreement , illustrated by the examples in figure [ fig : birds ] .
( 120,38 ) ( 30,0)(0,0)[b ] ( 30,-1)(0,0)[t](a ) ( 95,0)(0,0)[b ] ( 95,-1)(0,0)[t](b ) one extreme viewpoint on diversity is that preservation of species is all that matters : ` biodiversity ' simply means the number of species present ( as is common in the ecological literature as well as the media ) .
since no attention is paid to the abundances of the species present , rare species count for exactly as much as common species . from this viewpoint , community ( a ) of figure [ fig : birds ]
is more diverse than community ( b ) , simply because it contains more species .
the opposite extreme is to ignore rare species altogether and consider only those that are most common .
( this might be motivated by a focus on overall ecosystem function . ) from this viewpoint , community ( b ) is more diverse than community ( a ) , because it is better - balanced : ( a ) is dominated by a single common species , whereas ( b ) has three common species in equal proportions . between these two extremes , there is a spectrum of intermediate viewpoints , attaching more or less weight to rare species .
different scientists have found it appropriate to adopt different positions on this spectrum for different purposes , as the literature amply attests . rather than attempting to impose one particular viewpoint
, we will consider all equally .
thus , we use a one - parameter family of diversity measures , with the ` viewpoint parameter ' @xmath10 $ ] controlling one s position on the spectrum . taking @xmath15 will give rare species as much importance as common species , while taking @xmath16 will give rare species no importance at all .
there is one important dimension missing from the discussion so far .
we will consider not only the varying abundances of the species , but also the varying similarities between them .
this is addressed in the next section .
in this section and the next , we give a brief introduction to the diversity measures of leinster and cobbold @xcite .
we have two tasks .
we must build a mathematical model of the notion of ` biological community ' ( this section ) .
then , we must define and explain the diversity measures themselves ( next section ) . in brief
, a biological community will be modelled as a finite set ( whose elements are the species ) equipped with both a probability distribution ( indicating the relative abundances of the species ) and , for each pair of elements of the set , a similarity coefficient ( reflecting the similarities between species ) .
let us now consider each of these aspects in turn .
first , we assume a community or system of individuals , partitioned into @xmath41 species .
the word ` species ' need not have its standard meaning : it can denote any unit thought meaningful , such as genus , serotype ( in the case of viruses ) , or the class of organisms having a particular type of diet .
it need not even be a biological grouping ; for instance , in @xcite the units are soil types . for concreteness , however , we write in terms of an ecological community divided into species
. the division of a system into species or types may be somewhat artificial , but this is mitigated by the introduction of the similarity coefficients ( as shown in @xcite , p.482 ) .
second , each species has a relative abundance , the proportion of organisms in the community belonging to that species .
thus , listing the species in order as @xmath42 , the relative abundances determine a vector @xmath43 .
this is a probability distribution : @xmath44 for each species @xmath21 , and @xmath45 .
abundance can be measured in any way thought relevant , e.g. number of individuals , biomass , or ( in the case of plants ) ground coverage .
critically , the word ` diversity ' refers only to the _ relative _ , not absolute , abundances .
if half of a forest burns down , or if a patient loses @xmath46 of their gut bacteria , then it may be an ecological or medical disaster ; but assuming that the system is well - mixed , the diversity does not change . in the language of physics
, diversity is an intensive quantity ( like density or temperature ) rather than an extensive quantity ( like mass or heat ) , meaning that it is independent of the system s size .
the third and final aspect of the model is inter - species similarity .
for each pair @xmath20 of species , we specify a real number @xmath47 representing the similarity between species @xmath21 and @xmath22 .
this defines an @xmath18 matrix @xmath48 . in @xcite
, similarity is taken to be measured on a scale of @xmath49 to @xmath50 , with @xmath49 meaning total dissimilarity and @xmath50 that the species are identical .
thus , it is assumed there that@xmath51 in fact , our maximization theorems will only require the weaker hypotheses @xmath52 together with the requirement that @xmath19 is a symmetric matrix .
( in the appendix to @xcite , matrices satisfying were called ` relatedness matrices ' . ) just as the meanings of ` species ' and ` abundance ' are highly flexible , so too is the meaning of ` similarity ' : the simplest similarity matrix @xmath19 is the identity matrix @xmath53 .
this is called the * naive model * in @xcite , since it embodies the assumption that distinct species have nothing in common .
crude though this assumption is , it is implicit in the diversity measures most popular in the ecological literature ( * ? ? ?
* table 1 ) . with the rapid fall in the cost of dna sequencing , it is increasingly common to measure similarity genetically ( in any of several ways ) .
thus , the coefficients @xmath47 may be chosen to represent percentage genetic similarities between species .
this is an effective strategy even when the taxonomic classification is unclear or incomplete @xcite , as is often the case for microbial communities @xcite .
[ eg : sim - phylo ] given a suitable phylogenetic tree , we may define the similarity between two present - day species as the proportion of evolutionary time before the point at which the species diverged .
[ eg : sim - taxo ] in the absence of more refined data , we can measure species similarity according to a taxonomic tree .
for instance , we might define @xmath54 [ eg : sim - metric ] in purely mathematical terms , an important case is where the similarity matrix arises from a metric @xmath55 on the set @xmath28 via the formula @xmath56 .
thus , the community is modelled as a probability distribution on a finite metric space .
( the naive model corresponds to the metric defined by @xmath57 for all @xmath58 . )
the diversity measures that we will shortly define can be understood as ( the exponentials of ) rnyi - like entropies for such distributions .
here we state the definition of the diversity measures of @xcite , which we will later seek to maximize .
we then explain the reasons for this particular definition . as in section
[ sec : model ] , we take a biological community modelled as a finite probability distribution @xmath17 together with an @xmath18 matrix @xmath19 satisfying . as explained in section [ sec : spectrum ] , we define not one diversity measure but a family of them , indexed by a parameter @xmath10 $ ] controlling the emphasis placed on rare species .
the * diversity of order @xmath0 * of the community is @xmath59 ( @xmath60 ) . here
@xmath61 is the support of @xmath24 ( conventions , p. ) , @xmath62 is the column vector obtained by multiplying the matrix @xmath19 by the column vector @xmath24 , and @xmath63 is its @xmath21th entry . the hypotheses imply that @xmath64 whenever @xmath65 , and so @xmath23 is well - defined .
although this formula is invalid for @xmath66 , it converges as @xmath14 , and @xmath67 is defined to be the limit .
the same is true for @xmath68 .
explicitly , @xmath69 the applicability , context and meaning of definition are discussed at length in @xcite . here
we briefly review the principal points .
first , the definition includes as special cases many existing quantities going by the name of diversity or entropy .
for instance , in the naive model @xmath70 , the diversity @xmath71 is the exponential of the rnyi entropy of order @xmath0 , and is also known in ecology as the hill number of order @xmath0 .
( references for this and the next two paragraphs are given in ( * ? ? ?
* table 1 ) . ) continuing in the naive model @xmath70 and specializing further to particular values of @xmath0 , we obtain other known quantities : @xmath72 is species richness ( the number of species present ) , @xmath73 is the exponential of shannon entropy , @xmath74 is the gini
simpson index ( the reciprocal of the probability that two randomly - chosen individuals are of the same species ) , and @xmath75 is the berger parker index ( a measure of the dominance of the most abundant species ) . now allowing a general @xmath19 ,
the diversity of order @xmath76 is @xmath77 .
thus , diversity of order @xmath76 is the reciprocal of the expected similarity between a random pair of individuals .
( the meaning given to ` similarity ' will determine the meaning of the diversity measure : taking the coefficients @xmath47 to be genetic similarities produces a genetic notion of diversity , and similarly phylogenetic , taxonomic , and so on . ) up to an increasing , invertible transformation , this is the well - studied quantity known as rao s quadratic entropy .
given distributions @xmath24 and @xmath25 on the same list of species , different values of @xmath0 may make different judgements on which of @xmath24 and @xmath25 is the more diverse .
for instance , with @xmath70 and the two distributions shown in figure [ fig : birds ] , taking @xmath15 makes community ( a ) more diverse and embodies the first ` extreme viewpoint ' described in section [ sec : spectrum ] , whereas @xmath16 makes ( b ) more diverse and embodies the opposite extreme .
it is therefore most informative if we calculate the diversity of _ all _ orders @xmath10 $ ] .
the graph of @xmath23 against @xmath0 is called the * diversity profile * of @xmath24 .
two distributions @xmath24 and @xmath25 can be compared by plotting their diversity profiles on the same axes .
if one curve is wholly above the other then the corresponding distribution is unambiguously more diverse .
if they cross ( as in figure [ fig : main](b ) , p. ) , then the judgement as to which is the more diverse depends on how much importance is attached to rare species .
the formula for @xmath23 can be understood as follows .
first , for a given species @xmath21 , the quantity @xmath78 is the expected similarity between species @xmath21 and an individual chosen at random .
differently put , @xmath63 measures the ordinariness of the @xmath21th species witin the community ; in @xcite , it is called the ` relative abundance of species similar to the @xmath21th ' . hence the mean ordinariness of an individual in the community
is @xmath79 .
this measures the _ lack _ of diversity of the community , so its reciprocal is a measure of diversity .
this is exactly @xmath80 .
to explain the diversity of orders @xmath81 , we recall the classical notion of power mean .
let @xmath17 be a finite probability distribution and let @xmath82 , with @xmath34 whenever @xmath83 . for real @xmath84 , the * power mean * of @xmath37 of order @xmath85 , weighted by @xmath24 , is @xmath86 (
* chapter ii ) .
this definition is extended to @xmath87 and @xmath88 by taking limits in @xmath85 , which gives @xmath89 now , when we take the ` mean ordinariness ' in the previous paragraph , we can replace the ordinary arithmetic mean ( the case @xmath90 ) by the power mean of order @xmath91 . again taking the reciprocal , we obtain exactly .
that is , @xmath92 for all @xmath24 , @xmath19 , and @xmath10 $ ] .
so in all cases , diversity is the reciprocal mean ordinariness of an individual within the community , for varying interpretations of ` mean ' .
the diversity measures @xmath23 have many good properties , discussed in @xcite .
crucially , they are * effective numbers * : that is , @xmath93 for all @xmath0 and @xmath2 .
this gives meaning to the quantities @xmath23 : if @xmath94 , say , then the community is nearly as diverse as a community of @xmath95 completely dissimilar species in equal proportions . with the stronger assumptions on @xmath19
, the value of @xmath23 always lies between @xmath50 and @xmath2 .
diversity profiles are decreasing : as less emphasis is given to rare species , perceived diversity drops . more precisely : [ propn : div - dichotomy ]
let @xmath24 be a probability distribution on @xmath28 and let @xmath19 be an @xmath18 matrix satisfying . if @xmath63 has the same value @xmath96 for all @xmath65 then @xmath97 for all @xmath10 $ ] .
otherwise , @xmath23 is strictly decreasing in @xmath10 $ ] .
this is immediate from equation and a classical result on power means ( * ? ? ?
* theorem 16 ) : @xmath98 is increasing in @xmath85 , strictly so unless @xmath99 has the same value @xmath96 for all @xmath100 , in which case it has constant value @xmath96 .
so , any diversity profile is either constant or strictly decreasing . the first part of the next lemma states that diversity profiles are also continuous : [ lemma : div - cts ] fix an @xmath18 matrix @xmath19 satisfying . then : 1 .
[ part : cts - q ] @xmath23 is continuous in @xmath10 $ ] for each distribution @xmath24 ; 2 . [ part : cts - p ] @xmath23 is continuous in @xmath24 for each @xmath101 . see propositions a1 and a2 of the appendix of @xcite . finally , the measures have the sensible property that if some species have zero abundance , then the diversity is the same as if they were not mentioned at all . to express this
, we introduce some notation : given a subset @xmath102 , we denote by @xmath103 the submatrix @xmath104 of @xmath19 .
[ lemma : abs ] let @xmath19 be an @xmath18 matrix satisfying .
let @xmath102 , and let @xmath24 be a probability distribution on @xmath28 such that @xmath105 for all @xmath106 .
then , writing @xmath25 for the restriction of @xmath24 to @xmath107 , @xmath108 for all @xmath10 $ ] .
this is trivial , and is also an instance of a more general naturality property ( lemma a13 in the appendix of @xcite ) .
* for the rest of this work , fix an integer @xmath41 and an @xmath109 symmetric matrix @xmath19 of nonnegative reals whose diagonal entries are positive * ( that is , _ strictly _ greater than zero ) . also write @xmath110 for the set of probability distributions on @xmath28 . to prove the main theorem , we begin by making two apparent digressions .
let @xmath111 be any matrix .
a * weighting * on @xmath111 is a column vector @xmath112 such that @xmath113 is the column vector whose every entry is @xmath50 .
it is trivial to check that if both @xmath111 and its transpose have at least one weighting , then the quantity @xmath114 is independent of the choice of weighting @xmath112 on @xmath111 ; this quantity is called the * magnitude * @xmath115 of @xmath111 ( * ? ? ?
* section 1.1 ) . when @xmath111 is symmetric ( the case of interest here ) , @xmath115 is defined just as long as @xmath111 has at least one weighting . when @xmath111 is invertible , @xmath111 has exactly one weighting and @xmath115 is the sum of all the entries of @xmath116 .
the second digression concerns the dichotomy expressed in proposition [ propn : div - dichotomy ] : every diversity profile is either constant or strictly decreasing .
we now ask : which distributions have constant diversity profile ?
this question turns out to have a clean answer in terms of weightings and magnitude . to state it
, we make some further definitions .
a probability distribution @xmath24 on @xmath28 is * invariant * if @xmath117 for all @xmath118 $ ] .
let @xmath102 , and let @xmath119 be a nonnegative vector .
then there is a probability distribution @xmath120 on @xmath28 defined by @xmath121 in particular , let @xmath107 be a nonempty subset of @xmath28 and @xmath112 a nonnegative weighting on @xmath122 .
then @xmath123 , so @xmath120 is defined , and @xmath124 for all @xmath125 .
[ lemma : invt ] the following are equivalent for @xmath126 : 1 .
[ part : invt - invt ] @xmath24 is invariant ; 2 .
[ part : invt - supp ] @xmath127 for all @xmath128 ; 3 .
[ part : invt - ext ] @xmath129 for some nonnegative weighting @xmath112 on @xmath103 and some nonempty subset @xmath102 .
moreover , in the situation of ( [ part : invt - ext ] ) , @xmath130 for all @xmath10 $ ] .
( [ part : invt - invt])@xmath131([part : invt - supp ] ) is immediate from proposition [ propn : div - dichotomy ] . for ( [ part : invt - supp])@xmath132([part : invt - ext ] ) ,
assume ( [ part : invt - supp ] ) . put @xmath133 and write @xmath134 for any @xmath125 .
then @xmath135 , so we may define @xmath136 by @xmath137 ( @xmath125 ) .
evidently @xmath129 and @xmath112 is nonnegative .
furthermore , @xmath112 is a weighting on @xmath103 , since whenever @xmath125 , @xmath138 finally , for ( [ part : invt - ext])@xmath132([part : invt - supp ] ) and ` moreover ' , take @xmath107 and @xmath112 as in ( [ part : invt - ext ] ) .
then @xmath139 , so for all @xmath140 , @xmath141 hence @xmath142 for all @xmath10 $ ] by proposition [ propn : div - dichotomy ] .
we now prove a result that is much weaker than the main theorem , but will act as a stepping stone in the proof .
[ lemma:01 ] for each @xmath143 , there exists an invariant distribution that maximizes @xmath29 .
let @xmath143 .
then @xmath29 is continuous on the compact space @xmath144 ( lemma [ lemma : div - cts]([part : cts - p ] ) ) , so attains a maximum at some point @xmath24 .
take @xmath145 such that @xmath146 is least and @xmath147 is greatest . by lemma [ lemma : invt ] , it is enough to prove that @xmath148 .
define @xmath149 by taking @xmath150 to be the kronecker delta @xmath151 , and @xmath152 similarly .
then @xmath153 for all real @xmath85 sufficiently close to @xmath49 , and @xmath154 where holds because @xmath24 is a supremum , is a routine computation , and follow from the defining properties of @xmath22 and @xmath155 , and uses the symmetry of @xmath19 .
equality therefore holds throughout , and in particular in .
hence @xmath148 , as required .
an alternative proof uses lagrange multipliers , but is complicated by the possibility that @xmath29 attains its maximum on the boundary of @xmath144 .
the result we have just proved only concerns the maximization of @xmath29 for specific values of @xmath0 , but the following lemma will allow us to deduce results about maximization for all @xmath0 simultaneously .
a probability distribution on @xmath28 is * maximizing * if it maximizes @xmath29 for all @xmath10 $ ] .
[ lemma : invt - varying - q ] for @xmath156 , any invariant distribution that maximizes @xmath157 also maximizes @xmath29 . in particular , any invariant distribution that maximizes @xmath158 is maximizing .
let @xmath156 and let @xmath24 be an invariant distribution that maximizes @xmath157 .
then for all @xmath159 , @xmath160 since diversity profiles are decreasing ( proposition [ propn : div - dichotomy ] ) .
[ thm : main ] there exists a probability distribution on @xmath28 that maximizes @xmath29 for all @xmath10 $ ] .
moreover , the maximum diversity @xmath161 is independent of @xmath162 $ ] .
an equivalent statement is that there exists an invariant maximizing distribution . to prove this ,
choose a decreasing sequence @xmath163 in @xmath164 converging to @xmath49 .
by lemma [ lemma:01 ] , we can choose for each @xmath165 an invariant distribution @xmath166 that maximizes @xmath167 .
since @xmath144 is compact , we may assume ( by passing to a subsequence if necessary ) that the sequence @xmath168 converges to some point @xmath169 .
we will show that @xmath24 is invariant and maximizing .
we show that @xmath24 is invariant using lemma [ lemma : invt ] .
let @xmath170 .
then @xmath171 for all @xmath172 , so @xmath173 for all @xmath172 , and letting @xmath174 gives @xmath127 . to show that @xmath24 is maximizing , first note that @xmath175 maximizes @xmath167 whenever @xmath176 ( by lemma [ lemma : invt - varying - q ] ) . fixing @xmath177 and letting @xmath178 , this implies that @xmath24 maximizes @xmath167 , since @xmath167 is continuous ( lemma [ lemma : div - cts]([part : cts - p ] ) ) .
thus , @xmath24 maximizes @xmath167 for all @xmath177 . but @xmath179 as @xmath174 , and diversity is continuous in its order ( lemma [ lemma : div - cts]([part : cts - q ] ) ) , so @xmath24 maximizes @xmath158 .
since @xmath24 is invariant , lemma [ lemma : invt - varying - q ] implies that @xmath24 is maximizing .
the theorem can be understood as follows ( figure [ fig : main](a ) ) .
each particular value of the viewpoint parameter @xmath0 ranks the set of all distributions @xmath24 in order of diversity , with @xmath24 placed above @xmath25 when @xmath180 .
different values of @xmath0 rank the set of distributions differently .
nevertheless , there is a distribution @xmath181 that is at the top of every ranking .
this is the content of the first half of theorem [ thm : main ] .
alternatively , we can visualize the theorem in terms of diversity profiles ( figure [ fig : main](b ) ) .
diversity profiles may cross , reflecting the different priorities embodied by different values of @xmath0 .
but there is at least one distribution @xmath181 whose profile is above every other profile ; moreover , its profile is constant .
( 50,37 ) ( 25,-2)(0,0)[t](a ) ( 10,0)(0,0)[b]@xmath15 ( 10,5)(0,0)[b ] rank the set of distributions , and ( b ) in terms of diversity profiles.,title="fig : " ] ( 10,35.2)(0,0)[c]@xmath181 ( 10,28.9)(0,0)[c]@xmath24 ( 10,22.0)(0,0)[c]@xmath182 ( 40,0)(0,0)[b]@xmath183 ( 40,5)(0,0)[b ] rank the set of distributions , and ( b ) in terms of diversity profiles.,title="fig : " ] ( 40,35.2)(0,0)[c]@xmath181 ( 40,26.5)(0,0)[c]@xmath182 ( 40,14.5)(0,0)[c]@xmath24 ( 60,37 ) ( 30,-2)(0,0)[t](b ) ( 30,5)(0,0)[b ] rank the set of distributions , and ( b ) in terms of diversity profiles.,title="fig : " ] ( 1,15)(0,0)[r]@xmath29 ( 30,5)(0,0)[t]@xmath0 ( 10,18.5)(0,0)[c]@xmath24 ( 6,14)(0,0)[c]@xmath182 ( 16,23)(0,0)[c]@xmath181 theorem [ thm : main ] immediately implies : [ cor : max - invt ] every maximizing distribution is invariant .
this result can be partially understood as follows . for shannon entropy , and more generally any of the rnyi entropies ,
the maximizing distribution is obtained by taking the relative abundance @xmath184 to be the same for all species @xmath21 .
this is no longer true when inter - species similarities are taken into account .
however , what is approximately true is that diversity is maximized when @xmath63 , the relative abundance of species _ similar _ to the @xmath21th , is the same for all species @xmath21 .
this follows from corollary [ cor : max - invt ] together with the characterization of invariant distributions in lemma [ lemma : invt]([part : invt - supp ] ) ; but it is only ` approximately true ' because it is only guaranteed that @xmath127 when @xmath21 and @xmath22 both belong to the support of @xmath24 , not for all @xmath21 and @xmath22
. it may in fact be that some or all maximizing distributions do not have full support , a phenomenon we examine in section [ sec : elim ] .
the second half of theorem [ thm : main ] tells us that associated with the matrix @xmath19 is a numerical invariant , the constant value of a maximizing distribution : the * maximum diversity * of @xmath19 is @xmath185 , for any @xmath10 $ ] .
we show how to compute @xmath186 in the next section .
if a distribution @xmath24 maximizes diversity of order @xmath76 , say , must it also maximize diversity of orders @xmath50 and @xmath30 ?
the answer turns out to be yes : [ cor : irrelevance ] let @xmath24 be a probability distribution on @xmath28 . if @xmath24 maximizes @xmath29 for some @xmath187 $ ] then @xmath24 maximizes @xmath29 for all @xmath10 $ ] .
let @xmath187 $ ] and let @xmath24 be a distribution maximizing @xmath29
. then @xmath188 where the first inequality holds because diversity profiles are decreasing .
so equality holds throughout .
now @xmath189 with @xmath190 , so proposition [ propn : div - dichotomy ] implies that @xmath24 is invariant .
but also @xmath191 , so @xmath24 maximizes @xmath158 .
hence by lemma [ lemma : invt - varying - q ] , @xmath24 is maximizing .
the significance of this corollary is that if we wish to find a distribution that maximizes diversity of all orders @xmath0 , it suffices to find a distribution that maximizes diversity of a single nonzero order .
the hypothesis that @xmath1 in corollary [ cor : irrelevance ] can not be dropped .
indeed , take @xmath70 .
then @xmath72 is species richness ( the cardinality of @xmath61 ) , which is maximized by any distribution @xmath24 of full support , whereas @xmath73 is the exponential of shannon entropy , which is maximized only when @xmath24 is uniform .
the main theorem guarantees the existence of a maximizing distribution @xmath181 , but does not tell us how to find it .
it also states that @xmath192 is independent of @xmath0 , but does not tell us what its value is .
the following result repairs both omissions .
[ thm : comp ] 1 .
[ part : comp - value ] for all @xmath10 $ ] , @xmath193 where the maximum is over all @xmath102 such that @xmath103 admits a nonnegative weighting .
[ part : comp - dist ] the maximizing distributions are precisely those of the form @xmath120 where @xmath112 is a nonnegative weighting on @xmath103 for some @xmath107 attaining the maximum in .
let @xmath10 $ ] .
then @xmath194 where follows from the fact that there is an invariant maximizing distribution ( theorem [ thm : main ] ) , follows from lemma [ lemma : invt ] , and follows from the trivial fact that @xmath195 whenever @xmath103 admits a nonnegative weighting .
this proves ( [ part : comp - value ] ) .
every maximizing distribution is invariant ( corollary [ cor : max - invt ] ) , so ( [ part : comp - dist ] ) follows from lemma [ lemma : invt ] .
[ rmk : algorithm ] the theorem provides a finite - time algorithm for finding all the maximizing distributions and computing @xmath186 , as follows . for each of the @xmath196 subsets @xmath107 of @xmath28 , perform some simple linear algebra to find the space of nonnegative weightings on @xmath103 ; if this space is nonempty , call @xmath107 * feasible * and record the magnitude @xmath197 .
then @xmath186 is the maximum of all the recorded magnitudes . for each feasible @xmath107 such that @xmath198 , and each nonnegative weighting @xmath112 on @xmath103 , the distribution @xmath120 is maximizing .
this generates all of the maximizing distributions .
this algorithm takes exponentially many steps in @xmath2 , and remark [ rmk : no - quick - clique ] provides strong evidence that the time taken can not be reduced to a polynomial in @xmath2 .
but the situation is not as hopeless as it might appear , for two reasons .
first , each step of the algorithm is fast , consisting as it does of solving a system of linear equations .
for instance , in an implementation in matlab on a standard laptop , with no attempt at optimization , the maximizing distributions of @xmath199 matrices were computed in a few seconds .
second , for certain classes of matrices @xmath19 , we can make substantial improvements in computing time , as observed in section [ sec : examples - pos - def ] .
the next three sections give examples of the main results , beginning here with some simple , specific examples .
[ eg : naive ] first consider the naive model @xmath70 , in which different species are deemed to be entirely dissimilar . as noted in section [ sec : measures ] , @xmath71 is the exponential of the rnyi entropy of order @xmath0 .
it is well - known that rnyi entropy of any order @xmath1 is maximized uniquely by the uniform distribution .
this result also follows trivially from corollary [ cor : irrelevance ] : for clearly @xmath200 is uniquely maximized by the uniform distribution , and the corollary implies that the same is true for all values of @xmath1 .
moreover , @xmath201 .
[ eg : three ] for a general matrix @xmath19 satisfying , a two - species system is always maximized by the uniform distribution @xmath202 . when @xmath203 , however , nontrivial examples arise .
for instance , take the system shown in figure [ fig : three ] , consisting of one species of newt and two species of frog .
let us first consider intuitively what we expect the maximizing distribution to be , then compare this with the answer given by theorem [ thm : comp ] .
( 120,16)(0,-1 ) ( 0,7)(0,0)[l]@xmath204 ( 41,-1.5 ) ( 48,7)(0,0)[c]newt ( 117,12)(0,0)[r]frog species a ( 117,2)(0,0)[r]frog species b ( 108.5,7)(0,0)[l]0.9 ( 73,12)(0,0)[c]0.4 ( 73,2)(0,0)[c]0.4 if we ignore the fact that the two frog species are more similar to each other than they are to the newt , then ( as in example [ eg : naive ] ) the maximizing distribution is @xmath6 . at the other extreme ,
if we regard the two frog species as essentially identical then effectively there are only two species , newts and frogs , so the maximizing distribution gives relative abundance @xmath205 to the newt and @xmath205 to the frogs .
so with this assumption , we expect diversity to be maximized by the distribution @xmath206 .
intuitively , then , the maximizing distribution should lie between these two extremes . and
indeed , it does : implementing the algorithm in remark [ rmk : algorithm ] ( or using proposition [ propn : pos - def - max ] below ) reveals that the unique maximizing distribution is @xmath207 .
one of our standing hypotheses on @xmath19 is symmetry .
the last of our simple examples shows that if @xmath19 is no longer assumed to be symmetric , then the main theorem fails in every respect .
[ eg : nonsym ] let @xmath208 , which satisfies all of our standing hypotheses except symmetry .
consider a distribution @xmath209 . if @xmath24 is @xmath210 or @xmath164 then @xmath211 for all @xmath0 .
otherwise , @xmath212 from it follows that @xmath213 .
however , this supremum is not attained ; @xmath214 as @xmath215 , but @xmath216 .
equations and imply that @xmath217 with unique maximizing distributions @xmath218 and @xmath219 respectively .
thus , when @xmath19 is not symmetric , the main theorem fails comprehensively : the supremum @xmath220 may not be attained ; there may be no distribution maximizing @xmath221 for all @xmath0 simultaneously ; and that supremum may vary with @xmath0 . perhaps surprisingly ,
nonsymmetric similarity matrices @xmath19 do have practical uses .
for example , it is shown in ( * ? ? ?
* proposition a7 ) that the mean phylogenetic diversity measures of chao , chiu and jost @xcite are a special case of the measures @xmath23 , obtained by taking a particular @xmath19 depending on the phylogenetic tree concerned .
this @xmath19 is usually nonsymmetric , reflecting the asymmetry of evolutionary time .
more generally , the case for dropping the symmetry axiom for metric spaces was made in @xcite , and gromov has argued that symmetry ` unpleasantly limits many applications ' @xcite .
so the fact that our maximization theorem fails for nonsymmetric @xmath19 is an important restriction .
consider those matrices @xmath19 for which each similarity coefficient @xmath47 is either @xmath49 or @xmath50 .
a matrix @xmath19 of this form amounts to a ( finite , undirected ) reflexive graph with vertex - set @xmath28 , with an edge between @xmath21 and @xmath22 if and only if @xmath222 .
( that is , @xmath19 is the * adjacency matrix * of the graph . )
our standing hypotheses on @xmath19 then imply that @xmath223 for all @xmath21 , so every vertex has a loop on it ; this is the meaning of * reflexive*. what is the maximum diversity of the adjacency matrix of a graph ? to state the answer , we recall some terminology . vertices @xmath224 and @xmath225 of a graph are * adjacent * , written
@xmath226 , if there is an edge between them .
( in particular , every vertex of a reflexive graph is adjacent to itself . )
a set of vertices is * independent * if no two distinct vertices are adjacent .
the * independence number * @xmath227 of a graph @xmath228 is the maximal cardinality of an independent set of vertices of @xmath228 .
[ propn : graph - main ] let @xmath228 be a reflexive graph with adjacency matrix @xmath19 . then the maximum diversity @xmath186 is equal to the independence number @xmath227
. we will maximize the diversity of order @xmath30 and apply theorem [ thm : main ] .
for any probability distribution @xmath24 on the vertex - set @xmath28 , we have @xmath229 first we show that @xmath230 .
choose an independent set @xmath107 of maximal cardinality , and define @xmath126 by @xmath231 for each @xmath65 , the sum on the right - hand side of equation is @xmath232 .
hence @xmath233 , and so @xmath234 .
now we show that @xmath235 .
let @xmath126 .
choose an independent set @xmath236 with maximal cardinality among all independent subsets of @xmath61 .
then every vertex of @xmath61 is adjacent to at least one vertex in @xmath107 , otherwise we could adjoin it to @xmath107 to make a larger independent subset .
hence @xmath237 so there exists @xmath125 such that @xmath238 , where @xmath239 denotes the cardinality of @xmath107 .
but @xmath240 , so @xmath241 as required .
[ rmk : graph - proof ] the first part of the proof ( together with corollary [ cor : irrelevance ] ) shows that a maximizing distribution can be constructed by taking the uniform distribution on some independent set of largest cardinality , then extending by zero to the whole vertex - set . except in the trivial case
@xmath70 , this maximizing distribution never has full support .
we return to this point in section [ sec : elim ] .
[ eg : graph - lin3 ] the reflexive graph @xmath242 ( loops not shown ) has adjacency matrix @xmath243 .
the independence number of @xmath228 is @xmath76 ; this , then , is the maximum diversity of @xmath19 .
there is a unique independent set of cardinality @xmath76 , and a unique maximizing distribution , @xmath244 .
[ eg : graph - lin4 ] the reflexive graph @xmath245 again has independence number @xmath76 .
there are three independent sets of maximal cardinality , so by remark [ rmk : graph - proof ] , there are at least three maximizing distributions , @xmath246 all with different supports .
( the possibility of multiple maximizing distributions was also observed in the case @xmath183 by pavoine and bonsall @xcite . ) in fact , there are further maximizing distributions not constructed in the proof of proposition [ propn : graph - main ] , namely , @xmath247 and @xmath248 for any @xmath249 .
let @xmath55 be a metric on @xmath250 .
for a given @xmath251 , the * covering number * @xmath252 is the minimum cardinality of a subset @xmath253 such that @xmath254 where @xmath255 . the number @xmath256 is known as the * @xmath257-entropy * of @xmath55 @xcite .
now define a matrix @xmath258 by @xmath259 then @xmath258 is the adjacency matrix of the reflexive graph @xmath228 with vertices @xmath250 and @xmath260 if and only if @xmath261 .
thus , a subset of @xmath262 is independent in @xmath228 if and only if @xmath263 for every @xmath264 .
it is a consequence of the triangle inequality that @xmath265 and so by proposition [ propn : graph - main ] , @xmath266 recalling that @xmath267 extends the classical notion of rnyi entropy , this thoroughly justifies the name of @xmath257-entropy ( which was originally justified by vague analogy ) .
the moral of the proof of proposition [ propn : graph - main ] is that by performing the simple task of maximizing diversity of order @xmath30 , we automatically maximize diversity of all other orders . here is an example of how this can be exploited .
recall that every graph @xmath228 has a * complement * @xmath268 , with the same vertex - set as @xmath228 ; two vertices are adjacent in @xmath268 if and only if they are not adjacent in @xmath228 .
thus , the complement of a reflexive graph is irreflexive ( has no loops ) , and vice versa .
a set @xmath107 of vertices in an irreflexive graph @xmath269 is a * clique * if all pairs of distinct elements of @xmath107 are adjacent in @xmath269 .
the * clique number * @xmath270 of @xmath269 is the maximal cardinality of a clique in @xmath269 .
thus , @xmath271 .
we now recover a result of berarducci , majer and novaga ( * ? ? ?
* proposition 5.10 ) .
[ cor : capacity ] let @xmath269 be an irreflexive graph .
then @xmath272 where the supremum is over probability distributions @xmath24 on the vertex - set of @xmath269 , and the sum is over pairs of adjacent vertices of @xmath269 .
write @xmath28 for the vertex - set of @xmath269 , and @xmath19 for the adjacency matrix of the reflexive graph @xmath273 .
then for all @xmath169 , @xmath274hence by theorem [ thm : main ] and proposition [ propn : graph - main ] , @xmath275 it follows from this proof and remark [ rmk : graph - proof ] that @xmath276 can be maximized as follows : take the uniform distribution on some clique in @xmath269 of maximal cardinality , then extend by zero to the whole vertex - set .
[ rmk : no - quick - clique ] proposition [ propn : graph - main ] implies that computationally , finding the maximum diversity of an arbitrary @xmath19 is at least as hard as finding the independence number of a reflexive graph .
this is a very well - studied problem , usually presented in its dual form ( find the clique number of an irreflexive graph ) and called the maximum clique problem @xcite .
it is @xmath277-hard , so on the assumption that @xmath278 , there is no polynomial - time algorithm for computing maximum diversity , nor even for computing the support of a maximizing distribution .
the theory of magnitude of metric spaces runs most smoothly when the matrices @xmath19 concerned are positive definite @xcite .
we will see that positive ( semi)definiteness is also an important condition when maximizing diversity .
any positive definite matrix is invertible and therefore has a unique weighting .
( a positive semidefinite matrix need not have a weighting at all . ) now the crucial fact about magnitude is : [ lemma : pos - def - variational ] let @xmath111 be a positive semidefinite @xmath18 real matrix admitting a weighting .
then @xmath279 if @xmath111 is positive definite then the supremum is attained by exactly the nonzero scalar multiples @xmath37 of the unique weighting on @xmath111 .
this is a small extension of ( * ? ? ?
* proposition 2.4.3 ) .
choose a weighting @xmath112 on @xmath111 . by the cauchy
schwarz inequality , @xmath280 or equivalently @xmath281 for all @xmath35 .
equality holds when @xmath37 is a scalar multiple of @xmath112 , and if @xmath111 is positive definite , it holds only then .
finally , taking @xmath282 in and using positive semidefiniteness gives @xmath283 .
from this , we deduce : [ lemma : pos - def - sub ] let @xmath284 .
if @xmath19 is positive semidefinite and both @xmath19 and @xmath103 admit a weighting , then @xmath285 .
moreover , if @xmath19 is positive definite and the unique weighting on @xmath19 has full support , then @xmath286 .
the first statement follows from lemma [ lemma : pos - def - variational ] and the fact that @xmath103 is positive semidefinite .
the second is trivial if @xmath287 . assuming not ,
let @xmath288 be the unique weighting on @xmath103 ( which is positive definite ) , and write @xmath35 for the extension of @xmath289 by zero to @xmath28 .
then @xmath290 , @xmath291 , and @xmath292 but @xmath37 does not have full support , so by hypothesis , it is not a scalar multiple of the unique weighting on @xmath19 .
hence by lemma [ lemma : pos - def - variational ] , @xmath293 .
we now apply this result on magnitude to the maximization of diversity .
[ propn : pos - def - max ] suppose that @xmath19 is positive semidefinite . if @xmath19 has a nonnegative weighting @xmath112 , then @xmath294 and @xmath295 is a maximizing distribution .
moreover , if @xmath19 is positive definite and its unique weighting @xmath112 is positive then @xmath295 is the unique maximizing distribution .
this follows from theorem [ thm : comp ] and lemma [ lemma : pos - def - sub ] .
in particular , if @xmath19 is positive semidefinite and has a nonnegative weighting , then its maximum diversity can be computed in polynomial time .
[ cor : pos - def - support ]
if @xmath19 is positive definite with positive weighting , then its unique maximizing distribution has full support . in other words , when @xmath19 has these properties , its maximizing distribution eliminates no species . here
are three classes of such matrices @xmath19 .
[ eg : ultra ] call @xmath19 * ultrametric * if @xmath296 for all @xmath297 and @xmath298 for all @xmath21 .
( under the assumptions on @xmath19 , the latter condition just states that distinct species are not completely similar . )
if @xmath19 is ultrametric then @xmath19 is positive definite with positive weighting , by ( * ? ? ?
* proposition 2.4.18 ) .
such matrices arise in practice : for instance , @xmath19 is ultrametric if it is defined from a phylogenetic or taxonomic tree as in examples [ eg : sim - phylo ] and [ eg : sim - taxo ] .
let @xmath299 be a probability distribution of full support , and write @xmath19 for the diagonal matrix with entries @xmath300 .
then for @xmath301 , @xmath302 the right - hand side is the * rnyi relative entropy * or * rnyi divergence * @xmath303 ( * ? ? ?
* section 3 ) . evidently @xmath19 is positive definite , and its unique weighting @xmath304 is positive .
( in fact , @xmath19 is ultrametric . )
so proposition [ propn : pos - def - max ] applies ; in fact , it gives the classical result that @xmath305 with equality if and only if @xmath306 .
[ eg : id - pos - def ] the identity matrix @xmath70 is certainly positive definite with positive weighting . by topological arguments ,
there is a neighbourhood @xmath307 of @xmath53 in the space of symmetric matrices such that every matrix in @xmath307 also has these properties .
( see the proofs of ( * ? ? ? * propositions 2.2.6 and 2.4.6 ) . )
quantitative versions of this result are also available . for instance , in ( * ? ? ?
* proposition 2.4.17 ) it was shown that @xmath19 is positive definite with positive weighting if @xmath223 for all @xmath21 and @xmath308 for all @xmath309 .
in fact , this result can be improved : suppose that @xmath223 for all @xmath310 and that @xmath19 is * strictly diagonally dominant * ( that is , @xmath311 for all @xmath21 ) .
then @xmath19 is positive definite with positive weighting . since @xmath19 is real symmetric , it is diagonalizable with real eigenvalues . by the hypotheses on @xmath19 and
the gershgorin disc theorem ( * ? ? ?
* theorem 6.1.1 ) , every eigenvalue of @xmath19 is in the interval @xmath312 .
it follows that @xmath19 is positive definite and that every eigenvalue of @xmath313 is in @xmath314 .
hence @xmath315 is similar to a diagonal matrix with entries in @xmath314 , and so @xmath316 converges to @xmath317 .
thus , @xmath318 writing @xmath319 , the unique weighting on @xmath19 is @xmath320 .
the hypotheses on @xmath19 imply that @xmath321 has nonnegative entries and @xmath322 has positive entries .
hence by , @xmath323 entrywise , and so @xmath112 is positive .
thus , a matrix @xmath19 that is ultrametric , or satisfies and is strictly diagonally dominant , has many special properties : the maximum diversity is equal to the magnitude , there is a unique maximizing distribution , the maximizing distribution has full support , and both the maximizing distribution and the maximum diversity can be computed in polynomial time .
we saw in examples [ eg : graph - lin3 ] and [ eg : graph - lin4 ] that for certain similarity matrices @xmath19 , none of the maximizing distributions has full support .
mathematically , this simply means that maximizing distributions sometimes lie on the boundary of @xmath144 . but ecologically , it may sound shocking : is it reasonable that diversity can be increased by _ eliminating _ some species ?
we argue that it is .
consider , for instance , a forest consisting of one species of oak and ten species of pine , with each species equally abundant .
suppose that an eleventh species of pine is added , again with equal abundance ( figure [ fig : pine ] ) .
this makes the forest even more heavily dominated by pine , so it is intuitively reasonable that the diversity should decrease . but now running time backwards , the conclusion is that if we start with a forest containing the oak and all eleven pine species , eliminating the eleventh should _ increase _ diversity .
( 110,31 ) ( 55,-1)(0,0)[b ] ) and ten species of pine ( @xmath324 ) , to which one further species of pine is then added ( @xmath325 ) .
distances between species indicate degrees of dissimilarity ( not to scale).,title="fig : " ] ( 11.2,24.5)(0,0)[c]oak ( 87,28.5)(0,0)[c]pine to clarify further , recall from section [ sec : model ] that diversity is defined in terms of the _ relative _ abundances only . thus , eliminating species @xmath21 causes not only a decrease in @xmath184 , but also an increase in the other relative abundances @xmath326 .
if the @xmath21th species is particularly ordinary within the community ( like the eleventh species of pine ) , then eliminating it makes way for less ordinary species , resulting in a more diverse community . the instinct that maximizing diversity should not eliminate any species
is based on the assumption that the distinction between species is of high value .
( after all , if two species were very nearly identical or in the extreme , actually identical then losing one would be of little importance . ) if one wishes to make that assumption , one must build it into the model .
this is done by choosing a similarity matrix @xmath19 with a low similarity coefficient @xmath47 for each @xmath309 .
thus , @xmath19 is close to the identity matrix @xmath53 ( assuming that similarity is measured on a scale of @xmath49 to @xmath50 ) .
example [ eg : id - pos - def ] guarantees that in this case , there is a unique maximizing distribution and it does not , in fact , eliminate any species .
( the fact that maximizing distributions can eliminate some species has previously been discussed in the ecological literature in the case @xmath183 ; see pavoine and bonsall @xcite and references therein . )
we now derive necessary and sufficient conditions for a similarity matrix @xmath19 to admit at least one maximizing distribution of full support , and also necessary and sufficient conditions for _ every _ maximizing distribution to have full support .
the latter conditions are genuinely more restrictive ; for instance , if @xmath327 then some but not all maximizing distributions have full support . [ lemma : elim - main ] if at least one maximizing distribution for @xmath19 has full support then @xmath19 is positive semidefinite and admits a positive weighting . moreover , if every maximizing distribution for @xmath19 has full support then @xmath19 is positive definite .
fix a maximizing distribution @xmath24 of full support . maximizing distributions
are invariant ( corollary [ cor : max - invt ] ) , so by ( [ part : invt - invt])@xmath132([part : invt - ext ] ) of lemma [ lemma : invt ] , @xmath328 is a weighting of @xmath19 and @xmath329 .
in particular , @xmath19 has a positive weighting .
now we imitate the proof of proposition 3b of @xcite . for each @xmath330 such that @xmath331 , define a function @xmath332 by @xmath333 using the symmetry of @xmath19 and the fact that @xmath328 is a weighting , we obtain @xmath334 now @xmath335 and @xmath24 has full support , so @xmath336 for all real @xmath85 sufficiently close to zero . but @xmath337 for such @xmath85 , so @xmath338 has a local minimum at @xmath49 . hence @xmath339 .
it follows that @xmath338 is everywhere positive .
we have shown that @xmath339 whenever @xmath340 with @xmath335 . now take @xmath35 with @xmath341 .
put @xmath342 .
then @xmath335 , and @xmath343 hence @xmath19 is positive semidefinite . for ` moreover ' , assume that every maximizing distribution for @xmath19 has full support . by ,
we need only show that @xmath344 whenever @xmath345 with @xmath335 .
given such an @xmath346 , choose @xmath347 such that @xmath348 lies on the boundary of @xmath144 .
then @xmath348 does not have full support , so is not maximizing , so does not maximize @xmath349 ( by corollary [ cor : irrelevance ] ) .
hence @xmath350 , which by implies that @xmath344 .
we can now prove the two main results of this section .
[ propn : semi ] the following are equivalent : 1 .
[ part : semivec ] there exists a maximizing distribution for @xmath19 of full support ; 2 .
[ part : semidef ] @xmath19 is positive semidefinite and admits a positive weighting .
( [ part : semivec])@xmath132([part : semidef ] ) is the first part of lemma [ lemma : elim - main ] . for the converse , assume ( [ part : semidef ] ) and
choose a positive weighting @xmath112 . then @xmath351 , so @xmath352 is a probability distribution of full support .
we have @xmath353 for all @xmath0 , by lemma [ lemma : invt ] .
but the computation theorem implies that @xmath354 for some @xmath102 such that @xmath103 admits a weighting , so @xmath355 by lemma [ lemma : pos - def - sub ] .
hence @xmath24 is maximizing .
the following are equivalent : 1 .
[ part : vec ] every maximizing distribution for @xmath19 has full support ; 2 .
[ part : unique ] @xmath19 has exactly one maximizing distribution , which has full support ; 3 .
[ part : def ] @xmath19 is positive definite with positive weighting ; 4 .
[ part : crit ] @xmath356 for every nonempty proper subset @xmath107 of @xmath357 .
( the weak inequality @xmath358 holds for any @xmath19 , by the absent species lemma ( lemma [ lemma : abs ] ) . ) ( [ part : vec])@xmath132([part : def ] ) and ( [ part : def])@xmath132([part : unique ] ) are immediate from lemma [ lemma : elim - main ] and proposition [ propn : pos - def - max ] respectively , while ( [ part : unique])@xmath132([part : vec ] ) is trivial . for ( [ part : vec])@xmath132([part : crit ] ) , assume ( [ part : vec ] ) .
let @xmath359 .
choose a maximizing distribution @xmath25 for @xmath103 , and denote by @xmath24 its extension by zero to @xmath28 .
then @xmath24 does not have full support , so there is some @xmath10 $ ] such that @xmath24 fails to maximize @xmath29 .
hence @xmath360 where the second equality is by the absent species lemma . for ( [ part : crit])@xmath132([part : vec ] ) ,
assume ( [ part : crit ] ) .
let @xmath24 be a maximizing distribution for @xmath19 .
write @xmath133 , and denote by @xmath25 the restriction of @xmath24 to @xmath107 . then for any @xmath0 , @xmath361 again by the absent species lemma .
hence by ( [ part : crit ] ) , @xmath362 .
the main theorem , the computation theorem and corollary [ cor : irrelevance ] answer all the principal questions about maximizing the diversities @xmath29 . nevertheless , certain questions remain .
first , there are computational questions .
we have found two classes of matrix @xmath19 for which the maximum diversity and maximizing distributions can be computed in polynomial time : ultrametric matrices ( example [ eg : ultra ] ) and those close to the identity matrix @xmath53 ( example [ eg : id - pos - def ] ) . both are biologically significant . are there other classes of similarity matrix for which the computation can be performed in less than exponential time ?
second , we may seek results on maximization of @xmath23 under constraints on @xmath24 .
there are certainly some types of constraint under which both parts of theorem [ thm : main ] fail , for trivial reasons : if we choose two distributions @xmath24 and @xmath25 whose diversity profiles cross ( figure [ fig : main](b ) ) and constrain our distribution to lie in the set @xmath363 , then there is no distribution that maximizes @xmath29 for all @xmath0 simultaneously , and the maximum value of @xmath29 also depends on @xmath0 . but are there other types of constraint under which the main theorem still holds ? in particular , the distribution might be constrained to lie close to a given distribution @xmath24 .
the question then becomes : if we start with a distribution @xmath24 and have the resources to change it by only a given small amount , what should we do in order to maximize the diversity ?
third , we have seen that every symmetric matrix @xmath19 satisfying ( for instance , every symmetric matrix of positive reals ) has attached to it a real number , the maximum diversity @xmath186 .
what is the significance of this invariant ?
we know that it is closely related to the magnitude of matrices .
this has been most intensively studied in the context of metric spaces . by definition ,
the magnitude of a finite metric space @xmath269 is the magnitude of the matrix @xmath364 ; see @xcite , for instance . in the metric context ,
the meaning of magnitude becomes clearer after one extends the definition from finite to compact spaces ( which is done by approximating them by finite subspaces ) .
magnitude for compact metric spaces has recognizable geometric content : for example , the magnitude of a 3-dimensional ball is a cubic polynomial in its radius ( * ? ? ?
* theorem 2 ) , and the magnitude of a homogeneous riemannian manifold is closely related to its total scalar curvature ( * ? ? ?
* theorem 11 ) .
thus , it is natural to ask : can one extend theorem [ thm : main ] to some class of ` infinite matrices ' @xmath19 ? ( for instance , @xmath19 might be the form @xmath365 arising from a compact metric space . in this case , the maximum diversity of order @xmath76 is a kind of capacity , analogous to classical definitions in potential theory ; for a compact subset of @xmath366 , it coincides with the bessel capacity of an appropriate order @xcite . ) and if so , what is the geometric significance of maximum diversity in that context ?
there is already evidence that this is a fruitful line of enquiry . in @xcite , meckes gave a definition of the maximum diversity of order @xmath76 of a compact metric space , and used it to prove a purely geometric theorem relating magnitude to fractional dimensions of subsets of @xmath366 . if this maximum diversity can be shown to be equal to the maximum diversity of all other orders then further geometric results may come within reach . the fourth and final question concerns interpretation .
throughout , we have interpreted @xmath23 in terms of ecological diversity . however , there is nothing intrinsically biological about any of our results .
for example , in an information - theoretic context , the ` species ' might be the code symbols , with two symbols seen as similar if one is easily mistaken for the other ; or if one wishes to transmit an image , the ` species ' might be the colours , with two colours seen as similar if one is an acceptable substitute for the other ( much as in rate distortion theory @xcite ) . under these or other interpretations ,
what is the significance of the theorem that the diversities of all orders can be maximized simultaneously ?
we thank christina cobbold , ciaran mccreesh and richard reeve for helpful discussions .
this work was supported by the carnegie trust for the universities of scotland , the centre de recerca matemtica , an epsrc advanced research fellowship , the national institute for mathematical and biological synthesis , the national science foundation , and the simons foundation .
a. jeziorski , a. j. tanentzap , n. d. yan , a. m. paterson , m. e. palmer , j. b. korosi , j. a. rusak , m. t. arts , w. keller , r. ingram , a. cairns , and j. p. smol .
the jellification of north temperate lakes .
, 282:20142449 , 2015 .
a. rnyi . on measures of entropy and information . in j.
neymann , editor , _
4th berkeley symposium on mathematical statistics and probability _ , volume 1 , pages 547561 .
university of california press , 1961 .
l. wang , m. zhang , s. jajodia , a. singhal , and m. albanese . modeling network diversity for evaluating the robustness of networks against zero - day attacks . in
_ proceedings of the 19th european symposium on research in computer security ( esorics 2014 ) _ , pages 494511 , 2014 . | entropy , under a variety of names , has long been used as a measure of diversity in ecology , as well as in genetics , economics and other fields .
there is a spectrum of viewpoints on diversity , indexed by a real parameter @xmath0 giving greater or lesser importance to rare species .
leinster and cobbold @xcite proposed a one - parameter family of diversity measures taking into account both this variation and the varying similarities between species .
because of this latter feature , diversity is not maximized by the uniform distribution on species .
so it is natural to ask : which distributions maximize diversity , and what is its maximum value ? + in principle , both answers depend on @xmath0 , but our main theorem is that neither does .
thus , there is a single distribution that maximizes diversity from all viewpoints simultaneously , and any list of species has an unambiguous maximum diversity value .
furthermore , the maximizing distribution(s ) can be computed in finite time , and any distribution maximizing diversity from some particular viewpoint @xmath1 actually maximizes diversity for all @xmath0 .
+ although we phrase our results in ecological terms , they apply very widely , with applications in graph theory and metric geometry . |
in this section we discuss in more detail the derivation and the validity of the effective continuum model given in eq .
( 2 ) in the main text . for this analysis
we assume that the motion of atoms is described by classical trajectories @xmath135 , which is well justified for the parameter regime considered in this work . under this assumption the kinetic term appearing in eq .
( 1 ) in the main text is constant and the dynamics of the atoms can be reduced to their internal degrees of freedom . to achieve slow - light conditions we consider a periodically modulated waveguide with lattice constant @xmath0 and total length @xmath136 .
the hamiltonian describing the coupling of the two - level atoms and the 1d field in the waveguide can then be written in the general form @xmath137 @xmath138.\ ] ] by assuming periodic boundary conditions the field operator is given by @xmath139 where @xmath140 are bosonic annihilation operators obeying @xmath141=\delta_{bb ' } \delta_{kk'}$ ] and @xmath142 are bloch waves with band index @xmath143 and wavevector @xmath144 $ ] , @xmath145 .
the @xmath146 are periodic functions .
the slow - light limit that we are interested in corresponds to the tight - binding limit in electronic band structures . in this limit
the frequency separation between neighboring bands , @xmath147 , is much bigger than the variation of frequencies within one band . in this limit
the dispersion relation of a single band can be approximated by @xmath148 . by assuming in addition that also all the relevant frequency scales of the system dynamics ( see below ) are small compared to @xmath149
, we can restrict our analysis to a single band with @xmath150 . by omitting the band index we then obtain the single band model @xmath151 where @xmath152 . note that compared to the main text , we have considered here a system with a large , but finite number of lattice sites @xmath153 and @xmath154 is given in units of frequency .
therefore , eq . can be directly used for numerical simulations .
the limit @xmath155 can be obtained in the usual way by replacing @xmath156 and @xmath157 . due to their motion with constant velocity @xmath158 ,
the atoms induce a time periodic modulation of the coupling with frequency @xmath159 .
this modulation makes the dynamics of moving atoms coupled to slow - light waveguides in general quite involved . to obtain a simpler description , which mimics closely the coupling of moving atoms to photons in a continuous waveguide , we focus in this work primarily on the case of very fast atoms where @xmath160 . in this regime
the modulation is fast compared to the interaction timescale @xmath154 and we can replace the modulated couplings @xmath161 by the coupling @xmath31 averaged over one period , @xmath162 . to be more precise , let us assume that at time @xmath33 the @xmath163-th atom is located at the left border of one of the unit cells , i.e. , @xmath164 , and the atom then traverses @xmath165 lattice sites .
in the interaction picture with respect to the bare atom and photon hamiltonian , the integrated coupling between the atom and a photonic mode @xmath13 is then given by @xmath166 where @xmath167 .
we now use the fact that in the tight - binding limit , @xmath168 where @xmath169 is an approximately @xmath15-independent mode functions of one unit cell
. then , @xmath170 \times \sum_{n = n_1}^{n_1+\ell-1 } e^{i k a n } e^{i\delta_k t_i ( n+1/2 ) } .\ ] ] to proceed we neglect the variation @xmath171 in the integral in eq . .
this is valid for small detunings , @xmath172 , but also for strongly peaked @xmath169 . under this approximation
we can introduce the average coupling @xmath173 and obtain @xmath174 we see that this result is just the discretized version of the integral @xmath175 which we would obtain for the same quantity evaluated for the continuum model @xmath176 which is just the finite - length version of the effective model given in eq .
( 2 ) in the main text .
although in the derivation of eq .
several crude approximations have been involved , a numerical comparison with the full model given in eq .
shows that for all the parameters considered in this work very accurate qualitative and quantitative predictions with errors on the percent level can be obtained .
however , as we discuss in the following , there are also regimes , where such an effective description breaks down .
let us now discuss in more detail the range of validity for the continuum description .
to do so we consider the evolution of a state in the single excitation subspace , which we write as @xmath177|g\rangle|{\rm vac}\rangle$ ] . in the interaction picture with respect
to @xmath178 we obtain the following set of equations of motion @xmath179 where @xmath180 is the detuning of the atoms from the tilted band and @xmath181 . after inserting the second equation into
the first we obtain the following integro - differential equation @xmath182 the @xmath183 are periodic in time with period @xmath184 and can be expanded in a fourier series @xmath185 .
the resulting expression can be written as @xmath186^ * e^{ik(z_i - z_j ) } e^{i(\tilde\delta_k^i-\tilde \delta_k^j)t } e^{i(\omega_i n-\omega_j m ) t } \int_0^t dt ' e^{i\tilde\delta_k^j(t - t ' ) } e^{i\omega_j m ( t - t ' ) } c_e^j(t ' ) \\ & = -\frac{g^2a}{l } \sum_j \sum_k u^0_{ki } [ u^{0}_{kj}]^ * e^{ik(z_i - z_j ) } e^{i(\tilde\delta_k^i-\tilde \delta_k^j)t } \int_0^t dt ' e^{i\tilde\delta_k^j(t - t ' ) } c_e^j(t ' ) \ , + \ , { \rm rest}. \end{split}\ ] ] when evaluated under the same stroboscopic approximation as above , the first term in the second line of this equation simply corresponds to the evolution under the effective continuum model given in eq . .
a first set of corrections arises from terms with @xmath187 , but @xmath188 .
such terms lead to additional oscillating contributions to @xmath189 . however , as long as @xmath190 , these corrections remain small and do not affect the long time behavior of @xmath189 .
more crucial are terms with @xmath191 , which can generate non - oscillating contributions that affect the slowly varying dynamics of @xmath189 . to see this
more explicitly let us consider a single atom and a specific correction term @xmath192 c_e(t ) , \end{split}\ ] ] where in the second line we have made a markov approximation and replaced the integral over @xmath193 by a @xmath50-function in frequency . from this estimate
we see that the higher order correction terms can contribute additional decay channels with rates @xmath194 proportional to the density of states evaluated at the modulated frequency @xmath195 . to ensure that such corrections do not affect the dynamics , all the frequencies @xmath196 must lie outside the tilted propagation band , as illustrated in fig . [ figsup1 ]
this condition can be recast into the form @xmath197 where @xmath40 .
the boundary derived from this inequality is represented by the blue line in fig .
[ figsup1 ] a ) . and
the blue continuous line shows the boundary derived from condition , while the black dashed line represents the boundary derived from numerical simulations . on the right hand side of this line
the predictions for @xmath53 obtained from the full model and the effective model agree better than 10% .
see text for more details .
the dotted red lines indicte the lower and the upper edge of the tilted dispersion relation @xmath47 .
b)-d ) illustration of the higher order resonances shifted by @xmath198 for b ) @xmath199 and @xmath42 , c ) @xmath200 and @xmath201 and d ) @xmath202 and @xmath203.,scaledwidth=60.0% ] it turns out that under certain conditions eq .
is too strict and the continuum model can be applied even at lower velocities .
this is the case for @xmath7 tuned close to the lower band edge . as shown in fig .
[ figsup1 ] c ) , in this case the @xmath204 sideband can lie within the propagation band , but since the density of states near the band edge is much higher , the resulting correction is still negligible compared to effective evolution .
this explains why quasi - bound states near the lower band edge can be observed at finite velocities . however , since @xmath205 is non - zero , these bound states still decay on longer timescales .
to verify the validity of our analytic estimates we can compare the predictions from the effective model with a numerical simulation of the full model , which accounts for the exact time - dependence of the couplings .
for these numerical simulations we consider the specific example , where the local mode at each lattice site is represented by a gaussian wavepacket @xmath206 with a width @xmath207 .
this results in an average coupling @xmath208 for the simulation of the full model we assume @xmath209 and @xmath210 .
this corresponds to @xmath211 . for the cases : a ) @xmath41 and @xmath43 , b ) @xmath41 and @xmath44 and c ) @xmath45 and @xmath46 .
d ) decay of the excited state population corresponding to the previous cases .
e ) excitation transfer between three atoms with @xmath45 and @xmath46 . in all the plots
@xmath209 and @xmath210 is assumed . ,
scaledwidth=100.0% ] by using the parameters given above we compare in fig .
[ fig : validity ] the predictions of the effective model and the full model for some of the values of @xmath50 and @xmath51 assumed in fig .
( 2 ) and fig . ( 3 ) in the main text .
for all the cases we see an excellent agreement between the exact and the effective evolution , in particular for the evolution of @xmath53 .
further , to verify the validity condition given in eq .
above , we evaluate the evolution for the excited state probability @xmath53 up to a final time @xmath212 for a wide range of parameters @xmath50 and @xmath51 .
then , the discrepancy parameter @xmath213 can be used to quantify the validity of the effective model . in fig .
[ figsup1 ] a ) we plot the boundary set by @xmath214 .
we see that for most parameters it follows indeed the prediction from eq . .
however , near @xmath215 the region of validity extends to very low velocities , which can be explained by the qualitative arguments given above .
we also see an additional kick at @xmath216 and @xmath203 . as shown in fig .
[ figsup1 ] d ) , in this case also the @xmath204-sideband hits a divergency in the photonic density of states and therefore induces a significant perturbation .
in this section we derive the exact dynamics of a single atom moving at the maximum speed of light @xmath45 .
as explained in the main text , under this condition there appears a higher order divergence in the co - moving photonic density of states at a frequency @xmath217 . for the following analytic results we assume @xmath218 and approximate the dispersion relation by an infinite cubic band , i.e. , @xmath219 , where @xmath77 is the wave vector corresponding to the frequency @xmath220 . in the co - moving frame
the hamiltonian for a single atom reads @xmath221 where @xmath222 is the initial position of the atom that from now on , without loss of generalization , we will fix to : @xmath223 .
having approximated the dispersion as cubic , all the eigenstates of the system are scattering states with eigenvalues in the continuum labelled by the wavevector @xmath15 , i.e. , @xmath224 .
a generic eigenstate in the single - excitation sector of this hamiltonian has the form @xmath225|g\rangle|{\rm vac}\rangle$ ] and by plugging this ansatz into the schrdinger equation , @xmath226 , we obtain the coupled equations @xmath227 as previously mentioned the photonic wave function @xmath228 is a scattering state thus we need to use the lippmann - schwinger equation that in momentum representation reads @xcite @xmath229 where we used @xmath224 . in this way
is possible to write the atomic amplitude as @xmath230 where the integral @xmath231 is defined as @xmath232e^{-i\frac{\pi}{6 } { \text{sgn}}[(k - k_c)z]}}{ja^2(k - k_c)^2}\\ & \times \left(e^ { i\frac{(k - k_c)z}{2}-\frac{\sqrt{3}}{2}|(k - k_c)z|}+e^{i\frac{2}{3}\pi { \text{sgn}}[(k - k_c)z]-i(k - k_c)z}\right ) , \end{split}\ ] ] where @xmath233 is the sign function . in the evaluation of the previous integral
was taken the limit @xmath234 assuming @xmath235 for @xmath236 and @xmath237 for @xmath238 in order to have a convergent result .
the generic eigenstate can be written in momentum space as @xmath239|g\rangle|{\rm vac}\rangle\mbox{.}\ ] ] in real space the eigenstates read @xmath240|g\rangle|{\rm vac}\rangle,\ ] ] where the atomic and photonic amplitudes are given by @xmath241-i(k - k_c)z}\right)\right],\ ] ] @xmath242}\left[1+e^{i\frac{2}{3}\pi { \text{sgn}}[(k - k_c)z]}\right]},\ ] ] @xmath243}}{(\tilde\omega_k-\bar \delta)ja^ 2(k - k_c)^2+\bar g^2{\text{sgn}}(k - k_c)e^{-i\frac{\pi}{6 } { \text{sgn}}[(k - k_c)z]}\left[1+e^{i\frac{2}{3}\pi { \text{sgn}}[(k - k_c)z]}\right]}.\ ] ] the derivation of the eigenstates for multiple atoms can be obtained similarly . using these exact solutions we can investigate the spontaneous emission of an initially excited atom .
the time evolution of the atom - field dressed state @xmath244 can be expressed in terms of the eigenstates of the system according to @xmath245 from that is possible evaluate the emitted photonic wave packet shown in fig . 2 d ) of the main text , @xmath246 the probability of finding the atom in the excited state as function of time can also be computed by evaluating @xmath247 and is given by the following expression @xmath248^ 2+\frac{\eta^ 4}{4}}\right|^2,\ ] ] where the dimensionless parameters @xmath249 , @xmath250 and @xmath251 were introduced .
to obtain a simpler approximate result we can instead consider the laplace transform of the atomic amplitude @xmath252 where @xmath253 by assuming to be exactly on resonance with the flat part of the band , @xmath254 , we can rewrite @xmath255 as @xmath256 where in the last we made a single pole approximation . under this approximation
the inverse laplace transform of @xmath255 results in a damped cosine function as given in eq .
( 4 ) of the main text .
the coefficients @xmath257 represent the characteristic oscillation frequency and the damping rate , respectively .
in the main text we describe the implement of a slow - light waveguide qed system using rydberg atoms flying above a superconducting coplanar waveguide ( cpw ) resonator array @xcite .
the geometry of the resonator is shown in fig .
( 4 ) of the main text . for a transition dipole moment @xmath258 ( where @xmath259 is the bohr radius ) between two neighboring rydberg levels @xmath260 and @xmath261 , the atom - field coupling strength is @xmath262 where @xmath263 is the electric field of the resonator mode per photon . in a quasi 1d waveguide resonator
the electric field is of the form @xmath264\cos{\left(\frac{m\pi x}{l_x}\right)},\ ] ] where @xmath265 is the length of the resonator and the same coordinate as in fig
. 4 in the main text have been assumed . to estimate the transverse electric field profile we consider the limit of an infinitely flat cpw resonator of length @xmath265 and an additional ground plate located at a distance @xmath266 below the central track . according to ref .
@xcite , if the width of the two external ground planes is bigger than twice the gap between them , they can be treated as infinitely extended . with this assumption , an exact solution for the electric field above the central electrode of the cpw resonator
can be obtained using conformal mapping technics @xcite .
in particular , the two transversal field component are given by @xmath267 and @xmath268 , where @xmath269 with @xmath270 and @xmath271 related to the size of the middle electrode and to the ground planes separation [ see fig .
4b ) in the main text ] . in the previous expression we introduced the electric field per photon , @xmath272 , where @xmath273 is the vacuum dielectric constant and @xmath274 the mode volume of the resonator given by @xmath275.\ ] ] for the parameters considered in the main text and @xmath276 m , the volume below the central electrode is negligible compare to the mode volume above and only the latter
must be considered for the evaluation of the field strength @xmath277 .
, @xmath278 a)-b ) and @xmath279 , @xmath280 c)-d ) . in picture a)-c ) is shown the excited state population for the cases @xmath281 ( continuous line ) , @xmath282 ( long dashed ) , @xmath283 ( small dashed ) and @xmath284 ( dotted ) . in picture b)-d ) are shown the comparison of the emitted photon wavepackets ( evaluated at @xmath285 ) for the cases with ( dashed ) and without disorder ( continuous ) .
in all the plot is assumed @xmath54 .
, scaledwidth=100.0% ] periodic photonic structures are usually subject to imperfections and disorder in the lattice fabrication .
this results in random frequency offset between neighboring lattice sites , which in turn can lead to localization . we can model the effect of such disorder by adding a random frequency offset @xmath286 for each local resonator mode , i.e. , @xmath287 . in momentum space this corresponds to the hamiltonian @xmath288 where @xmath289 . for numerical simulations we chose the @xmath286 randomly from a uniform distribution in the interval @xmath290 $ ] . in fig .
[ fig : disorder ] a ) and b ) we plot the atomic population and the wave function of the emitted photon for the case @xmath291 and @xmath43 and for different values of the disorder strength @xmath292 . from the plots we see that the system dynamics is almost unaffected by the disorder up to values of about @xmath293 . above this value
the signatures of a co - moving bound state are washed out . in fig .
[ fig : disorder ] c ) and d ) we plot the analog results for the critical coupling conditions @xmath279 and @xmath46 . here
we find that the atomic decay is even less affected by disorder .
however , for @xmath294 the emitted wavefunction is becomes again significantly distorted , which would affect the dynamics of excitation transfer processes . in conclusion
, we find that all the effects described in the main text are robust with respect to disorder up to a strength of the order of @xmath295 . in the cpw resonators array implementation
this would corresponds a variation in the frequency of each resonator of about @xmath296 mhz .
this accuracy is achievable with the state of the art fabrication technology @xcite . for further discussions on the effect of disorder in coupled resonator waveguides
see also ref .
@xcite and the supplementary material of @xcite . | we describe the coupling of moving atoms to a one dimensional photonic waveguide in the regime where the atomic velocities are comparable to the effective speed of light .
such conditions could be achieved , for example , in photonic crystals or coupled resonator arrays , where the maximal photonic group velocity is significantly reduced compared to free space . in this case
the interplay between a velocity - induced directionality and the emergence of effective divergencies in the photonic density of states gives rise to a range of novel phenomena and non - perturbative effects in the emission of photons and the resulting dipole - dipole interactions between moving atoms .
we show that apart from potential nanophotonic implementations , rydberg atoms flying above a coupled array of superconducting microwave resonators provide a versatile platform for exploring this new regime of pseudo - relativistic atom - light interactions under accessible experimental conditions .
the radiation of charged particles moving close to the speed of light is a long - studied problem of classical electrodynamics @xcite with practical relevance for accelerator physics or the development of intense x - ray sources @xcite . in contrast , the coupling of neutral atoms and molecules to classical or quantized fields is usually investigated under opposite conditions , where particle velocities are low and mainly affect the emission and absorption of photons via a linear doppler shift .
this effect , however , is the basis for laser cooling @xcite and successive atom trapping techniques @xcite , which enabled the study of strong light - matter interactions even at the single photon level @xcite . based on the same techniques , it has recently become possible to interface cold atoms with one dimensional ( 1d ) waveguides @xcite and photonic crystal structures @xcite ,
inside which the propagation velocity of photons can be significantly reduced @xcite .
this opens up new possibilities for exploring an intriguing regime of atom - light interactions , where individual atoms and photons move at comparable velocities , while still interacting strongly with each other .
the coupling of atoms or other emitters @xcite to 1d waveguides , photonic crystals or coupled - cavity arrays has already attracted considerable attention . in this context , many interesting phenomena like single and multi - photon scattering @xcite , atom - photon bound states @xcite , self - organization @xcite and quantum many - body effects @xcite , or unidirectional emission @xcite have been explored , usually under the assumption of stationary or slowly moving emitters . in this work
we go beyond this conventional scenario and study atom - photon interactions in slow - light waveguides under conditions , where the atomic velocities approach , match or even exceed the effective speed of light . in particular , we show how a bounded group velocity introduces different types of directionalities and novel non - perturbative effects in photon emission and photon - transfer processes .
we further describe a new realization of a slow - light waveguide qed system consisting of rydberg atoms flying across a coupled array of planar microwave resonators @xcite .
this setting provides a versatile experimental platform for exploring such pseudo - relativistic light - matter interactions and study , for example , classical effects like cherenkov radiation @xcite in a quantum optical context , where each individual photon is interacting strongly with the moving atoms .
are strongly coupled to the evanescent field of a 1d waveguide . by introducing a sufficiently strong spatial modulation with periodicity
@xmath0 , photons are slowed down by bragg - reflection and the usual linear dispersion relation shown in b ) separates into individual bands of finite width @xmath1 , as indicated in c ) . within one band
the maximal group velocity is bounded by an effective speed of light , @xmath2 , which can be comparable to typical atomic velocities . , scaledwidth=48.0% ] _ model._we consider a setup as shown in fig .
[ fig : setup ] a ) , where @xmath3 two - level atoms are coupled to the field of a 1d photonic waveguide .
the atoms have a mass @xmath4 and the internal ground state @xmath5 and excited state @xmath6 are separated in frequency by @xmath7 .
restricted to the atomic motion along the waveguide direction , @xmath8 , the generic hamiltonian for this system reads @xmath9 here the @xmath10 and @xmath11 are the atomic position and momentum operators , @xmath12 , and the @xmath13 , where @xmath14=\delta(k - k')$ ] , are the bosonic annihilation operators for photons with wavevector @xmath15 . in conventional waveguides and within the relevant frequency range , @xmath16 and @xmath17 , where @xmath18 is the speed of light in vacuum and @xmath19
is the refractive index of the material .
therefore , in this case the photonic group velocity , @xmath20 m / s , is much larger than typical atomic velocities , @xmath21 m / s . to achieve a condition @xmath22
, we consider in this work a waveguide with a spatially periodic modulation , for example , of its width , refractive index , etc .
provided that this modulation is sufficiently strong , the propagation of photons is slowed down by bragg - reflection @xcite and the spectrum separates into individual bands with an approximate dispersion relation @xmath23 , where @xmath0 is the lattice constant and @xmath1 is the total width of the band .
this introduces an effective speed of light , @xmath24 , corresponding to the maximal propagation velocity within the band .
it can be tuned close to zero by taking the limiting case of a weakly tunnel - coupled array of individual resonators , making the regime @xmath25 experimentally accessible .
it is important to keep in mind that by following this approach one changes from a continuum to a bloch - wave description and introduces as well a periodic variation of the coupling , @xmath26 . for atoms moving along classical trajectories @xmath27
, this translates into a time - periodic modulation of the coupling and makes the dynamics of slow - light systems in general quite involved .
however , for a large range of parameters , in particular for the relevant case of high velocities , @xmath28 , and @xmath29 within the propagation band , we can derive an effective model @xmath30 where only the coupling averaged over one unit cell , @xmath31 , appears .
this model now mimics very closely the coupling of fast moving atoms to arbitrarily slow photons in a 1d continuum and will be used as a starting point for the following analysis .
details on the derivation of eq . and
its range of validity are summarized in the supplementary material @xcite .
_ photon emission in the quasi - relativistic regime._we first consider the emission of a photon from a single atom moving at a constant velocity @xmath32 . by assuming that at time @xmath33 the atom is in state @xmath6 and the waveguide in the vacuum state @xmath34
, the state of the whole system can be written as @xmath35|g\rangle|{\rm vac}\rangle$ ] , where @xmath36 is the excited state population and @xmath37 is the wavefunction of the emitted photon in @xmath15-space .
figure [ fig : emission ] a ) shows snapshots of the resulting wavefunction in the discretized position space , @xmath38 , for different ratios @xmath39 and different atom detunings @xmath40 .
and for the parameters @xmath41 , @xmath42 , @xmath41 , @xmath43 , @xmath41 , @xmath44 and @xmath45 , @xmath46 .
the insets show the shape of the tilted dispersion relation , @xmath47 , and the blue shaded area indicates the light cone , @xmath48 .
b ) plot of the directionality parameter @xmath49 as function of @xmath50 and @xmath51 .
the solid red lines indicate the minimal and maximal frequency of the tilted band . for each detuning
, the dashed line shows the minimal velocity @xmath52 , above which the effective model provides an accurate description of the dynamics @xcite .
c ) decay of the excited state population @xmath53 for different sets of parameters .
the dashed line indicates the approximate analytic solution given in eq . for @xmath45 and @xmath46 .
in all plots a coupling of @xmath54 is assumed . ,
scaledwidth=48.0% ] to obtain a better understanding of these emission patterns , it is convenient to change into a co - moving frame via the unitary transformation @xmath55 , where @xmath56 . in this new representation the time - dependence in the exponentials is eliminated , @xmath57 , but we obtain instead a tilted dispersion relation @xmath58 . for frequencies @xmath7 inside this tilted band and for sufficiently small @xmath31 , the coupling to the waveguide will result in an approximately exponential decay of @xmath53 with total rate @xmath59 , where @xmath60 here @xmath61 and @xmath62 are left- and right - propagating wavevectors defined by the resonance condition @xmath63 , and @xmath64 is the group velocity in the co - moving frame . in the usual case , where @xmath65 and @xmath66 , eq .
implies that photons are emitted with doppler - shifted wavevectors @xmath67 , but with the same group velocity in the laboratory frame and approximately the same rates @xmath68 .
in the quasi - relativistic regime @xmath28 , this picture changes significantly . as illustrated in example in fig .
[ fig : emission ] a ) , the tilting of the finite propagation band opens a frequency window , within which only backward propagating modes exist .
this gives rise to a purely unidirectional emission opposite to the atomic motion .
a different mechanism for directionality is shown in example , where instead forward emission is enhanced near a divergence of the photonic density of states , i.e. @xmath69 , associated with the tilted right band edge .
the combination of these effects can give rise to strong variations in the direction , the rate and the group velocity of the emitted photons as a function of the atomic velocity and detuning .
this is summarized in fig .
[ fig : emission ] b ) , where we plot the directionality parameter @xmath70 @xcite .
note that for this plot we have assumed a finite decay rate @xmath71 for the photons , which allows us to generalize eq .
to frequencies outside the band and to avoid unphysical divergencies @xcite . the exponential decay with rates @xmath72 given in eq .
assumes a perturbative treatment of the atom - light coupling , which is strictly valid only for @xmath73 . for static atoms
it is well known that even for @xmath74 , the divergencies in the photonic density of states near the band - edges result in non - perturbative effects , like an oscillatory decay @xcite and the emergence of atom - photon bound states @xcite .
examples and in fig .
[ fig : emission ] a ) show that such effects prevail even at very high velocities .
a large fraction of the photon is coherently reabsorbed and remains bound to the moving atom for a long time .
note , however , that for moving atoms true bound states only exist within the effective continuum model .
zeros in the modulated coupling @xmath75 in the full model will eventually lead to a complete decay of the atom @xcite .
_ photon emission at the speed of light._in addition to the above mentioned band edge effects , which are already present at low atomic velocities @xcite , a new and very unique situation occurs when the atomic velocity matches the speed of light , @xmath45 ( see example in fig .
[ fig : emission ] ) . in this case one obtains a cubic dispersion relation @xmath76 for wavevectors around @xmath77 .
this results in a higher order divergency @xmath78 in the effective photon density of states , as a consequence of the strong interactions between the moving atom and the co - propagating cherenkov photon .
however , in contrast to usual band - edge effects , this velocity - induced singularity lies in the middle of the band and for moderate @xmath31 there is no bound state associated with it . as a result
, the atom approximately decays as @xcite @xmath79 where coherent atom - photon oscillations and decay occur with the same characteristic rate @xmath80 . for a given @xmath74 ,
this is also the fastest rate at which the atom can be completely deexcited by emitting a highly localized photonic wave packet , which closely follows the atom in forward direction . at even higher velocities , @xmath81 , the density of states , and therefore also the emission rate , decreases again . as shown in fig .
[ fig : emission ] b ) , the directionality of the emitted photon is then only determined by the condition @xmath82 . for @xmath83 an extended photon , which is essentially stationary in the laboratory frame ,
is produced .
and relative separations @xmath84 and @xmath85 are considered .
the parameters in a ) are @xmath41 and @xmath42 and in b ) @xmath45 and @xmath46 . for the plot in c ) two atoms with @xmath86 , slightly different velocities @xmath87 and @xmath88 and a detuning of @xmath44 have been assumed . in all plots @xmath54.,scaledwidth=48.0% ] _ excitation transfer processes._compared to conventional cherenkov radiation studied in higher dimensional structures @xcite , a key feature of the current setting is that all emitted photons are confined to one dimension and can thus be efficiently reabsorbed by other atoms . for conventional waveguides , this emission and reabsorption of photons
gives rise to almost instantaneous , long - range and bidirectional dipole - dipole interactions , which depend only on the relative atomic positions , @xmath89 @xcite . in the quasi - relativistic regime
various qualitatively different excitation transfer processes arise , depending strongly on the velocities and detunings of the involved atoms . to illustrate this point we first consider in fig .
[ fig : transfer ] a ) and b ) the case of three atoms moving at the same speed @xmath32 , with the first atom to the right being initially prepared in the excited state . at rest the maximal population transfer to atoms 2 and 3 would be limited by the fraction of population trapped in the sub - radiant state @xcite , and @xmath90 .
this limit no longer applies at moderate and fast velocities , where the emission becomes highly directional and therefore also results in a much more efficient transfer of excitations from the right to the left .
note that for the parameters assumed in fig .
[ fig : transfer ] a ) , the retardation time @xmath91 is still short compared to the timescale of the transfer dynamics @xmath92 and the whole process is well described by a unidirectional markovian master equation @xcite in the co - moving frame .
this is no longer the case for the critical condition @xmath45 and @xmath93 assumed in fig .
[ fig : transfer ] b ) . here
we observe a very rapid transfer of the initial excitation from atom 1 to atom 2 , which is mediated by the highly compressed cherenkov photon shown in example in fig .
[ fig : emission ] a ) .
this photon then slowly falls behind and at a much later time overlaps with atom 3 .
the resulting oscillation between the photon and atom 3 with roughly the same rabi frequency @xmath94 as identified in eq . , are again a clear signature of the highly non - markovian nature of the transfer process . finally , in fig .
[ fig : transfer ] c ) we consider another example , where two initially co - propagating atoms exchange excitations via a co - moving bound photon , which remains exponentially localized around the atomic positions @xmath95 . by introducing a small velocity difference , i.e. @xmath96 , the atoms slowly separate and gradually decouple . depending on the detailed choice of parameters ,
the photon may remain with one of the atoms or be split into two separated atom - photon bound states @xcite .
_ implementation._while the approach for realizing the slow - light hamiltonian is very generic and could potentially be implemented with optimized nanophotonic crystal structures , the conditions for reaching the pseudo - relativistic regime are still very challenging . in particular , local disorder must be small compared to @xmath1 to avoid photon localization @xcite , and fast moving atoms must be guided very close to the surface such that the emission into the waveguide exceeds the remaining decay into free space . to overcome these problems we propose here an equivalent system in the microwave regime , where rydberg atoms @xcite fly above a coupled array of superconducting transmission line resonators , as shown fig .
[ fig : implementation ] a ) .
the resonators of length @xmath97 cm exhibit standing wave modes with frequencies @xmath98 in the ghz regime , while in @xmath99-direction the electric field is strongly confined to a few tens of @xmath100 m determined by the size of the middle electrode , @xmath101 [ see fig . [
fig : implementation ] b ) ] . by considering an array of parallel resonators separated by @xmath102 m
, we obtain a closely spaced resonator array with a tunnel coupling @xmath103 that can be fully adjusted by additional capacitive couplings at the end - points @xcite . above an array of parallel cpw resonators coupled by capacitances @xmath104 .
b ) cross section of a single resonator and sketch of the transverse field distribution .
c ) atom - field coupling strength as a function of the atom - surface distance @xmath105 for the example described in the text .
here @xmath106 denotes the maximal coupling evaluated at @xmath107 and @xmath31 the corresponding average coupling , which is used in the effective continuum model .
see ref .
@xcite for more details.,scaledwidth=48.0% ] as a specific example , we consider a microwave resonator with dimensions @xmath108 cm , @xmath109 m , @xmath110 m and @xmath111 ghz . by identifying two rydberg states with neighboring principle quantum numbers @xmath112 ,
i.e. @xmath113 and @xmath114 , we obtain a resonant coupling @xmath115 for @xmath116 .
the resulting maximal atom - field coupling @xmath117 @xcite , where @xmath118 is the transition dipole moment and @xmath119 the electric field per photon , as well as the corresponding averaged coupling , @xmath31 , are plotted in fig .
[ fig : implementation ] c ) as a function of the atom surface distance , @xmath105 . for an atomic beam of velocity @xmath120
m / s @xcite we choose a capacitive coupling @xmath121 mhz to obtain @xmath122 and still being larger than the typical level of on - site disorder @xcite .
therefore , already at convenient distances of @xmath123
m we obtain @xmath124 and for the validity of the effective model we can chose a smaller value of @xmath125 , by adjusting the position of the beam along the @xmath126-direction . on the relevant timescale set by @xmath127 khz , both the decay rate of the rydberg states , @xmath128 khz , and the photon loss rate @xmath129 khz of a high - q microwave resonator @xcite are still negligible .
note that for an array of @xmath130 resonators with a total length of @xmath131 cm , the condition on the transverse spread during the flight time , @xmath132 , requires a transverse cooling of the atom beam to a few mk , or alternatively , on - chip guiding systems for rydberg atoms @xcite .
this shows that microwave - based slow - light systems are well within reach of existing experimental capabilities . _
conclusions._in summary , we have described atom - photon interactions in the pseudo - relativistic regime , where the atomic velocities are comparable to the maximal speed of light in a waveguide structure . under such conditions
the atomic motion drastically changes photon emission and transfer processes and induces a strong directionality and novel non - perturbative effects .
importantly , we have shown that rydberg atoms flying above a coupled array of microwave resonator provide a versatile platform , where this combination of quasi - relativistic and strong - coupling quantum optical effects is experimentally accessible . _
acknowledgments._the authors thank d. chang , p. schneeweiss , j. volz , a. rauschenbeutel and f. ciccarello for valuable feedback .
this work was supported by the austrian science fund ( fwf ) through sfb foqus f40 , dk coqus w 1210 and the start grant y 591-n16 .
99 j. d. jackson , _ classical electrodynamics _ ( john wiley @xmath133 sons , 1962 ) . z. huang and k. j. kim , _ review of x - ray free - electron laser theory _ , phys . rev .
st accel .
beams * 10 * , 034801 ( 2007 )
. s. stenholm , _ the semiclassical theory of laser cooling _
, rev .
mod .
phys . * 58 * , 699 ( 1986 ) .
h. j. metcalf and p. van der straten , _ laser cooling and trapping _ ( springer , 1999 ) .
k. m. birnbaum , a. boca , r. miller , a. d. boozer , t. e. northup , and h. j. kimble , _ photon blockade in an optical cavity with one trapped atom _ , nature ( london ) * 436 * , 87 ( 2005 ) .
t. g. tiecke , j. d. thompson , n. p. de leon , l. r. liu , v. vuletic , and m. d. lukin , _ nanophotonic quantum phase switch with a single atom _ , nature ( london ) * 508 * , 241 ( 2014 ) .
j. volz , m. scheucher , c. junge , and a. rauschenbeutel , _ nonlinear @xmath134 phase shift for single fibre - guided photons interacting with a single resonator - enhanced atom _ , nat .
photon .
* 8 * , 965 ( 2014 ) .
b. hacker , s. welte , g. rempe , and s. ritter , _ a photon - photon quantum gate based on a single atom in an optical resonator _
, nature ( london ) * 536 * , 193 ( 2016 ) .
d. reitz , c. sayrin , r. mitsch , p. schneeweiss , and a. rauschenbeutel , _ coherence properties of nanofiber - trapped cesium atoms _ , phys .
rev .
lett . * 110 * , 243603 ( 2013 ) .
r. yalla , m. sadgrove , k. p. nayak and k. hakuta , _ cavity quantum electrodynamics on a nanofiber using a composite photonic crystal cavity _ , phys .
rev .
lett . * 113 * , 143601 ( 2014 ) .
j. d. thompson , t. g. tiecke , n. p.
de leon , j. feist , a. v. akimov , m. gullans , a. s. zibrov , v. vuletic , and m. d. lukin , _ coupling a single trapped atom to a nanoscale optical cavity _
, science * 340 * , 1202 ( 2013 ) .
a. goban , c .-
. hung , s. p.
yu , j. d. hood , j. a. muniz , j. h. lee , m. j. martin , a. c. mcclung , k. s. choi , d. e. chang , o. painter , and h. j. kimble , _ atom - light interactions in photonic crystals _ , nat . commun . * 5 * , 3808 ( 2014 ) .
t. f. krauss , _ slow light in photonic crystal waveguides _
, j. phys .
d : appl .
phys . * 40 * , 2666 ( 2007 ) .
p. lodahl , s. mahmoodian , and s. stobbe , _ interfacing single photons and single quantum dots with photonic nanostructures _ ,
rev .
mod .
phys . * 87 * , 347 ( 2015 ) .
j. t. shen and s. fan , _ strongly correlated two - photon transport in a one - dimensional waveguide coupled to a two - level system _ , phys .
rev .
lett . * 98 * , 153003 ( 2007 ) .
l. zhou , z. r. gong , y. x. liu , c. p. sun , and f. nori , _ controllable scattering of a single photon inside a one - dimensional resonator waveguide _ ,
phys .
rev .
lett .
* 101 * , 100501 ( 2008 ) .
t. shi and c. p. sun , _ lehmann - symanzik - zimmermann reduction approach to multiphoton scattering in coupled - resonator arrays _
, phys .
rev .
b * 79 * , 205111 ( 2009 ) .
p. longo , p. schmitteckert , and k. busch , _ few - photon transport in low - dimensional systems : interaction - induced radiation trapping _ , phys .
rev .
lett . * 104 * , 023602 ( 2010 ) .
p. bykov , _ spontaneous emission from a medium with a band spectrum _ , sov .
j. quant .
electron . * 4 * , 7 ( 1975 ) .
s. john and j. wang , _ quantum electrodynamics near a photonic band gap : photon bound states and dressed atoms _
, phys .
rev .
lett . * 64 * , 2418 ( 1990 ) . s. john and t. quang , _ spontaneous emission near the edge of a photonic band gap _ ,
phys .
rev .
a * 50 * , 1764 ( 1994 ) .
a. g. kofman , g. kurizki , and b. sherman , _ spontaneous and induced atomic decay in photonic band structures _ , j. mod
. opt . * 41 * , 353 ( 1994 ) .
p. lambropoulos , g. m. nikolopoulos , t. r. nielsen , and s. bay , _ fundamental quantum optics in structured reservoirs _ ,
rep .
. phys . * 63 * , 455 ( 2000 ) .
f. lombardo , f. ciccarello , and g. m. palma , _ photon localization versus population trapping in a coupled - cavity array _ ,
phys .
rev . a * 89 * , 053826 ( 2014 ) .
g. calajo , f. ciccarello , d. chang and p. rabl , _ atom - field dressed states in slow - light waveguide qed _ , phys .
rev .
a * 93 * , 033833 ( 2016 ) .
t. shi , y - h .
wu , a. gonzalez - tudela and j. i. cirac , _ bound states in boson impurity models _ , phys .
rev .
x * 6 * , 021027 ( 2016 ) .
y. liu and a. houck , _ quantum electrodynamics near a photonic bandgap _ , nature phys .
( 2016 ) ; doi:10.1038/nphys3834 .
d. e. chang , j. i. cirac , and h. j. kimble , _ self - organization of atoms along a nanophotonic waveguide _
, phys .
rev .
lett . * 110 * , 113606 ( 2013 ) .
t. grieer and h. ritsch , _ light - induced crystallization of cold atoms in a 1d optical trap _ , phys .
rev .
lett . * 111 * , 055702 ( 2013 ) .
m. j. hartmann , f. g. s. l. brando and m. b. plenio , _ strongly interacting polaritons in coupled arrays of cavities _
, nature phys .
* 2 * , 849 ( 2006 ) .
a. d. greentree , c. tahan , j. h. cole , and c. l. hollenberg , _ quantum phase transitions of light _ , nature phys .
* 2 * , 856 ( 2006 ) .
d. g. angelakis , m. f. santos , and s. bose , _ photon - blockade - induced mott transitions and xy spin models in coupled cavity arrays _ , phys .
rev .
a * 76 * , 031805(r ) ( 2007 ) .
j. petersen , j. volz , and a. rauschenbeutel , _ chiral nanophotonic waveguide interface based on spin - orbit interaction of light _
, science * 346 * , 67 ( 2014 ) .
r. mitsch , c. sayrin , b. albrecht , p. schneeweiss , and a. rauschenbeutel , _ quantum state - controlled directional spontaneous emission of photons into a nanophotonic waveguide _ , nat . commun . * 5 * , 5713 ( 2014 ) .
i. sllner , _ et al .
_ , _ deterministic photon - emitter coupling in chiral photonic circuits _ , nature nanotech . * 10 * , 775 ( 2015 ) .
k. stannigel , p. rabl and p. zoller , _ driven - dissipative preparation of entangled states in cascaded quantum - optical networks _
, new j. phys .
* 14 * , 063014 ( 2012 ) .
p. lodahl , s. mahmoodian , s. stobbe , p. schneeweiss , j. volz , a. rauschenbeutel , h. pichler , p. zoller , _ chiral quantum optics _ , arxiv:1608.00446 ( 2016 ) .
d. petrosyan and m. fleischhauer , _ quantum information processing with single photons and atomic ensembles in microwave coplanar waveguide resonators _ , phys . rev . lett . * 100 * , 170501 ( 2008 ) . s. d. hogan , j. a. agner , f. merkt , t. thiele , s. filipp , and a. wallraff , _ driving rydberg - rydberg transitions from a coplanar microwave waveguide _ , phys
. rev .
lett . * 108 * , 063004 ( 2012 ) . c. hermann - avigliano , r. celistrino teixeira , t. l. nguyen , t. cantat - moltrecht , g. nogues , i. dotsenko , s. gleyzes , j. m. raimond , s. haroche , and m. brune , _ long coherence times for rydberg qubits on a superconducting atom chip _ , phys . rev . a * 90 * , 040502(r ) ( 2014 ) .
m. a. beck , j. a. isaacs , d. booth , j. d. pritchard , m. saffman , r. mcdermott , _ optimized coplanar waveguide resonators for a superconductor - atom interface _ , appl .
. lett . * 109 * , 092602 ( 2016 ) .
d. underwood , w. e. shanks , j. koch , and a. a. houck , _ low disorder microwave cavity lattices for quantum simulation with photons _ , phys .
rev .
a * 86 * , 023837 ( 2012 ) .
p. a. cherenkov , _ visible radiation produced by electrons moving in a medium with velocities exceeding that of light _ , dokl .
akad .
nauk * 2 * , 451 ( 1934 ) [ phys . rev . *
52 * , 378 ( 1937 ) ] . j. v. jelly , _ cerenkov radiation and its applications _
( pergamon , london , 1958 ) . c. luo , m. ibanescu , s. g. johnson , j. d. joannopoulos , _ cerenkov radiation in photonic crystals _
, science * 299 * , 5605 ( 2003 ) . c. luo , m. ibanescu , e. j. reed , s. g. johnson , j. d. joannopoulos , _ doppler radiation emitted by an oscillating dipole moving inside a photonic band - gap crystal _ , phys .
rev .
lett . * 96 * , 043903 ( 2006 ) .
see the supplemental material for more detailed derivations of the main results presented . c. gonzalez - ballestero ,
e. moreno , f.j .
vidal , a. gonzalez - tudela , _ non - reciprocal few - photon devices based on chiral waveguide - emitter couplings _ , phys .
rev .
a 94 , * 94 * , 063817 ( 2016 ) .
d. e. chang , l. jiang , a. v. gorshkov , and h. j. kimble , _ cavity qed with atomic mirrors _ , new j. phys .
* 14 * , 063003 ( 2012 ) .
a. gonzlez - tudela and d. porras , _ mesoscopic entanglement induced by spontaneous emission in solid - state quantum optics _ , phys .
rev .
lett . * 110 * , 080502 ( 2013 ) . c. gonzalez - ballestero , f. j. garcia - vidal and e. moreno , _ non - markovian effects in waveguide - mediated entanglement _
, new j. phys . * 15 * , 073015 ( 2013 ) .
s. longhi , _ bound states in the continuum in a single - level fano - anderson model _ , eur .
phys .
j. b * 57 * , 45 ( 2007 ) .
p. facchi , m. s. kim , s. pascazio , f. v. pepe , d. pomarico and t. tufarelli , _ bound states and entanglement generation in waveguide quantum electrodynamics _
, phys .
rev .
a * 94 * , 043839 ( 2016 ) .
j. c. hernndez - herrejn , f. m. izrailev and l. tessieri , _ anomalous localization in the aperiodic kronig - penney model _ journal of physics a : mathematical and theoretical * 43 * , 425004 ( 2010 ) .
l. frunzio , a. wallraff , d. schuster , j. majer , and r. schoelkopf , ieee trans .
. supercond . * 15 * , 860 ( 2005 ) .
p. lancuba and s. d. hogan , _ electrostatic trapping and in situ detection of rydberg atoms above chip - based transmission lines _ , j. phys .
b : at . mol .
. phys . * 49 * , 074006 ( 2016 ) . * supplemental material for : + atom - photon interactions at the speed of light * this supplemental material contains additional details on the model and the derivation of various results stated in the main part of the paper . |
a great deal of attention has been devoted to non - contact friction between nanostructures , including , for example , the frictional drag force between two - dimensional quantum wells @xcite , and the friction force between an atomic force microscope tip and a substrate @xcite . in non - contact friction
the bodies are separated by a potential barrier thick enough to prevent electrons or other particles with a finite rest mass from tunneling across it , but allowing interaction via the long - range electromagnetic field , which is always present in the gap between bodies .
the presence of inhomogeneous tip - sample electric fields is difficult to avoid , even under the best experimental conditions @xcite .
for example , even if both the tip and the sample were metallic single crystals , the tip would still have corners present and more than one crystallographic plane exposed .
the presence of atomic steps , adsorbates , and other defects will also contribute to the inhomogeneous electric field .
the electric field can be easily changed by applying a voltage between the tip and the sample . the electromagnetic field
can also be created by the fluctuating current density , due to thermal and quantum fluctuations inside the solids .
this fluctuating electromagnetic field is always present close to the surface of any body , and consist partly of traveling waves and partly of evanescent waves which decay exponentially with the distance away from the surface of the body .
the fluctuating electromagnetic field originating from the fluctuating current density inside the bodies gives rise to the well - known long - range attractive van der waals interaction between two bodies @xcite .
if the bodies are in relative motion , the same fluctuating electromagnetic field will give rise to a friction which is frequently named as the van der waals friction .
van der waals friction can be considered as mediated by photon exchange between the bodies : one body emit a photon , and the other absorbs it , thus transferring momentum between the bodies , resulting in a friction force . at large distances between the bodies , the main contribution to friction comes from photon exchange , corresponding to the propagating electromagnetic waves .
however this contribution is very small because the photons corresponding to propagating waves carry very small momentum , no larger than @xmath0 .
the photons , corresponding to the evanescent electromagnetic waves , carry the momentum @xmath1 .
thus for distances @xmath2 between two bodies smaller characteristic distance @xmath3 , which depends on temperature ( at room temperature @xmath4 ) , the main contribution to friction comes from the evanescent electromagnetic field . in analogy with electron tunneling , this mechanism of momentum transfer can be considered as associated with the photon tunneling .
although the dissipation of energy connected with the non - contact friction always is of electromagnetic origin , the detailed mechanism is not totally clear , since there are several different mechanisms of energy dissipation connected with the electromagnetic interaction between bodies .
first , the electromagnetic field from one body will penetrate into the other body , and induce an electric current . in this case friction
is due to ohmic losses inside the bodies .
the fluctuating electromagnetic field can also excite the vibrations of the adsorbates or other surface localized modes , e.g. surface plasmons and polaritons . in this case friction
is due to energy relaxation of the surface modes .
another contribution to friction from the electromagnetic field is associated with the time - dependent stress acting on the surface of the bodies .
this stress can excite acoustic waves , or induce time - dependent deformations which may result in a temperature gradient .
it can also induce motion of defects either in the bulk , or on the surface of the bodies .
the contribution to friction due to non - adiabatic heat flow , or motion of defects , is usually denoted as internal friction .
it is very worthwhile to get a better understanding of different mechanisms of non - contact friction because of it practical importance for ultrasensitive force detection experiments .
this is because the ability to detect small forces is inextricably linked to friction via the fluctuation - dissipation theorem .
for example , the detection of single spins by magnetic resonance force microscopy , which has been proposed for three - dimensional atomic imaging @xcite and quantum computation @xcite , will require force fluctuations to be reduced to unprecedented levels .
in addition , the search for quantum gravitation effects at short length scale @xcite and future measurements of the dynamical casimir forces @xcite may eventually be limited by non - contact friction effects .
recently gotsmann and fuchs @xcite observed long - range non - contact friction between an aluminum tip and a gold ( 111 ) surface .
the friction force @xmath5 acting on the tip is proportional to the velocity @xmath6 , @xmath7 .
for motion of the tip normal to the surface the friction coefficient @xmath8 , where @xmath2 is the tip - sample spacing and @xmath9 @xcite .
later stipe _ _ et.al.__@xcite observed non - contact friction effect between a gold surface and a gold - coated cantilever as a function of tip - sample spacing @xmath2 , the temperature @xmath10 , and the bias voltage @xmath11 . for vibration of the tip parallel to the surface they found @xmath12 , where @xmath13 and @xmath14 at 295for the spacing @xmath15 100they found @xmath16 which is @xmath17500 times smaller that reported in ref .
@xcite at the same distance using a parallel cantilever configuration . in a recent letter , dorofeev _ et.al .
_ @xcite claim that a the non - contact friction effect observed in @xcite is due to ohmic losses mediated by the fluctuating electromagnetic field .
this result is controversial , however , since the van der waals friction has been shown @xcite to be many orders of magnitude smaller than the friction observed by dorofeev _ et.al . _ presently , the origin of the difference in magnitude and distance dependence of the long - range non - contact friction effect observed in @xcite and @xcite is not well understood . in order to improve the basic understanding of non - contact friction
, we present new results for van der waals friction .
in @xcite we developed a theory of van der waals friction for surfaces in parallel relative motion .
here we generalize the theory to include also the case when the surfaces are in normal relative motion , and we show that there is drastic difference between these two cases .
thus , for normal relative motion of clean good conductor surfaces , the friction is many orders of magnitude larger than for parallel relative motion , but still smaller than observed experimentally .
another enhancement mechanism of the non - contact friction can be connected with resonant photon tunneling between states localized on the different surfaces .
recently it was discovered that resonant photon tunneling between surface plasmon modes give rise to extraordinary enhancement of the optical transmission through sub - wavelength hole arrays @xcite .
the same surface modes enhancement can be expected for van der waals friction if the frequency of these modes is sufficiently low to be excited by thermal radiation . at room temperature
only the modes with frequencies below @xmath18 can be excited . for normal metals surface plasmons
have much too high frequencies ; at thermal frequencies the dielectric function of normal metals becomes nearly purely imaginary , which exclude surface plasmon enhancement of the van der waals friction for good conductors .
however surface plasmons for semiconductors are characterized by much smaller frequencies and damping constants , and they can give an important contribution to van der waals friction .
other surface modes which can be excited by thermal radiation are adsorbate vibrational modes . especially for parallel vibrations
these modes may have very low frequencies .
all information about the long - range electromagnetic interaction between two non - contacting bodies is , in principle , contained in the reflection factors of the electromagnetic field . at present time
very little is known about the reflection factors for large wave vectors and for extremely small frequencies . in the calculations of the reflection factors one must take into account the non - local response of the electron gas on the external electromagnetic field .
there are two correlation length which determine this nonlocal response .
first is the skin depth , which determines the long - range length scale of the nonlocality in the volume , and the second is the screening length .
the latter length scale determines the short range correlation length for non - local response in the surface region . in our previous calculations of the van der waals friction
@xcite we mostly considered good conductors . in this case
it was shown that the important contribution comes from the non - local optic effects in the surface region .
however it was shown that the van der waals friction becomes much larger for high resistivity material , for which the volume contribution from non - local effects is also important .
it is easy to see that within local optic approximation the van der waals friction diverge when the conductivity of materials tend to zero .
this means that the local optic approximation breaks down for high - resistivity materials .
this situation is completely different from the heat transfer between bodies via photon tunneling @xcite , where the heat flux is maximal at conductivities corresponding to semi - metals . in order to clarify the situation we study the dependence of the van der waals friction on the dielectric properties of the materials within the non - local dielectric approach , which was proposed some years ago for the investigation of the anomalous skin effects @xcite .
we consider two semi - infinite metals * 1 * and * 2 * having parallel flat surfaces .
we introduce a coordinate system with @xmath19 plane in the surface of body * 1 * , and the @xmath20 axis along the upward normal .
the surface of body * 2 * is located at @xmath21 , performing small amplitude vibrations along the @xmath22 axes with displacement coordinate @xmath23 since the system is translation invariant in the @xmath24 plane , the electromagnetic field can be represented by the fourier integral @xmath25 where @xmath26 and @xmath27 are the electric and magnetic induction field , and @xmath28 is the two - dimensional wave vector in @xmath29 plane .
after the fourier transformation it is convenient to choose the coordinate axis in the ( x , y ) plane along the vectors @xmath28 and @xmath30 $ ] . in the vacuum gap between the bodies
the electric field @xmath31 , and the magnetic induction field @xmath32 , can , to the linear order in the vibrational coordinate , be written in the form @xmath33 @xmath34e^{-ipz}+ [ \mathbf{k}^{+}\times\mathbf{w}_{0}]e^{ipz}\right ) \right .
\nonumber \\ \left .
+ \frac{1}{\omega + \omega _ 0}\left ( [ \mathbf{k}^{-}_+\times\mathbf{v}_{1}]e^{-ip^+z}+ [ \mathbf{k}^{+}_+ \times\mathbf{w}_{1}e^{ip^+z}]\right)e^{-i\omega _ 0 t } \right ] e^{-i\omega t } \label{two}\end{aligned}\ ] ] where @xmath35 at the surfaces of the bodies the amplitude of the outgoing electromagnetic wave must be equal to the amplitude of the reflected wave plus the amplitude of the radiated wave .
it is convenient to decompose the electromagnetic field into the @xmath36- and @xmath37 - polarized electromagnetic waves . for @xmath36-polarized electromagnetic waves the electric field is in the incident plane determined by the vectors @xmath28 and @xmath38 , and for @xmath37-
polarized electromagnetic waves the electric field is normal to the incident thus the boundary conditions for the electromagnetic field at @xmath39 can be written in the form @xmath40 @xmath41 where @xmath42 is the reflection factor for surface * 1 * for @xmath43 - polarized electromagnetic field , and where @xmath44 are the components of the fluctuating electric field outside the surface * 1 * in the absence of the body * 2*. the boundary condition at the surface of the body * 2 * must be written in the reference frame where the body * 2 * is at rest .
the electric field in this reference frame is determined by a lorentz transformation . performing a lorentz transformation of the electric field to linear order in @xmath45 gives @xmath46 } c \label{five}\ ] ] for the @xmath36-polarized electromagnetic waves the second term in ( [ five ] ) is of the order of magnitude @xmath47 relative to the first one and
can be neglected for the most practical cases . however , for the @xmath37-polarized electromagnetic waves the second term is of the order of magnitude @xmath48 and can be of the same order of magnitude as the first term . in the rest frame of body * 2 * there is also mixture of @xmath37- and @xmath49 polarized electromagnetic waves . in @xcite
it was was shown that this gives contribution of the order @xmath50 and thus can be neglected . after performing lorentz transformation to linear order in @xmath51 and @xmath52 we get @xmath53 @xmath54 @xmath55 the boundary conditions for the electromagnetic field at @xmath56 in the rest frame of body * 2 * can be written in the form @xmath57
@xmath58 @xmath59 where @xmath60 is the reflection factor for surface * 2 * for @xmath43 - polarized electromagnetic field , and where @xmath61 are the components of the fluctuating electric field outside the surface * 1 * in the absence of the body * 1*. from ( [ three],[four ] ) and ( [ six]-[eight ] ) we get @xmath62 @xmath63 @xmath64 @xmath65 @xmath66 where @xmath67 and @xmath68 .
other components of the fluctuating electromagnetic field can be found from the transversality conditions @xmath69 the fundamental characteristic of the fluctuating electromagnetic field is the correlation function , determining the average product of components @xmath70 .
accordingly to the general theory of the fluctuating electromagnetic field ( see for a example @xcite ) these correlation function are given by @xmath71 \label{corfs}\ ] ] @xmath72 \label{corfp}\ ] ] where @xmath73 denote statistical average over the random field , and where the bose - einstein factor @xmath74 we note that @xmath36 is real for @xmath75 ( propagating waves ) , and purely imaginary for @xmath76 ( evanescent waves ) .
thus for @xmath75 and @xmath76 the correlation functions are determined by the first and the second terms in eqs .
( [ corfs])and ( [ corfp ] ) , respectively .
the frictional stress @xmath77 which act on the surfaces of the two bodies can be obtained from @xmath78 component of the maxwell stress tensor @xmath79 , evaluated at @xmath39 : @xmath80 @xmath81 @xmath82 _ { z=0 }
\label{3thirteen}\ ] ] to linear order in the vibrational coordinate @xmath83 and the frequency @xmath84 the stress acting on the surface * 1 * can be written in the form @xmath85 here the first term determines the conservative van der waals stress and the second term is the adiabatic change of the conservative van der waals stress during vibration .
the last term determines the frictional stress with friction coefficient @xmath86 . for normal relative motion ( see appendix a ) we obtain the friction coefficient @xmath87 , where the contribution to the friction coefficient from the propagating electromagnetic waves is given by @xmath88 @xmath89 @xmath90\frac 1{\left|1-e^{2ipd}r_{1p}r_{2p}\right|^4 } + [ p\rightarrow s ] , \label{3seventeen}\ ] ] and where the contribution to the friction from the evanescent electromagnetic waves is given by @xmath91 @xmath92 @xmath93 \frac 1{\left|1-e^{-2kd}r_{1p}r_{2p}\right|^4 } + [ p\rightarrow s ] , \label{3eighteen}\ ] ] where @xmath94 .
the symbol @xmath95 $ ] in eqs .
( [ 3seventeen ] ) and ( [ 3eighteen ] ) denotes the term which is obtained from the first one by replacement of the reflection factors @xmath96 for @xmath49 polarized waves , by the reflection factors @xmath97 for @xmath98 polarized waves .
the friction coefficient for two flat surfaces in parallel relative motion was obtained by us before , @xcite and can be written as @xmath99 , where the contribution to the friction coefficient from the propagating electromagnetic waves is given by @xmath100 @xmath101 , \label{3nineteen}\ ] ] and where the contribution to the friction from the evanescent electromagnetic waves is given by @xmath102 @xmath103 .
\label{3twenty}\ ] ] there is a principal difference between the friction coefficient for normal and parallel relative motion , related to the denominator in the formulas for the friction coefficient .
the resonant condition corresponds to the case when the denominator of the integrand in eqs .
( [ 3seventeen]-[3twenty ] ) , which is due to multiple scattering of the evanescent electromagnetic waves from the opposite surfaces , is small . for two identical surfaces and @xmath104 ,
where @xmath105 and @xmath106 are the imaginary and real part , respectively , this corresponds to the resonant condition @xmath107 . at resonance
the integrand in eqs .
( [ 3nineteen ] ) and ( [ 3twenty ] ) has a large factor @xmath108 , in sharp contrast to the case of parallel relative motion , where there is no such enhancement factor .
the resonance condition can be fullfiled even for the case when @xmath109 because for evanescent electromagnetic waves there is no restriction on the magnitude of the real part or the modulus of @xmath110 .
this open up the possibility of resonant denominators for @xmath111 . to estimate the friction coefficient @xmath112 for an atomic force microscope tip we can use an approximate formula @xcite @xmath113 where it is assumed that the tip has cylinder symmetry . here
@xmath114 denotes the tip - surface distance as a function of the distance @xmath115 from the tip symmetry axis , and the friction coefficient @xmath116 is determined by the expressions for the flat surfaces .
this scheme was proposed in @xcite for the calculation of the conservative van der waals interaction .
the error of these scheme is not larger than 5 - 10% in practice in an atomic force microscopy experiment , and 25% in a worst case satiation @xcite .
although this scheme was proposed for the conservative van der waals interaction , we assume that the same scheme is also valid for the calculation of the van der waals friction .
we assume that the tip has a paraboloid shape given [ in cylindrical coordinates ( @xmath117 ) ] by the formula:@xmath118 , where @xmath2 is the distance between the tip and the flat surface , and where @xmath110 is the radius of curvature of the tip . in the case of the power dependence @xmath119 we get @xmath120 in a more general case one must use numerical integration . in the local optic approximation , where the dielectric function is assumed to depend only on the frequency @xmath121 , the reflection factors @xmath122 and @xmath123 for flat surfaces , covered by an adsorbate layer , are given by @xcite : @xmath124 } { p+s/\epsilon-4\pi in_aq[s\alpha_{\parallel}/\epsilon+q \alpha_{\perp } ] } , \label{3twentyone}\ ] ] @xmath125 where @xmath126 and where @xmath127 and @xmath128 are the polarizabilities of adsorbates in a direction parallel and normal to the surface , respectively . here
@xmath129 is the bulk dielectric function and @xmath130 is the concentration of adsorbates . for clean surfaces @xmath131 , and in this case formulas ( [ 3twentytwo,3twentythree ] ) reduce to the well - known fresnel formula . at @xmath132 and @xmath133 , where @xmath134 is the electron mean free path , and where @xmath135 and @xmath136 are the fermi velocity and fermi wave number , respectively , the system will be characterized by non - local dielectric function @xmath137 . in this paper
we use the non - local optic dielectric approach , proposed some years ago for the investigations of the optical properties of a semi - infinite electron gas @xcite . accordingly to @xcite , the reflection factor for @xmath36 - polarized electromagnetic field , incident on the flat surface ,
is determined by @xcite @xmath138 where the surface impedance @xmath139 is given by @xmath140 where @xmath141 is the finite- life- time generalization of the longitudinal lindhard dielectric function which accordingly to @xcite can be written as : @xmath142 @xmath143 @xmath144\ln \frac{z - u+1}{z - u-1}% + [ 1-(z+u)^2]\ln \frac{z+u+1}{z+u-1}\right ) , \label{rfive}\ ] ] where @xmath145 , @xmath146 is the plasma frequency , @xmath147 is the drude relaxation time , and where @xmath135 and @xmath136 are the fermi velocity and wave vector , respectively . for @xmath98 polarization
the reflection factor is determined by @xmath148 where @xmath149 ^ 2\ln \frac{z - u^{\prime } + 1}{z - u^{\prime } -1}\right .
\nonumber \\ & & \left .
+ [ 1-(z+u^{\prime } ) ^2]^2\ln \frac{z+u^{\prime } + 1}{z+u^{\prime } -1}% \right ) \label{rnine}\end{aligned}\ ] ] with @xmath150 .
we show in sec.4 that the maximum of the electromagnetic friction is reached for small electron densities , where the electron gas becomes non - degenerate ( the electro gas is degenerate for @xmath151 and non - degenerate for @xmath152 , where @xmath153 is the fermi energy ) .
for non - degenerate electron gas we use the following classical expressions for dielectric functions @xcite @xmath154 \label{rten } \\ \epsilon _
t(\omega , q ) & = & 1+\frac{\omega _ p^2}{\omega ( \omega + i\gamma ) } % f\left ( \frac{\omega + i\gamma } { \sqrt{2}qv_t}\right ) \label{releven}\end{aligned}\ ] ] where the function @xmath155 is defined by the integral @xmath156 and @xmath157 , where @xmath158 is the electron mass .
by a well - conducting metal we mean one whose dielectric function @xmath159 ( @xmath77 is the conductivity ) has an absolute value much larger than unity . for good conductors at thermal frequencies @xmath160 and @xmath161 .
thus an enhancement in friction is possible only for very small @xmath162 .
it is convenient to write the friction coefficient for the two flat surfaces in the form @xmath163 taking into account that @xmath164 , from eq .
( [ 3eighteen ] ) for normal relative motion of clean surfaces within local optic approximation we get contribution to friction from evanescent @xmath36-and @xmath37-polarized electromagnetic waves @xmath165 ^ 2 \big[\big[(k^2+|s/\epsilon|^2)\mathrm{cosh}kd + 2k\big[\mathrm{im } ( s/\epsilon)\big]\mathrm{sinh}kd\big]^2\ ] ] @xmath166\frac{1}{\left|\big((s/\epsilon)^2-k^2\big ) \mathrm{sinh}kd
+ 2ik(s/\epsilon)\mathrm{cosh}kd\right|^4 } \label{4one}\ ] ] @xmath167 ^ 2 \big[\big[(k^2+|s|^2)\mathrm{cosh}kd + 2k\mathrm{im } s\mathrm{sinh}kd\big]^2\ ] ] @xmath168\frac{1}{\left|\big(s^2-k^2\big )
\mathrm{sinh}kd + 2iks\mathrm{cosh } kd\right|^4 } \label{4two}\ ] ] as @xmath169 , there is no singularity in @xmath170 , and for @xmath171 , for @xmath172 , than taking into account that in this limit @xmath173 and @xmath174 , we get @xmath175 @xmath176 where the surface impedance @xmath177 . for @xmath178 , @xmath179 becomes slowly dependent on @xmath2 : @xmath180 for @xmath181 we get @xmath182 for the propagating electromagnetic waves , taking into account that @xmath183 , we get @xmath184 @xmath185 for @xmath178 the contribution to friction from propagating wave is negligibly small in the comparison with the contribution from the evanescent waves . for @xmath181 as @xmath186 ,
the integral @xmath187 has no singularity , and we get for @xmath188 @xmath189 in addition , @xmath190 has singularities at the other zeroes of @xmath191 , i.e. , near the values @xmath192 ( @xmath193 is an integer ) .
in the vicinity of @xmath194 , putting @xmath195 , we have @xmath196 and @xmath197 , @xmath198 @xmath199 the number @xmath158 of such contribution is obviously equal to the integer part of the quantity @xmath200 ( @xmath201 $ ] ) , so that all @xmath194 ( with the exception of @xmath202 ) make a summary contribution @xmath203 @xmath204 @xmath205 \label{4ten}\ ] ] in the integral @xmath187 , the contribution from the vicinity of the point @xmath194 is @xmath206 @xmath207 and consequently @xmath208 @xmath209 at @xmath210 , when we can assume @xmath211 , the @xmath98 and @xmath49 wave contribution are approximately equal , and for the total contribution from propagating electromagnetic waves in this limit we get @xmath212 the above formulas in this section were obtained from the general eqs .
( [ 3seventeen]-[3twenty ] ) assuming absence of spatial dispersion of the dielectric function .
but these formulas contain only surface impedance @xmath213 that describe the ratio of the tangential components of the electric and magnetic fields on the boundary of the body .
thus , the results in this section remain in force also in the presence of spatial dispersion , provided only that the surface impedance of the medium is small enough .
we would have arrived at the same formulas also if we were to assume from the very beginning that the leontovich boundary condition @xmath214 is satisfied on the surface of the metal . at not too low temperatures ,
the impedances of metals are given by the formula for normal skin effect @xmath215 where @xmath77 is the conductivity .
in the local optic approximation it is assuming that there is no dependence of @xmath77 on @xmath28 . in the wien region of frequencies
it is also good approximation to neglect by the frequency dependence of @xmath77 . in this approximation using ( [ 4three ] ) for @xmath216 ( @xmath217 ) we get @xmath218 for the comparison the @xmath36-wave contribution for parallel relative motion for @xmath219 is given by ( @xcite ) @xmath220 it is interesting to note that for normal relative motion in contrast to the parallel relative motion practically for all @xmath221 the main contribution to friction comes from retardation effects because eqs . (
[ 4fourteen ] ) in contrast to eq .
( [ 4fifteen ] ) contains the light velocity . from eq .
( [ 4four ] ) we get @xmath37-wave contribution to friction for @xmath222 @xmath223 for parallel relative motion the @xmath37-wave contribution is in two times smaller . for @xmath224 , taking into account that eq .
( [ 4five ] ) is valid only for @xmath225 , we get @xmath226 from eq .
( [ 4twelve ] ) we get distance independent contribution to friction from propagating electromagnetic waves for @xmath227 @xmath228 figures 1 - 2 show the calculated contribution to the friction coefficient @xmath229 from evanescent electromagnetic waves for two semi - infinite solids , with parameters chosen to correspond to copper ( @xmath230 , @xmath231 at @xmath232 , for parallel ( fig.1 ) and normal ( fig.2 ) relative motion .
results are shown separately for both the @xmath37- and @xmath36- wave contribution .
the dashed line show the result when the local ( long - wavelength ) dielectric function @xmath233 is used , where @xmath234 in this case the integration in ( [ rtwo ] ) and ( [ rseven ] ) can be performed analytically and we get fresnel formulas . fig .
1 shows that the non - local optic effects become important for parallel relative motion for the @xmath49 wave contribution for sufficiently small separations ( @xmath235 ) .
however , for the @xmath98 wave contribution for both parallel and normal motion the non - local optic effects are negligibly small for practically all separations . for normal relative motion for @xmath36-wave contribution
the non - local optic effects are less important than for the parallel relative motion . in the present calculations we use the non - local dielectric approach which take into account the non - local optic effects on the length scale of the skin - depth .
there are also the short range non - local optic effects coming from the non - local nature of the screening response near the surface .
this gives the surface non - local contribution which we investigated in our previous publications @xcite .
comparing our previous calculations with the present one , we find that for @xmath236 the volume contribution from the non - local effects is of the same importance as the surface contribution . for @xmath237 the main contribution to the friction coefficient @xmath229 comes from @xmath37-polarized waves . in
particularly , at @xmath238 the @xmath37-wave contribution @xmath239 , so that with the surface area @xmath240m@xmath241 ( typical for probe scanning microscopy ) , the friction coefficient is @xmath242 .
the @xmath98wave contribution is characterized by weak distance dependence for @xmath243 , and @xmath244 for @xmath245 . for good conductors like copper , even for very short distances ,
the main contribution to the friction coefficient comes from the @xmath98 polarized electromagnetic waves .
this difference between @xmath49 and @xmath98 polarized waves results from screening effects : good conductors are good reflectors for @xmath49 polarized field , which implies that they are ineffective in the emission and absorption of evanescent @xmath49 polarized waves .
however these screening effects are less important for @xmath98 polarized waves . as pointed out in @xcite , the @xmath36 -wave contribution increase and the @xmath98wave contribution decrease when the free electron density decrease . within the local optic approximation the force of friction diverges when one go to the limit of zero conductivity .
this situation is different from the radiative heat transfer , where , even in the case of local optics , a maximum in the heat transfer occurs for conductivities corresponding to semi - metals .
fig.3 shows the dependence of coefficient of friction on the electron density .
when the electron density decreases there is transition from a degenerate electron gas to a non - degenerate electron gas at the density @xmath246 . at @xmath232 the transition density @xmath247 .
for @xmath248 we use the ( non - local ) dielectric function appropriate for a degenerate electron gas , while for @xmath249 we use an expression corresponding to a non- degenerate electron gas . in the calculations we used the electron mean free path @xmath250 . at @xmath251 the maximum value @xmath252 is obtained for @xmath253 , corresponding to the dc conductivity @xmath254
we rewrite the denominator in eq .
( [ 3eighteen ] ) in the form @xmath255 ^ 2\ ] ] @xmath256 ^ 2 \label{5one}\ ] ] where @xmath106 and @xmath105 are real and imaginary part of @xmath110 , respectively ( @xmath257 ) .
let us suppose that @xmath258 . in this case
resonant conditions are determined by equation @xmath259 close to resonance we can write @xmath260\ ] ] @xmath261 \label{5three}\ ] ] where @xmath262 which leads to the following contribution to the friction coefficient : @xmath263 } \label{5four}\ ] ] the parameter @xmath264 in this expression defines the value of @xmath265 where the two poles approximation is valid . to proceed further
let us make the following simplifications .
close to the poles we can use approximation @xmath266 where @xmath267 is a constant .
then from resonant condition ( [ 5two ] ) we get @xmath268 for the two poles approximation to be valid the difference @xmath269 . from this condition
we get @xmath270 . for the short distances
the parameter @xmath264 defines the value of @xmath265 where the solution of eq .
( [ 5two ] ) ceases to exit . for @xmath271 and @xmath272 from eq .
( [ 5four ] ) we get @xmath273 for the parallel relative motion using the same approximation we get @xmath274 interesting , the explicit @xmath2 dependence has dropped out of eq .
( [ 5eight ] ) .
however , @xmath275 is still @xmath2 dependent , through the @xmath2 dependence of @xmath264 . for the small distances one can expect that @xmath264 is determined by the dielectric properties of the material and does not depend on @xmath2 . in this case
the friction will be also distance independent . probably , the weak distance dependence observed in @xcite can be explained by the resonant photon tunneling
. resonant photon tunneling enhancement of the van der waals friction is possible for two semiconductor surfaces which can support low - frequency surface plasmon modes . as an example we consider two clean surfaces of silicon carbide ( sic ) .
the optical properties of this material can be described using an oscillator model @xcite @xmath276 with @xmath277 and @xmath278 the frequency of surface plasmons is determined by condition @xmath279 and from ( [ three ] ) we get @xmath280 . in fig.2
we plot the friction coefficient @xmath281 : note that the friction between the two semiconductor surfaces is several order of magnitude larger than between two clean good conductor surfaces .
another enhancement mechanism is connected with resonant photon tunneling between adsorbate vibrational modes localized on different surfaces . as an example , let us consider ions with charge @xmath282 adsorbed on metal surfaces .
the polarizability for ion vibration normal to the surface is given by @xmath283 where @xmath284 is the frequency of the normal adsorbate vibration , and @xmath285 is the damping constant . in eq .
( [ 3twentyone ] ) the contribution from parallel vibrations is reduced by the small factor @xmath286 .
however , the contribution of parallel vibrations to the van der waals friction can nevertheless be important due to the indirect interaction of parallel adsorbate vibration with the electric field , via the metal conduction electron @xcite .
thus , the small parallel component of the electric field will induce a strong electric current in the metal .
the drag force between the electron flow and adsorbates can induce adsorbate vibrations parallel to the surface .
this gives the polarizability : @xmath287 where @xmath193 is the conduction electron concentration .
as an illustration , in fig.3 we show coefficient of friction for the two cu(001 ) surfaces covered by a low concentration of potassium atoms ( @xmath288 ) . in the @xmath289
integral in eqs.([3eighteen],[3twenty ] ) we used the cut off @xmath290 ( where @xmath291 is the inter - adsorbate distance ) because our microscopic approach is applicable only when the wave length of the electromagnetic field is larger than double average distance between the adsorbates . in comparison ,
the friction between two clean surface at separation @xmath292 is seven order of magnitude smaller . at @xmath292
the friction coefficient @xmath112 for an atomic force microscope tip with @xmath293 is @xmath294 ( @xmath295 , see fig.2 ) ; this is of the same order of magnitude as the observed friction @xcite .
we have calculated the van der waals friction between two flat surfaces for normal relative motion and found a drastic difference in the comparison with parallel relative motion .
this difference is connected with resonance produced by the multiple scattering of the electromagnetic waves from the opposite surfaces . in the case of sharp resonance
it gives much larger contribution to friction in the case of normal relative motion than for parallel relative motion .
we have studied in the detail the friction between two good conductors and have found that for normal relative motion even for very small distances the main contribution to friction comes from the retardation effects .
the non - local optic effects are very important for @xmath36-wave contribution to friction for parallel relative and much less important for normal relative motion . for @xmath37-wave contribution
the non - local optic effects are unimportant for both direction of relative motion . in the case of van der waals friction
we have found that for distances between bodies @xmath296 , for good conductors with a high free electrons concentration , the main contribution to friction is associated with the @xmath37-polarized electromagnetic waves .
for @xmath243 this mechanism gives a friction coefficient per unit area @xmath297kgs@xmath298m@xmath299 nearly independent of the distance @xmath300 while for @xmath245 the friction coefficient @xmath229 depends on distance as @xmath301 for an atomic force microscope tip with the near substrate area @xmath302m@xmath303 we got the friction coefficient @xmath304kgs @xmath298 for @xmath243 .
when the concentration of electrons decreases , the @xmath98contribution to the friction decreases while the @xmath49 contribution increases . at @xmath251 and with the electron lifetime @xmath305s ,
the @xmath49 contribution reaches maximum @xmath306kgs@xmath298m@xmath307 at the electron concentration @xmath308m@xmath309 which corresponds to the conductivity @xmath310(@xmath311m)@xmath312 we have shown that the van der waals friction can be enhanced by several orders of magnitude in the case of resonant photon tunneling between low - frequency surface plasmon modes and adsorbate vibrational modes . in the case of friction for two cu(100 ) surfaces covered by a low concentration of potassium atoms at @xmath313 we have found friction of the same order of magnitude as it was observed in experiment @xcite .
however , the distance dependence in this case is more stronger than it was observed in @xcite .
further experiments with simple and well defined composition of the tip and sample must be performed to elucidate different energy dissipation mechanisms in the non - contact friction .
the obtained results should have broad application in non - contact friction microscopy , and in design of new tools for studying adsorbate vibrational dynamics and optical properties of the surface plasmons .
a.i.v acknowledges financial support from dfg and the russian foundation for basic research ( project no .
01 - 02 - 16202 ) b.n.j .
p acknowledges financial support from bmbf .
we thank r.o .
jones for help in the numerical calculations .
after substituting ( [ one ] ) and ( [ two ] ) into formula ( [ 3thirteen ] ) to linear order in vibrational coordinate @xmath52 and frequency @xmath51 we get @xmath314 @xmath315 + \left(\frac{c}{\omega}\right)^2p\big[(p+p^{*})(<\left|w_{0y}\right|^2 > + < \left|v_{0y}\right|^2>)\ ] ] @xmath316 + \big(\frac{p^+}{q^2}\big[(p+p^{*})(<w_{1z}w_{0z}^{*}>+<v_{1z}v_{0z}^{*}>+c.c.)\ ] ] @xmath317+\frac{c^2}{\omega(\omega + \omega_0)}p^{+}\big[(p+p^{*})(<w_{1y}w_{0y}^{*}>\ ] ] @xmath318 \big)e^{-i\omega_0t}\big ) \label{aone}\ ] ] from eqs .
( [ 3fifteen ] ) and ( [ aone ] ) it follow that the friction coefficient is determined by the formula @xmath319 @xmath320+\frac{c^2}{\omega(\omega + \omega_0)}p^{+}\big[(p+p^{*})(<w_{1y}w_{0y}^{*}>\ ] ] @xmath321 \big)\big)_{\omega_0=0 } \label{atwo}\ ] ] using eqs .
( [ nine]-[thirteen],[corfs],[corfp ] ) we get @xmath322 @xmath323 \label{athree}\ ] ] @xmath324 @xmath325 @xmath326e^{-2|p|d } \label{afour}\ ] ] other similar expressions for the @xmath37-wave contribution can be obtained from eqs .
( [ athree ] ) and ( [ afour ] ) by replacement of the reflection factors @xmath327 for @xmath36-polarized wave by the reflection factors @xmath328 for @xmath37-polarized wave . after substituting eqs .
( [ athree ] ) and ( [ afour ] ) , and similar expression for @xmath37-polarized waves in eq .
( [ atwo ] ) we get the friction coefficient for normal relative motion which is determined by formulas ( [ 3seventeen]- [ 3eighteen ] ) .
1 . the friction coefficient for two flat surfaces in parallel relative motion as a function of separation @xmath2 at @xmath329k with parameter chosen to correspond to copper ( @xmath330 ) .
the contributions from the @xmath98 and @xmath49polarized electromagnetic field are shown separately .
the full curves represent the results obtained within the non - local optic dielectric formalism , and the dashed curves represent the result obtained within local optic approximation .
( the log - function is with basis 10 ) fig .
2 . the friction coefficient for two flat surfaces in normal relative motion as a function of separation @xmath2 at @xmath329k with parameter chosen to correspond to copper ( @xmath330 ) .
the contributions from the @xmath98 and @xmath49polarized electromagnetic field are shown separately .
the full curves represent the results obtained within the non - local optic dielectric formalism , and the dashed curves represent the result obtained within local optic approximation .
( the log - function is with basis 10 ) fig .
3 . the friction coefficient for two flat surface as a function of the free electron density @xmath193 at @xmath329k .
the full curve was obtained by interpolation between the result ( dashed lines ) obtained within the non - local optic dielectric approach , with dielectric functions corresponding to a degenerate electron gas for @xmath331m@xmath332 , and to a non - degenerate electron gas for @xmath333 the calculation were performed with the damping constant @xmath230 , separation @xmath251 and @xmath334 ( the log - function is with basis 10 ) fig.4 .
the friction coefficient for two clean semiconductor surfaces in ( a ) normal and ( b ) parallel relative motion , as a function of the separation @xmath2 . @xmath335k and with parameters chosen to correspond to a surfaces of silicon carbide ( sic ) ( see text for explanation ) ( the log - function is with basis 10 ) fig .
5 . the friction coefficient for two surface covered by adsorbates in ( a ) normal and ( b ) parallel relative motion , as a function of the separation @xmath2 . @xmath336k and with parameters chosen to correspond to k / cu(001 ) @xcite ( @xmath337 ) ( the log - function is with basis 10 ) | we calculate the van der waals friction between two semi - infinite solids in normal relative motion and find a drastic difference in comparison with the parallel relative motion .
the case of the good conductors is investigated in details both within the local optic approximation , and using a non - local optic dielectric approach .
we show that the friction may increase by many order of magnitude when the surfaces are covered by adsorbates , or can support low - frequency surface plasmons . in this case
the friction is determined by resonant photon tunneling between adsorbate vibrational modes , or surface plasmon modes .
the theory is compared to atomic force microscope experimental data . |
the cern large hadron collider ( lhc ) and the upcoming international linear collider ( ilc ) are expected to perform precision tests of the standard model ( sm ) and explore the new physics at the tev scale @xcite .
the large extra dimensions ( led ) model is one of the scenarios beyond the sm which are proposed to solve the hierarchy problem @xcite .
the led model has only one fundamental scale , @xmath5 tev , and it may induce predictable collider phenomena at both the lhc and the ilc . up to now , many works on both the virtual kaluza - klein ( kk ) graviton exchange and the real kk - graviton production have been presented ; for example , the @xmath6 and @xmath7 processes were studied in the led model in refs.@xcite-@xcite .
in fact , the triple gauge boson ( tgb ) production processes are sensitive to the quartic gauge couplings ( qgcs ) and thus related to the electroweak symmetry breaking mechanism @xcite .
any deviation from the sm prediction hints at the existence of new physics , such as the higgsless or extra dimension signals @xcite . in discriminating physics beyond the sm
, we should investigate the potential contributions from the extension models .
compared with the thoroughly studied diboson production processes in extra dimension models , the tgb productions have been fully studied in the sm @xcite but gained less attention in the led model .
not long ago , the neutral tgb production processes at the lhc , @xmath8 , @xmath9 , @xmath10 and @xmath11 , were studied in the framework of the led model in ref.@xcite . in this paper
, we investigate the possible contributions of the virtual kk - graviton exchange to the @xmath12 and @xmath13 productions at both the lhc and ilc .
the motivation for this work is shown in two fields : firstly , the @xmath12 and @xmath13 processes are directly related to the sm qgcs , namely @xmath14 , @xmath15 and @xmath16 , which are different from the absence of the neutral qgcs in the sm at the tree level @xcite . secondly , although the experimental precision is limited by our understanding of strong qcd background , the lhc can provide more precision measurements of the qgcs than the existing data from lep ii and tevatron searches due to its very high energy and luminosity @xcite . furthermore , in the future , the qgcs can be further probed with higher precision at the ilc due to its cleaner environment @xcite . in this sense , the lhc and the upcoming ilc will provide complementary studies on the tgb production channels . the paper is organized as follows : in section ii , we present the related theory of the led model used in our calculations . in section iii ,
the calculation strategies are presented .
the numerical results and analyses for the @xmath12 and @xmath13 production processes at both colliders are provided in section iv , and a short summary is given in the last section .
in the led model @xcite , the spacetime is @xmath17 with @xmath18 being the number of extra dimensions .
the only fundamental scale @xmath19 unifying the gravity and the gauge interactions is at the tev . to explore the phenomenological effects
, one can extract the low - energy effective theory by the kk reduction in the brane - bulk picture @xcite . in this scenario ,
the sm particles are confined on a @xmath20-dimensional brane world volume while the gravity can propagate in the @xmath21-dimensional bulk .
after the assumed torus compactification of the extra dimensions @xmath18 , the usual planck scale @xmath22 in the @xmath20 spacetime is related to the fundamental scale @xmath19 as @xmath23 @xcite , where @xmath24 is the radius of the @xmath18 torus . in the following calculations , we adopt the de donder gauge for the kk - graviton part , while the feynman gauge ( @xmath25 ) is used for the sm part .
we assume all the momenta flow to the vertices , except that the fermionic momenta are set to be along with the fermion flow directions
. then we list the feynman rules for the relevant vertices and the propagator of spin-2 kk - graviton in the led model below @xcite , where @xmath26 , @xmath27 , @xmath28 , @xmath29 and @xmath30 represent the fields of the graviton , fermion , @xmath31 boson , @xmath32 boson and photon , respectively .
* @xmath33 @xmath34\end{aligned}\ ] ] * @xmath35 @xmath36 * @xmath37 @xmath38\end{aligned}\ ] ] * @xmath39 @xmath40\end{aligned}\ ] ] * @xmath41 @xmath42\end{aligned}\ ] ] * @xmath43 @xmath44\end{aligned}\ ] ] where @xmath45 , @xmath46 is the fine - structure constant , @xmath47 is the electric charge of fermion , @xmath48 are sine ( cosine ) of the weinberg angle , the vector and axial vector couplings of the @xmath32-boson , i.e. , @xmath49 and @xmath50 , are the same as in the sm , and @xmath51 is related to the reduced planck mass as @xmath52 , where @xmath53 is the newton constant .
the tensor coefficients @xmath54 , @xmath55 and @xmath56 are expressed as @xcite @xmath57 , { \nonumber}\\ e^{\mu \nu \rho \sigma}(k_{1},k_{2 } ) & = & \eta^{\mu \nu}(k_1^{\rho } k_1^{\sigma } + k_2^{\rho } k_2^{\sigma } + k_1^{\rho } k_2^{\sigma } ) - \left [ \eta^{\nu \sigma } k_1^{\mu } k_1^{\rho } + \eta^{\nu \rho } k_2^{\mu } k_2^{\sigma } + ( \mu \leftrightarrow \nu)\right ] .
{ \nonumber}\end{aligned}\ ] ] after summation over kk states the spin-2 kk - graviton propagator can be expressed as @xcite @xmath58,\end{aligned}\ ] ] where @xmath59,\ ] ] and @xmath60 the integral @xmath61 contains an ultraviolet cutoff @xmath62 on the kk modes @xcite .
it should be understood that the point @xmath63 has been removed from the integration path , and we set the ultraviolet cutoff @xmath62 to be the fundamental scale @xmath19 routinely .
the real part proportional to @xmath64 in eq.([res ] ) is from the narrow resonant production of a single kk mode with @xmath65 and the imaginary part @xmath66 is from the summation over the many nonresonant states .
the @xmath12 and @xmath13 productions at the lhc arise form the quark - antiquark annihilation and the gluon - gluon fusion subprocesses at the parton level : @xmath67 the @xmath68 processes at the ilc can be denoted as @xmath69 in reactions ( [ channel-1 ] ) , ( [ channel-2 ] ) and ( [ channel-3 ] ) , @xmath70 , and @xmath71 @xmath72 represent the four - momenta of initial and final particles .
the leading order ( lo ) feynman diagrams with kk - graviton exchange for ( [ channel-1 ] ) and ( [ channel-2 ] ) channels are shown in figs.[fig1 ] and [ fig2 ] , respectively , while the lo additional feynman diagrams in the led model for the process ( [ channel-3 ] ) are depicted in fig.[fig3 ] . from these feynman diagrams , one can find that the kk - graviton couples not only to the fermion pair , vector boson pair , and fermion - antifermion - vector boson ( @xmath73 ) , but also to the tgb including charged gauge boson , which is absent in the neutral tgb production processes as shown in ref.@xcite .
therefore , it is natural to expect that kk - graviton in the led model may induce considerable effects at the tev scale on the tgb production processes concerning charged gauge bosons at the lhc and the future ilc .
we express the feynman amplitudes for the partonic processes @xmath74 and @xmath75 as @xmath76 where @xmath77 @xmath78 is the amplitude contributed by the sm - like diagrams , while @xmath79 and @xmath80 are the amplitudes with kk - graviton exchange .
the feynman amplitude for the @xmath81 process can be expressed as @xmath82 where @xmath83 stands for the amplitude mediated by the sm - like particles , while @xmath84 is mediated by the kk - graviton .
the total cross sections for the partonic process @xmath85 can be expressed as @xmath86 where @xmath87 is the three - momentum of one initial parton in the center - of - mass system ( c.m.s ) , the summation is taken over the spins or colors of initial and final particles , the prime on the sum recalls averaging over initial spins or colors , and @xmath88 is the three - body phase space element expressed as @xmath89 by convoluting @xmath90 with the parton distribution functions ( pdfs ) of the colliding protons , the total cross section for the @xmath91 parent process can be written as @xmath92 , { \nonumber}\\\end{aligned}\ ] ] where @xmath93 @xmath94 represents the pdf of parton @xmath95 in proton @xmath96 , @xmath97 is the factorization scale , @xmath98 and @xmath99 describe the momentum fractions of parton ( quark or gluon ) in protons @xmath100 and @xmath101 , respectively .
the expression for the total cross section for @xmath81 is @xmath102 where @xmath87 is the three - momentum of the incoming @xmath103 ( or @xmath104 ) in the c.m.s of @xmath105 collider .
the prime on the sum means averaging over initial spin states as declared for eq.([int - ppvvv ] ) .
in this section , we present the numerical results in both the sm and the led model at the lhc and ilc . for the calculations at the lhc , we use the cteq6l1 pdfs @xcite with @xmath106 and @xmath107 ; the factorization scale is set to be @xmath108 and @xmath109 for the @xmath110 and @xmath111 processes , respectively .
the active quarks are taken as massless , i.e. , @xmath112 , the ckm matrix is set to be the unit matrix .
the other related input parameters are taken as @xcite @xmath113 since the led model is an effective low - energy theory , it breaks down in the non - perturbative region where @xmath114 or above . in order to make reliable and viable phenomenological predictions , we take the hard and conservative truncation scheme as setting the cut @xmath115 for proton - proton collision and the limit @xmath116 tev @xmath117 for the @xmath105 collision , where @xmath118 and @xmath119 are the partonic and @xmath103-@xmath104 c.m.s energies , respectively . in evaluating the @xmath120 processes , we put a transverse momentum cut @xmath121 gev and a rapidity cut @xmath122 on the final photon in order to get rid of the ir singularity at the tree level .
recently , the led parameters @xmath123 and @xmath18 have obtained more severe constraints by the lhc experiments .
the atlas collaboration provided @xmath124 confidence level lower limits on @xmath123 in the range of @xmath125 tev depending on the number of extra dimensions @xmath18 in the range of @xmath126 to @xmath127 @xcite .
the diphoton searches at cms set @xmath128 tev @xmath129 tev @xcite , and the dilepton experiments at cms set the limit on @xmath123 as @xmath130 tev @xmath131 tev with the number of extra dimensions @xmath18 varying from @xmath126 to @xmath127 at @xmath124 confidence level @xcite . in our calculations
we take @xmath132 tev and @xmath133 unless otherwise stated . for the verification of the correctness of our numerical calculations
, we use both the feynarts 3.5 @xcite and comphep 4.5.1 @xcite packages to calculate the integrated cross sections for the @xmath134 and @xmath91 processes at the @xmath135 gev ilc and the @xmath136 tev lhc in the sm separately .
we take the input parameters , pdfs , and the event selection criteria as mentioned above .
the numerical results are listed in table [ tab1 ] .
it demonstrates that the results from the two packages are in good agreement within the calculation errors . .
[ tab1 ] the integrated cross sections for the processes @xmath137 and @xmath91 in the sm at the @xmath135 gev ilc and the @xmath136 tev lhc by using feynarts 3.5 @xcite and comphep 4.5.1 @xcite packages separately . [ cols="^,^,^",options="header " , ]
in this paper , we study the led effects induced by the virtual kk - graviton on the @xmath3 and @xmath4 production processes at both the lhc and ilc .
the comparison between the results for these productions at both colliders are made .
we present the transverse momentum and rapidity distributions of final particles , and find that the led contributions remarkably affect the observables of these processes , particularly in either the high transverse momentum region or the central rapidity region ( @xmath138 ) .
we see that the relative led discrepancy becomes larger with the increment of the invariant mass of final @xmath31 pair , and the integrated cross section in the led model decreases when @xmath19 goes up for a fixed @xmath18 at both the ilc and lhc .
the @xmath139 discovery and @xmath140 exclusion limits on the fundamental scale @xmath19 are also obtained at both the ilc and lhc .
we also find that the relative led discrepancy for the @xmath0 production at the lhc is generally larger than that at the ilc due to the additional led contributions via @xmath1 fusion subprocesses and the kk - graviton exchanging resonant effect induced by the continues large colliding energy available in the lhc . from the aspect of effectively exploring the led signal in measuring @xmath0 production ,
we conclude that the @xmath3 and @xmath4 productions at the lhc could have the distinct advantage over at the ilc .
* acknowledgments : * this work was supported in part by the national natural science foundation of china ( grants no.11075150 , no.11005101 ) , and the specialized research fund for the doctoral program of higher education ( grant no.20093402110030 ) .
nima arkani - hamed , savas dimopoulos , gia dvali , phys . lett . *
b429 * , 263(1998 ) , arxiv : hep - ph/9803315 ; phys . rev .
* d59 * , 086004(1999 ) , arxiv : hep - ph/9807344 . i. antoniadis , n. arkani - hamed , s. dimopoulos , g. dvali , phys . lett .
* b436 * , 257(1998 ) , arxiv : hep - ph/9804398 .
n. agarwal , v. ravindran , v.k .
tiwari , a. tripathi , nucl .
b830 * , 248(2010 ) , arxiv:0909.2651 ; phys . rev . *
d82 * , 036001(2010 ) , arxiv:1003.5450 .
kumar , p. mathews , v. ravindran , a. tripathi , phys .
* b672 * 45(2009 ) , arxiv:0811.1670 .
s. ask , eur .
j. * c60 * , 509(2009 ) , arxiv:0809.4750 .
xiangdong gao , et al . , phys . rev .
* d81 * , 036008(2010 ) , arxiv:0912.0199 .
kumar , p. mathews , v. ravindran , s. seth , nucl .
b847 * , 54(2011 ) , arxiv:1011.6199 .
a. lazopoulos , k. melnikov , f. petriello , phys . rev . *
d76 * , 014001(2007 ) , arxiv : hep - ph/0703273 . v. hankele , d. zeppenfeld , phys . lett . *
b661 * , 103(2008 ) , arxiv:0712.3544 .
g. bozzi , et al .
, phys . rev .
* d81 * , 094030(2010 ) , arxiv:0911.0438 .
aleph and delphi and l3 and opal and lep electroweak working group collaborations , cern - ph - ep/2005 - 051 , arxiv : hep - ex/0511027 .
measurement of the w+w - gamma cross section and direct limits on anomalous quartic gauge boson couplings at lep l3 collaboration , phys .
b490 * , 187(2000 ) , arxiv : hep - ex/0008022 . | we investigate the effect induced by the kaluza - klein ( kk ) graviton in the @xmath0 production in the framework of the large extra dimensions ( led ) model at both the cern large hadron collider ( lhc ) and the international linear collider ( ilc ) .
the integrated cross sections and various kinematic distributions in the led model are presented and compared with those in the standard model .
the results show that the contributions from kk - graviton exchange remarkably affect the observables of the triple gauge boson ( @xmath0 ) production processes at both the ilc and lhc , particularly either in the high transverse momentum region or in the central rapidity region .
we also find that the relative led discrepancy for the @xmath0 production at the lhc is generally larger than that at the ilc due to the additional led contribution via @xmath1 fusion subprocess and the kk - graviton exchanging resonant effect induced by the continuous large colliding energy in @xmath2 collision .
we conclude that the @xmath3 and @xmath4 productions at the lhc could have the distinct advantage over at the ilc from the aspect of effectively exploring the led signal in measuring @xmath0 production .
= 0.32 in addtoresetequationsection |
quasi - one dimensional ( 1d ) metals are interesting materials because of their tendency to undergo a phase - transition inevitably due to the inherent instability called the peierls instability @xcite .
this instability , originating from the electron - phonon interaction leads to a metal - insulator transition at reduced temperature , where the insulating state may be characterized by a lattice deformation in conjunction with a charge modulation and the opening of a single particle excitation gap in the electronic spectrum.@xcite one of the striking features of these materials are their unusual non - linear electrical conduction properties @xcite , which arises due to the collective motion of the charge - density wave ( cdw ) .
along with the conventional single particle excitations , these materials also have collective excitations , _ viz . _ ,
amplitudons and phasons , which dominate the low energy physical properties @xcite .
the raman active collective amplitudon mode , _
i.e. _ the transverse oscillation of the amplitude of the coupled charge - lattice wave , has been observed by various researchers @xcite .
blue bronze , a@xmath0moo@xmath1 ( a = alkali metal ) is one of the best known quasi-1d materials which undergoes a cdw transition@xcite via the peierls channel .
the first evidence of the cdw transition in k@xmath0moo@xmath1 by raman measurements came from the experiments by travaglini and wachter@xcite . in this paper a mode around 50 @xmath2 was observed at temperatures below t@xmath3 @xmath4 k that showed a strong temperature dependence of , in particular , its frequency when the temperature approached tcdw from below . as it was also found that this is a fully symmetric ( @xmath5-mode ) breathing mode it was assigned to the amplitude oscillation of the charge density wave .
further evidence came from the observation that the line width tends to diverge as the @xmath6 was approached from below .
this anomalous behaviour of the line width could be due to the fluctuation effects that become pronounced in the vicinity of the phase transition .
mean - field theory of a generalized second - order phase transition predicts that , due to a temperature dependent order parameter , the frequency ( and the amplitude ) of the soft mode tends to zero as the temperature is approaches t@xmath3.@xcite however , in the the study by travaglini _
et al_. @xcite the softening of the amplitudon mode with increasing temperature was found to be incomplete .
the observed softening of only @xmath713 % was ascribed to 3d - correlations .
the 3d - correlations are nothing but the long range coulomb interaction between density - modulated electrons , which might effectively screen the electron - phonon interaction @xcite , thereby reducing the expected complete mode - softening .
the effects of strong local electron - electron interaction leads to both dynamical and static screening processes dressing the electron - phonon coupling .
these screening effects are thought to be a generic feature of strongly correlated electron systems .
recently a raman study of the anomalous profile of the amplitudon mode in k@xmath0moo@xmath1 was reported @xcite in which for temperatures below 100 k the amplitudon mode showed an anomalous `` fin''-like profile toward the stoke s side of it s central frequency .
this was interpreted as splitting of the amplitudon mode into two or more closely spaced harmonic oscillator type modes .
the splitting was claimed to be due to a strong perturbation of the charge density wave by impurity potentials . at very low temperatures where there are practically no thermally excited electrons the impurity potentials remain unscreened , leading to a spatial variation of the cdw amplitude and a splitting of the amplitudon mode .
the present paper revisits the vibrational and electronic excitations in the blue bronzes through a comparative study of rb@xmath0moo@xmath1 and k@xmath0moo@xmath1 in order to examine the behaviour of the amplitudon mode in both the compounds together with a detailed temperature and polarization dependent study of the low energy phonons .
at room temperature the k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 bronzes crystallize in the monoclinic c__2/m _ _ space group @xcite with lattice constants _ _ a__=16.23 , _ _
b__=7.5502 , _ _
c__=9.8 , and @xmath8=94.89@xmath9 for k@xmath0moo@xmath1 and _ _ a__=16.36 , _ _
b__=7.555 , _ _
c__=10.094 , and @xmath8=93.87@xmath9 for rb@xmath0moo@xmath1 .
there are 20 formula units per unit cell .
the structure is build up by edge and corner sharing clusters of ten strongly distorted moo@xmath10 octahedra stacked along the [ 010 ] direction .
these clusters form layers in the ( 102 ) planes , which are separated by the alkali ions . at the charge density wave transition temperature
the structure transforms in an incommensurately modulated structure with a temperature dependent incommensurate wave vector @xmath11 close to ( but not equal to ) @xmath12.@xcite at low temperature ( @xmath13 k ) the incommensurate wave vector becomes temperature independent , even though this does not seem to correspond to a lock in transition , _
i.e. _ the modulation remains incommensurate.@xcite k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 single crystals with typical size @xmath14 mm@xmath15 were prepared by the temperature gradient flux method@xcite . after polishing ,
samples were mounted in an optical flow cryostat ( temperature stabilization better than 0.1 k ) and backscattering raman spectra were recorded ( frequency resolution @xmath72 @xmath2 ) using a triple grating monochromator in subtractive mode equipped with a liquid nitrogen cooled ccd detector . a solid - state diode - pumped , frequency - doubled @xmath16 laser system is used as excitation source ( wavelength 532 nm , spot size 10 @xmath17 m ) .
the power density was kept below 100 w/@xmath18 in order to minimize heating effects .
the polarization was controlled on both the incoming and outgoing beam , giving access to all the polarizations allowed for by the back - scattering configuration .
in general both parallel ( ( @xmath19 ) , where @xmath20 is in the [ 20 - 1 ] direction , and ( @xmath21 ) in porto notation ) and perpendicular polarizations have been measured .
the perpendicular spectra did not yield appreciable scattering intensity and will therefore not be discussed here .
fig . [ ram1 ] shows the raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 above t@xmath3 for ( @xmath21 ) and ( @xmath19 ) polarization . above t@xmath3 the raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1
show very similar raman vibrational features .
only a few modes are observed in the ( @xmath21 ) spectra ( fig . [ ram1 ] , left panel ) since here the polarizations are along the metallic direction .
there is a strong mode near 475 @xmath2 , a triplet with frequencies near 320 , 365 , and 400 @xmath2 , and some weaker modes below 200 @xmath2 .
the symmetric ( @xmath22 ) modes above 300 @xmath2 originate from bending vibrations of the moo@xmath23 units , as has already been noted in previous studies on blue bronzes@xcite , and is consistent with the interpretation of the raman spectra of pure @xmath24@xcite .
there are quite a few different modes originating from the moo@xmath23 units , due to the large distortion of the moo@xmath23 `` octahedra '' ; the metal - oxygen distances in blue bronzes typically range from 1.7 to 2.3 @xcite .
the mo - o stretching modes appear at higher frequencies ( between 900 to 1000 @xmath2 ) @xcite as reported by shigeru _
et al . , _ @xcite and massa @xcite . in general , the frequencies for the modes observed in rb@xmath0moo@xmath1 are somewhat lower than the corresponding modes in k@xmath0moo@xmath1 .
this is consistent with the fact that rb@xmath0moo@xmath1 has slightly larger lattice constants in the @xmath25 and @xmath26 directions@xcite resulting from the larger ionic radius of the rubidium ion .
a small variance in the various mo - o distances between k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 has also been observed in the study of schutte _ et al . ,
_ @xcite .
polarized @xmath27 and @xmath28 raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 above t@xmath3 .
spectra have been given an offset for clarity . ]
the right panel of fig.[ram1 ] shows the raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 ( @xmath29t@xmath3 ) for the polarizations of the incident and scattered light in the @xmath20 direction , _
i.e. _ perpendicular to metallic direction .
as expected , the scattering efficiency in this direction is substantially higher than in the metallic direction , and quite a few more modes are observed . in both rb@xmath0moo@xmath1 and
k@xmath0moo@xmath1 there are about 20 modes that are resolved in the present experiment .
again , the overall spectra are qualitatively very similar and many of the rb@xmath0moo@xmath1 vibrational modes are shifted toward slightly lower frequency in comparison to the k@xmath0moo@xmath1 modes .
for instance the triplet near 150 @xmath2 in the k@xmath0moo@xmath1 bronze are found some 10 - 15 @xmath2 lower in energy in the rb@xmath0moo@xmath1 spectra , and the 480 @xmath2 k@xmath0moo@xmath1 mode is observed at 473 @xmath2 in rb@xmath0moo@xmath1 . finally , note that group theory predicts 71 a@xmath30 modes for the high temperature phase of the bronzes .
the experimental spectra , however , reveal only 35 modes .
the `` missing '' modes most likely result from accidental degeneracies and from weak intensity and/or screening , in particular in the metallic direction .
for the predicted 55 b@xmath31 modes ( ( @xmath32 ) geometry ) the intensity is to weak to be detected in the present experiments .
( @xmath21 ) polarized raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 below t@xmath3 .
spectra have been given an offset for clarity . ] ( @xmath19 ) polarized raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 below t@xmath3 .
spectra have been given an offset for clarity .
the insets show details of the splitting of some of the modes .
the peak at 60 @xmath2 marked by an asterix is an experimental artefact originating from air . ]
the low temperature phase shows a much richer spectrum than the high temperature one ( see figs .
[ ram4 ] and [ ram3 ] ) , showing many quite narrow raman modes .
the ( @xmath21 ) polarized low temperature spectra ( fig .
[ ram4 ] ) exhibit 32 distinct modes for the k@xmath0moo@xmath1 compound , and 42 in the rb@xmath0moo@xmath1 compound .
for the ( @xmath19 ) polarization ( fig .
[ ram3 ] ) the number of observed modes is 38 and 44 for k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 , respectively .
again , the spectra for the two compounds are quite similar .
the main difference is the number of observed modes which seems to originate from an additional splitting of modes in the rb@xmath0moo@xmath1 case .
for instance the doublet observed near 75 @xmath2 in k@xmath0moo@xmath1 is seen as a triplet ( see also left inset fig.[ram3 ] ) .
we note that , in a previous raman study @xcite a band near 80 @xmath2 was observed in rb@xmath0moo@xmath1 which was ascribed to a zone - folded mode arising from the folding of the original brillouin - zone due to cdw transition .
the picture of zone - folding - active phonons is corroborated by previous neutron scattering studies by pouget _
et al . , _
@xcite in which a flat dispersion - less phonon branch was observed . in the present study
three modes are observed in the k@xmath0moo@xmath1 compound in the 40 - 100 @xmath2 region , and five in the rb@xmath0moo@xmath1 compound .
as is discussed in section [ tdep ] , the lowest modes at 56 @xmath2 ( k@xmath0moo@xmath1 ) and 53 @xmath2 ( rb@xmath0moo@xmath1 ) are the amplitudon mode of the cdw .
note that only a single mode is observed here , and no evidence is found for the `` fin '' like structure as discussed by shigeru _
also the mo - o bending modes show an additional splitting in the the rb@xmath0moo@xmath1 compound , which is exemplified by the right inset of fig .
[ ram3 ] for the 480 @xmath2mode . as a general rule ,
splitting of raman modes occurs mainly due to two mechanisms .
the first one occurs when a static distortion transforms equivalent atomic positions to inequivalent ones @xcite .
a second mechanism is the correlation field or davydov@xcite splitting that occurs due to coupling of vibrations of molecular units at different equivalent sites in the unit cell , and is sensitive to transitions involving multiplication of the unit cell size ( like for instance the peierls distortion ) .
the incommensurate nature of the blue bronzes may also activate vibrational modes in raman spectra@xcite .
modes with wave vector * k*@xmath33*q * become active in the incommensurate phase , where * q * is the incommensurate modulation vector , and their scattering strength depends strongly on the modulation amplitude .
the temperature dependence of the low frequency modes of both compounds is depicted in fig .
[ ram45 ] for a few selected temperatures .
as mentioned before , in k@xmath0moo@xmath1 there are only two clear phonons observed in this region at 74 @xmath2 , and 85 @xmath2 , whereas the rb@xmath0moo@xmath1 compound shows four phonon modes in this region centred at 64 , 75 , 83 , and 90 @xmath2 .
this , as well as the previously noted additional splitting in some of the high frequency modes in rb@xmath0moo@xmath1 , strongly indicates that the low temperature structures of the two compounds are not completely identical . in general
it is thought that the crystal structures of rb@xmath0moo@xmath1 and k@xmath0moo@xmath1 are , apart from the unit cell dimensions , the same@xcite .
the only relevant small difference lies in the incommensurate modulation .
both structures have their incommensurate modulation along the * b@xmath34 * direction .
the amplitude of the modulation , however , is found to be substantially larger for the rb@xmath0moo@xmath1 compound@xcite .
therefore , we tentatively conclude that the observation of a larger number of modes in rb@xmath0moo@xmath1 , as compared to k@xmath0moo@xmath1 , is due to the larger displacement amplitude of the incommensurate modulation in the low temperature phase .
polarized raman spectra of k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 at selected temperatures .
panel a ) and b ) : ( @xmath19 ) spectra ; panel c ) and d ) : ( @xmath21 ) spectra .
the spectra have been given an offset for clarity . ] upon decreasing temperature the modes in rb@xmath0moo@xmath1 show a continuous behaviour , and no further indications for a possible additional phase transition are found .
this in contrast to the situation in k@xmath0moo@xmath1 . upon lowering the temperature ,
some of the modes show an abrupt change in their energy in the between 80 k and 100 k. this is exemplified in fig .
clearly some of the modes show a discontinuous shift to higher energy of a few wave numbers indicating a phase transition .
it should be noted that not all modes exhibit the frequency-kink , as can for instance be seen for the 75 @xmath2 mode in fig.[ram7 ] . in general , a temperature - dependent frequency - jump is a signature of a first - order transition . in the present case
, it could be due to a lock - in transition to a commensurate state .
however , the low temperature phase of k@xmath0moo@xmath1 has been studied in quite some detail and temperature dependent neutron scattering experiments@xcite failed to show a lock - in transition in k@xmath0moo@xmath1 .
it was found that the incommensurate wave vector is strongly temperature dependent down to @xmath35 k. below this temperature the wave vector becomes temperature independent , but remains incommensurate .
still , the present study strongly indicates that there is indeed a first order phase transition in k@xmath0moo@xmath1 below 100 k. details of the ( @xmath19 ) polarized raman spectra of k@xmath0moo@xmath1 showing a discontinuous energy shift between 80k and 100k for some of the modes .
the spectra have been given an offset for clarity . ]
the most salient feature in the low frequency spectra is the appearance of the amplitudon in both compounds which shows a softening and broadening upon approaching the phase transition ( see fig.[ram45 ] ) .
the low temperature frequencies of these modes are 56 @xmath2 in the k@xmath0moo@xmath1 compound , and 53 @xmath2 in rb@xmath0moo@xmath1 .
the interpretation of these modes as amplitudon modes is consistent with the previous neutron study by pouget _
et al . , _ @xcite , and raman scattering studies by travaglini _
et al . , _ @xcite and massa @xcite .
recently this mode has also been observed in k@xmath0moo@xmath1 as coherent excitations in pump - probe transient reflectivity experiments @xcite , in addition to coherent phonon excitations@xcite at 74 @xmath2 , and 85 @xmath2 also observed in the present data .
temperature dependent energy ( left panel ) and line width ( right panel ) of the amplitudon mode in k@xmath0moo@xmath1 and rb@xmath0moo@xmath1 . ]
the amplitudon modes disappear from the spectra already around 150 k. this is due to the strongly over damped nature of the amplitudon in the vicinity of the phase transition to the metallic state originating from strong fluctuations of the charge density wave .
these fluctuations lead to a diverging line width ( see fig .
[ linewidth ] , right panel ) , which in turn makes it impossible to observe complete amplitude softening .
the observable amplitudon softening upon approaching the phase transition amounts only about 10 % , as is shown in fig .
[ linewidth ] ( left panel ) .
line width of the phonon modes in k@xmath0moo@xmath1 as a function of their energy for @xmath36 k , 100 k , and 170 k. ] finally , turn back to the phonon modes observed in the cdw phase .
the phonon modes observed at low temperatures are rather narrow , and often resolution limited .
this is in particular true for the phonons below 400 @xmath2 .
fig.[ram11 ] displays the measured line width of the phonons as a function of their frequencies in k@xmath0moo@xmath1 for various temperatures .
as can be seen from the figure , at 6 k the line width of the modes are very narrow up to @xmath7350 @xmath2 . in contrast , the modes above 350 @xmath2 show a progressive increase in their line width . in other words , there is a sharp onset near 350 @xmath2 in the behaviour of the line widths . as temperature
is increased this onset shifts to lower frequencies .
a similar behaviour is observed in rb@xmath0moo@xmath1 . the observed anomalous line broadening hints to a coupling of the phonons to electronic excitations with an energy above @xmath7350 @xmath2 , which , given the temperature dependence , scale with the cdw gap .
since the cdw energy gap in k@xmath0moo@xmath1 is about 900 @xmath2 it can not be a coupling to the usual quasi - particle excitations .
it has been argued in the literature that a common feature of incommensurate cdw materials is the presence of so called midgap states@xcite .
therefore the anomalous broadening of the phonons above @xmath7350 @xmath2 is tentatively assigned to a coupling of the phonons to midgap state excitations .
in summary , we have studied two representatives of the prototypical quasi - one - dimensional charge density wave system , blue bronze , using temperature and polarization dependent raman scattering experiments .
the cdw transition lead to the activation of a large number of modes in the low temperature phase of both the bronzes .
the amplitudon mode , found at nearly equal energy in rb@xmath0moo@xmath1 and k@xmath0moo@xmath1 ( there is a 2 @xmath2 difference between them ) , shows a rapid exponential broadening upon approaching the phase transition from below making it impossible to observe a full softening of this mode .
the observation of a number of additional modes in the rb@xmath0moo@xmath1 compound is assigned to the stronger incommensurate modulation in this compound .
this conclusion is , however , based on the limited amount of low temperature structural studies , and it would be interesting to perform additional studies to confirm the present interpretation .
this also holds for the observed changes in the k@xmath0moo@xmath1 spectra between 80 and 100 k , which indicate a first order phase transition .
this is in line with the recently discussion of the maximum in the threshold field for cdw conduction which was interpreted as due to a incommensurate - commensurate phase transition ( lock - in transition).@xcite this transition is not related to the recently reported observation of a glass transition in k@xmath0moo@xmath1 by staresinic _
et al._@xcite at @xmath37 k , for which no further evidence was found in the present study .
finally , evidence is found for a coupling of the high frequency phonons to mid - gap excitations .
* acknowledgements * this work is part of the research programme of the stichting voor fundamenteel onderzoek der materie ( fom) , which is financially supported by the nederlandse organisatie voor wetenschappelijk onderzoek ( nwo). | we report a comparative raman spectroscopic study of the quasi - one - dimensional charge - density - wave systems a@xmath0moo@xmath1 ( a = k , rb ) .
the temperature and polarization dependent experiments reveal charge - coupled vibrational raman features .
the strongly temperature - dependent collective amplitudon mode in both materials differ by about 3 @xmath2 , thus revealing the role of alkali atom .
we discus the observed vibrational features in terms of charge - density - wave ground state accompanied by change in the crystal symmetry .
a frequency - kink in some modes seen in k@xmath0moo@xmath1 between t = 80 k and 100 k supports the first - order lock - in transition , unlike rb@xmath0moo@xmath1 .
the unusually sharp raman lines(limited by the instrumental response ) at very low temperatures and their temperature evolution suggests that the decay of the low energy phonons is strongly influenced by the presence of the temperature dependent charge density wave gap . |
the complexity in the study of the non - equilibrium dynamics in isolated many - body quantum systems is at first sight overwhelming . at the same time
, many interesting questions naturally arise when a system is brought out of equilibrium , for example regarding the nature of quantum decoherence or relaxation processes . in the past decade , these issues have been at the center of a fruitful theoretical research activity , partially motivated by the experimental achievements in the physics of ultra - cold atoms , such as the possibility to realize highly isolated and controlled many - body quantum systems with reduced effective dimensionality @xcite . in particular , due to the exact solvability of several one - dimensional quantum systems , they represent an ideal framework where exact theoretical results can be compared with experimental data .
many studies have already shown that integrable one - dimensional many - body hamiltonians can be experimentally engineered and their non - equilibrium dynamics probed in exquisite detail @xcite .
arguably , the simplest protocol to bring a system out of equilibrium is that of a quantum quench @xcite : a well - defined initial state ( usually the ground state of some hamiltonian @xmath2 ) is prepared at time @xmath3 and subsequently time evolved with some hamiltonian @xmath4 . after the quench , it is expected that an extended system should act as an infinite thermal bath on its own finite subsystems and , accordingly , local properties should relax to stationary values described by a gibbs ensemble . for a generic ( i.e. non - integrable ) model
, this picture turns out to be correct and the time - dependent local correlations indeed approach thermal stationary values @xcite . a different behavior is observed in integrable systems . in this case , local correlations still approach stationary values but these are not in general given by a thermal ensemble .
the reason lies in the existence of an infinite set of conservation laws which constrain the dynamics at all times .
the thermal gibbs ensemble then has to be replaced by a generalized gibbs ensemble ( gge ) , which is built taking into account all the local and quasi - local conserved operators ( or charges ) of the system @xcite .
the importance of a notion of locality for the charges was not realized immediately and it was first pointed out in refs .
the gge has been first considered for integrable models admitting a free particles representation @xcite and only after for genuinely interacting integrable models ( see the volume @xcite as a collection of reviews on the subject ) .
the efforts to better understand the range of validity of the gge for interacting integrable models has already produced important conceptual and technical advances in our understanding of non - equilibrium physics in isolated quantum systems .
in particular , one promising development is the so called quench action method @xcite , which is an integrability based analytical approach to study the time evolution following a quench .
one of its main results is that the effective long - times stationary state can be represented by an highly excited eigenstate of the post - quench hamiltonian .
this description is complementary to the gge and it is analogous to the one employed in the standard treatment of the thermodynamic bethe ansatz .
indeed , in the latter it is well known that the thermal gibbs ensemble can be effectively replaced by a maximal entropy representative eigenstate @xcite .
while integrable models are by definition special , they can be engineered to a great level of accuracy in cold atomic systems ( as e.g. in refs .
@xcite ) and furthermore understanding their quench dynamics leads to remarkable insights also for the dynamics of more generic systems @xcite . the quench action method has proven to be extremely versatile , with applications to quantum quenches in various models ranging from spin chains @xcite , to bose gases @xcite and integrable quantum field theories @xcite , see ref .
@xcite for a recent review .
it is , however , limited to those initial states for which the overlaps with the eigenstates of the post - quench hamiltonian are known exactly .
these quantities are in general extremely difficult to compute and so far no general scheme has been developed to tackle this problem which has been solved only in a few cases @xcite . in the quantum quench literature ,
the xxz spin-@xmath0 heisenberg chain has served as the simplest interacting many - body quantum system where analytical predictions could be tested with high precision against numerical methods @xcite . even in this prototypical model ,
the quench dynamics from simple families of product states can not in general be analyzed by means of the quench action approach because the overlaps with the eigenstates are still missing . physically interesting examples of such states can be found amongst those considered in ref .
@xcite , where it was investigated how symmetries that are initially broken can be restored at late times after a quench .
the heisenberg spin chain has also been the prototypical example which allowed to understand that the known ( ultra-)local charges in integrable models are in general not complete and new independent families of quasi - local conserved operators exist .
these were firstly introduced in the context of transport problems @xcite , where their consequences have been explored in great detail @xcite , see @xcite for a recent review .
it was only later understood that quasi - local conservation laws have to be taken into account in the gge construction @xcite .
it is worth mentioning that independent works have also shown the existence and physical relevance of conservation laws with a generalized notion of locality in free xy spin chains @xcite and quantum field theories @xcite .
it has been explicitly shown that the quasi - local charges in the xxz heisenberg chain form a complete set of conservation laws @xcite : fixing their expectation value uniquely determines a single excited state of the hamiltonian in the thermodynamic limit .
this result has a natural application in the study of quantum quenches . indeed , assuming the existence of a post - quench representative eigenstate , this can be completely fixed by imposing that the expectation values of the quasi - local conserved charges are equal to those computed in the initial state .
for those cases for which the overlaps were known , it was shown @xcite that this approach leads to the same results obtained by applying the quench action method .
in particular , in the xxz spin-@xmath0 chain a full analytical characterization of the post - quench stationary state was recovered for two particular initial states , namely the nel and the majumdar - ghosh state , as first derived in refs .
@xcite . in this work
we consider simple families of physically relevant product states in spin-@xmath0 and spin-@xmath1 integrable chains , both in the gapped regime and at the isotropic point .
our study include the states considered in ref .
@xcite for which the overlaps are not known . by applying the techniques of ref .
@xcite , we show that even in these cases a closed - form analytical characterization of the post - quench stationary state can be explicitly exhibited . using these results , we compute the asymptotic value of the time - dependent local correlations after the quench and test our predictions against the tdmrg and itebd data of ref .
@xcite .
one of the main physical motivations for our work is to explore the features of post - quench steady states when initial states other than spin-@xmath0 nel and dimer states are considered .
as we will see , qualitative different properties emerge for ferromagnets and antiferromagnets , such as different bound - state compositions . in turn
these might be experimentally probed , for example by observing light - cones spreading of local perturbations on the post - quench steady state @xcite .
the organization of the rest of this manuscript is as follows . in section [ sec : setup ]
we introduce the models that we consider and briefly review their bethe ansatz description .
we then provide an overview of some of the ideas developed recently within the quench action approach , such as the possibility of characterizing the post - quench steady state in terms of a representative eigenstate .
for the sake of exposition , a summary of our results is given in section [ sec : results ] , where we exhibit the analytical characterization of the post - quench steady state for the families of initial states studied in this work .
the derivation of our results is then reported in section [ sec : derivation ] , while section [ sec : correlations ] is devoted to the comparison with the numerical data of ref .
@xcite for the whole time - evolution of local correlators . in section [ sec :
general_states ] more general initial states are discussed , while our conclusions are presented in section [ sec : conclusions ] .
the first model that we consider is the xxz spin-@xmath0 heisenberg chain .
it is defined on the hilbert space @xmath5 , where @xmath6 is the local space associated with site @xmath7 , @xmath8=2 $ ] .
the hamiltonian reads h_xxz^(1/2)&= & _ j=1^n , [ hamiltonian1 ] where periodic boundary conditions are assumed , @xmath9 . here
, @xmath10 are local spin operators which are related to the pauli matrices by @xmath11 . in this work
we restrict to the regime @xmath12 and we introduce the parameters & = & ( ) , + q&=&^. [ def : q ] we further employ the following notation for @xmath13-deformed numbers _ q = = . [ q_deformed ]
the hamiltonian can be diagonalized by bethe ansatz , as it will be reviewed in section [ sec : bethe_ansatz ] .
the generalization of the xxz heisenberg chain to the spin-@xmath1 case leads to the so called zamolodchikov - fateev model @xcite .
it is defined on the hilbert space @xmath14 , where @xmath15 is the local space associated with site @xmath7 and @xmath16=3 $ ] .
the hamiltonian reads h_zf^(1)&=&- 4n ^2 ( ) + _ j=1^n\{+2 + & & -_a , ba_ab()s^a_js^b_js^a_j+1s^b_j+1 } , [ hamiltonian2a ] where the indices @xmath17 in the second sum take the values @xmath18 , @xmath19 , @xmath20 and where periodic boundary conditions are assumed , @xmath9 .
the coefficients @xmath21 are defined by @xmath22 and a_xx = a_yy = 1 , a_zz = ( 2),a_xy=1 , a_xz = a_yz=2 - 1 , [ coefficients ] while @xmath23 plays the role of the anisotropy parameter along the @xmath20-direction . here , with a slight abuse of notations , we employed the symbols @xmath24 , already used for the spin-@xmath0 case in , for the spin-@xmath1 operators . in the spin-@xmath1 case
they are defined as s^x= ( c c c0&1&0 + 1&0&1 + 0&1&0 ) , s^y= ( c c c0&-i&0 + i&0&-i + 0&i&0 ) , s^z= ( c c c1&0&0 + 0&0&0 + 0&0&-1 ) .
[ spin_op ] finally , we also define for future reference the local spin-@xmath1 basis as |= ( c1 + 0 + 0 ) , |0 =( c0 + 1 + 0 ) , |= ( c0 + 0 + 1 ) . note that the hamiltonian obtained from by a straightforward substitution of the spin-@xmath0 with the spin-@xmath1 operators results in a non - integrable model .
while the zamolodchikov - fateev model is not the only spin-@xmath1 integrable chain , it is arguably the simplest . in particular
, the form of the hamiltonian greatly simplifies at the isotropic point @xmath25 , where it coincides with the well - known babujian - takhtajan hamiltonian @xcite h^(1)_bt=_j=1^n , [ hamiltonian2b ] where @xmath26 denotes the vector @xmath27 . in this work
we will consider both the gapped regime of the hamiltonian ( namely @xmath12 , @xmath28 ) , and the gapless isotropic hamiltonian .
in this section we sketch the aspects of the bethe ansatz formalism that will be used in this work , focusing in particular on the thermodynamic limit .
we refer the reader to the specialized literature for a more systematic treatment @xcite . for a bethe ansatz integrable model
, the energy eigenstates are parametrized by complex quasi - momenta @xmath29 , satisfying a set of non - linear quantization conditions , namely the so called bethe equations . in the gapped regime of the hamiltonians , , namely for @xmath28
, they can be written in a compact form as @xcite ^n = _ ^m , [ eq : bethe_eq ] where @xmath30 , @xmath31 for spin-@xmath0 and spin-@xmath1 respectively . here
, @xmath32 is the number of flipped - spins w.r.t . to the ferromagnetic state with all the spins up .
up to an overall global additive constant , the energy associated with an eigenstate corresponding to the rapidities @xmath29 is simply given by e&=&-(2s)_j=1^m .
the solutions of organize themselves into mutually disjoint patterns in the complex plane called strings " .
intuitively , a @xmath33-string solution corresponds to a bound state of @xmath33-magnons , i.e. spin flips w.r.t .
the ferromagnetic reference state .
the rapidities within a @xmath33-string are parametrized as _
^(j , m)= ^(m ) _
+ i ( j-)+^(j , m)_j=1 , , m. [ string_hp ] here @xmath34 is a real number called the string center , satisfying ^(m)_. the numbers @xmath35 are deviations from a perfect string which are in general vanishing in the thermodynamic limit . in particular , for the spin-@xmath0 case they are exponentially vanishing with the system size , while it is known that in the spin-@xmath1 case they exhibit a slower power law decay @xcite . within the so called string hypothesis , all these deviations are neglected . in some cases , solutions of that do not satisfy are known @xcite ; however it is widely believed that their contributions to the computation of physical quantities is vanishing in the thermodynamic limit .
the validity of the string hypothesis will always be assumed in this work . substituting into , we obtain the following set of bethe - takahashi equations 2i^(m ) _ = n ^(s)_m(_^(m ) ) - _ n _ _ m , n(^(m)_- ^(n ) _ ) , [ eq : bethetaka ] where as before we used the index @xmath36 to distinguish between the spin-@xmath0 and spin-@xmath1 case .
here we used the following definitions ^(1/2)_m ( ) & = & _ m ( ) , + ^(1)_m ( ) & = & _ m-1()+_m+1 ( ) , + _
m , n()&=&(1-_nm)_|n - m|()+2_|n - m|()+ +2_n+m-2()+_n+m ( ) , where _ n ( ) = 2,n1 , and @xmath37 . in the thermodynamic limit
n & & , + m & & , + & = & d , the number of rapidities for a given eigenstate grows to infinity .
the @xmath38-string centers then form a dense set in the interval @xmath39 $ ] and can be described by smooth distribution functions @xmath40 .
one also needs to introduce the hole distribution functions @xmath41 : they are a generalization to the interacting case of the hole distributions of an ideal fermi gas at finite temperature @xcite . following the well - known steps of the standard treatment of bethe ansatz @xcite , the thermodynamic limit of eqs . is readily obtained and reads ^(s)_m()+^(s)_h , m()=b^(s)_m()-_n=1^(a_mn^(s)_n ) ( ) .
[ eq : thermo_bethe ] here we have once again introduced the index @xmath36 to distinguish between spin-@xmath0 and spin-@xmath1 and we have defined a_mn()=(1-_mn)a_|m - n|()+2a_|m - n|()+ +2a_m+n-2()+a_m+n ( ) , where a_n()= _ n()= , [ def : a_function ] and b_n^(1/2 ) ( ) & = & ^(1/2)_n()= , [ b_1_function ] + b_n^(1 ) ( ) & = & ^(1)_n()= ( + ) .[b_2_function ] we furthermore used the following notation for the convolution of two functions ( fg)()=_-/2^/2df(-)g ( ) .
[ convolution ] the bethe equations characterize the eigenstates of the model in the thermodynamic limit .
given a particular solution corresponding to the set @xmath42 , the density of quasi - particle ( magnons ) @xmath43 and the energy per unit length @xmath44 can be directly computed as d^(s)&=&_n=1^_-/2^/2dn ^(s)_n(),[eq : density ] + e^(s)&=&_n=1^_-/2^/2d ^(s)_n()^(s)_n()[eq : energy ] , where ^(s)_n()=-(2s)(2s)_l=1^n.[def : varepsilon ] in the following , it will also be useful to define the density and energy of the quasi - particles forming @xmath38-strings ( namely @xmath38-particle bound states ) as d_n^(s)&=&n_-/2^/2d_n(),[density_string ] + e_n^(s)&=&_-/2^/2d_n()_n^(s)().[energy_string ] at the isotropic point , namely for @xmath45 , the picture outlined so far requires slight modifications . in particular ,
the support of the distribution functions @xmath40 , @xmath41 extends in this case to the whole real line .
furthermore , it can be seen that all the results concerning the thermodynamic limit of the isotropic spin chains can be obtained from the corresponding anisotropic regimes by the scaling & & 0 , + & & , + _ n ( ) & & _
n( ) , + _
h , n ( ) & & _ h , n( ) .
the bethe equations are then still valid , provided that @xmath46 and @xmath47 in - are substituted with the rational functions a_n()&= & , + b^(1/2)_n()&= & , + b^(1)_n()&=&. analogously , the energy per unit length in the isotropic case is given by ^(s)_n()=-(8s^2)_l=1^n .
we conclude this section with two standard definitions , which we report here for future reference .
they read ^(s)_t , n()&=&^(s)_n()+^(s)_h , n ( ) , + ^(s)_n()&=&. [ def_eta ] as we have already mentioned , a promising analytical approach , the quench action method , has been recently introduced in ref . @xcite and successfully applied in the study of quantum quenches in many integrable models @xcite .
one of its main achievements has been to show the existence of a representative eigenstate which effectively captures , in the thermodynamic limit , the local properties of the system at long times after a quench .
more specifically , given an initial state @xmath48 and under mild assumptions , it was shown in @xcite that there exists an excited eigenstate @xmath49 of the post - quench hamiltonian such that for any _ local _ observable @xmath50 _ t= , [ eq : asymptotic ] where we employed the standard notation @xmath51 for time evolved operators in the heisenberg picture .
note that @xmath49 depends on @xmath48 but not on @xmath50 .
given @xmath48 , the quench action approach provides a prescription on how to determine the representative eigenstate @xmath49 . an essential role is played by the computation of the overlaps between @xmath48 and the eigenstates of the post - quench hamiltonian ( or more precisely , their thermodynamically leading part ) .
once the representative eigenstate is known , the asymptotic values of the time dependent local correlation functions can be in principle computed from . in general
, the computation of the overlaps remains very hard and so far has been only performed in a few special instances @xcite .
when these are not known , the prescriptions of the quench action method can not be applied .
however , the nature of the limitations arising in these cases is believed to be purely technical and one still expects the existence of a representative eigenstate satisfying eq . .
in this work we employ a method originally developed in refs .
@xcite to determine the post - quench representative eigenstate which can also be applied when the overlaps are not known .
this approach is based on the knowledge of a complete set of quasi - local charges and will be explained in detail in section [ sec : derivation ] .
it is important to stress that it is at the moment limited to those integrable spin chains for which these quasi - local charges are known . from the discussions of the previous sections
, it follows that in the models considered here the characterization of an excited state in the thermodynamic limit is equivalent to providing a set of rapidity distribution functions @xmath52 .
the goal is then to write down these functions for the post - quench steady state from a given initial state . in the next section we present an explicit closed - form solution of this problem for simple families of physically interesting initial states .
the first family of initial states that we considered is that of the tilted ferromagnet .
it is defined as |;=^n .
[ ferromagnet ] the angle @xmath53 is chosen to be 0</2 , which corresponds to restricting to the sector of states with @xmath54 , where @xmath55 is the density of magnons .
the analytical expressions for @xmath56 and @xmath57 of the post - quench steady state are _ 1()&=&-1 + [ tf_eta1 ] , + _ h,1()&= & ( . + & & .
-),[tf_rhoh1 ] where t_1 ( ) & = & ( ) ( 4 ( ) -2 ( 2 ) ^2 + 3 ( 2 ) +
1 ) , + ( ) & = & 2 ^2 ( + i ) ( - i ) , + @xmath58 and @xmath41 for @xmath59 are directly related to @xmath56 and @xmath57 through the analytical relations _ n()=-1 , [ eq : eta_relation ] _ h , n()=_h , n-1(+i/2)[1+^-1_n-1(+i/2)]+_h , n-1(-i/2)[1+^-1_n-1(-i/2)]-_h , n-2 ( ) [ eq : holes_relation ] , where we used the conventions @xmath60 , @xmath61 . finally , the functions @xmath40 are trivially obtained by @xmath62 . for the tilted ferromagnet state for @xmath63 .
the plots correspond to @xmath64 and @xmath65 : @xmath66 , @xmath67 : @xmath68 .
the functions are shown for @xmath69 being symmetric w.r.t .
@xmath70 . ] of the quasi - particles forming @xmath38-strings [ as defined in ] for the tilted ferromagnet state .
plot @xmath65 corresponds to @xmath71 and shows the dependence on the angle @xmath53 , while plot @xmath67 corresponds to @xmath68 and shows the dependence on the anisotropy parameter @xmath72 .
] plots for the rapidity distribution functions @xmath40 are reported in fig .
[ fig:1a ] .
interestingly , we observe that they are peaked around @xmath73 , which can be heuristically understood as follows . in the gapped regime @xmath74 , the ground state of the hamiltonian displays antiferromagnetic order , as opposed to the tilted ferromagnet state .
then , one might think to obtain the latter from the ground state by adding an infinite number of spin wave excitations of minimum wave length , which corresponds to the maximum allowed rapidity @xmath73 .
note that for the representative eigenstate of the nel state , which instead exhibit antiferromagnetic order , the rapidity distribution function @xmath75 is peaked around @xmath70 as expected ( cf .
section [ subsec : tilted_neel ] ) . in fig .
[ fig:1b ] the ( normalized ) densities @xmath76 for the tilted ferromagnet state [ as defined in ] are displayed for different values of the angle @xmath53 and the anisotropy @xmath72 .
in general we see that @xmath38-string with @xmath59 are not negligible and indeed the values of @xmath72 and @xmath53 can be tuned in such a way that the largest contribution to the density is given by @xmath38-strings with @xmath59 .
a similar picture was found in the context of quantum quenches to the attractive lieb - liniger bose gas from a non - interacting condensate @xcite
. there , different values of the final interaction parameter yielded different compositions of the post - quench steady state in terms of @xmath38-particle bound states . note that in the case studied here , the densities of bound states in the post - quench steady state can be also varied by tuning the parameter @xmath53 of the initial state , and not only the value of the final interaction @xmath72 . as we will discuss in section [ sec : correlations ] the large densities of @xmath38-strings have consequences on the asymptotic values of local correlation functions after the quench .
the second family of initial states that we consider is that of the tilted nel state .
it is defined as |;=()^n/2 .
[ tilted_neel ] the analytical expressions for @xmath56 and @xmath57 are _ 1()&=&-1 + , + _
h,1 ( ) & = & - x_1 ( + i ) - x_1 ( - i ) , + where t_1 ( ) & = & - ( ) , + ( ) & = & ( 2 + i ) , + the functions @xmath58 , @xmath41 with @xmath59 are explicitly obtained by eqs .
( [ eq : eta_relation ] ) and ( [ eq : holes_relation ] ) .
note that for @xmath77 , we recover the known analytical results for the nel state @xcite . , @xmath78 for the tilted nel state for different values of @xmath53 .
the plots correspond to @xmath79 .
the functions are shown for @xmath69 being symmetric w.r.t .
@xmath70 . ] of the quasi - particles forming @xmath38-strings [ as defined in ] for the tilted nel state .
plot @xmath65 corresponds to @xmath79 and shows the dependence on the angle @xmath53 , while plot @xmath67 corresponds to @xmath80 and shows the dependence on the anisotropy parameter @xmath72 . ]
the rapidity distribution functions @xmath40 are displayed in fig .
[ fig:2a ] while we report in fig .
[ fig:2b ] the ( normalized ) densities @xmath76 as defined in .
we see that the post - quench steady state for tilted nel state maintains the qualitative features of the one corresponding to the nel state ( which is obtained for @xmath77 ) .
in particular , for all the values of @xmath53 the majority of the quasi - particles remains unbound while @xmath38-strings provide smaller contributions to the total density .
this is a very important physical difference between quenches starting from tilted ferromagnets and antiferromagnets which also strongly affects the expectation values of observables in the stationary state .
the simplest initial state to be considered in the spin-@xmath1 integrable chain is the following zero - magnetization product state @xmath56 and @xmath57 are _ 1 ( ) & = & + _ h,1 ( ) & = & .
the functions @xmath58 , @xmath41 with @xmath59 are analytically obtained by the recursive relations and . ,
@xmath78 for the spin-@xmath1 zero - magnetization product state state for different values of @xmath23 .
the functions are shown for @xmath69 being symmetric w.r.t .
@xmath70 . ] of the quasi - particles forming @xmath38-strings [ as defined in ] for @xmath65 : the spin-@xmath1 zero magnetization product state , @xmath67 : the spin-@xmath1 nel state . ]
the rapidity distribution functions @xmath75 , @xmath78 are displayed in fig .
[ fig:3a ] for different values of the anisotropy parameter @xmath23 , while the densities @xmath76 defined in are reported in fig .
[ fig:3b ] .
we see that the composition of the post - quench steady state is dominated by @xmath81-strings , as it is also the case for the spin-@xmath1 nel state discussed in the next section .
this can be easily understood as follows . in the limit @xmath82 ,
the eigenspace with lowest energy of the hamiltonian is @xmath83-fold degenerate ; it is generated by the state @xmath84 and the two realizations of the spin-@xmath1 nel state ( which are obtained one from the other by a @xmath1-site shift translation ) .
then , for large values of @xmath23 the string content of the post - quench steady states for @xmath84 and the spin-@xmath1 nel state will be similar to that of the ground state of the hamiltonian . in turn , this is composed solely of @xmath81-strings @xcite .
the physical interpretation of this lies in the antiferromagnetic order of the ground state : in the spin-@xmath1 case , antiparallel ordering of the spins can be heuristically thought of as bound states of two down spins every other site .
as expected , we also see from fig .
[ fig:3b ] that decreasing the value of the anisotropy parameter @xmath23 , the density of @xmath81-strings in the post - quench steady state decreases , even though it remains dominant .
the second initial state that we consider in the spin-@xmath1 integrable chain is a straightforward generalization of the spin-@xmath0 nel state .
it is defined as |= ( | | ) ^n/2 .[spin_1_neel ] the analytical expressions for @xmath56 and @xmath57 are simply _ 1 ( ) & = & , + _ h,1 ( ) & = & .
the functions @xmath58 , @xmath41 with @xmath59 are obtained by the recursive relations and .
, @xmath78 for the spin-@xmath1 nel state for different values of @xmath23 .
the functions are shown for @xmath69 being symmetric w.r.t .
@xmath70 . ]
the rapidity distribution functions @xmath75 , @xmath78 are displayed in fig .
[ fig:3c ] for different values of the anisotropy parameter @xmath23 , while the densities @xmath76 defined in are reported in fig .
[ fig:3b ] .
we refer the reader to the previous subsection for a discussion of the features displayed by the post - quench steady state rapidity distributions .
we now focus on the isotropic limit @xmath45 of , namely the babujian - takhtajan hamiltonian ( [ hamiltonian2b ] ) . in this regime ,
the zero - magnetization product state becomes an eigenstate of the hamiltonian and has therefore no quench dynamics .
this is not the case for the spin-@xmath1 nel state .
the corresponding analytical expressions for @xmath56 and @xmath57 are _ 1 ( ) & = & , + _ h,1 ( ) & = & . the functions @xmath58 , @xmath41 with @xmath59 are now analytically obtained as _ n()=-1 [ eq : eta_relation_iso ] _ h , n()=_h , n-1(+i/2)[1+^-1_n-1(+i/2)]+_h , n-1(-i/2)[1+^-1_n-1(-i/2)]-_h , n-2 ( ) , [ eq : holes_relation_iso ] where @xmath60 , @xmath61 . once again , the rapidity distribution functions are then trivially obtained as @xmath62 .
the method employed to derive the analytical results presented in the previous section was introduced in the recent works @xcite .
there , it was also shown that it leads to the same results as the quench action approach in the cases where the overlaps are known and the latter can be applied . assuming the existence of a post - quench representative eigenstate ( cf . section [ sec : representative_eigenstate ] ) , the idea is that the latter can be uniquely fixed by the constrains resulting from all the conservation laws of the model .
a crucial requirement is that the conserved operators that are considered must form a complete set . as we already discussed in the introduction , the known local charges for bethe ansatz integrable models are not complete and quasi - local conserved operators have to be taken into account . in this section
we briefly review the construction of quasi - local charges and how their expectation values can be exploited to directly obtain the rapidity distributions @xmath40 of the post - quench representative eigenstate .
we then discuss the closed form analytical formulas presented in the previous section and comment on the possibility of extending these results for more general states .
the logic behind the construction of quasi - local conserved operators is similar to that underlying the well known derivation of local charges .
the central object is the so called lax operator @xmath85 , acting on the tensor product of two local hilbert spaces @xmath86 .
the space @xmath87 is the local physical space associated with a single spin in the chain , while @xmath88 is the auxiliary space . in the standard algebraic bethe ansatz construction ,
@xmath87 and @xmath88 are isomorphic and this defines the fundamental @xmath85-operator .
conversely , in the construction of quasi - local conserved charges one allows for more general auxiliary spaces resulting in a non - fundamental @xmath85-operator . in the integrable spin-@xmath0 and spin-@xmath1 chains considered here , the generic @xmath85-operator can be directly written down in a compact form . in the spin-@xmath0 case , in the local spin basis @xmath89 of the physical space @xmath87 , the lax operator is written as @xmath90_q & s^{j}_-\\ s^{j}_+ & [ -i\lambda/\eta - s^{j}_z]_q \end{array } \right)\ , , \label{non_fund_1/2}\end{aligned}\ ] ] where we employed the notation in .
analogously , in the spin-@xmath1 case , the lax operator @xmath91 can be written as a @xmath92 matrix in the local spin basis @xmath93 , whose entries @xmath94 are operators acting on the auxiliary space .
they are given by l^(1,j)_11()&=&[-i/+1/2 + s^j_z]_q [ -i/-1/2 + s^j_z]_q,[l_1 ] + l^(1,j)_12()&=&[2]_q^1/ 2s^j_-[-i/-1/2 + s^j_z]_q,[l_2 ] + l^(1,j)_13()&=&(s^j_-)^2,[l_3 ] + l^(1,j)_21()&=&[2]_q^1/ 2s^j_+[-i/+1/2+s^j_z]_q,[l_4 ] + l^(1,j)_22()&=&s^j_+ s^j_- + [ -i/+ 1/2 + s^j_z]_q [ -i/-1/2- s_z]_q,[l_5 ] + l^(1,j)_23()&=&[2]_q^1/ 2s^j_-[-i/+1/2-s^j_z]_q,[l_6 ] + l^(1,j)_31()&=&(s^j_+)^2,[l_7 ] + l^(1,j)_32()&=&[2]_q^1/ 2s^j_+[-i/-1/2-s^j_z]_q,[l_8 ] + l^(1,j)_33()&= & [ -i/+1/2 - s^j_z]_q [ -i/-1/2- s^j_z]_q.[l_9 ] the lax operator @xmath91 defined above for the spin-@xmath1 case can be obtained from using the so called fusion procedure as explained for example in @xcite .
the operators @xmath95 [ not to be confused with the spin operators of the hamiltonians , ] act on the auxiliary space @xmath96 which for the purposes of this work has to be chosen as a unitary representation of the quantum group @xmath97 @xcite .
in particular , one considers the space @xmath98 generated by the vectors @xmath99 , with m=- , , , and = j+1 . here
@xmath7 labels the different representations corresponding to the auxiliary space and has to chosen as a positive integer , @xmath100 .
the operators @xmath95 then act on the basis vectors @xmath101 as s^j_z |m&= & m |m , + s^j_+ |m&= & |m+1 , + s^j_- |m&= & |m-1 . in the following
we will also make use of the following compact notation ^(s , j ) ( ) = _ a , b l_ab^(s , j ) ( ) e^ab , where as usual @xmath36 is the spin label , while @xmath102 and @xmath103 take the values @xmath104 . in the above expression @xmath105 are @xmath106 matrices acting on the local physical space and are defined by ( e^ab)_cd=_ac_bd . the construction of quasi - local charges proceeds by introducing the transfer matrix corresponding to the lax operators defined above , namely t^(s)_j ( ) = \ { _ 0 n^(s , j ) ( ) _ 0 1^(s , j ) ( ) } = _ \{a_i},\ { b_i } \
{ l^(s , j)_a_n b_n() l^(s , j)_a_1 b_1 ( ) } _
i=1^n e_i^a_i , b_i .
[ eq : productformtj ] here the trace is over the auxiliary space @xmath96 while the sum appearing in the r.h.s . is over all the sequences @xmath107 , @xmath108 with @xmath109 .
notice that for @xmath110 one recovers the known transfer matrix used in the standard algebraic bethe ansatz construction @xcite .
further , it can be seen that = 0 .
[ commutation ] the quasi - local conserved charges can then be defined as x^(s)_j ( ) & = & _ , [ def_charges ] where we used t_0^(1/2)()&=&^n_q , + t_0^(1)()&=&^n_q^n_q . due to
eq . , the operators @xmath111 commute with the hamiltonian of the model and are thus conserved .
they are not local in the sense that they can not be written as a sum over the chain of finitely supported operator densities .
however , they are quasi - local in the domain -<im()<. in this work , we do nt enter into the issue of providing a precise definition of quasi - locality . the only property that will be important for our purposes
is that the charges are extensive .
more precisely , their expectation value on eigenstates grows linearly with the system size , analogously to the case of previously known local charges .
in particular , the expectation value of the charges on bethe states was explicitly computed in ref .
@xcite for the spin-@xmath0 case , and it is straightforwardly generalized to the spin-@xmath1 case . the final result is expressed as follows _
n & = & _ m=1^ _ - ^ _ m^(s)()g_j , m(- ) , [ eq : exp_charges ] where g_j , m()=_k=1^min(j , m)a_|j - m|-1 + 2k ( ) , and where @xmath112 is defined in .
here we indicated with @xmath113 an eigenstate of the finite system of length @xmath114 which , in the thermodynamic limit , corresponds to the rapidity distribution functions @xmath40 .
remarkably , it was shown in @xcite that relation can be inverted .
in particular , from the expectation values of the quasi - local charges @xmath115 one can directly determine the corresponding rapidity distribution functions @xmath40 .
explicitly , one has _
j^(s ) ( ) & = & x_j^(s)(+i)+x_j^(s)(-i)-x_j+1^(s)()-x_j-1^(s ) ( ) .
[ eq : rhox ] the hole rapidity distribution functions can also be explicitly written as _ h , j^(s ) ( ) & = & b_j^(s)()- x_j^(s)(+i ) - x_j^(s)(-i ) , [ eq : holesx ] where @xmath116 is defined in , for @xmath30 and @xmath31 respectively . in equations and we used with a slight abuse of notation the same symbols for the operators @xmath117 and their expectation values .
equations and are a key result of the method developed in @xcite .
they state that the rapidity and hole distribution functions of the representative eigenstate are explicitly obtained in terms of the expectation values of the quasi - local charges . in the next section
we review the procedure to compute these expectation values on simple initial product states .
given an initial state @xmath48 , it is now evident from equations , that the problem of determining the post - quench steady state is reduced to the computation of the expectation value of the quasi - local charges .
we now focus on the families of initial states considered in this work .
they are product states of the form | _ 0 = |_0 ^(n / n_p ) , [ periodic_product ] where @xmath118 is a vector in the tensor product of local hilbert spaces @xmath119 . for these states ,
the computation of the expectation values of the charges was performed in @xcite using the methods previously applied in @xcite .
we now briefly review this computation , which is straightforwardly generalized also to the spin-@xmath1 case .
first , it is convenient to introduce the operators ^(s)_j ( ) & = & , [ eq : defxj ] where ^(1/2)_j ( ) & = & _ q_q , + ^(1)_j ( ) & = & _ q_q _ q_q , [ eq : epsilon12 ] and where we used the notation . in the large @xmath114 limit @xmath120 can be related to @xmath121 through the so - called inversion relation @xcite , [ inversion_relation ] which can be established by showing that for @xmath122 ||- constant while @xmath123 denotes the hilbert - schmidt norm ; for a generic operator @xmath124 acting on the hilbert space @xmath125 the latter is defined as space @xmath125 and where @xmath126 $ ] .
equation can be derived with calculations analogous to those presented in @xcite for the spin-@xmath0 and spin-@xmath1 cases . using it
is now straightforward to obtain the following relation which holds in the large @xmath114 limit , - _ , [ eq : aux_2 ] where _ 0^(1/2 ) ( ) & = & _ q + _
0^(1 ) ( ) & = & _ q _ q. the expectation value of @xmath120 can now be performed using standard techniques @xcite . in particular , for the state one can easily derive _
0 | ^(s)_j ( ) |_0 = |_x=0 _ j j \ { ^(s)_j ( ^- , ^++x ) } ^n / n_p , [ eq : aux ] where we introduced the notation ^=i , [ lambda_pm ] and where the matrix @xmath127 acts on the tensor product @xmath128 of two auxiliary representations . it is defined as ^(s)_j ( , ) = _ \ { a_},\{b _ } , \{c _ } ( _ = 1^n_p l^(s)_a_b_()l^(s)_b_c _ ( ) ) _ 0 | e_1^a_1 c_1
e_n_p^a_n_p c_n_p @xmath129 , @xmath130 , @xmath131 with @xmath132 . at large values of @xmath114 , the trace in the r.h.s . of is dominated by the eigenvalue of @xmath133 with the largest absolute value .
this observation leads to the possibility of an explicit expression in the limit @xmath122 . in refs .
@xcite this was explicitly obtained in terms of the so called jacobi formula .
here we simply report the final result , which reads _ j ( ) , [ eq : aux_3 ] where _
j()= , [ eq : jacobi ] and where we employed notation .
here we defined [ m]_ij=(-1)^i+jmin[m]_ji , where @xmath134_{lm}$ ] is the determinant of the matrix obtained from @xmath32 by removing line @xmath135 and column @xmath33 . putting everything together , equations , and
explicitly yield the expectation values of the charges @xmath136 on the initial state in the thermodynamic limit . from the results of sections [ sec : quasi_local ] and [ sec :
expectation ] , one can obtain , at least numerically , all the distribution functions @xmath40 and @xmath41 of the post - quench steady state for a given initial state of the form .
however , even for simple product states the use of jacobi formula becomes increasingly time - consuming for large @xmath38 .
this represents a non - negligible technical issue if one is interested in the local correlation functions on the post - quench steady state .
indeed , the computation of the latter to sufficient numerical precision typically requires the knowledge of the functions @xmath137 for large values of @xmath38 .
the quench from the nel state was first analyzed by means of the quench action method @xcite . using the latter ,
an additional analytical relation was established for the functions @xmath58 characterizing the post - quench steady state @xcite .
n()=-1 , [ y_system ] where one sets @xmath60 .
this relation allows one to obtain directly the functions @xmath58 from the single function @xmath56 .
it corresponds to the so called y - system , which is an ubiquitous structure in integrable models @xcite . using the quench action approach ,
the recursive relation was found to be valid also for the majumdar - ghosh state and its @xmath13-deformed version @xcite .
interestingly , the same relation was also derived for the steady state in the quench from a non - interacting initial state to the attractive lieb - liniger model @xcite .
it is then natural to conjecture that such a relation can be verified for more general classes of initial states .
this idea was first stated in ref @xcite . for the states of interest in this work
, we verified numerically the validity of equations .
in particular , we first computed numerically @xmath58 with the procedure outlined in the previous section up to @xmath138 .
we then compared each @xmath58 with the function obtained by exploiting equation , finding perfect agreement . once the functions @xmath58 are known , the bethe equations can be used to obtain an expression for the rapidity and hole distribution functions @xmath40 and @xmath41 .
in particular , the bethe equations can be cast in the partially decoupled form @xcite ^(s)_n(1+^(s)_n)=_n,(2s)+(^(s)_n-1^(s)_n-1+^(s)_n+1^(s)_n+1 ) , [ decoupled ] where @xmath139 and ( ) = ( 1 + 2_k=1^ ) .
equation can now be rewritten in the functional form @xcite _ h , n()=_h , n-1(+i/2)[1+^-1_n-1(+i/2)]+_h , n-1(-i/2)[1+^-1_n-1(-i/2)]-_h , n-2 ( ) [ eq : holes_relation_new ] , where we dropped the index @xmath36 and where we set @xmath140 . given the knowledge of @xmath56 and @xmath57 , eqs . and completely determine all the functions @xmath58 and @xmath41 for @xmath59 . in turn , @xmath56 and @xmath57 can be obtained from the method reviewed in the last section . for the simple states considered in this work
, the evaluation of the jacobi formula for @xmath141 can be performed analytically using the program mathematica .
making finally use of , one obtains the analytical expressions presented in section [ sec : results ] .
a non - trivial numerical check of our results is available : the density of magnons and the energy per unit length computed in the post - quench representative eigenstate from and have to be equal to the corresponding quantities of the initial state .
note that the calculations for the initial states can be done straightforwardly due to their simple product form , while the evaluation of the r.h.s . of , can be easily performed by numerical integration . in all cases
, we verified that the two calculations yield the same result within the expected numerical error .
we conclude this section by anticipating that the recursive relation does not hold for arbitrary initial states , as we will discuss in section [ sec : general_states ] .
the next section is instead devoted to the computation of local correlations in the post - quench steady state .
@xmath142 , @xmath67 @xmath68 .
the value of the anisotropy parameter is @xmath143 .
the solid lines represent the itebd data of ref .
the exact results for the asymptotic stationary values computed using the method described in the text ( dashed lines ) are displayed together with the ultra - local gge predictions computed in ref .
@xcite ( symbols ) .
different colors are used to distinguish different correlators . ] the set of rapidity distribution functions @xmath42 completely characterize the local properties of the corresponding state in the thermodynamic limit .
the results of section [ sec : results ] combined with eq .
, then allow in principle to compute the asymptotic value of all the local correlation functions at long times after the quench .
in practice , even in the presence of integrability , the computation of correlation functions is notoriously hard . in the xxz spin-@xmath0 model ,
many remarkable results have been obtained during the past decades for thermal and ground states @xcite .
however , only in the past few years this problem has started to receive attention for generic excited states , mainly because of its interest in the study of quantum quenches . in particular , to this end , integral formulas were recently presented in ref .
they were conjectured on the basis of a formal analogy between the results for nearest neighbor correlators from the bethe ansatz method and the transfer matrix approach @xcite .
although a rigorous proof is still missing , their validity has been numerically tested to high precision in ref .
@xcite and already applied to the study of quantum quenches in refs .
@xcite , where they were always found to be in excellent agreement with independent numerical methods . for completeness
, the integral formulas of @xcite are reported in appendix [ sec : app_correlators ] .
we applied the latter to provide predictions for the asymptotic values at long times after the quench of the local correlators ^_j^_j+k , k=1,2,3,=x , y , z , [ eq : correlators ] where @xmath144 are the pauli matrices . we compared our results with the numerical data of ref .
@xcite , which were obtained using the time - dependent density matrix renormalization group ( tdrmg ) @xcite and the infinite time - evolving block - decimation ( itebd ) @xcite algorithms . .
the value of the anisotropy parameter is @xmath143 .
the itebd results of ref .
@xcite ( solid lines ) are shown together with the exact results for the asymptotic stationary values ( dashed lines ) and the ultra - local gge predictions computed in ref .
@xcite ( symbols ) . ] in ref .
@xcite the numerical data for the time evolution of the correlators were compared with the predictions of the gge constructed using only the local charges of the system . in what follows
, we shall refer to the latter as the ultra - local gge .
we stress here that the existence of quasi - local charges relevant for relaxation processes after a quench was not known at the time . even though very good agreement was generally found for the states considered , in some cases deviations from the ultra - local gge predictions
were observed .
it is now clear that such deviations are due to the missing contributions of quasi - local charges , as it was established in a series of subsequent works @xcite .
the largest deviations from the ultra - local gge predictions were observed for the tilted ferromagnetic state . in figs .
[ fig:1 ] , [ fig:2 ] we show the numerical data of ref .
@xcite together with our predictions for the long time limit using the results of section [ sec : results ] and the integral formulas of ref .
our results are found to be in excellent agreement with the numerical data .
also , our predictions are significantly different to the values obtained using the ultra - local gge ( which are also displayed in the figures ) .
the large deviations of the gge predictions can in this case be understood in terms of high densities of @xmath38-strings with @xmath59 for the tilted ferromagnet state , cf .
[ fig:1b ] .
indeed , it was shown in ref . @xcite that fixing the value of ultra - local charges uniquely determine the hole distribution function @xmath57 .
if the rapidity distribution functions @xmath40 with @xmath59 are suppressed , then the bethe equations also yield @xmath75 if @xmath57 is known . in this case
it is reasonable to expect that the ultra - local gge provides a good approximation for the asymptotic values of time - dependent local correlation functions . while for the nel state at large @xmath72 higher strings are indeed suppressed , cf .
[ fig:2b ] , this is by no means true for the tilted ferromagnet , cf . fig .
[ fig:1b ] .
note further that the tilted ferromagnet state breaks @xmath145-invariance . on the other hand ,
the representative eigenstate is an eigenstate of the @xmath145-symmetry , since the hamiltonian is left invariant under rotation along the @xmath20-axis .
hence , the transverse correlators @xmath146 and @xmath147 are equal when computed on the representative eigenstate , namely the post - quench steady state . accordingly , as in ref .
@xcite we predict that transverse correlators should approach the same values at large times after the quench .
this is displayed in fig .
[ fig:2 ] : we observe that at the times accessible to the itebd algorithm @xmath145-invariance is not completely restored , even though transverse correlators are clearly oscillating around the same value .
@xmath148 , @xmath67 @xmath142 .
the value of the anisotropy parameter is @xmath149 .
the solid lines correspond to the tdmrg data of ref .
@xcite , which were computed for a chain of @xmath150 sites .
they are displayed together with the exact results for the asymptotic stationary values ( blue dotted lines ) and the ultra - local gge predictions computed in ref .
@xcite ( red dashed lines ) . ] for the quench from the nel state , it was shown in ref .
@xcite that the predictions of the ultra - local gge and the quench action method coincide at least up to the second order in @xmath151 . in the gapped regime
then the difference between the two is very small and tdmrg data can not in general be used to test the validity of one and not the other , as it was also observed in ref .
@xcite . in fig .
[ fig:3 ] , instead , we show that for the tilted nel state the difference between the two prediction becomes significant when the tilting angle @xmath53 increases . already in the time interval accessible to the tdmrg computations of ref .
@xcite , our results seem to be in better agreement with the latter when compared with the ultra - local gge values . here
we mention that also for the majumdar - ghosh state , studied in ref .
@xcite using the quench action method , a significant discrepancy between the predictions of the latter and the ultra - local gge was observed . in particular , it was found that tdmrg data were compatible with the quench action method predictions but not with those of the gge . for the spin-@xmath1 hamiltonians , ,
no result is available at present for the computation of correlation functions for arbitrary excited states . in the isotropic chain
integrable formulas were recently obtained within the quantum transfer matrix formalism for thermal states in refs .
accordingly , it is possible that the conjecture of ref .
@xcite could be generalized to the spin-@xmath1 case .
the results of section [ sec : results ] would then have an immediate application to provide predictions for the asymptotic value of local correlators .
this issue goes however beyond the scope of this work , and remains an interesting subject for future investigations .
in the work @xcite , an additional family of initial states was considered .
namely , the domain wall states defined on a chain of @xmath152 sites ( with @xmath153 , @xmath154 positive integers ) as |dw , p=|_p_p _p_p .
[ domain_wall ] for @xmath155 we recover the nel state , while for large @xmath154 this family might be of interest in the study of geometric quenches @xcite . note that this state is of the form .
hence , all the distribution functions @xmath40 , @xmath41 can be in principle numerically obtained using the procedure described in section [ sec : derivation ] .
however , we found that for @xmath156 the additional relation does not hold .
in particular , given @xmath56 the application of leads to the wrong result for the functions @xmath58 with @xmath59 , both at quantitative and qualitative level .
this is shown in fig .
[ fig:12 ] where we report the case @xmath157 as an example . in the figure
we compare the function @xmath78 as correctly computed using the general method reviewed in sections [ sec : quasi_local ] and [ sec : expectation ] with the result obtained assuming the validity of the @xmath158-system and it is evident that the two are very different .
the state then provides a physically interesting example for which the @xmath158-system does not hold . as a consequence , the computation of the distribution functions @xmath40
, @xmath41 necessarily becomes significantly time - consuming for large @xmath38 , since they can only be obtained after application of the jacobi formula .
also , this example shows that one can not a priori expect the validity of the @xmath158-system for arbitrary initial states . at this stage
it remains an open problem to understand which physical property of the initial states allow the additional analytical structure to hold ( see also ref .
@xcite , were the existence of states not satisfying the @xmath158-system was first observed ) .
the answer might be ultimately related to the structure of the overlaps with the eigenstates of the post - quench hamiltonian . for the domain wall state for @xmath157 .
the exact function computed using the method of sections [ sec : quasi_local ] , [ sec : expectation ] is compared with the result obtained assuming the @xmath158-system .
the plot corresponds to @xmath159 . ]
we have considered quantum quenches in spin-@xmath0 and spin-@xmath1 integrable chains from simple families of physically interesting initial states .
we focused on the cases for which the overlaps with the eigenstates of the model are not known and studied them using a recently developed approach based on the knowledge of complete sets of quasi - local charges .
namely we considered tilted ferromagnets , tilted antiferromagnets , and domain walls states for spin-@xmath0 chains and nel and zero magnetization product states for the spin-@xmath1 case . for all these initial states ( with the exception of the domain wall states ) , we provided a closed - form analytical characterization of the post - quench steady state which also allows for an efficient numerical evaluation of its rapidity distribution functions . in the spin-@xmath0 case , we computed the stationary value of local correlation functions at long times after the quench and found excellent agreement with the tdmrg and itebd calculations of ref . @xcite . in some cases a significant deviation of our results from the predictions of the ultra - local gge was observed ; this difference could be resolved by the tdmrg and itebd data within the accessible time scales of these methods . in the spin-@xmath1 case ,
no formula is yet available for the computation of local correlation functions in arbitrary excited states .
we hope that the existence of explicit results for the post - quench steady state presented in this work might stimulate investigations in this direction .
in particular , given the integral formulas for thermal correlators recently found in @xcite , one might expect the possibility of generalizing the results for the spin-@xmath0 hamiltonian presented in @xcite ( see also the very recent work @xcite ) . here
we mention that it should now be clear that the construction of unitary quasi - local charges considered in this work ( and first developed in @xcite ) , could be extended also to higher spin generalization of the spin-@xmath0 and spin-@xmath1 integrable chains that we have studied .
the main technical obstacle in the investigation of quantum quenches in these models then remains the computation of correlation functions in the post - quench steady state . on a more technical level
, it would be interesting to understand for which states a closed - form analytical solution of the quench problem , similar to those presented in section [ sec : results ] , exists .
as we have shown , this can not be expected in general .
the answer might be related to the structure of the overlaps between the initial state and the eigenstates of the model and remains at present an open problem .
finally , the efficiency of the method developed in @xcite and applied in this work should serve as a motivation to intensify the search for quasi - local conservation laws in models defined on the continuum .
this could have important ramifications in the study of quantum quenches in systems such as the attractive lieb - liniger gas @xcite or the sine - gordon field theory @xcite , where bound states exist .
indeed , it is at present not known whether quasi - local conserved charges analogous to those exploited here can be constructed in these models @xcite .
we would like to thank enej ilievski for insightful discussions during the last stage of this work and comments based on his unpublished results .
some of these were independently derived by us and have been presented in this work .
we thank mario collura for providing the tdmrg and itebd data presented in ref .
e.v . and p.c .
acknowledge the financial support by the erc under starting grant 279391 edeqs .
in this appendix we provide for completeness the integral formulas for local correlation functions which were used in our work , as reported in @xcite .
these formulas only require as an input the functions @xmath58 characterizing the post - quench steady state .
first , we introduce the set of auxiliary functions @xmath160 , @xmath161 as the solution of the following system of integral equations _
n^(a)()&=&-s^(a)_n()-_m=1^(_nm)(),[rho_a_aux ] + ^(a)_n()&=&^(a)_n()+_m=1^(_nm)()-_m=1^(_nm)(),[sigma_aux ] where we employed the notation for the convolution of two functions and where s^(a)_n()&=&()^a s^(0)_n ( ) , + ^(a)_n()&=&()^a ^(0)_n ( ) , + s^(0)_n()&= & , + ^(0)_n()&=&- , and _
jk()&=&- , + _ jk()&=&- .
next , one defines the parameters _ ab&=&-2_n=1^s^(b)_n , + _
ab&=&2(_n=1^_n^(b)+_n=1^s_n^(b ) ) , + _
ab&=&-(-1)^(a+b)/2_ab-(-1)^b()^a+bk()|_=0 , + w_ab&=&-(-1)^(a+b-1)/2_ab+(-1)^b()^a+b()|_=0 , where we introduced the notation fg=_-/2^/2df()g ( ) , and where k()&= & , + ( ) & = & .
finally , local correlators are given in terms of algebraic expressions involving the parameters @xmath162 and @xmath163 which are identical to those of the thermal case @xcite .
for nearest and next - nearest neighbors , they read _ 1^z_2^z&=&()_00+w_10 , + _ 1^z_3^z&=&2(2)_00+w_10+()(_20 - 2_11)-^2()w_21 , + _
1^x_2^x&=&--w_10 , + _ 1^x_3^x&=&--w_10-()(2)(_20 - 2_11)+^2()w_21 .
the formulas for @xmath164 with @xmath165 are analogous but significantly longer and we do nt report them here .
they are formally the same as in the thermal case , and can be found in @xcite .
note that these formulas involve sums over an infinite number of strings , which have to be truncated when numerically evaluated . in general the number of strings required for high numerical precision might change drastically from the different initial states and with the parameters of the post - quench hamiltonian .
for example , even at large values of @xmath72 , for the titled nel state we found that an increasing number of strings is required as the tilting angle @xmath53 grows ; in particular , the calculations corresponding to the results reported in fig . [ fig:3 ] required to consider @xmath166 strings .
in these cases where large numbers of strings have to be kept , auxiliary functions are most efficiently computed using the equivalent partially decoupled form of , which can be found in @xcite .
p. calabrese , f. h. l. essler , and m. fagotti , phys .
lett . * 106 * , 227203 ( 2011 ) ; + p. calabrese , f. h. l. essler , and m. fagotti , j. stat . mech .
( 2012 ) p07016 ; + p. calabrese , f. h. l. essler , and m. fagotti , j. stat .
( 2012 ) p07022 .
t. kitagawa , a. imambekov , j. schmiedmayer , and e. demler , new j. phys .
* 13 * 073018 ( 2011 ) ; + t. langen , m. gring , m. kuhnert , b. rauer , r. geiger , d. a. smith , i. e. mazets , and j. schmiedmayer , eur .
j. special topics * 217 * , 43 ( 2013 ) .
n. kitanine , j. m. maillet , and v. terras , nucl .
b * 567 * , 554 ( 2000 ) ; + n. kitanine , j. m. maillet , n. a. slavnov , and v. terras , j. phys .
a : math . gen . * 35 * , l753 ( 2002 ) ; + n. kitanine , j. m. maillet , n. a. slavnov , and v. terras , nucl .
b * 641 * , 487 ( 2002 ) .
a. klmper , _ integrability of quantum chains : theory and applications to the spin-1/2 xxz chain _ , in quantum magnetism , ed .
u. schollwck et al . , lecture notes in physics * 645 * , 349 , springer berlin heidelberg ( 2004 ) ; + a. klmper , arxiv : cond - mat/0502431 ( 2004 ) . | the study of quantum quenches in integrable systems has significantly advanced with the introduction of the quench action method , a versatile analytical approach to non - equilibrium dynamics .
however , its application is limited to those cases where the overlaps between the initial state and the eigenstates of the hamiltonian governing the time evolution are known exactly .
conversely , in this work we consider physically interesting initial states for which such overlaps are still unknown . in particular
, we focus on different classes of product states in spin-@xmath0 and spin-@xmath1 integrable chains , such as tilted ferromagnets and antiferromagnets .
we get around the missing overlaps by following a recent approach based on the knowledge of complete sets of quasi - local charges .
this allows us to provide a closed - form analytical characterization of the effective stationary state reached at long times after the quench , through the bethe ansatz distributions of particles and holes .
we compute the asymptotic value of local correlations and check our predictions against numerical data . |
darwinian adaptation spontaneously emerges from the processes of mutation , reproduction and competition , and these features need to be mirrored in any model of adaptation . in models , spontaneous mutations can be represented by a stochastic jump process in a `` fitness space '' .
reproduction is naturally described by a branching process by which individuals give birth at certain fitness - dependent rates @xcite . in combination , reproduction and mutations
thus generate a branching random walk @xcite , which by itself would lead to diverging population sizes . to avoid this unrealistic outcome , models of adaptation also encode a constraint on population sizes to account for the competition for finite resources .
the resulting process is a branching random walk subject to a global constraint , which we now frame mathematically .
the state of the population at time @xmath0 is described by a function @xmath1 representing the number density of individuals with fitness @xmath2 . in this context
, fitness refers to an individual s net - growth rate in the absence of competition for resources .
the population is assumed to evolve in discrete timesteps of size @xmath3 , which is eventually sent to zero in order to obtain a continuous - time markov process .
each timestep consists of two sub - steps .
the first substep realizes reproduction and mutations and the second substep implements competition .
the combined effect of reproductions and mutations can be described by the stochastic equation @xmath4 which takes the number density @xmath5 to an intermediate value @xmath6 .
the term @xmath7 represents the _ expected _ change in density due to reproduction and mutations .
this term is linear in the number density because the number of offspring and mutants per timestep is proportional to the current population density .
the term @xmath8 represents all sources of noise arising in this setup .
we will now discuss separately the precise meaning of both terms , and give natural alternatives for their form . the liouville operator @xmath9 depends on how the mutational process is modeled , and various examples are discussed in the following .
a particularly simple example is provided by @xmath10 , which has been used to model asexual evolution on a continuous fitness landscape @xcite . here
, the diffusion constant @xmath11 quantifies the fitness variance per generation , generated by an influx of novel mutations . to account for the notorious observation that most mutations are deleterious , a drift term @xmath12 is often included .
the linear `` reaction '' term @xmath13 in @xmath9 simply accounts for the fact that individuals with higher growth rate @xmath2 grow faster .
the term @xmath14 refers to the mean fitness of the population , which separates the population with a positive net growth rate @xmath15 from the less fit part of the population with @xmath16 .
one can not generally assume that ( biased ) diffusion is a good model for discrete mutational events because of the presence of the reaction term favoring highly fit individuals .
the diffusion approximation requires that mutation rates are higher than the typical fitness effects of novel mutations .
this may apply to rapidly mutating organisms , such as viruses , or close to a dynamic mutation - selection balance @xcite .
it may also effectively apply in island models with low migration rates , where the fitness effect of a mutation is reduced by potentially low migration rates .
but , in well - mixed populations , the diffusion approach breaks down when beneficial mutation rates are much smaller than their typical effect , which has been confirmed for a number of microbial species when they adapt to new environments @xcite .
more generally , asexual adaptation may , therefore , be cast into the form @xmath17 where the mutational process is described by the operator @xmath18 , which conserves particle numbers , i.e. , describes a pure jump process .
for instance , one may have one of the time - independent kernels @xmath19\ ; , & \text{staircase kernel } \\\int \mu(y ) \left [ c_t(x - y)-c_t(y ) \right ] dy\ ; , & \text{general mutational kernel}\end{matrix } \;\right.\ ] ] here @xmath20 is a mutation rate and @xmath21 is a characteristic scale for the mutational effect .
the diffusion kernel is the simplest of these kernels because it is characterized by only one compound parameter , the diffusion constant @xmath22 , rather than two in the staircase kernel or an entire function @xmath23 in the general case . the stochastic term @xmath24 in equation accounts for all random factors that influence the reproduction process .
the function @xmath25 represents standard white noise , i.e. a set of delta correlated random numbers , @xmath26 where @xmath27 denotes the ensemble average of a random variable @xmath28 , and @xmath29 and @xmath30 are the dirac delta function and the kronecker delta , respectively . the amplitude @xmath31 of the noise term in eq .
is typical for number fluctuations : due to the law of large number , the expected variance in population numbers from one timestep to the next is proportional to the number @xmath32 of expected births or deaths during one timestep .
the numerical coefficient @xmath33 is the variance in offspring number per individual .
for instance , when we assume that offspring have nearly matching division and death rates one finds @xmath34 .
the variance in offspring number is typically assumed to be of order one , but could become much larger if offspring distributions are highly - skewed , as it is the case when few individuals produce most of the offspring @xcite . because the reproduction step generally changes population numbers , another sub - step , following the branching process ,
is required to enforce a constant population size denotes the _ mean fitness _ of the population , the action of the liouvillean @xmath35 does not change the _ expected _ number of individuals in the population .
however , fluctuations in the reproduction implemented by genetic drift will result in slight deviations from this expected outcome .
these deviations accumulate over time and either lead to extinction or an ever increasing population size .
] , @xmath36 in most models and experiments @xcite , this step is realized by a random culling of the population : individuals are eliminated at random from the population until the population size constraint is restored .
mathematically , the population control step can be cast into the form @xmath37 where @xmath38 represents the fraction of the population that has to be removed to comply with the population size constraint .
the second sub - step completes the computational timestep , and takes the concentration field from the intermediate state @xmath39 to the properly constrained state @xmath40 .
the above standard model of adaptation with fixed population size represents a branching random walk subject to the constraint that the total population size is fixed . enforcing this constraint leads to the non - linearity that makes the associated model difficult to solve .
note that , in the above formulation , it is assumed that all noise comes from birth - death processes .
we have ignored , for simplicity , additional sources of noise due to , e.g. the mutational jump processes , which are sub - dominant in large populations .
while it is necessary to constrain the population dynamics to avoid an exponential run - away , there is no particular biological reason to strictly fix the population size - in fact , most population sizes fluctuate over time @xcite .
as we will see , there are , however , mathematical reasons to consider constraints of particular form , which greatly simplify the analysis . as a key step towards these tuned models , we note the fixed population size constraint eq .
can be viewed as one member of a whole class of linear constraints , @xmath41 that one could formulate with the help of a suitable weighting function @xmath42 .
any such constraint will be able to limit the population size , and thus defines , together with the liouvillean @xmath43 , a particular model of adaptation .
our main result will be the observation that there are an entire set of weighting functions for which the dynamics of the model becomes simple ( cf .
[ sec : main - results ] ) .
note that one recovers the fixed population size constraint of one chooses the weighting function @xmath42 to be a constant , @xmath44 . for any other choice
, the population size will not be fixed . at best , one obtains a steady state with a population size fluctuating around its mean value , @xmath45 , which may change depending on the time - dependence imposed on the weighting function @xmath46 .
note that culling is does not discriminate among individuals of different fitness .
it is only the _ amount _ of culling that depends on the distribution and type of individuals if @xmath42 is @xmath2dependent .
one advantage of using the general form eq . for a global constraint
is that many types of traveling waves arising in ecology and evolution can be cast in the same mathematical form , if an appropriate liouville operator and weighting function are used
. indicated in green .
note that for conventional models of adaptation with fixed population size , the growth rates are increasing with fitness and the weighting function is a constant .
by contrast , for models of species invasion or range expansion without allee effect @xcite , the growth rates and weighting function have a sigmoidal shape , saturating in the tip of the wave .
this ensures that individuals in the tip of the wave have the highest growth rates ( because of least competition ) and that their total number stays finite . ]
we would like to give an example of this assertion .
[ fig : noisy - wave - schema]b illustrates the expansion of a population in a real landscape , which may describe an advantageous gene spreading through a population distributed in space , or the invasion of virgin territory by an introduced species .
models of such real - space waves require the following features : i ) populations reproduce and die `` freely '' in the tip of the wave , where population densities are small , ii ) individuals move in real - space according to some jump process and iii ) the net - growth vanishes in the bulk of the wave , where resources are sparse . features i ) and ii ) again generate a branching - random walk in the tip of the wave , however , according to different position - dependent growth and jump rates than in the case of evolution .
feature iii ) requires a finite population size in the tip of the wave .
this non - linearity keeps the branching - random walk away from proliferating to infinite densities , where mean - field models apply . to generate an adequate branching - random walk
, one has to use an appropriate liouville operator . in one dimensions ,
one can choose , for instance , @xmath47 now , @xmath2 refers to the location in one - dimensional real space .
@xmath48 refers to the position of the cross - over to the bulk of the wave .
the operator @xmath18 generates a jump process . in the simplest case , again , the jump process may be approximated by a diffusive process , @xmath49 .
the growth term does not need to be modeled by a strict step - function , any sigmoidal function works in the limit of large population sizes @xcite .
importantly , the net reproduction rates monotonously increase in @xmath2 and saturates at some finite value @xmath21 at @xmath50 . finally , to limit the population size in the front of the wave , we need to use a non - constant weighting function @xmath51 in the global constraint , for instance , @xmath52 , which ensures that the growth region contains precisely @xmath53 individuals . the qualitative difference between the traveling waves in real and fitness space turns out to rely on the growth rates in the nose of the wave : while growth rates are saturating in the case of real - space waves
, it is increasing without bound in the case of waves of adaptation .
this makes models of adaptation even more sensitive to the effects of noise to the extent that a mean - field limit ( neglecting noise ) does not even yield finite velocity waves .
noise serves a crucial function in these models as it is required to regularize the wave dynamics .
these waves have therefore been called `` front - regularized waves '' @xcite .
in the same way as exemplified above for models of adaptation and invasion , one can frame many other eco - evolutionary scenarios , in their essence , as constrained branching random walks .
these models are , ultimately , defined by an operator @xmath9 generating the branching - random walk and a weighting function @xmath46 defining the global constraint . in this paper , we show that , in fact , for any given @xmath9 there are _ natural _ ways of choosing the weighting function .
the associated models , which we call _ tuned _ models , have desirable properties , including closed and linear moment equations , greatly facilitating their analysis .
we first state our main results on how to construct tuned models at any desired level of moments .
we will then provide an interpretation of these tuned models and provide simulation results , which we compare to fixed population size models .
detailed analytical derivations are given in later sections .
to characterize fluctuations in the makeup of the population it is convenient to consider the so - called @xmath54-point correlation function @xmath55 , which is the noise - average of the product @xmath56 of @xmath54 number density fields , @xmath57 , evaluated at the same time @xmath0 at various locations @xmath58 .
note that , here and henceforth , we use the notations @xmath59 and @xmath60 for a space and time - dependent function interchangeably . from many studies over the last 15 years , we know a lot about the first moment , @xmath61 , of several models of noisy traveling waves .
this provides access to the _ expected _ shape and velocity of the traveling wave , as well as the _ expected _ fate of individual mutations , or sub - populations .
however , wave shape and velocity fluctuations , as well , as the decay of genetic diversity requires access to higher moments , @xmath62 , for which there is no systematic approach so far .
our main result is that the dynamics of the @xmath63 moment ( and of all lower moments ) becomes analytically accessible if one chooses in the global constraint eq .
the weighting function @xmath46 to have the form @xmath64 for any positive integer @xmath54 , where @xmath65 satisfies @xmath66{w}\ ] ] with the adjoint operator @xmath67 of @xmath43 .
the arbitrary function @xmath68 controls the mean total population size as a function of time . for the special weighting function eq .
, the equation of motion for the @xmath63 moment becomes _ closed _ and _ linear _ , @xmath69 the resulting models may be called _ tuned _ , because for all other choices of the weighting function , the equation of motion for @xmath55 involves @xmath70 resulting in an infinite hierarchy of moments .
the moment closure of the tuned models is exact and not due to a truncation or approximation of higher moments .
the two terms @xmath71 are fluctuation - induced terms .
the last , positive term generates correlations at equal space arguments , which are then dissipated by the first , strictly negative , term over longer time scales .
the positive term exists only for @xmath62 and indicates that the effect of genetic drift on higher moments is more complicated than a cut - off in the growth term .
moreover , examining the behavior of differently labeled , but otherwise identical , subpopulations shows that tuned models can be interpreted naturally in the framework of population genetics .
first , the weighting function of all tuned models is a fixation probability function : the probability that descendants of individuals at location @xmath2 and time @xmath0 will take over the population on long times is given precisely by @xmath72 in the model tuned for the @xmath63 moment .
it is remarkable , in this context , that equation eq . for @xmath65 is precisely the equation governing the survival probability of an _ unconstrained _ branching random walk @xcite , with @xmath34 .
secondly , the higher moments provide access to the statistics of the genetic makeup of the population . according to the principle tenet of population genetics that , without mutations
, the genetic diversity of a population decreases with time : fixation and extinction of subtypes needs to be maintained by an appropriate influx of mutations .
the decay of genetic diversity fundamentally depends on higher moments the first moment captures fixation probabilities but not the time to fixation .
the decay of genetic diversity in the absence of mutations can be quantified by the cross - correlation function @xmath73 between subtypes @xmath74 . here , the number densities @xmath75 is the number density field of type @xmath76 .
the sum of all sub - types makes up the total population , @xmath77 .
if all @xmath78 are different and @xmath62 , it is clear that @xmath79 must continuously decay in the absence of mutations because one of the subtypes will take over in the presence of random genetic drift .
the case @xmath80 is the most prominent one : @xmath81 is proportional to the so - called heterozygosity at location @xmath2 and @xmath0 , which is the probability that two individuals sampled with replacement are of different type @xcite .
now it turns out that , in the @xmath63 -tuned model , the cross - correlation function can be expressed in the simple form @xmath82 with @xmath83 satisfying a one - dimensional linear equation @xmath84 subject to the initial conditions @xmath85 .
notice that the decay of genetic diversity in the @xmath63 -tuned model is described by the spectrum of the operator appearing on the right - hand side of eq . , which also occurs in eq .
describing the fluctuations of the total population .
consistent with our population genetic interpretation , one can show quite generally that this operator only has decaying relaxation modes ( see fig .
[ fig : spectrum - si ] ) .
we consider eqs . and to be our results with the most immediate applicability because they provide a feasible approach to resolving a key question in population genetics ( maintenance and decay of genetic diversity ) that _ relies _ on having access to higher moments .
the above formalism includes the case @xmath86 ( closed first moment ) presented in ref .
@xcite ( in which the tuned weighting function @xmath87 was denoted by @xmath88 ) . in constrained random walk models ,
one can generally retrieve lower moments from higher moments by contraction with the weighting function @xmath89 , @xmath90 where we invoked the global constraint eq . in going from the second to the third line .
to simplify the notation , we have here introduced the short hand @xmath91 to denote @xmath92 , i.e. , that the @xmath93th space variable should be omitted in the arguments .
the contraction rule eq
. also shows that contraction in the last term in the dynamical equation for @xmath94 simply generates the next lower moment @xmath95 .
moreover , if one obtains the @xmath63 moment by solving eq . , one generally has access to all lower - rank moments as well ( but not to the larger moments ) .
thus , the model tuned to be linear at the @xmath63 moment gives access to all moments up to and including the @xmath63 .
it is remarkable that the governing equations for @xmath96 and @xmath55 are independent of the offspring number variation @xmath33 .
the only effect of @xmath33 is to reduce the fixation probability @xmath97 and scale up the population sizes @xmath98 .
this means in particular that noise - induced terms @xmath99 in the moment equation are not scaled by @xmath33 , and thus do not become small in the small noise limit .
the lack of a potentially small parameter has important consequences for the use of perturbation theory in this context .
we now illustrate our main results by explicit numerical solutions and stochastic simulations for models tuned to be closed at the first and second moment . the two biological phenomena we consider , adaptation and invasion , both give rise to compact traveling waves in real space and fitness space , respectively . they fundamentally differ in their location - dependent growth rates : while in adaptation models , growth rates are linear increasing towards the tip of the waves , they saturate in invasion models . as a consequence ,
invasion waves have a well - defined infinite population size limit in contrast to adaptation waves .
the detailed behavior of simple models of asexual adaptation models varies with the assumptions made about how the mutation process influences the growth rates that define the branching process . in our numerical work ,
we focus on the diffusion kernel in eq . , which assumes that a growth rate variance @xmath11 is acquired per generation due to mutations .
the advantage of the diffusion kernel is that it only contains one parameter , the diffusion constant @xmath11 , and matters when mutation rates are large compared to the ( effective ) rate of selection @xcite .
[ [ numerical - approach . ] ] numerical approach .
+ + + + + + + + + + + + + + + + + + + to solve our framework of tuned models , we used a multi - dimensional newton - raphson method to determine traveling wave solutions corresponding to tuned models of degree @xmath54 .
to this end , we first determined a traveling steady state solution @xmath100 of eq . defining the set of tuned weighting function @xmath101 ( eq . ) .
this weighting function enforces a traveling steady - state for the @xmath63 moment @xmath102 of the number densities , which satisfies the linear time - independent equation eq . .
as a result , we obtain for a given @xmath54 , a weighting function @xmath103 and the @xmath63 moment .
details of the numerical scheme and the explicit equations solved for @xmath86 and @xmath80 are described in appendices [ appendix : explicit ] and [ appendix : numerics ] .
* numerical solution of the model of adaptation tuned to be closed at the first moment ( @xmath86 ) , and comparison with stochastic simulations .
* averages from stochastic simulations ( brown ) coincide with numerical solutions of the mean stationary population density @xmath104 ( blue line ) for small speeds ( * a , b * ) and large speeds ( * c , d * ) , respectively
. the weighting function @xmath105 ( green line ) exhibits a sharp increase towards the nose of the wave , before crossing over to its asymptotic linear solution , @xmath106 , for large fitness .
this is consistent with the interpretation of @xmath107 as fixation probability , tested in fig .
[ fig : measurefixation ] . models tuned to be closed at higher moments ( @xmath62 ) display the same features qualitatively and quantitatively , see fig .
[ fig : highermoments ] . ] [ [ model - tuned - to - the - first - moment - n1 . ] ] model tuned to the first moment ( @xmath86 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + to set the stage and to reproduce earlier results from ref .
@xcite , we first present numerical results for the model closed at the first moment .
[ fig : numericalsolution ] shows , for two wave speeds , the weighting function and the mean population density in a stationary comoving frame . while the population distribution is , except for an exponential decay in the wave - tip , close to a gaussian for large speeds , it is markedly skewed for lower wave speeds . note that the exact numerical results are in near perfect agreement with stochastic simulations , confirming our approach .
[ [ model - tuned - to - the - second - moment - n2 . ] ] model tuned to the second moment ( @xmath80 ) .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + fig .
[ fig : numericalsolution ] characterizes the behavior of @xmath80 models in comparison with @xmath86 tuned models .
since @xmath80 models provide access to the second moment , one can of course also obtain the first , by contraction with the weighting function @xmath108 ( cf .
[ fig : numericalsolution]b shows the mean population densities for both @xmath86 and @xmath80 models for the same velocity and weighting function @xmath96 which is identical to their respective weighting functions @xmath87 and @xmath109 up to an @xmath54dependent factor , cf .
the mean population densities can hardly be distinguished in the figures - generally , the agreement of the two different models increases with increasing wave velocity or , equivalently , population size .
moreover , the stochastic simulations track the predicted mean in near perfect agreement , providing numerical support for our analysis of the tuned closure at higher moments .
[ fig : numericalsolution]b shows , in 3d plot , the second moment @xmath110 in a semi - logarithmic plot .
while this plot is mostly gaussian , it reveals a distinctly exponential decay in the front and back of the wave .
this indicates the importance of fluctuations in these low - density regions .
higher order correlation functions can also be used to directly investigate fluctuation properties of the noisy adaptation wave .
the pearson - product - moment - correlation , defined as @xmath111/\sqrt{\mathrm{var}[c_t(x)]\mathrm{var}[c_t(y)]}$ ] shows a clear anticorrelation signal in the nose of the wave : usually , a ( stochastic ) rise in population number in the nose of the wave leads to an overall decrease in population size .
such a very fit population has a much larger weight in the tuned constraint , eq . , forcing the bulk population to be culled .
thus , bulk population size and nose population size are anticorrelated @xcite .
[ [ fixation - probabilities . ] ] fixation probabilities .
+ + + + + + + + + + + + + + + + + + + + + + + note from figs .
[ fig : numericalsolution ] and [ fig : highermoments ] that the weighting functions @xmath101 strongly increase towards the tip of the wave where it crosses over to a linear increase .
the functional form is consistent with the interpretation of a fixation probability : the success probability of an individual should be much larger in the tip of the wave rather than the bulk because it has to compete with less and less equally or more fit individuals .
the fixation probability approaches a linear branch beyond some cross - over fitness because individuals there are so exceptionally fit that they merely have to survive random death to fix .
the competition with conspecific is minimal - there is only competition with their own offspring . to test our prediction that @xmath112 can be interpreted , exactly , as an fixation probability , we have carried out the following test : we ran simulations for @xmath86 and @xmath80 tuned models , in which we labeled the tip of the wave such that the predicted fixation probability is exactly @xmath113 , cf .
[ fig : measurefixation ] .
the measured fixation probability is shown in the insets of fig .
[ fig : measurefixation ] . within the statistical error ,
the agreement is very good .
we would like to point out that the fact that @xmath114 is strongly increasing towards the wave tip means that highly fit individuals have a large impact on what fraction of the population is culled per time step .
for instance , if a single individual arises far out in the tail of the wave , it may have a large fixation probability . to keep the total fixation probability at one ,
this means that a significant fraction of the individuals need to be cleared from the population .
importantly , however , the culling itself is independent of the identity of individuals , i.e. , a poorly fit individual is equally likely to die as a highly fit individual .
this distinguishes our model from models that regularize the population size by removing cells preferentially at the wave tip .
such a procedure strongly modifies the wave dynamics and , in particular , reduces the amount of fluctuations .
* fixation in traveling wave models of adaptation sensitively depend on individuals in the nose of the wave . *
one of our key results is that tuned models generally have the convenient feature that fixation probabilities can be computed exactly they are given by the tuned weighting function @xmath115 .
here , we provide a test of this prediction .
first , we have generated @xmath116 independent start configurations , and labeled the wave tip such that the labeled population is predicted to have a @xmath113 fixation probability , @xmath117 .
panel * a * shows one start configuration and the labeled subpopulation in orange .
( the shown density profile ( light blue area ) is momentarily smaller than the mean density ( blue line ) due to random population size fluctuations , cf .
[ fig : popsizespeed ] ) . for each of the labeled start configurations ,
we ran @xmath118 simulations up until the ( fixation ) time where either the labelled subpopulation is extinct or has taken over . from counting the number of extinction and fixation events
, bayesian inference allows to compute the posterior distribution assuming a flat prior @xcite . * b * for the model with @xmath86 , we find that @xmath119 of the @xmath120-confidence intervals of this beta - distributed posterior include the expected fixation probability of @xmath121 ( black intervals ) , while @xmath122 intervals scatter considerably due to lattice effects in simulations ( red intervals ) .
* c * when closing the moment hierarchy at @xmath80 results exhibit stronger scattering .
however , the measured confidence intervals for @xmath80 reproduce the expected outcome to reasonable extent . ]
[ [ velocity - population - size - relationships ] ] velocity - population - size relationships + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + fig .
[ fig : popsizespeed ] shows the relation between velocity and mean population size in various tuned models . as comparison
, we also show the corresponding relationships between mean velocity and population size for fixed population size models .
all these models correspond to different statistical ensembles , yet for large population sizes the curves approach each other very well . the differences between models at finite population sizes
is explained mainly by population size fluctuations : if we force the fixed population size model to fluctuate precisely as a given realization of a tuned model , we obtain very accurate agreement .
conversely , if we plot velocity vs. @xmath123 shows excellent agreement between all models .
this agreement is due to a timescale separation , discussed below .
* fluctuations in adaption waves suggest the adaptation speed is controlled by the _
mean logarithmic population size_. * * a * stochastic simulations reveal that , for a given speed of adaptation , tuned models have a larger mean population size @xmath124 ( blue line ) than models of fixed population size ( green line ) .
the small discrepancy is much lower between these two models than to the best available cutoff theory ( gray line @xcite ) .
most of the remaining discrepancy is due to population size fluctuations ( very nearly exponentially distributed , see inset * b * ) : when we measure fluctuations in a stochastic realization of the tuned model and impose exactly the same population size time series in repeated stochastic simulations , we obtain adaptation speed vs. mean population size relation that coincides with the one obtained for the tuned models ( red circles ) .
importantly , fixed population size and tuned models have almost the same mean logarithmic population size , @xmath123 , for the same wave speeds .
it can be shown that this is due to a time scale separation between slowly decaying population size fluctuations ( coalescence time scale ) and the fast relaxation of the wave speed ( determined by a mixing time in the tip of the wave ) .
thus , the key dynamical quantity in traveling wave models with fluctuating population sizes is @xmath125 instead of @xmath126.,scaledwidth=50.0% ] [ [ decay - of - genetic - diversity ] ] decay of genetic diversity + + + + + + + + + + + + + + + + + + + + + + + + + + next , we have solved numerically for the mode spectrum that governs the decay of heterozygosity . fig .
[ fig : decaytimes ] shows the behavior of the lowest two eigenvalues as a function of the velocity of the wave and the population size , respectively .
it can be clearly seen that a time scale separation arises : the frequency of the first mode decays slowly to @xmath127 , following @xmath128 to a good approximation .
the frequency of the next higher mode , on the other hand , approaches a constant value of order @xmath129 .
this means that coalescence takes much longer than the time until a subpopulation has forgotten its the initial condition of its spatial distribution .
this time - scale separation not only helps in analytically finding the coalescence time in sec .
[ sec : time - scale - separ ] .
but it also underlies the ensuing bolthausen - sznitman coalescence in many models of adaptation and invasion waves of the fisher - kolmogorov type @xcite .
* time scales for the decay of genetic diversity in models of adaptation . * in models with limited population size , the ultimate fate of a subpopulation of neutral mutants is to either go extinct or to fixation .
hence , the diversity of the population will gradually decay unless new mutations come in , and quantifying this decay is an important population genetic challenge .
this figure quantifies the leading decay times of the function @xmath130 correlating the densities of labeled and unlabeled individuals .
the decay times were found from a spectral analysis of the equation of motion of @xmath131 , which is linear in the model tuned to be closed at the level of two - point correlation functions ( @xmath80 ) . *
a * our numerical results suggest that , for the considered model with diffusion kernel , the slowest decay time approaches @xmath132 for large speeds ( gray line indicates @xmath133 ) .
by contrast , the second slowest decay time @xmath134 approaches a constant for large populations .
this indicates an important time scale separation , as we argue in the main text .
also note that the numerical results approach , for large @xmath53 , the approximation @xmath135 in eq .
, which is based on the time scale separation of the two lowest eigenvalues emerging for large population sizes .
panel * b * shows the same quantities as panel * a * as a function of the population size @xmath136 .
, scaledwidth=50.0% ] [ [ sec : fkpp waves ] ] invasion waves + + + + + + + + + + + + + + after changing the liouville operator the one in eq . , and following the same numerical pipeline as described above , we obtain analogous results for invasion waves , see fig . .
note that the spatial co - ordinate now corresponds to real space rather than a fitness landscape .
our data reproduce the universal velocity - population size relationship and the coalescence time scaling that have been established for fkpp waves over the last 20 years @xcite .
an explicit form of the equations we solved numerically can be found in appendix [ appendix : explicit ] . * invasion waves . * here , we summarize our numerical results for models of invasion tuned to be closed at the second moment , @xmath80 . * a * wave speed as a function of population size .
the inset * b * compares our results with the leading order cut - off correction @xmath137 .
* c * shows the two longest decay times , @xmath138 , of the second moment .
the increasing gap between both decay times manifests a time scale separation .
the double logarithmic plot in inset * d * shows that the longest decay time follows @xmath139 ( upper gray lines ) asymptotically and , hence , behaves as the predicted coalescence time in fkpp waves @xcite .
by contrast , the second decay time follows @xmath140 ( lower gray line ) . ]
in the following we proceed with the derivation of our main results quoted above . to this end , we will first recapitulate the derivation of the stochastic differential equation governing the dynamics of branching random walks , as was done in ref . @xcite .
we will then discuss the consequences of this stochastic dynamics for @xmath54-point correlation functions , which will allow us to identify the natural choice of the weighting function @xmath141 .
subsequently , we will discuss how to modify our basic model to account for different subtypes within the population .
we will now determine the stochastic dynamics obtained in the limit @xmath142 , which is in general non - linear and hence not solvable .
we then use the ensuing stochastic differential equation to identify special models with closed moment hierarchies .
we will find that these tuned models are not only solvable , but also allow for a natural interpretation of the constraint in terms of fixation probabilities .
the stochastic dynamics of a constrained branching random walk ( cbrw ) was derived in ref .
@xcite and may be summarized as follows .
the state of the system is described by the number density @xmath1 of random walkers at position @xmath2 and time @xmath0 . at any time
, the distribution of random walkers has to satisfy a global constraint defined by eq . .
the combination of eqs . and can be written as a fraction , @xmath143 in the continuous - time limit ( small enough @xmath3 is required to ensure that the denominator of the fraction is never far from @xmath144 ) .
note that the expression in eq . evidently satisfies the constraint , @xmath145 moreover , in the continuous - time limit
, we only need to retain terms up to order @xmath146 @xcite .
thus , expanding eq . , we obtain @xmath147 \\ & & ~~+\epsilon\left[{{\mathrsfs}{l}}c_t - \sqrt{{b}c_t}\eta_t\bigl\langle u_t\vert \sqrt{{b}c_t}\eta_t\bigr\rangle - c_t\bigl\langle\partial_t u_t\vert c_t \bigr\rangle - c_t \bigl\langle u_t\vert \label{eq : c - fractionexpansion}{{\mathrsfs}{l}}c_t \bigr\rangle + c_t \bigl\langle u_t\vert \sqrt{{b}c_t}\eta_t \bigr\rangle^2 \right]\;.\end{aligned}\ ] ] in eq . , we required @xmath46 to change only deterministically , @xmath148 , i.e. it has no stochastic component . finally , we replace products of order @xmath149 in the deterministic term of expansion eq . with their averages , eq .
, to arrive at the following stochastic differential equation : the temporal change @xmath150 of the concentration field @xmath5 from time @xmath0 to @xmath151 can be written as @xmath152 which consists of a deterministic change @xmath153 of order @xmath146 and a stochastic change @xmath154 of order @xmath155 .
these are given by @xmath156 thus , we have arrived at the continuous - time stochastic process for a constrained branching random walk @xcite , which summarizes the combined effect of the original two - step algorithm , ( i ) `` branching random walk '' and `` enforce constraint '' .
the related concept of forcing the solution of a sde onto a manifold has been analyzed in @xcite .
this section introduces the hierarchy of moment equations of cbrws that characterize the mean and the fluctuations of the concentration field of the random walkers .
consider the dynamics of the products of @xmath157 .
our goal is to determine how the function @xmath157 changes as time marches forward . to this end
, we express the time increments of the @xmath54point products @xmath157 in terms of the changes of the single fields @xmath5 , @xmath158\;,\ ] ] using the deterministic and stochastic time increments @xmath159 computed in .
next , we expand the product up to order @xmath146 , @xmath160\;.\nonumber\end{aligned}\ ] ] note that the last term arises only for @xmath161 .
inserting the time increments eq .
of the single fields into eq . yields @xmath162\\ & & + \epsilon \left [ \sum^n_{j=1 } \left.\left({{\mathrsfs}{l}}- { b}u_t\right ) \right|_{x_j } c^{(n)}_t - n \bigl\langle \left(\partial_t+{{\mathrsfs}{l}}^\dagger-{b}u_t\right)u_t \mid c_t \bigr\rangle c_t^{(n ) } \right ] \nonumber \\ & & + \epsilon \sum_{j=1}^{n-1}\sum_{k = j+1}^{n } { b}\sqrt{c_t(x_j)c_t(x_k)}\eta_t(x_j)\eta_t(x_k)c^{(n-2)}_t(\backslash x_j,\backslash x_k ) \nonumber\\ & & -\epsilon ( n-1 ) \sum^n_{j=1 } \sqrt{{b}c_t(x_j)}\eta_t(x_j)\bigl\langle u_t \mid \eta_t \sqrt{{b}c_t}\bigr\rangle c^{(n-1)}_t(\backslash x_j ) \nonumber \\ & & + \epsilon \frac{n ( n-1)}2 \bigl\langle u_t \mid \eta_t \sqrt{{b}c_t } \bigr\rangle^2 c^{(n)}_t \;.\nonumber\end{aligned}\ ] ] within the deterministic @xmath146 terms , we again replace products of noises in terms of their averages , as given by eq . plus a stochastic component .
if such terms quadratic in the noise arise in the deterministic @xmath146 part of any stochastic equation , one may simply ignore their stochastic component , as it would lead to ( in the limit @xmath163 ) negligible contributions .
we then obtain @xmath164\\ \nonumber & & + \epsilon\left[\sum^n_{j=1 } \left .
\left({{\mathrsfs}{l}}-{b}n u_t \right)\right|_{x_j } c^{(n)}_t + { b}\sum_{j=1}^{n}\sum_{k = j+1}^{n}\delta(x_j - x_k)c^{(n-1)}_t(\backslash x_k ) \right ] \\
\label{eq : stoch - eqn - for - nth - moment } & & -\epsilon n\left\langle \left(\partial_t+{{\mathrsfs}{l}}^\dagger - b(n+1)/2\,u_t\right)u_t\mid c_t\right\rangle c^{(n)}_t\;.\end{aligned}\ ] ] upon averaging and sending @xmath3 to zero , we obtain an equation of moment for the @xmath54th moment , @xmath165 from we observe that the coupling to higher moments @xmath166 is mediated via the dynamics of the weighting function @xmath46 in the last term . the key result of ref .
@xcite was that a particular choice of the weighting function @xmath46 exists , for which the first moment equation is closed .
we now show that exact closures can be found for higher moments as well .
suppose , the solution @xmath167 of @xmath168u_t^{(n)}\ ] ] exists , and we choose @xmath169 as the weighting function .
for this particular model , the dependence on the @xmath170th moment in eq . disappears identically : @xmath171 thus , the hierarchy of the first @xmath54 moments is closed .
in fact , this closed set of @xmath54 differential equations can be summarized by a single integro - differential equation , eq . , because contracting @xmath172 with @xmath141 reduces the order of the moments by virtue of the constraint , cf .
eq . .
the final form of our results eq.s , are obtained upon substituting @xmath173 which is the initially quoted eq . .
in summary , starting from a linear operator @xmath43 , we have identified an algorithm to construct a constrained branching walk model solvable up to the @xmath54th moment : first , identify the weighting function @xmath114 for which the hierarchy of moment equation closes at the @xmath54th moment . to this end , solve equation , which is deterministic nonlinear equation for the weighting function @xmath114 depending on a space and a time variable .
second , solve the corresponding moment equation , which is a linear equation for the function @xmath55 that depends on @xmath54 space variables and a time variable .
once the function @xmath55 has been obtained , any lower - order moment follows by contraction with @xmath174 , as described in eq . .
we now extend our model to account for @xmath175 different types of individuals .
this enables studying questions such as how does mutator strain take over in an evolving population of bacteria even if it confers a direct fitness detriment , or , how does a faster dispersing mutant spreads during a growing tumor even if it might be slower growing ?
moreover , we can discuss the decay of genetic diversity and arrive at the very important conclusion that @xmath114 is always the fixation probability of a neutral mutation arising at position @xmath2 and time @xmath0 by considering exchangeable subtypes . to this end
, we define the dynamics of the subtypes analogously to our original constrained branching walk model for the total population in sec .
[ sec : derivations ] : the number density of individuals of type @xmath76 at position @xmath2 at time @xmath0 shall be given by a density field @xmath176 .
hence , the entire state of the system is described by the vector @xmath177 . in each time step , a given subtype undergoes a step of branching random walk subject to their own linear dynamics encoded by an operator @xmath178 and their own fluctuations generated by a noise field @xmath179 , @xmath180 the @xmath76dependence of the linear operators @xmath178 encode the _ phenotypic _ differences between types .
e.g. if type @xmath76 would refer to a mutator type , @xmath178 would include a particular mutational operator @xmath181 characterizing the mutator phenotype .
for instance , if mutations are modeled by diffusion , the corresponding diffusion constant of a mutator would be larger than that of the wild type .
the different subtypes are coupled only by the second computational substep @xmath182 which ensures a global constrained defined by a weighting function vector @xmath183 , @xmath184 notice that we have merely added another ( discrete ) dimension to the problem - the type degree of freedoms .
it may be checked that our arguments to arrive at an effective stochastic differential equation and for closing the moment hierarchy in sec .
generalize to any number of dimensions .
thus , we can immediately restate our central results for the extended model accounting for sub - types . in particular , if we choose the weighting function vector to be @xmath185 with @xmath186{w}_i(x , t)\;,\ ] ] then equation of motion for the @xmath63 moment will be closed . if we choose the notation @xmath187 with @xmath78 being the type of the @xmath188 number density field in the product on the right - hand - side .
the equation of motion for the @xmath63 moment is given by @xmath189 notice that the correlations indicated by the last term only arise for a subpopulation with itself , @xmath190 .
now , let us focus on the special case where types follow the same dynamics in the statistical sense , @xmath191 for such `` exchangeable '' subtypes , it is easy to see that the equation of motion eq . for the @xmath63 moment of the total population , @xmath192 ,
is obtained upon summing left and right - hand side of eq . over all type indices , i.e. by carrying out @xmath193 .
we can single out one particular subpopulation , say @xmath194 ( @xmath195 for labeled ) , by summing the equation of motion over all other type indices , @xmath196 .
this yields @xmath197 note that the correlation function @xmath198 satisfies the same linear equation as @xmath55 does eq . , which we abbreviate as @xmath199 defining a linear ( integro - differential ) operator @xmath200 .
imagine solving for the left eigenvector of @xmath200 corresponding to eigenvalue 0 , @xmath201 where @xmath202 is a function of @xmath54 variables just like @xmath94 .
then , contracting eq
. with this new function one obtains @xmath203 the second equality is a key step .
it holds because @xmath204 on long times if fixation occurs and @xmath205 , otherwise .
hence , the fixation probability @xmath206 of the labeled subpopulation with initial density @xmath207 at time @xmath127 is given by @xmath208 fortunately , the left eigenvector of @xmath200 is easily constructed : if we fully contract eq . with @xmath54 factors of the weighting function @xmath209 , we have to get @xmath127 on the lhs because of the constraint eq . .
hence , the sought - after left eigenvector can be written as @xmath210 since this eigenvectors contracts to 1 with the total population , eq .
becomes @xmath211 thus , as announced in section [ sec : main - results ] , a single labeled mutant at a certain location @xmath2 has probability @xmath112 that its descendants will take over the population on long times .
the diversity of labels in a population will inevitably decline , because of the rise and ultimate fixation of one of the labels initially present .
one can capture the gradual loss of genetic diversity by studying the expectation of the product of density fields that correspond to different types .
for instance , if there are just two types , @xmath212 , the correlation function @xmath213 satisfies @xmath214 notice that the right - hand side of eq . misses the positive @xmath215function term of eq . , which characterized the @xmath63 moment of the total population density .
assuming that @xmath216 ( as well as @xmath217 and @xmath218 ) have a stable stationary solution , we can conclude that @xmath130 will decay to @xmath127 at long times because of the lacking source term .
this is to be expected because the gradual fixation of one of the two types implies that @xmath130 has to approach @xmath127 on long times .
discerning relaxation times of the time - evolution is related to a key question in population genetics : how fast do lineages of two individuals coalesce ? the function @xmath219 is closely related to the so - called heterozygosity in population genetics .
the heterozygosity in the population is the probability that two randomly chosen individuals are of different type , @xmath220 . for our purposes , it is much more convenient and natural to consider a variant of that quantity , @xmath221 with @xmath222 and @xmath223 .
these quantities have nice properties . at any time , both @xmath224 and @xmath225 represent the probability of fixation of the respective subpopulations .
accordingly , we have @xmath226 , ensured by the global constraint .
thus , @xmath227 is similar to a heterozygosity in neutral populations and , under certain conditions discussed below , even converges against the actual heterozygosity .
an equation of motion for the expectation @xmath228 can be derived by contracting eq . twice with @xmath141 and using the equation of motion of @xmath141 , eq .
, @xmath229 notice that the right - hand side is negative always .
thus , again , we see that the heterozygosity will necessarily decay . as in eq . for @xmath80 , one can easily see that the equation of motion for @xmath230 is separable for the @xmath54 components if @xmath231 .
it therefore admits a solution of the simple form @xmath232 with @xmath83 satisfying a one - dimensional linear equation @xmath233 subject to the initial conditions @xmath85 . by contracting with @xmath96 and using its dynamics , @xmath234 , @xmath235 we see that @xmath60 must be continuously decaying with time . on long times
, we can assume that @xmath236 , where @xmath237 is the eigenfunction to the largest eigenvalue of eq . .
inserting this asymptotic behavior into eq .
yields an exponential decay @xmath238 of the mode amplitude @xmath239 with a decay time given by @xmath240 intuitively , this decay time describes how long it takes until significant fraction of the population has coalesced .
thus , one expects @xmath241 to depend on the fundamental parameters of the considered model in just the same way as the population coalescence time
. the numerical coefficient , of course , will be different by a factor of order 1 . due to the one - dimensional nature of eq .
, it is possible to obtain good approximations to the longest relaxation time for various models , either by directly solving the eigenvalue problem or by guessing the function @xmath242 in eq . .
we will provide a heuristic calculation for the case of a diffusive kernel in the limit of large populations ( or fast waves ) . in most models of noisy traveling waves ,
one has found empirically a time scale separation : for large @xmath243 , the longest relaxation time of the mean density field is much shorter than the coalescence time . in the case of adaptation waves and invasion waves with diffusion kernel ,
this is evident from gap in the two lowest relaxation times in our numerical results in fig .
[ fig : decaytimes ] , [ fig : fisherwaves]b ( also see si fig .
[ fig : spectrum - si ] ) . in the presence of such a time - scale separation , we can approximate @xmath244 and @xmath245 . for the purpose of using these approximations in the terms involving @xmath246 in eq .
[ eq : heterozygosity - var - eom ] , we need them to be good in the high fitness tail .
then , we obtain @xmath247 suggesting that the time scale for coalescence scales as @xmath248 this approximation indeed seems to approach the correct time scale for large @xmath243 , as can be appreciated from figs . [
fig : decaytimes ] and [ fig : fisherwaves]b .
the ecological and evolutionary fate of populations often depends on a small number of `` pioneers '' , distinguished by their growth rates , migration rates , location , or other characteristics correlated with long - term survival .
most analyses of these inherently stochastic problems have focussed on the mean behavior of the population , which sensitively depends on fluctuations in the pioneer populations . yet
, the mean behavior says little about any given realization , the variability between realizations and their correlation times . here
, we have shown that fluctuations can be analyzed in principle , if one relies on minimal models that reduce the dynamics to two essential ingredients : ( 1 ) birth , death and jumps give rise , effectively , to a branching random walk .
( 2 ) a non - linear population regulation makes sure that those branching processes do not get out of control . for
such constrained branching random walks , we have provided a general route towards analyzing fluctuations .
the basic idea of the method is to adjust the population control , an essential non - linearity , in such a way that the equations describing correlation functions of order @xmath54 are closed .
our method can be used to elucidate variability between replicates in evolution experiments as well as the genetic diversity within a population . to provide specific results
, we have focussed on simple models of adaptation and of invasions . in both cases
, we have found that the decay of genetic diversity scales as a power of the logarithm of the population size for large population sizes .
higher moments show a marked anticorrelation between the dynamics in the tip and the bulk of the wave .
moreover , we found that , for the models analyzed , the time scale for the decay of higher order correlations , such as the genetic heterozygosity , is much longer than the time the population wave needs to equilibrate at a given speed or population size .
the presence of such a time scale separation simplifies the analysis of coalescence times considerably .
the ensemble of the resulting tuned models is complementary to established models of adaptation .
while the latter have fixed population and fluctuating speeds of adaptation , the former has a fluctuating population size but fixed wave speeds .
the resemblance of both ensembles relies on the fact that the fluctuations occur on time scales long compared to the relaxation time of the population wave .
quantities that only depend on the mean logarithm of the population size , such as the wave speed or the coalescence time , thus agree asymptotically in both ensembles , see e.g. fig . [
fig : popsizespeed ] , [ fig : fisherwaves ] .
one might wonder about the net - effect of noise on models of adaptation and other traveling waves .
if one is only concerned with the mean , many previous works have assumed that the effect of noise can be summarized by an effective cutoff in the tip of the wave @xcite .
this cutoff effect can be explicitly seen in the closed first moment equation of tuned models , as was already pointed out in ref .
however , what is the effect of noise on higher - order correlations ?
our general formulation in eq . of the @xmath63 moment exhibits , in general , two terms with different signs that are unexpected in a deterministic framework .
one term tends to generate positive correlations between nearby individuals ( in fitness space ) .
these correlations then dissipate over distances due to the term with negative sign .
importantly , the correlation term becomes dominating in the tip of the wave due to its density dependence .
the net - effect of fluctuations on the correlations emphasizes the complex nature of fluctuations , which only to the lowest order can be captured by a simple cutoff term in an effective liouville operator .
the branching random walk contains a parameter @xmath249 , the variance in offspring numbers per generation , that effectively measures the strength of genetic drift .
surprisingly , the noise - induced terms do not depend on this parameter @xmath249 .
this means that the noise - induced terms are not small , in general , even if the parameter @xmath249 is small , such that a controlled small - noise perturbation analysis is not possible .
this reinforces the observation that noise is a singular perturbation that fundamentally impacts the outcome of ecological and evolutionary processes .
our method of model tuning is quite versatile as it applies to any branching random walk subject to a global constraint .
this includes models that combine ecology and evolution @xcite or epistatic models in which mutations are not simply additive but might interact @xcite .
however , for more complex scenarios of interest to evolutionary biologists , one would like to introduce additional non - linearities .
for instance , sex and recombination is a quadratic non - linearity as it depends on the the probability density of two different individuals finding each other and mating @xcite . in evolutionary game theory ,
one is interested in mutants that have a frequency - dependent advantage @xcite .
the fitness of producers of a common good depends on the frequency of producers .
this , again , introduces a non - linearity , which is quadratic in the simplest case .
such non - linearities can not be included in an exact way because they generate higher - order terms .
however , it may be a useful approach to build them in and truncate the moment hierarchy at an appropriate order provided one can show that the neglected terms really are small .
we believe that such reasoning should work , typically , if the non - linearities do not strongly influence the dynamics in the small density regions where the noise strength is large
. a truncation scheme would , in this case , amount to matching a stochastic but linear description of the wave tip with a deterministic but non - linear bulk of the wave .
we would welcome future work to examine these possibilities .
thanks to peter pfaffelhuber for useful discussions and making us aware of ref . @xcite .
this work was partially supported by a simons investigator award from the simons foundation ( o.h . ) , the deutsche forschungsgemeinschaft via grant ha 5163/2 - 1 ( o.h . ) . 10 l. j. allen . .
pearson education new jersey , 2003 .
n. barton , a. etheridge , and a. vber . modelling evolution in a spatial continuum .
, 2013(01):p01002 , 2013 .
e. brunet and b. derrida .
shift in the velocity of a front due to a cutoff .
, 56(3):2597 , 1997 .
e. brunet , b. derrida , a. mueller , and s. munier .
phenomenological theory giving the full statistics of the position of fluctuating pulled fronts .
, 73(5):056126 , 2006 .
e. cohen , d. a. kessler , and h. levine .
fluctuation - regularized front propagation dynamics in reaction - diffusion systems . , 94(15):158302 , 2005 .
e. cohen , d. a. kessler , and h. levine .
front propagation up a reaction rate gradient .
, 72(6):066126 , 2005 .
m. m. desai and d. s. fisher .
beneficial mutation selection balance and the effect of linkage on positive selection .
, 176(3):17591798 , 2007 .
m. m. desai , d. s. fisher , and a. w. murray . the speed of evolution and maintenance of variation in asexual populations .
, 17(5):385394 , 2007 .
m. m. desai , a. m. walczak , and d. s. fisher .
genetic diversity and the structure of genealogies in rapidly adapting populations .
, 193(2):565585 , 2013 . d. s. fisher .
asexual evolution waves : fluctuations and universality . , 2013(01):p01011 , 2013 .
r. frankham .
effective population size / adult population size ratios in wildlife : a review .
, 66(02):95107 , 1995 .
p. j. gerrish and r. e. lenski .
the fate of competing beneficial mutations in an asexual population .
, 102:127144 , 1998 .
l. geyrhofer . .
phd thesis , mpi for dynamics and self - organization , and university of gttingen , 2014 .
l. geyrhofer and o. hallatschek .
stochastic delocalization of finite populations .
, 2013(01):p01007 , 2013 .
l. geyrhofer and o. hallatschek .
oscillations in noisy traveling waves . , 2015 .
b. h. good , i. m. rouzine , d. j. balick , o. hallatschek , and m. m. desai .
distribution of fixed beneficial mutations and the rate of adaptation in asexual populations .
, 109(13):49504955 , 2012 . i. gordo , l. perfeito , and a. sousa .
fitness effects of mutations in bacteria .
, 21(1 - 2):2035 , 2011 .
s. goyal , d. j. balick , e. r. jerison , r. a. neher , b. i. shraiman , and m. m. desai .
dynamic mutation selection balance as an evolutionary attractor .
, 191(4):13091319 , 2012 .
p. haccou , p. jagers , and v. a. vatutin . .
number 5 .
cambridge university press , 2005 .
o. hallatschek . the noisy edge of traveling waves .
, 108(5):17831787 , 2011 .
o. hallatschek and d. r. nelson .
gene surfing in expanding populations .
, 73(1):158170 , 2008 .
d. hedgecock and a. i. pudovkin .
sweepstakes reproductive success in highly fecund marine fish and shellfish : a review and commentary . , 87(4):9711002 , 2011
. g. s. katzenberger .
solutions of a stochastic differential equation forced onto a manifold by a large drift .
, pages 15871628 , 1991
. e. kussell and m. vucelja . non - equilibrium physics and evolution
adaptation , extinction , and ecology : a key issues review .
, 77(10):102602 , 2014 . s. f. levy , j. r. blundell , s. venkataram , d. a. petrov , d. s. fisher , and g. sherlock .
quantitative evolutionary dynamics using high - resolution lineage tracking . , 2015 .
r. a. neher
. genetic draft , selective interference , and population genetics of rapid adaptation .
, 44:195215 , 2013 .
r. a. neher and o. hallatschek .
genealogies of rapidly adapting populations . , 110(2):437442 , 2013 .
r. a. neher , t. a. kessinger , and b. i. shraiman .
coalescence and genetic diversity in sexual populations under selection .
, 110(39):1583615841 , 2013 .
r. a. neher , b. i. shraiman , and d. s. fisher .
rate of adaptation in large sexual populations .
, 184(2):467481 , 2010 .
a. ohagan and j. forster . ,
volume 2b .
new york halsted press , 1994 .
f. pelletier , d. garant , and a. p. hendry .
eco - evolutionary dynamics . , 364(1523):14831489 , 2009 .
l. perfeito , l. fernandes , c. mota , and i. gordo .
adaptive mutations in bacteria : high rate and small effects . , 317(5839):813815 , 2007 . w. h. press , s. a. teukolsky , w. t. vetterling , and b. p. flannery . .
cambridge university press , 2007 .
m. reiter , s. rulands , and e. frey .
range expansion of heterogeneous populations .
, 112(14):148103 , 2014 .
i. m. rouzine , .
brunet , and c. o. wilke .
the traveling - wave approach to asexual evolution : muller s ratchet and speed of adaptation .
, 73(1):2446 , 2008 .
i. m. rouzine , j. wakeley , and j. m. coffin . the solitary wave of asexual evolution .
, 100(2):587592 , 2003 .
s. schiffels , g. j. szllsi , v. mustonen , and m. lssig .
emergent neutrality in adaptive asexual evolution .
, 189(4):13611375 , 2011 .
o. tenaillon .
the utility of fisher s geometric model in evolutionary genetics .
45:179201 , 2014 . l. s. tsimring , h. levine , and d. a. kessler . rna virus evolution via a fitness - space model . , 76(23):4440 , 1996 .
n. g. van kampen . , volume 1 .
elsevier , 1992 .
w. van saarloos .
front propagation into unstable states .
, 386(2):29222 , 2003 .
j. wakeley . , volume 1 .
roberts & company publishers greenwood village , colorado , 2009 . * relaxation rate spectra for models of adaptation and invasion .
* this figure quantifies the relaxation rates of correlations among labeled subpopulations in models of adaptation ( left ) and invasion ( right ) .
the relaxation rates @xmath250 for the @xmath63 model were found from a spectral analysis of the equation of motion for @xmath54 separable subpopulations in the @xmath63 correlation function , cf .
all eigenvalues are real and negative , i.e. , they lead to the decay of correlations .
each relaxation rate @xmath250 corresponds to a relaxation time @xmath251 . *
d * and * f * show the behavior of the two slowest decay rates @xmath252 and @xmath253 together with an approximation for @xmath252 .
the ratio @xmath254 controls the time scale separation between coalescence ( slow ) and wave profile relaxation ( fast ) .
in addition , the effects of chosing a particular @xmath54 for the closure of the moment hierarchy vanishes for large population sizes ( or wave speeds ) .
relaxation time scales of eigenmodes become more and more similar . figs .
[ fig : decaytimes ] and [ fig : fisherwaves ] depict only the first two timescales @xmath252 and @xmath253 ( here shown as bold dots in the upper panels and emphasized again in lower panels ) and concentrate on the case @xmath80 . ]
in the main text , we presented general equations for the correlation functions @xmath55 and the weighting function @xmath96 .
numerical results have been computed for the two cases @xmath86 and @xmath80 . for reference , here we state the explicit equations of motion for both , adaptive and invasive , waves .
the first step is always to calculate the weighting function @xmath96 in a comoving frame with speed @xmath255 : @xmath256 here , the term with multiplication of fitness , @xmath257 , indicates the selection term in adaptive waves , which has to replaced with @xmath258 for invasive waves , see below . after obtaining the general form of the weighting function @xmath96 ,
the mean stationary population density follows as @xmath259 a result that has already been described in ref .
@xcite .
if we choose @xmath80 , the model is tuned to have a closed 2-point correlation function @xmath110 , which is governed by @xmath260 in this case , the mean stationary population density is obtained by contraction , @xmath261 . for invasion waves the weighting function @xmath96 is the solution to @xmath262 which again sets the speed @xmath255 of the comoving frame as parameter .
in addition , tuning the model for the first moment , @xmath86 , yields the equation of motion for the stationary mean population density @xmath263 , @xmath264 choosing @xmath80 for invasion waves leads to @xmath265
although asymptotic analyses of noisy traveling wave models are often possible , closed form solution of either the fixation probability @xmath96 or correlation function @xmath55 are usually out of reach .
numerical methods can help alleviating this problem , as with tuned models we are able to state at least _ exact _ moment equations , which do not need an further approximations or assumptions . in order to solve the governing equations of our tuned model
numerically , we implemented an algorithm based on a multi - dimensional newton - raphson ( nr ) iteration scheme @xcite . here , we recount the basic steps of this scheme .
for the strictly one - dimensional case , @xmath86 , the numerical solution involves the roots of a set of @xmath266 equations in @xmath266 variables , @xmath267 which represent the steady state equations of motion , e.g. , eq . and eq . for the case of adaptation and @xmath86 .
the variables @xmath268 denote the value of the desired weighting function or correlation function at lattice point @xmath76 , respectively . for a small deviation @xmath269 in one of the variables
, one can expand @xmath270 into a series , @xmath271 in the nr scheme , one iterates the current guess for the solution @xmath272 to obtain @xmath273 .
the small difference @xmath274 is extrapolated by assuming the new solution @xmath275 fulfills the equation of motion , @xmath276 , while truncating after the _ linear _ term .
this leads to the equation @xmath277 .
thus , a single step comprises of evaluating the expression @xmath278 for each lattice point . in order to ensure a better convergence ,
the solutions @xmath279 for even and odd indices are computed consecutively . in the notation of
we also made use of a simplifying fact : the diffusion approximation ( for the mutation process in adaptive wave and movement in the invasive waves ) leads only to a `` local '' coupling , such that the equations of motion only depend on the values of @xmath280 at the focal lattice site @xmath76 and the two neighboring ones , @xmath281 .
for the case of adaptation waves , we discretize eq . and obtain the required expressions for the weighting function @xmath96 , @xmath282 which have to be inserted into eq . .
for obtaining solutions , the lattice spacing @xmath283 has to be chosen small enough , that the right - hand - side of is negative on the whole lattice and does not change sign for any @xmath284 .
the range of @xmath76 has to be adjusted to fit all characteristic features of the profiles onto the @xmath266 lattice points .
similar expressions to and hold for the mean stationary population density @xmath285 after discretizing . for the higher dimensional correlation functions @xmath55 ,
@xmath62 , the method can be extended in a straightforward fashion .
for instance , for @xmath80 one has @xmath286 variables @xmath287 and @xmath286 functions @xmath288 .
each dimension only adds two additional variables in the equation of motion , @xmath289 .
note that the discrete limit of the dirac - delta for correlations in the last term of eq .
is given by @xmath290 .
an improvement of this algorithm , that utilizes not only the `` local '' derivative @xmath291 , but the whole jacobian with entries @xmath292 , is often needed for extended mutation kernels @xmath23 in @xcite . in these cases ( and for reasonable parameter values ) , eq .
changes its sign twice on any lattice , regardless of the choice of lattice spacing @xmath283 , which renders this `` local '' approximation unusable .
however , suffices for all present purposes . in figure
[ fig : spectrum - si ] we displayed the spectral decomposition of the linear operator governing the equation of motion for the stationary mean population density . in this case , the governing ( discretized ) equations are given by the linear equation @xmath293 , with coefficients @xmath294 obtained from discretizing eq .
[ eq : sep - ansatz : dynamics ] .
the eigenvalues @xmath250 are obtained by a schur decomposition of the matrix @xmath294 , which leads to an ( quasi ) upper triangular matrix : along its diagonal it has @xmath295 blocks with real eigenvalues and @xmath296 blocks with its complex conjugate eigenvalues .
the existence of complex eigenvalues depends on the mutation ( or migration ) scheme . for a diffusion scheme ( i.e. a second derivative ) in eq .
one obtains only real eigenvalues numerically .
the code itself is based on the already implemented routines in the gnu scientific library ( gsl ) , in particular centered around the function ` sl_ei`en_nonsymmv to provide input and parse output .
the numerical code , implemented in c , is freely available from the authors .
for the stochastic simulations of adaptation waves , the continuous density @xmath1 in fitness space is discretized into bins on a regular one - dimensional lattice .
all individuals in the interval @xmath297 $ ] with @xmath298 are counted in the _ occupancy vector _ @xmath299 , @xmath300 the occupancies are updated in discrete time steps of length @xmath3 , which encompass the dynamics in the stochastic equation and the subsequent step to limit population sizes , or . the action of these equations consists of three sub - steps in the algorithm , indicated by superscripts in subsequent equations . in each of those steps the occupancies ( can ) change . first , the mean ( deterministic ) change due to a comoving frame , mutations and selection is applied , @xmath301 the first term for the comoving frame is only used in simulations with a tuned constraint .
the next term represents modifications in fitness due to mutations , while the last term represents growth due to selection . for the latter
, we have to distinguish again a fixed population size constraint and our tuned constraint .
the offset @xmath14 in the selection term is either set to the mean fitness @xmath302 in the fixed population size constraint , or set to @xmath303 as we incorporated the change in mean fitness already with the comoving frame . for the case of invasion waves ,
the selection term is replaced by @xmath304 . there
the offset @xmath14 is the position of the front , which we define as @xmath305 for a fixed population size constraint .
again , we have @xmath303 for the tuned constraint in its comoving frame . in the next sub - step
, the randomness due to birth and death events ( i.e. genetic drift ) further modifies all occupancies @xmath299 , @xmath306 where @xmath307 is a poisson - distributed random number with parameter @xmath299 .
this particular form of the noise ( with @xmath299 already updated from mutations and selection ) , ensures that ( i ) occupancies @xmath299 do not drop below zero , ( ii ) the mean value of the noise is zero and ( iii ) the variance at a lattice site @xmath76 amounts to @xmath308 per time step @xmath3 . while the introduction of occupancies @xmath299 instead of a density @xmath309 is irrelevant for most of the algorithm , it is convenient for this last feature ( iii ) .
the correlations of the noise are given by the expression @xmath310 in the discretized simulations , where the factor @xmath311 is then scaled already into the @xmath299 . in the last sub - step ,
the population is scaled uniformly to comply with its constraint , @xmath312 for the fixed population size constraint in section [ sec : second - subst - comp ] , we simply set @xmath313 for all @xmath314 ( or , for invasion waves @xmath315 for @xmath316 and @xmath317 otherwise ) . in the case of tuned models , we first solve the equation of motion for the fixation probability in a comoving frame . to obtain this numerical solution for @xmath96
, we utilize the code presented in appendix [ appendix : numerics ] .
simulation code , written in c , is available upon request from the authors .
in fig . [ fig : measurefixation]*b , c * we presented confidence intervals for the fixation probability , obtained via measurements of fixation and extinction events . _
a priori _ , counting events leads to an average value for the fixation probability , which might or might not be close to the expected ( theoretical ) value . in order to compute confidence intervals , additional assumptions have to be made , explained below . after generating different starting conditions @xmath5 ( snapshots from stochastic simulations ) , we label subpopulations @xmath318 in the nose of the wave , such that @xmath319 there are , of course , many ways to label a subpopulation such that the expected fixation probability is @xmath113 .
the simplest way is to label @xmath113 of the population in each bin , which unsurprisingly yields a fixation probabilities very close to @xmath113 ( cf .
[ fig : measurefixation : half ] ) .
however , this naive labelling protocol does not test our predictions for the spatial dependence of the fixation probability . to test the accuracy of our predictions in the spatially varying region of the fixation probability ,
we label the population in the tip of the wave , using the following form @xmath320 here , @xmath321 determines the position of the labelling and @xmath215 its steepness . to avoid artifacts associated with the discreteness , we choose the length scale @xmath215 such that on the order of @xmath322 bins , typically , contain both , labelled and unlabelled , subpopulations .
the crossover @xmath321 is iteratively adjusted until the condition in eq .
is met with sufficient accuracy ( usually , we demand a value close to machine precision , @xmath323 ) . * fixation and extinction measurements with the trivial labelling of @xmath324 . * with such a labelling
we obviously expect half of all simulations to end in fixation and half to end in extinction of the subpopulation @xmath318 .
the obtained confidence intervals serve as indicator for the noisiness of such measurements , when comparing results to the position depend labelling ( cf .
eq . ) presented in fig .
[ fig : measurefixation ] . in general , however , the initial hypothesis of a fixation probability of @xmath121 is corroborated by simulation results ( panels * bc * ) . ] from this configuration , we run the stochastic time evolution of the population until the labelled subpopulation either reaches fixation or goes extinct . in accordance with our interpretation of @xmath141 as fixation probability , we expect the labeled population to fix in half of the simulation runs and to go extinct otherwise . for definiteness , we abort simulations when one of the thresholds @xmath325 or @xmath326 is exceeded .
we count such events as fixation and extinction , respectively . after having amassed such simulation evidence , we can use bayesian inference to check if our assumption of the interpretation of @xmath141 as fixation probability is consistent @xcite .
the distribution of the fixation probability @xmath206 of the sub - population , given the ( simulation- ) data is computed as @xmath327 \sim \mathbb p\bigl[\,\mathrm{data}\,\vert\,{p_\ell^{\text{fix}}}\,\bigr]\,\mathbb p\bigl[\,{p_\ell^{\text{fix}}}\,\bigr]\;,\ ] ] using bayes theorem . here ,
the likelihood @xmath328 $ ] of observing either an extinction or fixation event is a simple binomial distribution : when having @xmath53 trials with @xmath329 fixation events , the likelihood is @xmath330= \binom{n}{x } ( { p_\ell^{\text{fix}}})^x(1-{p_\ell^{\text{fix}}})^{n - x}$ ] .
furthermore , we assume a flat prior @xmath331 $ ] for the fixation probability @xmath206 , ignoring any knowledge about its value at the beginning . such a flat prior can also be cast as a beta distribution ( incidentally a conjugate prior @xcite ) , which in its general form is given by @xmath332 using the two hyperparameters @xmath333 and @xmath334 we arrive at the uniform ( flat ) distribution on @xmath335 $ ] .
thus , the posterior distribution for the fixation probability @xmath206 of the sub - population can also be written as beta - distribution : @xmath336\sim ( { p_\ell^{\text{fix}}})^{x+\alpha-1}(1-{p_\ell^{\text{fix}}})^{n - x+\beta-1}\;,~~\alpha=\beta=1\;,\ ] ] up to a normalization factor . from this posterior
we can evaluate the @xmath120 confidence interval , and check if our assumption @xmath337 is within its range .
increasing the value of @xmath338 would increase the certainness of our initial hypothesis , @xmath337 , that we put into the model , narrowing the distribution . | the dynamics of adaptation is difficult to predict because it is highly stochastic even in large populations .
the uncertainty emerges from number fluctuations , called genetic drift , arising in the small number of particularly fit individuals of the population .
random genetic drift in this evolutionary vanguard also limits the speed of adaptation , which diverges in deterministic models that ignore these chance effects .
several approaches have been developed to analyze the crucial role of noise on the _ expected _ dynamics of adaptation , including the mean fitness of the entire population , or the fate of newly arising beneficial deleterious mutations .
however , very little is known about how genetic drift causes fluctuations to emerge on the population level , including fitness distribution variations and speed variations . yet , these phenomena control the replicability of experimental evolution experiments and are key to a truly predictive understanding of evolutionary processes . here , we develop an exact approach to these emergent fluctuations by a combination of computational and analytical methods . we show , analytically , that the infinite hierarchy of moment equations can be closed at any arbitrary order by a suitable choice of a dynamical constraint . this constraint regulates ( rather than fixes ) the population size , accounting for resource limitations .
the resulting linear equations , which can be accurately solved numerically , exhibit fluctuation - induced terms that amplify short - distance correlations and suppress long - distance ones .
importantly , by accounting for the dynamics of sub - populations , we provide a systematic route to key population genetic quantities , such as fixation probabilities and decay rates of the genetic diversity .
we demonstrate that , for some key quantities , asymptotic formulae can be derived . while it is natural to consider the process of adaptation as a branching random walk ( in fitness space ) subject to a constraint ( due to finite resources )
, we show that other noisy traveling waves likewise fall into this class of constrained branching random walks .
our methods , therefore , provide a systematic approach towards analyzing fluctuations in a wide range of population biological processes , such as adaptation , genetic meltdown , species invasions or epidemics .
many important evolutionary and ecological processes rely on the behavior of a small number of individuals that have a large dynamical influence on the population as a whole .
this is , perhaps , most obvious in the case of biological adaptation : future generations descend from a small number of currently well - adapted individuals .
the genetic footprint of the large majority of the population is wiped out over time by the fixation of more fit genotypes .
these dynamics can be visualized as a traveling wave in fitness space , see fig .
[ fig : noisy - wave - schema]a . at any time , the currently most fit `` pioneer '' individuals reside in the small tip of the wave . as time elapses , the wave moves towards higher fitness and the formerly rare most fit individuals dominate the population . by that time
, however , a new wave tip of even more fit mutants has formed and the cycle of transient dominance continues .
the principle of `` a few guiding the way for many '' also characterizes the motion of flocks of birds , which can be controlled by just a few leaders , or the expansion of an invasive species , which depends on pioneers most advanced into the virgin territory .
the overall dynamics of these processes can become highly erratic even in large populations because the behavior of the entire population is influenced by strong number fluctuations , called genetic drift , occurring in the small subset of `` pioneer '' individuals .
such propagation processes with an extreme sensitivity of noise have also been called `` pulled '' waves , because they are pulled along by the action of the most advanced individuals . if one ignores the fluctuations at the population level and is interested only in the expected dynamics of the population , one might be tempted to simply ignore genetic drift in models of pulled waves .
however , it turns out that mean - field models ignoring genetic drift drastically overestimate the speed of traveling waves , to the point that they predict an ever accelerating rather than a finite speed of adaptation .
it took 70 years since the first formulation of traveling wave models by fisher and kolomogorov , to realize that genetic drift influences both the expectation and the variation in singular ways @xcite .
the _ expected _ behavior of pulled waves has since been analyzed at great length .
many results were first obtained for waves of invasion , noisy versions of the classical `` fkpp '' model by fisher , kolmogorov , petrovskii , piskunov @xcite . in recent years , however , there has been a particularly strong research focus on models of adaptation .
these models aroused widespread interest because they can be applied to several types of data , including genomic data derived from experimental evolution experiments and from natural populations that undergo rampant adaptation , such as bacteria and viruses @xcite .
we now have analytical predictions for a number of valuable analytical or semi - analytical results for observables such as the mean speed , probability of fixation , distribution of fixed mutations in the asymptotic regime of large populations @xcite .
the great value of these results is that they show and rationalize which parameter combination chiefly influence the dynamics , and through which functional form .
importantly , it has been generally found that the overall dynamics depends logarithmically on population size and mutation rate .
the weak functional dependence is in fact at the root of universality observed in many such wave models : these predictions are independent of the precise details of the models , including the form of the non - linear population size regulation .
the basic challenge in analyzing noisy traveling arise from an essential non - linearity that is required to control population growth .
ignoring such a dynamical control of the population size leads to long - term exponential growth or population extinction .
progress in describing the _ mean _ behavior of front - sensitive models has been achieved by at least three different approaches : one can either heuristically improve the mean - field dynamics by setting the net growth rate equal to zero in regions where the population densities are too small @xcite .
such an ad - hoc approach , based on a growth rate cut - off , correctly reproduces the wave speed to the leading order but does not reveal other universal next - to - leading order corrections or the wave diffusion constant .
one can also invoke a branching - process approximation for the tip of the wave , thereby neglecting effects of the non - linear population size control , and then match this linearized description with a deterministic description of the bulk of the population @xcite . finally , there is also the possibility to invoke a particular dynamical constraint with the property that the dynamics exhibits a closed linear equation for the first moment .
importantly , this method , which has been called `` model tuning '' @xcite , reproduces the universal features of noisy traveling waves , which are independent of the chosen population control , ultimately because of the weakness of the population size dependence .
while understanding the mean behavior of noisy traveling waves has been an important achievement , the actual stochastic dynamics is characterized by pronounced fluctuations at the population level .
no two realizations of an evolution experiment , for instance , will exhibit the same time - dependent fitness distribution because of the chance effects involved in reproduction and mutations .
measuring the mean behavior requires many replicates in which the entire environment is accurately reproduced .
even if one has access to many replicates , as is possible in highly parallelized well - mixed evolution experiments , one can potentially learn a lot from the variability between replicates .
thus , a predictive understanding of the variability in evolutionary trajectories would greatly improve our quantitative understanding of how evolution works .
some exact results on fluctuations at the population level are available for a special model of fkpp waves @xcite .
still , we currently lack a systematic approach that can be applied to a wide range of models . here
, we fill this gap by extending the method of `` model tuning '' @xcite to the analysis of higher correlation functions : we show that it is possible to choose a constraint in such a way that the hierarchy of moments is closed at any desired level .
the resulting linear equations can be solved numerically and are amenable to asymptotic analytical techniques . as an important application of this approach , we show how the coalescence time can be computed within traveling wave models .
although our main results are applicable to a wide range of models , we focus our attention on simple models of adaptation . beyond simply grounding our discussion , there are two reasons to focus on these models . on the one hand ,
models of adaptation are simply important and have become an indispensable tool as a null model for evolutionary dynamics in microbial population . on the other hand , models of adaptation
manifestly exhibit a particular mathematical structure , which we call _ constrained branching random walks_. as we will argue below , this mathematical structure , to which all our formal results apply , can be identified as the essence of a wide range of models arising in physics , ecology and evolution . |
in the mainstream economy , the cornerstone for mathematical modeling of financial markets is the efficient market hypothesis ( emh)@xcite . a market is said to be efficient if all the existing information has affected the prices and any new random information influences the asset prices randomly @xcite .
the emh hypothesis is based on some assumptions such as rationality of the investors .
empirical studies have shown that the real prices do not completely follow a random walk which is the criterion of emh @xcite .
some methods such as behavioral models , try to explain the real behavior of asset prices . on the other hand ,
one of the behavioral aspects of markets is the entanglement of their conditions that lead to converting the local concepts to the global ones .
for example , the today market conditions have a dependency to the previous conditions .
so , researchers try to use new methods in order to model the behavior of the markets that consider the coupling structure of them .
one of the latest and most powerful approaches for this purpose is implementation of the concepts of quantum mechanics .
the previous studies have shown the usefulness of this approach for analyzing the financial markets .
application of quantum formalism in finance and economy has been shown by seagal et.al @xcite where they used quantum mechanics for option pricing . the work was completed by the paper of accardi et al @xcite .
baaquie has used the quantum concepts in different domains of finance @xcite .
haven has embedded the black- scholes option pricing model in a quantum physics setting@xcite .
one of the latest quantum formalisms for analyzing the financial markets is using the concept of bohmian mechanics@xcite . in the bohmian framework
, there is a potential quantum term that shows entanglements and interactions between particles very well .
so , this potential term is beneficiary for presenting the structure of the coupled systems such as financial markets that have different participant agents .
choustova @xcite presented a model for describing the structure of the financial markets based on the bohmian mechanics .
the model can describe the complexity of the financial markets by considering the psychology of agents of these markets in an information field .
in essence , this model can explain the behavioral aspects of markets by using the bohmian quantum mechanics @xcite . in this paper
, we have applied the chostova s proposed method in order to extract the information content of financial return time series .
we have used different databases ; covering securities from the tehran stock exchange ( tse ) as an emerge market @xcite , dow jones industrial average ( djia30 ) , standard and poor s 500 ( spx ) indices from a mature market , and gold & west texas intermediate oil prices as commodity markets .
we have analyzed the daily change of these markets from the 1st of jan 1996 until 1st of jan 2011 .
the paper is organized as follows : in section ii we explain the relation between the bohmian mechanics and finance theory . in section
iii we present our findings before stating our conclusions in section iv .
bohemian quantum mechanics is a special point of view to quantum theory with a causal interpretation .
this view was first presented by louis de broglie in 1927 , which was later rediscovered by david bohm in 1952 @xcite . in bohmian mechanics ,
a system of particles is described by a wave function that evolves according to schr@xmath0dinger s equation and a deterministic motion part @xcite .
the essential part of bohmian quantum is based on two special assumptions names as the non - locality and the deterministic trajectory .
this deterministic trajectory means that there is a guiding equation stemmed from newton equation that presents the actual positions of the particles .
the non - locality means that the behavior of any particle through the wave function which depends on the whole configuration of the universe relates to the other particles statues .
this property is equivalent to the notion of the entanglement which plays an important role in financial markets .
the bohemian wave function is as of the form @xmath1 where @xmath2 is the amplitude of the wave function and @xmath3 is its phase . substituting @xmath4 in the schr@xmath0dinger s equation before differentiating twice , two equations for @xmath5 can be obtained : @xmath6 the second term in eq .
2 contains an additional potential term besides the classical potential ( @xmath7 ) which was named by bohm as a quantum potential @xcite .
@xmath8 the hard financial conditions of markets such as natural resources and hard relations between traders can be described by the classical potential @xcite .
choustova used the quantum potential as a method for describing the mental factors or psychological aspects of the markets @xcite .
traders at the financial market behave stochastically due to free wills of individuals , where the combinations of a huge number of free wills lead to an additional stochastic term at the market that could not be described by classical potential . in the following section , by applying bohmian approach to financial markets and finding the quantum potential of the market returns , we try to find the psychological aspects of the markets .
this means that the today market return is entangled by the previous returns , where by considering this relationship , we obtain fine information about the market conditions .
the quantum potentials of different markets have been found by assuming @xmath9 based on the statistical inference of the probability density functions of the markets .
fig.1 shows the quantum potential of return time series of s&p500 index in different time scales . as we can see in daily scale , the maximum probable increase in returns is almost 7 percent . on the other hand ,
the maximum probable decrease in returns is about @xmath10 . in order to compare these data with real data , it is interesting to mention that there are three days that their returns are much more greater than the maximum limit ( @xmath11 ) and only there is one day ( black monday- 19 oct 1987 ) that its return is smaller than @xmath10 . as we can see for each market , there are two potential limits where the probability of occurrence of any returns outside these limits is very low .
in essence , these limits act as boundary to prevent the change of return time series .
based on the previous formulas , the quantum potential has an inverse relation with the probability distribution functions of the time series of returns . by comparing the behavior of a white noise with spx index , it is obvious that between the two limits , the real market has much higher probability for occurrence .
this phenomenon is stemmed from the fat - tail of the distributions of real markets . on the other hand , outside of these limits ,
the probability of occurrences for the white noise is much more than the spx . in fig .
2 , we have depicted the differences between two extreme limits for some mature and emerging markets during different time scales .
it is obvious that the slope of the scaling behaviors in the mature market indices ( spx and djia ) vs. scale approximately is about 0.4 . on the other hand ,
the value of the slope for the emerge market index ( tse ) is greater than 0.5 .
this behavior reminds us about the behavior of the hurst exponent in emerging and matures markets . as shown in previous studies , the hurst values for emerge and mature markets are greater and less than 0.5 respectively .
fig .3 shows that there is a cut - off in the slopes of the commodity markets ( oil and gold ) .
the slopes of these markets have two different values where they are the same as mature markets . before the cut - off points , in short scales the important factor that affects the structure of the markets is the competing nature of the firms in terms of supply and demand .
so , the commodity markets in this region have the same nature as the mature markets .
but , on the other hand in the large scales the most influential parameters that shape the behavior of the markets are the political and macro subjects .
so , the behavior of the commodity markets in large scales is different from the short scales .
in this paper , we have used the bohmian quantum approach to analyze the financial markets . by using this approach
we can describe the behavioral aspects of the market indices such as the coupling structure of markets returns during different times in terms of a quantum potential .
one of the important findings of this work is that the probable returns for each market lay in specific boundaries during different time scales .
also , by comparing the real market with a white noise , it is obvious that a real fat - tail distribution has a much more probability for occurrence within the boundaries . on the other hand , for a white noise
, the occurrence for any return could be made possible by increasing the quantum potential , while for the real market this issue is out of the question . by estimating the differences between maximum and minimum returns for some markets
, it is found that the quantum potentials of the return series possess scaling behaviors . also , the slopes of this scaling behavior in the mature and commodity markets are less than 0.5 and on the other hand , the slope of the emerge market ( tse ) is greater than 0.5 .
in addition , the cut - off observed in the curves of the commodity markets indicates the different efficiency of local and global effects .
the finding would remind us about the behavior of the hurst exponent , that was shown in previous work for mature markets , where the values where greater than 0.5 for emerge markets and less than 0.5 for mature markets .
based on the conclusions of this work , the bohmian quantum approach is a useful technique for analyzing the financial markets .
the authors would like to thank sara zohoor and dr .
s. vasheghani farahani for helping to edit the manuscript .
campbell , a.s .
low , and a.c .
mackinlay , the econometrics of financial markets .
princeton unoversity press ( 1997 ) .
harrison and s. pliska , martingales and stochastic integrals in the theory of countinous trading , stochastic process and their applications , 11 ( 1981 ) .
mantegna , h.e .
stanley , introduction to econophysics , cambridge university press , cambridge ( 2000 ) .
w. seagal , i.e. seagal , the black - scholes pricing formula in the quantum context .
proceeding of national academy of sciences of the usa 95 ( 1998 ) 4072 - 4075 .
l. accardi , a.boukas , the quantum black- scholes equation , gjpam 2(2 ) ( 2007 ) 155 - 170 .
baaquie , quantum finance , cambridge university press , cambridge , ( 2007 ) .
baaquie , m.srikant , and m. warachka , a quantum field theory term structure model applied to hedging , international journal of theoretical and applied finance 6(5 ) ( 2003 ) 443 .
baaquie and s. marakani , finite hedging in field theory models of interest rates , physical review e 69 ( 2004 ) 036130 .
baaquie and m. srikant, comparison of field theory models of interest rates with market data , physical review e 69 ( 2004 ) 036129 .
baaquie , c. coriano , and m. srikant , hamiltonian and potentials in derivative pricing models : exact results and latice simulations , physica a 334 ( 2004 ) 531 - 557 .
baaquie , quantum field theory of forward rates with stochastic volatility , physical review e 65 ( 2002 ) 056122 .
pouria pedram , the minimal length uncertainty and the quantum model for the stock market , physica a 391 ( 2012 ) 2100 - 2105
. e. haven , a discussion on embedding the black- scholes option pricing model in a quantum physics setting , physica a 304 ( 2002 ) 507 - 524 .
e. haven , a black - scholes shrodinger option price : bit versus qubit , physica a 324 ( 2003 ) 201 - 206 .
e. haven , an
h - brownian motion of stochastic option prices , physica a 344 ( 2003 ) 151 - 155 .
e. haven , pilot - wave theory and financial option pricing , international journal of theoretical physics 44 ( 11 ) ( 2005 ) 1957 - 1962 .
o. choustova , quantum bohmian model for financial markets , physica a 347 ( 2006 ) 304 - 314 .
o. choustova , quantum probability and financial market , information sciences 179 ( 2009 ) 478 - 484 .
o. choustova , quantum model for the price dynamics : the problem of smoothness of trajectories , j. math .
( 2008 ) 296 - 304 .
o. choustova , bohmian mechanics for financial process , j. modern optics 51 ( 2004 ) 1111 .
p. norouzzadeh , g. r. jafari , physica a 356 ( 2005 ) 609 - 627 ; g. r. jafari , a. bahraminasab , p. norouzzadeh , international journal of modern physics c 18 ( 2007 ) 1223 - 1230 .
d. bohm , quantum theory , prentice - hall , englewood cliffs , new jersey ( 1952 ) .
d. bohm , b. hiley , the undivided universe : an ontological interpretation of quantum mechanics , routledge and kegan paul , london ( 1993 ) . | the bohmian quantum approach is implemented to analyze the financial markets . in this approach
, there is a wave function that leads to a quantum potential .
this potential can explain the relevance and entanglements of the agent s behaviors with the past .
the light is shed by considering the relevance of the market conditions with the previous market conditions enabling the conversion of the local concepts to the global ones .
we have shown that there are two potential limits for each market .
in essence , these potential limits act as a boundary which limits the return values inside it . by estimating the difference between these two limits in each market
, it is found that the quantum potentials of the return time series in different time scales , possess a scaling behavior .
the slopes of the scaling behaviors in mature , emerging and commodity markets show different patterns .
the emerge market having a slope greater than 0.5 , has a higher value compared to the corresponding values for the mature and commodity markets which is less than 0.5 .
the cut - off observed in the curve of the commodity market indicates the threshold for the efficiency of the global effects . while before the cut - off , local effects in the market are dominant , as in the case of the mature markets .
the findings could prove adequate for investors in different markets to invest in different time horizons . |
concept inventories are used to assess learning in many areas of physics education @xcite . when used to determine the effectiveness of educational innovations they may contribute to the teaching development cycle . since the literature on special relativity education research does not include a concept inventory we have developed the relativity concept inventory ( rci ) , available from the supplemental appendix to this paper .
special relativity is interesting in a physics education research context because of its combination of deeply challenging concepts and simple mathematics .
this is in contrast with quantum mechanics , which has a more complex mathematical structure . nevertheless , the amount of physics education research on special relativity is small @xcite .
the rci has been validated by feedback from discipline experts and its validity and reliability established by standard methods @xcite .
these include the self - referential statistics of classical test theory , and benchmarking against traditional assessment such as homework and an exam .
we have also developed and applied monte carlo simulation techniques suitable for the analysis of correlations in data with small sample size . in the next section
we describe the process used to develop the rci . in section [ the students ]
we characterize the students the rci was administered to . in section [ data analysis methods ] we describe the methods used to analyse the collected data , including the use of : item response theory to control for the effect of student ability on correlations between questions , and monte carlo modeling to estimate the statistical significance of correlations . in section [ results ]
we present misconceptions diagnosed by the rci and evidence for its gender bias .
finally , in the conclusions , we suggest revisions of the rci . we also argue that understanding the gender bias in concept inventories is a significant problem for physics education research .
the development of the rci followed adams and wieman @xcite insofar as our six month project schedule allowed .
in particular , student interviews were not relied on as much as suggested by them . the only previous attempt to develop a concept inventory for special relativity
is reported in gibson s doctoral thesis @xcite .
we first formulated a list of concepts that captured the learning goals of the introductory relativity instruction in the physics 2 course at the the australian national university ( anu ) .
these concepts were also informed by relevant textbooks @xcite and the physics education research literature .
expert feedback on each of fourteen draft concepts of introductory relativity was obtained from thirty international respondents @xcite using an online survey .
agreement with the the appropriateness of the concepts in our list ranged from 100% for the first postulate to 50% . after individual consideration ,
concepts with agreement below 75% were dropped from the list .
the final list of nine concepts is given in table [ concepts ] .
[ cols="<,<,<",options="header " , ]
classical test theory suggests that the rci may be too easy and , perhaps consequently , insufficiently discriminating .
however , we do not recommend revisions , other than those suggested below , until data from a wider range of students has been analysed .
in section [ item response theory ] we concluded that question 24 , concerning the concept of mass - energy equivalence , should be dropped from the rci .
it has zero discrimination , and is the only question having a negative normalised gain between the pre and post - tests .
it was also found to have a strong negative correlation with an apparently unrelated question .
if dropped , the concept of mass - energy equivalence would not be tested by the rci .
in section [ misconceptions ] we concluded that a frame symmetrical pair of length contraction questions is desirable , mirroring the symmetrical pair of time dilation questions .
hence we recommend that a partner question be added to the existing rci length contraction question 13 .
however , any such question would require validation along the lines described in sections [ the development process ] and [ data analysis methods ] .
the evidence presented in section [ gender differences ] suggests that the rci is gender biased .
previous work has shown similar biases in the force concept inventory and in other concept inventories .
concept inventories are useful because they can help evaluate innovation and hence improve teaching .
however if their evaluations are biased with respect to certain student groups there is a risk that improved learning for some comes at the expense of the learning of others .
it is a task for future physics education research to investigate and understand this interesting and important problem .
the authors would like to thank the academics and students who helped develop the rci .
we would especially like to acknowledge a. wilson for her help with the student interviews , and p. francis for his advice on statistical analysis .
99 assessment instrument information page , nc state physics education r and d group , http://www.ncsu.edu/per/testinfo.html , accessed february 26 , 2013 .
m. wegener , t. mcintyre , d. mcgrath , c. savage , m. williamson , developing a virtual physics world , australasian journal of educational technology , * 28 * ( special issue , 3 ) , 504 ( 2012 ) .
d. mcgrath , m. wegener , t. mcintyre , c. savage , m. williamson , student experiences of virtual reality - a case study in learning special relativity , am .
* 78 * , 862 ( 2010 ) . c. m. savage , a. searle , l. mccalman , real time relativity : exploratory learning of special relativity , am .
* 75 * , 791 ( 2007 ) .
peter w. hewson , a case study of conceptual change in special relativity : the influence of prior knowledge in learning , eur .
educ . * 4 * , 61 ( 1982 ) .
george j. posner , kenneth a. strike , peter w. hewson , and william a. gertzog , accommodation of a scientific conception : toward a theory of conceptual change , sci . educ . * 2 * , 211 ( 1982 ) .
a. villani , and j. l. a. pacca , students spontaneous ideas about the speed of light , int .
* 9 * , 55 ( 1987 ) .
n. d. mermin , lapses in relativistic pedagogy , am . j. phys .
* 62 * , 11 ( 1994 ) .
r. e. scherr , p. s. shaffer , and s. vokos , student understanding of time in special relativity : simultaneity and reference frames , phys .
res . , am .
. suppl . * 69 * , s24 ( 2001 ) .
r. e. scherr , p. s. shaffer , and s. vokos , the challenge of changing deeply held student beliefs about the relativity of simultaneity , am . j. phys . *
70 * , 1238 ( 2002 ) .
m. belloni , w. christian , and m. h. darcy , teaching special relativity using physlets , phys .
42 * , 8 ( 2004 ) .
r. e. scherr , modeling student thinking : an example from special relativity , am . j. phys . *
75 * , 272 ( 2007 ) .
stephanie v. chasteen , rachel e. pepper , marcos d. caballero , steven j. pollock , and katherine k. perkins , colorado upper - division electrostatics diagnostic : a conceptual assessment for the junior level , phys .
8 * , 020108 ( 2012 ) .
wendy k. adams and carl e. wieman , development and validation of instruments to measure learning of expert - like thinking , int
* 33 * , 1289 ( 2011 ) .
k. gibson , special relativity in the classroom , doctoral thesis , arizona state university ( 2008 ) .
edwin f. taylor and john archibald wheeler , spacetime physics : introduction to special relativity , w. h. freeman & co. , new york , 1982 ; nathaniel david mermin , it s about time , princeton university press , new jersey , 2005 ; sean carroll , spacetime and geometry , addison - wesley , san francisco , 2004 ; wolfgang rindler , relativity , special , general and cosmological , oxford university press , 2006 .
expert input was solicited by email from the members of the australasian society for general relativity and gravitation , and from members of the education group of the australian institute of physics .
responses were also obtained through a posting to the matter and interactions yahoo group .
kirk allen , teri reed - rhoads , and robert terry .
work in progress : assessing student confidence of introductory statistics concepts . in : proceedings - frontiers in education - 36th annual conference , pgs .
ieee ( 2006 ) .
kirk allen , andrea stone , teri reed - rhoads , and teri j. murphy , the statistics concepts inventory : developing a valid and reliable instrument , proceedings , american society for engineering education annual conference & exposition ( 2004 ) .
kirk allen , teri reed - rhoads , and terry robert , work in progress : assessing student confidence of introductory statistics concepts , proceedings .
frontiers in education .
36th annual conference ( 2006 ) .
manjula devi sharma and james bewes .
self - monitoring : confidence , academic achievement and gender differences in physics .
journal of learning design * 4 * , issue 3 ( 2011 ) .
this project was approved by the anu human research ethics committee : human ethics protocol 2012/380 .
real time relativity website , http://realtimerelativity.org , accessed 26 february 2013 . the atar score ranks graduating secondary school students .
the physics 2 median atar score of 95 indicates that the median student scored higher than 95% of other year 12 students .
nate silver . the signal and the noise . the penguin press , new yory , 2012 . chapters 5 and 8
. william h. press , brian p. flannery , saul a. teukolsky , william t. vetterling .
numerical recipes : the art of scientific computing .
cambridge university press , new york , 1st edition , 1986 .
dennis d. wackerly , william iii mendenhall , and richard l. scheaffer .
mathematical statistics with applications .
thomson , belmont , 7th edition , 2008 .
lin ding , ruth chabay , bruce sherwood , and robert beichner , evaluating an electricity and magnetism assessment tool : brief electricity and magnetism assessment , phys .
. res . * 2 * , 010105 ( 2006 ) .
lin ding and robert beichner , approaches to data analysis of multiple - choice questions , phys .
. res . * 5 * , 020103 ( 2009 ) .
richard r. hake , interactive - engagement versus traditional methods : a six - thousand - student survey of mechanics test data for introductory physics courses , am .
* 66 * , 64 ( 1998 ) .
we used the multinomialdistribution function of mathematica 8 to generate our random samples .
this idea was suggested to us by dr .
paul francis of anu .
maja planinic , lana ivanjek , and ana susac , rasch model based analysis of the force concept inventory , phys .
res . * 6 * , 010103 ( 2010 ) .
we used ministep to do the rasch analysis : http://www.winsteps.com/ministep.htm , accessed 23 february 2013 .
harry h. harman .
modern factor analysis .
university of chicago press , chicago , 3rd edition 1976 .
terry f. scott , daniel schumayer , and andrew r. gray , exploratory factor analysis of a force concept inventory data set , phys .
8 * , 020105 ( 2012 ) .
robert c. maccallum , keith f. widaman , shaobo zhang , and sehee hong , sample size in factor analysis , psychological methods * 4 * , 84 ( 1999 ) .
daniel j. mundfrom , dale g. shaw , and tian lu ke , minimum sample size recommendations for conducting factor analyses , int . j. testing * 5 * , 159 ( 2005 )
. j.c.f .
de winter , d. dodou , and p. a. wieringa , exploratory factor analysis with small sample sizes , multivariate behavioral research * 44 * , 147 ( 2009 ) .
we used the statistical analysis package spss for our factor analysis .
it calculates question communalities .
sudhir panse , jayashree ramadas , and arvind kumar , alternative conceptions in galilean relativity : frames of reference , int .
* 16 * , 63 ( 1994 ) .
laura mccullough , gender , context and physics assessment , j. int .
women s studies , * 5 * , 20 ( 2004 ) .
r. d. dietz , r. h. pearson , m. r. semak , and c. w. willis , gender bias in the force concept inventory ?
in , aip conference proceedings 1413 : physics education research conference , pgs .
171 - 174 ( 2011 ) .
jennifer docktor and kenneth heller , gender differences in both force concept inventory and introductory physics performance . in , aip conference proceedings 1064 : physics education research conference , pgs .
15 - 18 ( 2008 ) .
vincent p. coletta , jeffery phillips , and jeffery j. steinert , fci normalized gain , scientific reasoning ability , thinking in physics , and gender effects . in , aip conference proceedings 1413 : physics education research conference , pgs .
23 - 26 ( 2011 ) .
lauren e. kost - smith , steven j. pollock and noah d. finkelstein , gender disparities in second - semester college physics : the incremental effects of a smog of bias , phys .
* 6 * , 020112 ( 2010 ) .
mercedes lorenzo , catherine h. crouch , and eric mazur , reducing the gender gap in the physics classroom , am .
* 74 * , 118 ( 2006 ) .
l. e. kost , s. j. pollock and n. d. finkelstein , characterizing the gender gap in introductory physics , phys .
res . * 5 * , 010101 ( 2009 ) .
robert j. beichner , testing student interpretation of kinematics graphs , am . j. phys . *
62 * , 750 ( 1994 ) . w. k. adams , k. k. perkins , n. s. podolefsky , m. dubson , n. d. finkelstein , and c. e. wieman , new instrument for measuring student beliefs about physics and learning physics : the colorado learning attitudes about science survey , phys . rev .
. res . * 2 * , 010101 ( 2006 ) .
j. mckendree and c. smith , faq - i have heard that multiple - choice questions ( mcqs ) are biased in favour of males .
what is the evidence for this ? , http://www.heacademy.ac.uk/resources/detail/subjects/ medev
/ faq - i - have - heard - that - multiple - choice - questions - are - biased - in - favour - of - males , accessed 23 february 2013 .
n. cole , the ets gender study : how females and males perform in educational settings .
educational testing service technical report ( 1997 ) , http://www.eric.ed.gov / ericwebportal / detail?accno= ed424337 , accessed 23 february 2013 .
this is the version of the rci that was used in the post - test
. + + * instructions : * * _ some of the questions are multiple choice , with an additional confidence scale similar to the example below . for each of these questions ,
circle the answer that you agree most with , and mark on the scale how confident you are in your choice .
_ * _ some of the questions are in the form of statements with which you may agree or disagree .
circle the response that most closely corresponds to your position on the question .
_ * _ in all of the following questions , the symbol @xmath1 represents the speed of light in a vacuum , @xmath2 .
_ * _ answer all of the questions to the best of your knowledge . _
+ _ _ in the following two questions , alice is standing in a train moving at velocity @xmath3 from * left to right * relative to bob , who is standing on a platform .
as alice passes bob , she drops a bowling ball out of the train s window : 1 .
consider a spaceship travelling from earth towards a distant star at a constant high velocity @xmath3 relative to earth .
the spaceship sends a light pulse back to earth . on earth
, the speed of this pulse is measured to be : 1 .
@xmath1 2 .
@xmath4 3 .
@xmath5 in the following two questions , abbey is in a spaceship moving at high speed relative to brendan , who is standing on an asteroid ( a very small piece of rock floating in space ) .
she flies past him so that at @xmath6 , she is momentarily adjacent to brendan . 1 . at the instant that abbey s ship passes brendan , she sends two light pulses to him from her ship . if the light pulses are emitted a nanosecond ( @xmath7 seconds ) apart
according to abbey s clock , what will be the time interval between the pulses according to brendan ? 1 .
greater than one nanosecond 2 .
equal to one nanosecond 3 .
less than one nanosecond 1 . also while abbey s ship passes brendan , brendan sends two light pulses to abbey .
if brendan sends the light pulses a nanosecond ( @xmath7 seconds ) apart according to his clock , what will be the time interval between the pulses according to abbey ? 1 .
greater than one nanosecond 2 .
equal to one nanosecond 3 .
less than one nanosecond 1 .
it is known that our galaxy is of the order of @xmath8 light - years in diameter .
true or false : `` travelling at a constant speed that is less than , but close to , the speed of light , in principle it is possible for a person to cross the galaxy within their lifetime . '' 1 .
false 1 .
the olympic games is a two - week long sports competition .
an interested alien astronomer watches the olympics from a distant planet moving at high speed relative to earth .
_ if the alien were to compensate for the time the light from earth takes to reach them _
, they would measure the length of the olympics to be : 1 .
greater than two weeks 2 .
equal to two weeks 3 .
less than two weeks in the following two questions , the scenario is as follows :
alex and his friend bianca decide to set off on separate voyages in identical spaceships .
they each speed away from earth in opposite directions - alex at @xmath9 to the left , and bianca at @xmath9 to the right , relative to an observer on earth . 1 .
if alex measures the rate at which his distance to bianca is increasing , he will obtain a value that is : 1 .
equal to @xmath10 2 .
greater than @xmath1 but less than @xmath10 3 .
equal to @xmath1 4 .
greater than @xmath11 but less than @xmath1 5 .
equal to @xmath11 1 .
if cameron , an observer on earth , measures the rate at which the distance between alex and bianca is increasing , he will obtain a value that is : 1 .
equal to @xmath10 2 .
greater than @xmath1 but less than @xmath10 3 .
equal to @xmath1 4 .
greater than @xmath11 but less than @xmath1 5 .
equal to @xmath11 in the following four questions , amanda is standing on a train travelling at high speed past bryan , who is standing on a platform .
as she passes bryan , she drops two bowling balls out of the window at the same time ( amanda s time ) , and from an arm s span apart . 1 .
bryan stands on the platform and watches the balls fall to the ground . _
if he compensates for the time that the light from the impacts takes to reach him _
, in what order does bryan measure the balls hitting the ground ? 1 .
at the same time 2 .
one ball before the other 1 .
charlotte is another passenger on the train with amanda . _
if she compensates for the time that the light from the impacts takes to reach her _
, in what order does charlotte measure the balls hitting the ground ? 1 .
at the same time 2 .
one ball before the other 1 .
amanda has an arm span of @xmath12 meters at rest .
if bryan performs a measurement of amanda s arm span as she passes him , he will obtain a value : 1 .
greater than @xmath12 2 .
equal to @xmath12 3 .
less than @xmath12 1 .
amanda also has a height of @xmath13 meters at rest .
if bryan performs a measurement of amanda s height as she passes him , he will obtain a value : 1 .
greater than @xmath13 2 .
equal to @xmath13 3 .
less than @xmath13 1 .
two separate light bulbs emit flashes of light , distant from an observer .
this observer receives the light from both flashes at the same time . from this alone it is possible to conclude that : 1 .
the flashes occurred at the same time for all observers 2 .
the flashes occurred at the same time for the observer at that location 3 .
the flashes occurred at the same time if the observer is not moving relative to the light bulbs 4 .
it is not possible to make any of the above conclusions 1 . in the following thought experiment , you are in a high speed train travelling along a railway . true or false : `` if you measure the dimensions of the train compartment , you will obtain different values than if the train were at rest . '' 1 .
false 1 .
consider a futuristic space station that specialises in constructing fast spaceships .
once the ships are built , they leave the station at high speed for testing .
as they leave the station at speed , a serial number is stamped instantaneously on the side of the ship by a machine on the station .
this serial number has length @xmath12 as measured by a builder on the space station .
after the ship has finished its test run , it returns to the station and is parked in the garage .
what is the length of the serial number now , as measured by the builder on the space station ?
greater than @xmath12 2 .
equal to @xmath12 3 .
less than @xmath12 1 .
adam is in a spaceship moving at @xmath14 relative to our galaxy .
adam wants to measure the mass of his ship by observing how resistant the ship is to acceleration .
if adam exerts a force on the ship ( by turning on a rocket engine , for example ) and measures ( with an accelerometer inside the ship ) the acceleration that results , he will obtain a value that is : 1 .
greater than what he would measure if his ship were at rest relative to the galaxy .
equal to what he would measure if his ship were at rest relative to the galaxy .
less than what he would measure if his ship were at rest relative to the galaxy . 1 . in the following thought experiment , you are in a high speed train travelling along a railway . true or false
: `` if you measure the rate at which your watch is ticking , you will obtain a different value than if the train were at rest . ''
false 1 .
you are in a well equipped physics lab without windows or ways of interacting with the outside world .
it is known that the lab is in uniform motion .
how do you determine the velocity of the lab ? 1 .
you throw a ball across the lab and measure its change in velocity 2 .
you shine a laser beam across the lab and measure its change in velocity 3 . either ( a ) or ( b ) 4 .
it is not possible to determine the lab s velocity by experiment 1 .
you observe a set of distant , spatially separated clocks that are synchronised in their rest frame .
you are at rest relative to the clocks , and you observe ( through a telescope ) that the times read on the clocks are different .
this is due to : 1 .
time dilation 2 .
length contraction 3 .
relativity of simultaneity 4 . none of the above 1 .
if two events are separated in such a way that an observer can be present at both events , which relationship(s ) between the two events are the same for all observers ? 1 .
the time between the two events 2 . the distance between the two events 3 .
the order in which the events occur 4 .
none of these relationships are the same for all observers 1 . if two events are separated in such a way that * no * observer can be present at both events , which relationship(s ) between the two events are the same for all observers ? 1 .
the time between the two events 2 . the distance between the two events 3 .
the order in which the events occur 4 .
none of these relationships are the same for all observers 1 .
consider a closed box , containing an equal amount of matter and antimatter .
the total mass of this box and its contents is initially @xmath15 .
the matter and antimatter are then allowed to annihilate inside the box , turning into photons in the process . what is the total mass of the box and its contents _ after _ the annihilation ? 1 .
greater than @xmath15 2 .
equal to @xmath15 3 .
less than @xmath15 | we report on a concept inventory for special relativity : the development process , data analysis methods , and results from an introductory relativity class .
the relativity concept inventory tests understanding of kinematic relativistic concepts .
an unusual feature is confidence testing for each question .
this can provide additional information ; for example high confidence correlated with incorrect answers suggests a misconception .
a novel aspect of our data analysis is the use of monte carlo simulations to determine the significance of correlations .
this approach is particularly useful for small sample sizes , such as ours .
our results include a gender bias that was not present in other assessment , similar to that reported for the force concept inventory . |
this work was supported by research funds of chonbuk national university ( 2004 ) and the korea research foundation grant ( moehrd ) ( r14 - 2002 - 059 - 01000 - 0 ) ( hh ) , and by the research grants council of the hksar under project 2017/03p and hong kong baptist university under project frg/01 - 02/ii-65 ( lht ) .
k. binder , in _ finite - size scaling and numerical simulation of statistical systems _ , edited by v. privman ( world scientific , singapore , 1990 ) , p. 173 ; k. binder and d. w. heermann , _ monte carlo simulation in statistical physics .
an introduction _ , 3rd ed .
( springer , berlin , 1997 ) .
y. kuramoto , in _ proceedings of the international symposium on mathematical problems in theoretical physics _ , edited by h. araki ( springer - verlag , new york , 1975 ) ; y. kuramoto , _ chemical oscillations , waves , and turbulence _
( springer - verlag , berlin , 1984 ) ; y. kuramoto and i. nishikawa , j. stat . phys . * 49 * , 569 ( 1987 ) . | the binder cumulant ( bc ) has been widely used for locating the phase transition point accurately in systems with thermal noise . in systems with quenched disorder
, the bc may show subtle finite - size effects due to large sample - to - sample fluctuations .
we study the globally coupled kuramoto model of interacting limit - cycle oscillators with random natural frequencies and find an anomalous dip in the bc near the transition .
we show that the dip is related to non - self - averageness of the order parameter at the transition .
alternative definitions of the bc , which do not show any anomalous behavior regardless of the existence of non - self - averageness , are proposed .
the characterization of phase transitions relies mainly on the singularity structure of physical quantities at the transition , which can be quantified by critical exponent values . in numerical efforts ,
the accuracy of the estimated exponents heavily depends on the precision of locating the phase transition point . in the case of most thermal systems , the binder cumulant ( bc )
is widely believed to provide one of the most accurate tools for estimating the transition point @xcite . the critical bc value at the transition is also believed to be universal , even though there is still controversy over its universality @xcite . in some complex systems @xcite
, the bc shows an anomalous negative dip in finite systems , which represents a rugged landscape ( multi - peak structure ) in the probability distribution function ( pdf ) of the order parameter .
great care is required in analyzing numerical data to see whether the dip will vanish in the thermodynamic limit .
if it does , the negative dips in the finite systems can be attributed to long - living metastable states .
otherwise , a nonvanishing negative dip usually implies that the transition is not continuous , but is of the first order . in systems with quenched disorder
, the disorder fluctuation may also generate an anomalous negative dip in the conventional bc , which is defined as the ratio of the disorder - averaged moments of the order parameter . in this case
, the negative dip may be related to the non - self - averageness ( nsa ) of the order parameter , which usually implies an extended and/or multi - peak structure in the disorder - averaged pdf @xcite .
we consider a typical nonequilibrium dynamical system with quenched disorder , such as the kuramoto model of interacting limit - cycle oscillators with random natural frequencies @xcite .
the dynamic synchronization transition is dominated by space - time fluctuations of the order parameter .
the quenched disorder is , by definition , perfectly correlated in the time direction , so it may generate strong disorder fluctuations similar to quantum systems with random defects @xcite .
in fact , we recently showed that the disorder fluctuation was anomalously strong near the synchronization transition @xcite .
we take the globally coupled kuramoto model , which can be solved analytically to some extent .
the model is defined by the set of equations of motion @xmath0 where @xmath1 represents the phase of the @xmath2th limit - cycle oscillator @xmath3 .
the first term @xmath4 on the right - hand side denotes the natural frequency of the @xmath2th oscillator , where @xmath4 is assumed to be randomly distributed according to the gaussian distribution function @xmath5 characterized by the correlation @xmath6 and zero mean @xmath7 .
we note that the natural frequency @xmath4 plays the role of _ quenched disorder"_. the second term of eq .
( [ eq : model ] ) represents global ( all - to - all ) coupling with equal coupling strength @xmath8 .
the sine coupling form is the most general representation of the coupling in the lowest order of the complex ginzburg - landau ( cgl ) description@xcite , and its periodic nature is generic in limit - cycle oscillator systems .
we consider the ferromagnetic coupling ( @xmath9 ) , so the neighboring oscillators favor their phase difference being minimized .
the scattered natural frequencies and the coupling of the oscillators compete with each other .
when the coupling becomes strong enough to overcome the dispersion of natural frequencies , macroscopic regions in which the oscillators are synchronized by sharing a coupling - modified common frequency @xmath10 may emerge .
collective phase synchronization is conveniently described by the complex order parameter defined by @xmath11 where the amplitude @xmath12 measures the phase synchronization and @xmath13 indicates the average phase .
when the coupling is weak , each oscillator tends to evolve with its own natural frequency , resulting in the fully random desynchronized phase ( @xmath14 ) . as the coupling increases
, some oscillators with @xmath15 become synchronized , and their phases @xmath16 start to show some ordering ( @xmath17 ) .
equation ( [ eq : model ] ) can be simplified to @xmath18 decoupled equations @xmath19 where @xmath12 and @xmath13 are to be determined by imposing self - consistency . in the steady state ( @xmath20 ) , the self - consistency equation reads @xmath21 with @xmath22 and @xmath23 @xcite .
this equation has a nontrivial solution only when @xmath24 : @xmath25 with @xmath26 .
we note that the exponent @xmath26 corresponds to the mean field ( mf ) value for systems of locally coupled oscillators @xcite .
now , we perform numerical integrations of eq .
( [ eq : model ] ) by using heun s method @xcite for various system sizes of @xmath27 to @xmath28 . for a given distribution of disorder @xmath29
, we average over time in the steady state after some transient time . after the time average
, we also average over disorder .
typically , we take the time step @xmath30 , the maximum number of time steps @xmath31 , and the number of samples @xmath32 . for convenience ,
we set @xmath33 ( unit variance ) ; then , the corresponding critical parameter value is @xmath34 .
figure [ fig : gl_m ] shows the behavior of the phase synchronization order parameter @xmath12 against the coupling strength @xmath35 for various system sizes @xmath18 . in the weak coupling region ( @xmath36 )
, we find that the order parameter approaches zero as @xmath37 , which is a characteristic of the fully random phase . in the strong coupling region @xmath38 ,
@xmath12 saturates to a finite value , indicating a phase transition at @xmath39 in the thermodynamic limit @xmath40 , which is consistent with the analytic result . to pin down the transition point @xmath41
precisely , we use the binder cumulant method @xcite .
the fourth - order cumulant of the order parameter , the binder cumulant ( bc ) , is defined in thermal systems as @xmath42}{3[\langle \delta^2 \rangle]^2 } b_{\delta } = 1-\frac{\langle \delta^4 \rangle } { 3\langle \delta^2 \rangle^2 } , \label{eq : gl_bm}\ ] ] where @xmath43 represents the thermal ( time ) average . in systems with quenched disorder ,
on the other hand , we should consider the disorder average besides the thermal one .
we may first consider the bc as the disorder - averaged moment ratio @xcite @xmath44}{3[\langle \delta^2 \rangle]^2 } , \label{eq : gl_bm_old}\ ] ] where @xmath45 $ ] denotes the disorder average , i.e. , the average over different realizations of @xmath29 .
figure [ fig : gl_bm_old ] displays @xmath46 as a function of the coupling strength @xmath35 for various system sizes @xmath18 . in the region of weak coupling @xmath47 , we expect the random nature of the oscillator phases @xmath48 to yield an asymmetric poisson - like probability distribution function ( pdf ) characterized by @xmath49 with a constant @xmath50 , which leads to @xmath51 . on the other hand , in the strong - coupling region ,
the pdf becomes a @xmath52-like function with a very narrow variance , which leads to @xmath53 .
the numerical data in fig .
2 are consistent with our predictions
. however , near the transition , the @xmath46 shows a big anomalous _ dip " _ on the desynchronized side . as the system size increases , the dip develops initially with a broad width and then becomes sharper and also deeper .
the dip s position moves toward the transition point .
the crossing points seem to nicely converge to the critical point @xmath54 . however , as the system size increases , the presence of the dip starts to hinder us in locating the critical point accurately . in this letter , we explain why the dip develops in this system and propose alternative definitions of the binder cumulant that do not show any dip in the same system .
we measure the disorder ( sample - to - sample ) fluctuations defined as @xmath55}{[\langle{\cal o}\rangle]^2}-1 , \label{eq : a}\ ] ] where @xmath56 is any observable , such as @xmath12 and @xmath57 , in a system .
this quantity is positive definite and is supposed to vanish in the thermodynamic limit in _ self - averaging _ systems and to remain finite in non - self - averaging systems @xcite . as one can see in fig .
[ fig : gl_a ] , the disorder fluctuation @xmath58 is quite sizable in the range of @xmath35 where the dip appears ( @xmath59 shows a similar behavior ) .
in other words , the @xmath46 shows a dip where the system is not well self - averaged .
a careful finite - size analysis on @xmath58 reveals that it vanishes as @xmath60 away from criticality , but saturates to a finite value at criticality .
the non - self - averageness at criticality is not surprising because the quenched randomness in natural frequencies should be relevant at this transition .
strong disorder fluctuations may cause non - negligible spreading of the _ effective _ coupling constants over different realizations of disorder @xcite .
figure [ fig : pdf ] shows for 20 independent samples , the pdf of @xmath12 just below the transition and obtained from the time series of @xmath12 after the system had reached the steady state .
indeed , a large part of the sample - to - sample variations can be interpreted as a shift in the @xmath41 of individual samples .
the two quantities @xmath61}$ ] and @xmath62 $ ] in eq .
( [ eq : gl_bm_old ] ) can be considered as the second and the fourth moments of the disorder - averaged pdf , which is much broader than the individual pdfs near the transition .
one can easily see that broadening yields a larger value for the ratio @xmath62/[\langle\delta^2\rangle]^2 $ ] and , hence , a smaller bc .
the effect is particularly pronounced on the small @xmath35 side of the transition , where @xmath12 itself is small , in which case a shift in @xmath41 has a stronger influence on the moments .
an alternative definition for the binder cumulant for systems with quenched disorder ( especially non - diminishing disorder fluctuations ) is@xcite @xmath63 .
\label{eq : gl_bm_new}\ ] ] we note that the disorder average is performed over the ratio of the time - averaged moments .
the moment ratio is calculated for each sample first and , is then averaged over disorder .
it is clear that this definition of the binder cumulant should eliminate the most dominant contribution from the disorder fluctuations , i.e. , the anomaly caused by the spreading of the effective coupling constants .
this definition has been adopted mostly in quantum disorder systems , where strong disorder fluctuations are anticipated @xcite .
figure [ fig : gl_bm_new ] displays @xmath64 versus @xmath35 .
we note that the dip shown in fig .
[ fig : gl_bm_old ] disappears and that the crossing points nicely converge to @xmath41 , implying that @xmath64 should serve better for locating the transition point than the conventional one , which is confirmed numerically ( not shown here ) .
yet another definition of the binder cumulant is @xmath65}{3[\langle \delta^2 \rangle^2]}. \label{eq : gl_bm_newer}\ ] ] we expect that @xmath66 may also behave smoothly near the transition because it does not involve disorder fluctuation terms such as @xmath45 ^ 2 $ ] included in @xmath46 .
figure [ fig : gl_bm_newer ] displays @xmath67 versus @xmath35 .
as expected , we find no anomalous behavior in @xmath66 .
we can directly relate @xmath46 and @xmath66 through the disorder fluctuation @xmath58 .
simple algebra leads to @xmath68 as the disorder fluctuation @xmath58 becomes larger , @xmath69 shows a bigger dip .
this explains quantitatively the size and the location of the dip in @xmath69 .
the critical value of @xmath70 ( @xmath71 ) provides additional information on the temporal variations of @xmath12 .
one can show that @xmath72/[\langle\delta^2\rangle^2]$ ] , where @xmath73 .
our numerical result indicates that the relative temporal fluctuations are almost negligible even at criticality . in this case
, @xmath70 is not practically useful in locating the transition point accurately . in summary , we studied binder cumulants in the quenched disorder system . for the kuramoto model
, we found that the conventionally defined bc shows a big anomalous dip near the transition .
this dip is shown to be directly related to the disorder fluctuation ( non - self - averageness ) .
alternative definitions of the bc , which did not show any anomalous behavior were proposed and may be useful in locating the transition point accurately in general systems with quenched disorder . |
the superior physical properties of graphene @xcite and its compatibility with two - dimensional lithographic processes have stimulated intense research of graphene - based electronics for `` beyond si - cmos '' technology @xcite .
one of the major challenges for realizing the graphene - based beyond si - cmos technology is the fabrication of high - quality large - area graphene sheets and the retention of superior electronic characteristics of graphene .
the apparent degradation of carrier mobility and significant variations in the electronic characteristics when graphene comes in contact with various dielectrics @xcite suggest significant susceptibility of the single - layer carbon atoms to the surrounding environment . on the other hand
, the strong susceptibility of graphene to external influences also provides opportunities for engineering unique properties of graphene .
for instance , it has been theoretically proposed that a designed strain aligned along three main crystallographic directions induces strong gauge fields that effectively act as a uniform magnetic field on the dirac electrons @xcite . in particular , for a finite doping level , the quantizing field results in an insulating bulk and a pair of counter - circulating edge states , which is similar to the case of topological insulators @xcite .
moreover , strained superlattices have been shown theoretically to open significant energy gaps in the electronic spectrum of graphene @xcite , which could be much more effective than current approaches to the bandgap engineering of graphene by either controlling the width of graphene nano - ribbons @xcite or applying magnetic fields to bi - layer graphene @xcite .
recent scanning tunneling microscopic and spectroscopic ( stm / sts ) studies of graphene nano - bubbles on pt(111 ) single crystal substrates @xcite have verified the theoretical prediction of strain - induced pseudo - magnetic fields , and significant field values in excess of 300 tesla have been observed . in the context of synthesis of large area graphene
, various growth techniques have been developed to date , including ultra - high - vacuum annealing to desorption of si from sic single crystal surfaces @xcite , deposition of graphene oxide films from a liquid suspension followed by chemical reduction @xcite , and chemical vapor deposition ( cvd ) on transition metals such as ru @xcite , ni @xcite , co @xcite , pt @xcite and cu @xcite . while samples as large as centimeter squared have been achieved , and characterizations of macroscopic physical properties such as raman scattering and mobility measurements have been reported @xcite , most large - area samples appear to exhibit thicknesses ranging from one to approximately twelve layers .
moreover , few investigations with the exception of refs .
@xcite have been made on correlating the local electronic properties with the microscopic structural variations of these large - area films . in this work
we report scanning tunneling microscopic and spectroscopic ( stm / sts ) investigations of the microscopic electronic and structural properties of cvd - grown graphene on copper and also on sio@xmath1 after transferred from copper .
we find that the cvd grown graphene is strongly strained so that both the lattice structure and the local electronic density of states ( dos ) of graphene are significantly affected @xcite .
the non - trivial strain results in significant pseudo - magnetic fields @xmath2 up to @xmath0 tesla , as manifested by the dos peaks at energies corresponding to quantized landau levels @xmath3 , where @xmath4 denotes both integers and fractional numbers at 0 , @xmath5 .
in addition , significant charging effects are found along the strongest strained areas of the sample , consistent with strain - induced scalar potential . on the other hand , for cvd - grown graphene transferred from copper to sio@xmath1 , both the pseudo - magnetic fields and charging effects diminish in most parts of the sample except for small regions containing topological ridges that can not be relaxed by changing the substrate .
the occurrence of conductance peaks at quantized integer and fractional landau levels in strained graphene is analogous to the magnetic field - induced integer and fractional quantum hall effects ( iqhe and fqhe ) , suggesting that the two - dimensional dirac electrons may be quantum confined into correlated many - body states by strain .
the strained - induced charging effects and quantized states also enable new possibilities of controlling the energy gaps and doping levels of graphene through nano - scale `` strain engineering '' .
our experimental approach to investigating the local electronic and structural correlations of graphene on dielectric or metallic substrates is to perform stm / sts studies using a home built cryogenic stm , which was compatible with high magnetic fields and also capable of variable temperature control from room temperature to 6 k , with a vacuum level of @xmath6 torr at the lowest temperatures . for studies reported in this work ,
the measurement conditions were at 77 k under high vacuum ( @xmath7 torr ) and in zero magnetic field .
the stm tips were made of pt / ir alloy and were produced by mechanical cleavage and electrochemical polishing .
the tips thus prepared were first characterized on graphite for atomic sharpness and then cleaned by gently touching the stm tips to an amorphous gold surface .
both topographic and spectroscopic measurements were performed simultaneously at every location in a ( @xmath8 ) pixel grid . at each pixel location , the tunnel junction was independently established so that the junction resistance of 1.5 g@xmath9 was maintained across the sample .
the differential conductance , ( @xmath10 ) , was calculated from the best polynomial fit of each tunneling current ( @xmath11 ) vs. bias voltage ( @xmath12 ) curve .
the data acquisition typically involved averaging ten @xmath11-vs.-@xmath12 spectra per pixel .
polynomial fitting to successive segments of the averaged @xmath11-vs.-@xmath12 curve at each pixel was made over a finite voltage range , and the @xmath13-vs.-@xmath12 spectrum per pixel was obtained by collecting the derivatives of the best fitted segments at all voltages .
the spectra thus derived are generally consistent with those obtained by superposing a modulated bias voltage with a lock - in amplifier on the @xmath11-vs.-@xmath12 measurements .
however , the modulated - bias technique generally requires much longer time to collect both the spectroscopic and topographical data pixel by pixel over extended areas with high spatial resolution , and so was not employed in this work .
the procedures for cvd growth of graphene on copper foils were similar to those described in literature @xcite .
briefly , graphene films were primarily grown on 25 @xmath14 m thick , highly polished polycrystalline copper foils in a furnace consisting of a fused silica tube heated in a split tube furnace .
the fused silica tube loaded with copper foils was first evacuated and then back filled with pure hydrogen gas , heated to @xmath15c , and maintained under partial hydrogen pressure with argon .
a flow of methane was subsequently introduced for a desired period of time at a total pressure of 500 mtorr .
finally , the furnace was cooled to room temperature , and the copper foils coated with graphene appeared brighter relative to the as - received copper foils , consistent with previous reports @xcite .
the graphene sheets thus prepared were found to be largely single- or double - layered based on raman spectroscopic studies . before the stm measurements , a graphene - coated copper foil was first cleaned and then loaded onto the stm probe .
the stm probe was then evacuated to high vacuum condition ( @xmath7 torr ) .
two different cvd - grown graphene samples on copper were studied at 77 k. both topographic and spectroscopic measurements on single - layer graphene regions were carried out on both samples from relatively large area surveys down to atomic - scale investigations . to transfer cvd - grown graphene samples to sio@xmath1 substrates ,
sio@xmath1 substrates were prepared by first thermally grown a 290 nm thick sio@xmath1 layer on the silicon wafer ( p - type @xmath16 ) , followed by gentle sonication of the substrate in acetone and then pure alcohol for about two minutes .
the substrate was then baked at 115@xmath17c on a hotplate , nitrogen blown dry while being cooled down .
for the cvd - grown graphene on copper , a layer of pmma was first deposited on top of the graphene sample as scaffolding , and then the copper substrate was removed with nitric acid .
next , the pmma / graphene sample was placed on a sio@xmath1 substrate with the graphene side down .
finally , the pmma was removed with acetone .
the transferred graphene was cleaned and annealed in an argon atmosphere at 400@xmath17c for 30 minutes .
finally , electrodes were created by thermally evaporating 2.5 nm chromium and 37.5 nm gold through an aluminum foil shadow mask .
prior to stm measurements , the sample was rinsed again gently with acetone followed by pure alcohol to remove possible organic surface contaminants .
the stm / sts studies on one of the transferred graphene samples on sio@xmath1 were all conducted on the single - layer region .
systematic studies of the cvd grown graphene on copper revealed large height variations everywhere , as exemplified by the topographic images figs .
1a - 1c and the corresponding height histograms in figs . 1d - 1f for one of the two samples .
the ripple - like height variations are consistent with previous reports @xcite and are primarily associated with two physical causes .
the first is the large difference in the thermal contraction coefficients of graphene and copper .
namely , upon cooling the sample from @xmath15c to room temperature after the cvd growth of graphene , the expanded graphene ( due to its negative thermal contraction coefficient ) can no longer remain flat on the thermally contracted copper foil , thus resulting in large ripples .
the second cause is the original surface roughness of the copper foils , as manifested in fig .
2 , where the surface morphology imaged by an atomic force microscope ( afm ) and the corresponding height histograms of a copper foil processed under the same conditions as those for the cvd - grown graphene on copper samples are shown is figs .
2a - 2d . comparing figs . 1 and 2 suggesting that the long - wavelength height variations ( @xmath18 ) of the cvd - grown graphene on copper may be associated with the surface morphology of the copper foil , and the averaged variation over a length scale @xmath19 is estimated at @xmath20 .
on the other hand , significant short - wavelength height variations , such as those exemplified in figs .
2e - 2f over a small @xmath21 nm@xmath22 area may be primarily attributed to the difference in the thermal contraction coefficients , where the height variations could reach as much as 1 nm over a @xmath21 nm@xmath22 area for both samples so that @xmath23 , which are much larger than those associated with the long - wavelength height variations and therefore suggest significant strain on graphene .
nm@xmath22 area , ( b ) a @xmath24 nm@xmath22 area and ( c ) a @xmath21 nm@xmath22 area . the white box in ( a ) represents the region that is enlarged into the image in ( b ) , and the white box in ( b ) represents the area that is enlarged into the image in ( c ) .
the corresponding height histograms for the images shown in ( a ) - ( c ) are given in ( d ) - ( f).,width=307 ] the large strain associated with the cvd - grown graphene - on - copper samples apparently incurred significant lattice distortion to the arrangements of carbon atoms . as exemplified in the atomically resolved topographic image in fig .
3a , different atomic arrangements were apparent in different regions of the @xmath21 nm@xmath22 area in view .
specifically , we found that in a more relaxed region of the graphene sample ( denoted as the @xmath25-region ) the atomic structure appeared to resemble the honeycomb or hexagon - like lattice with slight distortion . on the other hand ,
for spectra taken on the areas slightly below the ridge ( @xmath26-region ) of a ripple - like feature , the atomic arrangements deviated strongly from those of graphene , showing either nearly square - lattice or disordered atomic structure in the @xmath26-region .
additionally , we found certain areas that were completely disordered without atomic resolution , as exemplified in fig .
3c and denoted as the @xmath27-region .
the disordered region might be associated with amorphous carbon grown on defected features ( @xmath28 the bright features shown in the afm images in figs.2a and 2b ) on the copper substrate
. area .
( b ) an afm image of the surface morphology of the copper foil over a @xmath29 area .
( c ) the height histogram of the surface morphology along the red line in ( a ) .
( d ) the height histogram of the surface morphology along the red line in ( b ) . from ( a ) - ( d ) we find relatively long - wavelength height variations on the order of @xmath20 , where @xmath19 denotes the lateral length for total height variation @xmath18 .
( e ) the height variations along the dashed line in the stm topography in fig .
1c for a cvd - grown graphene on copper , showing atomic steps ( solid lines ) superposed on top of a gradual background ( dotted lines ) .
( f ) the height variations in ( e ) after removing the long - wavelength background , showing sharp steps of either @xmath30 nm or @xmath31 nm in height , comparable to one or two c - axis lattice constants of graphite . from ( e)-(f ) we find short - wavelength height variations on the order of @xmath23 , much larger than the long wavelength variations shown in ( a ) ( d).,width=307 ] to investigate the effect of lattice distortion on the local electronic dos of graphene , we compared the typical tunneling conductance ( @xmath10 ) vs. @xmath12 spectra obtained from different regions of atomic arrangements .
as shown in figs .
3b and 3d , the representative spectra for the more relaxed @xmath25-region of the graphene sample shown in fig .
1 resembled those found in mechanically exfoliated graphene except a substantially larger zero - bias conductance and additional conductance peaks . on the other hand
, the typical spectra associated with the strongly strained @xmath26-region appeared to be asymmetric relative to those in the @xmath25 region , with sharp conductance peaks more closely spaced .
additionally , for disordered @xmath27-regions without atomic resolution , nearly parabolic spectra that deviate fundamentally from the linear energy dependence of the dirac fermions were found , as exemplified in fig .
3d . for @xmath25- and
@xmath26-type spectra exemplified in fig .
3b , the energy interval between consecutive conductance peaks appear to decrease with increasing energy . on the other hand , for @xmath27-type spectra
, the conductance peaks seem randomly spaced .
similar results are confirmed in the second cvd - grown graphene - on - copper sample , as exemplified in figs .
3e - 3f , where the conductance peaks of the representative point spectra correspond to a pseudo - magnetic field @xmath32 tesla .
nm@xmath22 area , showing different atomic arrangements of the graphene sample in view , with slightly skewed honeycomb ( or hexagonal ) lattices in the @xmath25-region and a combination of nearly squared and distorted atomic arrangements in the @xmath26-region .
( b ) representative tunneling conductance @xmath10-vs.-@xmath12 point spectra in the @xmath25-region ( point 1 ) and @xmath26-region ( points 1 and 2 ) shown in ( a ) .
the spectra are slighted shifted relative to each other along the @xmath10 axis for clarify .
( c ) a representative topography of the same sample in ( a ) over a disordered @xmath21 nm@xmath22 area .
( d ) representative @xmath10-vs.-@xmath12 point spectra for the specific points in the @xmath27-region shown in ( c ) .
( e ) an atomically resolved topography of another cvd - grown graphene - on - copper sample over a @xmath33 nm@xmath22 area , showing distorted lattices .
( f ) representative @xmath10-vs.-@xmath12 point spectra of the sample shown in ( e ) .
the spectra for points 1 and 2 are shifted up slightly along the @xmath10 axis for clarity .
all data were taken at @xmath34 k.,width=316 ] to better understand the significance of the conductance peaks , we subtract off a parabolic conductance background associated with the copper contribution ( indicated by the dashed lines in figs .
3b and 3f ) and plot the resulting conductance against @xmath35 where @xmath36 is the dirac energy . as exemplified in figs .
4a and 4c for the point spectra after background subtraction , we find distinct peaks occurring at energies proportional to @xmath37 for @xmath38 and 6 .
furthermore , for stronger strained @xmath26-type spectra , sharp peaks at energies proportional to fractional numbers @xmath39 , @xmath40 and @xmath41 are clearly visible , in contrast to the weak `` humps '' occurring in the weaker strained @xmath25-type spectra at energies proportional to @xmath42 and @xmath43 . by plotting @xmath35 versus @xmath37 data for multiple spectra taken in figs .
4a and 4c , we find the @xmath35 data taken for each point spectrum in the atomically resolved strained regions all fall approximately on one linear curve , as exemplified in figs .
4b and 4d , and the slope for each @xmath35-vs-@xmath44 curve increases with increasing strain .
k : ( a ) the tunneling conductance @xmath13 spectra vs. @xmath35 after the subtraction of the parabolic conductance background shown in fig .
@xmath36 denotes the dirac energy .
distinct peaks at energies proportional to @xmath45 are clearly visible for @xmath46
. additional humps at fractional values @xmath47 , @xmath48 and @xmath49 are also identifiable .
( b ) @xmath50-vs-@xmath37 for the representative point spectra in ( a ) , showing that the data largely fall on linear curves that correspond to pseudo - magnetic fields ranging from @xmath51 tesla for the less strained @xmath25-region to @xmath52 tesla for the stronger strained @xmath26-region .
( c ) the @xmath13 spectra vs. @xmath35 after the subtraction of the parabolic conductance background shown in fig . 3f .
( d ) @xmath50-vs-@xmath37 for the representative point spectra in ( c ) , showing a pseudo - magnetic field @xmath32 tesla.,width=316 ] the results manifested in fig .
4 can be understood in terms of the presence of strain - induced pseudo - magnetic fields @xmath2 , with larger strain giving rise to larger @xmath2 values @xcite .
specifically , the pseudo - landau levels @xmath53 of dirac electrons under a given @xmath2 satisfy the following relation : @xmath54 where @xmath55 denotes the fermi velocity of graphene , @xmath56 , and the strain - induced gauge field @xmath57 is given by @xcite @xmath58 in eq .
( [ eq : gauge ] ) @xmath59 , where @xmath60 ev denotes the nearest - neighbor hopping constant , and @xmath61 is the nearest carbon - carbon separation .
additionally , the tensor components @xmath62 ( @xmath63 = @xmath64 ) are defined as @xmath65 , where @xmath66 is the @xmath67th component of the displacement field @xmath68 . using @xmath69 m / sec , we find that the pseudo - magnetic fields are approximately @xmath70 tesla for the @xmath25 region and @xmath52 tesla for the @xmath26 region .
these pseudo - magnetic fields are significant even though they are smaller than those found in graphene nano - bubbles @xcite , probably due to the spatial inhomogeneity of the pseudo - magnetic fields that leads to partial cancellations .
further , the magnetic lengths @xmath71 associated with the pseudo - magnetic fields are estimated to range from 3.5 to 5.5 nm according to the following formula : @xmath72 where @xmath73 is the flux quantum . )
nm@xmath22 area and for @xmath34 k : ( a ) @xmath74 , ( b ) @xmath75 , ( c ) @xmath76 , ( d ) @xmath77 , ( e ) @xmath78 , ( f ) @xmath79 , where the @xmath4 values are referenced to the strain - induced landau levels for the @xmath80 point spectrum . for smaller @xmath4 values ,
an approximately one - dimensional high - conductance `` line '' appears near the topographical ridge where the most abrupt changes in height occur , suggesting significant charging effects .
the confinement of high - conductance region diminishes with increasing @xmath4 .
( g ) spatial map of the dirac - point tunneling conductance for @xmath81 taken at spatially varying dirac energy @xmath82.,width=316 ] the strain found in the cvd - grown graphene on copper is not purely shear but also contains compression / dilation components .
the latter is theoretically predicted to gives rise to an effective scalar potential @xmath83 and therefore a static charging effect @xcite in addition to the aforementioned pseudo - magnetic field . here
@xmath83 is given by @xcite : @xmath84 where @xmath85 ev , and @xmath86 is the dilation / compression strain . for strain induced by corrugations along the out of plane direction ,
the typical strain is of order @xmath87 , where @xmath18 denotes the height fluctuations and @xmath19 is the size of the strained region .
although the charging effect associated with the scalar potential can be suppressed in a graphene layer suspended over a metal if @xmath88 @xcite , the mean value of @xmath18 over the @xmath89 nm@xmath22 area of the cvd - grown graphene on copper was @xmath90 nm , comparable to @xmath71 .
this quantitative consideration therefore suggests that significant charging effects may be observed . )
nm@xmath22 area of the cvd - grown graphene - on - copper sample shown in figure 5 , where the strain components are estimated by considering the spatial variations in the height @xmath91 : ( a ) @xmath92 , ( b ) @xmath93 , and ( c ) @xmath94.,width=316 ] to visualize the charging effect , we show in figs .
5a - 5f the conductance maps for one of the cvd - grown graphene - on - copper samples at different constant bias voltages over the same @xmath89 nm@xmath22 area .
the bias voltages are chosen so that they correspond to the pseudo - landau levels of the @xmath80 point spectrum for @xmath74 , 1 , 2 , 3 , 1/3 and 2/3 .
a nearly one - dimensional high - conductance region following closely to the most strained region immediately below the ridge is clearly visible in maps associated with smaller @xmath4 values , confirming the notion of strain - induced charging effects . for higher energies ( larger @xmath4 values )
, the high - conductance region becomes less confined , which is reasonable because of the higher confinement energies ( @xmath95 ) required for the dirac electrons .
we further note that the strain - induced charging effect also results in shifts in the dirac energy . as manifested by the spatial map in fig .
5 g for the tunneling conductance taken at the spatially varying dirac energy @xmath96 for the same graphene sample , the @xmath81 map for @xmath82 also exhibits excess conductance in strongly strained regions . to further manifest the compression / dilation strain - induced charging effect
, we show in figs .
6a - 6c spatial maps of effective compression / dilation strain components @xmath92 , @xmath93 and @xmath94 . comparing fig .
6c with the conductance maps in fig . 5 , particularly fig . 5 g
, we find apparent correlation between the magnitude of strain and the conductance at the dirac energy . for ( a ) a @xmath97 nm@xmath22 area and ( b ) a @xmath98 nm@xmath22 area .
both images were taken at 77 k. the corresponding height histograms for the images shown in ( a ) and ( b ) are given in ( c ) and ( d ) , respectively .
these height variations are much smaller than those found in fig .
1a - 1c.,width=316 ] next , we consider comparative studies of a cvd - grown graphene sample transferred from copper to sio@xmath1 . as
shown in figs .
7a - 7d , the height variations in the transferred graphene sample are much reduced relative to those shown in figs .
1a - 1c . the significant reduction in the average height variations of the ripples results in lattice structures and tunneling conductance spectra of the graphene sample similar to those found in mechanically exfoliated graphene on sio@xmath1 @xcite , as exemplified in figs .
8a and 8b .
moreover , the tunneling conductance spectra appear to be much smoother with no discernible conductance peaks in most of the transferred graphene - on - sio@xmath1 sample , as exemplified by the spectra in figs .
8b and 8e for the points marked in fig .
on the other hand , occasional tunneling spectra near strained regions of the transferred graphene still revealed conductance peaks at quantized energies , as exemplified by the point spectrum in figs .
8c and 8f .
these regions contain topological ridges that can not be relaxed by transferring the graphene sample from copper to sio@xmath1 substrate , so that the corresponding pseudo - magnetic field remains substantial , as exemplified in fig .
8d where @xmath99 tesla .
similarly , the reduced strain for the transferred graphene - on - sio@xmath1 sample leads to reduced charging effect , as manifested by the overall lower and more spatially homogeneous dirac - point conductance of the transferred graphene sample in fig .
9c relative to that of the cvd - grown graphene - on - copper sample in fig .
. further comparison of the effective lattice distortion @xmath100 between the transferred graphene and the graphene - on - copper in figs .
9b and 9d reveal that the compression / dilation strain is much reduced in the transferred sample .
for strongly strained graphene , the magnitude of the compression / dilation strain correlates well with the dirac - point conductance , as manifested by figs .
9a and 9b . on the other hand , for negligible compression / dilation strain as in the case of transferred graphene - on - sio@xmath1 , there is no longer direct correlation between the dirac - point conductance and the local strain , because the former may be largely dependent on the local charge impurities on substrates .
overall , our stm / sts studies of various graphene samples have manifested the strong influence of strain and substrates on the structural rippling effect and the electronic dos of graphene .
we have shown that significant differences between the thermal contraction coefficients of graphene and the transition - metal substrates for cvd growth graphene result in strong and non - uniform lattice distortion at the nano - scale , which provides a new ground for investigating the strain - induced charging and pseudo - magnetic field effects .
in particular , the strain - induced giant pseudo - magnetic fields enable studies of the integer and fractional quantum hall effects associated with the pseudo - spin ( @xmath101 , the valley ) degrees of freedom by exploiting the unique strength of the pseudo - spin - orbit coupling in graphene .
this realization is analogous to the situation of two - dimensional topological insulators where a gapped bulk state coexists with the counter - propagating states at the boundaries without breaking the time - reversal symmetry @xcite . while in topological insulators
the spin - orbit interaction is generally sizable ( @xmath102 ev ) , in graphene the spin - orbit coupling is very small so that the corresponding landau gaps are smaller than 1 @xmath14ev .
therefore , direct manifestation of the quantum spin hall effect in graphene can only be realized under extreme empirical conditions of ultra - low temperatures ( @xmath103 mk ) and ultra - high carrier mobilities ( @xmath104 cm@xmath22 v s@xmath106 ) . in contrast ,
the pseudo - magnetic field - induced quantum hall effect associated with the pseudo - spins is easily identifiable in strained graphene , as demonstrated by this work and by similar studies of graphene nano - bubbles on pt(111 ) @xcite .
k : ( a ) a ( @xmath107 ) nm@xmath22 spatial map of the dirac point ( @xmath108 ) tunneling conductance of the same cvd - grown graphene - on - copper sample as in figure 3a .
( b ) spatial map of the compression / dilation strain @xmath100 over the same sample area as in ( a ) .
the maps in ( a ) and ( b ) are clearly well correlated .
( c ) a spatial map of the tunneling conductance of the transferred graphene - on - sio@xmath1 sample taken at the dirac point @xmath36 over a ( @xmath107 ) nm@xmath22 area .
( d ) the spatial map of the effective compression / dilation strain @xmath100 over the same sample area as in ( c ) .
the effective strain in ( d ) is several orders of magnitude smaller than that in ( b ) .
hence , the lack of correlation between the maps in ( c ) and ( d ) implies that the charging effect in relaxed graphene is not directly due to strain and may be attributed to charge impurities on the substrate.,width=316 ] it is worth pointing out that our observation of pseudo - magnetic field - induced fractionally filled landau levels in graphene may be attributed to the complicated strain distribution at the nano - scale , which ensures both large pseudo - magnetic fields and ballistic behavior of the dirac electrons .
hence , strain - induced charge confinement in two dimensions ( similar to the external magnetic field - induced charge confinement ) gives rise to correlated many - body interactions .
moreover , the significantly more complicated strain distribution found in the cvd - grown graphene - on - copper differs from the controlled strain in the case of graphene nano - bubbles @xcite where the effective gauge potential leads to a relatively uniform magnetic field over each nano - bubble . therefore , further corrections to the hamiltonian of the dirac electrons are necessary in the case of cvd - grown graphene - on - copper , which may lead to additional terms in the effective gauge potential , similar to the presence of the chern - simons gauge potential terms in the case of fractional quantum hall ( fqh ) phenomena .
however , self - consistent theoretical studies will be necessary to account for the observation of fqh - like effects in the cvd - grown graphene - on - copper .
we further remark that the tunneling spectra acquired from highly strained cvd - grown graphene - on - copper samples contain not only spectral characteristics of the strained graphene but also excess conduction channels associated with the underlying copper substrate .
the excess conduction channels may be responsible for the commonly observed suppression of the zero - bias dos associated with the @xmath74 landau level .
however , the general effects of the underlying copper substrate on the graphene spectra are still unknown and await further theoretical investigation .
the findings reported in this work are highly relevant to the development of graphene - based nano - electronics .
in particular , proper `` strain engineering '' may be achieved by designing the substrate for graphene so that either local doping or energy - gap engineering associated with the pseudo - landau levels can be realized at the nano - scale .
ultimately , strain - induced charging effects and pseudo - magnetic fields may also be achieved at the macroscopic scale if quasi - periodic occurrences of these nano - scale effects can be constructed .
we have conducted spatially resolved topographic and spectroscopic studies of cvd - grown graphene - on - copper and transferred graphene - on - sio@xmath1 samples .
our investigation reveals the important influence of the substrate and strain on the carbon atomic arrangements and the electronic dos of graphene . for cvd - grown graphene remaining on the copper substrate ,
the monolayer carbon structures exhibit ripples and are strongly strained , with different regions exhibiting varying lattice structures and electronic density of states .
in particular , topographical ridges appear along the boundaries between different lattice structures , which also exhibit excess charging effects .
additionally , the large and non - uniform strain induces pseudo - magnetic fields up to @xmath0 tesla , as manifested by the tunneling conductance peaks at quantized energies that are associated with both integer and fractional pseudo - magnetic field - induced landau levels .
in contrast , for graphene transferred from copper to sio@xmath1 after cvd growth , the typical graphene structure is largely restored and the average strain on the whole is much reduced , so are the corresponding charging effects and pseudo - magnetic fields , with the exception of limited areas that are associated with topological defects and therefore the resulting strain can not be relaxed after the transfer of sample from copper to sio@xmath1 substrates .
our findings suggest feasible nano - scale strain engineering of the electronic states of cvd - grown graphene by proper design of the substrates and growth conditions .
k. v. emtsev , a. bostwick , k. horn , j. jobst , g. l. kellogg , l. ley , j. l. mcchesney , t. ohta , s. a. reshanov , j. rohrl , e. rotenberg , a. k. schmid , d. waldmann , h. b. weber , t. seyller nat .
( 2009 ) 203 . | atomically resolved imaging and spectroscopic characteristics of graphene grown by chemical vapor deposition ( cvd ) on copper are investigated by means of scanning tunneling microscopy and spectroscopy ( stm / sts ) . for cvd - grown graphene remaining on the copper substrate , the monolayer carbon structures exhibit ripples and appear strongly strained , with different regions exhibiting different lattice structures and electronic density of states ( dos ) .
in particular , ridges appear along the boundaries of different lattice structures , which exhibit excess charging effects . additionally , the large and non - uniform strain induces pseudo - magnetic field up to @xmath0 tesla , as manifested by the dos peaks at quantized energies that correspond to pseudo - magnetic field - induced integer and fractional landau levels .
in contrast , for graphene transferred from copper to sio@xmath1 substrates after the cvd growth , the average strain on the whole diminishes , so do the corresponding charging effects and pseudo - magnetic fields except for sample areas near topological defects .
these findings suggest feasible nano - scale `` strain engineering '' of the electronic states of graphene by proper design of the substrates and growth conditions .
scanning tunneling microscopy , scanning tunneling spectroscopies , chemical vapor deposition , quantum effects , surface structure , morphology , roughness , and topography , carbon |
in high mass x - ray binaries ( hmxbs ) a neutron star or black hole orbits a massive early - type star and accretes matter either via roche - lobe overflow or from the stellar wind which powers the x - ray emission ( for recent reviews see nagase 1989 , white et al .
1995 , bildsten et al .
one divides the class of hmxbs according to the stellar type of the mass donor star into supergiant x - ray binaries with luminosity class i - ii ob star and be / x - ray binaries with luminosity class iii - iv be star companions .
be / x - ray binaries form the larger sub - group of hmxbs .
balmer emission lines in the optical spectrum and a characteristic infrared excess are attributed to the presence of circum - stellar material , probably forming a disk in the equatorial plane of the be star .
be / x - ray binaries often show transient behaviour with two types of outbursts .
x - ray outbursts repeating with the orbital period are most likely associated with the passage of the neutron star through the circum - stellar disk in an eccentric orbit while giant outbursts , often lasting longer than a binary period , probably arise from an expansion of the disk .
currently about 100 hmxbs and candidates are known .
nearly one third were found in the magellanic clouds ( mcs ) from which the majority is located in the small magellanic cloud ( smc , coe 1999 ) .
most of the be / x - ray binaries in the smc were discovered in recent years by x - ray missions like asca , bepposax , rosat and rxte ( nagase 1999 ) . from 20
optically identified hmxbs in the smc only one is securely associated with a supergiant system ( the x - ray pulsar smcx1 ) and from 11 of the 19 be / x - ray binaries x - ray pulsations were detected .
five additional x - ray pulsars are yet to be identified , but are most likely also be systems .
the location of such a large number of hmxbs at a similar distance makes the smc ideally suited for statistical and in particular spatial distribution studies of the population of hmxbs in a galaxy as a whole .
recent surveys to look for h@xmath0 emission - line objects in the smc were performed by meyssonier & azzopardi ( 1993 , hereafter ma93 ) and murphy & bessell ( 1999 , mb99 ) .
the survey of ma93 mainly covers the main body and eastern wing of the smc and their catalogue lists 1898 emission - line stars .
the catalogue of mb99 covers nearly all the area where rosat pspc observations of the smc are available ( except the southern half of the most south - east observation ) but is less sensitive ( 372 objects , partially in common with ma93 ) .
a main goal of ma93 and mb99 was to identify planetary nebulae in the smc , however , the catalogues also contain be stars .
ma93 state that all three at the time of publication known b[e ] supergiants which were covered by the survey were detected .
very few be / x - ray binaries were known in the smc until 1993 and it was not noticed that the be stars proposed as optical counterparts for smcx-3 , 2e0050.17247 and 2e0051.17304 were listed in ma93 as emission - line stars ( lin 198 , azv 111 and azv 138 , respectively ) . a correlation of the larger sample of
be / x - ray binaries known today in the smc shows that most of them are found in the catalogues of ma93 and mb99 . in this paper
we use the identification of x - ray sources with emission - line stars to propose new very likely candidates for be / x - ray binaries in the smc ( sect . 2 ) .
x - ray source catalogues of the smc which we used for our correlations with the emission - line star catalogues were published by wang & wu ( 1992 ) based on einstein ipc observations ( seward & mitchel 1981 , inoue et al .
1983 , bruhweiler et al .
1987 ) and by haberl et al .
( 2000 , hfpk00 ) produced from rosat pspc data .
we also used a preliminary version of the rosat hri catalogue of sasaki et al .
( 2000b ) .
a cross - correlation of the 517 pspc x - ray sources in the smc region ( hfpk00 ) with the catalogue of h@xmath0 emission - line stars published by ma93 ( 1898 entries ) yielded 46 possible optical counterparts ( distance @xmath2 @xmath3 to account for systematic uncertainties in x - ray and optical positions , where r@xmath4 denotes the 90% statistical uncertainty of the x - ray position in arc seconds ) . an additional object correlating with a pspc source was found within the catalogue of candidate emission - line objects of mb99 . from this sample of 47 objects one
coincides with a known supernova remnant , three with supersoft sources and ten with optically identified be stars proposed as optical counterparts for the x - ray sources . extending the search by using additional smc x - ray sources such as from the rosat hri catalogue of sasaki et al .
( 2000b ) , the einstein ipc catalogue of wang & wu ( 1992 ) and x - ray pulsars discovered by instruments on asca , bepposax and rxte yields another three emission - line stars identified with known be stars .
our correlation results are summarized in table [ tab - ma93 ] which is sorted in right ascension .
columns 2 - 4 of table [ tab - ma93 ] give the source numbers in the x - ray catalogues for rosat pspc ( hfpk00 ) , rosat hri ( sasaki et al.2000b ) and einstein ipc ( wang & wu 1992 ) . for sources detected by rosat coordinates with statistical 90% error ( from hfpk00 when detected by the pspc or from sasaki et al .
2000b when detected by hri only ) are given in columns 5 - 7 .
for the group of ipc sources which were not detected by rosat , the ipc coordinates and a 40 error as published in ww92 is given .
the three digits in column 8 denote the number of pspc detections in the energy bands 0.1 - 0.4 kev , 0.5 - 0.9 kev and 0.9 - 2.0 kev .
with a few exceptions most of the sources were detected mainly in the higher energy bands which indicates a hard x - ray spectrum .
hfpk00 used count ratios in the different energy bands , the hardness ratios , for a spectral classification of the pspc sources . however , very hard sources without detection in the lower energy bands were not classified in hfpk00 because of large errors on the hardness ratios .
column 9 lists the maximum observed x - ray luminosity for be / x - ray binaries and candidates derived in this work .
the values are selected from literature or computed from rosat count rates using the conversion factor 1.67@xmath5 erg s@xmath6/cts s@xmath6 ( see kahabka & pietsch 1996 , hereafter kp96 ) , typical for x - ray binaries with hard spectrum at the distance of the smc .
hri count rates were converted to pspc rates using a multiplication factor of 3.0 ( sasaki et al .
luminosities derived from count rates are indicated with colon .
for all given x - ray luminosities we assume a distance of 65 kpc to the smc .
column 10 of table [ tab - ma93 ] lists the entry number of the nearest object in ma93 ( mb99 in one case ) and column 11 the ma93 classification type ( 2 = snr ; 5 = planetary nebula , pn ; 9 = late type star ) .
the distance between x - ray and optical position as listed in ma93 ( mb99 ) is listed in column 12 .
b , v and r magnitudes found in the literature for identified sources are given in columns 13 - 15 .
when available b and r obtained from the usno a2.0 catalogue are listed for the remaining sources . in the last column identifications are given together with references and new proposals are marked with ? behind the source type . from the total of 18 high mass x - ray binaries in the smc with known
be star as proposed counterpart 13 are found in the emission - line catalogues of ma93 and mb99 . for completeness the remaining five
be / x - ray binaries are also listed in table [ tab - ma93 ] .
the identification of most known be / x - ray binaries with stars in the emission - line catalogues of ma93 and mb99 suggests that un - identified x - ray sources with emission - line star counterparts are most likely also be / x - ray binary systems .
other objects like ssss and snrs can be recognized by their very soft x - ray spectrum ( in contrast to the be / x - ray binaries with hard spectrum ) or by their x - ray source extent , respectively . in the following section we summarize the 18 x - ray sources
optically identified as be / x - ray binary .
we then propose emission - line stars as likely be counterparts for the five un - identified pulsars and in addition for 25 hard x - ray sources .
three supersoft sources detected by rosat were identified with emission - line objects in the catalogue of ma93 .
two of them are associated with planetary nebulae ( pn ) while the remaining one is identified with a symbiotic star in the smc .
more detailed information on the individual sources and finding charts with x - ray error circles can be found in sect .
[ sect - notes].1 . for 19 x - ray sources in the smc
nearby be stars were optically identified and proposed as counterparts , suggesting a be / x - ray binary nature .
eighteen were covered by rosat observations and information on the individual sources is summarized in sect . [ sect - notes ] where also finding charts are found .
the heao source 1h0103762 was not observed with rosat ( see kp96 and references therein ) .
also we do not include the hmxb candidates rxj0106.27205 ( hughes & smith 1994 ) and exo0114.67361 in our summary . for rxj0106.27205 no optical spectrum from the suggested counterpart
is published yet , which would confirm its proposed be star nature .
for exo0114.67361 wang & wu ( 1992 ) propose the b0ia star azv488 as counterpart , however , azv477 , also a b0ia star is even closer to the x - ray position .
both candidates suggest a supergiant type of hmxb .
it is remarkable , that together with the only other known supergiant hmxb smcx1 , exo0114.67361 is located in the eastern wing of the smc .
fourteen of the identified be / x - ray binaries were detected by the rosat pspc and their x - ray properties can be found in hpfk99 .
axj0051722 , smcx3 and rxj0058.27231 were detected by the rosat hri and are listed in the catalogue of hri sources in the smc ( sasaki et al .
only 2e0051.17304 was not detected by rosat .
thirteen of the proposed be star counterparts are listed in the catalogues of ma93 and mb99 and only for five x - ray sources the be counterparts have no entry in ma93 and mb99 ( axj0049729 , smcx2 , rxj0032.97348 , rxj0058.27231 and xtej0111.2 - 7317 ) .
rxj0032.97348 was not covered by the ma93 survey .
five x - ray pulsars in the smc were reported for which no optical identifications are published up to day .
four of them were detected by rosat pspc and/or hri , yielding more accurate positions ( hfpk00 , sasaki et al .
2000b ) and for the fifth case , xtej0054720 , several rosat sources are found within the large rxte error circle . in or very close to the rosat error circles emission - line objects from ma93
are found and we propose these as optical counterparts .
literature , finding charts and other information on the x - ray binary pulsars is presented in sect .
[ sect - notes ] .
most of the be stars proposed as optical counterparts for x - ray sources in the smc , as summarized in the previous section , are found as emission - line objects in the catalogue of ma93 .
this strongly supports that the unknown emission - line objects within the error circles of the unidentified pulsars are also be star counterparts of the x - ray pulsars forming be / x - ray binaries . from the correlation of x - ray source and emission - line object
catalogues 34 hard x - ray sources were found with an h@xmath0 emission - line object as possible optical counterpart in the x - ray error circle ( see table [ tab - ma93 ] ) .
the 34 x - ray sources were investigated in detail to obtain more information which can help to identify the nature of the object .
finding charts and notes to the individual sources are compiled in sect .
[ sect - notes ] .
many sources were observed more than once by rosat and we looked for long - term time variability . in the case of the pspc we used the 0.9 2.0 kev band because of higher sensitivity for hard sources like be / x - ray binaries . to combine detections from the different instruments we convert hri to pspc count rates by multiplying with 3.0 and ipc to pspc count rates by multiplying with 1.1 ( appropriate for a 5 kev bremsstrahlung spectrum with 4.3@xmath7 @xmath8 absorption column density ) . given the uncertainties in the count rate conversions , variability is only treated as significant above a factor of 3 .
none of the sources was observed with sufficient counting statistics in order to perform a detailed temporal analysis on shorter time scales ( within an observation ) and to detect x - ray pulsations .
we discuss all un - identified x - ray sources with emission - line object in or close to the error circle in the following and indicate very promising candidates for be / x - ray binaries with
be / x ? " in the remark column of table [ tab - ma93 ] . to estimate the number of false identifications of x - ray sources with emission - line objects in ma93 we shifted the x - ray positions of the sources in an arbitrary direction and
cross - correlated again with the ma93 catalogue .
to get statistically more reliable results this was repeated with different distances between 1 10 arc minutes . for this purpose
we used the pspc catalogue which is most complete .
after application of our selection criteria for accepting be / x - ray binary candidates ( likelihood of existence for the x - ray source @xmath9 13 , no other identification , distance @xmath2 @xmath3 ) we find on average about seven expected chance coincidences between pspc sources and emission - line objects in ma93 .
we emphasize that these are mainly caused by the pspc sources with the largest position uncertainties .
the pspc sources in table [ tab - ma93 ] with the largest errors on the x - ray position and therefore most likely chance coincidences are 99 , 248 , 295 and 404 .
indeed three of them were rejected as be / x - ray binary candidates due to the presence of other likely counterparts .
similarly pspc sources 77 and 253 were disregarded .
other x - ray sources with large position errors in table [ tab - ma93 ] are the ipc sources which were not securely detected with rosat . also here
two were not regarded as be / x - ray binary candidates .
for the 25 new be / x - ray binary candidates we therefore estimate that about two to three may be misidentifications , most likely among those with position error r@xmath4 @xmath9 15 .
the 1st asca catalogue of x - ray sources in the smc was compiled by yokogawa ( 1999 ) .
the sources were classified according to their hardness ratios and be / x - ray binaries were detected as sources with hard x - ray spectrum in the asca energy band . from this classification yokogawa ( 1999 )
proposed eight new be / x - ray binary candidates ( binary pulsar candidates , bpc ) .
we correlated our list of be / x - ray binary candidates with the asca catalogue and find for five of the eight bpc a likely counterpart within 2 ( the maximum asca position uncertainty ) . in addition one out of the nine ( probably because of its low flux ) unclassified asca sources also correlates with a pspc be / x - ray binary candidate . in table
[ tab - asca ] the be / x - ray binary candidates with likely asca counterpart are summarized .
the first three columns show asca source number , classification and observed x - ray luminosity ( 0.7 10 kev , but corrected for the distance of 65 kpc used throughout this paper ) from yokogawa ( 1999 ) .
the next two columns contain pspc source number and distance between asca and pspc position and the last two columns give the same for hri sources .
the x - ray luminosities observed by asca are generally a factor of 1.2 2.4 higher than the rosat values ( table [ tab - ma93 ] ) which may only partly be explained by x - ray variability . finding a systematically higher luminosity with asca
is probably caused by the different sensitive energy bands and/or different intrinsic source spectrum .
in particular the pspc count rate to luminosity conversion is very sensitive to the assumed absorption . in the case of asca sources 36 ( pspc 279 ) and 7 ( pspc 468 )
the asca / rosat luminosity ratio is much higher than average ( 4.2 and 17 , respectively ) , indicating strong flux variability .
rlcrrrr & & + no & class & [ erg s@xmath6 ] & no & d [ ] & no & d [ ] + 2 & bpc & 3.5@xmath11 & 434 & 30.1 & & + 6 & bpc & 2.9@xmath11 & 511 & 39.8 & 28 & 27.8 + 7 & bpc & 4.6@xmath11 & 468 & 52.1 & & + 27 & bpc & 8.6@xmath11 & 159 & 15.6 & 95 & 19.0 + 28 & bpc & 2.6@xmath11 & 220 & 11.3 & 97 & 20.1 + 36 & un & 9.6@xmath12 & 279 & 56.5 & &
+ [ tab - asca ]
an optical identification campaign of a selected sample of hard rosat pspc sources from the catalogue of hfpk00 was started independently to the present work ( keller et al . in preparation ) . from three of the x - ray sources presented here ,
spectra were taken which in all cases revealed a be star nature of the proposed counterpart from the ma93 catalogue .
this confirms rxj0057.87207 ( pspc 136 , sect .
[ sect - notes].4.8 ) as be / x - ray binary and also the proposed counterpart for the pulsar axj0105722 ( pspc 163 , sect .
[ sect - notes].3.2 , see also filipovi et al .
2000a ) as be star .
rxj0051.97311 ( pspc 424 , sect .
[ sect - notes].2.8 ) was independently identified by schmidtke et al .
these results can be taken as further evidence that the emission - line objects we propose for counterparts of x - ray sources are indeed be stars .
the spatial distribution of the smc hmxbs including the new candidates from this work is shown in fig . [ fig - distrib ] .
nearly all new candidates are located along the main body of the smc where most of the optically identified be / x - ray binaries are concentrated .
only one new candidate is found in the eastern wing near the supergiant hmxb smcx1 , where already two other be / x - ray pulsars are known .
the distribution is not biased due to incomplete coverage , neither in the optical nor in x - rays and makes the strong concentration of be / x - ray binaries in certain areas of the smc more pronounced .
the x - ray luminosity distribution of be / x - ray binaries and candidates in the smc is compared to that of systems in the galaxy in fig .
[ fig - lx ] . to do this
we intensively searched the literature on galactic be / x - ray binary systems .
we derived 31 galactic sources with @xmath13 erg s@xmath6 which are summarized in table [ tab - begal ] ( which should be mostly complete ) . the luminosity estimates of galactic systems are often hampered by uncertain distances and different energy bands of the observing instrument .
this may cause luminosity uncertainties by a factor of @xmath1410 in some cases but should not change the overall distribution drastically .
the new candidates in the smc mainly raise the number of be / x - ray binaries with luminosities log ( ) @xmath2 35.5 ( 21 out of 24 are new candidates ) .
this can easily be explained by the high sensitivity of the rosat instruments which allowed to detect be / x - ray binaries in their low - state while most of the higher luminosity be / x - ray binaries were discovered during outburst .
x - ray luminosities derived from detectors sensitive at higher energies ( typically 0.5 10 kev ) might be up to a factor @xmath142 higher than those derived from rosat count rates ( see sect .
3 ) which would shift the low - luminosity end in fig . [ fig - lx ] by 0.3 dex to the right . however
, such a shift would not change the overall distribution significantly .
recently , in the galaxy several likely low - luminosity be / x - ray binaries were discovered by bepposax and asca ( 1saxj1324.4 - 6200 , angelini et al .
1998 ; axj1820.5 - 1434 , kinugasa et al . 1998 ; axj1749.2 - 2725 , torii et al . 1998 ; 1saxj1452.8 - 5949 , oosterbroek et al .
1999 ; axj1700062 - 4157 , torii et al .
rosat also contributed new low - luminosity systems ( rxj0440.9 + 4431 , rxj0812.4 - 3114 , rxj1037.5 - 5647 , rxj0146.9 + 6121 , motch et al .
however , the high absorption in the galactic plane makes the detection of low - luminosity x - ray sources in the rosat x - ray band and their optical identification difficult .
this might explain the smaller number of low - luminosity be / x - ray binaries discovered so far in the galaxy compared to the smc and might suggest that the luminosity distribution of be / x - ray binaries is very similar in the smc and our galaxy . in this case many more such systems are expected to be found in our galaxy which would significantly contribute to the hard x - ray galactic ridge emission ( warwick et al . 1985 ) .
various authors have suggested the existence of a population of low - luminosity systems which are usually persistent x - ray sources showing moderate outbursts and long pulse periods ( e.g. kinugasa et al .
1998 , mereghetti et al .
2000 ) , somewhat different to the high - luminosity systems with strong outbursts and shorter pulse periods .
the smc results suggest that the low - luminosity sources even dominate the be / x - ray binaries in number . from the fact that about one third of the already identified
be / x - ray binaries is not listed in current emission - line catalogues even more such systems are expected to be found in the rosat x - ray source catalogues of the smc . on the other hand some of them will be observed with higher maximum luminosity in future outbursts , but if they indeed form a class of low - luminosity be / x - ray binaries like xper , the outbursts are expected to be small changing the luminosity distribution immaterial . there is a large difference in the number of ob supergiant hmxbs between the galaxy and the smc . in the smc at most two such systems are identified ( smcx-1 and maybe exo0114.67361 ) resulting in an overall ratio of be to supergiant x - ray binaries of more than 20 . in the galaxy
this proportion is more of order 2 ( 12 supergiant systems in the galaxy are listed in the reviews of white et al .
1995 and bildsten et al .
it is remarkable that the smc supergiant hmxbs are all located in the eastern wing giving rise to a local be / supergiant hmxb ratio similar to that in the galaxy .
in contrast no supergiant hmxb is known in the smc main body making the difference more extreme .
one possible explanation is a different star formation history .
be / x - ray binaries evolve from binary star systems with typical total mass of @xmath1420 m@xmath15 within about 15 my ( van den heuvel 1983 ) while the more massive supergiant hmxbs are formed on shorter time scales .
the latter therefore would trace more recent epochs of star formation than the be / x - ray binaries .
the comparatively large number of be / x - ray binaries in the smc in this view suggests a burst of star formation about 15 my ago while relatively few massive early - type stars were born during the last few million years .
it is remarkable that the large magellanic cloud ( lmc ) may also have experienced a burst of star formation about 16 my ago as was derived from optical photometric surveys by harris et al .
lmc and smc resemble in the relative composition of their x - ray binary populations , both rich in hmxbs but very few old low - mass x - ray binaries ( cowley et al .
1999 ) , suggesting a common star formation history triggered by tidal interaction during close encounters of lmc , smc and milky way .
however , according to present day modeling the last encounter occurred @xmath140.2 gy ago ( gardiner & noguchi 1996 ) , too early for the formation of the be / x - ray binaries we see today in x - rays .
therefore , the event which caused the origin of the frequent smc ( and lmc ? ) be / x - ray binaries remains still unclear .
also the different numbers of hmxbs detected in lmc and smc relative to their total mass need to be explained .
in the following notes x - ray source numbers refer to the catalogues of wang & wu ( 1992 ) for einstein ipc , of hfpk00 for rosat pspc and of sasaki et al.(2000b ) for rosat hri . finding charts with x - ray error circles
are shown in figs .
[ fig - fcsss ] , [ fig - fcobex ] , [ fig - fcpbex ] , and [ fig - fcnbex ] , for ssss , already identified be / x - ray binaries , unidentified x - ray pulsars and new be / x - ray binary candidates , respectively .
the order within each group of sources follows table [ tab - ma93 ] and for faster access to the table entry the running number is given together with source name .
\7 ) rxj0048.47332 : the sss rxj0048.47332 was discovered by kahabka et al . (
1994 ) and identified as the symbiotic m0 star smc3 by morgan ( 1992 ) .
this star is listed in ma93 as object 218 and classified as late type star , consistent with the spectral type determined by morgan ( 1992 ) . the accurate hri position ( source 23 )
confirms the identification of the pspc source ( 512 ) .
\39 ) 1e0056.87154 : this sss was discovered in einstein data ( inoue et al .
1983 ) and was detected with rosat pspc ( source 47 ) and hri ( 79 ) .
it coincides in position with the smc planetary nebula n67 ( aller et al .
1987 ) which is listed as object 1083 in ma93 .
\41 ) rxj0059.67138 : this very soft source was discovered by hfpk00 ( pspc 51 ) and proposed as new sss due to its positional coincidence with the planetary nebula lin357 ( 1159 in ma93 ) in the smc .
\1 ) rxj0032.97348 : stevens et al .
( 1999 ) identified two be stars within the pspc error circle of rxj0032.97348 , discovered by kp96 as variable source with hard x - ray spectrum .
the a factor of @xmath145 smaller error radius obtained from a different pspc observation by hfpk00 ( source 567 ) , however , still contains both be stars which are very close to each other ( fig . [ fig - fcobex ] ) .
the two stars were not covered by the survey of ma93 .
\9 ) axj0049729 : yokogawa & koyama ( 1998a ) reported x - ray pulsations in asca data of this source .
kahabka & pietsch ( 1998 ) suggested the highly variable source rxj0049.17250 ( kp96 ) as counterpart .
stevens et al .
( 1999 ) identified two be stars , one only 3 from the x - ray position and one just outside the error circle given by kp96 .
the revised position of pspc source 351 in hfpk00 makes the more distant be star further unlikely as counterpart ( see fig . [ fig - fcobex ] ) .
none of the two be stars turns up in the list of emission - line objects of ma93 with the nearest entry ( 279 ) 58 away .
\13 ) axj0051733 : yokogawa & koyama ( 1998b ) discovered x - ray pulsations from this source in asca data .
the x - ray source was detected in einstein ipc , rosat pspc and hri archival data and the 18 year history shows flux variations by at least a factor of 10 ( imanishi et al .
cowley et al . identified already 1997 a be star as optical counterpart of the rosat hri source rxj0050.87316 ( hri 34 ) which is located within the asca error circle .
cook ( 1998 ) reported a 0.708 d period from this star using data from the macho collaboration .
the source was also detected by the pspc ( source 444 ) and coincides with object 387 in ma93 .
\16 ) axj0051722 : corbet et al .
( 1998b ) reported 91 s x - ray pulsations from asca observations of this pulsar which was originally confused with the nearby 46 s pulsar xtej0053 - 724 in xte data .
axj0051722 was not detected by the pspc .
an hri detection reduced the position uncertainty and stevens et al .
( 1999 ) identified a be star as likely optical counterpart .
the x - ray source is found as source 37 in the hri catalogue and the star is identical to the only emission - line object from ma93 ( 413 ) in the asca error circle .
lrrll name & pulse & d & & reference for + & period [ s ] & [ kpc ] & [ erg s@xmath6 ] & + 4u0115 + 63 & 3.6 & 4 & 3.0@xmath5 & tamura et al .
1992 + v0332 + 53 & 4.4 & 3 & 3.2@xmath5 & takeshima et al .
1994 + 2s1553 - 542 & 9.3 & 10 & 7.0@xmath16 & apparao 1994 + gs0834 - 430 & 12.3 & 5 & 1.1@xmath5 & wilson et al .
1997 + xtej1946 + 274 & 15.8 & 5 & 5.4@xmath16 & campana et al .
1999 + 2s1417 - 624 & 17.6 & 10 & 8.0@xmath16 & apparao 1994 + rxj0812.4 - 3114 & 31.9 & 9 & 1.1@xmath16 & reig & roche 1999a + exo2030 + 375 & 41.8 & 5 & 1.0@xmath17 & parmar et al .
1989 + gs2138 + 56 & 66.3 & 3.8 & 9.1@xmath11 & schulz et al .
1995 + groj1008 - 57 & 93.5 & 2 & 2.9@xmath11 & macomb et al .
1994 + gs1843 - 02 & 94.8 & 10 & 6.0@xmath16 & finger et al .
1999 + 4u0728 - 25 & 103 & 6 & 2.8@xmath11 & corbet et al .
1997 + a0535 + 26 & 105 & 2.4 & 2.0@xmath5 & apparao 1994 + axj1820.5 - 1434 & 152 & 4.7 & 9.0@xmath12 & kinugasa et al .
1998 + 1saxj1324.4 - 6200 & 171 & 10 & 9.8@xmath12 & angelini et al .
1998 + groj2058 + 42 & 198 & 7 & 2.0@xmath16 & wilson et al .
1998 + rxj0440.9 + 4431 & 203 & 3.2 & 3.0@xmath12 & reig & roche 1999b + axj1749.2 - 2725 & 220 & 8.5 & 2.6@xmath11 & torii et al .
1998 + gx304 - 1 & 272 & 2.4 & 1.0@xmath16 & apparao 1994 + 4u1145 - 619 & 292 & 0.5 & 7.4@xmath12 & apparao 1994 + 4u2206 + 54 & 392 & 2.5 & 2.5@xmath11 & saraswat & apparao 1992 + a1118 - 616 & 405 & 5 & 5.0@xmath16 & apparao 1994 + 1saxj1452.8 - 5949 & 437 & 9 & 8.7@xmath18 & oosterbroek et al .
1999 + axj1700062 - 4157 & 715 & 10 & 7.2@xmath12 & torii et al .
1999 + x persei & 836 & 0.83 & 1.9@xmath11 & haberl 1994 + rxj1037.5 - 5647 & 860 & 5.0 & 4.5@xmath11 & reig & roche 1999b + rxj0146.9 + 6121 & 1412 & 2.5 & 3.5@xmath11 & haberl et al .
1998 + a0114 + 65 & & 2.6 & 1.7@xmath12 & motch et al .
1997 + gamma cas & & 0.188 & 3.9@xmath12 & apparao 1994 + 1e0236.6 + 6100 & & 3.1 & 2.0@xmath12 & motch et al .
1997 + 1h0521 + 373 & & & 4.0@xmath18 & apparao 1994 + [ tab - begal ] \19 ) rxj0051.97311 : this x - ray source was detected by cowley et al .
( 1997 ) during rosat hri observations of einstein ipc source 25 and identified with a be star by schmidtke et al .
it is identical to pspc source 424 and hri 41 .
the be star is found as object 504 in ma93 .
\20 ) rxj0051.87231 : this source was reported as strongly x - ray variable by kp96 and is associated with the x - ray pulsar 2e0050.17247 ( israel et al .
1997 ) . observed x - ray luminosities range between 5@xmath12 erg s@xmath6 and 1.4@xmath16 erg s@xmath6 ( israel et al .
the star azv111 ( object 511 in ma93 ) was proposed as counterpart for 2e0050.17247 while israel et al .
( 1997 ) identified another h@xmath0 active star within their error circle of rxj0051.87231 which is larger than that of kp96 .
also the position error given for the corresponding pspc source 265 by hfpk00 is large .
the detection of the source in the pspc observation 600453p ( used by kp96 ) where the source was bright was rejected by the semi - automatic analysis of hfkp99 because the detection was close to the support structure of the detector entrance window . a careful analysis ( and using the latest processed data of 600453p ) of the photons of the source in the detector frame shows , however , that it moved nearly parallel to the window support structure and that it was not affected by it . in table
[ tab - ma93 ] therefore the parameters derived from this pspc observation are given .
they confirm the results of kp96 with small error circle ( see fig .
[ fig - fcobex ] ) . both azv111 and star 1 of israel et al .
( 1997 ) are outside this error circle which , however , contains a be star identified by stevens et al .
this star is found as object 506 in ma93 and is the most probable counterpart of rxj0051.87231 .
\22 ) smcx3 : this long - known x - ray source was not detected by the rosat pspc but is included in the hri catalogue as source 43 .
the be star counterpart ( e.g. crampton et al .
1978 ) corresponds to object 531 in ma93 .
\23 ) rxj0052.17319 : lamb et al .
( 1999 ) reported x - ray pulsations from the variable source rxj0052.17319 ( kp96 ) found in rosat hri and cgro batse data .
israel et al . ( 1999 ) identified a be star as likely optical counterpart .
it is found in ma93 as object 552 and was detected as x - ray source by ipc ( 29 ) , pspc ( 453 ) and hri ( 44 ) .
the strong x - ray variability by a factor of @xmath14200 between different hri observations ( kahabka 2000 ) strongly supports the identification as be / x - ray binary .
\24 ) 2e0051.17304 : for this source , listed as entry 31 in the einstein ipc source catalogue of wang & wu ( 1992 ) , the be star azv138 ( garmany & humphreys 1985 ) was proposed as optical counterpart .
azv138 corresponds to object 618 in ma93 .
2e0051.17304 was not detected in rosat observations .
\25 ) rxj0052.97158 : this source was detected as x - ray transient by cowley et al .
( 1997 ) during rosat hri observations of einstein ipc source 32 ( the largest circle in the finding chart of fig .
[ fig - fcobex ] .
upper limits derived from pspc observations imply flux variations by at least a factor of @xmath14350 ( cowley et al .
the strong variability and the hard x - ray spectrum imply a be / x - ray transient consistent with the suggested be star counterpart ( schmidtke et al .
the be star is identical to object 623 in ma93 .
the x - ray source was detected by rosat ( pspc 94 and hri 46 , the hri position is most accurate as indicated by the smallest error circle in the finding chart of fig .
[ fig - fcobex ] ) and is located near the edge of the error circle of xtej0054 - 720 . due to the large position uncertainty of xtej0054 - 720 it is , however , not clear if they are identical .
\28 ) smcx2 : the long known be / x - ray binary smcx2 was caught in outburst with 0.4 cts s@xmath6 by the rosat pspc ( source 547 , see kp96 and references therein ) .
another pspc observation yielded an upper limit indicating x - ray variability of more than a factor of 670 .
optical spectra of the be counterpart were taken by e.g. crampton et al .
the be star is located on the rim of the pspc error circle ( fig .
[ fig - fcobex ] ) and is not contained in the ma93 catalogue .
\31 ) xtej0055724 : x - ray pulsations from this source were discovered by rxte ( marshall & lochner 1998 ) and confirmed in a sax observation ( santangelo et al .
santangelo et al . ( 1998 ) also report pulsations from archival rosat data reducing the positional uncertainty .
stevens et al .
( 1999 ) identified a be star as optical counterpart which corresponds to object 810 in ma93 and which is inside the error circle of pspc source 241 and hri source 58 .
\38 ) rxj0058.27231 : rxj0058.27231 was detected as weak hri source by schmitdke et al .
( 1999 ) and identified with a be star .
it is contained in the hri catalogue ( source 76 ) but not found in the pspc catalogue of hfpk00 .
the be star is not detected in the emission - line star surveys of ma93 and mb99 .
\40 ) rxj0059.27138 : this transient x - ray pulsar with peculiar soft component in the x - ray spectrum was discovered by hughes ( 1994 ) during an outburst with a 0.2 2.0 kev luminosity of 3.5@xmath5 erg s@xmath6 .
the x - ray source was identified with a be star by southwell & charles ( 1996 ) as star 1 in their finding chart which is identical to the emission - line object 179 in mb99 .
\42 ) rxj0101.07206 : this source was suggested as x - ray transient by kp96 with a flux variability of at least a factor of 30 .
stevens et al .
( 1999 ) identified a be star as optical counterpart .
object 1 in their fig.1f corresponds to entry 1240 in ma93 .
\49 ) saxj0103.27209 : israel et al .
( 1998 ) reported x - ray pulsations from this source consistent in position with the einstein source 1e0101.57225 .
they confirm the be star suggested as counterpart for the einstein source by hughes & smith ( 1994 ) as the only object in the sax error circle showing strong h@xmath0 activity .
ogle observations presented by coe & orosz ( 2000 ) confirm this .
the be star corresponds to object 1367 in ma93 and was also detected by pspc ( source 143 ) and hri ( 101 ) .
\58 ) xtej0111.27317 : chakrabarty et al .
( 1998a ) reported x - ray pulsations found in rxte data from this source located about 30 from smcx1 .
wilson & finger ( 1998 ) confirmed the pulsations from cgro batse data and chakrabarty et al .
( 1998b ) derived an improved position from asca data .
two be stars were identified by israel et al .
( 1999 ) within or near the asca error circle of 30 .
the closer of the two was concluded as most likely counterpart of xtej0111.27317 by coe et al .
this be star has no counterpart in ma93 .
a week source with existence likelihood of 14.5 is found in the pspc catalogue ( 446 ) .
the large error circle of 61 overlaps with the asca one and includes the position of the be star .
there is an additional ma93 object ( 1731 ) within the rxte error circle and the second be star found by israel et al .
( 1999 ) is identical to object 1747 in ma93 but both are outside the asca and pspc confidence regions ( see fig .
[ fig - fcobex ] ) .
\59 ) rxj0117.67330 : similar to the previous source x - ray pulsations were discovered from the x - ray transient rxj0117.67330 ( clark et al .
1997 ) in rosat pspc and cgro batse data ( macomb et al .
1999 ) . between two pspc observations , about 8 months apart , the count rate diminished by a factor of 270 .
clark et al . (
1997 ) identified a be star counterpart which is identical to object 1845 in ma93 and also within the error circle of x - ray source 506 in the smc pspc catalogue .
\10 ) axj0049732 : axj0049732 was discovered as x - ray pulsar by imanishi et al .
filipovi et al .
( 2000b ) reported two hard x - ray point sources from the catalogue of hfpk00 as possible counterparts to the asca pulsar .
they suggest one of them ( pspc source 427 ) as the more likely counterpart due to its identification with an emission - line object in ma93 ( number 300 ) .
\25 ) xtej0054720 : the position of this x - ray pulsar could only be determined to an accuracy of 10 radius ( lochner et al .
there are at least five x - ray sources detected by the hri within that circle ( labeled 1 through 5 in fig .
[ fig - fcpbex ] which correspond to the catalogue sources 55 , 50 , 62 , 59 and 46 , respectively ) .
object 2 , 4 and 5 were also detected by the pspc ( 104 , 157 and 94 ) .
the southern of the three ( also detected by ipc , 36 ) is proposed as active galactic nucleus ( agn ) by hfpk00 and the northern ( pspc 94 , hri 46 , ipc 32 ? ) was identified as be / x - ray transient rxj0052.9 - 7158 ( see sect .
[ sect - notes].2.2 ) .
the be star counterpart of rxj0052.9 - 7158 coincides with object 623 in ma93 .
it is not clear if this be / x - ray binary is identical to the rxte pulsar .
a final identification requires the detection of pulsations from rxj0052.9 - 7158 .
\27 ) xtej0053724 : corbet et al .
( 1998a ) discovered this pulsar and report a rosat source within the error box .
the pulse period , originally confused with axj0051722 , was clarified by corbet et al .
( 1998b ) .
hfpk00 give source 242 as likely counterpart of xtej0053724 .
a single emission - line object from ma93 ( 717 ) is found inside the intersecting error circles of ipc source 34 and the pspc source .
\35 ) axj0058720 : x - ray pulsations from this source were discovered by yokogawa & koyama ( 1998b ) in asca observations .
the source was detected in archival einstein ipc , rosat pspc and hri data which span 18 years and showed flux variations by more than a factor of 100 ( tsujimoto et al .
this high variability already strongly suggests a be / x - ray binary .
a single emission - line object from ma93 ( 1036 ) is found within the pspc error circle ( source 114 ) which is also consistent with the hri position ( 73 ) .
it is not clear whether ipc source 41 originates from the same x - ray source .
it may also be associated with another emission - line object ( 1039 of ma93 ) closer to the ipc position or completely unrelated .
\53 ) axj0105722 : yokogawa & koyama ( 1998c ) reported axj0105722 as x - ray pulsar .
from several nearby objects in ma93 number 1517 is closest to the x - ray position of pspc source 163 .
this pspc source was identified as likely counterpart of the asca pulsar in an area of complex x - ray emission by filipovi et al .
( 2000a ) combining the rosat x - ray and radio data .
the star 1517 in ma93 is the northern and bluer component of a pair of stars close to the error circles of pspc and hri detection ( 110 ) .
the nearby ipc source 53 , 77 to the north - east is most likely associated to the snr dems128 ( filipovi et al .
2000a ) .
\2 ) rxj0041.27306 : hfpk00 classified pspc source 404 as foreground star based on the hardness ratios .
an emission - line object in the error circle is classified as planetary nebula by ma93 indicating a chance positional coincidence .
this makes the identification with the bright star just outside the error circle most likely .
\3 ) rxj0045.67313 : this source ( pspc 436 ) was detected once in the 0.9 2.0 kev band of the pspc .
an emission - line object in the error circle suggests an be / x - ray binary .
\5 ) rxj0047.37239 : the pspc error circle of rxj0047.37239 ( source 295 ) overlaps with that of ipc source 19 .
an emission - line object ( 168 in ma93 and classified as late type star ) and two radio sources from the catalogue of filipovi et al .
( 1998 ) are located in the x - ray confidence region . a point - like radio source as counterpart would favour an agn identification leaving the nature of rxj0047.37239 ambiguous .
\6 ) rxj0047.37312 : rxj0047.37312 ( pspc 434 ) is most likely identified with the emission - line star 172 in ma93 .
the fluxes derived from pspc detections show a factor of nine variations , supporting that the x - ray source is a be / x - ray binary .
rxj0047.37312 is probably identical to ipc source 18 , which showed an intensity within the range observed by the pspc .
it is also the likely counterpart of asca source 2 in yokogawa ( 1999 ; see sect .
3 ) , an x - ray binary candidate detected with similar intensity .
\8 ) rxj0048.57302 : the emission - line object 238 in ma93 is the brightest optical object in the error circle of rxj0048.57302 ( pspc 392 ) .
a be / x - ray binary is suggested .
\11 ) rxj0049.57331 : an hri detection ( source 28 ) with much improved x - ray position compared to the pspc ( source 511 ) confirms the identification with the emission - line object 302 in ma93 .
rxj0049.57331 is the probable counterpart of asca source 6 in yokogawa ( 1999 ; see sect .
3 ) further supporting the likely be / x - ray binary nature .
\12 ) rxj0049.77323 : this source ( pspc 468 ) was detected once in the 0.9 2.0 kev band of the pspc .
an emission - line object in the error circle suggests an be / x - ray binary .
rxj0049.77323 is also the likely counterpart of asca source 7 in yokogawa ( 1999 ) , classified as x - ray binary candidate ( see sect . 3 ) .
\14 ) rxj0050.77332 : rxj0050.77332 was only once detected by the pspc ( 514 ) and the emission - line object in the error circle suggests a be / x - ray binary identification .
\15 ) rxj0050.97310 : hri ( source 36 ) and pspc ( source 421 ) detections are consistent with the identification of rxj0050.97310 with the emission - line object 414 in ma93 , suggesting a be / x - ray binary .
\17 ) rxj0051.37250 : two close emission - line objects suggest rxj0051.37250 ( pspc 349 ) as be / x - ray binary , but make the identification ambiguous .
\18 ) rxj0051.87159 : the emission - line object 502 ( ma93 ) found in the error circle of rxj0051.87159 ( pspc 99 ) is classified as late type star in ma93 .
an active corona of this star may be producing the x - ray emission .
the large error circle contains , however , another bright object which could also be responsible for the x - rays .
the nature of rxj0051.87159 remains therefore unclear .
\21 ) ww 26 : two emission - line objects from ma93 are found near ipc source 26 ( hardness ratio 0.51 , ww92 ) .
object 521 is located inside the error circle while 487 can not be completely ruled out as counterpart .
no rosat detection could improve on the position .
a be / x - ray binary nature is suggested .
\26 ) rxj0053.47227 : a precise hri position ( source 48 at the rim of the error circle of pspc 246 ) with the emission - line star 667 ( ma93 ) as brightest object in the error circle makes rxj0053.47227 a likely be / x - ray binary .
\29 ) rxj0054.57228 : the uncertainty in the position of rxj0054.57228 ( pspc 248 ) is relatively large and six emission - line objects from ma93 are found as possible counterparts to the x - ray source .
it is therefore a likely be / x - ray binary but the optical counterpart remains ambiguous .
\30 ) rxj0054.97245 : precise rosat x - ray positions ( pspc 324 = hri 57 ) include an emission - line star ( 809 in ma93 ) with typical be star magnitudes as brightest object in the error circles .
a factor of five x - ray flux variability ( the source was bright during a hri observation ) strengthens the identification as be / x - ray binary .
\32 ) ww 38 = 2e0054.47237 : an emission - line object ( 904 in ma93 ) is found inside the error circle of ipc source 38 suggesting a be / x - ray binary .
the source was not detected by rosat .
\33 ) rxj0057.27233 : this weak pspc source ( 270 ) was marginally detected once in the hard 0.5 2.0 kev band with a likelihood of 10.4 . unlike all other hard sources in table [ tab - ma93 ] it was not detected in the 0.9 2.0 kev band and therefore is unlikely a be / x - ray binary .
\34 ) ww 40 = 2e0055.87229 : the error circle of ipc source 40 contains two emission - line objects from ma93 .
object number 1021 is identified as be star azv111 while 1016 , located further north , is of unknown type .
rosat detected an x - ray source inside the ipc error circle ( hri 71 and pspc 117 with consistent positions ) which , however , is located between the two emission - line objects .
the relation between the rosat and the einstein source and the emission - line objects is unclear .
ipc , hri and pspc count rates are consistent within a factor of two , which may indicate that they come from the same x - ray source .
however , the accurate rosat positions make an association with one of the nearby objects from ma93 unlikely .
\36 ) rxj0057.87207 : again small error circles from rosat hri ( source 74 ) and pspc ( source 136 ) observations make the identification of rxj0057.87207 with an emission - line star ( 1038 in ma93 ) very likely .
pspc detections with factor of eight different intensities and an hri detection during an x - ray bright state which increases the variability to a factor of about 37 , make a be / x - ray binary nature highly probable .
\37 ) rxj0057.97156 : be / x - ray binary candidate from positional coincidence of pspc source 87 with emission - line object 1044 in ma93 which shows typical optical brightness .
\43 ) rxj0101.37211 : pspc detections of this source ( pspc 159 = hri 95 ) indicate flux variations by at least a factor of 15 and the source was not detected in other observations ( upper limit a factor of 100 below the maximum count rate ) . this high variability and the presence of an emission - line star ( 1257 in ma93 ) in the small x - ray error circles
likely exclude any other explanation than a be / x - ray binary .
it also is the likely counterpart of asca source 27 in yokogawa ( 1999 ; see sect .
3 ) , an x - ray binary candidate .
\44 ) rxj0101.67204 : two accurate positions from hri ( source 96 ) and pspc ( source 121 ) observations suggest the identification of rxj0101.67204 with object 1277 in ma93 .
the factor of three variability supports a be / x - ray binary nature of rxj0101.67204 which is probably identical to the ipc source 46 in ww92 .
\45 ) rxj0101.87223 : rxj0101.87223 ( pspc 220 = hri 97 ) shows x - ray flux variations of a factor of three .
the emission - line star 1288 ( ma93 ) exhibits magnitudes typical for a be star in the smc and is located near the overlapping area of hri and pspc error circles .
we suggest rxj0101.87223 as be / x - ray binary as it is also the probable counterpart of asca source 28 in yokogawa ( 1999 ; see sect .
3 ) , an x - ray binary candidate .
\46 ) rxj0102.87157 : this weak pspc source ( 92 ) was only once marginally detected in the broad 0.1 2.4 kev band .
the low detection likelihood of 10.5 and the non - detection in the hard bands indicates that it may not be real , or is at least not a hard source .
a be / x - ray binary nature is therefore unlikely .
\47 ) ww 49 : the ipc source 49 ( ww92 give a hardness ratio of 0.21 ) contains a faint emission - line object ( 1357 in ma93 ) classified as planetary nebula .
the spectral hardness of the ipc source is inconsistent with an sss interpretation .
the positional coincidence is likely by chance .
\48 ) rxj0103.17151 : this source was detected only once by the pspc ( source 77 ) and the lowest upper limit indicates variability by at least a factor of five , suggesting the detection of a single outburst .
the emission - line object near the rim of the pspc error circle is , however , the optically weakest ( see table [ tab - ma93 ] ) , unusual in comparison with identified be / x - ray binaries and candidates derived from this work .
we therefore do not regard rxj0103.17151 as prime candidate for a be / x - ray binary .
\50 ) rxj0103.67201 : small error circles from hri ( source 105 ) and pspc ( source 106 ) observations make the identification with object 1393 in ma93 very likely .
rxj0103.67201 shows variability by a factor of three between the rosat observations , consistent with a be / x - ray binary .
51 ) rxj0104.17243 : two emission - line objects and a radio source from the catalogue of filipovi et al .
( 1998 ) close to rxj0104.17243 ( pspc 317 ) make the identification somewhat ambiguous .
the most likely identification with emission - line star 1440 in ma93 suggests rxj0104.17243 as be / x - ray binary .
\52 ) rxj0104.57121 : this source was not detected by the pspc but the accurate hri position ( source 108 ) includes only the emission - line object 1470 from ma93 as bright object in the error circle .
rxj0104.57121 is therefore very likely a be / x - ray binary .
\54 ) rxj0105.77226 : an emission - line star ( 1544 in ma93 ) in the pspc error circle ( 737 ) suggests rxj0105.77226 as be / x - ray binary .
\55 ) rxj0105.97203 : a single bright object ( the emission - line star 1557 in ma93 ) is found in the small pspc error circle ( source 120 ) , which makes the identification of rxj0105.97203 as be / x - ray very likely .
\56 ) rxj0107.17235 : the probable pspc detection ( 279 ) of ipc source 56 improves the x - ray position and allows to identify it with the emission - line star 1619 in ma93 .
the source was a factor of 10 brighter during the einstein observation and is also the likely counterpart of asca source 36 in yokogawa ( 1999 ; see sect .
3 ) detected with a factor @xmath144 higher intensity .
a be / x - ray binary nature is likely .
\57 ) rxj0109.07229 : the emission - line object 1682 in ma93 is classified as planetary nebula .
x - ray sources associated with planetary nebulae appear as ssss which is not compatible with the hard spectrum of rxj0109.07229 ( pspc 253 ) .
the positional coincidence of rxj0109.07229 is therefore by chance and the nature of the x - ray source is unclear .
\60 ) rxj0119.67330 : this source ( pspc 501 ) was detected once in the 0.9 2.0 kev band of the pspc .
an emission - line object in the error circle suggests a be / x - ray binary .
we reviewed the identification of eighteen known be / x - ray binaries in the smc and found that thirteen of them are listed in emission - line object catalogues of meyssonier & azzopardi ( 1993 ) and murphy & bessell ( 1999 ) . from a general correlation of smc x - ray source and h@xmath0
emission - line object catalogues we propose optical counterparts for the five optically unidentified x - ray pulsars and present 25 new be / x - ray binary candidates together with their likely optical counterparts .
this more then doubles the number of know high mass x - ray binary systems in the smc .
the rosat project is supported by the german bundesministerium fr bildung , wissenschaft , forschung und technologie ( bmbf / dlr ) and the max - planck - gesellschaft .
the finding charts are based on photographic data obtained using the uk schmidt telescope .
the uk schmidt telescope was operated by the royal observatory edinburgh , with funding from the uk science and engineering research council , until 1988 june , and thereafter by the anglo - australian observatory .
original plate material is copyright ( c ) the royal observatory edinburgh and the anglo - australian observatory .
the plates were processed into the present compressed digital form with their permission .
the digitized sky survey was produced at the space telescope science institute under us government grant nag w-2166 .
rrrrccrrcrcrlllp75mml 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 & 14 & 15 & 16 + & no & no & no & ra & dec & r@xmath4 & det & l@xmath19 & no & t & dist & b & v & r & remarks + & rp & rh & ei & & ( j2000.0 ) & [ ] & & erg s@xmath6 & ma & & [ ] & & & & + 1 & 567 & & & 00 32 56.1 & -73 48 19 & 12.9 & 012 & 1.3@xmath5 & & & & & & & be / x rxj0032.9 - 7348 , 2 be stars ( kp96 , scb99 ) + 2 & 404 & & & 00 41 16.4 & -73 06 41 & 35.6 & 100 & & 22 & 5 & 31.5 & 18.4 : & & 17.0 : & [ fg star ] gsc 9141.4223 ? + 3 & 436 & & & 00 45 37.9 & -73 13 54 & 29.4 & 001 & 1.2@xmath11 : & 114 & & 23.4 & 18.3 : & & 16.9 : & be / x ? + 4 & 413 & & 16 & 00 47 12.2 & -73 08 26 & 6.8 & 012 & & 165 & 2 & 4.5 & & & & snr 0045 - 73.4 ( rlg94 ) 13 cm + 5 & 295 & & 19 & 00 47 18.4 & -72 39 42 & 49.8 & 001 & & 168 & 9 & 21.1 & & & & agn ?
radio smc b0045 - 7255 ( fhw98 ) 13 cm + 6 & 434 & & 18 & 00 47 23.4 & -73 12 23 & 4.0 & 012 & 2.4@xmath11 : & 172 & & 1.6 & & & & be / x ? [ hard ] + 7 & 512 & 23 & & 00 48 23.1 & -73 31 43 & 24.1 & 620 & & 218 & 9 & 15.1 & & & & sss rxj0048.4 - 7332 ,
symbiotic star m0 ( kp96 , m92 ) + 8 & 392 & & & 00 48 33.7 & -73 02 24 & 7.6 & 011 & 1.2@xmath11 : & 238 & & 6.5 & 14.6 : & & 16.9 : & be / x ?
+ 9 & 351 & & & 00 49 02.5 & -72 50 52 & 13.7 & 002 & 6.1@xmath16 & & & & 17.01 & 16.92 & & be / x axj0049 - 729 , 74.67 s pulsar ( yk98a , kp98 , scb99 , co00 ) + 10 & 427 & & & 00 49 29.6 & -73 10 56 & 5.5 & 002 & 4.1@xmath11 & 300 & & 2.6 & 18.1 : & & 15.6 : & be / x ? axj0049 - 732 , 9.132 s pulsar ( iyk98 , fph00b ) + 11 & 511 & @xmath2028 & & 00 49 30.7 & -73 31 09 & 1.8 & 011 & 1.2@xmath11 : & 302 & & 22.2 & & & & be / x ? + 12 & 468 & & & 00 49 43.8 & -73 23 02 & 14.9 & 001 & 2.7@xmath12 : & 315 & & 14.7 & 17.7 : & & 15.4 : & be / x ? + 13 & 444 & 34 & & 00 50 44.3 & -73 15 58 & 4.9 & 023 & 1.8@xmath16 & 387 & & 6.9 & 15.41 & 15.44 & & be / x axj0051 - 733 , 323.2 s pulsar ( csm97 , yk98b , sc98 , co00 ) + 14 & 514 & & & 00 50 46.8 & -73 32 47 & 33.7 & 010 & 2.4@xmath12 : & 393 & & 9.2 & & & & be / x ? + 15 & 421 & 36 & & 00 50 56.5 & -73 10 09 & 4.5 & 013 & 1.1@xmath11 : & 414 & & 3.5 & & & & be / x ? + 16 & & 37 & & 00 50 56.9 & -72 13 31 & 1.4 & & 2.9@xmath5 & 413 & & 2.7 & 15.8 : & & 16.4 : & be / x axj0051 - 722 91.12 s pulsar ( cml98b , l98 , scb99 ) + 17 & 349 & & & 00 51 19.5 & -72 50 43 & 15.6 & 001 & 3.6@xmath12 : & 447 & & 14.6 & & & & be / x ? + 18 & 99 & & & 00 51 51.4 & -71 59 50 & 47.7 & 001 & & 502 & 9 & 22.3 & 14.6 : & & 13.3 : & ac ? + 19 & 424 & 41 & 25 & 00 51 51.5 & -73 10 31 & 2.2 & 055 & 4.7@xmath11 : & 504 & & 4.2 & 13.1 : & 14.4 & 14.4 : & be / x rxj0051.9 - 7311 ( csm97 , scc99 ) + 20 & @xmath21265 & &27 & 00 51 53.1 & -72 31 50 & 2.0 & 112 & 1.4@xmath16 & 506 & & 1.3 & 13.12 & 13.4 & & be / x rxj0051.8 - 7231 , 8.9 s pulsar ( isa97 , scb99 ) + 21 & & & 26 & 00 51 54.2 & -72 55 36 & 40.0 & & 6.0@xmath12 & 521 & & 28.6 & & & & be / x ? + 22 & & 43 & & 00 52 05.4 & -72 25 55 & 15.8 & & 5.6@xmath5 & 531 & & 7.1 & 15.0 & & 15.4 : & be / x smc x-3 + 23 & 453 & 44 & 29 & 00 52 13.9 & -73 19 13 & 1.9 & 011 & 1.3@xmath5 & 552 & & 5.1 & & 14.62 & 14.54 & be / x rxj0052.1 - 7319 , 15.3 s pulsar ( lpm99 , isc99 ) + 24 & & & 31 & 00 52 52.7 & -72 48 22 & 40.0 & & 1.6@xmath11 & 618 & & 7.8 & 14.28 & 14.28 & 15.6 : & be / x 2e0051.1 - 7304 , azv138 ( gh85 ) + 25 & 94 & @xmath2046 & 32 & 00 52 54.7 & -71 58 08 & 2.0 & 004 & 2.0@xmath5 & 623 & & 13.2 & 15.39 & 15.46 & 14.7 : & be / x
rxj0052.9 - 7158 ( csm97,scc99 ) = ? xtej0054 - 720 169.3 s pulsar + 26 & 246 & @xmath2048 & & 00 53 24.1 & -72 27 14 & 2.1 & 001 & 6.4@xmath12 : & 667 & & 31.0 & 17.0 : & & 15.8 : & be / x ? + 27 & 242 & & 34 & 00 53 53.3 & -72 27 01 & 25.7 & 012 & 7.4@xmath16 & 717 & & 22.7 & & & & be / x ? xtej0053 - 724 , 46.63 s pulsar ( cml98b ) + 28 & 547 & & & 00 54 30.8 & -73 40 55 & 10.5 & 011 & 8.4@xmath5 & & & & 15.7 & 16.0 & & be / x smcx-2 ( kp96 ) + 29 & 248 & & & 00 54 33.1 & -72 28 08 & 45.7 & 001 & 1.5@xmath16 : & 772 & & 26.6 & & & & be / x ? [ hard ] + 30 & 324 & 57 & & 00 54 55.3 & -72 45 06 & 5.6 & 012 & 3.0@xmath11 : & 809 & & 4.7 & 16.8 : & & 14.8 : & be / x ? + 31 & 241 & 58 & 35 & 00 54 55.4 & -72 26 46 & 3.5 & 023 & 3.0@xmath5 & 810 & & 3.1 & 15.24 & 15.28 & & be / x xtej0055 - 724 , 59.07 s pulsar ( ml98 , sci98 , scb99 , co00 ) + + @xmath22 b denotes 11 detections + @xmath23 entry in catalogue of mb99 + @xmath21 parameters from observation 600453 when source was bright + @xmath20 hri position and error + @xmath24 asca position and error [ tab - ma93 ] rrrrccrrcrcrlllp75mml 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10 & 11 & 12 & 13 & 14 & 15 & 16 + & no & no & no & ra & dec & r@xmath4 & det & l@xmath19 & no & t & dist & b & v & r & remarks + & rp & rh & ei & & ( j2000.0 ) & [ ] & & erg s@xmath6 & ma & & [ ] & & & & + 32 & & & 38 & 00 56 03.3 & -72 21 32 & 40.0 & & 4.0@xmath12 & 904 & & 29.6 & 15.7 : & & 14.3 : & be / x ? + 33 & 270 &
& & 00 57 15.4 & -72 33 38 & 25.9 & 000 & & 993 & & 27.1 & 17.2 : & & 16.7 : & ? + 34 & 117 ? & 71 ? & 40 & 00 57 32.4 & -72 13 17 & 40.0 & & & 1021 & & 23.6 & 15.4 : & & 13.7 : & ? + 35 & 114 & 73 & 41 & 00 57 48.4 & -72 02 42 & 7.9 & 001 & 1.6@xmath16 & 1036 & & 9.5 & 16.8 : & & 17.7 : & be / x
? axj0058 - 720 , 280.4 s pulsar ( yk98b ) + 36 & 136 & 74 & & 00 57 50.1 & -72 07 56 & 5.1 & 012 & 4.3@xmath11 : & 1038 & & 1.3 & 16.7 : & & 15.8 : & be / x ?
[ hard ] + 37 & 87 & & & 00 57 59.5 & -71 56 37 & 19.2 & 001 & 5.7@xmath12 : & 1044 & & 21.1 & 17.1 : & & 15.5 : & be / x ? + 38 & & 76 & & 00 58 12.9 & -72 30 45 & 3.1 & & 2.1@xmath11 : & & & & 16.7 : & 14.9 & 15.5 : & be / x rxj0058.2 - 7231 ( scc99 ) + 39 & 47 & 79 & 43 & 00 58 37.2 & -71 35 50 & 1.3 & @xmath22b30 & & 1083 & 5 & 1.8 & 16.8 : & & 15.3 : & sss 1e0056.8 - 7154 , pn lin333 ( kp96 ) + 40 & 53 & & & 00 59 11.3 & -71 38 45 & 2.8 & 111 & 5.0@xmath5 & @xmath23179 & & 10.1 & 14.21 & 14.08 & 14.03 &
be / x rxj0059.2 - 7138 , 2.763 s soft pulsar ( h94 , sc96 ) + 41 & 51 & & & 00 59 41.7 & -71 38 15 & 14.5 & 300 & & 1159 & 5 & 5.7 & 16.8 : & & 16.5 : & sss ?
rxj0059.6 - 7138 , pn lin357 ( kpfh99 ) + 42 & 132 & 93 & & 01 01 01.1 & -72 06 57 & 2.8 & 011 & 1.3@xmath16 & 1240 & & 8.5 & & & & be / x rxj0101.0 - 7206 ( kp96 , scb99 ) + 43 & 159 & @xmath2095 & & 01 01 20.5 & -72 11 18 & 1.6 & 013 & 6.6@xmath11 : & 1257 & & 4.7 & & & & be / x ?
[ nonstar ] + 44 & 121 & 96 & 46 & 01 01 37.0 & -72 04 19 & 7.2 & @xmath2206b & 3.8@xmath11 : & 1277 & & 5.5 & 17.7 : & & 16.4 : & be / x ?
+ 45 & 220 & 97 & & 01 01 51.0 & -72 23 26 & 8.1 & 025 & 2.2@xmath11 : & 1288 & & 9.8 & 14.7 : & & 14.3 : & be / x ? + 46 & 92 & & & 01 02 51.3 & -71 57 43 & 18.2 & 000 & & 1338 & & 14.1 & 15.1 : & & 13.6 : & ? + 47 & & & 49 & 01 03 06.9 & -72 32 59 & 40.0 & & & 1357 & 5 & 4.8 & & & & ? + 48 & 77 & & & 01 03 07.1 & -71 51 47 & 11.8 & 001 & & 1365 & & 14.6 & 18.8 : & & 17.8 : & ? + 49 & 143 & 101 & 50 & 01 03 14.0 & -72 09 16 & 3.4 & 067 & 1.5@xmath16 & 1367 & & 1.1 & 14.73 & 14.80 & 14.74 & be / x saxj0103.2 - 7209 , 345.2 s pulsar ( isc98 , hs94 , co00 ) + 50 & 106 & 105 & & 01 03 37.0 & -72 01 39 & 5.0 & 013 & 3.0@xmath11 : & 1393 & & 7.2 & & & & be / x ? + 51 & 317 & & & 01 04 07.4 & -72 43 59 & 17.7 & 002 & 3.8@xmath12 : & 1440 & & 9.0 & 14.1 : & & 14.4 : & be / x ? or agn ?
13 cm + 52 & & 108 & & 01 04 35.6 & -72 21 43 & 2.3 & & 4.8@xmath12 : & 1470 & & 4.0 & 14.8 : & & 15.1 : & be / x ?
+ 53 & 163 & 110 & & 01 05 08.9 & -72 11 44 & 6.6 & 012 & 1.5@xmath11 & 1517 & & 7.7 & & & & be / x ? axj0105 - 722 , 3.343 s pulsar ( yk98c , fhp00a ) + 54 & 737 & & & 01 05 42.3 & -72 26 15 & 15.7 & 001 & 1.8@xmath12 : & 1544 & & 12.8 & 14.2 : & & 14.0 : & be / x ?
+ 55 & 120 & & & 01 05 54.8 & -72 03 54 & 5.5 & 011 & 6.5@xmath12 : & 1557 & & 3.7 & & & & be / x ? + 56 & 279 & & 56 & 01 07 10.9 & -72 35 36 & 11.1 & 001 & 2.3@xmath12 : & 1619 & & 10.1 & 16.6 : & & 15.4 : & be / x ?
+ 57 & 253 & & & 01 09 01.2 & -72 29 07 & 17.6 & 012 & & 1682 & 5 & 12.3 & & & & ? [ hard ] + 58 & @xmath24446 & & & 01 11 14.5 & -73 16 50 & 30.0 & & 2.0@xmath17 & & & & 15.24 & 15.32 & 15.37 & be / x xtej0111.2 - 7317 , 31.03 s pulsar ( clc98a , wf98 , isc99 , chr99 ) + 59 & 506 & & & 01 17 41.4 & -73 30 49 & 0.6 & 122 & 1.2@xmath17 & 1845 & & 4.3 & 14.1 : & 14.2 & 14.7 : & be / x rxj0117.6 - 7330 , 22.07 s pulsar ( mfh99 , crw97 ) + 60 & 501 & & & 01 19 37.6 & -73 30 06 & 12.9 & 001 & 1.5@xmath12 : & 1867 & & 8.3 & 15.1 : & & 15.8 : & be / x ? + references : + ( chr99 ) @xcite 1999 , ( clc98a ) @xcite 1998a , ( cml98b ) @xcite 1998b , ( co00 ) @xcite 2000 , ( crw97 ) @xcite 1997 , ( csm97 ) @xcite 1997 , ( fhw98 ) @xcite 1998 , ( fhp00a ) @xcite 2000a , ( fph00b ) @xcite 2000b , ( gh85 ) @xcite 1985 , ( h94 ) @xcite 1994 , ( hs94 ) @xcite 1994 ,
( isa97 ) @xcite 1997 , ( isc98 ) @xcite 1998 , ( isc99 ) @xcite 1999 , ( iyk98 ) @xcite 1998 , ( kp96 ) @xcite 1996 , ( kp98 ) @xcite 1998 , ( kpfh99 ) @xcite 1999 , ( l98 ) @xcite 1998 , ( lpm99 ) @xcite 1999 , ( m92 ) @xcite 1992 , ( mfh99 ) @xcite 1999 , ( ml98 ) @xcite 1998 , ( rlg94 ) @xcite 1994 , ( sc96 ) @xcite 1996 , ( sc98 ) @xcite 1998 , ( scb99 ) @xcite 1999 , ( scc99 ) @xcite 1999 , ( sci98 ) @xcite 1998 , ( wf98 ) @xcite 1998 , ( yk98a ) @xcite 1998a , ( yk98b ) @xcite 1998b , ( yk98c ) @xcite 1998c | a correlation of x - ray source and h@xmath0 emission - line object catalogues in the small magellanic cloud ( smc ) shows that more than two thirds of the optically identified be stars in be / x - ray binaries are found as emission - line objects in the catalogues . on the basis of this result
we propose up to 25 x - ray sources mainly from recent rosat catalogues as new be / x - ray binaries and give their likely optical counterparts . also for the five yet unidentified x - ray pulsars in the smc we propose emission - line stars as counterparts .
this more than doubles the number of known high mass x - ray binary systems in this nearby galaxy .
the spatial distribution of the new candidates is similar to that of the already identified be / x - ray binaries with a strong concentration along the smc main body and some systems in the eastern wing .
the new candidates contribute mainly to the low - luminosity end of the x - ray luminosity distribution of be / x - ray binaries .
a comparison with the luminosity distribution in the milky way reveals no significant differences at the high - luminosity end and the large number of low - luminosity systems in the smc suggests that many such systems may still be undetected in the galaxy
. the overall ratio of known be to ob supergiant x - ray binaries in the smc is an order of magnitude larger than in the galaxy , however , might show spatial variations . while in the eastern wing the ratio is comparable to that in the galaxy no supergiant x - ray binary
is currently known in the main body of the smc .
possible explanations include a different star formation history over the last @xmath1 my . |
when coherent optical or microwave wave trains are interchanged between three or more freely orbiting spacecraft , and phase or frequency differences of all incoming and outgoing beams recorded , there is a closure that allows subsequent elimination of all source fluctuations ( `` laser noises '' ) and so achieves sensitivity to propagating gravitational waves at much lower levels set by secondary noises . in doing this
, the precise propagation times between sources and readout locations enter the data reduction algorithms , but equality of arm lengths is not necessary .
we have called this scheme `` time delay interferometry '' or tdi , @xcite and subsequently analyzed several configurations and data types needed for further elimination of the most serious secondary phase noise sources , _ viz . _ those due to spacecraft non - geodesic motion ( `` optical bench noise '' ) and especially to fluctuations of the ultra - stable local oscillators , or frequency standards , used in fringe tracking ( `` uso noise '' ) @xcite .
fringe tracking is necessary when the lasers in the configuration have widely differing frequencies , and/or when the configuration does not move rigidly so that emitted and received signals are doppler shifted during slow spacecraft separation changes . the lisa space gravitational wave antenna will use tdi , and is to be roughly equilateral , which maximizes the overall response to low frequency gravitational waves over a wide band . reduction of secondary noise , and simplicity of system architecture , nevertheless suggest consideration also of other schemes ; here we consider the most extreme possible application of tdi , _ viz . _ to a linear orbiting array , or , in astronomical terms , three spacecraft flying in syzygy .
we have previously considered a linear array , or parallel beam interferometric detector of gravitational waves , in work showing that with multiple bounces a split antiparallel laser beam will amplify a gravitational wave signal while laser fluctuations are cancelled conventionally by precise control of equal arm lengths @xcite .
we now see the multiple readout scheme of tdi as much preferable , and equal spacing of the three spacecraft is no longer required .
the equations for combining the heterodyne data streams in tdi can be taken directly from our previous papers , which had general triangular configuration ; in this paper they will be used in the linear limit .
the ( nearly ) linear formation of syzygy has some advantages and disadvantages over the nominal equilateral triangular formation planned for the lisa mission .
the primary disadvantages of the linear formation are less sensitivity at the low frequencies ( for a given spacecraft separation distance ) and sensitivity to only one linear combination of gw polarizations .
the advantages arise from the potential for a more stable formation geometry .
the nominal lisa triangular formation involves spacecraft in orbits with semi - major axis a , inclination i , and eccentricity e related by @xmath6 and @xmath7 @xcite .
this choice of orbital elements produces a rotating triangular configuration with arm lengths that are constant to @xmath8 .
the time variation of the arm lengths causes a variable doppler shift on the laser beam signals , a variable direction from each spacecraft to the other two spacecraft , and a variable angle between the spacecraft velocity and the direction to the other spacecraft .
the ( almost ) linear array might be formed with three spacecraft moving in a single circular orbit about the sun . in the ideal case ( ignoring perturbations from other bodies ) the distance between spacecraft and the directions from each spacecraft to the others will remain constant . the constant distance between spacecraft removes the doppler shift on the laser signals .
the constant directions means that no mechanism is needed to account for angle variation .
the constant angle also means that a single faceted ( e.g. cubical ) test mass can be used in each spacecraft , instead of the two in the nominal lisa configuration .
( in the case of spherical test masses , the constant angle between the velocity direction and direction to other spacecraft removes some surface irregularity errors that would affect a triangular formation with spherical test masses . ) the constant angle between the velocity direction and direction to other spacecraft means that the point - ahead and point - back angles are constant .
( these angles arise from the need to point lasers to hit the location the other spacecraft will be at after a one - way light time . )
the linear combination needs less spacecraft propulsion to reach the desired orbits ( for a given location relative to earth and arm length l ) since no plane change is needed .
no laser metrology is needed for the proof masses relative to the optical benches at the outlying spacecraft ( e.g. , data we have denoted @xmath9 , @xmath10 , at spacecraft 2 ) .
( at the center spacecraft ( spacecraft 1 ) the intra - spacecraft metrology data denoted @xmath11 ( or @xmath12 ) in the tdi equations must still be taken to eliminate phase fluctuations due to vibrations of the optical benches . ) the procedure for elimination of uso noise introduces considerable complication in the lisa system , inasmuch as additional calibrating data must be taken and exchanged if present generation flight - qualified uso s are used .
the uso noise enters the tdi combinations multiplied by the internally generated fringe tracking frequencies @xmath13 that are required when there are inter - spacecraft doppler drifts and/or offsets of the laser center frequencies .
control of relative spacecraft motions will automatically be achieved by spacing syzygy along a single nearly - circular solar orbit , and laser frequency offsets can be limited either by atomic line stabilization or by use of optical transponders @xcite .
thus no wide - band fringe tracking data , uso calibration , and associated modulation of the laser beams are needed by syzygy , as will be shown in the next section . in the linear case
the gravitational wave antenna patterns of all the tdi data combination become simple and identical ( so - called 6-pulse response ) .
we give this pattern explicitly in section 2 .
the pattern will of course sweep over the sky as the spacecraft move along their common orbit .
we apply the tdi data combinations in the linear geometry , and show how the only data combinations that are sensitive to gws have the same signal response as the combination x , the unequal arm michelson interferometer .
we introduce two new independent combinations that have no gravitational wave response , denoted @xmath14 and @xmath15 ; they can be used to monitor system noises .
all previously identified laser - noise - free data combinations are given in terms of x , @xmath14 and @xmath15 .
in section 3 we derive the syzygy response functions for the remaining secondary noises : optical bench vibration , proof mass residual accelerations , optical path noises at the photodetector readouts , and uso noise entering in to the heterodyne measurements . in section 4
we give numerical calculations and plots for these .
averaging over the sky and polarization states , we finally plot the resulting gravitational wave sensitivities to incident sine waves , using one year observation time and several syzygy scale sizes .
compared with an equilateral lisa , a syzygy of the same scale has similar best and high - frequency gravitational wave sensitivities
. however the sensitivity is poorer at low - frequencies , suggesting applicability in the frequency band between those of ground - based interferometers and lisa .
we use the notation and results of @xcite . a straight interferometric array , shown in figure 1 , is obtained by opening up the general triangular configuration of figure 2 of reference @xcite , so that spacecraft 1 is in the center and @xmath16 .
we denote the linear orientation by the unit vector @xmath17 = @xmath18 . on each of the spacecraft the usual pair of optical benches
can be fused into one , so we put @xmath19 . at the outlying spacecraft 2 and 3
both incoming beams will be referenced to the same face of a single proof mass . at the center spacecraft ,
number 1 , tdi will require use of two opposed faces of a single proof mass . for all we will have @xmath20 and the optical bench and proof mass motion components that enter are @xmath21 and @xmath22 , respectively . the intra - spacecraft data @xmath23 at the three proof masses consequently
satisfy @xmath24 , and @xmath25 .
the first four of these then drop out of all the laser - noise - free data combinations and so in fact need not be recorded .
the last two remain , and so one of them needs to be taken , which amounts to implementing precision laser metrology solely on the central spacecraft . when all these specializations are substituted into the data combinations derived for general configurations , it is found that they degenerate into essentially one , six pulse , response to incident gravitational waves ( we use x ) plus two others that respond only to system noises ( we call them @xmath14 and @xmath15 ) . all three will further be uso - noise - free to acceptable limits if the laser frequency offsets are limited by doppler effects due to relatively stable spacecraft motion in the syzygy linear orbital configuration .
specifically , the new laser - noise - free combinations we use for a linear array are @xmath26 @xmath27 @xmath28 where a semicolon subscripted index @xmath29 is understood to mean not only time delay by @xmath30 but also multiplication by the doppler factor @xmath31 , which in this case can be dropped from spacecraft 2 to spacecraft 3 differ from @xmath32 from spacecraft 3 to spacecraft 2 , etc .
one then has for syzygy @xmath33 and @xmath34 .
the primes on the time - delay subscripts in these equations for x , @xmath14 , and @xmath15 have been inserted for that case , but we do not insist on that generality in any of the subsequent discussion . ] . note how interchanging indices 2 and 3 takes @xmath14 and @xmath15 into one another .
the previously used tdi basic combinations in this linear array limit ( and without motion ) are given in terms of x , @xmath35 and @xmath15 : @xmath36 @xmath37 @xmath38 @xmath39 proof mass and shot noise response functions for x and @xmath14 ( or @xmath15 ) are given in section 4 .
the proof - mass - plus - optical - bench assembly and laser beam paths for the central spacecraft is show schematically in figure 2 .
the photodetectors that generate the time series @xmath40 , @xmath41 , and @xmath11 are as indicated .
the outgoing light from spacecraft 1 to spacecraft 2 is routed from laser 1 on the optical bench using mirrors and beam splitters ; this beam does not interact with spacecraft 1 s proof mass .
conversely , the incoming light beam from spacecraft 2 is first bounced off the nearest face of proof mass 1 before being reflected onto the photodetector , where it is mixed with light from laser 1 .
this time series is @xmath41 in figure 2 .
analogously , the time series @xmath40 is producted from light transmitted to and received from spacecraft 3 .
the relative motion of the optical bench and proof mass is monitored by measuring the doppler shift of the light coming directly from the laser with the light bounced off one face of the proof mass ( this forms the metrology time series @xmath42 , as indicated in figure 2 .
figure 3 shows the light paths on spacecraft 3 .
the outgoing light beam from spacecraft 3 to spacecraft 1 and 2 is routed from laser 3 on spacecraft 3 s optical bench using mirrors and beamsplitters ; again , this beam does not interact with spacecraft 3 s proof mass .
the two incoming light beams from spacecraft 1 and 2 are bounced off the proof mass before bing reflected onto photodetectors where they are mixed with light from laser 3 .
the time series produced are @xmath43 and @xmath44 , as shown .
the configuration at spacecraft 2 is analogous , producing @xmath45 and @xmath46 . @xmath44 and
@xmath45 enter @xmath47 , @xmath14 , and @xmath15 , but not x. the equations for the noises entering the @xmath48 and @xmath42 can be developed from figures 2 and 3 .
consider first the situation on spacecraft 1 .
the photodetector producing @xmath41 ( moving with velocity @xmath49 ) reads that time series by mixing the beam originating from the optical bench on spacecraft 2 sent in direction @xmath50 ( laser noise @xmath51 and optical bench motion @xmath52 , delayed by propagation along @xmath53 ) , after one bounce off the proof mass ( @xmath54 ) , with the local laser light ( @xmath55 ) .
the time series @xmath40 is produced analogously .
the @xmath11 measurement is from light originating at the laser ( @xmath56 ) , bounced off the left face of the proof mass ( @xmath54 ) , and mixed with the direct laser light ( @xmath55 ) . for spacecraft 2 and 3 , @xmath46 , and @xmath43 are produced similarly .
metrology time series on the end spacecraft are not required .
we will see below that there is no gravitational wave contribution to the data @xmath14 and @xmath15
. it may however be desirable to measure these to monitor system responses and secondary noises . in that case
we would also need the data @xmath45 and @xmath44 .
the @xmath48 and the required metrology time series @xmath11 (= @xmath12 ) , including gravitational wave signals and shot noises , can be developed from the diagrams and by consulting the appendix of tinto _ et al .
_ @xcite : @xmath57 @xmath58 @xmath59 where @xmath60 .
other @xmath48 are determined from equations ( 1 ) and ( 2 ) by cyclic index permutation . inserting these into the equations for @xmath47 and @xmath14
we can verify that the laser phase / frequency noises cancel and that the optical bench noises and uso noises are , for reasonable parameters of syzygy , suppressed to acceptable levels .
noting that @xmath61 and @xmath62 , and its permutations , we then obtain the following responses to proof - mass , uso , and optical path noises in x and @xmath14 or @xmath15 .
@xmath63 @xmath64 \nonumber \\ & & - ( \omega_{31}/\nu_0 ) ( q_3 - q_{1;2 } ) \nonumber \\ & & + y_{31}^{opt .
\ path } - y_{23}^{opt
. \ path } - \delta_2 y_{31;2}^{opt .
\ path } \label{eq:6}\end{aligned}\ ] ] under the assumptions that all noises are independent , that optical path noise only depends on the distance between spacecraft pairs , and that for proof mass noise in x terms with coefficients @xmath65 are negligible compared with terms having coefficients of order unity , the spectra of the noises in x and @xmath14 are : @xmath66 s^{proof \ mass } \nonumber \\ & & + [ 4 ( \omega_{32}^2 + \omega_{31}^2)/\nu_0 ^ 2 \ \sin^2(2 \pi
f l_2 ) + 4 ( \omega_{23}^2 + \omega_{21}^2)/\nu_0 ^ 2 \
\sin^2(2 \pi f l_3 ) \nonumber \\ & & -8 ( \omega_{21 } \omega_{31})/\nu_0 ^ 2 \
\cos(2 \pi f ( l_2 - l_3 ) ) \ \sin(2 \pi f
l_2 ) \ \sin(2 \pi f
l_3 ) ] s^{uso } \nonumber \\ & & + 8 \sin^2(2 \pi f l_3 ) s_{23}^{opt . \ path } + 8 \sin^2(2 \pi f l_2 ) s_{31}^{opt .
\ path } \label{eq:7}\end{aligned}\ ] ] @xmath67
the gravitational wave signal response in tdi combination x was given for a general triangle in equation ( 24 ) of reference @xcite .
the syzygy configuration leads to considerable simplification . taking the limit
as the general triangle is distorted to a line , using @xmath68 , and remembering that @xmath69 , we obtain a six pulse response with a single amplitude depending only on the orientation of the linear array with respect to the source direction : @xmath70 \label{eq:9}\end{aligned}\ ] ] where @xmath71 is the gravitational wavevector and @xmath72 is ( see @xcite ) @xmath73 for a transverse traceless gravitational wave @xmath74(t ) = @xmath75 $ ] .
note that @xmath76 is zero for waves incident at right angles to syzygy ( @xmath77 = 0 ) and for waves parallel or antiparallel with it ( @xmath77 = @xmath78 or @xmath79 ) .
although syzygy will not operate exclusively in the long - wavelength limit ( lwl - @xmath80 ) , analytical results at low - frequency are useful . expanding the gravitational wave @xmath81 in series of spatial derivatives and substituting into @xmath76
gives @xmath82 where @xmath83 .
figure 4 shows root - mean - square @xmath76 , averaged over polarization and the celestial sphere @xcite , for 3 armlength cases : ( a ) @xmath84 light - seconds , ( b ) @xmath85 light - seconds , and ( c ) @xmath86 light - seconds . notice that in the lwl @xmath87 , as expected from the lwl expansion above .
figure 5 shows the noise power spectra of x , using proof mass and optical path noises only , for the same armlength cases as figure 4 .
( uso can be made negligible compared to the other noises through trajectory control : if the maximum relative speed is kept below 0.2 m / sec either through trajectory design or by periodically using thrusters to adjust the maximum doppler shift
uso noise will not contribute . ) in constructing the curves in figure 5 we used the same instrumental spectra as in our previous sensitivity analysis of lisa .
the proof masses were assumed to have independent noises with one - sided acceleration spectral density @xmath88 m sec@xmath89 hz@xmath90 ( this corresponds @xcite to a one - sided spectral density of relative doppler @xmath91[f/1 hz]@xmath92hz@xmath93 . following previous practice @xcite , we approximately account for all optical path noise , including beam pointing noise ( see @xcite and reference @xcite , table 4.1 ) , on a given link with @xmath94 [ f/1 hz]@xmath95r@xmath96m@xmath97hz@xmath93 , where r is the distance between spacecraft on that link .
the aggregate noise spectrum plotted uses these spectra and the x transfer function ( see equations 4 and 6 ) .
the gw sensitivity is proportional to the ratio of the rms noise to the rms signal .
figure 6 shows the sensitivity of syzygy expressed in the conventional way : sensitivity to isotropic sinusoidal gravitational radiation ( snr = 5 in one - year integration ) , @xmath98^{1/2}$]/rms @xmath76(f ) , where b = 1 cycle / year , for the three armlength cases of figures 4 and 5 .
syzygy offers some technical simplifications compared with other configurations .
each spacecraft utilizes a single proof mass , and laser metrology is needed only at the central spacecraft . because of the low relative velocities afforded by the syzygy orbit , the complexity needed for uso noise elimination ( laser beam modulation and additional calibration data ) @xcite is not required by syzygy .
the angles between the arms are much more constant in time than for lisa , so time - dependent articulation of the telescopes is not required .
syzygy combinations @xmath14 and @xmath15 have exactly zero gw response and can thus be used with no ambiguity to assess on - orbit noise performance .
deliberate resetting of arm lengths ( i.e. , different arm lengths in different portions of a long mission ) is easier in a linear configuration and could be used to tune syzygy s best response over its mission lifetime . for similar scale size
, syzygy achieves both its best sensitivity and high - frequency performance comparable to those of lisa ; at low - frequencies , syzygy has poorer sensitivity .
this suggests applicability in the fourier band between those of lisa and ground - based interferometers .
little attention has been given to the @xmath99 hz band .
the work that has been done suggests it to be relatively free of foreground ( astrophysical ) gw sources @xcite . a syzygy with @xmath100 1 light - second could test this with @xmath4 sensitivity @xmath101 in a year s integration .
this research was performed at the jet propulsion laboratory , california institute of technology , under contract with the national aeronautics and space administration .
p. bender , k. danzmann , & the lisa study team , _ laser interferometer space antenna for the detection of gravitational waves , pre - phase a report _ , @xmath102 ( max - planck - institt fr quantenoptik , garching ) , july 1998 . fig .
1 . syzygy configuration .
unit vectors @xmath103 point between spacecraft pairs with the indicated orientations .
the arrows on the @xmath30 indicate the sense of the un - primed light times .
each spacecraft has one optical bench and one proof mass .
one - sided noise spectra of x as a function of fourier frequency .
one - sided spectra of proof mass noise and optical path noise are taken to be the the same as we have used in previous calculation of lisa sensitivity @xcite : respectively , @xmath106 m sec@xmath107hz@xmath90 and @xmath108r@xmath109m@xmath110 m hz@xmath90 , where r is the distance between spacecraft .
center frequencies of the lasers are assumed to be equal and a maximum doppler shift ( longest arm case ) of 0.2 m / sec is assumed .
solid lines : spectra of x including shot and proof mass noises . dashed lines :
contribution to x noise spectrum from an uncalibrated uso having one - sided spectrum of fractional frequency fluctuations @xmath111[f/1hz ] ( see text ) .
. gravitational wave sensitivity of x for syzygy ( snr = 5 in a one year integration , sky averaged , sinusoidal signals ) .
arm lengths , gravitational wave responses , and noise spectra are as in figures 4 - 5 . | we consider a spaceborne gravitational wave ( gw ) detector formed by three spacecraft in a linear array , coherently exchanging laser beams and using the data combinations of time - delay interferometry ( tdi ) .
we previously showed how the measured time series of doppler shifts in the six one - way laser links between spacecraft pairs in a general unequal - arm triangular configuration can be combined , using tdi , to exactly cancel the otherwise overwhelming phase noise of the lasers while retaining sensitivity to gws . here
we apply tdi , unfolding the general triangular configuration , to the special case of a linear array of three spacecraft .
we show that such an array ( `` syzygy '' ) has , compared with an equilateral triangle gw detector of the same scale , degraded ( but non - zero ) sensitivity at low - frequencies ( f @xmath0 c/(array size ) ) but similar peak and high - frequency sensitivities to gws .
we develop the gw and noise responses , noting that in this geometrical case only one tdi combination is gw - sensitive showing a relatively simple 6-pulse response to an incident gw burst .
sensitivity curves are presented for syzygys having various arm - lengths .
a number of technical simplifications result from the linear configuration .
these include : only one faceted ( e.g. cubical ) proof mass per spacecraft , intra - spacecraft laser metrology needed only at the central spacecraft , placement in an appropriate orbit can reduce doppler drifts so that no laser beam modulation is required for ultra - stable oscillator noise calibration , and little or no time - dependent articulation of the telescopes to maintain pointing .
because syzygy s sensitivity falls off more sharply at low - frequency than that of an equilateral triangular array , it may be most useful for gw observations in the band between those of ground - based interferometers ( @xmath1 hz ) and lisa ( @xmath2 hz ) . syzygy with @xmath3 1 light - second scale could , for the same instrumental assumptions as lisa , make observations in this intermediate frequency gw band with @xmath4 sensitivity to sinusoidal waves @xmath5 in a year s integration . |
water is a key element for life , but nevertheless its properties arising from its unique structure are not fully understood yet .
a large amount of published works explored the several phases of water and many efforts have been spent by the scientific community in order to match experimental findings with theoretical predictions . from the chemical and biological points of view , a deep understanding of the properties of acqueous systems and therefore of liquid water is fundamental .
over the past years , the structural and dynamical properties of liquid water have been investigated relying on the state - of - the - art molecular dynamics techniques such as empirical force field methods @xcite and _ ab initio _ molecular dynamics ( aimd ) .
the latter methodology has been extensively employed for the study of liquid water @xcite as well as acqueous solution of biological interest @xcite ; within , this technique , the atomic force fields required as input by any molecular dynamics simulation are constructed from _ ab initio _ potential energy surfaces ( pess ) and therefore directly derived from first principles electronic structure calculations .
working out an accurate pes for liquid water has been a longstanding challenge in the scientific community ; this is due mainly to the absence of an _ ab initio _ method able to develop a reliable pes that describes breaking and formation of hydrogen bonds ; other issues are related to difficulties in simulating weak interaction and polarization effects in the network of polar molecules .
the simulation of liquid systems requires model clusters with a large amount of molecules and therefore a good scalability is fundamental in order to carry out simulations in a reasonable computational time ; therefore the technique which has been mostly employed over the past years is density functional theory ( dft ) which combines scalability and flexibility features .
nevertheless , molecular dynamics simulations which start from dft - based pes fail in reproducing basic properties of liquid water such as the melting point temperature @xcite and the oxygen - oxygen radial distribution function @xmath4 @xcite .
another major problem which has not been elucidated yet by aimd is the description of proton transfer ( pt ) in liquid water , phenomenon which goes under the name of grotthuss mechanism @xcite .
this process is extremely important in a wide range of biological systems and it influences many dynamical processes in material science , biochemistry and bioenergetics ; for example , pt is one of the main mechanism for charge transport through the cell membranes @xcite ; it plays an important role in several enzymathic reactions and photosynthesis @xcite and moreover proton has the role of mediator and promoter of many acid - base reactions in solution @xcite .
although over the past years many experimental and theoretical works have been devoted to the study of grotthuss reaction , an accurate quantitative explanation of the underlined mechanism is not fully reached yet .
theoretical investigations still fail in pt basic properties predictions , such as the activation barrier .
experimental measures of this quantity have been carried out @xcite and followed by many attempts of its numerical estimation @xcite , without a satisfactory accordance .
nevertheless , over the past years aimd simulations have given many insights on the physics of grotthuss mechanism @xcite .
it has been found that proton transport is abnormally high and it is not driven by any ordinary diffusion process .
the picture which is instead commonly accepted describes the translocation of excess positive charge along the hydrogen - bonded network as structural diffusion of the defects .
has very recently shed new lights on proton transfer and has highlighted the complex nature of the process which involves both structural and dynamical changings of the hydrogen - bonded network .
one of the main issues which sets back from a complete understanding of pt in water , is related to the very sensitive thermal behavior of pt .
the required pes precision of the order of few tenths of kcal / mol has been reached only recently by the state - of - the - art computational methods beyond dft , such as coupled cluster ( cc ) , multi - reference configuration interaction ( mrci ) and the full configuration interaction quantum monte carlo ( fciqmc ) methods ; on the other hand , they are characterized by a much poorer scalability with respect to dft and therefore they do not allow simulation of sufficiently large molecular clusters . over the past years , two main structural models has been proposed to explain proton hydration : m. eigen @xcite considered as core of pt the complex @xmath5 in which an hydronium @xmath6 is strongly bounded to three water molecules ; g. zundel @xcite proposed a different reaction which involves the simpler @xmath0 complex ( the `` zundel ion '' or `` protonated water dimer '' ) as core of the transfer of proton between two h@xmath7omolecules .
it has been demonstrated @xcite that these two models occur only as limiting or ideal structures for a more complex phenomenon which involves a large portion of the hydrogen - bonded network .
nevertheless they constitute a perfect testing ground for the most advanced simulation techniques .
in particular , the protonated water dimer @xmath0 is the smallest system in which an excess proton is shared between water molecules ; due to its simplicity , many studies have been carried out to elucidate its properties .
over the past years , several accurate theoretical works have appeared on the zundel ion to study its structure and energetics@xcite ; the fast development of spectroscopical instruments allowed to probe experimentally vibrational properties of ionic species and therefore many studies on this side have been published @xcite on @xmath0.also molecular dynamics simulations which include quantum effects via feynman path integrals have been performed ( see ref . as a relevant example ) and recently an accurate pes has been produced @xcite with state - of - the - art coupled cluster technique including single , double and perturbative triple excitations ( ccsd(t ) ) . in the present paper
, we report on an extensive study of the protonated water dimer using a quantum monte carlo ( qmc ) approach based on a highly correlated variational wave function for energy and geometry calculations .
this approach has a very good scalability with the number of particles with respect to other post - dft methods .
we demonstrate that it also insures the precision of a few tenths of kcal / mol required for an accurate study of the system .
these appealing targets are achieved by our jastrow correlated antisymmetrized geminal power wave function ( jagp ) , developed on an efficient contracted single particle basis set . while the jagp provides the necessary ingredients to treat both dynamical and static correlations , the contracted basis set , based on atomic hybrid orbitals , reduces dramatically the computational cost to optimize the jagp wave function .
thanks to these promising features , we hope that it will pave the way to a complete understanding of the underlined physics of pt and other properties of liquid water . the paper is organized as follows . in sec
[ comp_methods ] we introduce the computational techniques used in this work .
in particular , in subsec .
[ qmc_wf ] we describe the variational ansatz and basis set which are the optimal compromise between accuracy and computational cost for water clusters . in sec .
[ qmc_water_sec ] we present our results on the single h@xmath7o molecule , the subject of a vast amount of literature .
it represents thus the most suitable system for a first benchmark of our approach . for the single h@xmath7o
, we study the basis set and wave function convergence and show that we have an accuracy comparable to the latest qmc studies .
[ water_dimer ] is devoted to the main results of our paper on the protonated water dimer .
we performed structural calculations which aim at finding the ground state geometry of @xmath0 at the jagp level of theory . in order to work out a zero - temperature potential energy landscape , we chose the distance between the two oxygens as a natural reaction coordinate .
along this path , we investigated how the symmetric `` zundel '' configuration - namely the one with the proton evenly shared between the two oxygens - evolves to a symmetry - broken geometry , where the excess proton is more bounded to one of the two molecules .
we believe that this crossover is critical to understand the physics of pt in larger water clusters . here
we show how our approach describes this relevant region of the @xmath0 pes .
the evaluation of proton transfer static barriers at fixed @xmath8 distances shows that our qmc approach has a global accuracy close to the state - of - the - art ccsd(t ) calculations performed by huang et al .
@xcite , with a discrepancy of 0.2 kcal / mol obtained by diffusion monte carlo ( dmc ) on the variational monte carlo ( vmc ) geometries . in conclusion
( sec . [ conclusions ] ) , we show that the favorable scalability of our method allows one to simulate larger protonated water clusters in a reasonable computational cost without loosing the accuracy reached in the smaller zundel ion . for the sake of test
, we performed a structural optimization on a more realistic pt model , i.e. the cluster of 6 water molecules and one excess proton , to study the behavior of the computational cost with the system size in practice .
in this section we present how we performed dft - based calculations on the zundel ion .
they are carried out to obtain optimized starting geometries and wave functions for further quantum monte carlo calculations .
the dft geometry has been obtained by a car - parrinello dynamics performed in a plane waves basis set with the quantum espresso@xcite suite of codes .
the technical details of this calculation are reported in subsec .
[ car_parri_dft ] .
once the ground - state dft geometry is found , the kohn - sham ( ks ) orbitals , used as starting input of our qmc wave function , are determined by the dft code of the turborvb package @xcite which works in a localized gaussian basis set , as described in subsec .
[ gaussian_dft ] . with the aim at finding the optimal dft geometry for further qmc structural relaxation
( see sec . [ qmc_methods ] ) , we use the perdew - burke - ernzerhof @xcite ( pbe ) version of the generalized gradient approximation for the exchange - correlation functional .
it has been proven to be quite reliable in describing properties of liquid water@xcite .
core electrons are taken into account using a norm conserving trouiller - martins pseudopotential@xcite .
the kohn - sham orbitals are expanded over plane waves up to a cutoff of @xmath9 h for the wave function and @xmath10 h for the charge density in a periodic box of 30 bohr radii ( @xmath11 ) per side .
the zero - temperature geometry relaxation is performed with damped dynamics within the car - parrinello ( cp ) approach @xcite .
this method allows a very small force convergence threshold of @xmath12 h@xmath13 on each atomic component at the end of the relaxation .
the potential energy surface of the dimer is very flat @xcite and it develops on energy differences of less than 1 kcal / mol .
therefore such an accuracy on atomic forces is essential for the minimization to provide a reliable dft description of the system .
we perform gaussian dft calculations with both pseudopotentials and full electron - ion potentials .
the single - particle kohn - sham ( ks ) orbitals @xmath14 are used later on as optimized starting guess for the qmc wave function , as discussed in sec .
[ qmc_wave_function_ansatz ] . the one - body dft wave function
is expanded over a primitive basis of _ gaussian type orbitals _ ( gto s ) centered on the atom @xmath15 , defined up to a normalization constant as : @xmath16 with @xmath17 $ ] and @xmath18 $ ] , where @xmath19 are angular momentum quantum numbers , @xmath20 identifies the number of gaussians for each angular momentum shell , @xmath21 are the spherical harmonics , and @xmath22 and @xmath23 are the electron and ion positions , respectively .
our dft basis set is _ even tempered _ , namely for each angular momentum @xmath24 the exponents of the gaussians
are expressed as a power series : @xmath25 for @xmath26 .
this allows a reduction of the number of parameters , as for each shell @xmath24 the full @xmath27 series is fixed just by three values : @xmath28 , @xmath29 , and @xmath20 . denoting the set of gto quantum numbers as @xmath30 , the dft orbitals @xmath31 expanded over the primitive basis set
read : @xmath32 where @xmath33 identifies the number of atoms of the system , @xmath34 is the size of the gto set of the atom @xmath15 . thus the global dft wave function can be expressed as the determinant of the matrix @xmath35 : @xmath36 with @xmath37 the spatial and spin coordinates of the @xmath38 electrons in the system , and @xmath39 are the spin orbitals .
the ab - initio hamiltonian of a quantum system is characterized by the divergence of the coulomb potential at the electron - electron and electron - ion coalescence points . when the potential energy becomes infinite
, the wave function of the system must have a cusp to cure the corresponding singularity of the hamiltonian@xcite .
these singularities represent a serious issue when an approximated wave function ansatz is employed to solve the schrdinger equation of a quantum system . within a dft framework ,
the divergence created by electron - electron coalescence is intrinsically solved by the mean - field description of the system , which maps the interacting hamiltonian onto a one with independent electrons .
nevertheless , the electron - ion coalescence still represents a problem .
one way to cure the latter divergence is to employ a smooth pseudopotential .
the pseudo - interaction is chosen such that it eliminates the need of electron - ion cusps in the wave function , by leading to a faster basis set convergence . in this work
we employ an alternative way to solve this issue .
we introduce an additional factor to the dft wave function ansatz which automatically satisfies the electron - ion cusp conditions and cancels out the corresponding divergences of the hamiltonian .
we rewrite eq .
[ global_dft_wf ] as : @xmath40 where the function @xmath41 is called one - body jastrow factor , borrowed from the qmc notation .
it is defined by means of a simple function @xmath42 which contains only one variational parameter @xmath43 : @xmath44 the one - body jastrow factor @xmath41 then reads : @xmath45 where @xmath46 are the atomic coordinates corresponding to the atomic number @xmath47 . in the function @xmath48 , the multiplicative factor @xmath49 is set by the electron - ion cusp conditions , while the scaling length @xmath50 deals with the large distance behavior set by the random phase approximation@xcite .
[ corrected_dft_wf ] has the typical functional form of the jastrow - single determinant ( jsd ) variational ansatz for quantum monte carlo calculations .
indeed the one - body jastrow of eq .
[ jas_dft ] has been introduced in ref . to provide a solution of the electron - ion cusp conditions within a qmc framework . differently from qmc ,
our @xmath51 presents only the one - body jastrow factor since in dft the many - body electron - electron interaction is solved within a mean - field description . by letting the wave function to fulfill the electron - ion cusp conditions ,
we are able to perform calculations considering the bare interaction for both oxygen and hydrogen atoms ( full potential calculations ) . furthermore , at variance with conventional dft calculations based on gaussian localized orbitals , in our approach all the overlap matrix elements involved are computed as numerical integrals over a uniform mesh .
such integrals , due to the presence of the jastrow factor , converge very rapidly with the number of points in the mesh and thus can be evaluated with a reasonable computational effort.@xcite with the aim at reducing the computational cost , calculations are carried out also by replacing the 1s core electrons of oxygen with the pseudopotential approximation , while the hydrogen atoms are treated always with the full coulomb potential .
the oxygen pseudopotential that we used in both dft and qmc calculations is the burkatzki - filippi - dolg ( bfd ) pseudopotential , introduced in ref . .
it is built by means of an energy - consistent approach in order to reproduce the valence all - electron excitation energies for a number of different atomic configurations computed at a scalar relativistic hartree - fock ( hf ) level .
gaussian dft calculations are performed with a local density approximation ( lda ) for the exchange - correlation functional .
the optimal basis sets which approach the dft complete basis set limit are @xmath52 and @xmath53 for the pseudopotential and full potential cases , respectively .
expanding the wave function over these basis sets , the variational parameters ( @xmath28 , @xmath29 , @xmath20 for each angular momentum @xmath24 and the one - body jastrow parameter @xmath43 ) are optimized at the dft level of theory , by minimizing the total lda - dft energy .
quantum monte carlo ( qmc ) refers to several numerical techniques for electronic structure calculations of quantum systems .
the interest of the scientific community for these methods has remarkably grown over the last three decades since they were successfully applied to highly correlated electronic systems , and from small- to medium - size quantum chemistry systems .
the qmc techniques used in this work are wave function - based , hence they aim at finding the many - body wave function in a representation as close as possible to the true ground state of the system .
the analytic form of the wave function has to be square - integrable and computable in a finite amount of time , but it does not present any further restrictions .
this property allows us to introduce correlations in a compact and efficient way
. another appealing feature of qmc methods is the scaling with the number of particles .
indipendently of the technique , qmc provides an intrinsic scaling of @xmath54-@xmath3 , with @xmath38 the number of electrons , whereas state - of - the - art coupled cluster single and double ( ccsd ) is @xmath55 and coupled cluster single , double and triple ( ccsd(t ) ) methods are @xmath56 . therefore qmc approaches allow the treatment of larger molecular clusters with respect to other highly - correlated methods .
major efforts have been done to reduce the multiplicative scaling prefactor of qmc techniques , i.e. the cost of a single - point calculation .
the research has followed two main paths . from one side
more compact trial wave functions have been developed .
one of the most promising functions is the jastrow - antisymmetrized geminal power ( jagp ) used in this work .
a description of the wave function ansatz is given in subec .
[ qmc_wave_function_ansatz ] .
a second way to decrease the scaling prefactor relies on improving the quantum monte carlo estimators .
qmc total energy calculations are in general much more efficient than those of other observables , due to the _ zero - variance _ property of the energy estimator . generally speaking ,
estimators of other important observables , as the charge density , do not possess this property .
some other bare estimators , as the one for the atomic forces , could have infinite variance .
thus they affect the overall computational efficiency of the qmc calculation .
in particular , over the recent years a major improvement has been achieved in the development of an efficient estimator for nuclear forces@xcite . in the present work , we use a version of this estimator based on the space - warp coordinate transformation@xcite and implemented with an exact infinitesimal differentiation method@xcite .
this allows a single sample calculation of the ionic forces with a computational effort of the order of @xmath2 ; this scaling is comparable to the one for total energy .
moreover , the infinite variance of the space - warped force estimator in the proximity of the nodal surface of the wave function is solved as explained in refs . .
the latter reference provides also a nice review on the latest progress in the qmc nuclear forces evaluation .
therefore , by means of this state - of - the - art scheme force estimator we are able to perform efficient structural optimizations at the qmc level of theory also for not - so - light atoms as oxygen .
the typical quantum monte carlo wave function is made of a symmetric bosonic factor ( _ jastrow factor _ applied to an antisymmetric fermionic part ( _ determinantal part _ ) : @xmath57 where the set @xmath58 represents spatial and spin coordinates of the electrons , as in eq .
[ global_dft_wf ] .
the jastrow factor is a function of the electron - electron separation .
it has been proven to be a crucial ingredient in order to well reproduce correlations of the true many - body wave function ; it contributes mainly in the description of dynamical correlation effects arising from charge fluctuations .
furthermore , it has been shown @xcite that the jastrow factor is particularly suitable in the treatment of van der waals intermolecular forces .
they play an important role also in the physics of the protonated water dimer , therefore an efficient parametrization of the jastrow is essential .
we split this factor in one - body , two - body and three / four - body terms ( @xmath59 ) .
the one - body factor accounts for electron - ion interactions and it has been already introduced in eqs .
[ jas_u],[jas_dft ] .
the two - body term deals with the electron - electron interactions .
in complete analogy with the one - body jastrow , it is parametrized by a simple function of the electron - electron separation : @xmath60 then it reads : @xmath61 @xmath41 and @xmath62 satisfy the kato cusp conditions@xcite , therefore correcting the divergence of the coulomb potential energy at electron - ion and electron - electron coalescence points , respectively .
the function @xmath48 rapidly decays to a constant value as the electron - ion and electron - electron distances increase ; thus the large distance behavior of correlations is described by the three / four - body jastrow term @xmath63 .
it deals with electron - electron - ion ( if @xmath64 ) and electron - ion - electron - ion ( in the case @xmath65 ) correlations effects and it represents an essential part of our wave function ; therefore we parametrize it in a richer way than the other jastrow terms : @xmath66 where @xmath67 represents the number of gto s of the primitive jastrow basis for each atom .
the jastrow uncontracted orbitals @xmath68 have the same form as the primitive gto basis set @xmath69 for the determinantal part in eq .
[ gaussians ] .
the jastrow primitive gto basis used in this work is @xmath703s2p1d[h]2s1p$ ] .
the choice of the fermionic part of the wave function is more delicate . in the case of spin unpolarized systems ( @xmath71 ) of @xmath38 electrons
, we can express it in a general way as an antisymmetrized product of the _ geminals _ or pairing functions @xmath72 of the system : @xmath73 \nonumber \\ & = & { \rm pf}(\phi(\mathbf{x}_i,\mathbf{x}_j ) ) .
\label{global_wf } \end{aligned}\ ] ] the geminals are antisymmetric functions of two electron coordinates expressed as a product of a spatial symmetric part and a spin singlet : @xmath74 the simplest , but compuationally most expensive expansion of the geminal is over the _ uncontracted atomic orbitals _ , and it reads : @xmath75 where the orbitals @xmath76 have the form reported in eq . [ gaussians ] with @xmath77 .
@xmath78 represents the number of primitive gto s per atom .
the introduction of the many - body jastrow factor @xmath79 in the total qmc wave function ( eq . [ total_qmc_wf ] ) allows a reduction of the primitive basis set size without loss of accuracy .
this reduction has several advantages in the qmc framework . at
first it obviosly decreases the computational effort for the wave function calculation .
furthermore it also guarantees a more robust energy minimization .
indeed , given the fact that qmc energy derivatives are noisy , using a more compact primitive basis set reduces its redundancy and helps in finding more quickly the global minimum since the number of the effective directions in the hilbert space is smaller . the size - reduced primitive basis employed for the determinantal part is @xmath80 and @xmath81 for pseudopotential and all - electron calculations , respectively .
the number of variational parameters expanded over gto s atomic orbitals is @xmath82 , where @xmath83 is the total number of gto s in the whole system . in this work
we use two distinct functional forms for @xmath84 .
they are distinguished by the number of non zero eigenvalues -
i.e. the rank - of the matrix @xmath85 in eq .
[ atomic_expansion ] .
the first one is commonly referred as single determinant ( sd ) and it is recovered when the @xmath85 matrix has the lowest possible rank compatible with the number of electrons in the system , namely when its rank is equal to @xmath86 . it can be shown that it is equivalent to the slater determinant ansatz for hf calculations . in this case the general expansion ( [ atomic_expansion ] ) , after diagonalization of the matrix @xmath85 , is written as : @xmath87 where @xmath88 are _ molecular orbitals _ ( mo ) . in this work
we obtain them starting from the optimized kohn - sham orbitals of gaussian dft calculations discussed in sec .
[ gaussian_dft ] .
the number of variational parameters in this case is @xmath89 where @xmath83 is the number of linear coefficients for each mo , equal to the total number of gto s .
close to the cbs limit @xmath90 , and so in the sd there is a significant reduction of the number of parameters with respect to the fully uncontracted agp expansion in eq .
[ global_wf ] . multiplying
the sd by the jastrow factor one obtains the jsd wave function , which is optimized simultaneously in both the j and sd parts .
the sd represents a particular limit of the function in eq .
[ global_wf ] . letting the rank of @xmath85
be greater than @xmath86 , one introduces multiconfigurational states and goes beyond the single determinant representation . in this more general case the determinantal wave function
is called antisymmetrized geminal power ( agp ) .
the agp is the particle - conserving version of the bardeen - cooper - schrieffer ( bcs ) wave function and it accounts for static correlations in the system , namely deriving from nearly - degenerate electronic energy levels . together with the jastrow factor it forms the jastrow - antisymmetrized geminal power ( jagp ) wave function which represents a practical implementation of the resonance valence bond idea introduced by linus pauling for chemical systems @xcite .
the jagp wave function has been proven to be particulary accurate in describing a wide range of strongly correlated systems @xcite , and it is the second variational form we tested in our work after the jsd . in order to benefit from the agp ansatz without paying the cost of dealing with a too large number of variational parameters , which will make the calculation unfeasible for big systems , we develop the agp expansion on a contracted basis set .
we used atomic hybrid orbitals , also employed in ref .
, as contractions of the primitive gto s .
one of the features of the agp ansatz ( without jastrow ) is that the geminal is also the one - body density matrix of the system . with the aim at finding the most effective local ( atomic ) basis set to describe the whole system
, we project the full one - body density matrix on its local atomic constituents , by retaining in the expansion of eq .
[ atomic_expansion ] only the @xmath91 s with @xmath64 , and setting to zero the other terms . by diagonalizing the projected one - body density matrix
, we obtain a set of local natural orbitals for each atomic fragment .
the atomic orbitals obtained in this way are hybrid ( i.e. linear combination of primitive gaussians not restricted to a given angular momentum shell ) , as they describe the hybridization arising from the atomic embedding , which breaks the spherical symmetry around the nucleus .
we call them _ atomic hybrid orbitals_. thanks to this feature , for each atom we keep not only the information on the local electronic structure due to the nuclear charge , but also the information of the nuclear embedding in the compound , namely the impact of the environment on its electronic structure . the initial
agp useful to fix the hybrid basis can be determined by dft calculations ( in that case the rank of the agp is n/2 as it comes directly from an sd wave function ) , or by a previously optimized jagp wave funtion ( in the latter case the jastrow factor is disregarded and only the determinantal part is taken in the one - body density matrix determination ) . once the hybrid basis set is chosen , and the agp expanded upon it , the geminal reads : @xmath92 where @xmath93 are the contracted atomic hybrid orbitals , and @xmath94 as before .
@xmath95 is the number of atomic hybrid orbitals required for an accurate description of each atom .
after their first determination by dft or by previous jagp calculations , they are further optimized in the qmc energy minimization .
the number of variational parameters of the wave function is @xmath96 , where @xmath97 is the total nuber of hybrid gaussians in the whole system .
since @xmath98 , the hybrid orbitals represent the optimal basis set , which reduces at most the total number of parameters in the correlated agp framework . in order to find the best value of @xmath95 for oxygen and hydrogen , we carried out a detailed analysis of the variational energy versus the hybrid basis size for the single water molecule , reported in sec .
[ qmc_water_sec ] .
all qmc calculations of this work have been carried out with the turborvb program@xcite .
all the variational parameters of the jsd and jagp wave functions have been optimized by means of the stochastic reconfiguration method with hessian accelerator , also called `` linear method''@xcite .
the forces on the parameters and ionic positions have always finite variance thanks to a reweighting scheme for finite systems introduced in ref . to cure the variance explosion around the nodes of the wave function .
a systematic way of improving the quality of the vmc ansatz is to perform lattice regularized diffusion monte carlo ( lrdmc ) calculations@xcite , a projective technique which filters out the high - energy components yielding a more accurate evaluation of the correlation energy . in the case of non - local pseudopotentials , as the one for the oxygen , lrdmc goes beyond the locality approximation , by providing always a variational upper bound of the true ground state energy . in the present paper , we report on both vmc and lrdmc calculations of the protonated water dimer .
the bfd energy - consistent pseudopotential described in sec .
[ gaussian_dft ] has been employed to replace the oxygen - core 1s electrons .
statistical error bars are kept smaller than @xmath99 kcal / mol in all qmc final calculation .
we report the results of our qmc study of protonated water dimer .
this section is organized as follows . the first part ( sec . [ qmc_water_sec ] )
is devoted to benchmark computations on water molecule which aim at proving the quality of our qmc ansatz .
the second part ( sec . [ water_dimer ] ) presents the qmc study on the protonated water dimer .
different level of theory are compared with qmc calculations for both geometry and energetics outcomes in order to assess the accuracy of our approach in the study of proton transfer systems .
the single water molcule has been the subject of many numerical studies based on several quantum chemistry methods .
it is essential to have a good description of its structural and electronic properties in order to tackle the study of larger water clusters , as the intramolecular degrees of freedom will significantly affect the intermolecular environment due to the large water dipole moment and the strong directionality of the h bond .
here we report our pseudopotential and all - electron calculations on the water molecule with different wave function types and basis sets , with the aim at choosing the best compromise bewteen accuracy and efficiency in order to transfer the most convenient ansatz to the zundel complex and eventually to larger water clusters . & & & & & & + jsd ( uncontracted orbitals ) & -17.24821(7 ) & 0.2655(6 ) & 0.0 & 682 & 18 & 895^_a_^ + jagp ( hybrid orbitals : 4o 1h ) & -17.25013(8 ) & 0.2635(12 ) & -1.91(11 ) & 21 & 158 & 374 + jagp ( hybrid orbitals : 4o 5h ) & -17.25183(6 ) & 0.2510(6 ) & -3.62(10 ) & 105 & 238 & 538 + jagp ( hybrid orbitals : 8o 2h ) & -17.25267(7 ) & 0.2426(18 ) & -4.46(10 ) & 78 & 298 & 571 + jagp ( hybrid orbitals : 8o 5h ) & -17.25302(6 ) & 0.2412(34 ) & -4.89(10 ) & 171 & 358 & 724 + jagp ( uncontracted orbitals ) & -17.25389(6 ) & 0.2296(5 ) & -5.68(10 ) & 682 & 18 & 895 + & & & & & & + jsd ( uncontracted orbitals ) & -76.40025(8 ) & 1.412(3 ) & 0.0 & 1383 & 19 & 1819 + jagp ( hybrid orbitals : 9o 2h ) & -76.40504(9 ) & 1.399(6 ) & -4.79(12 ) & 91 & 361 & 870 + jagp ( uncontracted orbitals ) & -76.40660(7 ) & 1.374(3 ) & -6.35(11 ) & 1383 & 19 & 1819 + pseudo jsd & 0.9542(4 ) & 104.730 + pseudo jagp & 0.9549(4 ) & 104.549 + all - electron jsd & 0.9539(4 ) & 105.187 + all - electron jagp & 0.9557(4 ) & 105.101 + experiment@xcite & 0.95721(30 ) & 104.522(50 ) + @xmath101 ( h ) & -17.26280(6 ) & -76.42475(15 ) + @xmath102 ( h ) & -17.26528(6 ) & -76.42690(14 ) + @xmath103 ( mh ) & -2.48(9 ) & -2.15(21 ) + as reported in sec . [ comp_methods ] , the bfd pseudopotential@xcite has been used for oxygen , while the two hydrogens have been treated all - electron .
the primitive gaussian basis set is o(5s,5p,2d ) and h(4s,2p ) for the determinant , while for the jastrow factor it is o(3s,2p,1d ) and h(2s,1p ) .
note that the latter basis set has been recently claimed to be one of the most accurate in an extensive qmc study of single molecule water properties@xcite , which used the same jastrow ansatz as ours .
the inhomogeneous three- and four - body jastrow term ( eq . [ 3body ] ) has been developed directly on the uncontracted primitive basis set , whose flexibility guarantees a very good description of dynamical correlations . on the other hand , for the antisymmetric part we tested two main wave function forms , the single slater determinant ( obtained by using a geminal with rank @xmath86 ) , and the agp function .
the difference in energy between the jsd and the jagp wave functions , reported in tab .
[ table_pseudo_water_energy ] , shows the size of static correlations in the system , which amounts to 5 - 6 mh .
this leads also to better geometrical properties , as seen in tab .
[ table_geometry_h2o ] . the jagp geometry is closer to the experiment than the jsd one , in both the oh distance and the hoh angle .
the structural effects of static correlations in the water molecule have been already pointed out in ref . , where they were attributed mainly to a change in the local description of the oxygen atom . in order to analyze in deeper detail the role of the agp correlations in the water molecule
, we studied the natural orbital occupations coming from the diagonalization of the geminal .
one of the appealing features of the agp theory is that the geminal wave function is directly related to the one - body density matrix of the system ( without jastrow factor ) .
its diagonalization yields the molecular natural orbitals as eigenvectors and their weights are related to the modulus of the agp eigenvalues which are plotted in fig .
[ agp_eigenvalues ] .
the figure shows that indeed the orbitals above the homo ( highest occupied molecular orbital in the single slater determinant representation ) have a sizable weight , with a distribution which falls abruptly to zero only after the 40-th orbital ( gray area in the plot ) .
this reflects the multi determinant character of the water molecule , taken into account by the agp ansatz .
we thus believe that , although the entanglement of quantum levels at the origin of static correlations can come from the oxygen atom , its impact in water has a genuine molecular character .
last but not least , the multi determinant agp representation leads to a better description of the nodal surface of the true ground state , with a gain of about 2.5 mh in the fixed node lrdmc energy with respect to the one obtained by using the jsd as trial wave function , as reported in tab .
[ table_lrdmc_water_energy ] .
we turn now the attention on how to reduce the agp basis set in an effective way .
so far , both the jsd and jagp wave functions have been developed on the uncontracted primitive basis in order to exploit at most its flexibility .
thus , the total number of variational parameters is 895 ( see the last column of tab .
[ table_pseudo_water_energy ] ) , quite large for a single molecule , particularly if one would like to tackle the study of larger water clusters by means of qmc techniques .
the most important limitation of this approach is that the number of variational parameters corresponding to the matrix elements @xmath104 increases clearly as the square of the atomic basis size ; therefore this should be reduced at minimum in order to make this approach feasible for a large number of molecules . to this porpouse , as explained in sec .
[ qmc_wave_function_ansatz ] , we define a new smaller basis set by contracting the o(5s,5p,2d)/h(4s,2p ) gaussian primitive basis via atomic natural hybrid orbitals . each atom in the system is described by its own set of hybrid orbitals .
we study how the size of the contracted hybrid basis set affects the quality of the geminal expansion .
we compare it with its rigorous lowest energy limit provided by the uncontracted jagp reference previously computed .
the results are reported in tab .
[ table_pseudo_water_energy ] .
the smallest basis set which includes the 1s for h and the 2s and 2p orbitals for o , thus taking into account the 2s2p near degeneracy at the atomic o level , is the 4o 1h hybrid set ( in self - explaining notations ) .
it gives the poorest energy and variance among the hybrid basis sets considered , though being lower than the jsd ansatz .
the best energy is obtained with the largest hybrid basis tried here , namely the 8o 5h set .
it recovers a large fraction of static correlations and its energy is less than 1mh above the uncontracted jagp one .
however , the price to pay is that the parameter reduction is weak , the total number of parameters being close to the one of the full jagp expansion ( see last column of tab . [ table_pseudo_water_energy ] ) . indeed , while the number of @xmath104 is still significantly lower than the one for the uncontracted basis set , nevertheless the parameters in the contracted orbitals grow too much .
the best compromise between efficiency , i.e. total number of variational parameters , and accuracy , i.e. variational energy , is provided by the 8o 2h basis , as it yields a significant gain in energy with a small / moderate number of parameters .
this advantage will be remarkable for large number of atoms , as the number of variational parameters corresponding to this atomic natural hybrid orbitals grows only _ linearly _ with the number of atoms ; on the other hand the number of parameters corresponding to @xmath104 , grows instead quadratically , but it remains still affordable since it is dramatically reduced by this approach ( see tab .
[ table_allel_water_energy ] ) .
finally , we study how the agp spectrum changes with the contracted hybrid basis sets .
[ agp_eigenvalues ] shows that , after a complete wave function optimization , the natural orbital eigenvalues magnitude of the hybrid agp covers the @xmath105 range of the fully uncontracted agp expansion , except for the shortest 4o 1h basis , which clearly spans a too small hilbert space .
moreover , we checked that the jagp expanded on the optimal 8o 2h basis gives the same fixed node lrdmc energy as the full jagp , signalling that the nodal surface is properly described even by the hybrid 8o 2h contraction .
therefore we are going to use it in the study of the protonated water dimer presented after in the paper .
as reported in the method section , the gain in efficiency of the hybrid basis set is expected to be larger and larger as the system size increases , as the quadratic growth of the @xmath104 with the number of atoms depends strongly on the atomic basis size .
to study the accuracy of the pseudopotential approximation , we carried out also all - electron calculations of the water molecule .
with respect to the pseudopotential calculations , the primitive and contracted basis sets for o have been extended in order to account for the additional 1s electrons .
the primitive basis set is then o(6s6p2d ) and o(3s3p1d ) for the determinant and jastrow term respectively , while the optimal contracted hybrid basis set is 9o 2h . in tab .
[ table_allel_water_energy ] we report the variational energies for different wave functions .
the energy gain provided by the all - electron jagp wave function is very close to the one in pseudopotential calculations .
the substantial agreement between the two calculations is apparent also in fig .
[ agp_eigenvalues ] , where the eigenvalues of the higher energy molecular natural orbitals in the agp behave similarly .
the lrdmc energy difference between the jagp and jsd trial wave functions coincides within the error bars with the one with pseudopotentials .
the nodal contribution to the fixed - node energy is the same .
the jagp lrdmc energy is one of the best ever published in literature , in statistical agreement with the one computed by luchow and fink@xcite ( -76.429(1 ) ) h ) , who used 300 determinants in the trial wave function , and with the one by zen et al.@xcite ( -76.42660 h ) , who used the same ansatz as ours in the trial .
the jagp lrdmc projected energy is only 11 mh higher than the extrapolated exact result of 76.438 h@xcite . for @xmath106 and @xmath107 for @xmath108 .
in the agp representation , by diagonalizing the geminal we obtain the corresponding molecular orbitals ( eigenvecotors ) and their occupations @xmath109 ( eigenvalues ) . in the agp ,
also the orbitals above the homo are occupied , with a weight which jumps across the homo to lumo transition ( in going from the green to white region ) .
the gray area shows when the mo occupation tail falls rapidly to zero in the full agp ( expanded on a primitive basis set ) , signaling that the mo s above that threashold start to be irrelevant to describe the static correlations in the system .
] the all - electron calculations confirm the importance of including static correlations to have a better description of the geometry , as shown in tab .
[ table_geometry_h2o ] .
however the hoh angle turns out to be less accurate than the one obtained with pseudopotentials if compared to the experiment , most probably because it is a quantity very sensitive to the basis set convergence , which is harder to reach in all - electron calculations .
another drawback of all - electron calculations is of course the larger variance for an equivalent wave function ansatz , due to the 1s electron fluctuations , as one can easily evince from the comparison between tab .
[ table_pseudo_water_energy ] and tab .
[ table_allel_water_energy ] . due to the larger primitive basis set required in all - electron calculations ,
the parameter reduction allowed by the hybrid basis contraction ( 9o 2h ) has a great impact on the efficiency .
the total number of parameters is reduced by almost a factor of 4 in the determinantal part of the single molecule , without any significant loss of accuracy in the jagp total energy ( see tab .
[ table_allel_water_energy ] ) .
the use of the bfd pseudopotential for oxygen , and the jagp ansatz together with the hybrid basis set , tested in the water molecule , is transferred to the protonated water dimer .
the protonated water dimer represents the simplest model for proton transfer in acqueous systems . in our work
we focus on the energetics of @xmath0 related to proton transfer by choosing suitable reaction coordinate ( rc ) . selecting a rc
allows to reduce the complexity of the full - dimensional pes .
it projects it onto a single - dimensional subset which retains the most important physical features of the full hypersurface .
since it is not unique , a correct choice of rc is essential in order to filter out the physical features which suitably describe the considered phenomenon .
we propose a mechanism of proton transfer within the dimer which leads to a natural definition of our rc .
as modeled by our system , the proton transfer reaction takes place in three different steps . at first
the excess proton is bounded to one water molecule forming an @xmath110 complex . by thermal fluctuations
the oxygen - oxygen separation can get closer to the optimal distance of the zundel complex ( around @xmath111 ) ; at this stage the system assumes a `` zundel configuration '' with the proton equally shared between the two oxygens .
a further stretch of @xmath8 disfavors the zundel configuration and a new @xmath112 complex is produced .
the overall effect of this process is a transfer of a proton along the hydrogen bond between two oxygen atoms .
this mechanism suggests to choose the oxygen - oxygen separation as rc for the dimer potential energy curve .
we present in the following the geometry and energetics of the protonated water dimer .
results on the global minimum geometry are reported at first .
the potential energy curve is the result of a stretching of the oxygen - oxygen distance ; the behavior of the excess proton is thus investigated along the same path .
a particular attention is devoted to the behavior of the @xmath110 complex in the broken - symmetry region of the energy landscape .
proton transfer static barriers at different @xmath8 distances are calculated in order to further estimate the accuracy of our approach .
all qmc calculations are performed with the jagp wave function ansatz developed on an atomic hybrid basis set , as discussed in the previous sec .
[ qmc_water_sec ] .
the minimum energy structure of @xmath0 has been debated in the scientific community , as there are two candidates with competing energies : a @xmath113-symmetric structure , commonly addressed as zundel configuration , with the proton evenly shared between the two oxygens , and a @xmath114-symmetric one with the proton slightly localized on one @xmath100 molecule ( see fig . [ geometries ] ) .
several _ ab - initio _
investigations @xcite have shown that a better treatment of electron correlations turns out in a change of the ground state ( gs ) geometry from the @xmath114-sym to the @xmath113-sym configuration .
accurate highly - correlated studies @xcite has eventually confirmed that the global minimum is @xmath113 symmetric . at a qmc level of theory , the @xmath0 gs shows a @xmath113 zundel configuration with centrosymmetric excess proton ( left part of figure [ geometries ] ) , in agreement with the previous studies .
the main geometrical parameters of the global minimum are presented in tab .
[ table_geometry_c2 ] , both for the pseudopotential and all - electron calculations .
dft - pbe & 2.4111 & 1.2074 & 1.2074 & 173.661 & 0.9697 &
0.9691 + dft - b3lyp@xcite & & & 1.2172 & 173.6 & 0.9706 & 0.9701 + qmc - with pseudo & 2.3847(5 ) & 1.1930(5 ) & 1.1942(8 ) & 174.71(7 ) & 0.9605(8 ) & 0.9650(8 ) + qmc - all - electron & 2.3905(4 ) & 1.1944(6 ) & 1.1989(5 ) & 174.43(9 ) & 0.9630(7 ) & 0.9628(6 ) + ccsd(t)@xcite & 2.3864 & 1.1950 & 1.1950 & 173.730 & 0.9686 &
0.9682 + dft - b3lyp@xcite & & & 1.2507 & 175.4 & 0.9746 & 0.9741 + qmc - with pseudo & 2.3996(6 ) & 1.1154(8 ) & 1.2852(4 ) & 176.5(1 ) & 0.9641(7 ) & 0.9625(4 ) + qmc - all - electron & 2.3913(3 ) & 1.1285(5 ) & 1.2648(4 ) & 175.29(6 ) & 0.9635(4 ) & 0.9616(5 ) + ccsd(t)@xcite & 2.3989 & 1.1233 & 1.2720 & 175.646 & 0.9641 & 0.9645 + if compared with ccsd(t ) , the qmc ground state geometries show an agreement of up to @xmath115 in the atomic separations , and up to @xmath116 in the angles . as the angle is close to @xmath117 , in this case its determination is more delicate and could be affected by a larger statistical bias .
therefore we notice a slight discrepancy on this quantity between qmc and coupled cluster outcomes .
nevertheless , these differences do not affect the overall description of the gs and the energetics of the system . in the present work ,
also the @xmath114-symmetric structure has been taken into account ( left - hand side of fig .
[ geometries ] ) . tab .
[ table_geometry_cs ] , which reports the vmc optimized @xmath114 geometries , confirms the trend seen in tab .
[ table_geometry_c2 ] for the gs , although the discrepances in the bond lengths are slightly larger between different methods , the ccsd(t ) values being in between the all - electron and pseudopotential vmc results .
the energy difference between the @xmath114 configuration and the @xmath113-symmetric global minimum turns out to be @xmath118 kcal / mol at vmc level and @xmath119 kcal / mol with lrdmc .
this is in satisfactory agreement with previous results carried out with mller - plesset pertubation theory@xcite ( @xmath120 kcal / mol ) and ccsd(t)@xcite ( @xmath121 kcal / mol ) techniques .
the quality of pseudopotential is hence verified also for the zundel complex . in view of this ,
its pes along the chosen rc will be worked out mainly with this approximation due to its substantially lower computational demand .
comparison between different computational methods .
structural relaxation is performed at each level of theory .
each curve has its minimum as reference point . ] from the global minimum @xmath113-sym configuration we stretch the oxygen - oxygen distance in order to study the potential energy curve and to elucidate the proton transfer properties in the dimer .
the structural relaxation at the vmc level for fixed @xmath8 separation requires a careful procedure due to the flatness of the pes . starting from a pbe - dft optimized geometry ( see sec .
[ car_parri_dft ] for technical details ) and a jagp variational wave function fully optimized in the electronic part , the atomic coordinates are relaxed with the steepest descent method , considering them as additional variational parameters of the qmc wavefunction .
for the sake of comparison , we minimize a parametrized full - dimensional pes fitted from ccsd(t ) calculations@xcite to find the best coupled cluster estimates of energy and geometry . by means of the _ downhill simplex _ minimization technique , we find the configuration of lowest ccsd(t ) energy at the same constrained @xmath8 distance as the corresponding qmc and dft calculations . in fig .
[ landscape1 ] we plot the energy landscape along the rc for pbe - dft , ccsd(t ) , vmc and lrdmc , the latter computed at the vmc geometry . for vmc technique , we report also full potential calculations for some @xmath8 separations .
we notice a substantial agreement among all techniques in the region at the left of the global minimum of the curve , except for a rigid shift by 0.02 between the pbe - dft results and the others .
the pbe @xmath8 minimum is indeed located at 2.41 , while the minimum of the other methods turns out to be at @xmath122 2.39 .
hence this part of the energy curve is only slightly influenced by a better treatment of correlations .
on the other hand , the region at the right of the minimum , at intermediate @xmath8 distances ( @xmath123 ) , displays a different behavior .
the pbe - dft overestimates the slope of the curve with respect to the most accurate techniques .
lrdmc , which yields the best qmc correlation energy , shows a remarkably good agreement with the state - of - the - art ccsd(t ) results . in particular all ccsd(t )
outcomes are in the range of @xmath124 kcal / mol , three times the statistical error of the lrdmc calculations , as shown in fig .
[ blowup_landscape ] .
the curves reported in fig .
[ landscape1 ] are obtained with the minimum energy geometry at each level of theory ( except for the lrdmc , whose geometry is set at the vmc level ) . in order to obtain a more reliable comparison and avoid the bias coming from the use of different geometries , we carried out the same calculations employing the vmc - optimized structures for every technique . the result
is reported in fig .
[ landscape2 ] .
the trend displayed in fig .
[ landscape1 ] for method - optimized geometries is remarkably enhanced when the same configuration of the dimer is considered .
away from the minimum , the pbe - dft energies show a larger overestimation of the slope .
distance ( ) .
all the calculations are performed with vmc - optimized geometry .
the zero energy reference point corresponds to the minimum of each curve . ]
the slope of the dimer potential energy curve is related to the behavior of the excess proton in the system . in fig . [ oh_distances ] this property is elucidated .
we report in the same plot the separations between the excess proton and each of the two oxygens : @xmath125 and @xmath126 .
the structure of the dimer is relaxed at different level of theories with the same procedure carried out for the potential energy curve in fig .
[ landscape2 ] .
the plot clearly shows the appeareance of two distinct regimes of the dimer .
one is characterized by a symmetric zundel configuration with the proton evenly bonded to the two oxygens ; from the point of view of proton transfer physics , it is basically equivalent to the gs configuration . stretching the @xmath8 distance results in the formation of a @xmath110 complex with the proton localized on one water molecule .
these configurations belong to the _ asymmetric regime _ of the dimer . within this regime
, the initial @xmath113 point symmetry of the gs geometry is broken due to proton localization .
notice that the @xmath114-sym local minimum presented in sec .
[ global_minimum ] displays a localized proton at a distance of @xmath127 , smaller than the symmetry - breaking distance shown in fig .
[ oh_distances ]
. however , we verified that if one stretches the @xmath8 distance starting from this local minimum , the energies obtained are higher than the values plotted in fig .
[ landscape1 ] and fig .
[ landscape2 ] .
the distances obtained by the qmc relaxation of the atomic coordinates are in excellent agreement with the ccs(t ) calculations ; in particular the root mean square distance between the two data sets over the whole @xmath8 range is of @xmath128for both @xmath125 and @xmath129 .
electron correlation plays a key role in determining the stability of the symmetric configuration of the dimer .
the overestimation of the potential energy curve slope by pbe - dft in fig .
[ landscape1 ] corresponds to an overestimation of the symmetry - breaking @xmath8 separation by @xmath130 with respect to higher - level post - dft _ ab - initio _ methods , and to a very poor description od the geometry in the symmetry - broken region close to the symmetry - breaking point , where the discrepancy in the @xmath131 distances is the largest ( going up to 0.15 ) . the zero - temperature potential energy curves reported in fig . [ landscape1 ] and fig .
[ landscape2 ] seem to conflict with the proton transfer mechanism discussed in the introduction of sec .
[ water_dimer ] , since the configuration with centrosymmetric proton is energetically favored and therefore it represents a stable state rather than a transition state between two asymmetric configurations with localized proton . this contraddiction is only apparent . indeed , ref .
has shown that the introduction of thermal and polarization effects due to the physical environment , favors the asymmetric regime of the complex . indeed , at finite temperature the _ free energy _
landscape displays a global minimum shifted towards the asymmetric regime and the zundel - like structure does not represent anymore the energetically favored configuration .
moreover , a recent experimental result shows that the average @xmath8 distance in liquid water is of @xmath132 @xcite which clearly corresponds to a symmetry - broken configuration of the dimer ( see fig .
[ oh_distances ] ) . anyway
, in order to jump from a water molecule to one of its neighbors , the proton must pass through a zundel configuration .
describing correctly the energetics and geometry of the protonated water dimer in the symmetry - breaking transition region is therefore of paramount importance to hope having an accurate description of the pt in more realistic models . indeed , as already mentioned , fig .
[ oh_distances ] highlights that at @xmath133 k a better treatment of electronic correlations turns out in a stability of the asymmetric regime over a considerably wider range of @xmath8 distances with respect to dft results .
hence we expect that , at finite temperature and including the sorrounding physical environment , the value of the free energy activation barrier for proton transfer would be significantly higher than the pbe - dft prediction . in this section
we focus the attention on some relevant properties of the broken - symmetry region where the excess proton is localized on a water molecule . at the qmc level of theory
, the formation of a @xmath110 complex occurs at an oxygen - oxygen distance of @xmath134 in perfect accordance with ccsd(t ) results , as seen in fig .
[ oh_distances ] .
a quantity which has been extensively studied over the past years@xcite is the _ static proton transfer barrier _ , i.e. the barrier that the @xmath135 has to overcome in order to jump from one h@xmath7omolecule to the other at a fixed @xmath8 distance in the symmetry - broken regime .
this quantity does not provide a realistic comparison with the experimental activation barrier for pt , as the @xmath8 distance will shorten during the proton hopping . nevertheless , it is relevant in order to provide a further check of the accuracy of our qmc approach .
fixing the @xmath8 distance , the barriers are obtained as difference between the asymmetric configuration with localized proton and a structure with the excess proton in a centrosymmetric position .
calculations are performed in three representative @xmath8 separations ; the results are shown in tab .
[ static_barrier ] where they are compared with existent data in literature .
the first @xmath8 distance is @xmath136 , very close to the symmetry - breaking point of the dimer as displayed by highly correlated approaches ; on the contrary , at a dft level the configuration with centrosymmetric proton is still energetically favored , as shown by fig.[oh_distances ] .
it has been noticed@xcite that the pes in this region is particulary flat .
furthermore the potential energy curve of the dimer develops on very tiny energy differences around the symmetry - breaking point .
these issues make the calculations in this region of the pes extremely delicate since the stochastic noise can considerably affect the quality of the qmc predictions
. indeed , tab .
[ static_barrier ] shows that qmc and ccsd(t ) display a vanishing energy barrier of the order of @xmath99 kcal / mol as the hydrogen is displaced along the oxygen - oxygen axis .
the height of this barrier is slightly above the attained statistical error in our typical qmc run .
however , despite the very sensitive behavior of the dimer pes around the symmetry - breaking point , the accuracy of our force minimization algorithm allows to account for the tiny energy differences involved ; thus it ensures the necessary precision to describe the pt physics in the dimer .
as we discuss in the next section , a similar accuracy can be achieved in larger molecular clusters with a reasonable amount of computational time .
therefore our qmc framework guarantees a reliable description of the pt physics also for more realistic models .
.static proton transfer barriers ( kcal / mol ) at fixed @xmath8 separations .
comparison between different level of theory .
[ cols="^,^,^,^",options="header " , ] the other results are obtained at larger oxygens separations , further from the symmetry - breaking point . they confirm the general behavior already seen along the @xmath8 rc .
lrdmc and cc results are in a good agreement up to @xmath137 kcal / mol , whereas the vmc slightly overestimates the barrier . as well known from previous works , dft substantially underestimates the barrier with respect to post - dft methods which provide a better treatment of correlations . finally , let us analyze the extreme limit of the asymmetric zundel configuration , namely when @xmath138 , with the formation of one h@xmath7o and one hydronium .
the dissociation energy @xmath139 of @xmath0 is computed by setting the distance between the two oxygens to @xmath140 . with the ccsd(t ) pes , we checked that this is already in the large distance plateau .
we get a @xmath139 of @xmath141 kcal / mol by vmc , and @xmath142 kcal / mol by lrdmc , to be compared with the ccsd(t ) value of @xmath143 kcal / mol from ref . , while the pbe - dft gives @xmath144 kcal / mol .
the agreement between the lrdmc and ccsd(t ) is impressive , while it is already good for the vmc estimate.this is mainly due to the size consistency of the jagp ansatz , obtained once the jastrow is close to the complete basis set limit .
the vmc and ccsd(t ) dissociation energies are plotted in fig . [ blowup_landscape ] with respect to lrdmc values . & & + 1 & 0.05 & 0.15 + 2 & 0.24 & 2.13 + 6 & 6.49 & 164.35 + qmc methods present a favorable scalability with the number of particles with respect to other highly correlated approaches such as cc . with the aim at proving this feature also for our approach
, we performed a benchmark calculation on a more realistic pt model composed of 6 water molecules and one excess proton , which will be the subject of a further study . in table
[ comp_time ] we report a comparison of the computational time required to carry out typical vmc and lrdmc runs for different sizes of the protonated water cluster at fixed optimized variational parameters and geometry .
the calculations have been carried out on the hpc curie thin nodes ( 2.7 ghz 8 core intel sandy bridge processors ) , and performed with a target statistical error of @xmath145 kcal / mol in the total energy . by performing a simple fit on the data in table [ comp_time ]
, we notice that the simple vmc displays a @xmath54 scaling with the number of particles ; the more accurate lrdmc , carried out at a lattice space of @xmath146 , shows instead an almost perfect @xmath3 scaling .
however , the lrdmc calculations are still feasible in a reasonable computational time for the 6 h@xmath7o cluster , while the vmc is still cheap at that cluster size .
in this paper we presented an extensive study of the protonated water dimer by means of the vmc and lrdmc techniques .
the jagp ansatz employed in this work implements an accurate treatment of both static and dynamical correlations among electrons .
the expansion of the determinantal part over atomic hybrid orbitals ensures a drastic reduction of the number of variational parameters thus making the wave function optimization procedure efficient and robust even for large systems . the comparison with previous published qmc calculations on the single water molecule showed the quality of our wave function ansatz .
total energy calculations are performed with the less expensive variational monte carlo approach along with the more precise projective diffusion monte carlo .
the powerful minimization algorithm implemented in our qmc software , allows an efficient estimation of the forces acting on each atomic component with a reasonable computational cost .
hence both energetics and geometry calculations are perfomed within the qmc framework .
a simple mechanism of proton transfer in the dimer has been presented and exploited to choose a suitable reaction coordinate for the potential energy curve .
the energy landscape as a function of the oxygen - oxygen distance is computed and compared with density functional theory in the pbe approximation and with ccsd(t ) results ; lrdmc is in excellent agreement with ccsd(t ) calculations ( within @xmath137 kcal / mol ) , whereas minor differences ( up to about @xmath147 kcal / mol ) are shown by vmc .
ref . has recently reported a similar accuracy for small water clusters by diffusion monte carlo calculations .
geometrical properties of the excess proton are also investigated by vmc structural relaxations , which provide geometries remarkably close to the ones obtained by a ccsd(t ) fitted pes .
we show the presence of two distinct regimes of the dimer depending on the oxygen - oxygen distance : one with a centrosymmetric excess proton and the other with the proton localized on one of the water molecules .
the stability of these configurations crucially depends on the level of theory employed to describe the electronic structure of the system . a better treatment of electron correlation results in the stability of the asymmetric proton geometry over a wider range of @xmath8 distances .
these results , together with the proton transfer static barrier and the dissociation energy @xmath139 , show that our qmc approach has a global accuracy comparable with the state - of - the art coupled cluster in both geometry and energetics of the dimer , with the advantage of having a better scaling with the number of particles .
the accuracy shown by our qmc calculations combines with a favorable scalability with the number of particles with the respect to other highly correlated techniques , as demonstrated with a test simulation on a larger protonated water cluster .
these features make this approach a very promising candidate for the study of proton transfer in complex acqueous systems .
in particular , we have shown that the vmc method is cheap , provides very accurate geometries and a global accuracy of less than 1 kcal / mol in the most important region for the pt physics .
we hope that our work will inspire further studies on this direction , and pave the way for accurate highly - correlated simulations of more realistic proton transfer models which will eventually shed new insights onto the pt mechanism in water .
we thank joel m. bownman and xinchuan huang who provided us the coupled cluster potential energy surface of @xmath0 .
one of us ( michele casula ) acknowledges computational resources in the form of the genci grant number x2013096493 .
99 w. l. jorgensen , j. chandrasekhar , j. d. madura , r. w. impey , m. l. klein , j. chem . phys . * 79 * , 926 ( 1983 ) .
m. sprik , j. phys . chem . * 95*,2283 ( 1991 ) .
m. w. mahoney , w. l. jorgensen , j. chem .
phys . * 112 * , 8910 ( 2000 ) .
p. l. silvestrelli , m. parrinello , j. chem .
phys * 111 * , 3572 ( 1999 ) .
j. c. grossman , e. schwegler , e. w. draeger , f. gygi , g. galli , j. chem .
* 120 * , 300 ( 2004 ) .
j. vandevondele , f. m. and m. krack , j. hutter , m. sprik , m. parrinello , j. chem . phys . * 122 * , 014515 ( 2005 ) .
r. jonchiere , a. p. seitsonen , g. ferlat , a. m. saitta , r. vuilleumier , j. chem . phys . * 135 * , 154503 ( 2011 ) .
r. z. khaliullin , t. d. khne , arxiv/1303.2067 ( 2013 ) .
p. carloni , u. rothlisberger , m. parrinello , acc .
35 * , 455 ( 2002 ) .
r. a. friesner , b. d. dunietza , acc .
chem . res . *
34 * , 351 ( 2001 ) .
s. yoo , x. c. zeng , s. s. xantheas , j. chem . phys . * 130 * , 221102 ( 2009 ) . h. s. lee , m. e. tuckerman , j. chem .
phys . * 126 * , 164501 ( 2007 ) . c. j. d. von grotthuss , ann .
lviii * , p. 54
n. agmon , chem .
. lett . * 244 * , 456 ( 1995 ) .
d. marx , chemphyschem , * 7 * , 1848 ( 2006 ) .
t. e. decoursey , physiol rev . *
83 * , 475 ( 2003 ) .
e. freiera , s. wolfb , k. gerwerta , proc .
* 108*,11435 ( 2011 )
. j. deisenhofer , o. epp , i. sinning , h. michel , j. mol . biol . * 246 * , 429 ( 1995 ) .
d. lancaster , h. michel , b. honig , m. gunner , biophys .
j. * 70 * , 2469 ( 1996 ) .
m. eigen , angew .
int . ed . * 3 * , 1 ( 1964 ) .
o. f. mohammed , d. pines , j. dreyer , e. pines , e. t. j. nibbering , science * 310 * , 83 ( 2005 ) .
o. f. mohammed , d. pines , e. pines , e. t. j. nibbering , chem .
phys . * 341 * , 240 ( 2007 ) .
z. luz , s. meiboom , j. am .
soc . * 86 * , 4768 ( 1964 ) .
noam agmon , _ hydrogen bonds , water rotation and proton mobility . _ , j. chim .
( paris ) , * 93 * , 1714 ( 1996 ) .
u. w. schmitt , g. a. voth , j. chem . phys .
* 111 * , 9361 ( 1999 ) .
h. lapid , n. agmon , m. k. petersen , g. a. voth , j. chem .
* 122 * , 014506 ( 2005 ) .
d. marx , m. e. tuckerman , j. hutter , m. parrinello , nature * 397 * , 601 ( 1999 ) .
o. markovitch , h. chen , s. izvekov , f. paesani , g. a. voth , n. agmon , j. phys .
b * 112 * , 9456 ( 2008 ) .
d. marx , a. chandra , m. e. tuckerman , chem . rev . * 110 * , 21742216 ( 2010 ) .
a. m. saitta , f. saija , p. v. giaquinta , phys .
108 * , 207801 ( 2012 ) a. hassanali , f. giberti , j. cuny , t. d. khne , m. parrinello , proc .
usa * 110 * , 13723 ( 2013 ) .
e. wicke , m. eigen , h. ackermann , z. phys .
* 1 * , 340 ( 1954 )
. g. zundel , h. metzger , z. phys .
chem * 58 * , 225 ( 1968 ) .
y. xie , r. b. remington , h. f. schaefer , j. chem .
* 101 * , 4878 - 4884 ( 1994 ) .
e. f. valeev , h. f. schaefer , j. chem . phys . * 108 * , 7197 - 7201 ( 1998 ) .
d. j. wales , j. chem . phys . * 110 * , 10403 ( 1999 ) .
a. a. auer , t. helgaker , and w. klopper , phys .
chem . phys .
* 2 * , 2235 ( 2000 ) .
d. wei , d. r. salahub , j. chem .
* 101 * , 7633 ( 1994 ) .
m. e. tuckerman , d. marx , m. l. klein , m .
parrinello , science * 275 * , 817 ( 1997 ) . j. m. headrick , e. g. diken , r. s. walters , n. i. hammer , r. a. christie , j. cui , e. m. myshakin , m. a. duncan , m. a. johnson , k. d. jordan , science * 308 * , 1765 ( 2005 ) .
x. huang , b. j. braams , j. m. bowman , j. chem .
* 122 * , 044308 ( 2005 ) .
p. giannozzi _
_ , j. phys .
: condens .
matter * 21 * , 395502 ( 2009 ) .
j. p. perdew , k. burke , and m. ernzerhof , phys .
* 78 * , 1396 ( 1997 ) .
t. d. kh ne , m. krack , m. parrinello , j. chem .
theory comput .
* 5 * , 235 ( 2009 ) trouiller n. , martins j. , phys .
b , * 43 * , 1993 ( 1991 ) . r car , m parrinello , phys . rev
. lett . * 55 * , 2471 ( 1985 ) .
t. kato , comm .
pure appl . math . *
10*,151 ( 1957 ) .
m. holzmann , d. m. ceperley , c. pierleoni , and k. esler phys .
e * 68 * , 046707 ( 2003 ) .
s. sorella , m. casula , d. rocca , j. chem . phys .
* 127 * , 014105 ( 2007 ) .
s. azadi , c. cavazzoni , s. sorella , phys .
b * 82 * , 125112 ( 2010 ) m. burkatzki , c. filippi , m. dolg , j. chem . phys . * 126 * , 234105 ( 2007 ) .
r. assaraf and m. caffarel , j. chem . phys . * 119 * , 10536 ( 2003 ) .
s. chiesa , d. ceperley , s. zhang , phys .
lett . * 94 * , 036404 ( 2005 ) .
a. badinski , p. d. haynes , j. r. trail , r. j. needs , j. phys . : condens .
matter * 22 * , 074202 ( 2010 ) . c. filippi , c. j. umrigar , phys .
b * 61 * , 16291 ( 2000 ) .
s. sorella , l. capriotti , j. chem . phys . *
133 * , 234111 ( 2010 ) . c. attaccalite and s. sorella , phys .
lett . * 100 * , 114501 ( 2008 ) .
l. pauling , _ the nature of the chemical bond _
, 3rd edition , cornell university press , ithaca , new york .
d. nissenbaum , l. spanu , c. attaccalite , b. barbiellini , a. bansil , phys .
b * 79 * , 035416 ( 2009 ) . f. sterpone , l. spanu , l. ferraro , s. sorella , l. guidoni , j. chem .
theory comput . , * 4 * , 1428 ( 2008 ) .
m. marchi , s. azadi , m. casula , s. sorella , j. chem .
phys . * 131 * , 154116 ( 2009 ) .
a. zen , y. luo , s. sorella , l. guidoni , j. chem .
theory comput .
* 9 * , 4332 ( 2013 ) .
s. sorella , l. capriotti , phys .
b , * 61 * , 2599 ( 2000 ) .
s. sorella , phys .
b , * 64 * , 024512 ( 2001 ) . s. sorella , phys . rev .
b , * 71 * , 241103 ( 2005 ) .
c. j. umrigar , j. toulouse , c. filippi , s. sorella , r. g. hennig , phys .
lett . * 98 * , 110201 ( 2007 ) .
m. casula , c. filippi , s. sorella , phys .
lett . * 95 * , 100201 ( 2005 ) . m. casula , s. moroni , s. sorella , c. filippi , j. chem .
phys . * 132 * , 154113 ( 2010 ) .
w. s. benedict , n. gailar , e. k. plyler , j. chem .
phys . * 24 * , 1139 ( 1956 ) .
a. lchow and r. fink , j. chem . phys .
* 113 * , 8457 ( 2000 ) .
d. feller , c. boyle , and e. davidson , j. chem .
phys . * 86 * , 3424 ( 1987 ) .
m. park , i. shin , n. j. singh , k. s. kim , j. phys .
chem . a , * 111 * , 10692 - 10702 ( 2007 ) .
u. bergmann , a. di cicco , p. wernet , e. principi , p. glatzel et al , j. chem .
phys . * 127 * , 174504 ( 2007 ) .
s. s. xantheas , j. am .
* 117 * , 10373 ( 1995 ) .
m. pavese , s. chawla , d. lu , j. lobaugh , g.a .
voth , j. chem . phys . * 107 * , 7428 ( 1997 ) .
s. sadhukhan , d. munoz , c. adamo , g. e. scuseria , chem .
lett * 306 * , 83 - 87 ( 1999 ) .
e. neuscamman , phys .
lett . * 109 * , 203001 ( 2012 ) m. j. gillan , f. r. manby , m. d. towler , d. alf , j. chem . phys . * 136 * , 244105 ( 2012 ) | we report an extensive theoretical study of the protonated water dimer @xmath0 ( zundel ion ) by means of the highly correlated variational monte carlo and lattice regularized monte carlo approaches .
this system represents the simplest model for proton transfer ( pt ) and a correct description of its properties is essential in order to understand the pt mechanism in more complex acqueous systems .
our jastrow correlated agp wave function ensures an accurate treatment of electron correlations . exploiting the advantages of contracting the primitive basis set over atomic hybrid orbitals , we are able to limit dramatically the number of variational parameters with a systematic control on the numerical precision , a crucial ingredient in order to simulate larger systems . we investigate the energetics and geometrical properties of the @xmath0 as a function of the oxygen - oxygen distance , taken as reaction coordinate . in both cases ,
our qmc results are found in excellent agreement with state - of - the - art coupled cluster ccsd(t ) techniques .
calculations on proton transfer static barriers and dissociation energies display the same agreement .
a comparison with density functional theory results in the pbe approximation points out the crucial role of electron correlations for a correct description of the pt in the dimer .
finally , the ability of our method to resolve very tiny energy differences ( @xmath1 kcal / mol ) at which the proton hopping takes place and the corresponding structural variations optimized directly in the vmc framework is also proven .
our approach combines these features with a @xmath2-@xmath3 scaling with number of particles .
this value is favorable with respect to other highly correlated _ ab initio _ approaches and it allows the simulation of more realistic pt models ; a test calculation on a larger protonated water cluster is hence carried out .
the qmc approach used here represents a promising candidate to provide the first high - level _ ab - initio _ description of pt in water . |
in general , level splitting due to structure inversion asymmetry , sia , known as rashba s - o splitting@xcite is inversely proportional to the energy gap .
s - o coupling is zero for s - like conduction bands and strong in p - like hole states .
however , mixing of the conduction subbands with the valence subbands increases with decreasing energy gap .
it has been shown that electrons in narrow gap heterostructures based on hgte@xcite , exhibit strong rashba s - o coupling .
in addition to the small energy gap in hgte quantum wells , qw s , another important factor contributing to the large magnitude of the rashba s - o coupling is the inverted band structure of hgte qw s with well widths greater than 6 nm , in which the first conduction band has heavy hole character.@xcite for possible applications in spintronics,@xcite the rashba effect has recently been investigated in a number of narrow gap iii - v systems@xcite in which typical values of the rashba s - o splitting energy , @xmath7 , are 3 to 5 mev .
@xmath7 is appreciably larger in ii - vi hgte qw s , and values of 10 to 17 mev have been determined.@xcite zhang _
et al._@xcite demonstrated that the rashba s - o interaction is the dominant mechanism in such structures ; they studied the strong dependence of s - o splitting on gate voltage and its subsequent disappearance when the qw was symmetric as expected for the rashba effect .
compared to the observation of a series of nodes in shubnikov - de haas , sdh , oscillations for an in@xmath8ga@xmath9as heterostructure@xcite at @xmath10 t , similar beating patterns are observable at higher magnetic fields in hgte heterostructures@xcite due to its larger rashba effect . in this article
, we report on an investigation of beating patterns in the sdh oscillations in high quality @xmath11 type hgte / hg@xmath12cd@xmath13te qw s . up to 12 nodes have been observed in the beating pattern within a magnetic field range of 0.9 t @xmath146 t. a s - o splitting of @xmath15 mev due to the rashba effect has been directly deduced from the node positions .
this value is in good agreement with self - consistent hartree calculations . the observed sdh oscillations and beating patterns are also in good agreement with the density of states , dos , obtained from self - consistent @xmath16 calculations .
fully strained @xmath11 type hgte / hg@xmath0cd@xmath1te(001 ) qw s were grown by molecular beam epitaxy , mbe , on cd@xmath17zn@xmath18te(001 ) substrates in a riber 2300 mbe system .
details of the growth has been reported elsewhere.@xcite samples a and b are from the same chip , q1772 , which was modulation doped asymmetrically in the top barrier of the hgte qw structure using cdi@xmath19 as a doping material .
the hgte well width is 12.5 nm and the hg@xmath0cd@xmath1te barriers consist of a 5.5 nm thick spacer and a 9 nm thick doped layer . with a well width of 12.5 nm ,
the first conduction band in the qw has heavy hole character , i.e. , is a pure heavy hole state at @xmath20 , and following standard nomenclature is labeled @xmath21 .
standard hall bars were fabricated by wet chemical etching .
a 200 nm thick al@xmath19o@xmath22 film was deposited on top of the structure , which serves as an insulating layer .
finally al was evaporated to form a metallic gate electrode on sample b. a metallic gate was not fabricated on sample a , which accounts for the different two dimensional electron gas , 2deg , concentrations in these two samples .
ohmic indium contacts to the hall bars were formed by thermal bonding .
magneto - transport measurements were carried out in several different cryostats using dc techniques with currents of 1 to 5 @xmath23a in magnetic fields ranging up to 15 t and temperature from 1.4 to 35 k. during the measurement , the applied electric field was kept low enough to avoid excessive electron heating.@xcite
the band structure , landau levels , ll s , and rashba s - o splitting energy , @xmath7 were obtained from self - consistent hartree calculations based on an @xmath24 band structure model including all second order terms in the conduction and valence band blocks of the @xmath25 hamiltonian . in the calculations
the inherent inversion asymmetry of hgte and hg@xmath8cd@xmath9te has been neglected , because this effect has been shown to be very small in narrow gap systems.@xcite the envelope function approximation was used to calculate the subbands of the qw s and the influence of the induced free carriers has been included in a self - consistent hartree calculation . the valence band offset between hgte and cdte
was taken to be 570 mev,@xcite and to vary linearly with barrier composition.@xcite the band structure parameters of hgte and cdte at 0 k employed in this investigation are listed in and the model is described in detail elsewhere.@xcite the hgte qw s in this investigation have a well width of 12.5 nm and consequently the band structure is inverted . in other words , the first conduction band in the qw has heavy hole character , i.e. , is a pure heavy hole state at @xmath20 , and following standard nomenclature is labeled @xmath21 .
this has important consequences for the large magnitude of the s - o splitting of the @xmath26 and @xmath27 subbands .
cccccccccc & @xmath28 & @xmath29 & @xmath30 & @xmath31 & @xmath32 & @xmath33 & @xmath34 & @xmath35 & @xmath36 + & ev & ev & ev & & & & & & + hgte & -0.303 & 1.08 & 18.8 & 0 & 4.1 & 0.5 & 1.3 & -0.4 & 21 + cdte & 1.606 & 0.91 & 18.8 & -0.09 & 1.47 & -0.28 & 0.03 & -1.31 & 10.4 + [ table1 ]
[ t ] [ t ] typical sdh oscillations are shown in fig . [ q1772a - rxxdos ] for sample a with a hall concentration of @xmath37 @xmath3 and a mobility of @xmath38 @xmath39/vs at 1.4 k. oscillations can be resolved down to 0.8 t , indicating the excellent quality of the sample .
fast fourier transformation , fft , spectra of sdh oscillations are shown in fig .
[ q1772ab - fft](a ) at various temperatures for sample a. the 2deg concentrations of the s - o split @xmath26 and @xmath27 subbands , are 0.80 and @xmath40 @xmath3 , respectively , which are constant , within experimental uncertainties , for temperatures up to at least 35 k. the amplitudes of the two peaks have similar temperature behavior which can be described by@xcite @xmath41 where @xmath42 from the temperature dependence of the sdh oscillation amplitudes , the effective electron mass at the fermi level , @xmath43 , was deduced to be @xmath44 and @xmath45 for samples a and b , respectively , where @xmath46 is the free electron mass .
these values are in good agreement with calculated effective electron masses of 0.049 and @xmath47 , respectively . beating patterns in the sdh oscillations
are observed when @xmath48 t. in the presence of significant broadening of the ll s , the amplitude of the beat frequency will have a maximum in the vicinity of the intersection of two ll s .
a node between two maxima will appear where only one ll is present i.e. , @xmath49 with @xmath50 , where @xmath51 is the total spin splitting and @xmath52 is the landau level splitting.@xcite the three observable quantum hall plateaus directly below the node at 5.35 t correspond to even filling factors , whereas the three above correspond to odd filling factors .
this node is due to the crossing point at @xmath53 .
[ t ] sample b has a higher electron concentration due to deposition of an insulating layer and metallic gate electrode which results in a different work function between the semiconductor and surface . in fig .
[ q1772b - rxxdos ] the vertical arrows indicate the node positions of the sdh oscillations . total electron concentration from the fft of @xmath54 @xmath3 ( @xmath55 ) , shown in fig .
[ q1772ab - fft](b ) agrees well with the value of @xmath56 @xmath3 deduced from the low magnetic field hall coefficient . from the ratio of the magnetic field strengths of all observed nodes
, it has been determined that the node at 4.25 t for sample b corresponds to @xmath57 .
up to 12 beating nodes were observed in the magnetic field between 0.9 t and the highest field of 7.0 t. the second conduction subband , @xmath58 , is also occupied ; however , the expected weak splitting of this primarily @xmath59-like state of @xmath60 mev was less than the experimental resolution . the total spin splitting energy , @xmath51 , deduced from the node position in the beating patterns of the sdh oscillations is shown in fig .
[ delta - hwc ] as a function of landau splitting energy , @xmath52 . when the ll s from the @xmath21 subband intersect at or near the chemical potential as shown for sample a in fig .
[ ll - q1772a ] , a maximum in the amplitude of the beat frequency occurs .
@xmath51 can be determined from the intersection according to @xmath61 the two crossing points in fig .
[ ll - q1772a ] correspond to a @xmath62 of 2 and 3 .
the change in landau quantum number for all pairs of ll s which intersect at or near the chemical potential is given by 1 , 2 , 3 , 4 , 5 @xmath63 . even though the ll s with @xmath64 do intersect ,
they do so further removed from the chemical potential than the subsequent series of ll pairs . in order to increase the number of theoretical data points , the energy difference between appropriate ll s was employed when one ll was below the chemical potential and the other above .
these values are in excellent agreement with those obtained from the intersection of ll s . a similar series of ll s crossing points exist for sample b , and the analysis of the ll s in the vicinity of the chemical potential described above also resulted in a consistent set of data .
[ t ] values of @xmath51 obtained from the intersection of ll s and a least square fit all theoretical data for both samples are plotted as a function of @xmath52 in fig .
[ delta - hwc ] together with the experimental results .
obviously theory and experiment are in very good agreement with the exception of the @xmath65 node in the amplitude of the beat frequency for sample a , which corresponds to ll s with small filling factors .
the calculated rashba s - o splitting energies , @xmath7 , for sample b are 31.5 and 29.1 mev for the in - plane @xmath66 and @xmath67 vectors at the fermi surface , respectively , see fig .
[ delta - ek0 ] .
similarly the values for sample a are 27.5 and 25.4 mev , respectively .
@xmath7 values averaged over @xmath68 space of 26.5 and 30.4 mev for samples a and b are in good agreement with the experimentally determined total s - o splitting energies of @xmath69 and @xmath70 mev , respectively .
the experimental splitting is due to the rashba s - o effect , which results in the large population difference of 14 and 14.7 % for samples a and b , respectively , shown in the fft spectra in fig .
[ q1772ab - fft ] .
[ t ] the rashba s - o splitting energy of up to @xmath71 mev in these hgte qw s is almost one order of magnitude larger than the previously reported values in iii - v materials .
this is due to the unique band structure of the hgte system and in particular the inverted band structure .
this value is also larger than previously reported values by schultz _
et al . _
@xcite and zhang _
et al._@xcite for hgte based qw s .
this is mainly due to a larger 2deg concentration in the @xmath21 subband in the present qw s and the larger structure inversion asymmetry . in order to compare the results of our self - consistent hartree calculations with the measured longitudinal resistance ,
we have employed the following relationship to calculate the density of states , dos , from the landau level structure in the lowest order cumulant approximation according to gerhardts;@xcite @xmath72^{-1/2 } \exp \left[{-2\frac{(e_f-\varepsilon_n^\pm)^2}{\gamma _ n^2 } } \right]}\ ] ] here @xmath73 are the landau level energies which are the result of our self - consistent hartree calculations .
@xmath74 is the usual magnetic length , and @xmath75 is the landau level broadening and assumed to be a constant . the experimental sdh oscillations and the numerical simulations of the dos by means of eq . [ eq4 ] for sample
a and b are reproduced in figs .
[ q1772a - rxxdos ] and [ q1772b - rxxdos ] , respectively .
the calculated fermi energy was modified less than 1 % in order to align the sdh oscillations .
the best fit was obtained using @xmath76 mev for sample a , and @xmath77 mev for sample b.
in conclusion , the beating patterns in the sdh oscillations of modulation doped hgte / hg@xmath0cd@xmath1te qw s have been analyzed .
the s - o splitting energy which has been directly determined from the node positions to be as high as 30 mev , is almost one magnitude higher than that in ingaas heterostructures with similar carrier densities .
self - consistent hartree calculations based on an @xmath78 hamiltonian have demonstrated that the experimental zero field splitting energies are due to rashba s - o splitting .
furthermore good agreement between experimental sdh oscillations and calculated dos is evidence that the rashba term is the dominant mechanism of giant s - o splitting in hgte qw s with an inverted band structure .
this large @xmath7 in hgte qw s with an inverted band structure is caused by its narrow gap , the large spin - orbit gap between the bulk valence bands @xmath79 and @xmath80 and the heavy hole character of the first conduction subband .
furthermore , our calculations show that the method of directly deducing s - o splitting from node positions in sdh oscillations , is applicable even for a system with a strongly nonparabolic band structure .
the financial support of the major state basic research project no .
g001cb3095 of china , national natural science foundation of china ( no . 10374094 ) , the deutsche forschungsgemeinschaft ( sfb410 ) , bmbf , and the darpa spins program are gratefully acknowledged . | we have investigated beating patterns in shubnikov - de haas oscillations for hgte / hg@xmath0cd@xmath1te(001 ) quantum wells with electron densities of 2 to 3 @xmath2 @xmath3 . up to 12 beating nodes have been observed at magnetic fields between 0.9 and 6 t. zero magnetic field spin - orbit splitting energies up to @xmath4 mev have been directly determined from the node positions as well as from the intersection of self - consistently calculated landau levels .
these values , which exceed the thermal broadening of landau levels , @xmath5 , at room temperature , are in good agreement with rashba spin - orbit splitting energies calculated by means of an @xmath6 kane model .
the experimental shubnikov - de haas oscillations are also in good agreement with numerical simulations based on this model . |
asymptotic giant branch ( agb ) carbon stars are produced following @xmath1 dredge - up in thermally pulsing stars ( e.g. iben & renzini 1983 ) .
the star changes from oxygen- to carbon - rich when sufficient carbon has been mixed - in with the stellar mantle to yield an abundance ratio c / o@xmath2 .
the chemistry of the dust in the circumstellar envelope ( cse ) changes accordingly .
the change occurs at smaller core - mass or lower luminosity for lower metallicity stars .
clear evidence for this comes from observations of clusters in the large magellanic cloud ( lmc ) that contain both carbon and m - type stars ( lloyd evans 1984 ; marigo et al .
1996 ) .
surprisingly , silicate emission from oxygen - rich dust was discovered in the iras low resolution spectra of several galactic carbon stars ( little - marenin 1986 ; willems & de jong 1986 ) .
willems & de jong interpreted these `` silicate carbon stars '' as direct evidence for a fast transition of m - type agb stars into carbon stars , but timescales of decades for the silicate emission from an expanding detached oxygen - rich cse to fade away are difficult to reconcile with the lifetimes of silicate carbon stars ( little - marenin et al .
1987 ; le bertre et al .
hence the oxygen - rich material must be stored in a stationary component .
many galactic silicate carbon stars are @xmath3c - enhanced , j - type , carbon stars ( lambert et al .
1990 ) . unlike genuine , n - type , carbon stars that form on the agb , j - type carbon stars
are thought to have become carbon - enriched as a result of binary evolution .
the presence of a mass - losing oxygen - rich companion star has been ruled out observationally for a number of galactic silicate carbon stars ( noguchi et al .
1990 ; engels & leinert 1994 ) .
the presently most supported explanation for the silicate carbon star phenomenon is that of a keplerian disk of oxygen - rich material , surrounding a binary including a faint companion ( lloyd evans 1990 ) . the oxygen - rich dust may originate from mass loss at a time when the carbon star was still oxygen rich ( lloyd evans 1990 ) .
the dust - enshrouded agb star iras04496@xmath06958 was recently discovered to be a luminous carbon star in the lmc by van loon et al .
( 1998 , 1999 ) on the basis of ground - based ( ctio ) 3 @xmath4 m spectroscopy , after having been selected and confirmed to be an agb star by loup et al .
( 1997 ) and zijlstra et al.(1996 ) , respectively .
the carbon star nature of this object has been confirmed by groenewegen & blommaert ( 1998 ) using optical spectroscopy . at @xmath5 mag
it is the brightest known magellanic carbon star and very close to the maximum agb luminosity ( @xmath6 mag ) .
we here present compelling evidence for the presence of oxygen - rich dust close to this remarkable carbon star , making it the first known extra - galactic silicate carbon star . [
cols="<,<,<,<,<,<,<,<,<,<,<",options="header " , ]
the spectral energy distribution ( sed ) of this object peaks in the infrared . therefore in order to properly model the spectrum we obtained photometric and spectro - photometric observations with the european infrared space observatory ( iso , see kessler et al .
1996 ) , using the isocam ( cesarsky et al .
1996 ) and isophot ( lemke et al .
1996 ) instruments .
the photometric observations at 12 @xmath4 m ( using isocam filter lw10 ) and 25 @xmath4 m ( using isophot ) were obtained on april 22 , 1996 .
the 12 @xmath4 m observation was done using @xmath7 pixels , and a total on - source integration time of 50 s split in 25 two - s integration intervals .
the 25 @xmath4 m isophot observation was done using the p2 detector , a @xmath8 aperture , and triangular chopping with a chopper throw of @xmath9 .
the on - source integration time was 64 s. the 60 @xmath4 m photometry was obtained using the phot - c100 detector using two different methods .
one observation ( april 22 , 1996 ) was done using chopping mode with triangular chops and a chopping angle of @xmath10 and an on - source integration time of 64 s. another observation ( april 1 , 1998 ) was done with a @xmath11 raster map with @xmath12 raster steps and an integration time per pointing of 128 s , giving an effective on - source integration time of @xmath13 s. spectro - photometric observations of the source were obtained with isophot - s and the isocam cvf . the phot - s spectrum ( april 22 , 1996 )
was done in staring mode with an on - source integration time of 512 s. two isocam cvf spectra were obtained . the first isocam spectrum ( june 5 , 1997 ,
hereafter `` spectrum a '' ) spans the wavelength range from 7 to 14 @xmath4 m , the second ( april 1 , 1998 , hereafter `` spectrum b '' ) from 5 to 17 @xmath4 m .
the @xmath14 pixel field of view was used , with an integration time per spectral point of 50 and 70 s , respectively .
the data was processed using standard processing routines in the phot interactive analysis ( pia ) and cam interactive analysis ( cia ) software .
the cam - cvf spectra were constructed using a @xmath11 pixel@xmath15 software aperture and applying a correction for the wavelength dependence of the point spread function .
we corrected the phot - s spectrum for the background as derived from the cam - cvf data , accounting for the annual modulation of the zodiacal light using cobe / dirbe weekly all - sky maps ( see also trams et al .
the photometric observations are listed in table 1 .
the iso observations are supplemented with ground - based j , h , k , and l - band observations made at the south african astronomical observatory ( saao ) , interpolated to the same epochs as the various iso observations .
the spectra are presented in fig
. 1 . also plotted is a spectrum around 3 @xmath4 m obtained at ctio ( van loon et al .
1999 ) , after scaling to match the approximate continuum level in the phot - s spectrum .
we believe that the phot - s spectrum longward of @xmath16 @xmath4 m has been under - estimated , and possibly distorted , due to difficulties in determining the stabilised signal at such low flux density levels .
the phot - s and ctio spectra show the strong 3 @xmath4 m feature from hcn and c@xmath17h@xmath17 ( fig . 1 ) , but the long wavelength part of the phot - s spectrum is rather noisy .
the cam - cvf spectra , however , show a prominent emission feature between 9 and 12 @xmath4 m with a small dip near 11 @xmath4 m . for comparison we also plot in fig .
2 the iso sws spectrum ( @xmath18 ) of the galactic silicate carbon star v778 cyg , taken from yamamura et al .
( 1997 ) , which shows strong silicate emission from oxygen - rich dust around 10 @xmath4 m .
the feature in iras04496@xmath06958 extends to longer wavelengths than in v778 cyg , and closely resembles that of another galactic silicate carbon star , cs1003 ( hen 83 , iras08002@xmath03803 ; see little - marenin 1986 and willems & de jong 1986 ) .
we also plot in fig .
2 the ground - based ukirt spectrum ( @xmath19 ) of the galactic carbon star afgl2368 , taken from speck et al .
( 1997 ) , which shows a prominent silicon carbide ( sic ) emission feature around @xmath20 m that is common in carbon stars ( see e.g. little - marenin 1986 ; yamamura et al.1997 ) . the shape of the 9 - 12 @xmath4 m emission feature in iras04496@xmath06958 may be explained by assuming that the feature is a composition of the silicate and sic features .
alternative explanations include large silicate grains ( forrest et al .
1975 ; papoular & pgouri 1983 ) , crystalline olivines ( koike et al . 1981 ) and corundum ( alo ) grains ( onaka et al . 1989 ) .
we note that similarly shaped emission is observed in the spectra of a wide variety of objects : the s star rt sco and ms stars ( little - marenin & little 1988 ) , the sil@xmath21 and sil@xmath22 classes of m - type mira variables ( little - marenin & little 1990 ) , @xmath23 pictoris ( knacke et al .
1993 ; fajardo - acosta & knacke 1995 ) , inter - planetary dust particles ( sandford 1988 ) and comet halley ( campins & ryan 1989 ) .
absorption against the photosphere by hcn and c@xmath17h@xmath17 is seen at 3.1 , 3.8 and 8 @xmath4 m ( fig . 1 ) . on the long - wavelength end of the emission feature the 13.7 @xmath4 m absorption due to c@xmath17h@xmath17
is seen , which is commonly observed in the spectra of galactic carbon stars ( yamamura et al .
this absorption is seen against the dust continuum that dominates over the photospheric continuum at these long wavelengths indicating that the molecules are abundant throughout the dusty cse .
it is absent in v778 cyg .
this may be understood if the molecules - to - dust ratio is larger at lower metallicity , possibly because depletion of molecules into dust grains is less severe at smaller dust - to - gas ratios .
iras04496@xmath06958 is a long period variable with a period of @xmath24 d and a k - band amplitude of @xmath25 mag ( whitelock et al .
, in preparation ) .
the two cvf spectra are taken at different phases in the lightcurve , with spectrum a closer to maximum light .
their ratio is plotted in fig .
the maximum difference is reached between 9 and 10 @xmath4 m , and no difference is seen for wavelengths @xmath26 @xmath4 m .
this suggests that a significant part of the variability around 10 @xmath4 m is due to a variable emission feature , whilst beyond 13 @xmath4 m non - variable dust continuum emission dominates .
the cvf ratio around 10 @xmath4 m is 1.2 , which equals the ratio of l - band flux densities at these two epochs ( the k - band flux density ratio is 1.5 ) .
this may be compared to n- ( 10 @xmath4 m ) and l - band amplitudes observed in galactic ir - bright carbon stars , @xmath27 mag ( standard deviation 0.23 ) ( le bertre 1992 ) , and oxygen stars , @xmath28 mag ( standard deviation 0.20 ) ( le bertre 1993 ) .
the difference results from the fact that in carbon stars the l - band includes variable hcn+c@xmath17h@xmath17 absorption , whereas in oxygen stars the n - band includes variable silicate emission .
the @xmath29 mag of iras04496@xmath06958 suggests a contribution of silicate emission to the variability in the n - band .
the 10 @xmath4 m variability of iras04496@xmath06958 is additional evidence for the silicate carbon star nature of this star , and 10 @xmath4 m variability might provide a new means for finding or confirming silicate carbon stars .
all ir colours between 1 and 25 @xmath4 m of iras04496@xmath06958 are similar to those of carbon stars ( van loon et al .
1998 ; trams et al . 1999 ) , whereas v778 cyg has colours ( chen et al .
1998 ) more similar to oxygen - rich stars .
hence the oxygen - rich dust component represents only a minor fraction of the total dust mass that is contained in the cse of iras04496@xmath06958 .
its cse is considerably thicker than that of known galactic silicate carbon stars , judged from its very red near - ir colours ( lloyd evans 1990 ; chan & kwok 1991 ) .
hence the observation that all galactic silicate carbon stars are of j - type ( lambert et al . 1990 ) may be an observational bias against n - type carbon stars with massive carbon - rich cses for which an optical spectrum to determine the @xmath3c/@xmath30c ratio is relatively difficult to obtain .
[ [ the - origin - of - the - oxygen - rich - dust - around - the - carbon - star - iras04496 - 6958 ] ] the origin of the oxygen - rich dust around the carbon star iras04496@xmath06958 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ iras04496@xmath06958 is special because with @xmath31 mag it is a very luminous carbon star ( van loon et al . 1998 , 1999 ) .
the popular scenario for the formation of silicate carbon stars in which a j - type carbon star evolves from a less massive r - type star ( lambert et al .
1990 ) implies that silicate carbon stars should not be very luminous , which has some observational support ( barnbaum et al .
therefore , iras04496@xmath06958 could be a different class from the galactic silicate carbon stars .
nuclear burning at the base of the convective envelope ( `` hot bottom burning '' , hbb ) reduces the carbon abundance of the mantle by cycling it into nitrogen ( iben 1981 ; iben & renzini 1983 ; wood et al .
theoretical models show that it occurs for the most massive agb stars ( blcker & schnberner 1991 ) .
boothroyd et al .
( 1993 ) predict that hbb prevents the occurence of carbon stars above m@xmath32 mag , consistent with the observed absence of optically bright carbon stars more luminous than m@xmath33 mag ( iben 1981 ; costa & frogel 1996 ) .
the existence of luminous dust - enshrouded carbon stars ( van loon et al . 1999 ) like iras04496@xmath06958 in the lmc and iras00350@xmath07436 in the smc ( @xmath34 mag ; whitelock et al .
1989 ) is explained by mass loss reducing the stellar mantle to below a critical mass required for the pressure and temperature at the lower convective boundary to be sufficiently high to support hbb ( boothroyd & sackmann 1992 ) .
if such a star experiences another thermal pulse and accompanying dredge - up of carbon to its surface , it may become a carbon star , after all ( frost et al .
1998 ; marigo et al .
1998 ) . hence , iras04496@xmath06958 has been an oxygen - rich star not longer than a thermal pulse interval of @xmath35 yr ago ( see vassiliadis & wood 1993 ) .
this does not exclude the possibility that emission from the oxygen - rich dust is still observable around the recently formed carbon star , but the massive carbon - rich cse around iras04496@xmath06958 suggests a relatively long lapse of time since the mass loss was oxygen rich . although the isophot 60 @xmath4 m photometry is rather inaccurate , the high 60 @xmath4 m flux density of iras04496@xmath06958 suggest that its mass - loss rate was higher in the past , some @xmath36 yr ago , which would be consistent with an episode of increased mass loss during a thermal pulse followed by a considerable period of mass loss at a more moderate rate . hence it remains to be seen whether the silicate carbon star nature of iras04496@xmath06958 requires a companion star to have captured the oxygen - rich material in a circumbinary disk , or whether it resulted from single star evolution of a massive agb star .
barnbaum c. , kastner j.h . , morris m. , likkel l. , 1991 , a&a 251 , 79 blcker t. , schnberner d. , 1991 , a&a 244 , l43 boothroyd a.i .
, sackmann i .-
, 1992 , apj 393 , l21 boothroyd a.i .
, sackmann i .- j .
, ahern s.c . , 1993 ,
apj 416 , 762 campins h. , ryan e. , 1989 , apj 341 , 1059 carter b.s . , 1990 , mnras 242 , 1 cesarsky c.j .
, abergel a. , agnse p. , et al . , 1996 ,
a&a 315 , l32 chan s.j .
, kwok s. , 1991 , apj 383 , 837 chen p .- s . ,
xiong g .- z . , wang x .- h . , 1998 , acta astron . sin .
39 , 202 costa e. , frogel j.a . , 1996 , aj 112 , 2607 engels d. , leinert ch . , 1994 , a&a 282 , 858 fajardo - acosta s.b . , knacke r.f
, 1995 , a&a 295 , 767 forrest w.j . ,
gillett f.c . ,
stein w.a . , 1975 ,
apj 195 , 423 frost c.a . , cannon r.c . , lattanzio j.c
. , wood p.r . , forestini m. , 1998 , a&a 332 , l17 groenewegen m.a.t . , blommaert j.a.d.l . , 1998 , a&a 332 , 25 iben i. , 1981 , apj 246 , 278 iben i. , renzini a. , 1983 , ara&a 21 , 271 kessler m.f . , steinz j.a . , anderegg m.e . , et al . , 1996 ,
a&a 315 , l27 knacke r.f . ,
fajardo - acosta s.b . ,
telesco c.m . , et al . , 1993 , apj 418 , 440 koike c. , hasegawa h. , asada n. , hattori t. , 1981 , ap&ss 79 , 77 lambert d.l . ,
hinkle k.h . , smith v.v . , 1990 , aj 99 , 1612 le bertre t. , 1992 , a&as 94 , 377 le bertre t. , 1993 , a&as 97 , 729 le bertre t. , deguchi s. , nakada y. , 1990 , a&a 235 , l5 lemke d. , klaas u. , abolins j. , et al .
, 1996 , a&a 315 , l64 little - marenin i.r .
, 1986 , apj 307 , l15 little - marenin i.r .
, little s.j . , 1988 ,
apj 333 , 305 little - marenin i.r . , little s.j . , 1990 , aj 99 , 1173 little - marenin i.r . ,
benson p.j . ,
dickinson d.f . , 1987 ,
apj 330 , 828 lloyd evans t. , 1984 , mnras 208 , 447 lloyd evans t. , 1990 , mnras 243 , 336 loup c. , zijlstra a.a . , waters l.b.f.m . ,
groenewegen m.a.t . , 1997 , a&as 125 , 419 marigo p. , girardi l. , chiosi c. , 1996 , a&a 316 , l1 marigo p. , bressan a. , chiosi c. , 1998 , a&a 331 , 564 noguchi k. , murakami h. , matsuo h. , et al . , 1990 , pasj 42 , 441 onaka t. , de jong t. , willems f.j . , 1989 , a&a 218 , 169 papoular r. , pgouri b. , 1983 , a&a 128 , 335 sandford s.a .
, 1988 , in : dust in the universe , eds .
bailey & d.a .
williams , cambridge university press , p193 speck a.k . ,
barlow m.j .
, skinner c.j . , 1997 ,
mnras 288 , 431 trams n.r . ,
van loon j.th .
, waters l.b.f.m . , et al
. , 1999 , submitted to a&a van loon j.th .
, zijlstra a.a . ,
whitelock p.a.w . ,
et al . , 1998 , a&a 329 , 169 van loon j.th . , zijlstra a.a .
, groenewegen m.a.t . , 1999 ,
a&a in press vassiliadis e. , wood p.r . , 1993 ,
apj 413 , 641 whitelock p.a . ,
feast m.w . , menzies j.w . ,
catchpole r.m . , 1989 ,
mnras 238 , 769 willems f.j .
, de jong t. , 1986 , apj 309 , l39 wood p.r . ,
bessell m.s . ,
fox m.w . , 1983 ,
apj 272 , 99 yamamura i. , de jong t. , justtanont k. , cami j. , waters l.b.f.m . , 1997 , in : first iso workshop on analytical spectroscopy , eds .
a.m. heras , k.j .
leech , n.r .
trams & m. perry , esa sp-419 , p313 zijlstra a.a .
, loup c. , waters l.b.f.m . , et al . , 1996 , mnras 279 , 32 | we describe iso observations of the obscured asymptotic giant branch ( agb ) star iras04496@xmath06958 in the large magellanic cloud ( lmc ) .
this star has been classified as a carbon star .
our new isocam cvf spectra show that it is the first carbon star with silicate dust known outside of the milky way .
the existence of this object , and the fact that it is one of the highest luminosity agb stars in the lmc , provide important information for theoretical models of agb evolution and understanding the origin of silicate carbon stars . |
merging the exquisite tunability of electronic nanostructures with ferromagnetic materials in nanospintronic devices bears great potential for applications and fundamental investigations .
electronic devices using the electron spin in magnetic field sensing are very successful , for example in hard disks of computers .
however , to use the electron spin directly , for example in a spin - transistor @xcite or as quantum bits @xcite , it is necessary to fabricate nanostructures with the required long coherence times and electrical tunability .
carbon nanotubes ( cnts ) and graphene are in principle ideally suited for spintronic devices due to the large intrinsic coherence times , tunable electron density and large maximum current densities .
early electrically tunable spin valves on carbon nanotubes @xcite , or nonlocal spin - accumulation experiments on graphene @xcite demonstrate the great potential of carbon based nanostructures . to obtain an electrically tunable spin signal ,
one strategy is to fabricate a nanostructure with a gate - tunable conductance , e.g. a quantum dot ( qd ) @xcite .
however , this electrical tunability introduces additional complexity to the data analysis , since the signal now depends on the position , amplitude and broadening of a conductance feature , which all can vary with the magnetisations of the contacts , as will be discussed in the results sections of this paper .
while in nonlocal measurements the spin signal can in principle be separated from the charge signal , this is difficult in two - terminal qd devices because both signals are detected at the same contacts . despite considerable efforts , spin injection and detection in qd spin valves are not yet reproducible enough for more complex experiments or applications , e.g. as detectors of electron spin entanglement @xcite .
this lack of reproducibility can have several reasons .
the most fundamental spin transport device is a spin valve with two ferromagnetic ( f ) contacts to a non - magnetic material in - between . in fig .
1a such a device is shown schematically with a carbon nanotube ( cnt ) between the f - contacts . ideally , the contacts are either magnetized parallel or anti - parallel to each other , adjustable by an external magnetic field .
the normalized difference between the electrical conductance ( or resistance ) of these two configurations is called magnetoresistance ( mr ) . here
we address two more technical problems .
the first is limitations in the device design .
compared to other carbon based nanoscale devices with normal metal @xcite or superconducting leads @xcite , the contact material has to be chosen from a very limited range of readily available and processable magnetic metals , which limits the optimization of the contacts .
in addition , most ferromagnetic materials form oxides when exposed to air , which diminishes the electrical contact yield . to obtain low - ohmic contacts with non - magnetic materials ,
one often chooses large contact areas , which , however , is in conflict with using narrow contact geometries to control the shape anisotropy and thus the magnetic field at which the magnetisation is reversed ( switching field ) @xcite .
even the thickness of the deposited material is limited to avoid the formation of vertical , more complex magnetic domains .
no adhesion or contact layer can be used because the equilibrium spin polarization decays very rapidly in non - magnetic metals ( on the scale of the exchange interaction , typically @xmath0 nm ) .
our choice of contact material is the well - studied ni@xmath1/fe@xmath2 alloy permalloy ( py ) , for which one can obtain single - domain contacts and control over the magnetic easy axis by the shape of the contacts @xcite .
we demonstrate that the same electrical and magnetic characteristics for sub - micrometer scale py contacts can be obtained by sputter deposition and for thermal evaporation .
this opens up the large field of magnetic multi - layer structures to be used in carbon based nano spintronic device fabrication , e.g. anti - ferromagnetic exchange - bias layers @xcite .
the second technical problem we propose a solution to is resist residues and the unwanted formation of py nanoparticles near and on top of the device , which can strongly alter the device characteristics . in nanospintronics
the interface area between the ferromagnetic contact and the non - magnetic structure , for example a cnt , is usually very small , which makes it very susceptible to resist residues .
this not only compromises the spin and charge transport properties , but also the electrical stability of a device due to dielectric charge traps . here
we report the fabrication of cnt spin valves by electron beam lithography using an essentially residue free ( on sio@xmath3 ) low - density polymer that allows the fabrication of optimal polymer mask cross sections without resorting to multiple resist layers . using this recipe
we obtain electrical contacts with a significantly increased electrical stability and yield , for both , thermal and sputter deposition of py .
we present measurements from two devices to discuss the need for an extended data analysis in nanospintronic devices .
our cnts are grown by chemical vapor deposition at a temperature of @xmath4c using methane as source gas and fe / ru catalyst nano particles @xcite .
the substrate is a heavily doped si wafer acting as a backgate and a @xmath5 nm thermal oxide top layer .
as shown schematically in fig .
1a our approach to obtain reproducible magnetic domains and switching characteristics for the ferromagnetic contacts is to fabricate @xmath6 nm thick ferromagnetic permalloy ( py ) strips with a large aspect ratio ( @xmath7 ) @xcite .
these strips are fabricated by electron beam lithography with an electron sensitive polymer resist , followed by metal deposition and a lift - off procedure .
to deposit py we use two techniques : 1 ) thermal evaporation of py by an electron gun in a uhv chamber at a base pressure of @xmath8 mbar , sample cooling to @xmath9c and a deposition rate of @xmath10 / s .
2 ) dc sputter deposition using an ar plasma at the power of @xmath11w and an ar pressure of @xmath12 mbar in a uhv chamber with a base pressure of @xmath8 mbar .
the fabrication of nanostructures by sputter deposition is often difficult because the sputtered material is scattered at gas particles in the chamber , which leads to a large angular spread that can fill the lithographically defined polymer trench and lead to lift - off problems .
we obtain sufficient directionality for the sputter deposition of py by working with a relatively low ar pressure and no sample rotation .
the sample resides directly above the py target at a distance of @xmath13 cm from the plasma at room temperature .
we use a deposition rate of @xmath14 / s .
we have systematically investigated the morphology and magnetic properties of py strips fabricated by electron beam lithography with different resist systems and beam acceleration voltages and identify two fundamental problems rather specific to nanospintronic devices : 1 ) py nanoparticles form at the side walls of polymer structures with insufficient under - cuts and are deposited nearby or on the py strips in the lift - off procedure , 2 ) resist residues lead to a significant decrease in the yield of obtaining low - ohmic electrical contacts to cnts ( with @xmath15m@xmath16 two - terminal resistance at room temperature ) .
both problems are illustrated in scanning electron microscopy ( sem ) images in fig .
subfigure ( i ) shows a tilted side view of a @xmath17 nm thick poly(methyl methacrylate ) ( pmma ) mask of a strip after lithography and thermal py deposition .
the py strip forms at the bottom of the polymer trench . however , due to the large beam acceleration voltage of @xmath18kv used for this structure , the polymer trench is v - shaped with a thin metal film deposited also on the side walls , which often leads to a bad lift - off and large ferromagnetic residues .
subfigure ( ii ) shows a top view of the resulting py strip . while the strip appears well defined , we reproducibly find a large number of py nanoparticles on top and around the strip , as indicated by the black arrow .
such particles can be magnetic with very large characteristic fields , leading to seemingly non - symmetric low - field mr curves .
the polymer profile can be improved significantly by either reducing the sem acceleration voltage , e.g. to @xmath19kv , or by using an additional more sensitive resist layer , which both lead to a larger under - cut .
we have tested the copolymer system pmma / ma ( ma : methacrylic acid ) and pmma(950k)/pmma(50k ) with different pmma chain lengths .
all three methods can be optimised to obtain better undercuts and a significantly reduced number of py particles on the surface . from this finding
we conclude that the particles form at the side walls of v - shaped profiles , which is therefore an essentially geometric effect and independent of the polymer . the red ( bright ) arrow in subfigure ( ii ) of fig .
1b points out polymer residues , which we identify by the smaller sem contrast and the much shorter oxygen dry etching times required to remove them in test samples ( not shown ) .
we reproducibly find polymer residues for trenches fabricated in the polymer systems pmma , pmma / ma @xcite and pmma(950k)/pmma(50k ) , with methyl isobutyl ketone ( mibk ) and ipa ( 1:3 ) as developer and lift - off in warm acetone .
this is a well - known issue not specific to nanospintronic devices @xcite . in semiconductor device fabrication , or
when contacting metallic parts of the device , the residues can be removed before the next metal deposition by standard cleaning procedures like oxygen plasma etching or ar sputtering .
however , most of these procedures also remove or damage the cnt and graphene parts of a device .
we note that also the post - deposition structuring of py films frequently used for the fabrication of nanometer scaled magnetic devices , e.g. ion milling or chemical etching , also remove carbon with a large rate .
a close to ideal polymer mask cross section and negligible residues can be obtained using the copolymer resist zep 520a ( zep ) @xcite and @xmath20s development in n - amylacetate , stopped in a 9:1 solution of mibk and ipa , followed by rinsing in ipa .
after the metalization , a good lift - off is achieved in a @xmath21min n - methyl-2-pyrrolidone ( nmp ) bath at @xmath22c , followed by @xmath18min in acetone at @xmath23c and rinsing in ipa .
we use a @xmath24 nm thick zep layer , an electron acceleration voltage of @xmath25kv and a typical dose of @xmath26c/@xmath27 , for which we obtain undercuts with a narrow opening at the top of the polymer film , as demonstrated in fig . 1c ( i ) .
this undercut can be tuned systematically by the dose and acceleration voltage .
subfigure ( ii ) shows a resulting py strip obtained by thermal py evaporation .
we find no metallic particles or metal flakes and could not detect any resist residues .
we obtain similarly clean strips with a slightly increased surface roughness using the zep recipe and sputter deposition of py , as demonstrated in fig .
1d .
we fabricate long ( @xmath28 m ) , thin ( @xmath6 nm ) py strips with a small width @xmath29 , which forces the magnetisation of the ferromagnetic contacts to lie along the strip axis .
the direction can be inverted by an external magnetic field along the axis that switches the magnetisation to the opposite orientation at a characteristic switching field @xmath30 tunable by the width @xmath29 of the strip @xcite . to assess the magnetic properties and material quality of an individual py strip ,
it is contacted by pd contacts to measure the anisotropic magnetoresistance ( amr ) .
an example curve is shown in the inset of fig . 2 , where the resistance @xmath31 of a @xmath32 nm wide strip of sputtered py is plotted as a function of the external magnetic field @xmath33 along the strip axis .
sharp characteristic changes in the resistance at @xmath34mt indicate the reversal of the magnetisation @xcite .
the smooth background variation we attribute to a small ( @xmath35 ) misalignment between the field and the strip axis , which mixes in the large continuous mr signals obtained when the field is applied perpendicular to the strip axis .
figure 2 shows the switching fields @xmath36 as a function of @xmath29 for strips obtained by different fabrication techniques .
we find that the sputtered and thermally evaporated contacts defined using zep exhibit the same dependence on @xmath29 as the pmma processed and thermally evaporated py contacts @xcite .
the switching fields can be distinguished reliably for widths @xmath37 nm , for which @xmath36 increases strongly for smaller @xmath29 .
while the amr curves of individual py strips are very reproducible , the resulting mr in a spin valve are more problematic , as will be discussed below .
we note already here that amr experiments are sensitive to the bulk of the material , while in spin valve configurations the last few atomic layers are crucial .
in this section we demonstrate the need of extended data acquisition and analysis for magnetoresistance devices with non - trivial conductance characteristics .
the magnetoresistance ( mr ) of a spin valve device is defined in terms of its conductances @xmath38 and @xmath39 when the magnetisations in the two contact strips are either parallel ( p ) or anti - parallel ( ap ) .
similar expressions are easily obtained using the device resistances . here
we define @xmath40 which is more symmetric than the usual definitions and leads to smaller mr values of maximally @xmath41% .
this definition is more adequate for our purpose because it provides an equal measure for positive and negative mr . in a qd spin valve ,
the conductance depends on the gate voltage , which tunes the qd level energies .
the charging energy and level separation lead to characteristic coulomb blockade ( cb ) conductance maxima , with a strongly reduced conductance in between . in the color scale image in fig .
3 the qd conductance is plotted as a function of the backgate voltage @xmath42 and an increasing magnetic field @xmath33 ( up - sweep ) for a qd fabricated with standard pmma - based lithography .
the base temperature in all experiments presented here is @xmath43mk .
the qd conductance has a maximum at @xmath44v and decays rapidly away from this value .
a cross section at constant magnetic field is plotted in white . when the magnetic field is increased from negative values beyond @xmath45 ,
a first sharp change ( @xmath46mt ) in the conductance pattern occurs at @xmath47mt , and another at @xmath48mt .
these fields correspond well to the contact switching fields of the two py strips . at @xmath49 the amplitude of the cb resonance increases by a factor of @xmath50 and the peak position shifts by about @xmath51mv , which corresponds to an energy shift of @xmath52ev or to almost the resonance width . while the amplitude of the cb resonance increases by almost a factor of 2 at @xmath49 starting at the low field side , it does not change at @xmath53 and
is reduced slightly only at higher fields . at @xmath53
the resonance position switches back roughly to the same gate position as for @xmath54 . in a standard magnetoresistance ( mr )
measurement the conductance is recorded as a function of @xmath33 alone , which corresponds to cross sections in fig . 3 at a fixed gate voltage .
two examples for slightly off - resonance voltages are shown on top and below the main figure : at a more negative gate voltage ( green dashed line , 2 ) we find a decrease in conductance for the anti - parallel magnetisations , @xmath55 , which corresponds to an increased resistance and a positive mr of @xmath56% .
off - resonance for a more positive gate voltage ( blue dashed line , 1 ) the mr at fixed voltage is negative , @xmath57% .
these large values are almost exclusively due to the large shift of the resonance position . in the simplest model by jullire for tunneling mr @xcite
one would expect @xmath58% when using @xmath59 for the tunneling polarizations in the two f - contacts @xcite .
these values rather correspond to the amplitude modulation ( @xmath60% ) than to the mr observed in cross sections .
we will discuss shifts of the conductance features in the mr in more detail in the next section and only point out that while the mr at @xmath49 and @xmath53 might be described by a simple spin valve model , the increase of @xmath61 for @xmath62 with respect to @xmath63 is more difficult to explain since it suggests a difference between the two parallel configurations , a phenomenon possibly related to the single switching behavior reported before @xcite .
the electrical stability and reproducibility of the qd spin valve signals is considerably improved for devices fabricated using the zep recipe introduced above .
we analyze in more detail the data shown in fig . 4 measured on a sample with sputtered py contacts . in fig .
4a the qd spin valve conductance @xmath61 is plotted for a large backgate voltage interval at a base temperature of @xmath64mk .
the cb peaks occur in groups of four consistent with the spin and valley degeneracy of a cnt orbital .
such a pattern suggests that the cnt segment forming the qd is relatively clean @xcite . from charge stability diagrams ( not shown ) we find the lever arm of the backgate to the qd @xmath65 , a charging energy of @xmath66mev and a level spacing of @xmath67mev .
we estimate the source , drain and backgate capacitances as @xmath68af , @xmath69af and @xmath70af . from the cb maxima of @xmath71 ,
the average broadening of the peaks @xmath72mev and using the breit - wigner form for resonant tunneling at low temperatures ( @xmath73 ) , we find for the tunnel couplings of the qd to source and drain @xmath74mev and @xmath75mev , which gives a relatively small coupling asymmetry of @xmath76 ( we chose the larger value as @xmath77 ) .
we now focus on the four cb peaks highlighted in fig .
4a by the red rectangle , which originate from the same four - fold degenerate qd orbital . in fig .
4b and 4c the qd conductance @xmath61 is plotted for this gate voltage interval and as a function of an external magnetic field @xmath33 along the py contact strips . in fig .
4b the field is increased from negative values , while in 4c it is decreased , starting from positive values .
the magnetisations were saturated at @xmath78mt before the respective sweep . in the up sweep in fig . 4b , @xmath61 is larger for @xmath79mt , which usually is identified as the anti - parallel configuration of the contact magnetisations .
in the down sweep the magnetisation switching occurs at negative fields and we find an increased conductance for @xmath80mt .
the variation between the absolute values of the switching fields in the up and down sweeps are compatible with the variation observed in the corresponding amr curves . in fig .
4d shows the up and down sweeps at a fixed backgate voltage , indicated by the dashed lines in figs .
4b and 4c .
we find a sharp switching of the conductance at the py strip switching fields , which corresponds to a mr of @xmath81% .
the mr is negative for all gate voltages , which we now discuss in more detail . in devices with a variable conductance ,
e.g. as a function of the backgate voltage , the origin of the mr signal can lie in changes of the width , position and amplitude of the conductance feature . in fig .
5a we plot the cb oscillations indicated in fig .
4a as a function of @xmath42 for the different magnetisation configurations .
the two parallel configurations lead to identical conductances , which demonstrates the reproducibility of both , the magnetic and electronic structures in the device .
the anti - parallel configuration , however , deviates significantly from the parallel .
the resulting mr vs @xmath42 curve is plotted in fig .
5b ( full red line ) .
the mr is negative for almost all backgate voltages and shows a mr modulation of @xmath82% on an offset of about @xmath83% .
the modulation is correlated with the gradient @xmath84 of @xmath61 , i.e. it is largest at the gate voltages where @xmath61 has the largest slopes , which suggests that the mr is caused mainly by a shift of the cb resonances . in the next step
we fit the data with multiple lorenzians to extract the amplitude , width and position of the individual cb peaks ( no background is subtracted ) .
the resulting parameters are plotted in figs .
5c - e for the third cb peak highlighted by an asterisk in fig .
compared to the parallel magnetisation configurations , the anti - parallel shows an increase in amplitude and width by @xmath85% and @xmath86% , respectively , and a shift of @xmath87mv , which corresponds to @xmath88ev or @xmath89% of the peak width .
we obtain similar values for the other cb peaks .
all peaks are shifted by the same absolute value within experimental errors .
the extracted parameters allow us to investigate the respective impact on the mr , for example by calculating the mr from the measured curve for the parallel magnetisations and a shifted curve in the anti - parallel case .
the result is plotted in fig .
5b as blue dashed line ( cor .
mr ) and has mr maxima at gate voltages where also @xmath61 has maxima , as expected if the shifts were corrected precisely enough .
the mr variation on this curve is only @xmath90% with a slightly smaller negative offset than in the original data .
we generally find a better electrical stability and larger contact yield for qds fabricated with the presented recipes .
we note , however , that also with this method we obtain samples that show gate - dependent or temporal charge rearrangements , which we tentatively attribute to surface impurities on the substrate .
nevertheless , these methods lead to devices with reproducible mr with a yield of @xmath91% , a clear improvement compared to previous methods that yielded useful devices only rarely ( @xmath92% ) .
a periodic modulation of the mr with the cb oscillations was observed already earlier and modeled by spin dependent effective tunnel rates @xcite . in this simple model
one can construct negative mr signals for a strongly asymmetric qd coupling to the contacts , which can lead to a negative offset for strongly overlapping resonances .
the change of the effective tunnel couplings at the switching fields could in principle also result in a change of the resonance widths .
however , this model requires strongly asymmetric tunnel couplings and predicts that the mr maxima occur near the conductance minima , both in contrast to our observations .
this model does not produce shifts in the cb resonance energies , either .
more elaborate models @xcite predict major contributions to the mr from shifts of the cb resonances in an effective magnetic field , caused either by spin dependent electron scattering at the qd - contact interfaces or by a spin - dependent renormalization of the qd energy levels .
characteristic for both mechanisms is that the sign of the shifts depends on the spin state of the cb resonance .
specifically , of the four states in a cnt orbital , two should be shifted in energy opposite to the other two .
none of the models predict identical shifts for all four peaks , nor a negative offset of the mr .
another mechanism that results in a constant shift in the anti - parallel configuration is the magneto - coulomb effect ( mce ) @xcite .
the opposite zeemann shifts and the different density of states at the fermi energy of the majority and minority bands leads to a rearrangement of electrons between the bands , which is compensated by a change in the electrical potential .
we estimate an mce shift of the qd resonances in an external magnetic field @xmath33 of @xmath93v / t . in the last step we used @xmath94 as an upper limit of the ( thermodynamic ) py polarization in both leads , the land @xmath95-factor in thick ( @xmath96 nm ) py films of @xmath97 @xcite and the bohr magneton @xmath98 . with the same parameters one obtains a total change in position of @xmath99v when sweeping the field beyond both switching fields . the negligible slope observed for the peak positions is consistent with the small value obtained in these estimates .
however , the same parameters also predict a negligible change at the switching fields .
we note that also the qualitative curve shape observed in the experiments does not follow the triangular characteristics at the switching fields of the mce @xcite .
a quite natural explanation of our experimental findings is that one py strip does not couple directly to the qd , but rather to an anti - ferromagnetically coupled contact area .
such effects could occur at the chemical bonds between the metal / cnt interface @xcite , or in oxidized layers of the magnetic material that are strongly coupled to the bulk .
the latter coupling could depend on the thickness of the oxide and explain why both contacts are not coupled identically to the cnt , a phenomenon well known from non - magnetic metal contacts to cnts .
this scenario explains the sign reversal of the mr gate modulation and offset , but not the peak shifts at the switching fields .
a physically more intriguing scenario would be a spatially varying spin susceptibility ( rkky interaction ) in the cnt segments connecting the qd @xcite , which could lead to a rotation of the injected spins depending on the distance from the ferromagnetic contact - an effect that can persist over large length scales due to the low dimension of the cnt and the strong electron - electron interactions .
we report a recipe for essentially residue free electron beam lithography , useful to improve the fabrication of carbon based nanospintronic devices .
we obtain very reproducible magnetic and electrical contact properties for sputtered and thermally deposited permalloy when both processes are optimised individually . using these recipes
, we obtain an improved yield of electrical contacts and more reproducible features in the mr of carbon nanotube quantum dot spin valves .
since the mr in nanospintronic devices not only depends on the magnetic orientation of the contacts , but also on the electrostatic environment ( e.g. gates ) , it is necessary to expand the standard magnetic field sweeps to three - dimensional maps that also contain a variable gate voltage to track the origin of the observed mr .
we demonstrate this idea with two devices and show that the major contribution to the mr can be due to a shift in the conductance features . from the discussion of several mechanisms specific to nanospintronic devices
we tentatively conclude that interface properties might be crucial to explain the presented magnetoresistance characteristics .
we gratefully acknowledge useful discussions with b. hickey and m. elkin .
this work is financially supported by the swiss national science foundation ( snf ) , the swiss nccr qsit and nccr nano , the erc project quest and the fp7 project se2nd . | we report an improved fabrication scheme for carbon based nanospintronic devices and demonstrate the necessity for a careful data analysis to investigate the fundamental physical mechanisms leading to magnetoresistance .
the processing with a low - density polymer and an optimised recipe allows us to improve the electrical , magnetic and structural quality of ferromagnetic permalloy contacts on lateral carbon nanotube ( cnt ) quantum dot spin valve devices , with comparable results for thermal and sputter deposition of the material .
we show that spintronic nanostructures require an extended data analysis , since the magnetisation can affect all characteristic parameters of the conductance features and lead to seemingly anomalous spin transport .
in addition , we report measurements on cnt quantum dot spin valves that seem not to be compatible with the orthodox theories for spin transport in such structures . |
lets consider that @xmath0 is a compact riemann surface .
then we can define a green s function with respect to a metric and a divisor on @xmath0 .
suppose that @xmath1 is a hyperbolic 3 - manifold with infinite volume and having compact riemann surfaces @xmath2 as its conformal boundaries at infinity . also choose a geometry on @xmath1 which bears a metric of constant negative curvature .
the main new results of this paper express the green s functions on each @xmath3 in terms of the length of some certain geodesics in @xmath1 .
manin has done this provided that @xmath1 has one boundary component ( i.e. n=1 ) and is uniformized by a schottky group in @xcite . in this work
we will generalize manin s results in general case , @xmath1 has more than one boundary components and is uniformized by a fuchsian , quasi - fuchsian or kleinian group .
+ in this introduction we state the main definitions , motivation and the plan of doing the work .
we start by introducing the definition of a green s function on a compact riemann surface for a divisor and with respect to a normalized volume form on it .
let @xmath0 be a compact riemann surface and @xmath4 ( with @xmath5 integer number ) be a divisor on @xmath0 .
we show the support of @xmath6 by @xmath7 . also lets consider that @xmath8 is a positive real - analytic 2 - form on @xmath0 . by a green s function on @xmath0 for @xmath6 with respect to @xmath8 we mean a real analytic function @xmath9 satisfying the following conditions : + ( i ) _ laplace equation : _
@xmath10 where @xmath11 , and @xmath12 is the standard @xmath13 - current @xmath14 .
+ ( ii ) _ singularities : _ let @xmath15 be a complex coordinate in a neighborhood of the point @xmath16 . then @xmath17 is locally real analytic . + a function satisfying these two conditions is uniquely determined up to an additive constant .
and the third condition is \(iii ) _ normalization : _ @xmath18 which eliminates the remaining ambiguous constant .
@xmath19 is additive on @xmath6 and for @xmath20 , @xmath21 is symmetric , i.e. @xmath22 .
+ lets consider that @xmath23 is another divisor on @xmath0 that is prime to the divisor @xmath6 .
that means , @xmath24 . and put @xmath25 @xmath19 is additive and @xmath21 is symmetric , then @xmath26 is symmetric and biadditive in @xmath27 .
+ in general , the function @xmath28 depends on the metric @xmath8 , but in the case that both of the divisors are of the degree zero , from the condition ( i ) we see that @xmath29 depends only on @xmath30 .
notice that , as a particular case of the general kahler formalism , to choose @xmath8 is the same as to choose a real analytic riemannian metric on @xmath0 compatible with the complex structure .
this means that @xmath31 are conformal invariants when both divisors are of degree zero .
also in the case that the divisor @xmath6 is principal , i.e. @xmath6 is the divisor of a meromorphic function like @xmath32 , then @xmath33 is a 1 - chain with boundary @xmath34 .
the divisors of degree zero on the riemann sphere @xmath35 are principal . then this formula can be directly applied to divisors of degree zero on riemann sphere . + it is well known that the green s function of the degree zero divisors on a riemann surface of arbitrary genus can be expressed exactly via the differential of the third kind @xmath36 with pure imaginary periods and residues @xmath5 at @xmath16 when the divisor is @xmath37(see , @xcite , @xcite )
. then the generalization of the previous formula for arbitrary divisors @xmath27 of degree zero is @xmath38 in general , when the degree of the divisors are not restricted to zero , the basic green s function @xmath39 can be expressed explicitly via theta functions ( as in @xcite ) in the case when @xmath40 is the _ arakelov metric _ constructed with the help of an orthonormal basis of the differentials of the first kind on the riemann surface .
the result of manin in @xcite which relates the arakelov green function on a compact riemann surface to configurations of geodesics in a 3-dimensional hyperbolic handlebody with schottky uniformization , having the riemann surface as conformal boundary at infinity , was extremely innovative and influential and had a wide range of consequences in the arithmetic context of arakelov geometry as well as and in other contexts , ranging from p - adic geometry , real hyperbolic 3-manifolds , the holography principle and ads / cft correspondence in string theory , and noncommutative geometry .
a natural question is to what extent the result of manin can be generalized to cases where , instead of dealing with a single riemann surface , one has several riemann surfaces whose union is the boundary of a hyperbolic 3-manifold , uniformized no longer by a schottky group , but by a fuchsian , quasi - fuchsian , or more general kleinian group .
such a generalization is not only interesting because it is a very natural question to pass from kleinian - schottky groups to more general kleinian groups , but also for its potential applications to arakelov geometry , to the case of curves defined over number fields with several archimedean places , while manin s result was formulated for the case of arithmetic curves defined over the rationales .
we have focused on the formula in manin s work , that expresses the arakelov green s function on a compact riemann surface in terms of a basis of holomorphic differentials of the first kind and of differentials of the third kind . in the case of schottky uniformization , when the limit set has hausdorff dimension strictly smaller than one , one can construct such differentials in terms of averages over the schottky group .
while the same type of formula no longer holds in the fuchsian or quasi - fuchsian case , we use the canonical covering map relating fuchsian and schottky uniformization and the coding of limit sets for the fuchsian and schottky case , to express the green function in the fuchsian or quasi - fuchsian case in terms of the one in the schottky case .
the approach we follow for the more general kleinian case is via a decomposition of the uniformizing group as a free product of quasi - fuchsian and schottky groups and applying the results we obtained for these cases individually .
the paper consists of three chapters devoted to the cases that the hyperbolic 3-manifold @xmath1 have 1 , 2 and @xmath41 many boundary components at infinity respectively .
first chapter pays to the one boundary component case and includes four subchapter devoted to the foundations and the genera 0,1 and @xmath42 respectively .
also this chapter contains fundamental definitions and basic computations that are bases for the computations in the next chapters .
the results of this chapter are a summary of the reference @xcite .
all hyperbolic 3-manifolds with two boundary components with the same genera are uniformized by quasi - fuchsian groups ; the second chapter is devoted to this manifolds . in third chapter
we consider the general case i.e. the case that @xmath1 is uniformized by some kleinian groups and have @xmath41 boundary components .
finally , as a remark for the chapter three we show that for manifolds with @xmath43 many boundary components at infinity the situation is like the previous .
i am grateful to matilde marcolli who suggested this problem , supported me in max - planck and hausdorff institute in bonn , and supervised me in all process .
i also many thank saad varsaie for several useful and encouraging comments .
in this chapter , first we gather some fundamental definitions and some useful notations that will be used in all of the paper .
next we bring the computation of green s function for riemann sphere @xmath35 , the boundary at infinity of the hyperbolic space .
all 3-manifolds with one boundary component that is a compact riemann surface with genus @xmath44 , are uniformized by schottky groups .
the rest of the chapter is devoted to these manifolds .
the results of this chapter are a summary of the reference @xcite .
consider the hyperbolic space @xmath45 with the upper half space model and coordinate @xmath46 that comes from @xmath47 equipped by the hyperbolic distance function corresponding to the metric @xmath48 of constant curvature -1 .
the geodesics in this model are vertical half - lines @xmath49 and also vertical half - circles orthogonal to the plane at infinity @xmath50 ( i.e. for @xmath51 ) .
+ if we consider the end points of the geodesics ( including @xmath43 for the ends of the vertical half - lines ) , then we can consider the riemann sphere @xmath52 ( or @xmath53 in the unit boll model ) as the boundary at infinity of @xmath45 .
lets show the geodesic joining @xmath54 to @xmath55 in @xmath56 by @xmath57 ; by @xmath58 the point on the geodesic @xmath59 closest to the point @xmath54 ( i.e , the intersection point of @xmath59 and a geodesic passing through @xmath54 and orthogonal to @xmath59 ) ; by @xmath60 the distance from @xmath61 to the geodesic @xmath57 and by @xmath62 the oriented distance between two points lying on an oriented geodesic in @xmath45 ( figure [ fig1.10 ] ) .
also lets show by @xmath63 the angle ( at @xmath61 ) between the semi - geodesics joining @xmath61 to @xmath54 and @xmath61 to @xmath55 , and for @xmath64 in @xmath65 by @xmath66 the oriented angle between the semi - geodesics joining @xmath58 to @xmath54 and @xmath67 to @xmath55 . in order to calculate
@xmath66 we must first make the parallel translation along @xmath59 identifying the normal spaces to @xmath59 at @xmath58 and at @xmath67 .
the orientation of a normal space is defined by projecting it along oriented @xmath59 to its initial end into the tangent space to @xmath65 , which is canonically oriented by the complex structure . in this case we consider that the hyperbolic manifold @xmath1 is the hyperbolic space @xmath45 with riemann - sphere as its boundary at infinity .
for this case we have [ p1.5 ] for @xmath68 and @xmath69 in @xmath35 , denote by @xmath70 a meromorphic function on @xmath35 with the divisor @xmath71 . then we have @xmath72 and @xmath73 mobius transformations preserve hyperbolic distance and angel
. then both sides of ( [ f1.3 ] ) and ( [ f1.4 ] ) are invariant under these transformations .
hence it suffices to consider the case when @xmath74 in @xmath35 .
then the geodesic @xmath75 is the @xmath76 semi - axis in @xmath77 coordinate for @xmath45 and the geodesics joining the points @xmath78 and @xmath79 normally to this semi - axis are half circles passing from these points with center in @xmath80 and normal to @xmath50 ( figure [ fig1.11 ] )
. then we have @xmath81 and @xmath82 also from the properties of cross - ratio we have @xmath83 this gives ( [ f1.3 ] ) directly . to see ( [ f1.4 ] ) , we know that angles in @xmath45 can be calculated using the euclidean metric @xmath84 and also the parallel transport along @xmath76 semi - axis coincides with the euclidean one .
as it s shown in figure [ fig1.11 ] the vectors @xmath85 and @xmath86 tangent to the geodesics from the points @xmath15 and @xmath87 , are normal to @xmath88 . hence if we transport ( parallel ) the vector @xmath86 in the point @xmath89 to the point @xmath90 then both vectors are in an euclidean plane normal to @xmath76 semi - axis
. then the angel from the vector @xmath85 to the vector @xmath86 can be calculated via there angels from @xmath91 axis .
then we have @xmath92 this proves ( [ f1.4 ] ) .
the part @xmath93 in the proposition above is the classical cross - ratio of four points on the riemann sphere @xmath35 , for which it is convenient to have a special notation @xmath94 let @xmath68 and @xmath69 be in @xmath35 .
then we have + @xmath95 + from the formula ( [ f0.11 ] ) for the green s function in the case that the divisors are principal , we have + @xmath96 + then from proposition ( [ p1.5 ] ) and notation ( [ f1.5 ] ) we have ( [ f1.6 ] ) . in this case
the group that uniformize the hyperbolic 3 - manifold with its boundary at infinity is a cyclic group , and there is a nice explicit formula for the basic green s function on the boundary at infinity of 3-manifold ; but because we do not use it here we do not want to point to it here .
but one can find it in @xcite ( or for a physical point of view in @xcite ) .
of course the process in the next part can be used for this case too .
we have pointed some notes about this case in section [ s1.5 ] .
consider the complex projective linear transformations group @xmath98 . _
( i ) kleinian groups .
_ a subgroup @xmath99 of @xmath98 is called a kleinian group if @xmath99 acts on @xmath45 properly discontinuously . for a kleinian group @xmath99 and a point @xmath100 in @xmath45
lets denote by @xmath101 , the orbit of the point @xmath102 under the action of @xmath99 . since @xmath99 acts on @xmath45 properly discontinuously , @xmath101 has accumulation points only on @xmath35 .
they are independent of the choice of the reference point @xmath100 and are called the _
limit set _ of @xmath99 , which is denoted by @xmath103 .
equivalently , it is the closure of the set of all fixed points of the elements of @xmath99 other than the identity .
@xmath103 is the minimal non - empty closed @xmath99 - invariant set .
the complement of the limit set @xmath104 is denoted by @xmath105 and is called the _ region of discontinuity _ of @xmath99 .
+ the kleinian group @xmath99 is called of the _ first kind _ if @xmath106 and of the _ second kind _ if @xmath107 . in the case that @xmath99 is of the second kind @xmath99 acts on @xmath105 properly discontinuously .
therefor in this case @xmath105 is the maximal open @xmath99 - invariant subset of @xmath35 where @xmath99 acts properly discontinuously .
the quotient space @xmath108 has the complex structure induced from that of @xmath105 .
thus @xmath108 is a countable union of riemann surfaces lying at infinity of the complete hyperbolic 3 - manifold @xmath109 . for a torsion free kleinian group @xmath99 , a manifold @xmath110
possibly with boundary is denoted by @xmath111 and is called a _ kleinian manifold_. the interior @xmath109 of @xmath111 which admits the hyperbolic structure is denoted by @xmath112 . + _
( ii ) loxodromic , parabolic and elliptic elements . _ an element @xmath113 in @xmath98 is called loxodromic if it has exactly two different fixed points in @xmath35 .
these points are denoted by @xmath114 and @xmath115 and are called attracting one and repelling one . for any @xmath116 and @xmath117 , we have @xmath118 .
if we denote by @xmath119 the eigenvalue of @xmath113 on the complex tangent space to @xmath114 then there is a local coordinate for @xmath35 that in this coordinate @xmath113 is represented by @xmath120 .
for this reason @xmath119 is called the multiplier of @xmath113 , and we have @xmath121 , @xmath122 . also by definition @xmath113 is called a parabolic element if it has precisely one fixed point in @xmath35 , and elliptic if it has fixed points in @xmath45 .
in fact an elliptic element fixes all points of a geodesic in @xmath45 joining two fixed points in @xmath35 . for a loxodromic or elliptic element
@xmath113 the geodesic joining two fixed points in @xmath35 is called axis of @xmath113 .
an element is elliptic if and only if it has finite order , then elliptic elements cause a singularity in @xmath112 .
for this reason we consider kleinian groups without elliptic elements .
this means that the group is torsion free or equivalently acts freely on @xmath45 . + _
( iii ) schottky groups_. a finitely generated , free and purely loxodromic kleinian group is called a schottky group .
purely loxodromic means , all elements except the identity are loxodromic .
for such a group @xmath123 the number of a minimal set of generators is called the genus of @xmath123 .
a marking for a schottky group @xmath123 of genus @xmath100 by definition is a family of @xmath124 open connected domains
@xmath125 in @xmath35 and a family of generators @xmath126 with the following properties . for @xmath127 + a )
the boundary @xmath128 of @xmath129 is a jordan curve homeomorphic to @xmath130 and closures of @xmath129 are pairwise disjoint .
\b ) @xmath131 and @xmath132 .
+ we say that a marking is classical , if all @xmath128 are circles .
it is known that every schottky group admits a marking . in fact
each schottky group admit infinitely many marking but there are some schottky groups for which no classical marking exists . + _ ( iv ) @xmath123 - invariant sets of the schottky groups and there quotient spaces_. a schottky group @xmath123 is a kleinian group , then it acts on @xmath45 properly discontinuously .
this action is free too .
then the quotient space @xmath133 has a complete hyperbolic 3-manifold structure this means that it is a non - compact riemann space of constant curvature -1 .
topologically , if @xmath123 is of genus @xmath100 then @xmath133 is the interior of a handlebody of genus @xmath100 . + now lets consider the marking @xmath134 for the schottky group @xmath123 of genus @xmath100 and put @xmath135 the set @xmath136 is the region of discontinuity of @xmath123 and @xmath137 is a fundamental domain for the action of @xmath123 on @xmath35 .
then @xmath123 acts on @xmath136 freely and properly discontinuously .
so the quotient space @xmath138 is a complex riemann surface of genus @xmath100 .
all compact riemann surfaces can be obtained in this way ( see @xcite ) and every compact riemann surface admits infinitely many different schottky covers .
+ the cayley graph of a schottky group @xmath123 of genus @xmath100 is an infinite tree with multiplicity of @xmath124 at each vertex .
@xmath123 is free , then this tree is without loops and each path between two points is unique .
now , by the definition of a marking for @xmath123 , each generator @xmath139 takes @xmath140 to the inside of @xmath141 and again the generator @xmath142 takes this part to the inside of the second copy of @xmath143 inside @xmath141 .
this means that for each element @xmath113 in @xmath123 we can associate a path in the cayley graph that the points of @xmath140 moves along that path on a tubular neighborhood .
this express the action of @xmath123 on the set @xmath136 ( figure [ fig1.12 ] illustrating this for the case @xmath144 ) .
we can consider the riemann surface @xmath145 as the boundary at infinity of the manifold @xmath146 by identifying the points of @xmath145 with the points of the set of equivalence classes of unbounded ends of oriented geodesics in @xmath146 modulo the relation `` distance=0 '' .
+ by the definition of the limit set for a kleinian group , the complement @xmath147 is the limit set of @xmath123 .
for the cyclic schottky groups ( of genus 1 ) the limit set @xmath148 consists of two points which can be chosen as @xmath149 , but for genus@xmath150 this set is an infinite cantor set ( see @xcite ) .
+ lets consider the irreducible left - infinite words like @xmath151 . where @xmath152 , @xmath153 , and @xmath154 is a point in the fundamental domain @xmath137 . and put @xmath155 this is a well defined point of @xmath148 and is independent of the point @xmath154 . since @xmath123 is a free group , the map @xmath156 establishes a bijection as following irreducible left - infinite words in @xmath139 + @xmath157 + ends of the cayley graph of @xmath123 , @xmath158 + @xmath157 + points of @xmath159 + denote by @xmath160 the hausdorff dimension of the set @xmath148 .
it can be characterized as the convergence abscissa of any poincare series @xmath161 where @xmath15 is any coordinate function on @xmath35 with a zero and a pole in @xmath136 . for @xmath162
that re@xmath163 , this series converges uniformly on compact subsets of @xmath136 .
generally we have @xmath164 ( see @xcite ) .
in the next section we will consider @xmath165 to have the convergency of some infinite products defining some @xmath123 - automorphic functions .
this class of schottky groups was characterized by bowen @xcite in following way : @xmath165 if and only if @xmath123 admits a rectifiable invariant quasi - circle ( which then contains @xmath148 ) . + choose marking for @xmath123 of genus @xmath100 and denote by @xmath166 the image of @xmath128 in @xmath167 with induced orientation . also for @xmath127 choose the points @xmath168 in @xmath128 and denote by @xmath169 the images in @xmath167 of oriented pathes from @xmath168 to @xmath170 lying in @xmath137 .
these images are obviously closed paths and we can choose them in such a way that they do nt intersect . if we denote the classes of these pathes in 1- homology group @xmath171 by the same notations , then the set @xmath172 form a canonical basis of this group i.e. for all @xmath127 we have @xmath173 moreover the kernel of the map @xmath174 which is induced by the inclusion @xmath175 is generated by the classes @xmath166 .
lets consider the cyclic schottky group @xmath176 , for @xmath177 , @xmath178 . in this case
@xmath179 and consequently @xmath180 .
then a differential of the first kind on @xmath180 can be written as @xmath181 where @xmath154 is any point @xmath182 . and
for @xmath183 , this differential determines a differential of the first kind on @xmath167 . in general case for a schottky group @xmath123 of genus
@xmath100 we can make a differential of the first kind @xmath184 for any @xmath185 on @xmath136 and @xmath167 by an appropriate averaging of this formula .
lets consider a marking for @xmath123 and denote by @xmath186 a set of representatives of @xmath187 ; by @xmath188 a similar set for @xmath189 ; and by @xmath190 the conjugacy class of @xmath113 in @xmath123 .
then for any @xmath191 we have \(a ) if @xmath192 , the following series converges absolutely for @xmath193 and determines ( the lift to @xmath145 of ) a differential of the first kind on @xmath145 : @xmath194 this differential does not depend on @xmath154 , and depends on @xmath113 additively . also if the class of @xmath113 is primitive ( i.e. non - divisible in h:@xmath195 $ ] ) , @xmath184 can be rewritten as following @xmath196 ( b ) if @xmath139 form a part of the marking of @xmath123 , and @xmath166 are the homology classes described before , we have
@xmath197 it follows that the map @xmath113 mod @xmath198\mapsto \omega_g$ ] embeds @xmath199 as a sublattice in the space of all differentials of the first kind .
+ ( c ) denote by @xmath200 the complementary set of homology classes in @xmath201 as in before .
then we have for @xmath202 , with an appropriate choice of logarithm branches : @xmath203 and @xmath204 where @xmath205 is @xmath206 without the identity class . for the proofs , see @xcite and @xcite .
notice that our notation here slightly differs from @xcite ; in particular , @xmath207 here corresponds to @xmath208 of @xcite .
lets consider the points @xmath54 and @xmath55 in @xmath140 the fundamental domain of @xmath123 and put @xmath209 . then assuming @xmath165
, we see that this series absolutely converges and because it s @xmath123 - automorphic it gives us a differential of the third kind on @xmath145 with residues @xmath210 at the images of @xmath64 in @xmath145 . moreover , since both points @xmath64 are out of the circles @xmath128 , its @xmath166 - periods vanish .
now , if we consider the linear combination @xmath211 with real coefficients @xmath212 , then it will have pure imaginary @xmath166 - periods and if we find real coefficients @xmath212 so that the real part of @xmath169 - periods of the form @xmath213 vanish , we will be able to use this differential in order to calculate conformally invariant green s functions .
the set of the equations for calculating the coefficients @xmath214 are as following @xmath215 for @xmath127 .
the parts @xmath216 are calculated by means of formula ( [ f1.10 ] ) and ( [ f1.11 ] ) , and @xmath169 - periods of @xmath217 are given in @xcite .
+ now if we denote the points @xmath218 , and @xmath69 in @xmath145 and their images in the fundamental domain @xmath137 by the same notations , then we have @xmath219 and finally we have the green s function on @xmath167 as following @xmath220
all hyperbolic riemann surfaces are uniformized by fuchsian groups .
also fuchsian groups act on hyperbolic space by poincare extension and uniformize some hyperbolic 3-manifolds with two boundaries at infinity with the same genera .
but they do not uniformize all such manifolds ; in fact they are uniformized by an extension of fuchsian groups that are called quasi - fuchsian groups . in this chapter
first we intend to extend the method in previous chapter to fuchsian groups and then for quasi - fuchsian groups .
+ consider hyperbolic plane @xmath221 with the upper half plane model and coordinate @xmath222 with @xmath223 , equipped by the hyperbolic distance function corresponding to the metric @xmath224 of the constant curvature -1 .
the geodesics in this model are vertical half - lines @xmath225 and vertical half - circles orthogonal to the line at infinity @xmath226 ( i.e , for @xmath51 ) .
+ if we consider the end points of geodesics ( including @xmath43 for the end of vertical half - lines other than the end in @xmath226 ) then we can consider circle @xmath227 ( or @xmath228 in unit disc model ) as boundary at infinity of @xmath221 .
each subgroup @xmath229 of @xmath230 , the general projective linear group over @xmath226 , that acts freely and properly discontinuously on @xmath221 is called a fuchsian group .
similar to the kleinian groups the limit set @xmath231 and the region of discontinuity @xmath232 are defined but in @xmath233 , the boundary at infinity of @xmath221 .
and the properties are similar to them . also similar to the loxodromic elements ,
a hyperbolic element is an element @xmath234 with two fixed points @xmath235 in @xmath233 . for a fuchsian group @xmath229
the quotient space @xmath236 is a hyperbolic riemann surface possibly with boundary @xmath237 ( if @xmath238 ) .
+ let @xmath229 be a fuchsian group such that @xmath239 is a compact riemann surface with the genus @xmath240 .
then @xmath229 is a purely hyperbolic group of finite order @xmath124 ( for example see @xcite ) .
lets denote by @xmath241 the generators of @xmath229 and by @xmath242 the fixed point set of @xmath241 . also consider that @xmath102 is the fundamental polygon of @xmath229 with @xmath243 as it s sides , such that @xmath244 and @xmath245 for @xmath246 .
we represent @xmath102 as following : @xmath247 where @xmath248 and @xmath249 are the intersection points of @xmath166 and @xmath169 , @xmath250 and @xmath251 and so on . now we have @xmath252=i \rangle\ ] ] and also ( see for example @xcite ) .
@xmath253 if we denote by @xmath254 and @xmath255 the images of @xmath166 and @xmath169 in @xmath0 , then @xmath256 generate @xmath257 .
@xmath229 acts similarly , freely and properly discontinuously on lower half plane @xmath258 of @xmath50 too , and @xmath258 is @xmath229-invariant .
lets denote by @xmath229 the poincare extension of @xmath229 on @xmath45 too ( see @xcite or @xcite ) . by this extension @xmath229 can be considered as a kleinian group .
and we have @xmath259 ( see @xcite ) .
and @xmath260 has a hyperbolic structure , @xcite .
we know that @xmath231 the limit point set of @xmath229 is @xmath261 ( @xmath130 in unit disc model ) , @xcite .
then as a kleinian group @xmath262 and the region of discontinuity of @xmath229 considering as a kleinian group is @xmath263 and also the kleinian manifold is @xmath264.\end{aligned}\ ] ] now because @xmath0 is compact then @xmath265 .
consequently @xmath260 is a hyperbolic 3 - manifold with two compact boundary component at infinity @xmath266 and @xmath267 with the same genus @xmath100 . for convenience and
having a simple intuition lets consider the unit disc model for @xmath221 in this section .
we can code points of @xmath268 and geodesics with beginning and end points on @xmath268 as following : + lets mark semicircles including sides of p by @xmath269 in the counter clockwise direction around @xmath130 and put @xmath270 , and so on . in general for @xmath271 , @xmath272 . and label end points of @xmath273 on @xmath130 by @xmath274 and @xmath275 ( with @xmath276 ) with @xmath274 occurring before @xmath275 in the counter clockwise direction , ( see figure [ fig2.13 ] ) . and define @xmath277 @xmath278 then @xmath279 is a well defined map and is called a markov map related to the fuchsian group @xmath229 .
we have the following lemma the map @xmath279 and the group @xmath229 are orbit equivalent on @xmath130 , namely except for the pairs @xmath280 , for @xmath281 ; for each x , y in @xmath130 , @xmath282 for some @xmath283 in @xmath229 if and only if there exists nonnegative integers @xmath284 such that @xmath285 . ( see @xcite ) .
now label each arc @xmath286 by @xmath139 , and for each element @xmath16 in @xmath130 put @xmath287 . where the component @xmath288 is the label of the segment to which @xmath289 belongs .
we know that for each generator @xmath241 of @xmath229 , the fixed point @xmath290 is in the circle @xmath291 and @xmath292 is in @xmath293 then from the definition of markov map we have @xmath294 next , if @xmath295 be a geodesic with the beginning point @xmath296 and end point @xmath297 then put @xmath298 now for each @xmath299 let @xmath300 show the geodesic arc between @xmath301 and @xmath302 ( the axis of @xmath303 ) and @xmath304 the image of @xmath305 in @xmath260 .
then we have the following geodesics in @xmath260 * closed geodesics : a geodesic in @xmath260 is closed if and only if it s the projection of the axis of a hyperbolic element in @xmath229 , * images of the geodesics with beginning and end points in @xmath231 , * geodesics with beginning points in @xmath306 that are limit cycle to @xmath307 , for a generator @xmath241 of @xmath229 , * geodesics with beginning and end points in @xmath308 or @xmath309 .
, * geodesics with beginning point in @xmath308 and end point in @xmath309 and vis versa . for the riemann surface @xmath0 and associated fuchsian group @xmath229 as above let @xmath310 be the smallest normal subgroup of @xmath229 including @xmath241 for @xmath311 .
then it s obvious that the factor group @xmath312 is a free group generated by @xmath100 generators .
if we consider the covering @xmath313 associated to @xmath310 , then from normality of @xmath310 the group of deck transformations of this cover is @xmath312 and according to the classical koebe uniformization theorem @xcite ( see also @xcite)there is a planar region @xmath314 that is a region of discontinuity of a schottky group @xmath315 and @xmath316 .
@xmath317 is covering isomorphic to @xmath318 and for coverings @xmath319 , @xmath320 and @xmath321 we have @xmath322 i.e. the following diagram is commutative @xmath323 + @xmath324 + @xmath0 also we have @xmath325 for @xmath127 and @xmath326 for each @xmath283 in @xmath310 and each mappings is complex - analytic covering and @xmath123 is uniquely determined to within conjugation in @xmath327 @xcite .
+ for an extension of @xmath328 to a set some more than @xmath221 that we need it in the later computations , lets denote by @xmath329 the free subgroup of @xmath229 generated by @xmath330 .
we know that @xmath331 is a subset of @xmath231 .
@xmath331 is @xmath329 invariant and is out of the closer of the fundamental domain of @xmath329 , then all points of @xmath331 are in only the circles @xmath128 that are related to the generators of @xmath329 and their inverses .
this shows that the coding of each point in @xmath331 includes only the generators of @xmath329 and their inverses . also since each generator of @xmath229 like
@xmath283 is an isometry of @xmath221 , it s a composition of two reflections , then @xmath332 is not in the isometry circle of @xmath333 for each point @xmath16 in @xmath130 .
then by the definition of the markov map @xmath279 all the codings of the points of @xmath130 are irreducible .
this means that each point @xmath16 in @xmath331 can be coded by the irreducible formal combination @xmath334 for @xmath335 .
this fuchsian coding is suitable for expressing geodesics in @xmath260 or @xmath336 when the beginning and end points of the geodesics are in @xmath331 and also can be used for extending the map @xmath328 on @xmath331 and maybe on some more .
+ now we want to extend @xmath328 to @xmath337 ( onto @xmath338 ) . for this , we know that each element of @xmath148 can be coded by the infinite words @xmath339 for a constant @xmath154 in @xmath136 and @xmath340 . similarly , since @xmath329 is free and purely loxodromic group then it s a schottky group too and we can use schottky coding for it too . for conveniences in some proofs
we will use schottky codings for the points of @xmath331 .
then each point @xmath16 in @xmath331 can be coded by the infinite words @xmath341 for @xmath335 and a constant point @xmath342 in @xmath343 ( also @xmath344 )
. then we can consider @xmath342 in @xmath345 such that @xmath346 . also from ( [ f2.10 ] ) and using the properties of a loxodromic element and invariance of the limit set under @xmath229 ( or @xmath329 ) it is not hard to show that on @xmath331 we have the following relation between fuchsian and schottky coding @xmath347 and also this is true for repelling points too because of the relation @xmath348 for each @xmath283 in @xmath229 .
now , lets put @xmath349 in other words , by the definition of infinite words , the continuity of @xmath328 and the relation @xmath350 on @xmath221 @xmath351 then the map @xmath352 is well defined and onto . also from ( [ f2.9 ] ) we see that we can explain @xmath328 and consequently the functions ( spatially green s function ) on @xmath308 by the fuchsian coding , when these functions are expressed via this map . + for each point @xmath16 in @xmath331 and element @xmath283 in @xmath329 equivalent to @xmath59 ( i.e. @xmath353 , where @xmath335 and @xmath241 is replaced by @xmath354 and vice versa ) we have @xmath355 then @xmath356 and @xmath357 .
let @xmath329 be the free group generated by @xmath330 and @xmath358 be a divisor with support @xmath359 in @xmath221 .
put @xmath360 and again lets denote by @xmath361 a meromorphic function on @xmath35 with the divisor @xmath6 , and for an element @xmath342 in @xmath362 define : @xmath363 for the schottky group @xmath123 associated to the fuchsian group @xmath229 , if @xmath165 , then the product ( [ f2.91 ] ) converges absolutely and uniformly on any compact subset of @xmath221 , after deleting a finite number of factors that may have a pole or zero on this subset .
let @xmath364 be a compact subset of @xmath221 .
since @xmath229 acts on @xmath221 properly discontinuously the set @xmath365 is finite .
if we delete the set @xmath366 from the index set of the product ( [ f2.91 ] ) then for each @xmath367 when @xmath332 and @xmath368 lie outside a fixed compact neighborhood of @xmath359 we have @xmath369 where @xmath370 is a constant , @xmath371 , @xmath346 and @xmath372 ( and @xmath373 ) .
the last inequality comes from the equality @xmath374 .
now since @xmath165 the series @xmath375 converges uniformly on the compact subsets of @xmath136 . then the product ( [ f2.91 ] ) is convergent . in @xmath376 , if we change the point @xmath342 to @xmath377 , then we have @xmath378 . where @xmath379 is a nonzero complex number that depends on the points @xmath342 and @xmath377 and @xmath380
also , for @xmath381 and @xmath382 we have @xmath383 @xmath384 is a nonzero complex number multiplicative on @xmath385 and @xmath283 and also independent of @xmath342 .
we can see this as following @xmath386 then @xmath387 .
when @xmath123 is a cyclic group @xmath388 .
this shows that @xmath389 is not @xmath229 automorphic function on @xmath221 in general .
\a ) lets denote by @xmath390 a set of the representatives of @xmath391 .
then for @xmath392 $ ] in @xmath329 we have @xmath393 and by defining @xmath394 we have @xmath395 where @xmath396 is the set @xmath397 without the identity class .
+ b ) lets denote by @xmath398 a set of representatives of @xmath399 and by @xmath400 the conjugacy class of @xmath283 in @xmath329 .
then for some @xmath377 in @xmath221 such that @xmath401 stays in @xmath402 , we have @xmath403 and this is independent of @xmath377 .
lets put @xmath404 , @xmath405 and @xmath406 then we have @xmath407 the part ( [ f2.8 ] ) comes from the equation @xmath408 and the invariance of the cross - ratio on the action of mobius transformations .
the second part of a ) and b ) can be proved similarly . for b ) we should consider that @xmath409 . in above theorem in the case that the divisor is @xmath71 we have @xmath410 by previous theorem @xmath411 is a meromorphic function without any poles and zeroes in @xmath221 then it is a holomorphic function on @xmath221 . also as we see in the expression , it is independent of the point @xmath377 . for each @xmath283 in @xmath329 lets put @xmath412
however @xmath389 is not @xmath229- automorphic function on @xmath221 in general but we have @xmath413 and this shows that the differential @xmath414 is @xmath229 automorphic . then @xmath415 is a differential of the first kind on @xmath308 . + according to the classical theorem of cuts we can choose @xmath136 the region of discontinuity of the schottky group @xmath123 with the marking @xmath416 with @xmath417 such that @xmath418 the image of @xmath166 for @xmath127 in @xmath308 be coincident with the image of @xmath141 . and these together with @xmath419 the image of @xmath169 for @xmath127 in @xmath308
make a canonical base for @xmath420 i.e. @xmath421 \a ) @xmath422 is a riemann s basis for the space of differentials of the first kind on @xmath308 by choosing the previous base for @xmath423 i.e. @xmath424 b ) if we denote by @xmath390 a set of the representatives of @xmath391 and by @xmath425 , the set @xmath390 without the identity class , then for @xmath426 we have @xmath427 and for @xmath428 by defining @xmath429 we have @xmath430 we have @xmath431 where @xmath432 , @xmath401 and @xmath433 .
if we show this equality for @xmath434 by @xmath435 then for a ) we have @xmath436 the last equality comes from this fact that if @xmath428 since @xmath437 and @xmath438 , only for @xmath439 the first alternative and for all @xmath440 the third one is valid . if @xmath426 only the third alternative is right for all @xmath441 .
we can see these from the figure 1 .
+ for the first part of b ) .
we have shown by the point @xmath442 the intersection of @xmath443 and @xmath444 in the representation of @xmath102 , the fundamental domain of @xmath229 .
then @xmath445 is the image of the part of @xmath169 that is between the points @xmath442 and @xmath446 in the fundamental domain .
now , from ( [ f2.11 ] ) we have @xmath447 where the last equality comes from the definition in chapter one .
similarly we can reach to the second part of b ) using ( [ f2.12 ] )
. for @xmath448 lets denote by the same words the corresponding points in @xmath449 the fundamental domain of @xmath229 ( then @xmath450 and @xmath451 are in @xmath452 the fundamental domain of @xmath123 ) and put @xmath453 .
then @xmath217 is a differential of third kind on @xmath308 with the residues 1 and -1 at the images of @xmath54 and @xmath55 . also , because the points @xmath54 and @xmath55 are in the fundamental domain then they are out of the circles @xmath128
. then @xmath418 - periods of @xmath217 are zero , and from ( [ f2.13 ] ) the @xmath454 - periods are @xmath455 then we have @xmath456 now , we can reach to a differential of the third kind with pure imaginary periods by defining @xmath457 where the real coefficients @xmath458 are such that the set of the equations @xmath459 for @xmath460 are satisfied .
in fact the coefficients @xmath458 kill the real part of the @xmath419 - periods of @xmath217 .
notice that the new differential form @xmath461 probably have nonzero @xmath418 - periods , but since the coefficients @xmath458 are real and @xmath462 are pure imaginary then they are pure imaginary too
. + finally , if we denote the points @xmath218 , and @xmath69 in @xmath463 and the images of these points in the fundamental domain of @xmath229 by the same notations , then we have @xmath464 and by ( [ t2.11 ] ) @xmath465 then the green s function on @xmath308 can be computed as following @xmath466 because of the commutativity of the diagram @xmath467 @xmath468 @xmath469 the image of the point @xmath382 and @xmath470 are the same in @xmath308 .
then the above formula for the green s function on @xmath308 gives an expression via the points on @xmath308 .
in fact for the points @xmath64 in @xmath232 the images of the geodesics @xmath471 and @xmath472 in the kleinian manifold @xmath473 are the same .
we can see this by using the following covering spaces @xmath474 for all parts @xmath475 , and again we have @xmath476 .
also since @xmath328 is extended on @xmath331 in a natural way then for each @xmath477 the image of the geodesics @xmath478 and @xmath479 in @xmath260 are the same .
+ one can reach to a formula for the green s function on @xmath309 similarly by replacing @xmath258 and @xmath480 instead of @xmath221 and @xmath102 . a finitely generated , torsion free kleinian group @xmath481
is called quasi - fuchsian if the limit set @xmath482 be a jordan curve and each of the two simply connected components of @xmath483 be @xmath481-invariant . given a quasi - fuchsian group @xmath481 , there exist a fuchsian group @xmath229 and a quasiconformal diffeomorphism between kleinian manifolds @xmath484 and @xmath473 .
( see @xcite ) .
lets denote by @xmath485 and @xmath486 the simply connected components of @xmath483 for quasi - fuchsian group @xmath481 .
then there are two fuchsian groups @xmath487 and @xmath488 related to @xmath481 such that @xmath489 is homeomorphic to @xmath490
. then each @xmath491 is isomorphic to @xmath481 . also since @xmath492 is topologically conjugate to @xmath493 ( or @xmath494 in unit disk model )
then naturally there is a markov map @xmath495 like fuchsian case .
then all the statements for fuchsian groups can be extended to the quasi - fuchsian groups . in this case because the jordan curve @xmath482 is generally not smooth nor even rectifiable ( @xcite p.263 ) then the hausdorff dimension of the subgroup @xmath496 of @xmath481 ( like @xmath329 for fuchsian case @xmath229 ) may be not less than 1 in general . we have the following theorem for a finitely generated kleinian group @xmath99 the following conditions are equivalent + ( 1 ) @xmath497 is diffeomorphic to @xmath498 $ ] where @xmath0 is a component of @xmath499 ; + ( 2 ) @xmath99 is quasi - fuchsian ; + ( 3 ) @xmath105 has an invariant component that is a jordan domain ; + ( 4 ) @xmath105 has two invariant components ; ( see @xcite , page 125 ) . also two homeomorphic riemann surfaces can be made uniform simultaneously by a single quasi - fuchsian group ( theorem of bers on simultaneous uniformization ) .
then we have the computations for all hyperbolic 3 - manifolds with two compact riemann surfaces with the same genus as it s boundary components .
already , we have computed the green s function for the boundary components of a hyperbolic 3-manifolds with two boundary at infinity with the same genera . in this chapter
we want to do the problem for the most general case that the green s function can be defined i.e. for manifolds with @xmath41 boundary components that are compact riemann surfaces probably with different genera .
the case with two boundary is different genera are included in this case too .
such manifolds are uniformized by some kleinian groups and for the first , we will try to find a decomposition for such kleinian groups .
this , help us to find invariants of these groups that are necessary for computing the automorphic functions that will be used for computing the differentials and the green s functions . in all of this section we consider @xmath99 to be a finitely generated and torsion free kleinian group of the second kind ( i.e. @xmath501 ) . * ( the ahlfors finiteness theorem ) * let @xmath99 be a finitely generated and torsion free kleinian group
. then @xmath502 is a finite union of analytically finite riemann surfaces ( closed riemann surface from which a finite number of points are removed ) .
for the rest of the section we consider @xmath503 such that for each @xmath504 , @xmath3 is a compact riemann surface of genus @xmath505 and without cusped point .
then we should consider that @xmath99 does nt have any parabolic elements . also consider that the component @xmath506 is the one that for @xmath507 , @xmath508 . in this case
@xmath99 is a function group ( a finitely generated non - elementary kleinian group which has an invariant component in its region of discontinuity ) and there is an invariant component @xmath509 in @xmath105 .
we consider two cases for @xmath509 in the proposition above : + * case1 : * the component @xmath509 is simply connected .
then in this case @xmath99 is a b - group ( a function group with simply connected invariant component ) .
and we know that a b - group with a compact boundary component say @xmath511 is a quasi - fuchsian group ( see @xcite page 411 ) . + * case2 : * the component @xmath509 is not simply connected . in this case
we have the following theorem .
let @xmath99 be a function group with invariant component @xmath509 that is not a simply connected component
. then @xmath99 has a decomposition into a free product of subgroups as following @xmath512 where each @xmath513 is a b - group , @xmath514 is a free abelian group of rank two with two parabolic generators , @xmath515 is the cyclic group generated by the parabolic transformation @xmath505 and @xmath516 is the cyclic group generated by the loxodromic transformation @xmath241 .
furthermore , each parabolic transformation in @xmath99 is conjugate in @xmath99 to one a listed subgroup .
@xmath99 has finite - sided fundamental polyhedron if and only if all the groups @xmath513 do .
this theorem shows that in our special case we have @xmath517 for b - groups if one boundary component be compact then it s a quasi - fuchsian group . and
@xmath518 is a free and purely loxodromic group and consequently it s a schottky group @xcite .
hence we have @xmath519 where @xmath520 is a quasi - fuchsian group and @xmath123 is a schottky group .
then by van kampen theorem we have @xmath521 and if @xmath522 then @xmath523 we will denote by @xmath524 the genus of schottky group @xmath123 in the decomposition of the group @xmath99 .
we intend to analyze the structure of the region of discontinuity of @xmath99 . for the first
, we consider the case that we do nt have the schottky group @xmath123 in the decomposition of @xmath99 .
lets denote by @xmath129 and @xmath525 the simply connected components of the quasi - fuchsian group @xmath520 . from proposition ( [ p3.11 ] ) all other components than @xmath509 of @xmath105 are simply connected .
now , since @xmath531 and from lemmas ( [ l3.11 ] ) , ( [ l3.12 ] ) and the decomposition of @xmath499 we see that @xmath532 is a complete system of representatives of the equivalent classes ( @xmath533 and @xmath534 are equivalent or conjugate if and only if @xmath535 is conjugate in @xmath99 to @xmath536 ; and since @xmath537 then @xmath533 and @xmath534 are equivalent or conjugate if and only if there is a @xmath113 in @xmath99 such that @xmath538 ) of the components of @xmath105 . since @xmath509 is @xmath99 - invariant the class of @xmath509 has one member
. then we have in this case we have @xmath549 + now , lets consider the general case that there is a schottky group @xmath550 ( of order @xmath524 ) in the decomposition of the group @xmath99 .
again we know that all other components of @xmath105 than @xmath509 is simply connected
. then @xmath551 can not be glued to the components other than @xmath506 in @xmath499 .
also this means that in this case we have @xmath552 and @xmath553 in this case if @xmath554 be a marking for the schottky group @xmath555 , then the fundamental domain is @xmath556 and also for the genus of the component @xmath506 we have @xmath557 .
lets denote by @xmath310 the smallest normal subgroup of the group @xmath560 including the generators @xmath561 for @xmath507 and @xmath562 and @xmath563 be the smallest normal subgroup of the group @xmath520 including the generators @xmath561 for @xmath562 .
then we have @xmath564 now if @xmath310 be the smallest normal subgroup of the group @xmath99 including the generators @xmath561 for @xmath507 and @xmath562 , then by the previous lemma one can see that @xmath568 is isomorphic to the group @xmath569 then @xmath568 is a free group of order @xmath570 .
+ now lets consider the covering space map @xmath571 .
since @xmath310 is a normal subgroup of @xmath99 then the covering space @xmath572 is regular and is between @xmath573 with the covering transformations group @xmath574 that is a free group of order @xmath570 . then similar to the previous section we have a schottky group @xmath575
( we can arrange the generators of the group @xmath123 in this way because of the isometry ) and a complex - analytic covering mapping @xmath576 such that the following diagram is commutative and for @xmath507 and @xmath579 , @xmath580 and for @xmath581 , @xmath582 and @xmath583 ( we have considered the same notations for the generators of @xmath584 and @xmath123 ) . as a matter of fact there is a fuchsian group @xmath229 and there are normal subgroups @xmath585 and @xmath586 of @xmath229 such that the sequence @xmath587 is the composition of analytic covering maps .
then we have the composition of covering maps @xmath588 and since @xmath589 is a free group , like the previous section we have @xmath590 for some schottky group @xmath123 .
in fact , for the first , because of the isometries @xmath591 we can choose the fuchsian group @xmath592=i\quad \rangle\end{aligned}\ ] ] such that the equations @xmath593 for @xmath507 and @xmath594 are satisfied . where @xmath595 are the generators of @xmath99 .
then like the previous section we can choose uniquely up to conjugation in @xmath596 the schottky group @xmath123 that satisfies the equations @xmath597 then since @xmath598 is onto , the relations @xmath599 and @xmath600 are satisfied automatically for @xmath507 and @xmath594 .
this means that the following diagram is commutative in all loops @xmath601 @xmath602 @xmath601 this gives a proof for the existing the of covering space @xmath136 and schottky group @xmath123 satisfying the properties that we need and also shows the relations between the fuchsian group @xmath229 , kleinian group @xmath99 and the schottky group @xmath123 associated to @xmath506 . + now lets put @xmath603 .
where @xmath604 is the free subgroup of @xmath520 generated by the elements @xmath605 . for the extension of @xmath328 to @xmath606 i.e. defining the map @xmath607
, we know that each @xmath604 is free and purely loxodromic then it is a schottky group . like the previous section we can code an element of @xmath608 by schottky coding @xmath609 for @xmath610 and a point @xmath342 in @xmath611 . for each element @xmath113 in @xmath612 lets denote by @xmath59 the element in @xmath123 corresponding to @xmath113 , such that @xmath613 .
now put @xmath614 @xmath615 where @xmath346 . and
for @xmath507 put @xmath616 @xmath617 and finally lets define @xmath618 @xmath619 + then like the previous section we can show that for each element @xmath113 in @xmath612 and @xmath59 the element of @xmath123 corresponding to @xmath113 and each @xmath16 in @xmath606 we have @xmath620 and @xmath621 . also like the previous chapter we can consider the fuchsian coding for the points of @xmath606 and from ( [ f2.9 ] ) we can explain @xmath328 and consequently the green s function on @xmath506 via the fuchsian coding .
if we consider the condition @xmath165 and use the subgroup @xmath622 instead of @xmath329 in the previous section then one can bring all of the definitions like before with some changes , and compute the green s function on @xmath506 and the other parts in the same way .
in this case we should notice that if @xmath623 be a set that makes a base for @xmath624 then each member of it has a class of images in @xmath509 .
but we can consider that the representatives that are used in the computations are in the fundamental domain .
as its shown in figure [ fig3.14 ] . in this case
these images are in @xmath625 for @xmath507 and also in @xmath626 for @xmath627 .
then the representatives for @xmath3 and @xmath506 are coincide if we uniformize @xmath3 by @xmath628 , i.e. identifying both components of @xmath629 .
final formula for the green s function on @xmath506 is as following @xmath630 when we have infinitely many boundary components in the boundary of @xmath112 that are riemann surfaces without puncture , according to the ahlfors finiteness theorem , @xmath99 is an infinitely generated kleinian group . and
hence one of the boundary components of @xmath111 is with infinity genus and is not a compact riemann surface then for this component the green s function is not defined .
but for the other components we can bring the computations similar to the previous condition using some subgroups of @xmath99 that are isomorphic to the kleinian groups in the previous sections .
30 l. v. ahlfors .
`` finitely generated kleinian groups '' .
amer.j . math .
86 ( 1964),413 - 429 and 87(1965),759 .
`` uniformization , moduli and kleinian groups '' . bull .
london math .
, 4(1972),257.300 .
+ r. bowen .
`` hausdorff dimension of quasi - circles '' publ . math .
ihes 50,11 - 25(1979 ) .
c. series .
`` markov maps associated with fuchsian groups '' .
publications mathmatiques de li.h.e.s.,tome 50(1979),p.153 - 170
. g. cornel , j.h .
`` arithmetic geometry '' .
springer verlag ( 1986 ) .
g. faltings .
`` calculating on arithmetic surfaces '' .
119 , 387 - 424 ( 1982 ) .
l. r. ford .
`` automorphic functions '' .
mcgraw - hill 1929 .
d. hejhal .
`` on schottky and teichmuller spaces '' .
15 , 133 - 156 ( 1975 ) .
`` compact riemann surfaces '' .
`` uber die uniformisierung der algebraischen kurven '' .
iv , math . ann . 75 ( 1914 )
`` introduction to arakelov theory '' .
new york berlin heidelberg : springer 1988 .
`` three - dimensional hyperbolic geometry as @xmath43 - adic arakelov geometry '' . invent .
104 ( 1991 ) , no .
2 , 223243 . yu .
i. manin , v. drinfeld , `` periods of p - adic schottky groups '' .
j. reine angew . math .
262/263 ( 1973 ) , 239247 .
i. manin , matilde marcolli .
`` holographi principle and arithmetic curves '' .
k. matsuzaki , m. taniguchi .
`` hyperbolic manifolds and kleinian groups '' .
oxford university press , 1998 .
c. mcmullen .
`` riemann surfaces and the geometrization of 3 - manifolds '' . bull .
27 ( 1992 ) , no .
2 , 207216 .
a. marden .
`` the geometry of finitely generated kleinian groups '' .
the annals of mathematics , 2nd ser .
99 , no . 3 ( may , 1974 ) , pp . 383 - 462
. j. g. ratcliffe .
`` foundations of hyperbolic manifolds '' springer verlag ( 1991 ) .
p. zograf , l. takhtajan .
`` on the uniformization of riemann surfaces and on the weil - petersson metric on the teichmuller and schottky spaces '' . math .
ussr - sb .
60 ( 1988 ) , no.2 , 297 - 313 . | the work is motivated by a result of manin in @xcite , which relates the arakelov green function on a compact riemann surface to configurations of geodesics in a 3-dimensional hyperbolic handlebody with schottky uniformization , having the riemann surface as conformal boundary at infinity .
a natural question is to what extent the result of manin can be generalized to cases where , instead of dealing with a single riemann surface , one has several riemann surfaces whose union is the boundary of a hyperbolic 3-manifold , uniformized no longer by a schottky group , but by a fuchsian , quasi - fuchsian , or more general kleinian group .
we have considered this question in this work and obtained several partial results that contribute towards constructing an analog of manin s result in this more general context . |
this work is supported in part by the national natural science foundation of china .
see , e.g. , t. fukuyama and h. nishiura , hep - ph/9702253 ; r. n. mohapatra and s. nussinov , _ phys .
d _ * 60 * , 013002 ( 1999 ) ; z. z. xing , _ phys .
d _ * 61 * , 057301 ( 2000 ) ; _ phys .
d _ * 64 * , 093013 ( 2001 ) ; _ phys .
d _ * 74 * , 013010 ( 2006 ) .
s. luo , j. w. mei and z. z. xing , _ phys .
d _ * 72 * , 053014 ( 2005 ) ; s. luo and z. z. xing , _ phys .
b _ * 632 * , 341 ( 2006 ) ; _ phys . lett .
b _ * 637 * , 279 ( 2006 ) ; z. z. xing and h. zhang , _ commun .
* 48 * , 525 ( 2007 ) . | assuming the majorana nature of massive neutrinos , we generalize the friedberg - lee neutrino mass model to include cp violation in the neutrino mass matrix @xmath0 . the most general case with all the free parameters of @xmath0 being
complex is discussed .
we show that a favorable neutrino mixing pattern ( with @xmath1 , @xmath2 , @xmath3 and @xmath4 ) can naturally be derived from @xmath0 , if it has an approximate or softly - broken @xmath5-@xmath6 symmetry .
we also point out a different way to obtain the nearly tri - bimaximal neutrino mixing pattern with @xmath7 and non - vanishing majorana phases . recently
, a novel neutrino mass model has been proposed by friedberg and lee ( fl).@xcite the neutrino mass operator in the fl model is simply given by @xmath8 where the parameters @xmath9 , @xmath10 , @xmath11 and @xmath12 are all assumed to be _ real _ , and the charged - lepton mass matrix is taken to be diagonal .
a salient feature of @xmath13 is its partial gauge - like symmetry ; i.e. , its @xmath9 , @xmath10 and @xmath11 terms are invariant under the transformation @xmath14 ( for @xmath15 ) with @xmath16 being a space - time independent constant element of the grassmann algebra.@xcite from eq .
( 1 ) , one can directly write down the neutrino mass matrix : @xmath17 two interesting features can be inferred from the diagonalization of @xmath0 .
first , the neutrino mass matrix takes a magic form,@xcite in which the sums of rows and columns are all equal to @xmath12 .
the unitary matrix used to diagonalize @xmath0 must have one eigenvector with three equal components @xmath18 .
second , when @xmath19 holds , it is very easy to check that the neutrino mass operator @xmath13 has the exact @xmath5-@xmath6 symmetry ( i.e. , @xmath13 is invariant under the exchange of @xmath5 and @xmath6 indices).@xcite in addition , one may consider to remove one degree of freedom from @xmath13 or @xmath0 ( for instance , by setting @xmath20).@xcite to include cp or t violation into the fl model , one may insert the phase factors @xmath21 into eq .
( 1 ) by replacing the term @xmath22 with the term @xmath23.@xcite the resultant neutrino mass matrix is no longer symmetric , hence it describes dirac neutrinos instead of majorana neutrinos .
however , in most of the realistic models , the majorana nature is preferable to the dirac nature of neutrinos .
hence , in this work , we aim to generalize the fl model to include cp and t violation for massive majorana neutrinos .
let us start from the generic analysis with all the parameters of @xmath0 in eq .
( 2 ) being complex . for majorana neutrinos
, @xmath0 is symmetric and can be diagonalized by the transformation @xmath24 , in which @xmath25 ( for @xmath26 ) stand for the neutrino masses .
after a straightforward calculation , the neutrino mixing matrix @xmath27 turns out to be @xmath28 where the explicit expressions of @xmath29 and @xmath30 are @xmath31 and the definitions of @xmath32 , @xmath33 and @xmath34 can be found in ref . .
furthermore , three mass eigenvalues of @xmath0 and two majorana phases of @xmath27 are found to be @xmath35 and @xmath36 again , the expressions of @xmath37 have also been listed in ref . .
we proceed to consider two special but interesting scenarios of the generalized fl model and explore their respective consequences on three neutrino mixing angles and three cp - violating phases .
scenario ( a ) : @xmath9 and @xmath12 are real , and @xmath38 are complex .
note that the @xmath5-@xmath6 symmetry of @xmath0 is softly broken in this case , because @xmath39 holds . by using the generic results given in eqs .
( 3)-(6 ) , one can easily arrive at @xmath40 $ ] , @xmath41 , @xmath42 ^ 2 + 3 \left [ { \rm i m } \left ( b \right ) \right ] ^2 } ~ - a + { \rm re } \left ( b \right ) \ ; , \ \ \ \ \
m^{}_2 = m^{}_0 \ ; , \nonumber \\ m^{}_3 & = & \sqrt{\left [ m^{}_0 + a + 2{\rm re } \left ( b \right ) \right ] ^2 + 3 \left [ { \rm i m } \left ( b \right ) \right ] ^2 } ~ + a - { \rm re } \left ( b \right ) \ ; , % ( 7)\end{aligned}\ ] ] together with @xmath43 . comparing our results with the well - known standard
parametrization,@xcite we immediately obtain @xmath44 , @xmath45 , and @xmath46 $ ] .
the leptonic jarlskog parameter @xmath47 , which is a rephasing - invariant measure of cp violation in neutrino oscillations,@xcite reads @xmath48 .
if @xmath49 holds , the tri - bimaximal neutrino mixing pattern ( with @xmath50 or @xmath51 , @xmath2 and @xmath52)@xcite will be reproduced .
one can see that the soft breaking of @xmath5-@xmath6 symmetry leads to both @xmath53 and @xmath54 , but it does not affect the favorable result @xmath2 given by the tri - bimaximal mixing pattern . on the other hand , @xmath55 is an excellent approximation , since @xmath29 must be small to maintain the smallness of @xmath56 . in view of @xmath57,@xcite
we obtain @xmath58 and @xmath59 .
it is possible to measure @xmath60 in the future long - baseline neutrino oscillation experiments .
the neutrino masses in scenario ( a ) rely on four real model parameters @xmath12 , @xmath9 , @xmath61 and @xmath62 .
thus it is easy to fit the neutrino mass - squared differences @xmath63 and @xmath64.@xcite such a fit should not involve any fine - tuning , because ( a ) the number of free parameters is larger than the number of constraint conditions and ( b ) three neutrino masses have very weak correlation with three mixing angles .
a detailed numerical analysis can be found in ref . , and a remarkable feature is that only the normal neutrino mass hierarchy ( @xmath65 ) is allowed in this scenario .
scenario ( b ) : @xmath9 , @xmath10 and @xmath11 are all real , but @xmath12 is complex . by using eqs .
( 3)-(6 ) , we obtain @xmath66 , and @xmath67 ^ 2 + \left [ { \rm i m } \left
( m^{}_0 \right ) \right ] ^2 } \;\ ; , \ \ \ \ \ m^{}_2 = \left
|m^{}_0 \right | \ ; , \nonumber \\ m^{}_3 & = & \sqrt{\left [ m^{}_{+ } + { \rm re } \left ( m^{}_0 \right ) \right ] ^2 + \left [ { \rm i m } \left ( m^{}_0 \right ) \right ] ^2 } \;\ ; , % ( 8)\end{aligned}\ ] ] where @xmath68 .
two majorana phases @xmath69 and @xmath70 are given by @xmath71 although @xmath69 and @xmath70 have nothing to do with the behaviors of neutrino oscillations , they may significantly affect the neutrinoless double - beta decay.@xcite comparing our formulae with the standard parametrization , we arrive at @xmath72 , @xmath73 , @xmath74 together with @xmath7 for the dirac phase of cp violation .
the results for @xmath75 and @xmath56 in this scenario are the same as those obtained in scenario ( a ) , but the jarlskog parameter @xmath47 is now vanishing .
because of the @xmath5-@xmath6 symmetry breaking , @xmath76 may somehow deviate from the favorable value @xmath77 .
given @xmath78 corresponding to @xmath57 , @xmath76 is allowed to vary in the range @xmath79 .
the neutrino masses depend on five real model parameters @xmath9 , @xmath10 , @xmath11 , @xmath80 and @xmath81 .
hence there is sufficient freedom to fit two observed neutrino mass - squared differences @xmath82 and @xmath83 . our careful numerical analysis , which has been done in ref .
, shows that both normal ( @xmath65 ) and inverted ( @xmath84 ) neutrino mass hierarchies are allowed , and the majorana phases ( @xmath69 , @xmath70 ) are less restricted in scenario ( b ) .
although our discussions about the generalized fl model are restricted to low - energy scales , it can certainly be extended to a superhigh - energy scale ( e.g. , the gut scale or the seesaw scale ) . in this case
, one should take into account the radiative corrections to both neutrino masses and flavor mixing parameters when they run from the high scale to the electroweak scale.@xcite we conclude that the @xmath5-@xmath6 symmetry and its slight breaking are useful and suggestive for model building .
we expect that a stringent test of the generalized fl model , in particular its two simple and instructive scenarios , can be achieved in the near future from the neutrino oscillation and neutrinoless double - beta decay experiments . |
the existence of non - baryonic dark matter is supported by varied cosmological measurements . of great relevance
are the values of the matter and energy densities of the universe at the present time .
these densities can be determined by means of several cosmological data : the temperature fluctuations in the cosmic microwave background ( cmb ) , the distance - luminosity relation for supernovas , the distribution of galaxies on large scales , the abundance of light elements ( primordial nucleosynthesis ) , etc .
the density values so obtained are compatible with all current astrophysical and cosmological observations , from the internal motions of galaxies and galaxy clusters to studies of weak gravitational lensing .
@xcite finds the following values @xcite for the current matter and energy densities , @xmath0 , in units of @xmath1 g/@xmath2 ( i.e. 18.79 yg / m@xmath3 or 1.689 nj / m@xmath4 ) : * -0.5em a negligible density in relativistic particles ( `` radiation '' ; e.g. , the cmb photons contribute only @xmath5 ) ; * -0.5em @xmath6 in a smoothly distributed component ( dark energy ) ; * -0.5em @xmath7 in non - relativistic particles ( `` matter '' ) , of which * * -1em @xmath8 in protons and neutrons ( baryons ) , * * -1em @xmath9 ( 95% cl ) in non - baryonic hot dark matter , * * -1em @xmath10 in non - baryonic cold dark matter .
it is the excess of total matter density ( @xmath11 ) over baryonic matter density ( @xmath12 ) that constitutes the evidence for non - baryonic dark matter .
none of the known elementary particles can account for non - baryonic dark cold matter .
the obvious standard model candidates would be neutrinos , but the measurements of the neutrino mass squared differences , @xmath13 ev@xmath14 , and the experimental upper bound of 3 ev on the mass of the neutrino produced in tritium beta decay , impose that the masses of all three known neutrinos are @xmath15 ev .
neutrinos so light constitute _ hot _ dark matter , and have to be included in @xmath16 ( 95% cl ) .
thus no known particle is a candidate for non - baryonic cold dark matter .
scores of hypothetical particles have been proposed as cold dark matter candidates over the past several decades .
they range from new particles in well - founded extensions of the standard model , to possible particles inspired by recent theoretical ideas . to the first category belong an extra heavy neutrino , the axion , and the lightest supersymmetric particle ( the neutralino , the gravitino , or the sneutrino ) . in the second category are particles like wimpzillas , solitons ( b - balls and q - balls ) , self - interacting dark matter , string - inspired dark matter , kaluza - klein dark matter , etc .
a class of non - baryonic dark matter candidates is of interest here because of several recent claims to their detection : weakly interacting massive particles , or wimps .
wimps are appealing because of a simple mechanism that can produce the observed value of their cosmic density .
assume that in the early universe wimps were in thermal and chemical equilibrium with the rest of the matter and radiation . as the universe expanded and cooled down , the chemical reactions coupling wimps to the rest of the world slowed down and eventually stopped , leaving a constant number of wimps in a volume that expands with the universe .
numerically , the correct present - time density of wimps is obtained for matter - wimp couplings of the order of electroweak couplings , and wimp masses in the 1 gev100 tev range .
these characteristics give these particles their name .
examples of wimps are a heavy neutrino and the lightest neutralino .
the latter arises in supersymmetric extension of the standard model , and is one of the most popular candidates for non - baryonic dark matter .
signals from dark matter wimps can be either direct or indirect .
direct signals are due to collisions of dark matter wimps with nuclei in a detector
. a very sensitive low - background detector records the amount of energy deposited by wimps in collisions with nuclei , and in the future also the direction of motion of the struck nucleus .
indirect signals are due to the products of wimp reactions in planets , stars , or galaxies .
the most common reaction is wimp annihilation : wimps can annihilate with anti - wimps , if present , or with themselves , if , like the neutralino , they are their own anti - particle . out of the products of wimp annihilation , neutrinos , positrons , anti - protons , and high - energy gamma - rays are those of most interest because they are rarely produced by usual astrophysical processes .
wimp annihilations occur at a detectable rate where wimps are concentrated , as in the center of the sun , the center of the earth , and the inner regions of galactic halos , ours in particular .
neutrino telescopes , gamma - ray telescopes , and cosmic ray detectors are used to search for wimps indirectly .
the next three sections discuss signals that can or have been attributed to wimps , stressing the importance of exploiting distinctive signatures to ascertain their wimp origin .
in two separate balloon flights with different detectors , the heat collaboration @xcite has observed more cosmic ray positrons above @xmath177 gev than expected in current models of cosmic ray propagation in our galaxy . in these models , positrons arise as secondary particles in the interactions of primary cosmic rays with interstellar matter .
modifications of these models could in principle account for the extra positrons ( and similar extra photons observed in egret @xcite ) , but no proposed modification can yet reproduce all observable cosmic ray data ( see discussion in @xcite , e.g. ) .
wimp annihilation can r[0pt]0.49 -2 also be invoked to explain the extra positrons , as illustrated in figure [ fig : heat ] .
do the positron data distinguish between the suggested origins for the excess ? not by themselves , in that they lack a clear and unique signature that they are due to wimps .
the positron spectrum predicted by wimp annihilation lacks any discriminating feature , with the exception of a reduction in flux and then a cut - off as the energy increases towards the wimp mass .
any such flux reduction can however be pushed beyond any foreseeable maximum detectable energy by simply raising the wimp mass . in the absence of a distinguishing feature from wimps ,
it is hard to draw conclusions on the origin of the positron excess .
in principle , gamma - rays from wimp annihilation offer a characteristic signature in their spectrum : a gamma - ray line @xcite .
the line originates in the annihilation of wimps into a pair of photons , each photon carrying an energy equal to the wimp mass , between 10 gev and 100 tev .
no astrophysical process is known to produce a gamma - ray line at these energies .
this makes the wimp gamma - ray line an ideal signature for wimps .
r0.48 -2 -0.5 searches for the wimp gamma - ray line are continuing , but no line has been detected yet .
the challenge is twofold : both a large number of photons and a fine energy resolution are needed .
figure [ fig : glast ] shows that the space - born gamma - ray telescope glast , scheduled for launch in 2006 , is expected to have such capabilities . in the meantime ,
another source of gamma - rays from wimps has been used to claim their detection : the gamma - ray continuum . these are gamma - rays generated in the decay of secondary products , such as pions , produced by wimp annihilation into quarks , w , or z bosons .
contrary to the gamma - ray line , continuum gammas from wimps lack a characteristic feature , except for a flux reduction and cut - off near the wimp mass . in this respect , they are similar to cosmic ray positrons from wimps . raising the wimp mass pushes the flux reduction beyond the observable energies .
the possibilities are limitless .
for example , the early-2004 cangaroo l0.4 -1.5 -0.5 report of high - energy gamma - rays from the galactic center @xcite has been interpreted as due to annihilations of @xmath17 1 tev wimps ( see figure [ fig : hooper ] ) .
the cangaroo data can also be explained by appropriate modeling of accretion flows around the black hole at the galactic center @xcite .
the mid-2004 hess observation of gamma - rays from the same region @xcite has a very different spectrum from cangaroo s , but it can also be interpreted either with appropriate ( but different ) accretion flows @xcite or as due to wimp annihilation , this time with @xmath17 20 tev wimps ( figure [ fig : horns ] ) . if this mass seems too high to supersymmetry aficionados , it may be amusing to see that even minimal supergravity models allow for @xmath17 10 tev neutralinos compatible with cosmological and astrophysical bounds and the hess data ( figure [ fig : hbg ] ) .
the lack of an unmistakable wimp signature in the gamma - ray continuum makes it unsuitable as a primary indicator of the presence of wimps .
a gamma - ray line would be desirable .
an excellent signature for wimps in direct detection has been known for years : the annual modulation @xcite .
the earth motion periodically changes the relative speed of earth and wimps , causing the wimp flux on earth , and thus the wimp detection rate , to vary in time and to repeat itself once every year .
the details of the annual pattern depend on the wimp velocity distribution .
for example , the date of maximum rate is set by the most common arrival direction of the wimps , and happens in june for the canonical halo model with maxwellian velocity distribution , but may occur in december for sikivie s cold infall model ( see @xcite ) .
similarly , the amplitude of the modulation depends on the halo model .
the dama collaboration has observed an annual modulation in their l0.55 -2.5 - sodium iodide data ( figure [ fig : dama ] ) and has attributed it to wimps .
no valid alternative explanation has been put forward yet , but no wimp signal has been observed in any other direct detection experiment either .
however , comparison of the various experimental results , which are obtained with different targets , is marred by the need of uncertain theoretical assumptions about the wimp mass , interaction , and halo model .
in fact , the expected event rate depends on the product of the wimp - nucleus cross section and the wimp flux on earth .
the cross section scales differently with the atomic mass of the target nucleus according as the wimp interacts with the nuclear mass or the nuclear spin .
the flux depends on the distribution of wimp velocities , which is probably more complicated than an arbitrarily - assumed maxwellian .
indeed , current hierarchical models of galaxy formation entail halo substructure and streams of dark matter .
( further details on the direct detection rate can be found in @xcite . )
-0.5 -0.5[fig : cdms ] neglect of the theoretical assumptions involved has given rise to many controversies .
for example , the first dama analysis @xcite was artificially restricted to wimps heavier than 30 gev , on the basis of some theoretical prejudice on the nature of the wimps . under the assumption of a spin - independent cross section ,
the latest cdms data @xcite exclude the original dama region ( figure [ fig : cdms]a ) , but relaxing the artificial restriction on the wimp mass there remains the possibility of a ( now admittedly narrow ) region at low wimp masses ( figure [ fig : cdms]b ) . the moral is : do not use theoretical prejudices when analyzing data .
another example in which the theoretical assumptions play a subtle role is the case of spin - dependent cross sections . for a canonical maxwellian halo ,
if wimps interact predominantly with neutrons , existing data exclude the dama region , while if wimps interact predominantly with protons , the dama region may still be allowed , but only under further assumptions .
as shown in figure [ fig : sgf ] , the most stringent limit on wimp - proton spin - dependent interactions comes from the absence of indirect neutrino signals l0.55 -1em -0.01 - 1em -0.25 from wimp annihilation in the sun .
this places an upper limit of @xmath1710 gev on the wimp mass ( and a further analysis of lower energy neutrinos may lower this limit even further ) .
however , there may be no anti - wimps in the sun with which the wimps can annihilate , and as a consequence there may be no neutrino bounds on the spin - dependent wimp - proton cross section .
this would make the dama region allowed up to wimp masses of @xmath17100 gev .
notice that the neutralino , a fashionable candidate , is its own anti - particle , and thus for it the neutrino bounds can not be avoided ; restricting attention to the neutralino is however a theoretical prejudice . in the face of all these difficulties of interpretation , other wimp signatures in direct detection
would be helpful .
very promising are detectors that can record not only the energy deposited by the wimp but also the direction of motion of the nucleus after the collision .
one such detector , drift , is currently under construction @xcite
. it will be possible to use the recoil direction of the nuclei to discriminate a wimp signal from background , for instance more wimps should come from the direction of motion of the solar system than from the opposite direction .
another signature for wimps has been proposed in @xcite in case a stream of dark matter passes through the solar system .
streams are common in simulations of galaxy formation , and have already been observed in our galaxy .
one of them in particular , the stream associated with the tidal disruption of the sagittarius dwarf galaxy , may pass by the solar system .
streams through the solar system may even enlarge the possibilities for the dama region , provided they come from roughly ahead of the solar system motion ( see figure [ fig : ggdama ] ) . the new signature proposed in @xcite is a combined annual modulation of the rate and of the highest energy that wimps in the stream can impart to a nucleus .
this highest energy shows as a step - like feature in the nucleus recoil spectrum , and the location of the step is predicted to vary in energy with a period of one year ( figure [ fig : fgn ] ) .
no background spectrum is expected to behave in this way .
with additional experimental signatures for wimps in direct detection we may be able to understand the origin of a claimed signal .
the best signals for wimp discovery are characterized by special features that make them recognizable as due to wimps and nothing else .
for example , gamma - rays from wimp annihilation should show a gamma - ray line in correspondence to the wimp mass ; high - energy ( @xmath18gev ) neutrinos from the sun or the earth can not be produced by anything else but wimps ; the direct detection rate , and for wimps in a stream , the highest recoil energy , should follow a modulation with a period of a year ; etc .
it is these features one should look for in searching for dark matter wimps .
history has shown however ( e.g. for the dama annual modulation ) that a single feature , although accountable for in theoretical models , may not be enough to make sure that we have detected wimps .
this is because the theory of wimps and their distribution is still very uncertain , and allows for many possibilities .
what is needed is _ experimental _ confirmation from a variety of detectors , and perhaps from different kinds of signals that can be explained within the _ same _ wimp model .
different kinds of signals will in any case be needed to determine all the interesting wimp properties , such as mass , cross section , density , etc . finally , to be really convinced , we will probably have to produce wimps in the laboratory , perhaps with high energy colliders . | more and more claims of having detected wimp dark matter are being put forward . some are discussed here , stressing the importance of exploiting distinctive signatures to ascertain their wimp origin .
the best signals for wimp discovery are characterized by special features that make them recognizable as due to wimps and nothing else . sometimes , however , a single feature , although accountable for in theoretical models , may not be enough to make sure that we have detected wimps .
this is because the theory of wimps and their distribution in the galaxy is still very uncertain , and allows for many possibilities .
what are needed are _ experimental _ verifications of the claimed signals , either by discovering _
unmistakable _ features , or by detecting several kinds of signals that can all be explained by the _
same _ wimp model .
one of the most intriguing results to come out of recent cosmological observations is that about 90% of the mass of the universe is not made of protons , neutrons , electrons , or any other known particle , but of something unknown that does not shine .
discovering the composition of this so - called non - baryonic dark matter is one of the big challenges of modern physics and cosmology .
proposals as to the nature of non - baryonic dark matter do not lack .
is it made of axions ( neutral particles suggested to explain the smallness of cp violation in the strong interactions ) ? or is it made of wimps ( weakly interacting massive particles that arise naturally in extensions of the standard model of particle physics , such as supersymmetry ) ?
or is non - baryonic dark matter made of something else , or a combination of all that ? following the tradition of experimental science , the way to find out the nature of non - baryonic dark matter is to detect its constituents , either directly by recording their collisions with a detector , or indirectly by observing products of their reactions in planets , stars , or galaxies .
the last ten years have seen more and more claims of having detected dark matter in the form of weakly interacting massive particles ( wimps ) .
three of these claims will be described below : ( 1 ) a distinctive signal variation in a direct detection experiment , ( 2 ) high - energy gamma - rays from the center of our galaxy , and ( 3 ) an excessive flux of positrons in cosmic rays .
however , explanations that do not invoke wimps exist for gamma - rays and positrons , and other direct detection experiments have not observed any signal from wimps ( although straightforward comparisons are difficult ) .
so , has wimp dark matter been detected ? what is the real origin of the detected signals ? or more proactively : how can we make sure we detect wimp dark matter
? this question has been asked repeatedly in the past , and several methods have been proposed to distinguish a dark matter signal from an ordinary one . in recent years , however , an excessive reliance on theory has interfered with an open - minded but critical interpretation of the experimental results .
current theories of wimps still leave a lot of possibilities as to their particle properties ( mass , couplings , etc . ) and astrophysical characteristics ( density distribution , velocity distribution , etc . ) . rather than theoretical arguments , what are needed
are _ experimental _ verifications of the claimed signals , either by discovering _
unmistakable _ features which can only be explained by the presence of wimps , or by detecting several kinds of wimp signals that can all be explained by the _
same _ theoretical model of wimps . |
supernova remnants ( snrs ) have long been believed to be the source of hadronic galactic cosmic rays ( gcrs ) up to energies of the ` knee ' , near @xmath1 ev , of the cosmic ray ( cr ) spectrum .
supernova explosions forming collisionless shock waves induce the shocked gas and relativistic particles ( hereafter cosmic rays ) that produce multi - wavelength thermal and nonthermal emission .
diffusive shock acceleration ( dsa ) is the most promising mechanism for converting the kinetic energy of a supernova explosion into energetic particles @xcite and plays an important role in nonthermal emission during the overall process , ( e.g. , @xcite ) . in the framework of dsa , an
individual charged particle experiences many collisions with background electromagnetic waves and gains energy by shock crossing .
this leads to a nonthermal cr spectrum of the power - law form @xmath2 .
shock acceleration by dsa in snr shocks is associated with transport processes , and some of the highest energy crs eventually escape from their acceleration sites by a so - called diffusion process due to interactions with turbulent magnetic fields . to analyse the diffusion process we have to determine the effective diffusion coefficient @xmath3 , and the understanding of @xmath3 is necessary to interpret many astronomical observations .
several studies of escaping crs have already been done and a strong spatial correlation between tev emission and the molecular gas distribution at the galactic center has been observed @xcite . the pion - decay signature in snrs is believed to be evidence for protons accelerated in middle - aged snrs interacting with molecular clouds , @xcite .
recent observations suggest that gamma - ray emission and co+hi emission are spatially correlated in young snrs _
rx j1713.7 - 3946 _ and _ rx j0852.0 - 4622 _ ( @xcite , see also @xcite . nevertheless , the identification of pion - decay gamma rays is difficult because high - energy electrons also produce gamma rays via bremsstrahlung and inverse compton scattering ( leptonic model ) @xcite .
x - ray observations show that electrons are accelerated to highly relativistic energies in snr shocks @xcite .
therefore , in order to understand the acceleration sites of crs , it is crucial to distinguish gev - tev emission from inverse compton scattering by cr electrons and the decay of neutral pions produced by inelastic collisions between cr protons and ambient thermal nuclei . in this paper , we investigate the diffusion of crs using a hydrodynamics simulation of a strong shock wave propagating in a realistic multiphase interstellar medium and a one - phase medium .
the organization of the paper is as follows . in section 2
, we describe the three - dimensional hydrodynamics simulations and the resulting configuration of electromagnetic field used .
we also briefly introduce the process of bohm diffusion .
the results of test particle simulations performed in these environments are shown in section 3 .
we investigate the properties of escaping crs in terms of the diffusion coefficient in both energy and configuration space , and finally we summarize and discuss our findings in section 4 .
the interstellar medium ( ism ) is an open system in which radiative cooling and heating are effective .
it is an inhomogeneous , multiphase system in which gases of different temperatures , densities , and ionization fractions can coexist in approximate pressure equilibrium .
diffuse warm gas ( diffuse intercloud gas ) with t @xmath4 10@xmath5 k and hi clouds ( interstellar clouds ) with t @xmath4 10@xmath6 k are approximately in pressure equilibrium in a typical ism environment . as a consequence of the thermal instability driven by external compressional events such as shock waves due to expanding hii regions or very late phase snrs
, unstable gas evolves into diffuse gas and hi clouds @xcite .
therefore , inhomogeneities inevitably emerge and remain ubiquitous in the ism . the characteristic length scale of an inhomogeneity can be expressed in terms of the field length " , which is the critical wavelength of the thermal instability @xcite .
the field length depends on density and temperature and can be smaller than 1 pc .
a blast wave generated by supernova expansion sweeps up the dense and clumpy hi clouds of the multiphase ism , which eventually generates strong velocity shear in the magnetic fields .
magnetic fields undergo amplification from their typical strength of @xmath7 g to mg due to the turbulent dynamo in the post - shock region ( @xcite ) .
this process may explain the existence of magnetic fields of mg strength investigated by y. uchiyama et al .
( 2007 ) @xcite .
t. inoue et al . ( 2012 ) performed ideal three - dimensional magnetohydrodynamic ( mhd ) simulations of a strong shock wave ( @xmath8 2500 km s@xmath9 ) propagating in a realistic multiphase ism as the pre - shock region .
we use the data for the perpendicular shock of t. inoue et al .
( 2012 ) at @xmath10 = 750 years as the background ism to set up the electromagnetic field for our microscopic particle simulations . to generate a multiphase ism , t. inoue et al .
( 2012 ) solved the ideal mhd equations including cooling , heating , and thermal conduction , which determine the unstable scale of thermal instability .
they considered a net cooling function and photoelectric heating , and generated an inhomogeneous medium via thermal instability .
the simulation considered ideal gas and used an adiabatic index of @xmath11 = 5/3 .
the mean number density , initial thermal pressure , and initial magnetic field strength were taken to be @xmath12 = 2.0 @xmath13 , @xmath14 = 2887 k @xmath13 , and @xmath15 = 5.0 @xmath7 g , respectively , at the @xmath16 = 0 boundary plane .
for the density , they imposed random density fluctuations with a thermally unstable state in the range 10 k @xmath17 t @xmath17 @xmath18 k for effective cooling and heating . in the resulting clumpy cloud , they induced a high mach number shock wave using a hot plasma with @xmath19 = 10@xmath20 k @xmath13 and @xmath21 = 0.1 @xmath13 .
t. inoue et al .
( 2013 ) performed ideal three - dimensional magnetohydrodynamic ( mhd ) simulations to investigate the interaction between blast wave ( @xmath22 km @xmath23 ) and interstellar density fluctuations .
they investigated the magnetic field amplification and the magnetic field distribution of turbulent snrs driven by the richtmyer - meshkov instability ( rmi ) .
they assumed an adiabatic gas with adiabatic index @xmath11=5/3 and used a high mach number shock wave .
density fluctuations superposed by sinusoidal functions were included and followed an isotropic power - law spectrum with random phases .
the power spectrum of the density fluctuations was shown to be described by an isotropic power law for the wavenumber _
k _ in the inertial range of turbulence : @xmath24 where @xmath25 is the fourier component of the density . the mean number density , the initial thermal pressure , and the initial magnetic field strength were taken to be @xmath12 = 0.5 @xmath13 , @xmath14 = 4 @xmath26 10@xmath27 k @xmath13 , and @xmath28 = 3.0 @xmath29 g , respectively .
the parameters represent typical values in the diffuse ism @xcite . to induce the blast wave ,
a hot plasma is set up as follows : @xmath19 = 2 @xmath26 10@xmath30 k @xmath13 , @xmath31 = 0.05 @xmath13 , and @xmath15 = 3.0 @xmath29 g at the @xmath32 = 0 boundary plane .
this creates the primary shock wave whose normal vector is perpendicular to the mean magnetic field .
the snr is modelled as a young snr ( age = 10@xmath33 years ) with a late with velocity of 1800 km @xmath23 .
we used the data for the perpendicular shock on model 1 in @xcite at @xmath34 = 700 years as the background ism .
a simulation box with size @xmath35 = 2 pc is used , and the system resolution is @xmath36/(number of grid cells ) = @xmath37 pc , where the number of uniform grid cells is @xmath38 .
periodic boundary condition is used for the _ yz_-plane . the above simulation resolution and boundary condition
are applied for both media .
the fourier power spectra of snr turbulence for both media are given in figure [ fig : fourierpowerspectrum ] .
( for more details see @xcite . )
the magnetized turbulent medium flows along the ( positive ) _
we calculate the trajectories of cr particles using a snapshot of the mhd simulation data .
we consider a collisionless environment for crs because the mean free path , @xmath39 , of relativistic particles is much larger than the gyroradius of cr particles . in the framework of non - relativistic ideal mhd , electromagnetic fields can be expressed as + @xmath40 where @xmath41 is the fluid velocity , @xmath42 is the local magnetic field at the position of the particles , and @xmath43 is the speed of light .
when we compute the lorentz force acting on each cr particle , we interpolate @xmath41 and @xmath42 at the position of the particle before we calculate @xmath44 using equation ( [ eq : e ] ) .
to trace the trajectories of cr , we solve the relativistic equation of motion using a fourth - order runge - kutta method for updating particle information .
the momentum of a particle is defined as + @xmath45 where @xmath46 is the velocity of a particle .
@xmath47 is the particle lorentz factor , and @xmath48 is the mass of particle @xmath49 .
the equations governing the motion of relativistic charged particles are expressed as @xmath50 where @xmath51 is the charge of particle @xmath49 .
the gyroradius , @xmath52 , of particle @xmath49 can be expressed by @xmath53 the time discretization , @xmath54 , for each particle is as follows : @xmath55 @xmath56 @xmath57 where @xmath58 = @xmath59 denotes the particle gyrofrequency and @xmath60 is the size of a grid cell .
@xmath61 is the number of divisions which determines the resolution of a particle trajectory .
the time discretization is determined by inverse of gyrofrequency as it shows in equation ( [ eq : time ] ) . to trace the particle trajectory ,
@xmath54 should be smaller than gyrofrequency .
@xmath62 is selected in our simulation when the particle energy is sufficiently small such that @xmath63 , whereas @xmath64 is selected for higher energy cases satisfying @xmath65 . for more details on @xmath66 , see figure [ fig : multi - rg ] in section 3.2 .
we obtain @xmath54 by considering @xmath67 and the velocity components ( @xmath68 , @xmath69 , and @xmath70 ) of a particle as well as using a sufficiently large value of @xmath61 for tracing the trajectories of particles .
for example , @xmath54 was about 250 s for @xmath71 = @xmath72 tev . in all calculations
, we used @xmath61 = 10@xmath27 , which is the value for keeping the errors of runge - kutta time - integration sufficiently small ( @xmath73 ev ) for energy conservation .
to calculate the transport of particles , we released sets of 10@xmath33 protons with the same initial energies but isotropic velocity distributions in the computational frame .
the range of the initial energies ( @xmath71 ) considered is 10@xmath74 ev to 10@xmath75 ev at intervals of @xmath76 @xmath71 = 0.5 ev .
we assigned particle positions randomly in a cubic box region in the post - shock region ( see tables 1 and 2 )
. initially the directions of velocities are also random . to obtain the background fluid velocity and magnetic field we applied a linear interpolation to the mhd simulation data because the magnetic and fluid fields are defined at discrete points .
we use periodic boundary conditions in the _ yz_-direction as was done in the mhd simulation .
we do not follow the time integration for orbits of particles that have escaped the @xmath32-boundary .
first , we show the results of a multiphase medium case . in a turbulent medium , second - order fermi acceleration
is expected where some particles gain energy stochastically for a sufficiently long time interval .
this can be considered to be a random walk in energy space ( @xmath77 ) .
we calculated up to @xmath78 years for @xmath78 tev and @xmath79 tev protons .
figure [ fig : multiphase - energygain ] show the magnitude of the particle energy , @xmath80 , as a function of @xmath81 .
dashed lines in both figures represent energy gain proportional to @xmath82 .
this demonstrates the stochastic behaviour of energy gain in our simulation .
however , the energy gain rate is very small within 10@xmath6 years , less than @xmath83 of the initial energy .
thus , we describe the energy in this paper with initial energy ( @xmath71 ) .
gyroradius ( equation ( [ eq : rg ] ) ) in units of pc for different values of @xmath71 shows in figure [ fig : multi - rg ] .
the values are calculated for 10@xmath27 particles at five different values of @xmath71 at intervals of @xmath84 = 0.5 ev .
each symbol of the solid lines corresponds to the gyroradius at a different time .
black open and filled circles correspond to @xmath10 = 10 years and 10@xmath85 years , respectively .
red stars indicate for an initial time .
note that the smallest ( @xmath86 ) and largest ( @xmath87 ) turbulent scales in our simulation correspond to a few grids and a few hundreds grids , respectively @xcite .
the grid size corresponds to @xmath88 = 6.03 @xmath26 10@xmath75 cm .
particles are supposed to be reflected by the turbulent field when @xmath89 . for the case of @xmath90 , particles only see an almost uniform field due to grid based method , and thus are expected to undergo @xmath91 drift . in this case
the travel distance is proportional to @xmath10 at early times . in figure
[ fig : multi - rg ] , particles with @xmath71 = 10@xmath92 ev , 10@xmath93 ev , 10@xmath94 ev , and 10@xmath75 ev satisfy @xmath89 .
we do not consider the case of @xmath71 = @xmath95 ev for our study of diffusion process in configuration space , because the number of escaped particles in our simulation is large . to analyze the diffusion in configuration space we measure the displacement of transported particles .
the displacement of a cr particle is expressed by @xmath96 where the notation @xmath97 denotes the particle @xmath97 and @xmath98 , @xmath99 .
the dispersion of displacement r , hereafter @xmath100 , is expressed by @xmath101 where @xmath102 denotes the total number of particles . here
, we show the transport property of particles in our simulations as a function of @xmath81 , and analyze the diffusion in configuration space .
the lines in figure [ fig : multi - rms - main ] represent @xmath100 for protons with energies of @xmath71 = @xmath103 ev , @xmath104 ev , and @xmath105 ev , corresponding to @xmath106 ( gyroradius ) @xmath107 @xmath88 ( grid size ) ( see figure [ fig : multi - rg ] ) . the values are averaged over @xmath102 = 10@xmath27 particles for each set of energies ( @xmath71 ) .
we calculated up to @xmath10 = 10@xmath6 years . to study the effect of changing the initial location , we defined several boxes for the initial cr particle locations .
table [ table : physical - multi ] shows the physical values of the background mhd data in the specific regions , defined as three - dimensional boxes . for the study of diffusion in a multiphase medium
, we describe the particle distribution and background information for the case of a box of size ( 0.3 pc)@xmath108 corresponding to region - i , as given in table [ table : physical - multi ] .
the evolution of @xmath100 in figure [ fig : multi - rms - main ] is described by standard diffusion , the so - called bohm diffusion , that is the slowest process to be proportional to @xmath82 for all energies in our simulations , @xmath71 = @xmath103 ev , @xmath104 ev , and @xmath105 ev . the black dotted line in figure [ fig : multi - rms - main ] is the slope @xmath82 . for efficient acceleration of crs , small diffusion coefficient is required .
thus , our results can demonstrate that the diffusion process of crs in snrs is can be approximated by bohm diffusion , which may support the assumptions in @xcite . in this section ,
we quantify the diffusion coefficient of cr particles . according to the standard assumption of the theory of dsa , accelerated particles diffuse out in the plasma , where the diffusive flux obeys fick s law :
the flux of particles is proportional to the gradient of flux density @xcite .
it is based on the assumption that charged particles are diffuse isotropically by turbulent electromagnetic fields . in general
, the diffusion coefficient can be expressed as @xmath109 where @xmath110 and @xmath111 are defined as the mean scattering length and the particle velocity , respectively .
the frequently used choice for the diffusion coefficient in dsa is that of bohm diffusion , expressed as @xmath112 this depends on the degree of turbulence ( @xcite ) . in our mhd simulation
, however , the field strength shows the relation @xmath113 , as shown in table 1 .
therefore we simply use a simplified expression for bohm diffusion with @xmath114 : @xmath115 this model , @xmath116 , has been used by previous work ( e.g. , @xcite ) . in order to quantify the diffusion of cr particles in our background field , we define @xmath117 as @xmath118 we compare this quantity to the bohm diffusion coefficient .
if @xmath119 is proportional to @xmath10 , @xmath117 is constant in time and can be used as the diffusion coefficient in the standard diffusion equation for cr particles . strictly speaking , @xmath117 is not the diffusion coefficient if it is not constant in time . even in that case , however , @xmath117 is a measure of the diffusion of cr particles .
figure [ fig : multi - bohm ] shows the ratio of @xmath120/@xmath121 for different values of @xmath71 calculated with equations ( [ eq : dbohm ] ) and ( [ eq : dsim ] ) .
the values are calculated for 10@xmath27 particles at five different values of @xmath71 at intervals of @xmath84 = 0.5 ev .
black open and red filled circles correspond to @xmath10 = 10 years and 10@xmath85 years , respectively .
both diffusion coefficients , @xmath120 and @xmath121 , are summarized in table [ table : dsim / dbohm - multi ] .
the value of @xmath122 decreases with increasing @xmath123 and remains of order unity , 13.5 ev < log @xmath71 < 14.5 ev owing to bohm diffusion process in our simulation .
thus , the bohm diffusion coefficient description for cr diffusion can be justified by our test particle simulations , at least for the shocked snrs .
as we mentioned in section 2.2 , the electric field in the comoving frame of the ism vanishes in the ideal mhd limit .
the strength of the electric field in the computation frame is smaller than the magnetic field strength by a factor of @xmath124 ( see equation ( [ eq : e ] ) ) . therefore , the force due to the electric field is small in comparison to the force due to magnetic field . in this section ,
we study the effect of electric field on the diffusion of crs in the ism .
figure [ fig : multi - noe ] shows the results of diffusion calculations with and without the background plasma ( @xmath125 ) .
it shows that inclusion of @xmath125 results in slightly larger diffusion but the difference is rather limited ( @xmath126 34@xmath127 ) .
therefore , electric fields are less important for studying diffusion in configuration space , but they are critical to studying the diffusion in energy space . from this section , we describe the transport properties of particles in configuration space for a one - phase medium as a function of @xmath128 , and analyze the diffusion in configuration space .
the values are averaged over @xmath102 = 10@xmath27 particles for each set of energies ( @xmath71 ) .
we calculated up to @xmath10 = 10@xmath6 years .
we mainly demonstrate the particle distribution and background information of region - i as given in table [ table : physical - one ] .
this region is chosen to be in the dense , strong magnetic field post - shock region . the calculation method and
conditions for a one - phase medium are the same as a multiphase medium of section 3.1 .
the lines in figure [ fig : onephase - mainresult - rms - t ] represent @xmath100 for protons with energies of @xmath71 = @xmath103 ev , @xmath104 ev , and @xmath105 ev , corresponding to @xmath106 @xmath107 @xmath88 and @xmath89 ( see the gyroradius in figure [ fig : onephase - rg ] and @xcite ) .
the uppermost black dashed line in the figure represents the slope @xmath129 .
we observed that the evolution of @xmath100 for @xmath130 ev ( blue solid line ) and 10@xmath94 ev ( red solid line ) can be fitted by @xmath129 , corresponding to fast diffusion , so - called superdiffusion " .
superdiffusion has been studied by s. xu @xmath131 h. yan ( 2013 ) @xcite and a. lazarian @xmath131 h. yan ( 2014 ) @xcite .
one of the reasons for superdiffusion can be the wandering of magnetic field lines in a turbulent medium ( see a. lazarian @xmath131 h. yan ( 2014 ) for details ) .
we demonstrate about superdiffusion property in next section 3.2.3 . in figure 8 ,
@xmath100 for @xmath71 = @xmath103 ev shows standard diffusion , corresponding to a simple random walk , represented by black dotted line in figure [ fig : onephase - mainresult - rms - t ] .
this random walk can be characterized by a diffusion coefficient , as we will discuss in the section 3.1 .
figure [ fig : onephase - dbohm ] shows the ratio of @xmath120/@xmath121 for different energies ( @xmath71 ) , calculated using equations ( [ eq : dbohm ] ) and ( [ eq : dsim ] ) ( see section 3.1 for review ) . applying the bohm diffusion process ,
the differences in @xmath120/@xmath121 are less than a factor of 5 with energy range with 13.5 ev < log @xmath71 < 14.5 ev , at both times in figure [ fig : onephase - dbohm ] .
the differences in the ratio of @xmath120/@xmath121 between @xmath10 = 10 years and @xmath132 years have grown more than those in figure [ fig : multi - rg ] due to superdiffusive process .
however , because the differences in @xmath120/@xmath121 are significantly small , thus our result precisely justifies the bohm diffusion .
the diffusion coefficients for a one - phase medium are summarized in table [ table : one - phase - diffusioncoeff ] .
superdiffusion has been reported in simulations by @xcite that is expected to occur below the injection scale .
( see the injection scale of one - phase medium case in figure [ fig : fourierpowerspectrum ] ) .
we compare our results with s. xu & h. yan ( 2013 ) to investigate the existence of superdiffusion in our simulations .
the pre - shock magnetic field lines in our simulations are laminar and not turbulent , and the turbulence is created by the rmi @xcite driven by the propagation of a blast wave .
in contrast , s. xu & h. yan ( 2013 ) drove turbulence by solenoidal forcing in fourier space .
we calculated the dispersion @xmath133 as a function of @xmath134 up to 10@xmath85 years , the result of which is shown in figure [ fig : xu ] , where @xmath135 is perpendicular to the magnetic field ( * _ b _ * ) , and @xmath136 is the displacement along the magnetic field .
the upper black dashed lines represent @xmath133 @xmath137 @xmath129 , respectively . the slope close to 1.5 may a manifestation of wandering magnetic field lines analogous to richardson diffusion as discussed in a. lazarian & h. yan 2013 . this may the reason for the superdiffusion found in figure [ fig : onephase - mainresult - rms - t ] .
however , the larger energy particles do not show the 1.5 slope of figure [ fig : xu ] , because the gyroradius of larger energy particles is larger than the injection scale of turbulence .
young supernova remnants are accompanied by collisionless shock waves in which dsa is expected to occur .
gamma rays are emitted by pion decay caused by cr protons accelerated in snrs interacting with molecular clouds .
observations of gamma rays from the vicinity of snrs have shown strong evidence that galactic cr protons are accelerated by the shock waves of the snrs @xcite . in dsa theory , efficient acceleration of crs
requires a small diffusion coefficient .
indeed , the time - variability seen in synchrotron x - ray observations of _
rx j1713.7 - 3946 _ seems to need the slowest diffusion process @xcite .
t. inoue et al .
@xcite adopted bohm diffusion as a diffusion of crs in the same snr to explain both the gamma - ray spectrum and x - ray variability with magnetic field amplification up to mg . in previous research , bohm diffusion
is discussed by considering ( @xmath138 @xmath139 1 ) ( see . equation ( [ eq : skilling ] ) , @xcite , and @xcite ) . in the present study , we performed test particle simulations that describe the propagation of high - energy cr particles in snrs in the early evolutionary phase .
we investigate the diffusion of cr particles in a background medium that is created from numerical simulation of a realistic interstellar medium .
our results show that the richtmeyer - meshkov instability can provide enough turbulence downstream of the shock to make the diffusion coefficient close to the bohm level for energy larger than 30 tev .
thus , the present study indicates that we can use bohm diffusion coefficient in practical calculations of cr particle diffusion , which may support the assumption used in the interpretation of high - energy emission from snrs @xcite .
it is worth stressing that in this paper we do not deal with the diffusion of particles in the medium upstream of the shock .
we recall that the dsa mechanism can predict the maximum energy of particles needed to explain the observations , only if the diffusion coefficient is close to the bohm one also in the upstream .
we also identified superdiffusion in a one - phase medium , in which the travel distance is proportional to @xmath129 . because the period of superdiffusion is limited in time ,
the actual travel distance is not so different from that described by a bohm diffusion coefficient , at least for particles with energies above 30 tev .
we thank toshio terasawa , takeru k. suzuki , and jennifer stone for useful comments .
sr thanks kazunari iwasaki and takuma matsumoto for their kind assistance .
numerical computations were in part carried out on cray xc30 at center for computational astrophysics , national astronomical observatory of japan .
00 a. r. bell , mnras 182 ( 1978 ) 443 - 455 .
a. r. bell , mnras.182 ( 1978 ) 147 - 156 .
blandford , j. p. ostriker , astrophys . j. lett . 221 ( 1978 )
w. i axford , e. leer , g. skadron , in : proc .
15th icrc , vol .
11 ( 1977 ) pp .
132 - 137 . w. i axford , e. leer , j.f mckenzie , a@xmath131a . 111 ( 1982 ) 317 - 325 .
drury , j. h. volk , astrophys .
j. 248 ( 1981 ) 344 - 351 .
f. aharonian et al . ,
nature 439 ( 2006 ) 695 - 698 .
e. wommer , f. melia , m. fatuzzo , mnras 387 ( 2008 ) 987 - 997 .
d. r. ballantyne et al .
, astrophys .
j. 657 ( 2007 ) l13-l16 .
m. ackermann et al . , science 339 ( 2013 ) 807 - 881 .
a. a. abdo et al . , fermi lat collaboration , astrophys . j. 706 ( 2009 ) l1-l6 .
a. a. abdo et al . , fermi lat collaboration , astrophys . j. 712 ( 2009 ) 459 - 468 .
a. a. abdo et al .
, fermi lat collaboration , astrophys . j. 718 ( 2010 ) 348 - 356 .
a. a. abdo et al .
, fermi lat collaboration , science 327 ( 2010 ) 1103 .
m. tavani , et al .
j. 710 ( 2010 ) l151-l155 .
y. ohira et al . ,
mnras 427 ( 2010 ) 91 - 102 . y. fukui et al . , astrophys .
j. 746 ( 2012 ) 82 .
y. fukui et al . ,
cosmic rays in star - forming environments .
34 ( 2013 ) p. 249 .
a. a. abdo et al . , astrophys .
j. 734 ( 2011 ) 28 .
t. inoue et al . , astrophys .
j. 774 ( 2010 ) 71 .
s. gabici , f.a .
aharonian , mnras 445 ( 2014 ) l70-l73 .
f. w. stecker , nature 222 ( 1969 ) 865
. g. g. fazio , annual review of astronomy and astrophys 5 ( 1967 ) 481 .
s. hayakawa , prog .
phys 8 ( 1952 ) 571 - 572 . t. naito , f. takahara , journal of physics g 20 ( 1994 ) 477 - 486 .
y. hanabata et al . , astrophys .
j. 786 ( 2014 ) 145 .
t. inoue , r. yamazaki , s. inutsuka , y. fukui , astrophys .
j. 744 ( 2012 ) 71 . v. n. zirakashvili , f. a. aharonian , f. a. astrophys . j. 708 ( 2010 ) 965 - 980 .
k. koyama et al . ,
nature 378 ( 1995 ) 255 - 258 .
s. inutsuka , h. koyama , t. inoue , aipc 784 ( 2005 ) 318 - 328 . t. inoue , s. inutsuka , astrophys .
j. 687 ( 2008 ) 303 - 310 .
t. inoue , s. inutsuka , astrophys .
j. 759 ( 2012 ) 35 .
h. koyama , s. inutsuka , astrophys . j. 602 ( 2004 ) l25-l28 . g. b. field , astrophys . j. 142 ( 1965 ) 531 - 567 .
t. inoue , r. yamazaki , s. inutsuka , astrophys .
j. 723 ( 2010 ) l108-l112 .
t. inoue , j. shimoda , y. ohira , r. yamazaki , astrophys . j. 772 ( 2013 ) l20 . j. giacalone , j. r. jokipii , astrophys . j. 663 ( 2007 ) l41-l44
. y. uchiyama et al . ,
nature 449 ( 2007 ) 576 - 578 .
p. c. myers , astrophys .
j. 225 ( 1978 ) 380 - 389 .
r. beck , space sci .
, 99 ( 2001 ) 243 - 260 .
s. xu , h. yan , astrophys .
j. 779 ( 2013 ) 140 .
a. lazarian , h. yan , astrophys . j. 784 ( 2014 ) 38 .
j. g. kirk , r. o. dendy , nucl . part .
27 ( 2001 ) 1589 - 1595 .
a. shalchi , aph 31 ( 2009 ) 237 - 242 .
m. hussein , a. shalchi , astrophys .
j. 785 ( 2014 ) 31 .
j. skilling , mnras 172 ( 1975 ) 557 - 566 .
.physical values in different simulation regions for the multiphase medium .
`` region - i '' and post - shock region " indicate the volume sizes occupied by the initial positions of particles in our simulation .
region - i and post - shock region have volumes ( 0.3 pc)@xmath33 for the range 1.5 - 1.8 pc in @xmath140 configuration space and ( 2 pc)@xmath33 in 1.6 - 1.8 pc of @xmath32 , 0.7 - 0.9 pc of @xmath141 , and 1.4 - 1.6 pc of @xmath142 configuration space . in this paper
, we show the results using region - i for a multiphase medium . [
cols="^,^,^,^,^",options="header " , ] , as a function of @xmath81 for protons with an initial energy of @xmath71 = @xmath78 tev ( left ) and @xmath71 = @xmath79 tev ( right ) .
we show several cases ( solid line ) for this figure the uppermost dotted line in figure shows the slope @xmath82.,title="fig : " ] , as a function of @xmath81 for protons with an initial energy of @xmath71 = @xmath78 tev ( left ) and @xmath71 = @xmath79 tev ( right ) .
we show several cases ( solid line ) for this figure the uppermost dotted line in figure shows the slope @xmath82.,title="fig : " ] ) in units of pc as a function of cosmic ray initial energy ( @xmath71 ) .
the results at 10 years and 10@xmath85 years are shown by black open and filled circles , respectively .
red star symbol is for the case @xmath10 = 0 . ] ) of cosmic ray particles ( @xmath71 ) as a function of @xmath143 .
green , blue , and red lines indicate the curves of @xmath71 = @xmath103 ev , @xmath104 ev and @xmath105 ev , respectively . all lines are similar to the uppermost black dashed line that is proportional to t@xmath144 , corresponding to the slowest process .
the travel distances with -0.5 < @xmath143
< 0.5 are proportional to @xmath10 . ] ) .
vertical axis shows the ratio between the simulation values ( @xmath120 ) and the estimated values ( @xmath121 ) from equations ( [ eq : dbohm ] ) and ( [ eq : dsim ] ) , shown as a function of @xmath84 by solid lines .
the results at 10 years and 10@xmath85 years are shown by open black and filled red circles , respectively . ] ) in units of pc as a function of cosmic ray initial energy ( @xmath71 ) .
the results at 10 years and 10@xmath85 years are shown by black open circles and filled circles , respectively .
red star symbol is for the case @xmath10 = 0 . ] ) of cosmic ray particles ( @xmath71 ) as a function of @xmath143 .
green , blue , and red lines indicate the curves of @xmath71 = @xmath103 ev , @xmath104 ev and @xmath105 ev , respectively .
the red and blue lines are similar to the uppermost black dashed line that is proportional to t@xmath145 , corresponding to the fast process ( superdiffusive ) . and green line is similar to the uppermost black dashed line that is proportional to t@xmath144 , corresponding to the slowest process . ] ) and
the estimated values ( @xmath121 ) from equations ( [ eq : dbohm ] ) and ( [ eq : dsim ] ) , shown as a function of @xmath84 by solid lines .
the results at 10 years and 10@xmath85 years are shown by open black and filled black circles , respectively .
stars are for the case @xmath10=0 . ] | supernova remnants ( snrs ) are one of the most energetic astrophysical events and are thought to be the dominant source of galactic cosmic rays ( crs ) .
a recent report on observations from the fermi satellite has shown a signature of pion decay in the gamma - ray spectra of snrs .
this provides strong evidence that high - energy protons are accelerated in snrs .
the actual gamma - ray emission from pion decay should depend on the diffusion of crs in the interstellar medium . in order to quantitatively analyse the diffusion of high - energy crs from acceleration sites , we have performed test particle numerical simulations of cr protons using a three - dimensional magnetohydrodynamics ( mhd ) simulation of an interstellar medium swept - up by a blast wave .
we analyse the diffusion of crs at a length scale of order a few pc in our simulated snr , and find the diffusion of crs is precisely described by a bohm diffusion , which is required for efficient acceleration at least for particles with energies above 30 tev for a realistic interstellar medium .
although we find the possibility of a superdiffusive process ( travel distance @xmath0 ) in our simulations , its effect on cr diffusion at the length scale of the turbulence in the snr is limited .
bohm diffusion , galactic cosmic rays , superdiffusion , supernova remnant |
the accelerated expansion of the universe , which is today observed , shows that cosmological dynamic is dominated by the so called dark energy which gives a large negative pressure .
this is the standard picture , in which such new ingredient is considered as a source of the _ right hand side _ of the well known friedman - robertson - walker field equations @xcite .
it should be some form of non - clustered and non - zero vacuum energy which , together with the clustered dark matter , drives the global dynamics .
this is the so called `` concordance model '' ( lambda cold dark matter , @xmath0cdm ) which gives , in agreement with the cosmic microwave background radiation ( cmbr ) , large structure scale ( lss ) and supernovae ia ( sneia ) data , a good framework of the today observed universe , but presents several shortcomings as the well known `` coincidence '' and `` cosmological constant '' problems @xcite .
an alternative approach is changing the _ left hand side _ of the field equations , seeing if observed cosmic dynamics can be achieved extending general relativity @xcite . in this different context , it is not required to find out candidates for dark energy and dark matter , that , till now , have not been found , but only the `` observed '' ingredients , which are curvature and baryon matter , have to be taken into account .
considering this point of view , one can think that gravity is not scale - invariant and a room for alternative theories is present @xcite . in this picture ,
even the sensitive detectors for gravitational waves , like bars and interferometers ( i.e. those which are currently in operation and the ones which are in a phase of planning and proposal stages ) @xcite , could , in principle , be important to confirm or to rule out the physical consistency of general relativity or of any other theory of gravitation .
this is because , in the context of extended theories of gravity , some differences between general relativity and the others theories can be pointed out starting from the linearized theory of gravity @xcite . in principle , the most popular dark energy and dark matter models can be achieved considering @xmath1 theories of gravity @xcite , where @xmath2 is the ricci curvature scalar . in the tapestry of extended theories of gravity , it is quite important to understand why we are going to select @xmath1 theories of gravity among several existing modified theories of gravity @xcite .
the motivation on such a choice have been recently well emphasized in the interesting analysis in @xcite .
thus , following the review @xcite , we recall that some @xmath1 theories of gravity are not excluded by requirements of cosmology and solar system tests , i.e are
_ viable_. on the same time , they have quite rich cosmological structure : they may naturally describe the effective ( cosmological constant , quintessence or phantom ) late - time era with a possible transition from deceleration to acceleration thanks to gravitational terms which increase with scalar curvature decrease .
in other words , @xmath1 theories are well tuned with an accelerating standard frw universe .
these theories may be also rewritten as scalar - tensor gravity ( einstein frame ) with the aid of conform transformations .
starting by these considerations , in @xcite two specific viable models have been considered : the one with @xmath3 and @xmath4 terms , and another with @xmath5 terms .
it has been shown that both of the models may lead to the ( cosmological constant or quintessence ) acceleration of the universe as well as an early time inflation . moreover
, the first model seems to pass the solar system tests , i.e. it has the acceptable newtonian limit , no instabilities and no brans - dicke problem ( decoupling of scalar ) in scalar - tensor version @xcite .
these versions of @xmath1 theories may also describe cosmic speed - up at late universe .
@xmath1 theories can be taken into account also in the analysis of the sbgws which , together with the cmbr , would carry , if detected , a huge amount of information on the early stages of the universe evolution @xcite ) . a key role for the production and the detection of this graviton background is played by the adopted theory of gravity @xcite in this case too . in the second section of this review ,
the sbgws is analysed with the aid of the wmap data @xcite .
we emphasize that , in general , in previous works in the literature about the sbgws , old cobe data were used ( see @xcite for example ) . in the third section
we want to face the problem of how the sbgws and @xmath1 gravity can be related , showing , vice versa , that a revealed sbgws could be a powerful probe for a given theory of gravity . in this way
, it will also be shown that the conform treatment of sbgws can be used to parametrize in a natural way f(r ) theories . in this framework
, we will discuss like particular case @xcite some interesting viable @xmath1 theories that have been analysed in refs .
@xcite . in this way ,
a discussion which started in @xcite will be integrated .
the presence and the potential detection of the sbgws is quite important in the framework of the debate on high - frequency gravitational waves ( hfgws ) too .
recently , the importance of hfgws has been emphasized in some papers in the literature @xcite .
we emphasize that this is a review of previous works on the sbgws which have been discussed in various peer - reviewed journals and international conferences .
from our analysis , it will result that the wmap bounds on the energy spectrum and on the characteristic amplitude of the sbgws are greater than the cobe ones , but they are also far below frequencies of the earth - based antennas band . at the end of this section
a lower bound for the integration time of a potential detection with advanced ligo is released and compared with the previous one arising from the old cobe data .
even if the new lower bound is minor than the previous one , it results very long , thus for a possible detection we hope in the laser interferometric space antenna ( lisa ) and in a further growth in the sensitivity of advanced projects . the strongest constraint on the spectrum of the relic sbgws in the frequency range of ground based antennas like bars and interferometers , which is the range @xmath6 , comes from the high isotropy observed in the cmbr .
the fluctuation @xmath7 of the temperature of cmbr from its mean value @xmath8 k varies from point to point in the sky @xcite , and , since the sky can be considered the surface of a sphere , the fitting of @xmath7 is performed in terms of a laplace series of spherical harmonics @xmath9 and the fluctuations are assumed statistically independent ( @xmath10 , @xmath11 ) .
@xmath12 denotes a point on the 2-sphere in eq .
( [ eq : fluttuazioni cbr ] ) , while the @xmath13 are the multipole moments of the cmbr . for details about the definition of statistically independent fluctuations in the context of the temperature fluctuations of cmbr see @xcite .
the wmap data @xcite permit a more precise determination of the rms quadrupole moment of the fluctuations than the cobe data @xmath14 while in the cobe data we had @xcite @xmath15 a connection between the fluctuation of the temperature of the cmbr and the sbgws derives from the _ sachs - wolfe effect _ @xcite .
sachs and wolfe showed that variations in the density of cosmological fluid and gws perturbations result in the fluctuation of the temperature of the cmbr , even if the surface of last scattering had a perfectly uniform temperature @xcite .
in particular , the fluctuation of the temperature ( at the lowest order ) in a particular point of the space is @xmath16 the integral in eq .
( [ eq : null geodesic ] ) is taken over a path of null geodesic which leaves the current spacetime point heading off in the direction defined by the unit vector @xmath12 and going back to the surface of last scattering of the cmbr .
here @xmath17 is a particular choice of the affine parameter along the null geodesic . by using conform coordinates , we have for the metric perturbation @xmath18 and @xmath19 in eq .
( [ eq : null geodesic ] ) is a radial spatial coordinate which goes outwards from the current spacetime point .
the effect of a long wavelength gw is to shift the temperature distribution of cmbr from perfect isotropy . as the fluctuations are very small ( @xmath20 @xcite ) , the perturbations caused by the relic sbgws can not be too large .
the wmap results give rather tight constraints on the spectrum of the sbgws at very long wavelengths .
in @xcite we find a constraint on @xmath21 which derives from the cobe observational limits , given by @xmath22 we included a dimensionless factor @xmath23 which now is just a convenience ( in the past it came from an uncertainty in the value of @xmath24 ) . from the wmap data we have @xmath25 @xcite .
now , the same constraint will be obtained from the wmap data @xcite . because of its specific polarization properties , relic sbgws should generate particular polarization patterns of cmbr anisotropies , but the detection of cmbr polarizations is not fulfilled today @xcite .
thus , an indirect method will be used .
we know that relic gws have very long wavelengths of hubble radius size , then the cmbr power spectrum from them should manifest fast decrease at smaller scales ( high multipole moments ) .
but we also know that scalar modes produce a rich cmbr power spectrum at large multipole moments ( series of acoustic peaks , @xcite ) . then , the properties of tensor modes in the cosmological perturbations of spacetime metric can be extract from observational data using angular cmbr power spectrum combined with large scale structure of the universe .
one can see ( fig .
1 ) that in the range @xmath26 ( the same used in @xcite , but with the old cobe data @xcite ) scalar and tensor contributions are comparable . * figure 1 : the tensor to scalar ratio , adapted from c. corda * * mod
a 22 , 16 , 1167 - 1173 ( 2007 ) * from @xcite , the wmap data give for the tensor / scalar ratio @xmath19 the constraint @xmath27 .
@xmath28 @xcite obtained @xmath29 @xcite .
thus , if one remembers that , at order of hubble radius , the tensor spectral index is @xmath30 @xcite , it results @xmath31 which is greater than the cobe data result of eq .
( [ eq : limite cobe ] ) .
we emphasize that the limit of eq .
( [ eq : limite wmap ] ) is not a limit for any gws , but only for relic ones of cosmological origin , which were present at the era of the cmbr decoupling .
also , the same limit only applies over very long wavelengths ( i.e. very low frequencies ) and it is far below frequencies of the virgo - ligo ( laser interferometric gravitational observatory ) band .
the primordial production of the relic sbgws has been analysed in @xcite ( note : a generalization for f(r ) theories of gravity will be given in the next section following @xcite ) , where it has been shown that in the range @xmath32 the spectrum is flat and proportional to the ratio @xmath33 as the spectrum remains flat at high frequency , the relic sbgws are quite important also in the framework of hfgws , which have been emphasized in some works in the recent literature @xcite .
wmap observations put strongly severe restrictions on the spectrum , as we discussed above . in fig .
2 the spectrum @xmath34 is mapped : the amplitude ( determined by the ratio @xmath35 ) has been chosen to be _ as large as possible , consistent with the wmap constraint _ ( [ eq : limite wmap ] ) .
* figure 2 : the wmap bound , adapted from c. corda - astropart . phys .
, 30 , 209 - 215 ( 2008 ) * nevertheless , because the spectrum falls off @xmath36 at low frequencies @xcite , this means that today , at virgo and lisa frequencies , indicated in fig .
2 , @xmath37 while using the cobe data it was @xmath38(refs .
@xcite ) .
it is interesting to calculate the correspondent strain at @xmath39 , where interferometers like virgo and ligo have a maximum in sensitivity .
the well known equation for the characteristic amplitude @xcite can be used : @xmath40 obtaining @xmath41 then , as for ground - based interferometers a sensitivity of the order of @xmath42 is expected at @xmath39 , four order of magnitude have to be gained in the signal to noise ratio @xcite .
let us analyze smaller frequencies too .
the sensitivity of the virgo interferometer is of the order of @xmath43 at @xmath44 @xcite and in that case it is @xmath45 for a better understanding of the difficulties on the detection of the sbgws , a lower bound for the integration time of a potential detection with advanced ligo is released . for a cross - correlation between two interferometers the signal to noise ratio ( snr ) increases as @xmath46 where @xmath47 is the one - sided power spectral density of the @xmath48 detector @xcite and @xmath49 the well known overlap - reduction function @xcite . assuming two coincident co - aligned detectors @xmath50 with a noise of order @xmath51 ( i.e. a typical value for the advanced ligo sensitivity @xcite ) , one gets @xmath52 for @xmath53 using our result @xmath54 while it is @xmath52 for @xmath55 using previous cobe result @xmath56 . as the overlap reduction function degrades the snr , these results can be considered a solid upper limit for the advanced ligo configuration for the two different values of the spectrum .
the sensitivity of the lisa interferometer will be of the order of @xmath42 at @xmath57 @xcite and in that case it is @xmath58 then , a stochastic background of relic gravitational waves could be in principle detected by the lisa interferometer .
we also hope in a further growth in the sensitivity of advanced projects .
we emphasize that the assumption that all the tensor perturbation in the universe are due to a sbgws is quite strong , but our results ( [ eq : limite spettro wmap ] ) , ( [ eq : limite per lo strain ] ) , ( [ eq : limite per lo strain2 ] ) and ( [ eq : limite per lo strain3 ] ) can be considered like upper bounds . resuming , in this section the sbgws has been analysed with the aid of the wmap data , while previous works in the literature used the old cobe data , seeing that the predicted signal for these relic gws is very weak . from our analysis
it resulted that the wmap bound on the energy spectrum and on the characteristic amplitude of the sbgws are greater than the cobe ones , but they are also far below frequencies of the earth - based antennas band .
in fact , the integration time of a potential detection with advanced interferometers is very long , thus , for a possible detection , we have to hope in a further growth in the sensitivity of advanced ground based projects and in the lisa interferometer .
gws are the perturbations @xmath59 of the metric @xmath60 which transform as three - tensors . following @xcite ,
the gw - equations in the tt gauge are @xmath61 where @xmath62 is the usual dalembert operator and these equations are derived from the einstein field equations deduced from the einstein - hilbert lagrangian density @xmath63 @xcite .
clearly , matter perturbations do not appear in ( [ eq : dalembert ] ) since scalar and tensor perturbations do not couple with tensor perturbations in einstein equations .
the latin indices run from 1 to 3 , the greek ones from 0 to 3 .
our task is now to derive the analogous of eqs .
( [ eq : dalembert ] ) assuming a generic theory of gravity given by the action @xmath64 where , for a sake of simplicity , we have discarded matter contributions .
a conform analysis will help in this goal .
in fact , assuming the conform transformation @xmath65 where the conform rescaling @xmath66 has been chosen being the prime the derivative in respect to the ricci curvature scalar and @xmath67 the `` conform scalar field '' , we obtain the conform equivalent hilbert - einstein action @xmath68,\label{eq : conform}\ ] ] where @xmath69 is the conform scalar field contribution derived from @xmath70 and @xmath71 in any case , as we will see , the @xmath69-term does not affect the gws - tensor equations , thus it will not be considered any longer ( note : a scalar component in gws is often considered @xcite , but here we are taking into account only the genuine tensor part of stochastic background ) .
starting from the action ( [ eq : conform ] ) and deriving the einstein - like conform equations , the gws equations are @xmath72 expressed in the conform metric @xmath73 as no scalar perturbation couples to the tensor part of gravitational waves , it is @xmath74 which means that @xmath75 is a conform invariant . as a consequence , the plane wave amplitude @xmath76 where @xmath77 is the polarization tensor , are the same in both metrics . in any case
the dalembert operator transforms as @xmath78 and this means that the background is changing while the tensor wave amplitude is fixed . in order to study the cosmological stochastic background , the operator ( [ eq : quadratello ] )
can be specified for a friedman - robertson - walker metric @xcite , and eq .
( [ eq : dalembert conf ] ) becomes @xmath79 being @xmath80 , @xmath81 the scale factor and @xmath82 the wave number .
it has to be emphasized that eq .
( [ eq : evoluzione h ] ) applies to any f(r ) theory whose conform transformation can be defined as @xmath83 the solution , i.e. the gw amplitude , depends on the specific cosmological background ( i.e. @xmath81 ) and the specific theory of gravity .
for example , assuming power law behaviors for @xmath81 and @xmath84 that is @xmath85 it is easy to show that general relativity is recovered for @xmath86 while @xmath87 is the relation between the parameters for a generic @xmath88 where @xmath89 with @xmath90 @xcite .
equation ( [ eq : evoluzione h ] ) can be recast in the form @xmath91 whose general solution is @xmath92.\label{eq : sol ev h}\ ] ] @xmath93 s are bessel functions and @xmath94 while @xmath95 @xmath96 and @xmath97 are constants related to the specific values of @xmath98 and @xmath99 the time units are in terms of the hubble radius @xmath100 @xmath101 is a radiation - like evolution ; @xmath102 is a dust - like evolution , @xmath103 labels power - law inflationary phases and @xmath104 is a pole - like inflation . from eq .
( [ eq : m n ] ) , a singular case is for @xmath105 and @xmath106 it is clear that the conform invariant plane - wave amplitude evolution of the tensor gw strictly depends on the background .
now , let us take into account the issue of the production of relic gws contributing to the stochastic background .
several mechanisms can be considered as cosmological populations of astrophysical sources @xcite , vacuum fluctuations , phase transitions @xcite and so on . in principle
, we could seek for contributions due to every high - energy physical process in the early phases of the universe evolution .
it is important to distinguish processes coming from transitions like inflation , where the hubble flow emerges in the radiation dominated phase and processes , like the early star formation rates , where the production takes place during the dust dominated era . in the first case ,
stochastic gws background is strictly related to the cosmological model .
this is the case we are considering here which is , furthermore , also connected to the specific theory of gravity .
in particular , one can assume that the main contribution to the stochastic background comes from the amplification of vacuum fluctuations at the transition between an inflationary phase and the radiation - dominated era .
however , in any inflationary model , we can assume that the relic gws generated as zero - point fluctuations during the inflation undergoes adiabatically damped oscillations ( @xmath107 ) until they reach the hubble radius @xmath108 this is the particle horizon for the growth of perturbations .
on the other hand , any other previous fluctuation is smoothed away by the inflationary expansion .
the gws freeze out for @xmath109 and re - enter the @xmath110 radius after the reheating in the friedman era @xcite .
re - entering in the radiation dominated or in the dust - dominated era depends on the scale of the gw .
after the re - entering , gws can be detected by their sachs - wolfe effect on the temperature anisotropy @xmath111 at the decoupling @xcite . when @xmath67 acts as the inflaton @xcite we have @xmath112 during the inflation . considering the conformal time @xmath113 , eq .
( [ eq : evoluzione h ] ) reads@xmath114 where @xmath115 and derivation is with respect to @xmath116 inflation means that @xmath117 and then @xmath118 and @xmath119 the exact solution of ( [ eq : evoluzione h 3 ] ) is @xmath120.\label{eq : sol ev h2}\ ] ] inside the @xmath121 radius it is @xmath122 furthermore , considering the absence of gravitons in the initial vacuum state , we have only negative - frequency modes and then the adiabatic behavior is @xmath123 at the first horizon crossing ( @xmath124 ) , the averaged amplitude @xmath125 of the perturbations is @xmath126 when the scale @xmath127 grows larger than the hubble radius @xmath128 the growing mode of evolution is constant , that it is frozen .
this situation corresponds to the limit @xmath129 in eq .
( [ eq : sol ev h2 ] ) .
since @xmath67 acts as the inflaton field , it is @xmath130 at re - entering ( after the end of inflation ) .
then , the amplitude @xmath131 of the wave is preserved until the second horizon crossing after which it can be observed , in principle , as an anisotropy perturbation of the cbr .
it can be shown that @xmath132 is an upper limit to @xmath131 since other effects can contribute to the background anisotropy @xcite . from this consideration , it is clear that the only relevant quantity is the initial amplitude @xmath133 in equation ( [ eq : sol ev h3 ] ) which is conserved until the re - entering .
such an amplitude directly depends on the fundamental mechanism generating perturbations .
inflation gives rise to processes capable of producing perturbations as zero - point energy fluctuations .
such a mechanism depends on the adopted theory of gravitation and then @xmath111 could constitute a further constraint to select a suitable f(r)-theory . considering a single graviton in the form of a monochromatic wave , its zero - point amplitude is derived through the commutation relations @xmath134=i\delta^{3}(x - y),\label{eq : commutare}\ ] ] calculated at a fixed time @xmath135 where the amplitude @xmath136 is the field and @xmath137 is the conjugate momentum operator .
writing the lagrangian for @xmath136 @xmath138 in the conformal frw metric @xmath139 ( @xmath136 is a conform invariant ) , we obtain @xmath140 equation ( [ eq : commutare ] ) becomes@xmath141=i\frac{\delta^{3}(x - y)}{e^{2\phi}a^{3}}\label{eq : commutare 2}\ ] ] and the fields @xmath136 and @xmath142 can be expanded in terms of creation and annihilation operators @xmath143,\label{eq : crea}\ ] ] @xmath144.\label{eq : distruggi}\ ] ] the commutation relations in conform time are then @xmath145=i\frac{(2\pi)^{3}}{e^{2\phi}a^{3}}.\label{eq : commutare 3}\ ] ] putting ( [ eq : sol ev h3 ] ) and ( [ eq : ah ] ) , we obtain @xmath146 where @xmath147 and @xmath67 are calculated at the first horizon crossing and then @xmath148 which means that the amplitude of gws produced during inflation directly depends on the given f(r ) theory being @xmath149 explicitly , it is @xmath150 this result deserves some discussion and can be read in two ways . from one side
the amplitude of relic gws produced during inflation depends on the given theory of gravity that , if different from general relativity , gives extra degrees of freedom which assume the role of inflaton field in the cosmological dynamics @xcite . on the other hand ,
the sachs - wolfe effect related to the cmbr temperature anisotropy could constitute a powerful tool to test the true theory of gravity at early epochs , i.e. at very high redshift .
this probe , related with data a medium @xcite and low redshift @xcite , could strongly contribute to reconstruct cosmological dynamics at every scale , to further test general relativity or to rule out it against alternative theories and to give constrains on the sbgws , if f(r ) theories ares independently probed at other scales . for a better understanding of the issue , it is important to analyse in spirit of our discussion some particular viable model @xcite , for example the interesting viable models in @xcite . in this way
we will also integrate an analysis which started in @xcite .
let us start by the model proposed in @xcite . in this case
it is @xmath151 which , in our case , gives the conform scalar field @xmath152 here @xmath153 @xmath154 @xmath155
are positive constants and @xmath98 is a positive integer @xcite .
these equations can not satisfy the curvature @xmath156 in the present universe for which it is @xmath157 because this implies a singularity in the amplitude of the relic gws ( [ ah4 ] ) . in the model proposed in @xcite
it is @xmath158\label{eq : f2}\ ] ] and@xmath159 here @xmath153 and @xmath160 are positive constants @xcite .
thus , in the case of flat universe
@xmath161 , it is @xmath162 and @xmath163 on the other hand , in the case in which it is @xmath164 in the present universe it is @xcite @xmath165,\label{eq : f3}\ ] ] i.e. @xmath166 plays the role of an effective cosmological constant . in this case @xmath131 increases with increasing curvature .
@xmath131 has a maximum which is@xmath167 in the case in which @xmath168 two pretty realistic f(r ) theories have been proposed in @xcite .
the first is @xmath169 which gives the correspondent conform scalar field@xmath170 here @xmath98 is a positive integer @xmath171 where @xmath172 represents the effective cosmological constant and @xmath173 @xcite . in this case
the amplitude of the stochastic background admits a minimum at the point @xmath174 where @xmath175 satisfies @xmath176 and the minimum is given by @xmath177 the second proposal in @xcite is the theory @xmath178 which gives the correspondent conform scalar field @xmath179 here @xmath153 @xmath154 @xmath180 and @xmath181are constants , while @xmath182 is identified with the effective cosmological constant at the inflation era if @xmath183 the case of flat universe is interesting .
we get@xmath184 resuming , in this section it has been shown that amplitudes of tensor gws are conform invariant and their evolution depends ond the cosmological sbgws .
such a background is tuned by a conform scalar field which is not present in the standard general relativity . assuming that primordial vacuum fluctuations produce a sbgws , beside scalar perturbations , kinetic distortions and so on , the initial amplitude of these ones is function of the f(r)-theory of gravity and then the sbgws can be , in a certain sense , `` tuned '' by the theory .
vice versa , data coming from the sachs - wolfe effect could contribute to select a suitable f(r)-theory which can be consistently matched with other observations .
we have also considered some interesting examples which have been recently discussed in the literature @xcite .
however , further and accurate studies are needed in order to test the relation between sachs - wolfe effect and f(r ) gravity .
this goal could be achieved in the next future through the forthcoming space ( lisa ) and ground based ( virgo , advanced ligo ) interferometers and li - baker hfgw detectors .
this paper has been a review of previous work on the sbgws which has been discussed in various peer - reviewed journals and international conferences .
the sbgws has been analysed with the aid of the wmap data while , in general , in previous works in the literature about the sbgws , old cobe data were used .
after this , it has been shown how the sbgws and f(r ) gravity can be related , showing , vice versa , that a revealed sbgws could be a powerful probe for a given theory of gravity . in this way
, it has also been shown that the conform treatment of sbgws can be used to parametrize in a natural way f(r ) theories .
some interesting examples which have been recently discussed in the literature , see @xcite , have been analysed .
the presence and the potential detection of the sbgws is also quite important in the framework of the debate on high - frequency gravitational waves ( hfgws ) .
the importance of hfgws has been recently emphasized in some papers in the literature and the ground based li - baker hfgw has been proposed to be sensitive at @xmath185 ghz to @xmath186 amplitude hfgws @xcite .
i strongly thank the referees for helpful advices which permitted to substantially improve the paper .
capozziello s and corda c - int .
d * 15 * 1119 -1150 ( 2006 ) ; corda c - _ response of laser interferometers to scalar gravitational waves_- talk in the _ gravitational waves data analysis workshop in the general relativity trimester of the institut henri poincare - _ paris 13 - 17 november 2006 , on the web in www.luth2.obspm.fr/ihp06/workshops/gwdata/corda.pdf | this paper is a review of previous works on the stochastic background of gravitational waves ( sbgws ) which has been discussed in various peer - reviewed journals and international conferences .
the sbgws is analyzed with the aid of the wilkinson microwave anisotropy probe ( wmap ) data .
we emphasize that , in general , in previous works in the literature about the sbgws , old cosmic background explorer ( cobe ) data were used .
after this , we want to face the problem of how the sbgws and f(r ) gravity ( where f(r ) is a function of the ricci scalar r ) can be related , showing , vice versa , that a revealed sbgws could be a powerful probe for a given theory of gravity . in this way
, it will also be shown that the conform treatment of sbgws can be used to parametrize in a natural way f(r ) theories . some interesting examples which have been recently discussed in the literature will be also analysed .
the presence and the potential detection of the sbgws is quite important in the framework of the debate on high - frequency gravitational waves ( hfgws ) too .
recently , the importance of hfgws has been emphasized in some papers in the literature .
associazione scientifica galileo galilei , via pier cironi 16 , i-59100 prato italy and 0574news.it - sezione scientifico - tecnologica , via sante pisani 46 - 59100 prato , italy _ e - mail address : _
pacs numbers : 04.80.nn , 04.30.nk , 04.50.+h keywords : gravitational waves ; extended theories of gravity ; stochastic background ; high - frequency gravitational waves |
a prolonged , quiescent period followed the creation of the universe . eventually , minuscule dark matter density perturbations initiated the collapse of dark matter , forming gravitational potential wells and giving rise to the condensation of primordial gas clouds ( e.g. , * ? ? ? * ) .
this process furnished the universe with the earliest building blocks , laying the foundation for cosmological structure .
separate building blocks ( i.e. , gas fragments ) underwent independent chemical enrichment before conglomerating to form larger systems .
evidence of this hierarchical formation @xcite remains observable today through the accretion of satellite systems onto the milky way halo .
although the formation history of the galaxy is tangled and chaotic , it serves as an excellent and more importantly , accessible laboratory to investigate the evolution of the universe since the earliest times .
many stars in our galaxy have formed in - situ , but a significant fraction have also been added through the accretion of smaller systems onto the milky way ( for a review see * ? ? ? * ) .
as these satellite systems move in towards the galaxy s gravitational potential , they are disrupted by tidal forces , causing loosely gravitationally bound stars to be strewn in forward and trailing directions as tidal tails or ` stellar streams ' .
the observational evidence for ongoing accretion has significantly increased in the last decade since the discovery of the sagittarius dwarf spheroidal undergoing tidal disruption @xcite . amongst the known substructures in the halo , the orphan stream is particularly interesting .
@xcite and @xcite independently discovered the stream , spanning over @xmath2 in the sky from ursa major in the north to near sextans in the south , where the sdss coverage ends .
the stream has properties that are unique from other halo substructures .
it has an extremely low surface brightness , ranging from 32 - 40mag arcsec@xmath3 , and a full width half - maximum of @xmath02@xmath4 on sky . at a distance of @xmath020kpc @xcite
, this corresponds to @xmath0700pc .
the cross - section and luminosity of the stream are directly related to the mass and velocity dispersion of the parent satellite @xcite .
the orphan stream width is significantly broader than every known globular cluster tidal tail @xcite and larger than the tidal diameter of all known globular clusters ( * ? ? ?
* edition ) .
as @xcite notes , if the cross - section of the stream is circular , then in a simple logarithmic potential with @xmath5 km s@xmath6 a reasonable first - order approximation for the milky way the expected random velocities of stars would be required to be @xmath7 km s@xmath6 in order to produce the stream width .
such a velocity dispersion is significantly larger than the expected random motions of stars that have been weakly stripped from a globular cluster , implying that the orphan stream s progenitor mass must be much larger than a classical globular cluster .
photometry indicates the stream is metal - poor , implying that negligible star formation has occurred since infall began several billion years ago . for a stream of this length to remain structurally coherent over such a long timescale ,
the progenitor is also likely to be dark - matter dominated . in their discovery papers , @xcite and @xcite concluded that the likely parent of the orphan stream is a low - luminosity dwarf spheroidal ( dsph ) galaxy .
@xcite were able to map the distance and velocity of the stream across the length of the sdss catalogue using blue horizontal branch ( bhb ) and f - turnoff stars .
they found the stream distance to vary between 19 - 47kpc , extending the 20 - 32kpc distance measurements made by @xcite , and has been extended to 55kpc by @xcite .
@xcite note an increase in the density of the orphan stream near the celestial equator @xmath8 , proposing the progenitor may be close to this position .
the authors attempted to extend their trace of the stream using southern survey data ( e.g. supercosmos * ? ? ? * ) , but to no avail ; the stream s surface brightness is lower than the survey faint limit .
it is still unclear whether the stream extends deep into the southern sky . @xcite
observe an increase in surface brightness near the celestial equator , but given the stream is closest near the celestial equator @xcite , an increase in stream density may be expected given a constant absolute magnitude .
nevertheless , with sdss photometry and radial velocities from the segue catalog , @xcite were able to derive a prograde orbit with an eccentricity , apogalacticon and perigalacticon of @xmath9 , 90kpc and 16.4kpc respectively .
the ability to accurately trace the orphan stream to such extreme distances would make a powerful probe for measuring the galactic potential ( e.g. see * ? ? ?
* ) . from their simulations ,
@xcite find a halo and disk mass of @xmath10 , and similarly @xcite find @xmath11 , which is @xmath060% lower than that found by @xcite and @xcite , and slightly lower than the virial mass of @xmath12 derived by @xcite .
metallicities derived from segue medium - resolution spectra confirm photometric estimates indicating that the stream is metal - poor .
a mean metallicity of [ fe / h]@xmath13dex is found from bhb stars , with a range extending from @xmath141.3 to @xmath153dex @xcite .
@xcite also find a wide range in metallicities by tracing rr lyrae stars along the stream : [ fe / h ] = @xmath16 to @xmath17dex .
however if f - turnoff stars from the @xcite sample are included , the metallicity distribution function extends more metal - rich from @xmath153 to @xmath140.5dex .
the situation is complicated by interlopers and small number statistics , so the full shape of the metallicity distribution function ( mdf ) is unknown . to this end , @xcite observed low - resolution spectroscopy for hundreds of stars towards the stream at the celestial equator .
the authors targeted the less numerous k - giants ( * ? ? ?
* ; * ? ? ?
* where a mere 1.3 red giant branch stars are expected per square degree ) and found a very weak detection of the stream from kinematics alone . using wide selections in velocity , distance , proper motions , metallicities , and surface gravity , they identified nine highly likely orphan stream giants .
the velocity dispersion of their candidates is within the observational errors ( @xmath18 km s@xmath6 ) , suggesting the stream is kinematically cold along the line - of - sight .
like @xcite ( and independently confirmed by @xcite ) , they also found an extended range in metallicities : two stars below [ fe / h ] @xmath19 and two stars near [ fe / h ] @xmath20dex , all of which are consistent with stream membership .
the mean metallicity of their sample was [ fe / h]@xmath21dex , with a wide dispersion of @xmath22 } ) = 0.56$]dex .
it appears the orphan stream may have an extremely wide range in metallicity , consistent with the internal chemical enrichment typically observed in dwarf spheroidal ( dsph ) galaxies @xcite .
as the name suggests , the orphan stream s parent satellite has yet to be found . in an effort to identify a progenitor ,
a number of systems have been identified as being plausibly associated with the orphan stream .
these include the linear complex a hi clouds , as well as the globular clusters palomar 1 , arp 2 , terzan 7 and ruprecht 106 .
these systems were all identified along the great circle path of the stream .
the low - luminosity dwarf satellites segue 1 and ursa major ii also lie along the great circle , although segue 1 was considered an extended globular cluster until recently @xcite .
@xcite first noted a possible association between the orphan stream , ursa major
ii and complex a. @xcite and @xcite explored the possible association between the orphan stream with complex a and ursa major ii , respectively . in the best - fitting complex a model , the predicted heliocentric velocities
did not match those found by @xcite , or later observations by @xcite .
the expected distances were also under - estimated by a factor of @xmath233 when compared with observations , making the association with complex a tenuous at best . in the ursa major ii scenario , the stream s on - sky position was required to exactly overlap with a previous wrap , a somewhat contrived and unlikely scenario .
@xcite found that the ursa major ii - orphan stream connection was also not compelling , as the observed stream kinematics were not consistent with the ursa major ii model .
simulations involving the complex a and ursa major ii associations introduced an a priori assumption that the object ( e.g. ursa major ii or complex a ) was related to the orphan stream , and consequently found an orbit to match .
in contrast , @xcite approached the problem by fitting an orbit to a single wrap of the data , without assuming a parent satellite a priori .
their @xmath24-body simulations were inconsistent with either a complex a or ursa major ii association . instead
, the authors favor a progenitor with a luminosity @xmath25 or an absolute magnitude @xmath26 , consistent with the observation by @xcite of @xmath27 .
simulations by @xcite suggest the progenitor may be similar to the present - day ` classical ' milky way dwarfs like carina , draco , leo ii or sculptor , but would be very close to being fully disrupted , which they suggest has occurred over the last 5.3 gigayear .
time of infall is a critical inference .
longer timescales produce streams that are too wide and diffuse , whereas shorter timescales do not reproduce the @xmath2360@xmath4 stream length .
the degeneracies between these simulation parameters are important to note , but in any case , there are robust conclusions that can be drawn irrespective of those degeneracies . for example , @xcite note that a globular cluster has a central density too high ( @xmath28kpc@xmath29 ) to be fully disrupted along their orphan stream orbit within a hubble time . given this constraint , and the lower limit on luminosity ( @xmath30 ) , a globular cluster progenitor seems unlikely from their models .
in addition to the work by @xcite , @xmath24-body simulations by @xcite exclude all known halo globular clusters as possible progenitors .
they conclude with the postulation of two possible scenarios : the progenitor is either an undiscovered satellite located between @xmath31 and @xmath32 , or segue 1 is the parent system .
segue 1 , an ultrafaint dwarf galaxy @xcite , resides at a similar distance ( @xmath023kpc ) along the great circle path of the orphan stream .
segue 1 also shares velocities that are consistent with the orphan stream at its nearest point .
moreover , both segue 1 and the orphan stream exhibit low velocity dispersions : on the order of @xmath04 km s@xmath6 . the similarities in position , distance , and velocities between the two systems are indeed striking . however , it is not the only system alleged to be associated with the orphan stream .
given the extended apogalaction of 90kpc in the stream orbit found by @xcite , @xcite reasoned the stream may be the tidal tail of the massive globular cluster ngc 2419 .
this system is the most distant and luminous outer halo globular cluster known ( @xmath085kpc , @xmath33 ) .
it is unlike any other globular cluster in the milky way .
spectroscopic studies have confirmed photometric observations by @xcite that the system is metal - poor ( [ fe / h ] @xmath34 ) , and identified a remarkable anti - correlation between mg and k abundances @xcite .
the level of magnesium depletion ( [ mg / fe ] @xmath35 ) is not seen anywhere else in the galaxy or its satellite systems , neither is the enhanced potassium enrichment ( up to [ k / fe ] = 2 ) at the opposite end of this peculiar mg - k anti - correlation . if ngc 2419 is the parent of the orphan stream , then this unprecedented chemical signature ought to exist in disrupted stream members as an example of chemical tagging ( e.g. , see * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
lccccccccccc hd 41667 & 06:05:03.7 & @xmath1432:59:36.8 & 8.52 & 2011 - 03 - 13 & 23:40:52 & 1.01 & 90 & 272 & 297.8 & 1.0 + hd 44007 & 06:18:48.6 & @xmath1414:50:44.2 & 8.06 & 2011 - 03 - 13 & 23:52:18 & 1.03 & 30 & 239 & 163.4 & 1.3 + hd 76932 & 08:58:44.2 & @xmath1416:07:54.2 & 5.86 & 2011 - 03 - 14 & 00:16:47 & 1.16 & 30 & 289 & 119.2 & 1.2 + hd 122563 & 14:02:31.8 & @xmath3609:41:09.9 & 6.02 & 2011 - 03 - 13 & 07:15:04 & 1.28 & 30 & 230 & 23.4 & 1.0 + hd 136316 & 15:22:17.2 & @xmath1453:14:13.9 & 7.65 & 2011 - 03 - 14 & 09:37:26 & 1.12 & 90 & 335 & 38.2 & 1.1 + hd 141531 & 15:49:16.9 & @xmath3609:36:42.5 & 9.08 & 2011 - 03 - 14 & 09:52:00 & 1.31 & 90 & 280 & 2.6 & 1.0 + hd 142948 & 16:00:01.6 & @xmath1453:51:04.1 & 8.03 & 2011 - 03 - 14 & 09:45:12 & 1.11 & 90 & 271 & 30.3 & 0.9 + oss 3 ( l ) & 10:46:50.6 & @xmath1400:13:17.9 & 17.33 & 2011 - 03 - 14 & 01:51:07 & 1.36 & 4@xmath372500 & 48 & 217.9 & 1.0 + oss 6 ( h ) & 10:47:17.8 & @xmath3600:25:06.9 & 16.09 & 2011 - 03 - 14 & 00:25:37 & 2.00 & 3@xmath371600 & 59 & 221.2 & 1.0 + oss 8 ( h ) & 10:47:30.3 & @xmath1400:01:22.6 & 17.25 & 2011 - 03 - 14 & 04:44:04 & 1.16 & 5@xmath371900 & 49 & 225.9 & 1.0 + oss 14 ( h ) & 10:49:08.3 & @xmath3600:01:59.3 & 16.27 & 2011 - 03 - 15 & 00:32:17 & 1.88 & 4@xmath371400 & 48 & 225.1 & 1.0 + oss 18 ( m ) & 10:50:33.7 & @xmath3600:12:18.3 & 17.82 & 2011 - 03 - 15 & 02:12:46 & 1.30 & 4@xmath372100 & 31 & 247.8 & 1.2 in this study , we present an analysis of high - resolution spectroscopic observations for five orphan stream candidates . the observations and data reduction are outlined in section [ sec : observations ] . in section [ sec : analysis ] we describe the details of our analysis to infer stellar parameters and chemical abundances .
we discuss our results in section [ sec : discussion ] , including the implications for association between the orphan stream and the two currently alleged stream progenitors : segue 1 and ngc 2419 .
finally , we conclude in section [ sec : conclusions ] with a summary of our findings .
high - resolution spectra for five orphan stream candidates and seven well - studied standard stars have been obtained with the magellan inamori kyocera echelle ( mike ) spectrograph @xcite on the magellan clay telescope .
these objects were observed in march 2011 using a 1 wide slit in mean seeing of 0.9 .
this slit configuration provides a continuous spectral coverage from 333 nm to 915 nm , with a spectral resolution of @xmath38 in the blue arm and @xmath39 in the red arm .
a minimum of 10 exposures of each calibration type ( biases , flat fields , and diffuse flats ) were observed in the afternoon of each day , with additional flat - field and th - ar arc lamp exposures performed throughout the night to ensure an accurate wavelength calibration .
the details of our observations are tabulated in table [ tab : observations ] .
the signal - to - noise ( @xmath40 ) ratio for the standard stars exceeds 200 per pixel , and varies between 30 - 60 for the orphan stream candidates .
the candidates were chosen from the low - resolution spectroscopic study of @xcite . from their classification ,
three of the selected stream candidates have high probability of membership to the orphan stream .
one target was classified with medium probability , and another with a low probability of membership .
initially we planned to observe many more high - priority targets .
however , after our last exposure of target star oss 18 , inclement weather forced us to relinquish the remainder of our observing time .
the data were reduced using the carpy pipeline written by d. kelson .
every reduced echelle aperture was carefully normalised using cubic splines with defined knot spacing .
extracted apertures were stacked together and weighted by their inverse variance to provide a continuous normalised spectrum for each object .
each normalised , stitched spectrum was cross - correlated against a synthetic template to measure the radial velocity of each star .
this was performed using a python implementation of the @xcite method .
the wavelength region employed was from @xmath41 nm , and a synthetic spectrum of a metal - poor giant was used as the rest template .
heliocentric corrections have been applied to our radial velocity measurements , and the resultant heliocentric velocities are shown in table [ tab : observations ] .
atomic data for absorption lines has been taken from @xcite , and these data are listed with their measured equivalent widths ( ews ) in table [ tab : atomic - data ] .
ews for all atomic transitions were measured using the automatic profile fitting algorithm described in @xcite .
while this technique is accurate , and robust against blended lines as well as strong changes in the local continuum @xcite , every fitted profile was visually inspected for quality . spurious or false - positive measurements were removed and of order @xmath05 ew measurements ( of 528 atomic transitions in our line list as they were not significantly detected in any stars .
see section [ sec : chemical - abundances ] for details . ] ) were manually re - measured for each object .
we excluded transitions with reduced equivalent widths ( rew ) , @xmath42 , in order to avoid using lines near the flat region of the curve - of - growth .
given the range in metallicities for the orphan stream candidates ( see section [ sec : analysis - metallicity ] ) , these restrictions resulted in only 14 fei and 4 feii acceptable transitions for our most metal - poor candidate , oss 4 .
with so few lines available , minute changes in stellar parameters caused large trends and variations in the fe line abundances .
this resulted in a poor solution while performing the excitation and ionization equilibria . at this point
, we opted to supplement our line list with transitions from @xcite .
each additional transition was inspected in our most metal - rich candidate ( oss 18 ) to ensure that it was not blended with other features .
blended transitions were not added . as a result , the minimum number of acceptable fei and feii transitions for any star increased to 48 and 12 , respectively . clcrcccccccccccccc[t ! ] 630.030 & o i & 0.00 & 9.72 & & 14.8 & & 6.8 & & 38.2 & 35.4 & + 636.378 & o i & 0.02 & 10.16 & 11.6 & & & & 11.6 & 14.7 & 12.7 & + 568.819 & na i & 2.11 & 0.42 & 71.9 & 30.9 & 57.4 & & 29.3 & 38.1 & & 55.2 + 615.423 & na i & 2.10 & 1.53 & 10.7 & & 7.8 & & & 5.2 & 28.4 & + 616.075 & na i & 2.10 & 1.23 & 16.6 & & 13.4 & & 4.6 & 7.0 & 43.4 & 17.4
we employed the 1d plane - parallel model atmospheres of @xcite to infer stellar parameters .
these @xmath1-enhanced models assume that absorption lines form under local thermal equilibrium ( lte ) , ignore convective overshoot and any center - to - limb variations .
we have interpolated within a grid of these model atmospheres following the prescription in @xcite .
effective temperatures for all stars have been found by excitation balance of neutral iron lines . since all ew measurements were visually inspected , we generally identified no outlier measurements during this stage .
the most number of fei outliers ( @xmath44 ) removed while determining the effective temperature for any star was three . for each iteration in temperature
, a linear fit was made to the data ( @xmath45 , @xmath46(fei ) ) .
this slope was minimized with successive iterations of effective temperature .
final slopes less than @xmath47dex ev@xmath6 were considered to be converged .
the effective temperatures of our standard stars are in excellent agreement with the literature . on average ,
our effective temperatures are 13k cooler than the references listed in table [ tab : stellar - parameters ] .
the largest discrepancy exists for hd 122563 , a cool metal - poor giant . for extremely metal - poor stars ,
effective temperatures found through excitation balance are known to produce systematically cooler temperatures than those deduced by other methods ( e.g. * ? ? ?
we have chosen to remain consistent with the excitation balance approach , and accept the systematically cooler temperature of 4358k .
the other noteworthy temperature deviant is hd 142948 , where we find a temperature 337k hotter than that found by @xcite .
the reason for this discrepancy is not obvious .
microturbulence is necessary in 1d model atmospheres to represent large scale , 3d turbulent motions .
the correct microturbulence will ensure that lines of the same species that are formed in deep and shallow photospheric depths will yield the same abundance .
we have solved for the microturbulence by demanding a zero - trend in rew and abundance for all neutral iron lines .
the resultant gradient between rew and abundance is typically @xmath49dex , and the largest slope in any star is 0.004dex .
the surface gravity for all stars has been inferred through the ionization balance of neutral and single ionized fe lines .
we iterated on surface gravity until the mean fei abundance matched the mean feii abundance to within 0.01dex .
this process was performed in concert while solving for all other stellar parameters . as a consequence of a systematically cooler temperature in hd 122563 ,
we have obtained a somewhat lower surface gravity for this star , such that it sits above a metal - poor isochrone on a hertzsprung - russell diagram .
modulo hd 122563 , the surface gravities for our standard stars are in good agreement with the literature : a mean offset of 0.22dex is observed ( our @xmath50 values are lower ) .
this difference remains within the mutual 1@xmath51 uncertainties .
lcccccccccccl[h ! ] + + hd 41667 & 4643 & 1.54 & 1.81 & 1.18 & & & 4605 & 1.88 & 1.44 & 1.16 & @xcite + hd 44007 & 4820 & 1.66 & 1.70 & 1.69 & & & 4850 & 2.00 & 2.20 & 1.71 & @xcite + hd 76932 & 5835 & 3.93 & 1.42 & 0.95 & & & 5849 & 4.11 & & 0.88 & @xcite + hd 122563 & 4358 & 0.14 & 2.77 & 2.90 & & & 4843 & 1.62 & 1.80 & 2.54 & @xcite + hd 136316 & 4347 & 0.44 & 2.15 & 1.93 & & & 4414 & 0.94 & 1.70 & 1.90 & @xcite + hd 141531 & 4373 & 0.52 & 2.05 & 1.65 & & & 4280 & 0.70 & 1.60 & 1.68 & @xcite + hd 142948 & 5050 & 2.39 & 1.83 & 0.64 & & & 4713 & 2.17 & 1.38 & 0.77 & @xcite + + + oss 3 ( l ) & 5225 & 3.16 & 1.10 & 0.86 & & & & & & 1.31 & @xcite + oss 6 ( h ) & 4554 & 0.70 & 2.00 & 1.75 & & & & & & 1.84 & @xcite + oss 8 ( h ) & 4880 & 1.71 & 1.86 & 1.62 & & & & & & 1.62 & @xcite + oss 14 ( h ) & 4675 & 1.00 & 2.53 & 2.66 & & & & & & 2.70 & @xcite + oss 18 ( m ) & 5205 & 2.91 & 1.88 & 0.62 & & & & & & 0.90 & @xcite the final stage of iterative stellar parameter analysis is to derive the total metallicity . for these analyses
we adopt the mean [ fei / h ] abundance as the overall metallicity [ m / h ] .
a difference of @xmath53 } - \langle\textrm{[fe\,\textsc{i}/h]}\rangle| \leq 0.01}$]dex was considered an acceptable model for the data . excluding hd 122563 , the mean metallicity difference for all other standard stars is 0.03dex less than values in the literature , with a standard deviation of 0.05dex .
when the entire standard sample is included , these values change to [email protected] . acknowledging that our systematically low temperature for hd 122563 has resulted in differing stellar parameters , the rest of our standard stars exhibit excellent agreement with previous studies .
we note that this effect is likely significantly smaller for the metal - poor star oss 14 , as this star is further down the giant branch with an effective temperature @xmath0300k hotter than hd 122563 .
thus , we can be confident in the metallicity determination for oss 14 . the offsets in metallicities between the values we derive from high - resolution spectroscopy and those found by @xcite from low - resolution spectroscopy are noticeable .
for the high probability targets ( oss 6 , 8 and 14 ) the agreement is excellent : @xmath55[fe / h ] = + 0.04 , 0.0 , + 0.09dex , respectively .
we note that the uncertainties adopted in @xcite are of the order @xmath56dex ; the discrepancies are well within the 1@xmath51 uncertainties of _ either _ study . the largest difference between this study and that of @xcite is in the lowest probability target ( oss 3 ) , where we find a metallicity that is @xmath360.45dex higher .
it is reassuring that the metallicities of our high probability stars show the best agreement with the low - resolution measurements .
candidates were classified by @xcite to have a low , medium , or high probability of membership with the orphan stream .
this classification was dependent on a number of observables , including metallicity .
however , the metallicity determinations by @xcite were calculated with the implied assumption that these stars were at a distance of @xmath020kpc . the fact that our metallicities from high - resolution spectra are in excellent agreement with these initial metallicities indicates that our initial assumption was correct , and these high probability targets are at the approximate distance to the orphan stream . during excitation and ionization equilibria assessments ,
each fitted slope has an associated uncertainty due to the scatter in iron abundances .
we have varied the effective temperature and microturbulence to match the formal uncertainty in each fei slope .
although these parameters are correlated , we have independently varied each parameter to reproduce the slope uncertainty . in order to estimate the uncertainty in @xmath50
, the surface gravity has been adjusted until the mean difference in fei and feii matches the quadrature sum of the variance in fei and feii abundances .
these uncertainties are listed in table [ tab : stellar - parameter - uncertainties ] for all standard and program stars . as these uncertainties are calculated with the assumption that they are uncorrelated ,
they are therefore possibly underestimated . as such
we have assumed a minimum uncertainty of 125k in @xmath43 , 0.3dex in @xmath50 and 0.2 km s@xmath6 in @xmath48 .
thus , the adopted uncertainty for each star is taken as the maximum of these values and those listed in table [ tab : stellar - parameter - uncertainties ] . + lcccccccccccccl[h ! ] hd 41667 & 47 & 0.08 & 0.05 + hd 44007 & 51 & 0.08 & 0.17 + hd 76932 & 61 & 0.10 & 0.14 + hd 122563 & 15 & 0.01 & 0.07 + hd 136316 & 34 & 0.02 & 0.05 + hd 141531 & 40 & 0.03 & 0.05 + hd 142948 & 62 & 0.12 & 0.09 +
oss 3 ( l ) & 73 & 0.08 & 0.11 + oss 6 ( h ) & 39 & 0.04 & 0.15 + oss 8 ( h ) & 74 & 0.17 & 0.23 + oss 14 ( h ) & 51 & 0.05 & 0.13 + oss 18 ( m ) & 135 & 0.15 & 0.28 chemical abundances are described following the standard nomenclature ] , and comparisons made with reference to the sun = \log\epsilon{\left(\frac{n_x}{n_h}\right)}_{\rm star } - \log\epsilon{\left(\frac{n_x}{n_h}\right)}_{\rm sun}$ ] ] have been calculated using the solar composition described in @xcite .
we note that some transitions have been removed from our line list ( table [ tab : atomic - data ] ) when they were not detected in any star ( e.g. , sci , zri , zrii , rbi , and ceii transitions ) .
carbon has been measured from synthesizing the 431.3 nm and 432.3 nm ch features .
the best fitting abundances inferred from these two syntheses were within 0.05dex for every star .
nevertheless , have adopted a conservative total uncertainty of @xmath540.15dex for these measurements .
the mid - point of these two measurements is listed in table [ tab : chemical - abundances ] .
only the 432.3 nm region was synthesized for oss 18 , as significant absorption of atomic lines was present in the bluer band head , and we deemed the 432.3 nm region to yield a more precise determination of carbon abundance .
sodium , aluminium and potassium are primarily produced through carbon , neon and oxygen burning in massive stars , before they are ejected into the interstellar medium @xcite .
although our line list includes three clean , unblended na lines , not all were detectable in the orphan stream candidates .
generally only one na and al line was available for the orphan stream stars .
in these cases , a minimum standard deviation of 0.10dex has been assumed when calculating abundance uncertainties ( see section [ sec : chemical - abundance - uncertainties ] ) . wherever there were no detectable lines for a given element ,
an upper limit was determined by synthesis of the strongest transition in our line list .
at least one k line was detected in every standard and program star .
these lines at 766 nm and 769 nm are quite strong , but fall directly between a strong telluric band head . usually both lines were detected , but one was dominated by the earth s atmospheric absorption . in these cases we rejected the contaminated line and adopted the single , unaffected k line .
the @xmath1-elements ( o , mg , ca , si and ti ) are produced during hydrostatic burning of carbon , neon and silicon by @xmath1-particle capture . following type ii supernovae ,
the @xmath1-enriched material is dispersed into the interstellar medium and contributes to the next generation of star formation .
although ti is not formally an @xmath1-element ( @xmath57 ) , it generally tracks with @xmath1-elements , and has been included here to facilitate a comparison with the local group study of @xcite .
generally , the orphan stream candidates have lower @xmath1-element abundances compared to iron than their halo counterparts .
the situation is a little ambiguous for [ o / fe ] compared to other @xmath1-elements .
we have employed the forbidden lines at 630 nm and 636 nm to measure oxygen abundances in our stars . given
the weakness of these lines , oxygen was immeasurable in all of the orphan stream candidates . in place of abundances ,
upper limits have been determined from spectral synthesis of the region .
the synthesis line list includes the nii feature hidden within the 630 nm absorption profile @xcite . in the low @xmath40 ratio case of oss 18 ,
the forbidden oxygen region was sufficiently dominated by telluric absorption such that we deemed even a robust upper limit to be indeterminable .
the [ mg / fe ] abundance ratios for the orphan stream targets are noticeably lower than any of the other @xmath1-elements .
in fact , oss 18 and 3 exhibit solar or sub - solar [ mg / fe ] abundances , and the mean [ @xmath1/fe ] abundances in the lower panel of figure [ fig : alpha - fe ] are most affected by [ mg / fe ]
. nevertheless , [ ca / fe ] and [ si / fe ] abundances for orphan stream targets are systematically lower than their milky way counterparts . in figure
[ fig : alpha - fe ] , tiii has been adopted for [ ti / fe ] as few reliable tii transitions were available for analysis .
lower [ @xmath1/fe ] abundance ratios are less apparent for our lower probability orphan stream candidates , oss 18 and oss 3 .
although their mean [ @xmath1/fe ] @xmath0 + 0.20dex , with [ fe / h ] @xmath0 0.70dex they are significantly more metal - rich than expected for the orphan stream . therefore their [ @xmath1/fe ] abundance ratios are consistent with the [ @xmath1/fe]-[fe / h ] trend of the milky way .
as discussed further in section [ sec : stream - membership ] , these low probability targets are unlikely to be true orphan stream members , and are marked appropriately in figure [ fig : alpha - fe ] .
-element abundance ratios ( mgi , cai , tiii shown ) for field standards ( grey , filled ) , high - probability orphan stream candidates ( white squares ) , and lower - probability orphan stream candidates ( white squares with crosses ) that are deemed to be interlopers .
data compiled by @xcite for the present - day dsph galaxies ( grey squares ) and the milky way is shown , using the same colour scheme adopted in @xcite : thin disk ( red ) , thick disk ( green ) , halo ( cyan ) , high velocity halo ( dark blue ) , and retrograde halo stars ( black ) . ] the fe - peak elements ( @xmath58 to 30 ) are primarily produced by explosive nucleosynthesis during oxygen , neon , and silicon burning .
the ignition of these burning phases occurs both from type ii sn of massive stars , and once an accreting white dwarf exceeds the chandrasekhar limit , causing spontaneous ignition of carbon and an eventual type ia supernova .
the abundance of the fe - peak elements with respect to iron in the milky way are typically either flat ( i.e. , [ x / fe ] @xmath0 0 ; crii , ni ) or trend positively ( sc , cri , mn , cu ) with overall [ fe / h ] ( e.g. , * ? ? ?
* ; * ? ? ?
scandium absorption profiles have appreciable broadening due to hyperfine splitting .
as such , we have determined sc abundances for all stars from spectral synthesis with hyperfine splitting taken into account . however , for oss3 we found the broadening due to hyperfine structure to be negligible , and the variance in our synthesis measurements was larger than our abundance measurements . the mean scii abundance to be @xmath59 from synthesis of 5 lines , where the ews of 11 scii lines yields @xmath60 . therefore in the case of oss3 , we have adopted the scii abundances from ews .
other fe - peak elements with hyperfine structure ( e.g. v , mn , co , cu ) have been synthesised with the relevant isotopic and/or hyperfine splitting included .
the random measurement scatter in v , cr , and mn abundances is typically low ( @xmath610.05dex ) for all the standard and program stars .
however for co , there was a noticeable increase in line - to - line scatter for the orphan stream stars compared to the standard stars .
there is a factor of @xmath06 difference in @xmath40 ratio between the two samples which can explain this variance . in the orphan stream stars , no synthesised co profiles were sufficiently ` worse ' than each other to qualify exclusion .
the case was quite different for ni , where a plethora of clean lines ( without appreciable hyperfine structure ) are available , and the uncertainties in @xmath62 abundances due to the uncertainties in stellar parameters generally cancel with @xmath63 , yielding excellent measurements of [ ni / fe ] .
cu abundances and upper limits are derived from the synthesis of a single neutral cu line at 510.5 nm .
the inclusion of hyperfine structure was most important for cu , as a cu abundance measured directly from an ew produced a systematically higher @xmath64 , on the order of + 0.4dex .
zn abundances were calculated directly from the ews of the 472.2 nm and 481.0 nm transitions where available .
all of our fe - peak abundances with respect to iron are consistent with the observed chemical trends in the milky way @xcite .
the atomic absorption lines of these heavy elements have appreciable broadening due to hyperfine structure and isotopic splitting . where applicable , the relevant hyperfine and/or isotopic splitting has been employed , and solar isotopic compositions have been adopted . sr and y are elements of the first @xmath65-capture peak . while only one line ( 421.5 nm ) was synthesised for srii ,
generally three unblended y lines were available .
when no y lines were detected above @xmath66 , an upper limit was ascertained from the 520 nm line the strongest in our line list .
sr and y abundances generally agree with each other in our stars , with the exception of high probability stream candidates oss 6 and oss 8 . in these cases
only upper limits could be determined for yii , and those limits are @xmath00.6dex lower than our measured srii abundances .
the remaining high probability candidate , oss 14 , also has an upper limit for yii ( [ y / fe ] @xmath67 ) but this limit is much closer to our srii measurement ( [ sr / fe ] = @xmath68 ) . the y upper limits for our high probability orphan stream stars results in lower limits for [ ba / y ] , which for oss 6 and oss 8 , are well in excess of the milky way trend ( figure [ fig : ba - y ] ) .
lanthanum and neodymium abundances for our stars are consistent with the chemical evolution of the milky way .
europium , a @xmath69-process dominated element , has been measured by synthesising the 664.5 nm euii transition .
although the spread in [ eu / fe ] abundances for our stars is wide , no significant @xmath69-process enhancement is observed .
although the standard error about the mean ( @xmath70 ) abundance in table [ tab : chemical - abundances ] can be quite low , the uncertainties in stellar parameters will significantly contribute to the total error budget for any given abundance .
furthermore , the total uncertainty in abundance ratios ( e.g. , [ a / b ] ) depends on how the uncertainties in elements a and b are correlated with stellar parameters .
+ lcrcrrcclcrcrrc + species & n & @xmath71 & @xmath51 & [ x / h ] & [ x / fe ] & @xmath70 & & species & n & @xmath71 & @xmath51 & [ x / h ] & [ x / fe ] & @xmath70 + species & n & @xmath71 & @xmath51 & [ x / h ] & [ x / fe ] & @xmath70 & & species & n & @xmath71 & @xmath51 & [ x / h ] & [ x / fe ] & @xmath70 + & & + c ( ch ) & 2 & 6.95 & 0.20 & 1.48 & 0.24 & 0.15 & & c ( ch ) & 2 & 6.73 & 0.20 & 1.70 & 0.01 & 0.15 + o i & 1 & 7.85 & & 0.84 & 0.34 & & & o i & 1 & 7.43 & & 1.26 & 0.44 & + na i & 3 & 4.95 & 0.10 & 1.29 & 0.10 & 0.06 & & na i & 1 & 4.57 & & 1.67 & 0.03 & + mg i & 7 & 6.71 & 0.13 & 0.89 & 0.29 & 0.05 & & mg i & 5 & 6.29 & 0.06 & 1.31 & 0.40 & 0.03 + al i & 4 & 5.27 & 0.11 & 1.18 & 0.00 & 0.05 & & al i & 0 & & & & & + si i & 8 & 6.58 & 0.14 & 0.93 & 0.26 & 0.05 & & si i & 7 & 6.18 & 0.09 & 1.33 & 0.37 & 0.03 + k i & 2 & 4.66 & 0.00 & 0.38 & 0.81 & 0.00 & & k i & 2 & 4.43 & 0.05 & 0.60 & 1.10 & 0.04 + ca i & 18 & 5.52 & 0.07 & 0.82 & 0.37 & 0.02 & & ca i & 20 & 5.10 & 0.09 & 1.24 & 0.46 & 0.02 + sc ii & 14 & 2.10 & 0.10 & 1.05 & 0.13 & 0.03 & & sc ii & 12 & 1.45 & 0.08 & 1.70 & 0.00 & 0.02 + ti i & 25 & 3.90 & 0.16 & 1.05 & 0.13 & 0.03 & & ti i & 20 & 3.35 & 0.08 & 1.60 & 0.11 & 0.02 + ti ii & 28 & 4.16 & 0.18 & 0.79 & 0.40 & 0.03 & & ti ii & 36 & 3.57 & 0.13 & 1.38 & 0.32 & 0.02 + v i & 4 & 2.82 & 0.09 & 1.11 & 0.07 & 0.05 & & v i & 4 & 2.24 & 0.05 & 1.69 & 0.01 & 0.02 + cr i & 12 & 4.28 & 0.07 & 1.36 & 0.18 & 0.02 & & cr i & 15 & 3.75 & 0.06 & 1.89 & 0.19 & 0.02 + cr ii & 3 & 4.65 & 0.13 & 0.99 & 0.20 & 0.08 & & cr ii & 3 & 4.10 & 0.04 & 1.54 & 0.16 & 0.03 + mn i & 5 & 4.12 & 0.17 & 1.31 & 0.13 & 0.08 & & mn i & 7 & 3.30 & 0.10 & 2.13 & 0.43 & 0.04 + fe i & 72 & 6.32 & 0.11 & 1.18 & 0.00 & 0.01 & & fe i & 74 & 5.80 & 0.12 & 1.70 & 0.00 & 0.01 + fe ii & 17 & 6.32 & 0.08 & 1.18 & 0.00 & 0.02 & & fe ii & 17 & 5.80 & 0.12 & 1.70 & 0.01 & 0.03 + co i & 4 & 3.85 & 0.05 & 1.14 & 0.04 & 0.03 & & co i & 3 & 3.29 & 0.09 & 1.70 & 0.00 & 0.05 + ni i & 20 & 4.96 & 0.13 & 1.26 & 0.08 & 0.03 & & ni i & 23 & 4.46 & 0.12 & 1.76 & 0.05 & 0.03 + cu i & 1 & 2.86 & & 1.33 & 0.15 & & & cu i & 1 & 1.90 & & 2.29 & 0.59 & + zn i & 2 & 3.39 & 0.04 & 1.17 & 0.01 & 0.03 & & zn i & 2 & 2.88 & 0.08 & 1.68 & 0.02 & 0.06 + sr ii & 1 & 1.75 & & 1.12 & 0.06 & & & sr ii & 1 & 1.19 & & 1.68 & 0.02 & + y ii & 3 & 1.08 & 0.12 & 1.13 & 0.05 & 0.07 & & y ii & 3 & 0.35 & 0.06 & 1.86 & 0.15 & 0.04 + ba ii & 2 & 0.97 & 0.03 & 1.21 & 0.02 & 0.02 & & ba ii & 2 & 0.43 & 0.08 & 1.75 & 0.04 & 0.05 + la ii & 2 & 0.14 & 0.03 & 0.97 & 0.22 & 0.02 & & la ii & 2 & 0.42 & 0.00 & 1.52 & 0.18 & 0.00 + nd ii & 6 & 0.56 & 0.06 & 0.86 & 0.32 & 0.02 & & nd ii & 7 & 0.26 & 0.09 & 1.68 & 0.03 & 0.03 + eu ii & 1 & 0.16 & & 0.68 & 0.50 & & & eu ii & 1 & 1.16 & & 1.68 & 0.02 & + + + & & + c ( ch ) & 2 & 7.63 & 0.20 & 0.79 & 0.15 & 0.15 & & c ( ch ) & 2 & 5.26 & 0.20 & 3.17 & 0.27 & 0.15 + o i & 1 & @xmath618.33 & & @xmath610.36 & @xmath610.58 & & & o i & 1 & 6.15 & & 2.54 & 0.37 & + na i & 3 & 5.44 & 0.03 & 0.80 & 0.15 & 0.02 & & na i & 1 & @xmath613.48 & & @xmath612.76 & @xmath610.14 & + mg i & 8 & 6.94 & 0.18 & 0.66 & 0.30 & 0.06 & & mg i & 8 & 5.25 & 0.08 & 2.35 & 0.56 & 0.03 + al i & 4 & 5.56 & 0.08 & 0.89 & 0.07 & 0.04 & & al i & 1 & @xmath614.83 & & @xmath611.62 & @xmath611.28 & + si i & 9 & 6.80 & 0.18 & 0.71 & 0.24 & 0.06 & & si i & 1 & 5.21 & & 2.30 & 0.61 & + k i & 2 & 4.96 & 0.06 & 0.07 & 0.89 & 0.05 & & k i & 1 & 2.75 & & 2.28 & 0.63 & + ca i & 23 & 5.69 & 0.09 & 0.65 & 0.31 & 0.02 & & ca i & 19 & 3.81 & 0.05 & 2.53 & 0.39 & 0.01 + sc ii & 16 & 2.38 & 0.10 & 0.77 & 0.18 & 0.02 & & sc ii & 15 & 0.17 & 0.07 & 2.98 & 0.07 & 0.02 + ti i & 17 & 4.16 & 0.06 & 0.79 & 0.16 & 0.01 & & ti i & 17 & 2.06 & 0.04 & 2.89 & 0.02 & 0.01 + ti ii & 38 & 4.35 & 0.12 & 0.60 & 0.36 & 0.02 & & ti ii & 42 & 2.22 & 0.10 & 2.73 & 0.18 & 0.02 + v i & 4 & 3.28 & 0.12 & 0.65 & 0.31 & 0.06 & & v i & 1 & 0.81 & & 3.12 & 0.21 & + cr i & 18 & 4.58 & 0.07 & 1.06 & 0.11 & 0.02 & & cr i & 11 & 2.39 & 0.08 & 3.25 & 0.34 & 0.02 + cr ii & 3 & 4.86 & 0.04 & 0.78 & 0.18 & 0.02 & & cr ii & 3 & 2.83 & 0.03 & 2.81 & 0.10 & 0.02 + mn i & 9 & 4.31 & 0.10 & 1.12 & 0.17 & 0.03 & & mn i & 7 & 2.12 & 0.06 & 3.31 & 0.39 & 0.02 + fe i & 96 & 6.55 & 0.11 & 0.95 & 0.00 & 0.01 & & fe i & 165 & 4.59 & 0.11 & 2.91 & 0.00 & 0.01 + fe ii & 20 & 6.55 & 0.13 & 0.95 & 0.00 & 0.03 & & fe ii & 22 & 4.60 & 0.11 & 2.90 & 0.01 & 0.02 + co i & 4 & 4.09 & 0.12 & 0.90 & 0.05 & 0.06 & & co i & 5 & 2.23 & 0.14 & 2.76 & 0.15 & 0.06 + ni i & 25 & 5.32 & 0.11 & 0.90 & 0.05 & 0.02 & & ni i & 19 & 3.48 & 0.10 & 2.74 & 0.17 & 0.02 + cu i & 1 & 2.95 & & 1.24 & 0.29 & & & cu i & 1 & @xmath610.10 & & @xmath614.19&@xmath611.19 & + zn i & 2 & 3.71 & 0.02 & 0.85 & 0.10 & 0.01 & & zn i & 2 & 1.83 & 0.06 & 2.72 & 0.19 & 0.05 + sr ii & 1 & 2.00 & & 0.87 & 0.08 & & & sr ii & 1 & 0.64 & & 3.51 & 0.60 & + y ii & 2 & 1.29 & 0.04 & 0.92 & 0.03 & 0.02 & & y ii & 1 & 0.82 & & 3.03 & 0.12 & + ba ii & 2 & 1.34 & 0.04 & 0.84 & 0.12 & 0.03 & & ba ii & 2 & 1.94 & 0.06 & 4.12 & 1.20 & 0.04 + la ii & 1 & 0.69 & & 0.41 & 0.54 & & & la ii & 1 & @xmath611.51 & & @xmath612.61 & @xmath610.29 & + nd ii & 3 & 0.81 & 0.11 & 0.61 & 0.35 & 0.07 & & nd ii & 1 & @xmath611.83 & & @xmath613.25&@xmath610.35 & + eu ii & 1 & @xmath610.27 & & @xmath610.79 & @xmath610.15 & & & eu ii & 1 & @xmath611.72 & & @xmath612.24 & @xmath610.66 & + + + & & + c ( ch ) & 2 & 6.05 & 0.20 & 2.38 & 0.47 & 0.15 & & c ( ch ) & 2 & 6.31 & 0.20 & 2.12 & 0.42 & 0.15 + o i & 1 & 7.11 & & 1.58 & 0.35 & & & o i & 2 & 7.36 & 0.01 & 1.33 & 0.31 & 0.00 + na i & 2 & 4.21 & 0.08 & 2.03 & 0.10 & 0.06 & & na i & 3 & 4.40 & 0.07 & 1.84 & 0.19 & 0.04 + mg i & 8 & 6.09 & 0.29 & 1.51 & 0.42 & 0.10 & & mg i & 9 & 6.36 & 0.27 & 1.24 & 0.40 & 0.09 + al i & 1 & @xmath614.58 & & @xmath611.87 & @xmath610.04 & & & al i & 1 & 4.82 & & 1.63 & 0.01 & + si i & 8 & 5.93 & 0.17 & 1.58 & 0.35 & 0.06 & & si i & 7 & 6.08 & 0.15 & 1.43 & 0.21 & 0.06 + k i & 2 & 3.83 & 0.03 & 1.21 & 0.72 & 0.02 & & k i & 2 & 4.13 & 0.06 & 0.90 & 0.74 & 0.04 + ca i & 19 & 4.77 & 0.08 & 1.57 & 0.36 & 0.02 & & ca i & 17 & 5.02 & 0.06 & 1.32 & 0.32 & 0.01 + sc ii & 16 & 1.24 & 0.08 & 1.91 & 0.01 & 0.02 & & sc ii & 12 & 1.56 & 0.05 & 1.59 & 0.05 & 0.01 + ti i & 26 & 3.02 & 0.13 & 1.93 & 0.01 & 0.03 & & ti i & 26 & 3.34 & 0.13 & 1.61 & 0.03 & 0.03 + ti ii & 39 & 3.46 & 0.18 & 1.49 & 0.44 & 0.03 & & ti ii & 32 & 3.70 & 0.17 & 1.25 & 0.40 & 0.03 + v i & 6 & 1.88 & 0.08 & 2.05 & 0.12 & 0.03 & & v i & 6 & 2.24 & 0.10 & 1.69 & 0.05 & 0.04 + cr i & 18 & 3.52 & 0.13 & 2.12 & 0.20 & 0.03 & & cr i & 15 & 3.78 & 0.07 & 1.86 & 0.22 & 0.02 + cr ii & 2 & 3.83 & 0.03 & 1.81 & 0.11 & 0.02 & & cr ii & 1 & 4.14 & & 1.50 & 0.14 & + mn
i & 9 & 3.14 & 0.14 & 2.29 & 0.36 & 0.05 & & mn i & 7 & 3.44 & 0.15 & 1.99 & 0.35 & 0.06 + fe i & 100 & 5.57 & 0.12 & 1.93 & 0.00 & 0.01 & & fe i & 83 & 5.85 & 0.12 & 1.65 & 0.00 & 0.01 + fe ii & 16 & 5.57 & 0.08 & 1.93 & 0.00 & 0.02 & & fe ii & 17 & 5.85 & 0.08 & 1.65 & 0.00 & 0.02 + co i & 4 & 3.08 & 0.12 & 1.91 & 0.01 & 0.06 & & co i & 6 & 3.33 & 0.13 & 1.66 & 0.01 & 0.05 + ni i & 24 & 4.24 & 0.12 & 1.98 & 0.05 & 0.03 & & ni i & 25 & 4.50 & 0.15 & 1.72 & 0.07 & 0.03 + cu i & 1 & 1.72 & & 2.47 & 0.54 & & & cu i & 1 & 2.15 & & 2.04 & 0.40 & + zn i & 2 & 2.65 & 0.10 & 1.91 & 0.02 & 0.07 & & zn i & 2 & 2.79 & 0.04 & 1.77 & 0.13 & 0.03 + sr ii & 1 & 0.90 & & 1.97 & 0.04 & & & sr ii & 1 & 1.20 & & 1.67 & 0.03 & + y ii & 3 & 0.12 & 0.10 & 2.09 & 0.16 & 0.06 & & y ii & 3 & 0.35 & 0.10 & 1.86 & 0.21 & 0.06 + ba ii & 2 & 0.19 & 0.08 & 1.99 & 0.06 & 0.06 & & ba ii & 2 & 0.47 & 0.05 & 1.71 & 0.07 & 0.04 + la ii & 2 & 0.73 & 0.03 & 1.83 & 0.09 & 0.02 & & la ii & 2 & 0.54 & 0.07 & 1.65 & 0.00 & 0.05 + nd ii & 8 & 0.40 & 0.04 & 1.82 & 0.10 & 0.02 & & nd ii & 7 & 0.18 & 0.07 & 1.60 & 0.04 & 0.03 + eu ii & 1 & 1.12 & & 1.64 & 0.29 & & & eu ii & 1 & 0.93 & & 1.45 & 0.19 & + + + & & + c ( ch ) & 2 & 7.75 & 0.15 & 0.68 & 0.03 & 0.15 & & c ( ch ) & 2 & 7.54 & 0.15 & 0.89 & 0.00 & 0.15 + o i & 2 & 8.57 & 0.02 & 0.12 & 0.52 & 0.01 & & o i & 1 & @xmath618.41 & & @xmath610.27 & @xmath610.52 & + na i & 2 & 5.69 & 0.02 & 0.55 & 0.09 & 0.01 & & na i & 2 & 5.22 & 0.01 & 1.02 & 0.16 & 0.01 + mg i & 6 & 7.34 & 0.10 & 0.26 & 0.38 & 0.04 & & mg i & 6 & 6.70 & 0.14 & 0.90 & 0.04 & 0.06 + al i & 4 & 6.03 & 0.09 & 0.42 & 0.22 & 0.04 & & al i & 2 & 5.27 & 0.11 & 1.18 & 0.32 & 0.08 + si i & 8 & 7.10 & 0.15 & 0.41 & 0.23 & 0.05 & & si i & 8 & 6.72 & 0.16 & 0.79 & 0.07 & 0.06 + k i & 1 & 5.17 & 0.00 & 0.14 & 0.78 & 0.00 & & k i & 1 & 4.85 & & 0.18 & 0.68 & + ca i & 18 & 5.93 & 0.11 & 0.41 & 0.23 & 0.03 & & ca i & 21 & 5.71 & 0.13 & 0.63 & 0.23 & 0.03 + sc ii & 13 & 2.81 & 0.09 & 0.34 & 0.30 & 0.03 & & sc ii & 11 & 2.24 & 0.09 & 0.91 & 0.05 & 0.03 + ti i & 22 & 4.44 & 0.10 & 0.51 & 0.14 & 0.02 & & ti i & 19 & 4.17 & 0.15 & 0.78 & 0.08 & 0.03 + ti ii & 31 & 4.77 & 0.20 & 0.18 & 0.47 & 0.04 & & ti ii & 37 & 4.42 & 0.24 & 0.53 & 0.33 & 0.04 + v i & 4 & 3.44 & 0.09 & 0.49 & 0.15 & 0.05 & & v i & 2 & 2.97 & 0.05 & 0.88 & 0.02 & 0.04 + cr i & 11 & 4.82 & 0.05 & 0.82 & 0.17 & 0.01 & & cr i & 17 & 4.73 & 0.20 & 0.91 & 0.05 & 0.05 + cr ii & 2 & 5.11 & 0.01 & 0.53 & 0.11 & 0.00 & & cr ii & 3 & 4.89 & 0.26 & 0.75 & 0.11 & 0.15 + mn i & 8 & 4.74 & 0.18 & 0.69 & 0.05 & 0.06 & & mn i & 3 & 4.05 & 0.03 & 1.38 & 0.42 & 0.02 + fe i & 68 & 6.86 & 0.13 & 0.64 & 0.00 & 0.02 & & fe i & 97 & 6.64 & 0.20 & 0.86 & 0.00 & 0.02 + fe ii & 16 & 6.87 & 0.14 & 0.63 & 0.01 & 0.03 & & fe ii & 27 & 6.64 & 0.24 & 0.86 & 0.00 & 0.05 + co i & 5 & 4.49 & 0.08 & 0.50 & 0.14 & 0.04 & & co i & 4 & 4.27 & 0.26 & 0.72 & 0.14 & 0.13 + ni i & 21 & 5.62 & 0.16 & 0.60 & 0.04 & 0.03 & & ni i & 25 & 5.33 & 0.24 & 0.89 & 0.03 & 0.05 + cu i & 1 & 3.90 & & 0.29 & 0.35 & & & cu i & 1 & 2.53 & & 1.66 & 0.80 & + zn i & 2 & 4.13 & 0.00 & 0.42 & 0.22 & 0.00 & & zn i & 2 & 3.67 & 0.15 & 0.89 & 0.03 & 0.10 + sr ii & 1 & 2.40 & & 0.47 & 0.17 & & & sr ii & 1 & 1.95 & & 0.92 & 0.06 & + y ii & 3 & 1.74 & 0.24 & 0.47 & 0.17 & 0.14 & & y ii & 2 & 1.28 & 0.35 & 0.93 & 0.07 & 0.25 + ba ii & 2 & 1.44 & 0.02 & 0.74 & 0.09 & 0.01 & & ba ii & 2 & 1.52 & 0.13 & 0.66 & 0.20 & 0.09 + la ii & 2 & 0.61 & 0.10 & 0.49 & 0.15 & 0.07 & & la ii & 1 & @xmath610.57 & & @xmath611.53 & @xmath610.36 & + nd ii & 4 & 1.02 & 0.07 & 0.40 & 0.24 & 0.04 & & nd ii & 1 & 1.46 & & 0.04 & 0.90 & + eu ii & 1 & 0.25 & & 0.27 & 0.37 & & & eu ii & 1 & @xmath610.09 & & @xmath610.61 & @xmath610.28 & + + + & & + c ( ch ) & 2 & 6.39 & 0.20 & 2.04 & 0.37 & 0.15 & & c ( ch ) & 2 & 6.85 & 0.20 & 1.58 & 0.00 & 0.15 + o i & 1 & @xmath617.08 & & @xmath611.61 & @xmath610.06 & & & o i & 1 & @xmath617.46 & & @xmath611.23 & @xmath610.35 & + na i & 1 & 4.26 & & 1.98 & 0.23 & & & na i & 1 & 4.36 & & 1.88 & 0.26 & + mg i & 6 & 5.94 & 0.21 & 1.66 & 0.09 & 0.09 & & mg i & 6 & 6.10 & 0.15 & 1.49 & 0.13 & 0.06 + al i & 1 & 4.84 & & 1.61 & 0.14 & & & al i & 1 & @xmath615.21 & & @xmath611.24 & @xmath610.34 & + si i & 3 & 5.84 & 0.07 & 1.67 & 0.09 & 0.04 & & si i & 2 & 6.01 & 0.13 & 1.50 & 0.12 & 0.09 + k i & 1 & 3.86 & & 1.17 & 0.58 & & & k i & 1 & 4.24 & & 0.79 & 0.83 & + ca i & 21 & 4.78 & 0.15 & 1.56 & 0.19 & 0.03 & & ca i & 22 & 5.03 & 0.14 & 1.31 & 0.31 & 0.03 + sc ii & 5 & 1.10 & 0.10 & 2.05 & 0.31 & 0.05 & & sc ii & 16 & 1.47 & 0.15 & 1.68 & 0.06 & 0.04 + ti i & 23 & 3.09 & 0.18 & 1.86 & 0.11 & 0.04 & & ti i & 20 & 3.31 & 0.11 & 1.64 & 0.02 & 0.02 + ti ii & 35 & 3.33 & 0.19 & 1.62 & 0.13 & 0.03 & & ti ii & 37 & 3.59 & 0.21 & 1.36 & 0.26 & 0.03 + v i & 2 & 1.74 & 0.12 & 2.19 & 0.44 & 0.08 & & v i & 2 & 2.29 & 0.18 & 1.64 & 0.02 & 0.13 + cr i & 20 & 3.71 & 0.13 & 1.93 & 0.18 & 0.03 & & cr i & 17 & 3.82 & 0.12 & 1.82 & 0.19 & 0.03 + cr ii & 2 & 4.10 & 0.06 & 1.54 & 0.21 & 0.04 & & cr ii & 3 & 4.23 & 0.04 & 1.41 & 0.21 & 0.02 + mn i & 2 & 3.23 & 0.01 & 2.20 & 0.45 & 0.01 & & mn i & 5 & 3.53 & 0.18 & 1.90 & 0.28 & 0.08 + fe i & 66 & 5.75 & 0.11 & 1.75 & 0.00 & 0.01 & & fe i & 60 & 5.88 & 0.15 & 1.62 & 0.00 & 0.02 + fe ii & 15 & 5.74 & 0.13 & 1.76 & 0.00 & 0.03 & & fe ii & 12 & 5.88 & 0.16 & 1.62 & 0.00 & 0.05 + co i & 3 & 3.00 & 0.36 & 1.99 & 0.23 & 0.21 & & co i & 4 & 3.35 & 0.14 & 1.64 & 0.02 & 0.07 + ni i & 22 & 4.39 & 0.13 & 1.83 & 0.08 & 0.03 & & ni i & 20 & 4.52 & 0.15 & 1.70 & 0.08 & 0.03 + cu i & 1 & 1.02 & & 3.17 & 1.42 & & & cu i & 1 & 1.68 & & 2.51 & 0.89 & + zn i & 2 & 2.46 & 0.14 & 2.09 & 0.34 & 0.10 & & zn i & 1 & 2.89 & & 1.67 & 0.05 & + sr ii & 1 & 0.73 & & 2.14 & 0.39 & & & sr ii & 1 & 0.79 & & 2.08 & 0.46 & + y ii & 1 & @xmath610.59 & & @xmath612.77&@xmath611.10 & & & y ii & 1 & @xmath610.41 & & @xmath612.60&@xmath611.00 & + ba ii & 1 & 0.08 & & 2.26 & 0.50 & & & ba ii & 2 & 0.04 & 0.20 & 2.14 & 0.51 & 0.14 + la ii & 1 & @xmath611.00 & & @xmath611.82&@xmath610.35 & & & la ii & 1 & 0.12 & & 1.22 & 0.40 & + nd ii & 2 & 0.71 & 0.08 & 2.13 & 0.38 & 0.06 & & nd ii & 1 & 0.08 & & 1.34 & 0.28 & + eu ii & 1 & @xmath611.45 & & @xmath611.97&@xmath610.30 & & & eu ii & 1 & @xmath610.62 & & @xmath611.14 & @xmath610.44 & + + + & & + c ( ch ) & 2 & 5.85 & 0.20 & 2.58 & 0.08 & 0.15 & & c ( ch ) & 1 & 7.57 & 0.20 & 0.86 & 0.10 & 0.15 + o i & 1 & @xmath616.40 & & @xmath612.09 & @xmath610.57 & & & o i & 0 & & & & & + na i & 1 & @xmath613.87 & & @xmath612.37 & @xmath610.29 & & & na i & 1 & 5.37 & & 0.87 & 0.25 & + mg i & 8 & 5.19 & 0.21 & 2.41 & 0.25 & 0.07 & & mg i & 5 & 7.01 & 0.25 & 0.59 & 0.03 & 0.11 + al i & 0 & & & & & & & al i & 1 & @xmath615.66 & & @xmath610.79&@xmath610.03 & + si i & 2 & 5.50 & 0.04 & 2.00 & 0.66 & 0.02 & & si i & 5 & 6.98 & 0.28 & 0.53 & 0.09 & 0.13 + k i & 1 & 2.66 & & 2.37 & 0.30 & & & k i & 1 & 4.91 & & 0.12 & 0.50 & + ca i & 16 & 3.89 & 0.15 & 2.45 & 0.21 & 0.04 & & ca i & 19 & 5.82 & 0.14 & 0.52 & 0.10 & 0.03 + sc ii & 9 & 0.13 & 0.13 & 3.02 & 0.35 & 0.04 & & sc ii & 11 & 2.64 & 0.18 & 0.51 & 0.11 & 0.05 + ti i & 11 & 2.24 & 0.12 & 2.71 & 0.05 & 0.04 & & ti i & 17 & 4.27 & 0.24 & 0.68 & 0.06 & 0.06 + ti ii & 32 & 2.37 & 0.18 & 2.58 & 0.09 & 0.03 & & ti ii & 25 & 4.44 & 0.33 & 0.51 & 0.11 & 0.07 + v i & 1 & @xmath611.61 & & @xmath612.32 & @xmath610.34 & & & v i & 4 & 3.26 & 0.01 & 0.67 & 0.05 & 0.00 + cr i & 10 & 2.54 & 0.23 & 3.10 & 0.43 & 0.07 & & cr i & 18 & 4.83 & 0.28 & 0.81 & 0.19 & 0.06 + cr ii & 1 & 3.06 & & 2.58 & 0.09 & & & cr ii & 1 & 5.29 & & 0.35 & 0.27 & + mn i & 4 & 2.20 & 0.20 & 3.23 & 0.56 & 0.10 & & mn i & 3 & 4.30 & 0.09 & 1.13 & 0.51 & 0.05 + fe i & 133 & 4.83 & 0.20 & 2.67 & 0.00 & 0.02 & & fe i & 48 & 6.88 & 0.21 & 0.62 & 0.00 & 0.03 + fe ii & 16 & 4.83 & 0.16 & 2.67 & 0.01 & 0.04 & & fe ii & 12 & 6.88 & 0.25 & 0.62 & 0.00 & 0.07 + co i & 3 & 2.32 & 0.48 & 2.67 & 0.01 & 0.28 & & co i & 3 & 4.65 & 0.41 & 0.34 & 0.28 & 0.24 + ni i & 5 & 3.51 & 0.07 & 2.71 & 0.04 & 0.03 & & ni i & 17 & 5.43 & 0.24 & 0.79 & 0.17 & 0.06 + cu i & 1 & @xmath610.77 & & @xmath613.42&@xmath610.75 & & & cu i & 1 & 3.14 & & 1.05 & 0.43 & + zn i & 1 & @xmath611.92 & & @xmath612.64 & @xmath610.02 & & & zn i & 1 & 3.87 & & 0.69 & 0.07 & + sr ii & 1 & 0.22 & & 3.09 & 0.42 & & & sr ii & 1 & 2.31 & & 0.56 & 0.06 & + y ii & 1 & @xmath610.86 & & @xmath613.10&@xmath610.40 & & & y ii & 1 & 1.92 & & 0.29 & 0.33 & + ba ii & 2 & 1.85 & 0.05 & 4.04 & 1.37 & 0.04 & & ba ii & 2 & 1.76 & 0.05 & 0.42 & 0.20 & 0.04 + la ii & 1 & @xmath610.91 & & @xmath612.01 & @xmath610.65 & & & la ii & 1 & 1.29 & & 0.19 & 0.81 & + nd ii & 1 & @xmath610.56 & & @xmath611.98 & @xmath610.68 & & & nd ii & 2 & 1.71 & 0.02 & 0.29 & 0.91 & 0.01 + eu ii & 1 & @xmath611.14 & & @xmath611.66 & @xmath611.00 & & & eu ii & 1 & 0.72 & & 0.20 & 0.82 & + lcccccc + + na i & + 0.06 & + 0.00 & 0.01 & 0.07 & 0.09 & 0.08 + mg i & + 0.07 & 0.02 & 0.02 & 0.06 & 0.09 & 0.18 + al i & + 0.04 & + 0.00 & + 0.00 & 0.08 & 0.09 & 0.10
+ si i & + 0.03 & + 0.01 & 0.01 & 0.06 & 0.07 & 0.08 + k i & + 0.07 & 0.03 & 0.04 & 0.10 & 0.13 & 0.23 + ca i & + 0.07 & 0.02 & 0.03 & 0.03 & 0.08 & 0.17 + sc ii & + 0.00 & + 0.03 & 0.03 & 0.03 & 0.05 & 0.19 + ti i & + 0.12 & 0.01 & 0.05 & 0.03 & 0.14 & 0.14 + ti ii & + 0.01 & + 0.02 & 0.06 & 0.04 & 0.07 & 0.15 + v i & + 0.12 & + 0.00 & 0.03 & 0.06 & 0.13 & 0.06 + cr i & + 0.11 & 0.01 & 0.06 & 0.05 & 0.13 & 0.13 + cr ii & 0.02 & + 0.04 & 0.04 & 0.15 & 0.16 & 0.15 + mn i & + 0.10 & 0.02 & 0.03 & 0.11 & 0.16 & 0.21 + fe i & + 0.08 & + 0.00 & 0.04 & 0.02 & 0.09 & + fe ii & 0.03 & + 0.04 & 0.03 & 0.05 & 0.07 & + co i & + 0.11 & 0.01 & 0.07 & 0.13 & 0.19 & 0.19 + ni i & + 0.07 & + 0.00 & 0.04 & 0.05 & 0.09 & 0.04 + cu
i & + 0.10 & + 0.01 & 0.05 & 0.10 & 0.15 & 0.06 + zn i & + 0.00 & + 0.02 & 0.04 & 0.10 & 0.11 & 0.11 + sr ii & + 0.03 & + 0.00 & 0.01 & 0.10 & 0.10 & 0.11 + y ii & + 0.01 & + 0.03 & 0.02 & 0.25 & 0.25 & 0.16 + ba ii & + 0.02 & + 0.01 & 0.06 & 0.09 & 0.11 & 0.05 + nd ii & + 0.04 & + 0.04 & 0.04 & 0.10 & 0.12 & 0.21 in order to investigate the abundance uncertainties due to stellar parameters , we have generated model atmospheres for a @xmath72 offset in each stellar parameter ( section [ sec : stellar - parameter - uncertainties ] ; @xmath43 , @xmath50 , @xmath48 ) independently , and calculated the resultant mean abundance offset from our ews .
the resultant abundance changes were added in quadrature with @xmath70 to provide a total uncertainty in [ x / h ] .
these total uncertainties are tabulated in table [ tab : chemical - abundance - uncertainties ] for all standard and program stars . while this yields an uncertainty for our abundance ratios in [ x / h ] , we are often interested in how elemental abundances vary with respect to iron .
we have calculated these uncertainties following @xcite : @xmath73 where the covariance between elements a and b ( @xmath74 ) is given by @xmath75\sigma_{t_{\rm eff},\log{g } } \\
+ \left[\left(\frac{\partial\log{\epsilon_{\rm a}}}{\partial\xi}\right)\left(\frac{\partial\log\epsilon_{\rm b}}{\partial\log{g}}\right ) + \left(\frac{\partial\log{\epsilon_{\rm a}}}{\partial\log{g}}\right)\left(\frac{\partial\log\epsilon_{\rm b}}{\partial\xi}\right)\right]\sigma_{\xi,\log{g}}\end{gathered}\ ] ] the covariance between effective temperature and surface gravity has been calculated by sampling about the effective temperature and performing an ionization balance for the adjusted temperature .
the resultant covariance is given as : @xmath76 covariance between @xmath48 and @xmath50 is calculated in the same manner .
the total abundance uncertainty with respect to h and fe ( e.g. , @xmath77[x / h ] ) , @xmath51([x / fe ] ) ) for all species is given in table [ tab : chemical - abundance - uncertainties ] .
these total uncertainties have been adopted in all figures throughout this text . due to the cancellation of systematic effects ,
some abundance ratios with respect to iron have lower uncertainties than their quoted absolute abundance uncertainties .
given the low surface brightness of the orphan stream , separating true members from interlopers can be particularly challenging . before inferring any properties of the undiscovered parent satellite from our sample
, we must examine whether our targets are stars truly from the disruption of the orphan stream progenitor .
when compared against the mean of our high priority targets , the velocities of the low and medium probability candidates are + 6.2 and @xmath78 km s@xmath6 different , respectively . given this region of the stream has a low intrinsic velocity dispersion @xcite , the significantly lower velocity of the medium - probability target , oss 18 , is intriguing .
perhaps more concerning is that the low and medium probability candidates are markedly more metal - rich than the high - priority targets : [ fe / h ] = @xmath79 and @xmath80dex for oss 3 and oss 18 , a difference of @xmath81 and @xmath360.28dex from the low - resolution measurements , respectively .
spectroscopic studies suggest the stream is significantly more metal - poor than [ fe / h ] @xmath82dex @xcite , making the association between the orphan stream and oss 3 or oss 18 tenuous .
it is also worth noting that these targets are the farthest candidates ( @xmath83@xmath4 ) from the best - fit orphan stream orbital plane deduced by @xcite . on the basis of the observables we deduce that the lower probability members , oss 3 and oss 18 , are unlikely to be disrupted orphan stream members . the three high - probability targets ( oss 6 , 8 and 14 ) have velocities within 2.4 km s@xmath6 of each other , consistent with the orphan stream velocity . these velocities from high - resolution spectra
confirm the very low line - of - sight velocity dispersion in this part of the stream .
the metallicities of these targets are also in reasonable agreement with those found from low - resolution spectroscopy , implying the candidates are @xmath020kpc away at approximately the same distance as the orphan stream . for the remainder of this discussion , we consider oss 6 , 8 and 14 , to be true disrupted members of the orphan stream .
although the sample size is small , the three orphan stream members have a wide range in metallicity , ranging from [ fe / h ] = @xmath84 to as metal - poor [ fe / h ] = @xmath85 . from three members not much can be said about the metallicity distribution , other than to say it is wide , and inconsistent with a mono - metallic population ( e.g. , a globular cluster ) .
@xcite and @xcite found the stream to have a mean metallicity of [ fe / h ] @xmath86 from bhb and rr lyrae stars . according to the metallicity gradient along the stream reported by @xcite , we might expect a higher mean metallicity than [ fe / h ] @xmath87 in our sample closer to the celestial equator .
we note that our metallicity spread of [ fe / h ] = @xmath84 to @xmath85 is consistent with that of @xcite : [ fe / h ] = @xmath16 to @xmath17 , and the somewhat wider spread found by @xcite : [ fe / h ] = @xmath88 to @xmath89 .
a total metallicity spread as wide as @xmath01dex , or a standard deviation of @xmath22 } ) = 0.56 $ ] @xcite , is consistent with the stochastic chemical enrichment of the present - day dsph galaxies @xcite .
this is also compatible with the constraints from the dynamical constraints placed inferred by the arc and width of the stream , which suggest the progenitor is a dark matter - dominated system .
the abundance trends of individual elements track the star formation history of a cluster or galaxy . in the milky way ,
this is most evident by the evolution of @xmath1-element abundances with increasing metallicity .
the @xmath1-elements are produced in massive stars before being ejected to the interstellar medium by type ii supernovae . at later times ,
type ia supernovae begin to contribute to the galaxy s chemical enrichment .
type ia supernovae expel iron ( like type ii ) , but do not produce significant amounts of @xmath1-elements . as such , a net decrease in [ @xmath1/fe ] is observed upon the onset of type ia supernovae .
this inflexion occurs in the milky way near [ fe / h ] @xmath90 , decreasing towards solar - like values .
@xcite first noted that stars in the present - day dsph galaxies were separated in [ @xmath1/fe ] from the majority of the milky way stars . in the dsph galaxies ,
a significant contribution of type ia supernovae is observable in [ @xmath1/fe ] abundance ratios near [ fe / h ] @xmath91dex , with the exception of draco , which shows low [ @xmath1/fe ] at all metallicities .
some dwarf spheroidal stars also have a large range in [ @xmath1/fe ] values , increasing up to values seen in field stars .
the orphan stream stars have low [ @xmath1/fe ] abundance ratios ( @xmath00.22 ) with respect to the milky way sample ( @xmath00.40dex ) .
the low- and medium - probability members which we deem to be field interlopers also have low [ @xmath1/fe ] abundance ratios , but given their overall metallicity these @xmath1-element ratios are consistent with the thick disk enrichment of the milky way . since our orphan stream members ( i.e. , the high - probability members oss 6 , 8 and 14 ) are metal - poor , we can be confident we are not confusing their low-@xmath1 signature for the standard chemical enrichment of the milky way .
we stress that [ @xmath1/fe ] , on its own , should never be used as a litmus test for accretion on to the milky way .
one must carefully select candidates that appear to be stream members based on all the observables , and then examine their detailed abundances in order to infer the chemical evolution of the disrupted host .
we are already quite confident these stars are truly disrupted orphan stream members , and as such , we can say that the low [ @xmath1/fe ] abundance ratios observed in the orphan stream are consistent with the low @xmath1-element enhancement of the present - day dsph galaxies .
differences in [ ba / y ] ratios between the present - day dsph stars and the milky way have been observed by a number of groups @xcite . a number of possible explanations
have been proposed to explain this offset , including changes in sne ii yields or by changing the frequency of sne ii explosions via adjustments to the initial mass function or altering the influence of @xmath1-rich freeze - out ( e.g. , see * ? ? ?
* for a discussion ) .
abundance ratios for the orphan stream stars and field standards . milky way and dsph data compiled by @xcite
is also shown .
markers and colours are the same described in figure [ fig : alpha - fe ] . ] arbitrarily adjusting sne ii yields between two different nucleosynthesis sites appears a somewhat unlikely scenario . by employing sn frequency corrections in addition to using adjusted yields
, @xcite can reproduce the low [ ba / y ] trends in the galaxy , but not the high [ ba / y ] ratios observed in the dsph galaxies . leaving sne ii yields unchanged
, these offsets could alternatively be reproduced by simply adjusting the frequency of sne ii events through the truncation of the upper initial mass function ( imf ) . in effect , massive stars would be less numerous , which are thought to be the primary production sites for the first @xmath65-capture peak elements ( e.g. , yttrium in this case ) . because higher @xmath1-element yields are also expected for massive stars
, this may affect both the [ ba / y ] abundance and the overall [ @xmath1/fe ] abundance .
however , whether the upper imf differs between the dsph and the milky way remains an open question .
we have lower limits in [ ba / y ] for all orphan stream stars .
this is because y was not detected in any of the stream stars .
the standard stars and lower probability stream targets all have [ ba / y ] ratios that are consistent with the milky way for their given overall metallicity .
this is shown in figure [ fig : ba - y ] .
although we have only lower limits for [ ba / y ] , in oss 6 and oss 8 these are @xmath0 + 0.5dex offset from the main component of the milky way . this is a similar offset as that observed in the present - day dsphs , as compiled by @xcite .
for the most metal - poor stream star , oss 14 , our limit on y abundance yields a weak but consistent lower limit of [ ba / y ] @xmath92dex .
we note that robust lower limits for [ ba / y ] in the most metal - poor dsph stars ( [ fe / h]@xmath93 ) are difficult to ascertain . a number of associations have been proposed between the orphan stream and known milky way satellites .
however , most have been shown to be either unlikely or implausible .
here we discuss the possibility of association between the orphan stream and ngc 2419 , as well as segue 1 .
the width and length of a tidal tail are a clear indication to the nature of the parent satellite .
for the case of the orphan stream , these characteristics favor a dark - matter dominated system ( e.g. , a dwarf galaxy ) .
however , given the large apogalacticon of 90kpc for the orphan stream , a relationship has been proposed between the orphan stream and ngc 2419 @xcite .
ngc 2419 is the most distant ( @xmath94kpc ) and most luminous ( @xmath95 ) globular cluster in the milky way halo ( @xmath96kpc ) . just like the orphan stream ,
the system is also quite metal - poor : [ fe / h ] = 2.15 stars in globular clusters are known to exhibit peculiar chemical patterns , most notably in an anti - correlation between sodium and oxygen abundances ( e.g. , * ? ? ?
following this , ngc 2419 is particularly special because it shows an unusual anti - correlation between magnesium and potassium , even stronger than the classical na - o anti - correlation .
this chemical signature is so far unique to ngc 2419 , and illustrates the most extreme cases of mg - depletion and k - enhancement seen anywhere in the milky way , or its satellite systems @xcite . ) and @xcite ( @xmath97 ) .
standard stars and orphan stream candidates are shown with the same markers used in figure [ fig : alpha - fe ] . ]
we do not find any orphan stream stars to be extremely depleted in [ mg / fe ] or enhanced in [ k / fe ] ( figure [ fig : mg - k ] ) , as found in ngc 2419 @xcite . if we consider two sub - populations in ngc 2419 : a mg - poor , k - rich and a mg - normal , k - poor sample , then a two sample kolmogorov - smirnoff test demonstrates that we can be at least 99.7% confident that our orphan stream members are inconsistent with being drawn from the mg - poor ngc 2419 population .
we can not make such statements about the mg - normal population , since abundances for that sample are typical for the milky way .
the mg - poor and mg - normal sub - populations each comprise about half of the observed ngc 2419 population @xcite .
given the uniqueness of the mg - k anti - correlation , a single orphan stream star with significant mg - depletion or k - enhancement would have provided an exciting hint for the orphan stream parent .
however , given the stream s large metallicity spread @xcite , this indicates the stream is unlikely to be associated with ngc 2419 given the cluster s small dispersion in overall metallicity ( 0.17dex ; * ? ? ?
we therefore firmly rule out any association between ngc 2419 and the orphan stream .
the only known satellite system with an ambiguous association with the orphan stream is segue 1 .
the system resides in a particularly crowded region of the sky : it lies extremely close to the perigalacton of the orphan stream , and just south of the bifurcated sagittarius stream .
the original discovery of segue 1 by @xcite suggested the system was an extended globular cluster , until recent studies showed the system was an ultrafaint dwarf galaxy @xcite .
the similarities between segue 1 and the orphan stream are striking .
in addition to being nearby on the sky , segue 1 s distance of 23kpc is consistent with the closest portion of the orphan stream .
the velocities between the two systems are also coincident : + 114 km s@xmath6 for segue 1 and @xmath0 + 120 km s@xmath6 for the nearest portion of the stream .
additionally , both systems exhibit extremely low velocity dispersions , on the order of 3 - 8 km s@xmath6 .
however , stars linking the two systems have yet to be found .
the association between the two is not obvious . as @xcite summarises , ` _ are they just ships passing in the dark night ? _ ' the characteristics of the orphan stream are clues to the nature of its progenitor : the arc length and intrinsic width suggests the parent system must be dark matter - dominated .
segue 1 is the most dark matter - dominated system known , such that it has become a prime focus for indirect dark matter detection experiments @xcite .
however , there are some doubts as to whether segue 1 has experienced any tidal effects @xcite .
segue 1 is well within the galactic tidal field , and initial indications of east - west tidal effects @xcite seem to have been complicated by the crowded region bordering the system @xcite .
presently , no stars or tails clearly connecting the two systems have been found .
the chemistry of segue 1 is particularly relevant here .
segue 1 hosts an extremely wide range in stellar metallicities : the mean is found to be [ fe / h ] @xmath98 and the spread extends from [ fe / h ] = @xmath99 to @xmath100dex @xcite .
there is significant overlap between the metallicity distributions reported for segue 1 and the orphan stream .
the high - resolution spectra in this study confirm the wide metallicity range of orphan stream members found by @xcite , and independently reported by @xcite ( [ fe / h ] @xmath101 to @xmath17 ) .
thus , in addition to on - sky position , distances , velocity , and velocity dispersion , segue 1 and the orphan stream share a wide and consistent range in metallicities .
in contrast to these observables , @xcite and frebel et al .
( submitted ) find segue 1 members to be extremely @xmath1-enhanced ( [ @xmath1/fe ] @xmath102 ) , unlike what we observe in disrupted orphan stream members .
we note that although the two systems appear to be related in some respect , if segue 1 were the ` parent ' of the orphan stream , the differing [ @xmath1/fe ] abundances and lack of tidal features surrounding segue 1 after extensive examination is somewhat puzzling .
furthermore if segue 1 is the disrupted host , and if the stream metallicity gradient presented by @xcite is correct , that implies the stream becomes more metal - rich at greater distances from the parent system , contrary to what is typically observed .
we present a chemical analysis of five orphan stream candidates , three of which we confirm are true members from the disrupted orphan stream parent satellite .
the two non - members were both originally identified to have ` low ' or ` medium ' probability of stream membership from the low - resolution spectroscopy by @xcite .
we encourage high - resolution spectroscopic follow - up of the remaining high - probability orphan stream members presented in @xcite .
this work demonstrates the first detailed chemical study of the orphan stream from high - resolution spectra .
a large metallicity spread is present in the orphan stream members , confirming the work by @xcite and @xcite that the orphan stream is not mono - metallic .
the spread in overall metallicity is consistent with the internal chemical evolution of the present - day dwarf galaxies .
detailed chemical abundances confirm this scenario .
low [ @xmath1/fe ] element ratios are observed in the stream stars , and lower limits of [ ba / y ] are ascertained , which sit well above the bulk component of the milky way for two of the three stream members .
thus , we present the first detailed chemical evidence that the parent of the orphan stream is a dwarf galaxy . on the basis of chemistry
, we exclude the extended globular cluster ngc 2419 as a plausible parent to the orphan stream . no firm link between segue 1 and
the orphan stream has been identified . while the wide range in metallicities adds to the phase - space similarities between the two systems , the substantial difference in [ @xmath1/fe ] abundance ratios places doubt on any association .
it appears the disrupting dwarf galaxy parent may reside in the southern sky , just waiting to be discovered .
a.r.c . is thankful to martin asplund , david yong , amanda karakas , ana bonaca and josh simon for fruitful discussions and commentary on this work .
australian access to the magellan telescopes was supported through the national collaborative research infrastructure strategy of the australian federal government .
acknowledges the financial support through the australian research council laureate fellowship lf0992131 .
s.k . and g.dac .
acknowledge the financial support from the australian research council through discovery program dp120101237 .
a.f . acknowledges support from nsf grant ast-1255160 .
+ , r. , shectman , s. a. , gunnels , s. m. , mochnacki , s. , & athey , a. e. 2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4841 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . m. iye & a. f. m. moorwood , 16941704 | we present the first high - resolution spectroscopic study on the orphan stream for five stream candidates , observed with the magellan inamori kyocera echelle ( mike ) spectrograph on the magellan clay telescope .
the targets were selected from the low - resolution catalog of @xcite : 3 high - probability members , 1 medium and 1 low - probability stream candidate were observed .
our analysis indicates the low and medium - probability target are metal - rich field stars .
the remaining three high - probability targets range over @xmath01dex in metallicity , and are chemically distinct compared to the other 2 targets and all standard stars : low [ @xmath1/fe ] abundance ratios are observed , and lower limits are ascertained for [ ba / y ] , which sit well above the milky way trend .
these chemical signatures demonstrate that the undiscovered parent system is unequivocally a dwarf spheroidal galaxy , consistent with dynamical constraints inferred from the stream width and arc . as such
, we firmly exclude the proposed association between ngc 2419 and the orphan stream .
a wide range in metallicities adds to the similarities between the orphan stream and segue 1 , although low [ @xmath1/fe ] abundance ratios in the orphan stream are in tension with high [ @xmath1/fe ] values observed in segue 1 .
open questions remain before segue 1 could possibly be claimed as the ` parent ' of the orphan stream .
the parent system could well remain undiscovered in the southern sky . |
one goal of relativistic heavy ion collision experiments is to explore and map the qcd phase diagram as a function of temperature and baryon chemical potential .
recent theoretical developments suggest that a key qualitative feature , namely a critical point which in a sense defines the landscape to be mapped , may be within reach of discovery and analysis by the cern sps or by rhic , as data is taken at several different energies @xcite .
the discovery of the critical point would in a stroke transform the map of the qcd phase diagram from one based only on reasonable inference from universality , lattice gauge theory and models into one with a solid experimental basis @xcite . in qcd with two massless quarks ( @xmath6 ; @xmath7 )
the phase transition at which chiral symmetry is restored as @xmath1 is increased with @xmath8 is likely second order and belongs to the universality class of @xmath9 spin models in three dimensions @xcite .
below @xmath10 , chiral symmetry is broken and there are three massless pions . at @xmath11 , there are four massless degrees of freedom : the pions and the sigma . above @xmath11 , the pion and sigma correlation lengths are degenerate and finite . in nature , the light quarks are not massless . because of this explicit chiral symmetry breaking ,
the second order phase transition is replaced by an analytical crossover : physics changes dramatically but smoothly in the crossover region , and no correlation length diverges .
this picture is consistent with present lattice simulations @xcite , which suggest @xmath12 mev @xcite .
[ fig : phasediagram ] arguments based on a variety of models @xcite indicate that the transition as a function of @xmath1 is first order at large @xmath2.-axis in two - flavor qcd @xcite .
] this suggests that the phase diagram features a critical point @xmath0 at which the line of first order phase transitions present for @xmath13 ends , as shown in figure 1 . would be a tricritical point , at which the first order transition becomes second order .
] at @xmath14 , the phase transition at @xmath15 is second order and is in the ising universality class @xcite .
although the pions remain massive , the correlation length in the @xmath16 channel diverges due to universal long wavelength fluctuations of the order parameter .
this results in characteristic signatures , analogues of critical opalescence in the sense that they are unique to collisions which freeze out near the critical point , which can be used to discover @xmath0 @xcite .
the position of the critical point is , of course , not universal .
furthermore , it is sensitive to the value of the strange quark mass .
@xmath14 decreases as @xmath17 is decreased @xcite , and at some @xmath18 , it reaches @xmath19 and the transition becomes entirely first order @xcite .
the value of @xmath18 is an open question , but lattice simulations suggest that it is about half the physical strange quark mass @xcite , although these results are not yet conclusive @xcite . of course
, experimentalists can not vary @xmath17 .
they can , however , vary @xmath2 .
the ags , with beam energy 11 agev corresponding to @xmath20 gev , creates fireballs which freeze out near @xmath21 mev @xcite .
when the sps runs with @xmath22 gev ( beam energy 158 agev ) , it creates fireballs which freeze out near @xmath23 mev @xcite .
rhic will make even smaller values of @xmath2 accessible . by dialing @xmath24 and thus @xmath2
, experimenters can find the critical point @xmath0 . finding @xmath0 would confirm that @xmath18 is less than the physical value of @xmath17 , and that the phase transition at lower chemical potentials ( @xmath25 ) is a smooth crossover .
predicting @xmath14 , and thus suggesting the @xmath24 to use to find @xmath0 , is beyond the reach of present theoretical methods because @xmath14 is both nonuniversal and sensitively dependent on the mass of the strange quark .
crude models suggest that @xmath14 could be @xmath26 mev in the absence of the strange quark @xcite ; this in turn suggests that in nature @xmath14 may have of order half this value , and may therefore be accessible at the sps if the sps runs with @xmath27 gev .
however , at present theorists can not predict the value of @xmath14 even to within a factor of two .
the sps can search a significant fraction of the parameter space ; if it does not find @xmath0 , it will then be up to the rhic experiments to map the @xmath28 mev region .
locating @xmath0 on the phase diagram can only be done convincingly by an experimental discovery .
theorists can , however , do reasonably well at describing the phenomena that occur near @xmath0 , thus enabling experimenters to locate it .
this is the goal of ref .
the signatures proposed there are based on the fact that @xmath0 is a genuine thermodynamic singularity at which susceptibilities diverge and the order parameter fluctuates on long wavelengths .
the resulting signatures are _ nonmonotonic _ as a function of @xmath24 : as this control parameter is varied , we should see the signatures strengthen and then weaken again as the critical point is approached and then passed .
the simplest observables to use are the event - by - event fluctuations of the mean transverse momentum of the charged particles in an event , @xmath29 , and of the total charged multiplicity in an event , @xmath30 .
the fluctuations observed in sps collisions by na49 are as perfect gaussians as the data statistics allow , as expected for freeze - out from a system in thermal equilibrium @xcite .
the data on multiplicity fluctuations show evidence for a nonthermodynamic contribution , which is to be expected since the extensive quantity @xmath30 is sensitive to the initial size of the system and thus to nonthermodynamic effects like variation in impact parameter .
the contribution of such effects to the fluctuations have now been estimated @xcite ; the combined thermodynamic and nonthermodynamic fluctuations are in satisfactory agreement with the data @xcite .
the width of the event - by - event distribution of the mean @xmath29 of the charged pions in a single event is in good agreement with predictions based on noncritical thermodynamic fluctuations @xcite .
that is , na49 data are consistent with the hypothesis that almost all the observed event - by - event fluctuation in mean @xmath29 , an intensive quantity , is thermodynamic in origin .
this bodes well for the detectability of systematic changes in thermodynamic fluctuations near @xmath0 .
one analysis described in detail in ref .
@xcite is based on the ratio of the width of the true event - by - event distribution of the mean @xmath29 to the width of the distribution in a sample of mixed events .
this ratio was called @xmath31 .
na49 has measured @xmath32 @xcite , which is consistent with expectations for noncritical thermodynamic fluctuations ..
bose effects increase @xmath31 by @xmath33 @xcite ; an anticorrelation introduced by energy conservation in a finite system when one mode fluctuates up it is more likely for other modes to fluctuate down decreases @xmath31 by @xmath33 @xcite ; two - track resolution also decreases @xmath31 by @xmath33 @xcite .
the contributions due to correlations introduced by resonance decays and due to fluctuations in the flow velocity are each much smaller than @xmath34 @xcite . ]
critical fluctuations of the @xmath16 field , i.e. the characteristic long wavelength fluctuations of the order parameter near @xmath0 , influence pion momenta via the ( large ) @xmath35 coupling and increase @xmath31 @xcite .
the effect is proportional to @xmath36 , where @xmath37 is the @xmath16-field correlation length of the long - wavelength fluctuations at freezeout @xcite . if @xmath38 fm , the ratio @xmath31 increases by @xmath39 , ten to twenty times the statistical error in the present measurement @xcite .
this observable is valuable because data on it has been analyzed and presented by na49 , and it can therefore be used to learn that pb+pb collisions at 158 agev do _ not _ freeze out near @xmath0 .
once @xmath0 is located , however , other observables which are more sensitive to critical effects will be more useful . for example
, a @xmath40 , defined using only the softest @xmath41 of the pions in each event , will be much more sensitive to the critical long wavelength fluctuations .
the higher @xmath29 pions are less affected by the @xmath16 fluctuations @xcite , and these relatively unaffected pions dominate the mean @xmath29 of all the pions in the event .
this is why the increase in @xmath31 near the critical point will be much less than that of @xmath40 .
the multiplicity of soft pions is an example of an observable which may be used to detect the critical fluctuations without an event - by - event analysis .
the post - freezeout decay of sigma mesons , which are copious and light at freezeout near @xmath0 and which decay subsequently when their mass increases above twice the pion mass , should result in a population of pions with @xmath42 which appears only for freezeout near the critical point @xcite . if @xmath43 , this population of unusually low momentum pions will be comparable in number to that of the `` direct '' pions ( i.e. those which were pions at freezeout ) and will result in a large signature .
the variety of observables which should _ all _ vary nonmonotonically with @xmath24 ( and should all peak at the same @xmath24 ) is sufficiently great that if it were to turn out that @xmath44 mev , making @xmath0 inaccessible to the sps , all four rhic experiments could play a role in the study of the critical point .
the purpose of ref . @xcite is to estimate how large @xmath37 can become , thus making the predictions of ref .
@xcite for the magnitude of various signatures more quantitative . in an ideal system of infinite size which was held at @xmath15 ; @xmath46 for an infinite time , the correlation length @xmath45 would be infinite . ref .
@xcite estimated that finite size effects limit @xmath45 to be about 6 fm at most .
we argue in ref .
@xcite that limitations imposed by the finite duration of a heavy ion collision are more severe , preventing @xmath45 from growing larger than about @xmath47 fm . the nonequilibrium dynamics analyzed in ref .
@xcite is _ guaranteed _ to occur in a heavy ion collision which passes near @xmath0 , even if local thermal equilibrium is achieved at a higher temperature during the earlier evolution of the plasma created in the collision .
we assume early thermal ( although not necessarily chemical ) equilibration , and ask how the system evolves out of equilibrium as it passes @xmath0 . for the present , assume that the system cools through the critical point @xmath0 as sketched in trajectory ( b ) of figure 1 .
if it were to cool arbitrarily slowly , @xmath48 would be maintained at all temperatures , and @xmath45 would diverge at @xmath49 .
however , it would take an infinite time for @xmath45 to grow infinitely large . indeed , near a critical point ,
the longer the correlation length , the longer the equilibration time , and the slower the correlation length can grow .
this critical slowing down means that the correlation length can not grow as fast as @xmath50 , and the system can not stay in equilibrium .
we use the theory of dynamical critical phenomena to describe the effects of critical slowing down on the time development of the correlation length @xmath51 , using the three - dimensional ising model to describe @xmath52 near @xmath0 @xcite . in the ising model , the order parameter and the correlation length are functions of the reduced temperature @xmath53 and the magnetic field @xmath54 .
( in the ising model , @xmath53 is defined as @xmath55 and is usually called @xmath56 ; we reserve the symbol @xmath56 for time , however . )
we assume that the @xmath53- and @xmath54-axes are as sketched in figure 1 , although in reality they will be deformed .
we consider cooling along trajectories that pass through @xmath0 at various angles , like ( a ) and ( b ) sketched in figure 1 and also consider trajectories like ( c ) which come near to but miss @xmath0 .
our results depend on the universal function describing @xmath57 , on the universal dynamical exponent @xmath58 describing critical slowing down ( perturbations away from equilibrium relax toward equilibrium on a timescale which scales with @xmath45 like @xmath59 @xcite ) , on the nonuniversal constant @xmath60 , the nonuniversal constants which relate @xmath53 and @xmath54 to @xmath61 and @xmath62 , on @xmath49 which we take to be @xmath63 mev , and finally on the cooling rate @xmath64 which we estimate to be 4 mev / fm @xcite .
our results have a number of consequences , detailed in the remainder of this section , which should be taken into account both in planning experimental searches for the qcd critical point , and in planning future theoretical work .
we do indeed find that because of the critical slowing down of the long wavelength dynamics near @xmath0 , the correlation length does not have time to grow as large as it would in equilibrium : we find @xmath65 fm for trajectories passing near @xmath0 .
although critical slowing down hinders the growth of @xmath45 , it also slows the decrease of @xmath45 as the system continues to cool below the critical point . as a result
, @xmath45 does not decrease significantly between the phase transition and freezeout .
our estimate that @xmath45 does not grow larger than @xmath66 is robust in three senses .
first , it depends very little on the angle with which the trajectory passes through @xmath0 .
second , it turns out to depend on only one combination of all the nonuniversal quantities which play a role .
we call this parameter @xmath67 ; it is proportional to @xmath68 .
third , our results do not depend sensitively on @xmath67 .
we show that the maximum value of @xmath45 scales like @xmath69 @xcite.he @xcite . ]
thus , for example , @xmath64 would have to be a factor of 25 smaller than we estimate in order for @xmath45 to grow to @xmath70 instead of @xmath66 .
although our results are robust in this sense , they can not be treated as precise because our assumption that the dynamics of @xmath45 in qcd is described by the universal classical dynamics of the three - dimensional ising model only becomes precise if @xmath71 , while our central result is that @xmath45 does not grow beyond @xmath72 .
a @xmath73-dimensional quantum field theoretical treatment of the interplay between cooling and the dynamics of critical slowing down is not yet available , but promising first steps in this direction can be found in ref .
@xcite .
a result which is of great importance in the planning of experimental searches is that one need not hit @xmath0 precisely in order to find it .
our analysis of trajectories like ( c ) of figure 1 demonstrates that if one were to do a scan with collisions at many finely spaced values of the energy and thus @xmath2 , one would see signatures of @xmath0 with approximately the same magnitude over a broad range of @xmath2 .
the magnitude of the signatures will not be narrowly peaked as @xmath2 is varied . as long as one gets close enough to @xmath0 that the equilibrium correlation length is @xmath74
, the actual correlation length @xmath45 will grow to @xmath72 .
there is no advantage to getting closer to @xmath0 , because critical slowing down prevents @xmath45 from getting much larger even if @xmath50 does .
data at many finely spaced values of @xmath2 is _ not _ called for .
over which @xmath75 fm is about @xmath76 mev for @xmath77 mev .
similar results can be obtained @xcite within a random matrix model @xcite .
it is likely over - optimistic to estimate @xmath76 mev when the effects of the strange quark are included and @xmath14 itself is reduced .
a conservative estimate would be to use the models to estimate that @xmath78 . ] knowing that we are looking for @xmath79 fm is very helpful in suggesting how to employ the signatures described in detail in ref .
the excess of pions with @xmath42 arising from post - freezeout decay of sigmas is large as long as @xmath80 , and does not increase much further if @xmath37 is longer .
this makes it an ideal signature .
the increase in the event - by - event fluctuations in the mean transverse momentum of the charged pions in an event ( described by the ratio @xmath81 of ref .
@xcite ) is proportional to @xmath36 .
the results of ref .
@xcite suggest that for @xmath38 fm , this will be a @xmath82 effect .
this is ten to twenty times larger than the statistical error in the present na49 data , but not so large as to make one confident of using this alone as a signature for @xmath0 .
the solution is to use signatures which focus on the event - by - event fluctuations of only the low momentum pions .
unusual event - by - event fluctuations in the pion momenta arise via the coupling between the pions and the sigma order parameter which , at freezeout , is fluctuating with correlation length @xmath37 .
this interaction has the largest effect on the softest pions @xcite .
@xmath40 , described in the previous section , is a good example of an observable which takes advantage of this .
depending on the details of the cuts used to define it , it should be enhanced by many tens of percent in collisions passing near @xmath0
@xcite suggests other such observables , and more can surely be found .
together , the excess multiplicity at low momentum ( due to post - freezeout sigma decays ) and the excess event - by - event fluctuation of the momenta of the low momentum pions ( due to their coupling to the order parameter which is fluctuating with correlation length @xmath37 ) should allow a convincing detection of the critical point @xmath0
. both should behave nonmonotonically as the collision energy , and hence @xmath2 , are varied .
both should peak for those heavy ion collisions which freeze out near @xmath0 , with @xmath38 fm .
as the matter created in a heavy ion collision cools through the qcd phase transition , it may be driven out of equilibrium .
the way in which this may happen is quite different depending on whether the phase transition is traversed with @xmath5 , @xmath3 or @xmath4 .
if @xmath13 , the phase transition is first order ; if @xmath62 is large enough , the transition may be sufficiently strongly first order that it proceeds via the nucleation of well - separated bubbles of hadronic matter , which then grow .
this bubbling results in a spatially inhomogeneous hadronic system .
if the system freezes out before it gets re - homogenized , the result will be large , non - gaussian fluctuations in hadron multiplicity as a function of rapidity @xcite .
these overdense and underdense regions in rapidity would provide a distinctive signature , and have not been seen . if @xmath3 , the phase transition is a smooth crossover .
this is the region of the diagram where the transition may be traversed most quickly , with the most rapid cooling rate .
( for @xmath4 , the specific heat is unusually large and therefore even though the rate of decrease of energy density is not unusual , the cooling rate @xmath64 is unusually small . for @xmath5 ,
neither of the two phases which coexist at the first order phase transition has an unusually large specific heat .
the system nevertheless spends a long time at the coexistence temperature @xmath10 because of the release of latent heat .
thus , @xmath64 is allowed to be largest for @xmath3 . )
although the smooth crossover with nondivergent specific heat allows a large @xmath64 in principle , the cooling rate which is actually achieved will depend on the system size and on the duration of time between the initial collision and the phase transition . if conditions can be found in which @xmath64 can be made large enough , the dynamics may then be well - modelled as a quench , in which most of the degrees of freedom in the system are imagined to cool arbitrarily rapidly . in this circumstance ,
the long wavelength modes of the pion field become unstable to exponential growth @xcite , leading to the formation of long wavelength pion oscillations .
once excited , these oscillations may be further amplified by parametric resonance @xcite .
these long wavelength disorientations of the chiral condensate result in characteristic fluctuations in the ratio of the number of low @xmath29 neutral pions to the number of low @xmath29 charged pions @xcite .
other signatures of dcc oscillations have also been discussed @xcite .
wa98 @xcite and na49 @xcite have looked for signatures of dcc oscillations in 158 agev pbpb collisions , but none have been seen .
the considerations above suggest that rhic collisions provide more favorable conditions for dcc production than sps collisions , because @xmath2 is smaller and one is more likely to have a smooth crossover which allows large @xmath64 .
however , the hadronic systems produced in rhic collisions will be large , will traverse the phase transition late , and will freeze out even later .
the late transition reduces the cooling rate , making dcc formation less likely , and the late freeze out makes wash - out of dcc signatures a worry .
this suggests that the most favorable conditions in which to look for dcc signatures may be rhic collisions of medium - sized ions , or auau collisions at rhic with impact parameters of order the nuclear radius .
( that is , midway between central and peripheral collisions . )
quite different considerations have lead asakawa , minakata and mller to propose that non - central rhic collisions provide the most favorable conditions for dcc formation @xcite : these collisions feature strong electromagnetic @xmath83 fields which may , through the anomaly , result in a coherent excitation of the @xmath84 field .
there are therefore three reasons to suggest that non - central rhic collisions may yield dcc signatures : the smooth crossover and small system size allow the system to cool through the transition faster , favoring the instability driven growth of dccs ; the electromagnetic kick may further drive their growth ; and , the smaller system size will hasten freezeout .
the first and third reasons also apply to rhic collisions of medium - sized ions . for @xmath4 ,
the specific heat is larger in the vicinity of the critical point than anywhere else on the phase diagram , and the cooling rate @xmath64 is correspondingly slow .
this slow cooling means that those collisions which create matter which cools through the transition region with @xmath4 are _ least _ likely to be driven far from equilibrium .
that is , we do not expect dcc signatures or signatures of bubbling in these collisions .
the interesting twist described in the previous section is that it is nevertheless true that such collisions can not stay _ precisely _ in equilibrium @xcite .
even though the system cools through the vicinity of @xmath0 slowly , the correlation length still grows much too slowly to become as long as it would in equilibrium .
we have discussed this phenomenon of slowing out of equilibrium at length , as it determines the magnitude of signatures of the critical point . here , we note that this non - equilibrium physics is much milder than quenching or bubbling .
quenching or bubbling create _ more _ fluctuations than would be obtained in thermal equilibrium .
they also create fluctuations which are non - gaussian in characteristic ways .
in contrast , slowing out of equilibrium serves to _ reduce _ the growth of fluctuations relative to what would be achieved if equilibrium were maintained .
the critical point @xmath0 will be signalled by an increase in the magnitude of suitably chosen event - by - event fluctuations , for example parametrized by the growth in @xmath31 and @xmath40 , but these fluctuations are expected to remain gaussian .
slowing out of equilibrium means that these gaussian fluctuations are smaller than they would be if equilibrium were maintained near @xmath0 .
they are nevertheless larger than the equilibrium thermal fluctuations far from @xmath0 and hence result in nonmonotonic signatures which peak for collisions which freezeout near @xmath0 .
will not be washed out before freezeout ; if necessary , however , smaller ions or non - central collisions can be used to further elevate the freezeout temperature @xcite . ]
we have learned much from the beautiful gaussian event - by - event fluctuations observed by na49 .
the magnitude of these flucutations are consistent with the hypothesis that the hadronic system at freezeout is in approximate thermal equilibrium .
these and other data show no signs that the system at freezeout has recently been bubbled or quenched .
there is also no sign of the larger , but still gaussian , fluctuations which would signal freezeout near the critical point @xmath0 . combining these observations with the observation of several tantalizing indications that the matter created in sps collisions is not well described at early times by hadronic models @xcite suggests that collisions at the sps may be exploring the crossover region to the left of the critical point @xmath0 ,
in which the matter is not well - described as a hadron gas but is also not well - described as a quark - gluon plasma .
this speculation could be confirmed in two ways .
first , if the sps is probing the crossover region then the coming experiments at rhic may discover direct signatures of an early partonic phase , which are well - described by theoretical calculations beginning from an equilibrated quark - gluon plasma .
second , if 158 agev collisions are probing the crossover region not far to the left of the critical point @xmath0 , then running at lower energies would result in the discovery of @xmath0 .
if , instead , rhic were to discover @xmath0 with @xmath44 mev , that would indicate that the sps experiments have probed the weakly first order region just to the right of @xmath0 .
regardless , discovering @xmath0 would take all the speculation out of mapping this part of the qcd phase diagram .
a. barducci , r. casalbuoni , s. decurtis , r. gatto , g. pettini , phys .
b231 * , 463 ( 1989 ) ; s.p .
klevansky , rev .
phys . * 64 * , 649 ( 1992 ) ; a. barducci , r. casalbuoni , g. pettini and r. gatto , phys . rev . * d49 * , 426 ( 1994 ) . s. mrowczynski and b. mller , phys .
* b363 * , 1 ( 1995 ) ; h. hiro - oka and h. minakata , phys .
* b425 * , 129 ( 1998 ) ; d. i. kaiser , phys .
rev . * d59 * , 117901 ( 1999 ) ; h.
hiro - oka and h. minakata , phys .
* c61 * , 044903 ( 2000 ) .
a. a. anselm , phys .
* b217 * , 169 ( 1989 ) ; a. a. anselm and m. g. ryskin , phys .
lett . * b266 * , 482 ( 1991 ) ; j .-
blaizot and a. krzywicki , phys .
* d46 * , 246 ( 1992 ) ; j. d. bjorken , int . j. mod .
* a7 * , 4189 ( 1992 ) ; j. d. bjorken , acta phys . pol . * b23 * , 561 ( 1992 ) ; k. l. kowalski and c. c. taylor , hep - ph/9211282 . c. greiner , c. gong and b. mller , phys .
lett . * b316 * , 226 ( 1993 ) ; z. huang , m. suzuki and x. wang , phys .
rev . * d50 * , 2277 ( 1994 ) ; s. gavin , nucl . phys . *
a590 * , 163 ( 1995 ) ; d. boyanovsky , h. j. de vega , r. holman and s. prem kumar , phys .
* d56 * , 5233 ( 1997 ) ; y. kluger , v. koch , j. randrup and x. wang , phys . rev . * c57 * , 280 ( 1998 ) ; c. chow and t. d. cohen , phys . rev . * c60 * , 054902 ( 1999 ) . h. minakata and b. mller , phys .
* b377 * , 135 ( 1996 ) ; m. asakawa , h. minakata and b. mller , nucl .
* a638 * ( 1998 ) 443c ; m. asakawa , h. minakata and b. mller , phys .
* d58 * , 094011 ( 1998 ) .
the earlier freeze out of smaller systems has been confirmed experimentally by observing the @xmath60-dependence of the freeze - out temperature via analyses of flow @xcite , coulomb effects ( h. w. barz , j. p. bondorf , j. j. gaardhoje and h. heiselberg , phys . rev . * c57 * ( 1998 ) 2536 ) , and pion interferometry ( u. heinz , nucl .
a638 * , 357c ( 1998 ) . ) | i review arguments for the existence of a critical point @xmath0 in the qcd phase diagram as a function of temperature @xmath1 and baryon chemical potential @xmath2 .
i describe how heavy ion collision experiments at the sps and rhic can discover the tell - tale signatures of such a critical point , thus mapping this region of the qcd phase diagram .
i contrast the different ways in which the matter produced in a heavy ion collision can be driven out of equilibrium : quenching out of equilibrium ( possible , but not guaranteed , if the transition region is traversed at @xmath3 ) vs. slowing out of equilibrium ( guaranteed for @xmath4 ) vs. bubbling out of equilibrium ( possible , but not guaranteed , for @xmath5 ) .
quenching or bubbling create and amplify distinct , detectable , non - gaussian fluctuations .
in contrast , slowing out of equilibrium reduces the magnitude of the specific , detectable , gaussian fluctuations which signal the presence of the critical point . |
in this paper we study fractal geometric aspects of the limit set @xmath3 of a normal subgroup @xmath0 of some given non - elementary kleinian group @xmath1 acting on @xmath4-dimensional hyperbolic space @xmath5 .
we always assume that @xmath2 contains a coset which is represented by some loxodromic isometry @xmath6 .
it is well known that in this situation @xmath3 coincides with the limit set @xmath7 of the larger group @xmath1 .
however , a comparison of finer aspects of these two limit sets usually turns out to be far more involved , as can be seen , for instance , in the work of brooks @xcite and rees @xcite .
+ in this paper we investigate the set @xmath8 of directions at some arbitrary point @xmath9 on the manifold @xmath10 associated with @xmath0 for which the resulting geodesic movement on @xmath10 eventually escapes from every compact region on @xmath10 , but which is nevertheless contained in the @xmath11-neighbourhood of some sequence of closed loops starting and ending at @xmath9 on @xmath10 , for each @xmath12 .
that is , we consider the transient limit set @xmath8 of @xmath0 , given by @xmath13 here , @xmath14 refers to the hyperbolic metric in @xmath5 and @xmath15 denotes a @xmath16-parametrisation of the geodesic ray from the origin to @xmath17 .
+ the following theorem gives the main result of this paper .
let @xmath1 and @xmath0 be given as above .
we then have @xmath18 this theorem gives a extension of results by fernndez and melin @xcite , who studied the set @xmath19 of escaping , not necessarily loop - approximable directions on a complete oriented non - compact riemann surface @xmath20 with fundamental group @xmath21 .
that is , @xmath22 , where @xmath23 refers to the boundary at infinity of @xmath24 and @xmath25 denotes the hyperbolic metric in @xmath24 .
the main result of ( * ? ? ?
* theorem 1 ) was to establish the following tricothomy . *
if @xmath20 has finite area , then @xmath19 is countable . * if brownian motion on @xmath20 is transient , then @xmath19 has full lebesgue measure . *
if @xmath20 has infinite area and brownian motion is recurrent , then @xmath19 has zero lebesgue measure , but its hausdorff dimension is equal to @xmath26 .
therefore , for hyperbolic manifolds which are normal coverings of some hyperbolic manifold and which posess a loxodromic representative @xmath27 , our main theorem extends the results by fernndez and melin to arbitrary dimensions and to the situation where the boundary at infinity of hyperbolic space is replaced by the limit set of the fundamental group .
however , let us emphasize that our proof does require the existence of a loxodromic @xmath27 , and hence our extension is restricted to normal coverings with this property .
our proof hinges on two renormalisation procedures .
these are used to locate a certain family of subsystems within the loop - approximable , non - recurrent dynamics on the manifold @xmath28 . at the boundary of the universal covering space these subsystems
are described by a family of cantor sets contained in @xmath3 , and here the key observation is that this family contains cantor sets whose hausdorff dimension is arbitrarily close to the hausdorff dimension of @xmath7 .
these cantor sets are constructed inductively using the following two renormalisation procedures .
the first of these employs a well - known construction by bishop and jones for computing the hausdorff dimension of bounded dynamics ( see @xcite , @xcite ) .
this construction gives rise to a certain weighted scaling law , which we then apply a sufficient number of times in order to prepare for the second renormalisation step . on the manifold @xmath28
this first step of the overall construction corresponds to a family of well separated quasi - geodesics within a bounded region of @xmath28 , each starting at the same point .
the second renormalisation procedure consists in prolonging each of these quasi - geodesics by a long geodesic segment which is contained in the projection of the axis @xmath29 of @xmath30 onto @xmath28 , and which is chosen such that it leads out of the bounded region which contained the original quasi - geodesics .
the resulting dynamical behaviour on @xmath28 is sketched in figure [ mnf ] .
we assume that the reader is familiar with the proof of bishop and jones result on the relationship between the exponent of convergence @xmath31 of a non - elementary kleinian group @xmath1 and the hausdorff dimension of the radial limit set @xmath32 of @xmath1 ( see @xcite and for a more detailed proof @xcite ) . here
, the reader might like to recall that the radial limit set @xmath32 represents those limit points @xmath17 for which the projection of a hyperbolic ray towards @xmath17 returns infinitely often to some compact part of the manifold associated to @xmath1 .
@xmath33 .,title="fig : " ] [ mnf ]
throughout , let @xmath1 be a non - elementary kleinian group acting on @xmath4-dimensional hyperbolic space @xmath5 . also , let @xmath0 be a normal subgroup of @xmath1 such that @xmath2 contains a coset @xmath34 $ ] , for some loxodromic @xmath6 .
moreover , we always assume without loss of generality that @xmath35 is an element of the axis @xmath36 of @xmath30 , and we let @xmath37 ( @xmath38 resp . ) refer to the repulsive ( attractive resp . )
fixed point of @xmath30 .
next , recall that to any arbitrary kleinian group @xmath21 we can associate its truncated poincar series @xmath39 , as well as its poincar series @xmath40 .
these series are given , for @xmath41 and @xmath42 , by @xmath43 the abzissa of convergence of the infinite series @xmath40 is called the exponent of convergence of @xmath21 , and it will be denoted by @xmath44 . here , note that bishop and jones @xcite ( see also @xcite ) showed that if @xmath21 is non - elementary , then we always have that @xmath45 also , we will make use of the following standard facts and notations for the poincar model @xmath46 of the @xmath4-dimensional hyperbolic space .
let @xmath47 refer to the hyperbolic ball centred at @xmath48 of radius @xmath49 , and let @xmath50 denote the radial projection from the origin to the boundary @xmath51 of hyperbolic space .
that is , for @xmath52 we have @xmath53 , where @xmath54 refers to the euclidean straight line between the origin and @xmath17 .
also , we require the following analogue of pythagoras s theorem for hyperbolic triangles . for this , consider a hyperbolic triangle with sides of finite lengths @xmath55 , @xmath56 and @xmath57 , and with @xmath58 denoting the angle opposite to the side of length @xmath55 .
a straightforward application of the hyperbolic cosine rule ( see e.g. @xcite ) then gives that there exists a constant @xmath59 , depending only on @xmath60 , such that @xmath61 moreover , we require the following additional observation from elementary hyperbolic geometry . for this , consider some arbitrary hyperbolic geodesic @xmath62 which does not contain the origin , and let @xmath63 denote the geodesic ray connecting the origin with one of the endpoints @xmath64 of @xmath65 .
also , let @xmath66 refer to the summit of the geodesic @xmath65 .
that is , @xmath66 is uniquely determined by @xmath67 .
a straightforward exercise in hyperbolic geometry then shows that there exists a universal constant @xmath68 such that @xmath69 in fact , an elementary calculation shows that @xmath70 is equal to @xmath71 , which is often referred to as schweikart s constant . for a sequence @xmath72 of distinct points in @xmath5 ,
let @xmath73\ ] ] denote the quasi - geodesic path obtained by connecting @xmath74 and @xmath75 with the unique geodesic arc between them , for each @xmath76 .
a standard observation from hyperbolic geometry then shows that if the lengths of these geodesic segments are uniformly bounded away from @xmath77 , and if each of the angles between adjacent geodesic segments is uniformly bounded from below by some @xmath78 , then @xmath79 $ ] is a quasi - geodesic ray towards a unique point at infinity .
that is , each @xmath74 is , with respect to the hyperbolic metric , uniformly bounded ( depending on @xmath60 ) away from the geodesic ray from @xmath80 towards the uniquely determined limit at the boundary at infinity of the sequence @xmath81 .
finally , note that we use the common notation @xmath82 if two sequences of positive real numbers @xmath83 and @xmath84 are comparable , that is , if the ratio @xmath85 is uniformly bounded from below by @xmath86 and from above by @xmath57 , for some @xmath87 and for all @xmath76 .
let us begin with by giving our first renormalisation procedure . here ,
a _ geodesic @xmath0-tree @xmath88 rooted at @xmath9 _ refers to an infinite tree whose set of vertices @xmath89 is contained in @xmath90 and whose edges are finite geodesic segments between the vertices , such that each vertex @xmath91 has a finite set @xmath92 of successors of cardinality at least @xmath93 and such that each element in @xmath94 has a unique predecessor . in the following , let @xmath95 be defined by @xmath96 , for each @xmath97 . note that our first renormalisation procedure is well known for the special case in which each @xmath95 lies in the orbit @xmath98 . in this situation ,
its outcome has already been obtained in ( * ? ? ?
* proposition 3.5 ) .
the novelty here is that for the normal subgroup @xmath0 the result of ( * ? ? ?
* proposition 3.5 ) continues to hold for each element of the orbit @xmath99 of the origin under @xmath100 . * recurrent renormalisation procedure ( [ rrp ] ) .
* @xmath101 _ for each @xmath102 , there exist @xmath103 , @xmath104 and @xmath105 such that for each @xmath106 and @xmath107 there exists a geodesic @xmath0-tree @xmath108 rooted at @xmath109 with the following properties . _ * if @xmath110 , then @xmath111 for each @xmath112 . * if @xmath112 for some @xmath110 , then @xmath113 . * if @xmath114 for some @xmath110 , then @xmath115 and @xmath116 . * for each @xmath110 we have @xmath117 we will say that the so derived family @xmath118 is obtained by applying the recurrent renormalisation procedure to @xmath110 .
as already mentioned before , for @xmath119 the assertion in this procedure has been obtained in ( * ? ? ?
* proposition 3.5 ) ( see also @xcite ) , and we refer to these papers for the proof in this case . in fact , note that the main idea of the proof of ( * ? ? ?
* proposition 3.5 ) consists of a geometrization of the rate of increase , for @xmath120 tending to infinity , of the truncated poincar series @xmath121 for @xmath122 . for the proof of the general situation , that is , for some arbitrary @xmath107 , note that the value of the truncated poincar series associated with @xmath0 does not change if we exchange the observation point @xmath123 by some arbitrary point in @xmath124 .
more precisely , since @xmath0 is normal in @xmath1 , we have for each @xmath97 and @xmath125 , @xmath126 using this observation , the assertion now follows from a straightforward adaptation of the arguments in the proof of ( * ? ? ?
* proposition 3.5 ) . * transient renormalisation procedure ( [ trp ] ) . * @xmath101 _ for @xmath127 , @xmath107 and @xmath128 , let @xmath108 denote the geodesic @xmath0-tree obtained in the recurrent renormalisation procedure _
[ rrp]_. then there exists a constant @xmath129 such that for each @xmath130 sufficiently large and for each @xmath106 with @xmath131 , we have that @xmath132 where @xmath133 .
we will say that the ball @xmath134 is obtained by starting at @xmath109 and applying the transient renormalisation procedure @xmath135 times .
_ @xmath33 .,title="fig : " ] [ etaplus ] let @xmath108 and @xmath106 be given as stated in the renormalisation procedure .
let us first show that @xmath136 lies always close to the summit @xmath137 of the geodesic @xmath138 , the image of the axis @xmath139 under @xmath140 .
indeed , since @xmath141 , the statement in ( 1 ) of [ rrp ] implies that there exists @xmath142 such that @xmath143 ( see also figure [ etaplus ] ) .
the elementary observation in ( [ axis ] ) shows that the distance from the summit of a geodesic to each of the two rays from the origin to the endpoints of the geodesic is less than @xmath70 .
thus , by construction , we have that @xmath144 recall that in the recurrent renormalisation procedure [ rrp ] we have already derived the existence of the parameter @xmath145 , which is the upper bound of the lengths of the edges in the tree @xmath146 .
next , consider the geodesic @xmath147 containing the points @xmath148 and @xmath109 ; one of its endpoints will be @xmath149 .
assume , by way of contradiction , that the distance between @xmath109 and the summit @xmath137 of @xmath147 is larger than @xmath150 .
projecting onto the manifold @xmath28 associated to @xmath0 and using the hyperbolic triangle inequality , we obtain a contradiction to the fact that @xmath151 .
it immediately follows that @xmath152 where the comparability constant depends only on the distance @xmath150 , and therefore , only on @xmath0 and @xmath153 .
the statement now follows by applying @xmath154 to the ball @xmath155 , which immediately gives that @xmath156 where @xmath157 .
let us remark that the constant @xmath158 in [ rrp ] and the constant @xmath70 in [ trp ] are independent of each other . also , the statements in [ rrp ] continue to hold if we replace @xmath158 by a smaller positive number , and the same holds for @xmath70 in [ trp ] .
therefore , for the remainder of this paper , when applying [ rrp ] and [ trp ] , we use @xmath159 instead of @xmath158 and @xmath70 .
let us first observe that it is sufficient to prove the assertion in the theorem for the case in which @xmath160 indeed , this can immediately be seen by way of contradiction as follows .
suppose that @xmath161 .
since diminishing a set by a subset of smaller hausdorff dimension does not alter the hausdorff dimension of that set and using the well known fact that @xmath162 ( see @xcite and @xcite ) , we obtain @xmath163 therefore , we can now assume , without loss of generality , that @xmath164 the rough strategy for proving the main theorem in this case is as follows . for some arbitrary given @xmath165 , we construct a certain cantor set @xmath166 , and then show that @xmath167 . by the arbitrary choice of @xmath153 , the theorem then follows .
the idea of the cantor set construction is to start at the origin and then to perform an alternating inductive process using both renormalisation procedures .
the building block of this process is that we first apply the recurrent renormalisation procedure [ rrp ] sufficiently many times until the resulting power of @xmath168 is large enough ( in fact , this number of times depends on the outcome of the step to come ) .
after that , we perform the transient renormalisation procedure [ trp ] sufficiently many times , without loosing the control on the distortion ( in particular , this step will guarantee that out cantor set contains only transient limit points ) .
more precisely , let @xmath169 be given .
then @xmath153 determines the width @xmath145 of the recurrent renormalisation procedure [ rrp ] . having fixed @xmath145
, we choose @xmath170 , the number of times we are going to apply the transient renormalisation procedure [ trp ] , so that @xmath171 this choice of @xmath135 will guarantee that the cantor set @xmath172 we are going to construct will be contained in @xmath173 .
finally , we choose @xmath174 to be minimal with respect to the property @xmath175 let us now come to the explicit construction of @xmath172 .
as already mentioned , the construction starts at the origin , and we set @xmath176 .
then , the first step is to apply the recurrent renormalisation procedure [ rrp ] @xmath177 times , starting at @xmath178 . according to [ rrp ] , this gives rise to a set of hyperbolic balls @xmath179 whose radial projections to the boundary @xmath51 are pairwise disjoint and of comparable diameter .
the set of centres of these balls in @xmath5 will be denoted by @xmath180 .
then , the second step is to apply the transient renormalisation procedure [ trp ] @xmath135 times to each element in @xmath180 .
the set of centres of the so obtained hyperbolic balls will be denoted by @xmath181 .
this represents the start of the induction , and we then continue as follows .
assume that the sets @xmath182 and @xmath183 have been constructed , for some @xmath76 .
to each of the points in @xmath183 we then apply the recurrent renormalisation procedure [ rrp ] @xmath177 times .
the set of centres of these so obtained hyperbolic balls gives the set @xmath184 .
next , we apply the transient renormalisation procedure [ trp ] @xmath135 times to each of the elements in @xmath184 .
the set of centres of these so obtained hyperbolic balls gives rise to the set @xmath185 .
this finishes our alternating inductive argument ( see also figure [ induct ] ) , and we can now use it to define our desired cantor set @xmath172 by @xmath186 here , @xmath187 refers to the constant which we specified at the end of section [ 3 ] . + next ,
observe that in this cantor set construction we have good control over the distortion , when going from one generation in the construction to the next .
that is , by using [ rrp ] @xmath188 , [ trp ] and the condition in ( [ constants ] ) , we have the following crucial estimate , for each @xmath76 , @xmath189 using a straightforward generalisation of the folklore arguments from fractal geometry of ( * ? ? ?
* lemma 2.5 ) and ( * ? ? ?
* corollary 2.6 ) , the latter estimate immediately gives that @xmath190 it remains to show that the set @xmath172 is contained in @xmath8 .
for this , note that , by viewing the construction of @xmath172 from within @xmath5 , the set @xmath172 gives rise to a geodesic @xmath1-tree which is rooted at the origin and whose vertex set is equal to @xmath191 . by construction
, this tree has the property that the lengths of the constituting geodesic edges and the angles formed by adjacent edges are uniformly bounded away from zero .
therefore , each path in this tree starting at the origin is a quasi - geodesic heading towards a uniquely determined point at infinity .
clearly , the projection of each of these quasi - geodesics onto the manifold @xmath10 gives some piecewise geodesic movement in @xmath10 which has the following properties .
if an edge in the tree starts at a vertex in @xmath192 and ends at a point in @xmath182 , then in @xmath10 this edge is represented by a geodesic loop of hyperbolic length at most @xmath193 .
obviously , this loop must then be contained in a bounded region of @xmath10 of diameter at most @xmath193 .
whereas , if an edge in the tree starts at a vertex in @xmath182 and ends at a point in @xmath183 , using [ trp ] , we then have that in @xmath10 this edge represents a geodesic segment in @xmath10 which starts in the previous bounded region and then heads straight towards the end of @xmath10 associated with the attractive fixed point @xmath38 of @xmath30 .
moreover , the condition in ( [ transient ] ) guarantees that the hyperbolic length of that segment is at least equal to @xmath194 , and this shows that it s end point is separated by at least @xmath195 from the previous bounded region .
@xmath196 m. rees , checking ergodicity of some geodesic flows with infinite gibbs measure , _ ergod . th .
@xmath197 dynam .
* 1 * ( 1981 ) , 107133 .
m. rees , divergence type of some subgroups of finitely generated fuchsian groups , _ ergod .
@xmath197 dynam .
sys . _ * 1 * ( 1981 ) , 209221 .
b. o. stratmann , the exponent of convergence of kleinian groups ; on a theorem of bishop and jones , _ progr . in probab . _
* 57 * ( 2004 ) , 93107 . | the main result of this paper is to show that if @xmath0 is a normal subgroup of a kleinian group @xmath1 such that @xmath2 contains a coset which is represented by some loxodromic element , then the hausdorff dimension of the transient limit set of @xmath0 coincides with the hausdorff dimension of the limit set of @xmath1 .
this observation extends previous results by fernndez and melin for riemann surfaces . *
ams classification : * 30 f 40 , 37 f 99 , 37 f 30 , 28 a 80 . *
keywords : * kleinian groups ; poincar exponent ; fractal geometry ; dissipative dynamics . |
the critical behavior of two - dimensional frustrated spin models with noncollinear or canted order has been the object of intensive theoretical and experimental studies , being of prime interest for those investigating layered magnetic systems with special structures and superconducting josephson - junction arrays in an external magnetic field . in physical magnets , frustration ( leading to noncollinear order ) may arise either because of the special geometry of the lattice or from the competition of different kinds of interactions .
an example of the first kind is the two - dimensional triangular antiferromagnet , whereas for the latter case typical models are the fully frustrated @xmath5 model on a square lattice and models with different nearest - neighbor and next - nearest - neighbor interactions . in these systems
the hamiltonian is minimized by the noncollinear configurations with a 120@xmath6 spin structure . as a consequence , at criticality
, there is a breakdown of o(@xmath0 ) symmetry in the high - temperature phase to o(@xmath7 ) symmetry in the ordered phase .
field - theoretical ( ft ) studies of systems with noncollinear order are based on the o(@xmath0)@xmath8o(@xmath9 ) symmetric landau - ginzburg - wilson ( lgw ) hamiltonian @xcite @xmath10 + { 1\over 4!}u_0 \left ( \sum_a \phi_a^2\right)^2 \right . \left . + { 1\over 4 ! } v_0 \sum_{a , b } \left [ ( \phi_a \cdot \phi_b)^2 - \phi_a^2\phi_b^2\right ] \right\ } , \label{lgwh}\end{aligned}\ ] ] where @xmath11 ( @xmath12 ) are @xmath9 sets of @xmath0-component vectors .
we will consider the case @xmath13 , that , for @xmath14 , describes frustrated systems with noncollinear ordering .
negative values of @xmath15 correspond to simple ferromagnetic or antiferromagnetic ordering , and to magnets with sinusoidal spin structures @xcite .
the physically relevant cases are described by the hamiltonian ( [ lgwh ] ) with @xmath16 ( frustrated @xmath5 model ) and @xmath17 ( frustrated heisenberg model ) . despite the very intensive theoretical and experimental efforts to fully understand the critical behavior of these models , the situation is still controversial . for the frustrated @xmath5 model
the strongly debated point is either the single critical temperature @xmath18 exists at which both the @xmath19 and the @xmath20 symmetries are simultaneously broken or there are two successive phase transitions at different critical temperatures . in the latter case , the order at which two phase transitions occur and the numerical values of the critical exponents are doubtful too .
obviously , studying the hamiltonian ( [ lgwh ] ) , we can not clarify the point just mentioned , since this hamiltonian describes only the chiral phase transition . in more detail , these and other relevant issues are reviewed in ref .
@xcite , where one can find a complete list of references ; we mention here only the very recent studies @xcite ( not quoted in ref .
@xcite ) .
the critical behavior of the frustrated heisenberg antiferromagnet in two dimensions is clearer : it undergoes one phase transition mediated by @xmath20 topological defects .
again , a complete list of references can be found in ref .
@xcite . here
the debated point concerns whether the lgw hamiltonian is able to keep the non trivial topological excitations present in this model , to the contrary of other approaches like the nonlinear @xmath21 model ( @xmath22 ) . in ref .
@xcite it is claimed that it may be the case and we will assume it in the following .
the lgw hamiltonian ( [ lgwh ] ) has been extensively studied in the framework of @xmath23 @xcite , and by means of the @xmath3 @xcite expansion .
the existence and the stability properties of the fixed points were found to depend on @xmath0 and on the spatial dimensionality @xcite .
within the @xmath24 expansion , for sufficiently large values of @xmath0 the renormalization group equations possess four fixed points : the gaussian fixed point ( @xmath25 ) , the @xmath26 ( @xmath27 ) one and two anisotropic fixed points located in the region @xmath28 and usually named chiral and antichiral .
the chiral fixed point is the only stable one .
there is a critical dimensionality @xmath29 where the chiral and antichiral fixed points coalesce and disappear for @xmath30 . in the last case , under the absence of any stable fixed point the system undergoes a weak first - order phase transition , since the associated rg flows run away from the region of stability of the fourth - order form in the free energy expansion .
the three - loop estimate of @xmath29 is @xcite @xmath31 that , after an appropriate resummation , results in @xmath32 in three dimensions .
this inequality leads to the conclusion that for the physical models with @xmath33 the three - dimensional chiral transition is first order , as corroborated also by some other rg studies @xcite . to the contrary , both for @xmath16 and @xmath17 the highest - order six - loop calculations in three dimensions @xcite reveal a strong evidence for a stable chiral fixed point that , however , is not related to its counterpart found within the @xmath24 expansion .
the four - loop two - dimensional analysis shows an equivalent topology of the fixed points location in the renormalized coupling constants plane @xcite .
furthermore , in a recent work @xcite we claimed that both in two and three dimensions the controversial situation may reflect the quite unusual mode of critical behavior of the @xmath0-vector chiral model under the physical values of @xmath0 .
it was shown that this critical behavior is governed by a stable fixed point which is a focus , attracting rg trajectories in a spiral - like manner .
approaching the fixed point in a nonmonotonic way , looking somewhat irregular , may result in a large variety of the crossover and near - critical regimes .
in this paper , we report the perturbative five - loop two - dimensional rg expansions and carry out their analysis .
some of the results for @xmath1 and @xmath2 was already anticipated in @xcite .
we point out the idea that the critical behavior for large @xmath0 is the one predicted by the @xmath24 expansion . with decreasing @xmath0 ,
the stable chiral fixed point becomes focuslike at some marginal value of @xmath0 , @xmath34 that differs , in principle , from @xmath29 where both chiral and antichiral fixed points disappear . with the precision of our calculations , we are not able to fully clarify if there is a region of values of @xmath0 , where the chiral transition is first order .
it is worthy to note that the parallel work on the three - dimensional models @xcite reveals a full analogy between chiral critical behaviors in two and three dimensions .
the paper is organized as follows . in sec .
ii we derive the perturbative series for the renormalization - group functions up to five - loop order .
the resummations methods and the results of the analysis are presented in sec .
iii . in sec .
iv we draw our conclusions .
the fixed - dimension field - theoretical approach @xcite represents an effective procedure in the study of the critical properties of systems that undergo second - order phase transitions @xcite . in the case under consideration ,
the expansion is performed in two quartic coupling constants of the hamiltonian ( [ lgwh ] ) .
the theory is renormalized by a set of zero - momentum conditions for the one - particle irreducible two - point correlation function , four - point correlation function , and two - point correlation function with an insertion of the operator @xmath35 : @xmath36 , \label{ren1}\ ] ] @xmath37 , \label{ren2}\ ] ] @xmath38 where @xmath39 is an appropriate combinatorial factor @xcite . the perturbative knowledge of the functions @xmath40 , @xmath41 , and @xmath42 allows one to relate the renormalized physical parameters ( @xmath43 ) to the bare ones ( @xmath44 ) .
the fixed points of the model are defined by the common zeros of the @xmath45 functions , @xmath46 the stability properties of these points are determined by the eigenvalues @xmath47 of the matrix : @xmath48 a fixed point is stable if both the eigenvalues have a positive real part . if the eigenvalues possess nonvanishing imaginary parts , the fixed point is called a focus and the corresponding rg trajectories are spirals . the eigenvalues @xmath47
are connected to the leading scaling corrections , which go as @xmath49 , with few exceptions as the two - dimensional ising model ( see for a discussion ref .
@xcite ) .
the values of the critical exponents @xmath50 , @xmath51 and @xmath52 are related to the rg functions @xmath53 and @xmath54 evaluated at the stable fixed point : @xmath55 @xmath56 ^{-1 } , \label{nu_fromtheseries}\end{aligned}\ ] ] where @xmath57 @xmath58 in this section we extend the four - loop perturbative series @xcite for the rg functions @xmath59 , @xmath60 , @xmath61 , and @xmath54 up to five - loop order using the numerical values of two - dimensional five - loop integrals evaluated in ref . @xcite . in order to obtain finite fixed point values in the limit of infinite components of the order parameter ( @xmath62 ) we use the rescaled couplings @xmath63 where @xmath64 .
the resulting rg series are @xcite : @xmath65 where @xmath66 only the five - loop coefficients @xmath67 , @xmath68 , @xmath69 and @xmath70 are reported in table [ t3 ] , since up to the fourth order they are reported in ref.@xcite .
note that , due to the last rescaling , the element of the @xmath71 matrix ( [ omega ] ) are two times the derivative of @xmath72 with respect to @xmath73 and @xmath74 .
we verify the exactness of our series by checking them in particular limits ( @xmath75 , @xmath76 , @xmath77 , @xmath62 ) as in ref .
@xcite . .
five - loop coefficients @xmath67 , @xmath68 , @xmath69 and @xmath70 of rg functions . [ cols="^ , < " , ] [ tabellone ] to further approve the above conclusions , we apply , along with the conformal - mapping based , the resummation by means of the pad - borel technique . in the region where the fixed point is expected
, many pad approximants turn out to be defective , i.e. , possessing dangerous poles .
we find that the approximants @xmath78 $ ] , @xmath79 $ ] , and @xmath80 $ ] are nondefective in the majority of the cases .
considering these three pad approximants with @xmath81 ( being the more stable under variation of the number of loops ) , we finally have nine approximants for each @xmath45 function .
the location of the chiral fixed point is reported in table [ tabellone ] , with all the defective cases discarded .
we also compute the exponents @xmath50 and @xmath54 with both the resummation techniques .
we consider a lot of approximants and finally we choose the more stable ones with varying the number of loops .
the found values are close to those of the @xmath82 model for @xmath83 , but differ from the exact ones coming from @xmath84 model @xmath85 and @xmath86 ( e.g. , we find @xmath87 for @xmath88 and @xmath89 for @xmath90 ) .
we think these discrepancies are due to nonanalyticities , as in the case of @xmath82 models . to conclude the analysis of the fixed points , we study the stability of the heisenberg fixed point by calculating the crossover exponent @xmath91 .
the five - loop results show that this fixed point is unstable , in agreement with the conclusion drawn earlier from the four - loop series @xcite .
in this subsection we demonstrate the structure of the rg flow for physical values of @xmath0 .
as it was already shown , a stable focus governs the critical behavior of these systems . in fig .
[ fign2d2 ] ( @xmath16 ) one can see the spiral - like approach of the stable fixed point which is peculiar for the fixed points with stability eigenvalues possessing nonvanishing imaginary parts .
the shaded area is the region where the resummation procedure is expected to fail ( i.e. , @xmath92 ) , since we do not take into account the singularity of the borel transform closest to the origin that is on the real positive axis ( cf .
eq ( [ bsing ] ) ) . consequently ,
the rg trajectories given by the resummed perturbative series are quantitatively correct only within the unshaded areas .
we report , nevertheless , the rg flows in other parts of the coupling constant plane in order to present a complete qualitative picture .
note , that the domain usually referred to as the sector of the first - order transitions , where the positiveness of the quartic form in the free - energy expansion breaks down , is given by the inequality @xmath93 .
in fact , however , because of the presence of the higher - order terms in this expansion keeping the system globally stable at any temperature , the true domains of the first - order transitions should be more narrow . from fig .
[ fign2d2 ] we see that for some initial values of @xmath94 the rg trajectories have the coordinate @xmath74 that grows very fast at the beginning and seems to reach the region of the first - order phase transitions , but just before arriving there these trajectories drastically curve moving toward the stable chiral fixed point .
these bizarre trajectories can imitate both the discontinuous phase transitions and strongly scattered effective critical exponents observed in numerous physical and computer experiments . to substantiate the former scenario ,
we calculate the numerical values which the effective exponent @xmath95 takes along the rg trajectories running from the gaussian fixed point to the chiral one .
corresponding curves are shown in fig .
as is seen from this figure , the exponent oscillates within a large range , even near the stable fixed point .
we demonstrate also the rg flows for @xmath17 ( fig .
[ fign3d2 ] ) and for @xmath96 ( fig .
[ f8 ] ) . in the former ( physical case ) the stable fixed point is a focus and phase trajectories approach it in a spiral - like manner . in the latter ( unphysical case )
the stable fixed point turns out to be a node and the trajectories are certainly not spirals .
also for @xmath17 we estimate the effective exponent @xmath95 along the rg trajectories running from the gaussian fixed point to the chiral one .
the corresponding curves ( cf . fig .
[ crn3 ] ) show big oscillations and sometimes reach negative values .
we note that these negative values are obtained only in the region where the resummation is not sure , i.e. , @xmath92 .
nevertheless , this strange crossover , if reproduced from the three - dimensional series @xcite , might explain some doubtful experimental results , where a negative @xmath50 was found . to conclude , it is worthy to note that earlier the focuslike stable fixed points were found on the rg flow diagrams of the model describing critical behavior of liquid crystals @xcite and of the @xmath97-symmetric systems undergoing first - order phase transition close to the tricritical point @xcite . in those cases , however ,
the independent coupling constants had different scaling dimensionality and played essentially different roles in critical thermodynamics .
moreover , recently complex subleading exponents were found from the high - temperature expansion of dyson s hierarchical ising model @xcite .
we studied the critical thermodynamics of the two - dimensional @xmath0-vector chiral model in the five - loop rg approximation . using the advanced resummation technique based upon borel transformation combined with conformal mapping and use of pad approximants
, it was shown that the fixed point governing the chiral critical behavior is a stable node only for large enough @xmath0 , @xmath98 . for smaller @xmath0 a chiral fixed point
turns out be a stable focus and the rg trajectories approach this point in a spiral - like manner
. of particular importance is the fact that a focus - driven critical behavior was found for systems with physical values of @xmath0 , @xmath99 . evaluating the critical exponent @xmath50 along some rg trajectories , we demonstrate that spiral - like approach of the chiral fixed point results in unusual crossover and near - critical regimes that may imitate varying critical exponents seen in physical and computer experiments .
we would like to thank ettore vicari for many useful discussions .
the financial support of the russian foundation for basic research under grant no .
01 - 02 - 17048 ( e.v.o . , a.i.s . ) and the ministry of education of russian federation under grant no .
e00 - 3.2 - 132 ( e.v.o . , a.i.s . )
is gratefully acknowledged .
a.i.s . has benefited from the warm hospitality of scuola normale superiore and dipartimento di fisica delluniversit di pisa , where the major part of this research was done .
r. melzi , p. carretta , a. lascialfari , m. mambrini , m. troyer , p. millet , and f. mila , phys .
lett . * 85 * , 1318 ( 2000 ) ; r. melzi , s. aldorovandi , f. tedoldi , p. carretta , p. millet , and f. mila , phys . rev .
b * 64 * , 024409 ( 2001 ) ; p. carretta , r. melzi , n. papinutto , and p. millet , phys .
lett . * 88 * , 047601 ( 2002 ) .
a. pelissetto , p. rossi , and e. vicari , phys .
b * 63 * 140414(r ) ( 2001 ) , _ ibid . _
* 65 * 020409(r ) ( 2002 ) .
p. calabrese , p. parruccini , and a. i. sokolov , e - print cond - mat/0205046 , phys .
b to appear .
p. calabrese , p. parruccini , and a. i. sokolov , in preparation .
a. pelissetto and e. vicari , nucl .
b * 519 * , 626 ( 1998 ) .
t. dombre and n. read , phys .
b * 39 * , 6797 ( 1989 ) ; p. azaria , b. delamotte , and t. jolicoeur , phys . rev . lett . * 64 * , 3175 ( 1990 ) ; h. kawamura , j. phys .
jpn . * 60 * , 1839 ( 1991 ) ; p. azaria , b. delamotte , f. delduc , and t. jolicoeur , nucl .
b * 408 * , 485 ( 1993 ) . | we analyze the critical behavior of two - dimensional @xmath0-vector spin systems with noncollinear order within the five - loop renormalization - group approximation .
the structure of the rg flow is studied for different @xmath0 leading to the conclusion that the chiral fixed point governing the critical behavior of physical systems with @xmath1 and @xmath2 does not coincide with that given by the @xmath3 expansion .
we show that the stable chiral fixed point for @xmath4 , including @xmath1 and @xmath2 , turns out to be a focus .
we give a complete characterization of the critical behavior controlled by this fixed point , also evaluating the subleading crossover exponents .
the spiral - like approach of the chiral fixed point is argued to give rise to unusual crossover and near - critical regimes that may imitate varying critical exponents seen in numerous physical and computer experiments .
-0.4 in 0.4 in 9.0 in |
the dating of the oldest stars in the milky way allows us to infer the early history of star formation in our galaxy , and hence , provides us with important information regarding the formation of the milky way .
a great deal of attention has been paid to the ages of globular clusters ( see for example the reviews by @xcite ) in the halo and there have been a number of studies which determine the ages of the oldest open clusters in the thin disk ( e.g. @xcite ; ( * ? ? ?
* hereafter cgl ) ) . typically , there appears to be a gap of several gyr between the ages of the globular clusters and the ages determined for the oldest stars in the thin disk .
we have elected to investigate this further , in order to quantify the age difference ( if any ) between the thin disk , the thick disk and the halo of the milky way . in recent years , numerous techniques have been employed to determine the ages of the oldest stars in the galactic ( thin ) disk .
@xcite examined the color magnitude diagram ( cmd ) of hipparcos field stars and arrived at a minimum disk age of 8 gyr ; @xcite used isochrone fits to determine the ages of several old open clusters . @xcite and @xcite used observations of local white dwarfs and theoretical white dwarf cooling curves to determine a local disk age of @xmath4 gyr and @xmath5 gyr respectively .
@xcite also used observations of white dwarfs to place a minimum age of 9.5 gyr on the galactic disk .
cgl studied the old open cluster ngc 6791 and determined an age of @xmath6 gyr .
globular clusters in the thick disk and halo have been the subject of a number of recent investigations .
for example , @xcite determined the ages of three thick disk globular clusters , 47 tuc , m71 , and ngc 6352 , and arrived at an age of 9.2 gyr for all three .
@xcite examined two halo clusters , ngc 2419 and m92 , in order to determine their relative and absolute ages .
their study determined that the clusters had the same relative age to within 1 gyr , and an absolute age of @xmath7 gyr . in this paper , the relative age of the thin and thick disks
is measured , comparing the open clusters ngc 188 , ngc 6791 and selected hipparcos field stars to the thick disk globular cluster 47 tuc .
the results of this analysis are related to the results of ( * ? ? ?
* hereafter csa ) , in which the relative age of 47 tuc to three halo clusters was determined using the same input physics as this study . compared to other studies which have investigated the relative age of the thin disk and the halo ( * ? ? ?
* eg . ) this study has the advantage of using the same assumptions and methodologies in dating the thin and thick disks and the halo .
this allows us to determine precise relative ages and to provide a reliable estimate of the error in the relative age estimates .
section [ secmodel ] of this paper will discuss the stellar model and isochrones used in this study .
section [ thindisk ] details the age estimates for ngc 188 and the hipparcos field stars ( thin disk ) .
the age of 47 tuc ( thick disk ) is determined in [ sec47tuc ] .
the results are summarized in [ secsumm ] where the relative age between the the thin and thick disks and the halo is discussed .
stellar evolution tracks were constructed using chaboyer s stellar evolution code ( cgl ) for a range of masses of @xmath8 in increments of 0.05 . the input physics , including low and high temperature opacities , nuclear reaction rates , helium diffusion coefficients and the equation of state were identical to those described in csa .
the models were evolved in 6000 time steps from the zero age main sequence through the red giant branch .
the values of @xmath1$}}$ ] for each of set of stellar models were chosen to reflect the observed values in each of the clusters . to determine the uncertainty in the age due to the uncertainty in
, stellar models were calculated for a range of values , centered upon the preferred value for each of the clusters .
most of the models used a scaled solar heavy element composition @xcite .
the models for 47 tuc were calculated with @xmath9 = 0.0 $ ] ( scaled solar ) and @xmath9 = + 0.40 $ ] .
the calibrated solar model for this study had a helium mass fraction of @xmath10 and a heavy element mass fraction of @xmath11 .
the primordial helium abundance was taken to be @xmath12 @xcite , and the helium abundance for each model was calculated using the relation @xmath13 with @xmath14 .
the isochrones were constructed for ages between 5 and 14 gyr , in 1 gyr increments for each of the compositions .
color transformations to the observational plane were done in a manner identical to that described in cgl .
figure [ figalpha ] compares the 10 gyr isochrone for @xmath1$}}= -0.70 $ ] for both the @xmath15-enhanced and the non @xmath15-enhanced cases .
the resulting change in the isochrone corresponds to a change in @xmath1$}}$ ] of nearly 0.20 dex .
these results are similar to those found in @xcite .
additional isochrones were generated in order to ascertain the effect that changes in the low temperature opacities would have on the isochrones .
figure [ figalex ] plots the 10 gyr isochrone for @xmath1$}}= -0.70 $ ] with @xcite opacities , as well as the 9 , 10 and 11 gyr isochrones with @xcite opacities used in this study .
this figure demonstrates that ages determined using isochrones generated with kurucz opacities will indicate ages about 0.5 gyr older than isochrones using alexander & ferguson opacities .
this analysis was done in order to compare the age of ngc 6791 as found in cgl , which used alexander & ferguson opacities in its models , with the ages of the clusters in this study and csa , which utilized kurucz opacities .
kurucz ( 1991 ) opacities were used in this study and by csa as these opacities are available for both scaled solar and @xmath15-element enhanced compositions .
the alexander & ferguson opacities were only available to us for scaled solar compositions .
isochrones were fit to the cmd for ngc 188 obtained from observations by ( * ? ? ? * hereafter s99 ) .
the cmd was cross - correlated to include only cluster members with a greater than 75 percent membership probability @xcite .
membership criteria was based upon a proper motion study by @xcite .
ngc 188 has @xmath1$}}= -0.05 $ ] based upon several determinations @xcite .
isochrones with @xmath1$}}= -0.05 $ ] were fit simultaneously in and using the main sequence , allowing the distance modulus to vary around the value established by s99 .
figure [ fign188 ] shows the 5 , 6 , and 7 gyr isochrones fit to the observational data .
the reddening was taken to be @xmath16 ( s99 ) and transformed to @xmath17 by @xmath18 .
the distance modulus which yielded an excellent main sequence fit was @xmath19 , well within the error bars for the value determined by s99 of @xmath20 .
the age indicated by both and fits was 6.5 gyr .
isochrones for a more metal poor composition of @xmath1$}}= -0.15 $ ] were fit to the cmd , again assuming a reddening of @xmath21 .
the fits produced a distance modulus of @xmath22 , slightly less than s99 , and an age of 7.5 gyr .
figure [ fign188lml ] shows the 6 , 7 , and 8 gyr isochrones for this lower metallicity fit to the observational data .
isochrones with a higher metallicity , @xmath1$}}= + 0.05 $ ] , were fit to the data as well and are shown in figure [ fign188uml ] , indicating @xmath23 and an age of 5.5 gyr . however , it would appear unlikely that a younger age as a result of higher metallicity is the case , as the cmd does not show a hook at the main sequence turn off which is seen in the isochrones . in order to determine the uncertainty due to reddening , the @xmath1$}}= -0.05 $ ] isochrones were fit for @xmath24 and @xmath25 .
these fits indicate ages of 7.5 gyr and 5.5 gyr , respectively .
the three sets of isochrone fits for ngc 188 are summarized in table [ tabn188 ] .
the best age estimate was found to be @xmath26 gyr , allowing for uncertainty in the metallicity and reddening of the cluster .
this error estimate does not include the uncertainty in the input physics in the stellar models and isochrones , and so should be viewed as the error on the relative age of ngc 188 .
berkeley 17 ( be 17 ) has been suggested to be the oldest open cluster @xcite .
it has a metallicity of @xmath1$}}= -0.29\pm 0.13 $ ] from moderate resolution spectroscopy @xcite , while @xcite estimate @xmath1$}}\sim -0.35 $ ] based upon the slope of the red giant branch in the infrared .
@xcite obtained bvi photometry of this cluster which was used in our isochrone fits .
a simultaneous fit to the and photometry was attempted , assuming @xmath27 .
we attempted to fit isochrones with @xmath1$}}= -0.10 , -0.24,$ ] and @xmath28 with @xmath9 = + 0.0 $ ] and @xmath29 ( a total of 6 different compositions ) without success . in all cases ,
the be 17 main sequence was redder than our isochrones in and bluer than our isochrones in .
our best attempt at a fit is shown in figure [ figbe17 ] .
as this is clearly not an acceptable fit to the data , we are unable to assign an age to be 17 .
there are a number of possible explanations for the inability of the isochrones to simultaneously fit the and data .
these include : ( 1 ) a non - standard extinction law in the direction of be 17 ( ie .
@xmath30 ) , ( 2 ) a helium abundance significantly different than in our models , ( 3 ) errors in our isochrones ( 4 ) errors in the photometry ( of order 0.05 mag in the color ) .
open clusters are expected to dissipate / disrupt in the galactic plane .
thus , only the most massive open clusters or those with orbits which keep them far away from the plane for most of their lifetimes are expected to survive for signficant periods of time .
for this reason , the age of the oldest field stars in the thin disk may provide a better estimate for the age of the thin disk than open clusters .
an analysis of the age of oldest thin disk stars in the solar neighborhood was made by comparing solar metallicity isochrones to a selected sample of field stars from the hipparcos database @xcite .
the hipparcos database was searched for single , non - variable stars with good parallaxes ( @xmath31 ) .
these stars were cross referenced with the 1996 catalogue @xcite to find stars with @xmath1$}}$ ] within 0.07 dex of solar . choosing stars with metallicities near the solar value
should ensure that these stars are members of the thin disk .
in addition , choosing a relatively narrow metallicity range makes the age determination process simple .
the selected stars were then plotted on a cmd and compared to solar metallicity isochrones .
as many of the stars do not have i band photometry , the comparison to the isochrones was only done in .
we are interested in determining the onset of star formation in the thin disk , and so concentrated our attention on the oldest stars in our sample for which it is possible to determine ages .
these are the faintest stars in the main sequence turn - off or subgiant branch regions of the cmd .
there were 21 stars which appeared to be old in our search of the hipparcos database .
these 21 stars were checked for other published metallicities and space velocities .
some stars were omitted from the sample which were shown in other sources to have metallicities of more than 0.07 dex from solar @xcite or shown to have space velocities which are not characteristic of the thin disk @xcite .
the resulting sample of stars are shown in figure [ fighipp ] along with our zero - age main sequence and 6 8 gyr isochrones ( all with @xmath1$}}= 0.0 $ ] ) .
figure [ fighipp]b reveals that there are 3 stars which are much brighter / redder than our main sequence isochrones .
there a few possible reasons for this discrepency , including that these stars are binaries , or are more metal - rich than indicated by the 1996 catalogue
. we will not determine the age of these three stars .
an inspection of the oldest stars in the main sequence turn - off and subgiant branch region in figure [ fighipp]b reveals that 2 stars lie on the 8 gyr isochrone , 4 stars lie on or very near the 7 gyr isochrone and 1 star lies between the 7 and 8 gyr isochrone .
the derived ages for the stars in the turn - off region are quite sensitive to possible errors in the colors of the stars .
for example , if the color of the 8 gyr star with @xmath32 was changed from and @xmath33 to @xmath34 then its age would be changed to 7 gyr .
given the uncertainties in the photometry , reddening , and in our color transformation we are reluctant to base our estimate for the age of the oldest stars in the thin disk on just two stars .
for this reason , we estimate that the oldest stars are @xmath35 gyr old .
isochrones for @xmath1$}}= -0.05 $ ] and + 0.05 yielded ages of 8.0 gyr and 7.0 gyr respectively and we conclude that the oldest solar metallicity stars in the thin disk have an age of @xmath36 gyr . in determining this error
, we simply added in quadrate the @xmath37 gyr error due to the uncertainty in the metallicity of the stars with the @xmath37 gyr uncertainty due to the photometry / reddening and color transformation .
there is a range in metallicities in the thin disk and @xcite found that ( in the local solar neighborhood ) the thin disk has metallicities ranging down to @xmath1$}}\simeq -0.20 $ ] .
for this reason , we elected to search the hipparcos database for single , non - variable stars with good parallaxes ( @xmath31 ) with @xmath38$}}\leq -0.25 $ ] ( values from the 1996 catalogue @xcite ) .
this sample of stars allows us to determine the age of the somewhat metal - poor stars in the thin disk . in order to focus on the oldest stars ,
color and magnitude cuts were made in order to select stars with @xmath39 and @xmath40 .
these values were chosen based upon an inspection of our 7 gyr and older isochrones with @xmath1$}}= -0.10 $ ] and @xmath1$}}= -0.24 $ ] .
these cuts were made rather generous to ensure that all potentially older stars are included in the sample .
these cuts resulted in a sample of 21 stars which were plotted on a cmd with isochrones of appropriate metallicities .
stars which were less than @xmath41 gyr old were deleted from further study .
this left a sample of 16 stars .
basic data for each of these 16 stars was retrieved from the simbad database at cds .
four stars which were classified as variables or spectrospic binaries were removed from list as it would not be possible to determine their ages .
the derived ages are quite sensitive to the metallicity of the star .
for this reason , the references were checked for each star and 3 stars whose recent ( post 1995 ) determinations were significantly different than those listed in the 1996 catalogue were removed from the list .
the basic data for each of these stars are given in table [ tabfield ] .
the parallaxes , proper motions and colors are from the hippacos catalog .
the values are from @xcite , and the radial velocities were obtained from the simbad database . the age of each of these nine stars was determined from our isochrones ( interpolating between the @xmath1$}}= -0.10 $ ] and @xmath1$}}= -0.24 $ ] isochrones ) .
the error in the derived age of each star is due to the error in the absolute magnitude ( due to the error in the parallax ) , an assumed color error of @xmath42 in and an assumed error in of @xmath43 dex .
the ages of these stars are given in table [ tabfield2 ] , which also includes some derived kinematic data and orbital parameters for each star . the kinematic data and orbital parameters allow us to separate thin and thick disk stars .
the space motion of each star was calculated using a routine kindly provided to us by dr .
dana dinescu .
the motions were calculated in a cylindrical coordinate system ( origin at the galactic center ) by adopting a solar radius of @xmath44 kpc and a rotation velocity of the local standard of rest ( lsr ) of @xmath45 km / s . in this cooridinate system ,
the @xmath46 component is positive outward from the galactic center , @xmath47 is positive in the direction of galactic rotation and @xmath48 is positive towards the north galactic pole .
errors in the derived velocities include errors in the proper motions , radial velocities and distances .
table [ tabfield2 ] includes some basic orbital parameters based upon intergation of the orbits in two models of the galaxy s potential .
these integrations were made by dr .
dana dinescu , and full details of the integration routine may be found in @xcite .
the orbits were integrated in galactic potential models given by @xcite and @xcite .
the two potentials yield similar orbital parameters and only the orbital parameters from the @xcite potential are shown in table [ tabfield2 ] . the orbital parameters listed in table [ tabfield2 ]
are : @xmath49 the @xmath50-component of the angular momentum ( a conserved quantity ) , pericentric ( @xmath51 ) and apocentric ( @xmath52 ) radii , the maximum distance from the plane @xmath53 , the eccentricity of the orbits @xmath54 and the inclination angle with respet to the galactic plane @xmath55 . the stars in table [ tabfield2 ] all have @xmath56$}}\le -0.25 $ ] and in this metallicity range one finds both thin and thick disk stars in the local solar neighborhood .
the dispersion in the kinematics of the thin and thick disk make it impossible to definitavily assign a single star to either the thin or thick disk .
for example , in the thin disk @xmath57 km / s while in the thick disk @xmath58 km / s @xcite .
thus , the @xmath48 velocity of hd 41330 ( @xmath59 km / s ) implies that it could be member of the thick or thin disk . by considering all of the kinematic and orbital parameters ( @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath54 and @xmath55 ) one can classify a star as either a thin or thick disk star , bearing in mind that these classifications will never be 100% accurate .
in general , thin disk stars have small @xmath46 and @xmath48 velocities , @xmath60 km / s , @xmath61 kpc km / s , @xmath62 kpc , @xmath63 and small values of @xmath55 . in table
[ tabfield2 ] we have indicated the most propobable classificiation of each star ( thin or thick disk ) _ based on the kinematic and orbital parameters alone_. hd 52711 and 210918 are prototypical thick disk stars with rotation ( @xmath47 ) velocities significantly different than the lsr , a low @xmath50-component to their orbital angular momentum ( @xmath49 ) and a high eccentricity . in contrast , hd 207129 is a prototypical thin disk star with a rotation velocity and @xmath49 similar to the lsr , a low eccentricity , a small @xmath53 and a small angle to the galactic plane . three stars which are likely thin disk stars ( hd 15335 , 202628 and 207129 ) have similar ages of @xmath64 gyr .
the oldest stars which appear to be thin disk stars are hd 32923 ( @xmath65 gyr ) and hd 41330 ( @xmath66 gyr ) . both of these stars have rotation velocities and @xmath49 similar to the lsr and low eccentricities .
the @xmath53 of these stars is not too large ( 0.46 and 0.29 kpc ) given that thin disk stars have @xmath67 kpc scaleheight exponentials @xcite .
the angle of their orbits to the galactic plane ( @xmath68 and @xmath69 ) are somewhat larger than typical for thin disk stars but not extremely so . as a result
, we believe that it is likely that at least one of these two stars is a true member of the thin disk .
these stars are located in a region of the cmd where the derived ages are relatively insensitive to the metallicities , colors and absolute magnitudes . consequently , the derived ages have very small error bars . averaging the ages of these two stars together
we find that the oldest , somewhat metal - poor thin disk stars in the solar neighborhood have an age of @xmath2 gyr .
the two stars with kinematic and orbital parameters most representive of the thick disk ( hd 52711 and 210918 ) have quite different ages of @xmath70 gyr and @xmath71 gyr .
the older age is similar to 47 tuc ( see below ) while the younger age would suggest that there is a considerable overlap in the ages of the thin and thick disks .
however , we are reluctant to reach such a conclusion based only on one star . two other stars which might be thick disk stars ( hd 11007 and 67458 ) are also fairly young ( @xmath72 and @xmath73 ) . however , one could argue that the kinematics and orbital parameters of these two stars are not too different from the thin disk , and so there identification as thick disk stars is debateable
. further age determinations of more metal - poor stars ( @xmath1 $ } } < -0.25 $ ] ) are needed before one can conclude that their is a significant spread in the age of the thick disk .
isochrone fitting ages were determined for the thick disk globular cluster 47 tuc .
the photometric data for the cluster was obtained from @xcite .
heavy element abundances in the literature indicate values of @xmath1$}}= -0.70\pm0.07 $ ] @xcite , and @xmath1$}}= -0.81 $ ] @xcite .
additionally , @xcite indicate an enhancement in @xmath15-capture elements of @xmath74$}}= 0.22 $ ] .
however , for the stellar models used in this study , opacities were only available for @xmath74$}}= 0.00 $ ] and @xmath75 .
thus isochrones were fit for abundances of @xmath1$}}= -0.70 $ ] and @xmath76 both with @xmath74$}}= 0.40 $ ] .
isochrones were fit in a manner identical to ngc 188 , simutaneously in b - v and v - i .
the value of reddening was fixed at @xmath77 @xcite and @xmath78 .
the distance modulus was varied around @xmath79 @xcite in order to obtain a good main sequence fit .
figure [ fig47tuc-070 ] shows the isochrone fits for @xmath1$}}= -0.70 $ ] and @xmath74$}}= 0.40 $ ] .
the poor simultaneous fit is likely due to the difference in @xmath15-enhancement by nearly 0.2 dex between the literature value and the models .
the fits for @xmath1$}}= -0.80 $ ] and @xmath74$}}= 0.40 $ ] , shown in figure [ fig47tuc-080 ] , likewise indicate a poor simultaneous fit to b - v and v - i . in order to better reflect the total heavy element abundance ( @xmath80 ) in the cluster ,
isochrones were generated with @xmath1$}}= -0.95 $ ] and @xmath74$}}= 0.40 $ ] .
this abundance was chosen to match the heavy element mass fraction for the values of and found in @xcite , corresponding to @xmath81 .
the result , shown in figure [ fig47tuc-095 ] , was an acceptable fit which indicated an age of @xmath82 gyr for a reddening of @xmath77 and distance modulus of @xmath83 .
both the reddening and distance modulus were in good agreement with the values quoted in @xcite .
uncertainty due to reddening was determined by producing fits with @xmath84 and @xmath85 .
these fits indicated ages of @xmath86 and @xmath87 gyr , respectively .
table [ tab47tuc ] summarizes the isochrone sets and parameters used in fitting to 47 tuc .
the age used for comparison to the other clusters and field stars was @xmath88 gyr , as indicated and includes the uncertainty due to reddening .
age determinations were also made using the @xmath89 method described in csa .
the @xmath89 determination method is theoretically robust , but requires a well defined horizontal branch not present in the cmds of open clusters . in order to compare the @xmath89 ages in csa with the isochrone fitted ages in this paper , the age of 47 tuc
was determined using both methods , and all ages were calculated relative to 47 tuc .
the observed value for @xmath89 in the cmd for 47 tuc was 3.18 @xmath00.04 mag .
the theoretical value of @xmath90 was calculated using @xmath91$}}+ 1.6 ) + 0.56 $ ] ( csa ) and corrected for the fact that the hb was only apparent redward of the rr lyrae instability strip in 47 tuc .
this correction was determined using the theoretical hb models of @xcite . using the @xmath1$}}= -0.95 $ ] ,
@xmath9 = + 0.40 $ ] isochrones yielded an age of @xmath92 gyr .
the @xmath1$}}= -0.70 $ ] , @xmath9 = + 0.40 $ ] isochrones yielded an age of 12.2 gyr .
47 tuc has @xmath9 = + 0.22 $ ] @xcite and @xmath1$}}= -0.70\pm0.07 $ ] @xcite which gives a @xmath80 value intermediate between the above 2 isochrones sets . for this reason , we average the 2 ages and adopt an age for 47 tuc of @xmath93 gyr
. the age determined from the is @xmath94 gyr older than that determined from the isochrone fits .
this small difference in age is well within the estimated error in our isochrone fit age .
as the age estimate is expected to yield more accurate absolute ages than isochrone fitting , we will assume an absolute age of @xmath95 gyr for 47 tuc .
however , when determining the relative age of 47 tuc to the old thin disk , we will use the isochrone fitting age of 47 tuc , as the open clusters and hipparcos field stars had their ages determined using isochrone fitting .
the age of 47 tuc ( @xmath96 ) gyr provides a reliable estimate for the age of the thick disk .
the results for the various relative stellar ages determined in this study are summarized in table [ tabresults ] .
this table includes the age for ngc 6791 from cgl and the ages of three inner halo globular clusters ( ngc 6652 , ngc 1851 and m107 ) investigated by csa .
the results from csa indicate that star formation in the inner halo began roughly at the same time as the thick disk .
all three of the studies discussed here ( csa , cgl , and this study ) used the same input physics and distance scale , allowing for a relative age comparison between all clusters in the studies .
the relative age between the thin disk stars and 47 tuc in this study and cgl were determined using the isochrone fitting ages .
relative ages between 47 tuc and the inner halo were determined using the @xmath89 method . the globular cluster 47 tuc was used as a `` bridge '' between the studies , as its age
was determined using both the isochrone fitting and @xmath89 methods .
the relative ages determined in this paper indicate that ngc 188 was formed @xmath97 gyr after 47 tuc .
the oldest solar metallicity hipparcos field stars were formed @xmath98 gyr after 47 tuc .
both of these results suggest that the solar metallicity stars in the thin disk were formed signifiantly later than the thick disk . simply averaging these results suggest that solar metallicity stars ( in the solar neighborhood ) started to form @xmath99 gyr after the thick disk .
in contrast , the slightly metal - poor ( @xmath1$}}\approx -0.22 $ ] ) solar neighborhood stars were formed @xmath3 gyr after 47 tuc .
as this difference in age is less than @xmath100 it is conceivable that thin disk stars were being born in the solar neighborhood at the same time , or shortly after thick disk objects like 47 tuc were forming .
the small relative ages between the thick disk and inner halo ( table [ tabresults ] ) provide evidence that star formation in these regions began at roughly the same time .
figure [ figfehage ] shows the relative ages versus metallicity of all the clusters discussed in this study .
it is interesting to note that the very metal - rich open cluster ngc 6791 is @xmath101 gyr younger than 47 tuc .
this old open cluster has an orbit which makes its classification as a thin disk object somewhat problematic .
in particular @xcite report that this cluster has a rotation velocity @xmath47 which is 60 km / s lower than the thin disk and it is moving radially outward with @xmath102 km / s resulting in a orbit with a large eccentricity .
however , this work assumed a considerably larger distance than was found by cgl ( 5.3 vs.4.2 kpc ) .
for this reason , we have calculated the kinematic and orbital parameters of ngc 6791 using the distance found by cgl ( 4.2 kpc ) , the radial velocity from @xcite ( @xmath103 km / s ) and the preliminary proper motions from @xcite .
we find the following : @xmath104 km / s ; @xmath105 km / s ; @xmath106 km / s ; @xmath107 kpc km / s ; @xmath108 kpc ; @xmath109 kpc ; @xmath110 kpc ; @xmath111 and @xmath112 . the orbit is fairly eccentric with a low rotation velocity and a reasonably large @xmath53 suggesting that ngc 6791 is a thick disk object .
however , @xcite caution that clusters like ngc 6791 may simply be in the wings of the kinematic distribution of the thin disk , and its peculiar orbit is what has allowed it to survive .
table [ tabresults ] includes the median distance of the stars / clusters from the galacic center ( @xmath113 ) based upon the orbits of the stars / clusters .
we see that all of the thick disk and thin disk objects we have discussed in this paper cover a relatively modest range in their distance from the galactic center ( @xmath114 ) .
thus , all of these object were born at reasonably similar galacto - centric distances at so their ages reflect the star formation in small region of the galaxy .
the 3 inner halo clusters discussed by csa turn out to have a fairly wide range in @xmath113 ( 2 to 18 kpc ) .
the fact that the 47 tuc has a similar age to these clusters suggests that globular cluster formation was occuring at about the same time over a wide range of galacto - centric distances there have been many proposed scenarios of galactic formation .
these scenarios include the free - fall collapse of a protogalactic cloud , originally proposed by @xcite , and the merger of independently formed stellar systems @xcite .
although no definative conclusions about galactic formation can be drawn from the results of this study , some possibilities can be presented .
the similarity in age between the inner halo and thick disk indicates that if the thick disk was formed from the gravitational collapse of the halo , this collapse occurred on short timescales ( @xmath115 gyr ) .
there is a clear age gap of several gyr between solar metallicity stars in the nearby thin disk and the thick disk ( 47 tuc ) .
the age gap between moderately metal - poor thin disk stars and the thick disk is much smaller , and the present uncertainties in the age determinations allow for the possibility that the thin disk formed immediately after ( @xmath116 gyr ) after the thick disk and halo .
we are extremely gratefull to dana dinescu for providing us with the routines to calculate space velocities of the field stars , for calculating the orbits of the stars and for cross - referencing her membership list for ngc 188 with the photometry from sarajedini et al .
we would like to think the anonymous referee and the editor jim liebert whose numerous suggestions led to a considerable improvement in this paper .
this research was supported by nasa through the nasa / new hampshire space grant , an ltsa award nag5 - 9225 ( bc ) , by a burke research grant from dartmouth college ( bc ) and used the simbad database , operated by operated at cds , strasbourg , france . bc acknowledges the hospitality of the aspen center of physics , where he revised the paper .
alexander , d. r. & ferguson , j. w. 1994 , , 437 , 879 bergeron , p. , ruiz , m. t. & leggett , s. k. 1997 , , 108 , 339 brown , j. a. & wallerstein , g. 1992 , , 104 , 1818 carraro , g. & chiosi , c. 1994 , , 288 , 751 carraro , g. , girardi , l. & chiosi , c. 1999 , , 309 , 430 carraro , g. , vallenari , a. , girardi , l. & richichi , a. 1999 , , 343 , 825 carretta , e. & gratton , r. g. 1997 , , 121 , 95 cayrel de strobel , g. , soubiran , c. , friel , e. d. , ralite , n. & francois , p. 1997
, , 124 , 299 chaboyer , b. , sarajedini , a. , & armandroff , t. e. 1999 , aj , submitted ( csa ) chaboyer , b. , green , e. m. & liebert , j. 1999 , , 117 , 1360 cudworth , k.m .
1994 , , 25 , 1454 demarque , p. , zinn , r. , lee , y. & yi , s. 2000 , , 119 , 1398 dinescu , d. i. private communication dinescu , d. i. , girard , t. m. & van altena , w. f. 1999 , , 117 , 1792 dinescu , d. i. , girard , t. m. , van altena , w. f. , yang , t. & lee , y. 1996 , , 111 , 1205 edvardsson , b. , andersen , j. , gustafsson , b. , lambert , d. l. , nissen , p. e. & tomkin , j. 1993 , , 275 , 101 eggen , o. j. 1998 , , 115 , 2397 eggen , o. j. , lynden - bell , d. & sandage , a. r. 1962 , , 136 , 748 the hipparcos and tycho catalogs , esa sp-1200 friel , e.d . , janes , k.a . , hong , l. , lotz j. , tavarez , m. 1995 , in the formation of the milky way , eds .
alfaro , e.j .
delgado ( cambridge : cambridge univ .
press ) , 189 grevesse , n. , & noels , a.1993 , in origin and evolution of the elements , ed .
n. prantzos , e.vangioni-flam & m. casse ( cambridge : cambridge univ .
press ) , 15 harris , w. e. , et al .
1997 , , 114 , 1030 harris , w. e. 1996 , , 112 , 1487 hobbs , l. m. , thorburn , j. a. & rodriguez - bell , t. 1990 , , 100 , 710 jimenez , r. , flynn , c. & kotoneva , e. 1998 , , 299 , 515 johnston , k.v . ,
spergel , d.n . & hernquist , l. 1995 , , 465 , 278 kaluzny , j. , wysocka , a. , stanek , k. z. & krzeminski , w. 1998 , acta astronomica , 48 , 439 knox , r. a. , hawkins , m. r. s. & hambly , n. c. 1999 , , 306 , 736 kurucz , r.l .
1991 , in stellar atmospheres : beyond classical models , ed .
l. crivellari , i. hubeny , d.g .
hummer , ( dordrecht : kluwer ) , 440 majewski , s.r .
1993 , , 31 , 575 mcwilliam , a. 1990 , , 74 , 1075 olive , k. , steigman , g. , & skillman , e.d .
1997 , apj , 483 , 788 oswalt , t. d. , smith , j. a. , wood , m. a. & hintzen p. 1995 , nature , 382 , 692 paczyski , b. 1990 , apj , 348 , 485 peterson , r. c. & green , e. m. 1998 , , 502 , l39 phelps , r. l. 1997 , , 483 , 826 salaris , m. & weiss , a. 1998 , , 335 , 943 sarajedini , a. , chaboyer , b. & demarque , p. 1997
, , 109 , 1321 sarajedini , a. , von hippel , t. , kozhurina - platais , v. & demarque , p. 1999 , , 118 , 2894 scott , j. e. , friel , e. d. & janes , k. a. 1995 , , 109 , 1706 searle , l. 1977 , in the evolution of galaxies and stellar populations , ed .
tinsley , r.b .
larson , ( new haven : yale univ .
press ) , p.219 .
the formation of the galaxy .
searle , l. & zinn , r. 1978 , , 225 , 357 stetson , p. b. , vandenberg , d. a. , & bolte , m. , 1996 , pasp , 108 , 560 thogersen , e. n. , friel , e. d. & fallon , b. v. 1993 , , 105 , 1253 twarog , b. a. , ashman , k. m. & anthony - twarog , b. j. 1997 , , 114 , 2556 ccccc @xmath117 & 11.43 & 0.09 & 0.1125 & 6.5 + @xmath118 & 11.34 & 0.09 & 0.1125 & 7.5 + @xmath119 & 11.48 & 0.09 & 0.1125 & 5.5 + @xmath117 & 11.31 & 0.07 & 0.0875 & 7.5 + @xmath117 & 11.52 & 0.11 & 0.1375 & 5.5 + rrrrrrrr 11007 & @xmath120 & 0.57 & @xmath121 & @xmath122 & @xmath123&@xmath124&@xmath125 + 15335 & @xmath126 & 0.59 & @xmath127 & @xmath128 & @xmath129&@xmath130&@xmath131 + 32923 & @xmath132 & 0.66 & @xmath133 & @xmath134 & @xmath135&@xmath136&@xmath137 + 41330 & @xmath138 & 0.60 & @xmath139 & @xmath140 & @xmath141&@xmath142&@xmath143 + 52711 & @xmath118 & 0.60 & @xmath144 & @xmath145 & @xmath146&@xmath147&@xmath148 + 67458 & @xmath138 & 0.60 & @xmath149 & @xmath150 & @xmath151&@xmath152&@xmath153 + 202628 & @xmath154 & 0.64 & @xmath155 & @xmath156 & @xmath157&@xmath158&@xmath159 + 207129 & @xmath118 & 0.60 & @xmath160 & @xmath161 & @xmath162&@xmath163&@xmath164 + 210918 & @xmath120 & 0.65 & @xmath165 & @xmath166 & @xmath167&@xmath168&@xmath169 + rrrrrcrrrrrr 11007 & @xmath170 & @xmath171&@xmath172&@xmath173 & 2006.4 & 11.83 & 7.71 & 0.84 & 0.211 & 4.85 & thick disk ?
+ 15335 & @xmath174 & @xmath175&@xmath176&@xmath177 & 2126.4 & 12.66 & 7.96 & 0.08 & 0.228 & 0.44 & thin disk + 32923 & @xmath178 & @xmath179&@xmath180&@xmath181 & 1680.8 & 8.20 & 7.24 & 0.46 & 0.063 & 3.45 & thin disk + 41330 & @xmath182 & @xmath183&@xmath184&@xmath185 & 1669.6 & 8.22 & 7.05 & 0.29 & 0.077 & 2.15 & thin disk + 52711 & @xmath186 & @xmath187&@xmath188&@xmath189 & 1255.2 & 8.00 & 4.26 & 0.02 & 0.306 & 0.22 & thick disk + 67458 & @xmath190 & @xmath191&@xmath192&@xmath193 & 1822.4 & 10.95 & 6.64 & 0.21 & 0.245 & 1.41 & thick disk ? + 202628 & @xmath194 & @xmath195&@xmath196&@xmath197 & 1886.4 & 9.39 & 8.00 & 0.23 & 0.080 & 1.51 & thin disk + 207129 & @xmath198 & @xmath199&@xmath200&@xmath201 & 1695.2 & 8.01 & 7.38 & 0.08 & 0.041 & 0.61 & thin disk + 210918 & @xmath202 & @xmath203&@xmath204&@xmath205 & 1137.6 & 8.19 & 3.60 & 0.01 & 0.390 & 0.09 & thick disk + @xmath76 & 0.40 & 13.45 & 0.04 & 0.05 & no simult .
fit + @xmath206 & 0.40 & 13.40 & 0.04 & 0.05 & no simult .
fit + @xmath207 & 0.40 & 13.50 & 0.04 & 0.05 & @xmath208 + @xmath207 & 0.40 & 13.28 & 0.03 & 0.0375 & @xmath209 + @xmath207 & 0.40 & 13.42 & 0.05 & 0.0625 & @xmath210 + llrcclc 47 tuc & thick disk & @xmath211 & 6.2 & 0 & iso .
fit & & this paper + field stars & thick disk & @xmath120 & 5.9 & @xmath212 & isochrone & this paper + ngc 6791 & thick disk ? & @xmath29 & 6.6 & @xmath213 & isochrone fit & cgl + field stars & thin disk & @xmath126 & 7.6 & @xmath214 & isochrone & this paper + field stars & thin disk & @xmath215 & 8.0 & @xmath216 & isochrone & this paper + ngc 188 & thin disk & @xmath117 & 10.1 & @xmath217 & isochrone fit & this paper + m107 & inner halo & @xmath218 & 2.8 & @xmath219 & & csa + ngc 1851 & inner halo & @xmath220 & 18.0 & @xmath221 & & csa + ngc 6652 & inner halo & @xmath222 & ( 2.0 ) & @xmath223 & & csa + | we determine the relative ages of the open cluster ngc 188 and selected hipparcos field stars by isochrone fitting , and compare them to the age of the thick disk globular cluster 47 tuc .
the best fit age for ngc 188 was determined to be 6.5 @xmath01.0 gyr .
the solar metallicity hipparcos field stars yielded a slightly older thin disk age , 7.5 @xmath00.7 gyr .
two slightly metal - poor ( @xmath1$}}= -0.22 $ ] ) field stars whose kinematic and orbital parameters indicate that they are members of the thin disk were found to have an age of @xmath2 gyr .
the age for 47 tuc was determined to be 12.5 @xmath01.5 gyr .
all errors are internal errors due to the uncertainty in the values of metallicity and reddening .
thus , the oldest stars dated in the thin disk are found to be @xmath3 gyr younger than 47 tuc .
furthermore , as discussed by @xcite 47 tuc has a similar age to three globular clusters located in the inner part of the galactic halo , implying that star formation in the thin disk started within @xmath3 gyr of star formation in the halo . |
due to the fact , that experimental setups are usually not suited to cover the complete complete phase space , event generators are very important tools for experiments .
practically , the experimental physicist needs a tool in hand which allows to control almost all kinematic variables with a manageable user interface and to exchange the physics models which have to be compared to measured data . in our contribution
, we present the software structure of the pluto event generator @xcite originally developed for the hades experiment @xcite but successfully used by other collaborations in the hadronic physics field as well .
its recent redesign partly discussed in this contribution enhanced its flexibility and provide new features which allow to meet new challenges coming with the detector studies for the new fair experiments panda @xcite and cbm @xcite .
pluto is a collection of c++ classes , adding up to the framework of a simulation package for hadronic physics interactions in the energy regime up to a few gev .
it is launched interactively within the root @xcite environment , and makes use of root only , without requiring additional packages .
the focus is on streamlining particle generators by providing the tools to set up and manipulate particles , reaction channels , and complex reactions , as well as applying experimental filters on the reaction products , such as geometrical acceptance and kinematical conditions .
typical simulations may be executed with a few lines of input , with no expertise required on the part of the user .
it standardizes simulations by providing a common platform that is accessible via the analysis environment ( root ) , with a straightforward user interface that does not inhibit non - experts in simulations .
the output may be analyzed on line , or further forwarded to a digitization package .
the pluto framework includes models for hadronic and electromagnetic decays , resonance spectral functions with mass - dependent widths , and anisotropic angular distributions for selected channels .
a decay - manager interface enables `` cocktail '' calculations .
an extensive particle data base is available , with capabilities to support user - defined ones .
various particle properties and decay modes are included in the data base .
thermal distributions are implemented , enabling multi - hadron decays of hot fireballs .
the general philosophy of pluto is that on one hand the users should be able to drive the package in a very easy way , but on the other hand are allowed to customize it without much restrictions .
this means that the steering application has the opportunity to add new decays / particles among with new models up to the incorporation of plugins .
structure and design of pluto : the ( orange ) boxes in the upper left show the user classes to set up and handle reactions ( and decay channels ) .
the ( yellow ) box on the left is the interface for accessing the particle stream while the event loop is running .
the ( blue shaded ) boxes in the lower right corner represent the data base for the particles and decays among with a user interface for customization .
the ( light blue ) boxes above are the distribution interface and the included physics models ( color online ) . ] the structure and design of pluto are sketched in fig .
[ fig : pluto_at_a_glance ] . the framework can be roughly divided into 3 parts : a user interface for event production , a physics package , and an event loop interface . in this report
we would like to concentrate on the latter two issues : the first one allows to add and/or select physics cases , where models can be either build - in models or runtime replacements , which will be discussed in sec .
[ sec : models ] .
this area of pluto works even outside the event loop for the calculation of parameters ( mass distributions , decay widths , ... ) .
the other block allows for external particle decay and acceptance filter classes ( see sec . [
sec : bulk ] ) and filling of on - line histograms ( see sec . [
sec : projector ] ) . as the interface for event production
is described in detail elsewhere @xcite in this context we will give only a short example .
after all models have been configured accordingly , an elementary reaction chain is defined with a character string containing the reaction products . to perform , e.g. , the @xmath4 production in the @xmath5 reaction with a consecutive dalitz decay as discussed in the following sections , the commands as defined below are sufficient : .... preaction * r = new preaction("1.25","p","p " , " p d+ [ p dilepton [ e+ e- ] ] ") ; r->print ( ) ; r->loop(10000 ) ; .... with the number 1.25 as the kinetic beam energy in gev , the beam and target particle indicated by following two options and finally a string describing the decay .
particles are defined via their data base name ( as `` p '' for a proton or `` d+ '' for a @xmath4 ) .
the setup of the reaction channels and the attachment of all selected distributions and models is done by the framework , with the possibility to dump the complete chain and its models via the print ( ) method .
finally , the command loop(n ) produces @xmath6 events .
then , event - event - by - event , the individual steps in each reaction are executed via a common interface of all classes making distributions ( see @xcite for details ) which do the calculation of final particle tracks ( also the unstable ones ) which are stored in an array .
in the pluto framework , each physics process is described by a number of objects , instantiated from dedicated classes . beside the objects which communicate with the event loop and set the 4-momenta of the resulting particle tracks ( to form the above - mentioned `` distributions '' ) , pluto offers methods to make calculations including various channels and their accompanied functions in advance to ( or event without ) the event - loop . for this purpose ,
a relative data base has been included , which connects particles and decays to `` models '' .
at least one ( so - called `` _ primary _ '' ) model is needed per decay in order to describe the respective mass shape ( usually calculated by a mass - dependent breit - wigner in a recursive way ) and the partial / total decay width . each selected model can be accessed from all locations of the software package as well as by the user .
the choice which of the allover offered physics models are linked to the data base can be done inside the production root - macro by calling a distribution manager ( for details about the handling see @xcite ) , which itself takes care that only one primary model per decay and particle is selected , respectively .
moreover , pluto offers the user the freedom to attach user - defined models based on its own sources or even new classes ( written in c++ ) at runtime , thereby replacing built - in models .
these features make pluto an extremely versatile and highly configurable tool for simulations , while guaranteeing overall consistency at all times .
one specific feature of pluto is that we do not use monolithic decay models only but allow for the splitting of the underlying physics process into different models in a very granular way ( e.g. , to exchange form factors or total cross sections ) .
this turned out to be a very important tool in order to check various scenarios along with measured data .
therefore pluto allows for the attachment of secondary models for all kinds of purposes . here
, a secondary model is an object for a particle / decay returning a ( complex ) number as a function of a defined number of values .
we do not attempt to have a base class for each kind of function and model ( e.g. , a total cross section base class , a base class for @xmath7 , etc . ) because this makes further extensions difficult .
our approach is as follows : for each class of secondary models a new entry ( defined by a unique name ) is added dynamically to the data base .
consequently , each model is bound to two data base keys : one to link it to a decay ( or particle ) and one more for the type of secondary model .
anyhow , there the user does not need to take care about these issues , as the following template example shows . in pluto ,
each model is inherited from the base class pchannelmodel with a lot of useful features . in order to offer the result of the calculation the getweight ( ) method
has to implemented .
let us create such an example class as a template for the following discussion and future developments : .... class phelperfunctiontest : public pchannelmodel { public : phelperfunctiontest(const char_t * i d , const char_t * de , int_t key ) ; double_t getweight(double_t * x , int_t * opt = null ) ; } .... with an realization of the ( in this case 1-dimensional ) function : .... double_t phelperfunctiontest::getweight(double_t * x , int_t * opt ) { double_t xx = x[0 ] ; val = ... return val ; } ; .... objects can be instantiated to work as a model for specific physics processes by the usage of a parser residing inside the constructor of the base class : .... phelperfunctiontest * obj = new phelperfunctiontest("unique_name@d+_to_p_dilepton / helper " , " test",-1 ) ; .... here , the unique_name is used to identify the object inside the distribution manager .
the string hereafter binds the new object to a decay which is in this case @xmath8 .
the optional appendix /helper enables the new object to be a secondary model of this type .
let us now assume , the primary model has to access its helper function . in this case
it is sufficient to execute the following command at a single place ( an init ( ) function is available for that ) : .... pchannelmodel * sec_model = getsecondarymodel("helper " ) ; .... and execute the daughter method sec_model->getweight(value ) whenever required .
the access to alien secondary models is possible as well but a further discussion on this topic would exceed the scope of this report .
the power of such a splitting can be demonstrated best with an real example .
for this purpose let us use the study of the electromagnetic @xmath9 dalitz decay @xmath10 .
as explained in the introduction pluto reads this as a multi - step process : in the first step , the @xmath9 mass is sampled using a mass - dependent width . in the second step ,
the already sampled @xmath9 mass @xmath11 is used by the following dalitz decay model to sample the di - electron ( = virtual photon ) mass @xmath12 . by using a dedicated plugin : .... makedistributionmanager()->exec("dalitz_mod : krivoruchenko " ) ; .... the following model for the di - electron mass spectrum is enabled , using the prescription of ref .
@xcite with : @xmath13 and the decay rate given by krivoruchenko _
et al _
@xcite : @xmath14 where the index @xmath15 refers to the produced nucleon , @xmath16 is the electron charge , and @xmath17 represents the electromagnetic transition amplitudes . here , from the software point of view , it is important to know that this factor is a separated object in pluto making the addition of new complementary descriptions very easy .
this avoids the usage of nasty flags which is error - prone .
e.g. , following the qed approach ( assuming that no hadronic transition form factor is involved ) from ref.@xcite gives : @xmath18 with @xmath19 , @xmath20 , and @xmath21 as the magnetic , electric and coulomb transition form factors .
alternatively , one can switch to the vmd - like form factor of iachello and wan @xcite which is already implemented and will be discussed in detail in a further , separate publication . after activating this plugin , both above - mentioned versions ( vmd and qed )
are offered , appearing as objects with their unique names in the directory - like distribution manager structure . ....
makedistributionmanager()->print("vmd " ) ; -------------------- pdistributionmanager -------------------- root root group : 494 objects ( of 504 ) , subgroups : 9 particle_models mass sampling of particles : 27 objects ( of 29 ) decay_models phase space mass sampling & decay partial widths : 219 objects ( of 225 ) polar_angles polar angles in elementary particle production : 10 objects ( of 10 ) genbod_models momentum sampling : 220 objects ( of 220 ) ... vmd vmd form factors : 0 objects ( of 2 ) ( ) d+_vmd_ff vmd ff for d+ - > p e+e- ( x ) d+_qed_ff qed ff for d+ - > p e+e- ... qed qed form factors : 2 objects ( of 2 ) .... possible conflicts between the qed and vmd models are indicated .
the user can now activate all models inside the directory via : .... makedistributionmanager()->enable("vmd " ) ; ....
the bulk interface . ]
the aim of the event loop interface is to allow access to the particle bulk before it is written to disk by providing an interface for bulk decays ( also using 3@xmath22 party event generators such as pythia @xcite ) , file access , implantation of embedded particles for detector studies , and filter classes .
it works - like indicated in fig .
[ fig : bulk_interface ] - by stacking individual bulk classes ( by preaction::addbulk(pbulkinterface * obj ) ) and making use of the base - class modify ( ) method which handles the embedding and removing of particles or mark them as decayed / invalid .
like in the previous section , user classes can be compiled - if needed - and added on - the - fly , which could be very helpful in the context of fair experiments studies , as the following applications show : particle tracks might be added from ( or written to ) user files in any self - defined format .
this feature makes pluto open for other collaborations and enables it to be used as an `` afterburner '' for externally created simulation data or output of transport models .
test particles can be embedded into a background reaction which is an important tool for detectors studies , e.g. to test tracking in a high multiplicity environment . moreover , among with the previous feature the mixing of embedded particles with either real events or 3@xmath22 party simulation is one of the applications . in the first case ,
the reconstructed vertex of the analyzed experimental data is used to simulate rare probes ( as the di - electrons ) at exactly the same vertex . in this way the efficiency of these probes can be calculated with the best background one can have , namely measured data .
in addition to a well defined reaction chain , pluto offers free `` bulk '' decays , i.e. the decay of ( embedded ) particles following the calculated mass - dependent branching ratios . at this stage
it is possible as well to employ 3@xmath22 party event generators . only to give an example
: pluto can generate embedded particles like resonances with a selected distribution and the pythia package ( or pluto itself ) can be used for their realistic free decay .
the particles created in the previous steps can be pre - filtered using a `` detector acceptance '' user class following a class definition as outlined below : .... class pmydetector : public pbulkinterface { private : double_t * acc ; public : bool_t modify(pparticle * * stack , int * decay_done , int * num , int stacksize ) ; } ; pmydetector::pmydetector ( ) { acc = makestaticdata()->getbatchvalue("acc_value " ) ; } ; .... with acc_value as a booked batch value ( see below for more information ) and , e.g. : .... bool_t pmydetector::modify(pparticle * * stack , ... ) { * acc=1 . ; for ( int i=0 ; i < * num ; i++ ) { pparticle * cur = stack[i ] ; if ( " check for cur " ) * acc=0 . ; } } ....
the usual way to analyze generated events ( beside a full monte - carlo production including digitization ) is to open the simulation file with an appropriate analysis macro , loop over the number of events , read the particle objects and do the required operations inside the loop .
e.g. , in hadronic physics , in particular for exclusive studies , usually missing and invariant masses are reconstructed , and possibly after boost operations angular correlations are investigated .
in addition , signal selection has to be done not only on a single track level but using particle correlations ( such as a invariant / missing mass selection ) .
this prevents to use the features provided by ttree::draw ( ) only , as these are fixed to single track properties .
overview of the pprojector design . ]
the main aim of the `` pprojector '' ( see fig . [
fig : pluto_pprojector ] ) is to offer a simplified analysis method following a similar idea as ttree::draw ( ) in order to project particles ( and their correlations ) onto on - line histograms and/or tntuples by making use of the bulk interface described in the previous section .
this means , no analysis macro is needed for a large number of typical observables .
there are two ways to include the projector into the event loop .
either one or more of its instances are added to the bulk interface ( for an alternating mixing with other classes ) , or , which is sufficient for the most cases , the short version preaction::do(char * command ) can be used , where command follows a c++-like ( but more simple ) syntax .
each of these commands - just forming steps in a batch - is executed inside the event loop . to fill histograms
one can use the extended method preaction::do(thx * histo , char * command ) where histo points to a 1/2/3-dimensional histogram . in this case
the command has to define the key variables _ x , _ y and _ x as the respective axis value , depending on the chosen dimension of histo .
the pluto particle array objects can be accessed via its pluto particle name inside brackets as with the optional num as the consecutive number of equal particles inside each event .
all objects can be combined with a number of build - in methods , moreover all browsable methods of the root - class tlorentzvector can be employed , as the following example indicates : .... r->do(histo1,"_x = [ d+]->m ( ) " ) ; .... which fills the 1-dimensional histo1 with the @xmath4 mass .
invariant masses ( composite particles ) can be formed by adding the objects : .... " _ x = ( [ e+ ] + [ e-])->m2 ( ) " .... and stored for a consequent use in semi - colon separated sub - commands without requiring a `` new '' operator : .... " dilepton = ( [ e+ ] + [ e- ] ) ; _ x = dilepton->m2 ( ) " .... let us now come to a more complex example , which is the calculation of the polar emission angle of one of the outgoing protons in a @xmath23 reaction ( n.b .
pluto treats the c.m .
system as a particle named p + p ) : .... " p1 = [ p,1 ] ; p1->boost([p + p ] ) ; _ x = cos(p1->theta ( ) ) " .... in this example , in the first sub - command the first appearing proton is copied to the object p1 which is furthermore boosted into the c.m .
energy frame and in the last step the cosine of its polar angle is written as a value to the x - axis .
the projection to tntuples turned out to be very useful as it allows to make additional selections ( and definition of the histogram axis ) afterwards which is a very simple task using the pprojector following the syntax as described above .
let us now assume , the momenta of simulated @xmath24 s should be stored .
then the first step would be to define the tntuple with branch names : .... tntuple * ntuple = new tntuple("ntuple"," ... ","eta_px : eta_py : eta_pz " ) ; .... hereafter , the do()-method has to be called like in the above - indicated example but with a pointer to a tntuple - object instead of a histogram and the variable names of the ntuple instead of the axis names : .... " eta_px = [ eta]->px ( ) ; eta_py = [ eta]->py ( ) ; eta_pz = [ eta]->pz ( ) " .... the filling of the ntuple is done by pluto , only ( but on purpose ) the opening and closing of the root - file has to be called externally .
in addition to the writing of tntuples it is also possible to read ( back ) any ntuple , modify its values and write it back or make histograms .
let us use the ntuple of the previous example which uses the ntuple ( usually from a file ) and defines an _ empty _ reaction , but attaches the ntuple as a 3rd party input : .... preaction my_reaction ; my_reaction.input(ntuple ) ; .... the next step is to use the branches of the ntuple in the batch command line to , e.g. , reconstruct the 4-momentum : .... " myeta = p3m(eta_px , eta_py , eta_pz,0.54745 ) " .... with p3 m as the short syntax for a vector constructor with 3 momenta and the particle mass as parameters . at this point
any kind of mathematical operation available via tformula can be applied as well . finally using loop ( ) without a number reads all events from the ntuple .
the employment of batch commands allows to define on - line filters on single particle properties as well as kinematic variables which are accessible as indicated above .
these filters can be used to select events for single histograms only or forwarded to the event loop to apply physical constraints , before a full event is written to disk , hence reducing the event file .
variables prepared in a detector class as outlined in sec . [ sec : filters ] can be integrated very easily .
this is realized by using the if - command , followed by the variable or a combined condition based on particle observables .
the remaining batch sub - commands are executed only under a positively evaluated condition .
it goes without saying that the condition statement can combine all methods among with the established tformula syntax : .... " if(obj->p ( ) > 0.3 & & obj->p ( ) < 0.6 ) ; dowhatever ... " .... in this case , histograms and ntuples in the do ( ) method are only filled under the defined condition , or , `` dowhatever '' can be used to re - define variables . in this context , it should be noted , that variables , starting with `` # '' are interpreted by the pluto event loop as a condition for the final event file . in our next case ,
an event is accepted if both the electron and the positron are produced with polar angles in the range of 18 to 85 degrees : .... " theta_ep = ( [ e+]->theta ( ) * 180.)/tmath::pi ( ) " " theta_em = ( [ e-]->theta ( ) * 180.)/tmath::pi ( ) " " # filter = 1 ; if(theta_ep<18 || theta_ep>85 || theta_em<18
theta_em>85 ) ; # filter = 0 " .... by this means it is only a single step to activate the detector class mydetector from sec .
[ sec : filters ] : .... " # acc_filter = acc_value " .... and optionally combine the prepared acceptance flag with our kinematical conditions .
in summary , we presented new developments for the pluto event generator which moves this package from being a pure event generator to a comprehensive framework .
the particular features are the bookkeeping of various models and the user interface for their simple selection .
the new event loop interface allows to mix several sources , to define and enable online selections , project to histograms and to make a lot of operations .
detector acceptance and resolution can be included without much overhead , which is of particular importance for detector studies in the context of the fair project . in the near future
, we will implement general purpose resolution scenarios using a common data format and file readers to transport models ( like urqmd @xcite and hsd @xcite ) .
this will allow to test very rapidly the performance of different detector configuration options .
this work was supported by the hessian loewe initiative through the helmholtz international center for fair ( hic for fair ) .
we would like to thank b. kmpfer for reading and commenting the manuscript .
g. agakishiev _ et al . _
[ the hades collaboration ] , arxiv:0902.3478 [ nucl - ex ] .
j. g. messchendorp [ panda collaboration ] , arxiv:0711.1598 [ nucl - ex ] ; b. seitz [ panda collaboration ] , _ prepared for 16th international workshop on deep inelastic scattering and related subjects ( dis 2008 ) , london , england , 7 - 11 apr 2008_. | we present the design of the simulation package pluto , aimed at the study of hadronic interactions at sis and fair energies .
its main mission is to offer a modular framework with an object - oriented structure , thereby making additions such as new particles , decays of resonances , new models up to modules for entire changes easily applicable .
overall consistency is ensured by a plugin- and distribution manager .
particular features are the support of a modular structure for physics process descriptions , and the possibility to access the particle stream for on - line modifications .
additional configuration and self - made classes can be attached by the user without re - compiling the package , which makes pluto extremely configurable .
@xmath0 institut fr kernphysik , goethe - universitt , 60438 frankfurt , germany + @xmath1 gsi helmholtzzentrum fr schwerionenforschung gmbh , 64291 darmstadt , germany + @xmath2 institut de physique nuclaire dorsay , cnrs / in2p3 , 91406 orsay cedex , france + @xmath3 smoluchowski institute of physics , jagiellonian university of cracow , 30 - 059 krakw , poland |
observations of black hole demographics locally is increasingly providing a strong constraint on models that explain the assembly and growth of black holes in the universe .
the existence of a tight relation between the velocity dispersion of bulges and the mass of the central black hole has been reported by several authors ( merritt & ferrarese 2001 ; tremaine et al .
2002 ; gebhardt et al .
this correlation is tighter than that between the luminosity of the bulge and the mass of the central black hole ( magorrian et al .
1998 ) . the physical processes that set up this correlation
are not fully understood at the present time , although there are several proposed explanations that involve the regulation of star formation with black hole growth and assembly in galactic nuclei ( haehnelt , natarajan & rees 1998 ; natarajan & sigurdsson 1998 ; silk & rees 1999 ; murray , quataert and thompson 2004 ; king 2005 ) .
recent work by several authors has suggested that umbhs are hereafter referred to as umbhs . ]
ought to exist : bernardi et al .
( 2006 ) show that the high velocity dispersion tail of the velocity distribution function of early - type galaxies constructed from the sloan digital sky survey ( sdss ) had been under - estimated in earlier work suggestive of a corresponding high mass tail for the central black hole masses hosted in these nuclei . as first argued by lauer et al .
( 2007a ) and subequently by bernardi et al .
( 2007 ) and tundo et al .
( 2007 ) , even when the scatter in the observed @xmath4 correlation is taken into account it predicts fewer massive black holes compared to the @xmath5 relation . while bernardi et al .
( 2007 ) argue that this is due to the fact that the @xmath6 relation in currently available samples is inconsistent with the sdss sample from which the distributions of @xmath7 or @xmath3 are based . from an early - type galaxy sample observed by hst , lauer et al .
( 2007b ) argue that the relation between @xmath5 is likely the preferred one for bcgs ( brightest cluster galaxies ) consistent with the harboring of umbhs as evidenced by their large core sizes .
the fact that the high mass end of the observed local black hole mass function is likely biased is a proposal that derives from optical data . deriving the mass functions of accreting black holes from optical quasars in the sloan digital sky survey data release 3 ( sdss dr3 ) , vestergaard et al .
( 2008 ) also find evidence for umbhs in the redshift range @xmath8 . in this paper
, we show that umbhs exist using x - ray and bolometric agn luminosity functions and for consistency with local observations of the bh mass density , an upper limit to their masses is required . to probe the high mass end of the bh mass function , in earlier works
the agn luminosity functions were simply extrapolated .
this turns out to be inconsistent with local estimates of the bh mass function .
here we focus on the high mass end of the predicted local black hole mass function , i.e. extrapolation of the @xmath4 relation to higher velocity dispersions and demonstrate that a self - limiting cut - off in the masses to which bhs grow at every epoch reconciles the x - ray and optical views .
the outline of this paper is as follows : in section 2 , we briefly summarise the current observational census of black holes at high and low redshift including constraints from x - ray agn . the pathways to grow umbhs
are described in section 3 .
derivation of the local black hole mass function from the x - ray luminosity functions of agn is presented in section 4 .
the arguement for the existence of an upper limit to black hole masses from various lines of evidence is presented in section 5 ; the prospects for detection of this population is presented in section 6 followed by conclusions and discussion .
we adopt a cosmological model that is spatially flat with @xmath9 ; @xmath10 .
the demography of local galaxies suggests that every galaxy hosts a quiescent supermassive black hole ( smbh ) at the present time and the properties of the black hole are correlated with those of the host . in particular , observational evidence points to the existence of a strong correlation between the mass of the central black hole and the velocity dispersion of the host spheroid ( tremaine et al .
2002 ; merritt & ferrarese 2001 , gebhardt et al .
2002 ) in nearby galaxies .
this correlation strongly suggests coeval growth of the black hole and the stellar component via likely regulation of the gas supply in galactic nuclei ( silk & rees 1999 ; kauffmann & haehnelt 2000 ; cattaneo 2001 ; bromley , somerville & fabian 2004 ; king 2003 ; murray , quataert & thompson 2005 ; sazonov et al .
2005 ; begelman & nath 2005 ; alexander et al .
2005 ) .
black hole growth is primarily powered by gas accretion ( lynden - bell 1969 ) and accreting black holes that are optically bright are detected as quasars .
the build - up of smbhs is likely to have commenced at extremely high redshifts .
indeed , optically bright quasars have now been detected at @xmath11 ( e.g. , fan et al . 2001a , 2003 ) in the sdss .
there are also indications that high redshift quasar hosts are strong sources of dust emission ( omont et al .
2001 ; cox et al .
2002 ; carilli et al .
2002 ; walter et al . 2003 ; reuland et al .
2004 ) , suggesting that quasars were common in massive galaxies at a time when galaxies were undergoing copious star formation .
the growth spurts of smbhs are also detected in the x - ray waveband .
the summed emission from these agn generates the cosmic x - ray background ( xrb ) , and its spectrum suggests that most black - hole growth is optically obscured ( fabian 1999 ; di matteo et al .
1999 ; mushotzky et al .
2000 ; hasinger et al . 2001 ; barger et al .
2003 ; barger et al . 2005 ; worsley et al .
there are clear examples of obscured black - hole growth in the form of ` type-2 ' quasars , and the detected numbers are in agreement with some recent xrb models ( treister & urry 2005 ; gilli et al .
2007 ) and have the expected luminosity dependence of the obscured fraction .
additionally , there is tantalizing recent evidence from infra - red ( ir ) studies that dust - obscured accretion is ubiquitous ( martinez - sansigre et al .
2005 , 2007 ) .
at present it is unknown what fraction of the total mass growth occurs in such an optically dim phase as a function of redshift .
the build - up of bh mass in the universe has been traced using optical quasar activity .
the current phenomenological approach to understanding the assembly of smbhs involves optical data from both high and low redshifts .
these data are used to construct a consistent picture that fits within the larger framework of the growth and evolution of structure in the universe ( haehnelt , natarajan & rees 1998 ; haiman & loeb 1998 ; kauffmann & haehnelt 2000 ; 2002 ; wyithe & loeb 2002 ; volonteri et al .
2003 ; di matteo et al . 2003 ; steed & weinberg 2004 ) .
black hole accretion histories derived from the quasar luminosity function ( e.g. soltan 1982 ; haehnelt , natarajan & rees 1998 ; salucci et al .
1999 ; yu & tremaine 2002 ; marconi et al . 2004 ; shankar et al . 2004
; merloni et al .
2004 ) , synthesis models of the xrb ( e.g. comastri et al . 1995 ; gilli et al . 1999
; elvis et al . 2002 ; ueda et al . 2003 ; barger et al . 2005 ; treister & urry 2005 ; gilli et al .
2007 ) , and observations of accretion rates in quasars at different redshifts ( vestergaard 2004 ; mclure & dunlop 2004 ) and composite models ( hopkins et al . 2005b ; 2006a ; 2006b ) suggest that supermassive black holes spend most of their lives in a low efficiency , low accretion rate state . in fact , only a small fraction of the smbhs lifetime is spent in the optically bright quasar phase , although the bulk of the mass growth occurs during these epochs . in this paper , we examine the consequences of such an accretion history for the high mass end of the local black hole mass function .
surveys at x - ray energies allow us to obtain a more complete view of the agn population , as they cover a broader range in luminosity and are simultaneously less affected by baises due to obscuration . while optical surveys of quasars , like the sdss or 2df , are used to obtain a large sample of unobscured and high - luminosity sources , it is with x - ray surveys that the obscured low - luminosity population can be well traced . in particular , surveys at hard x - ray energies , 210 kev , are almost free of selection effects up to columns of @xmath12 . in the work of ueda et al .
( 2003 ) the agn x - ray luminosity function is computed based on a sample of @xmath13250 sources observed with various x - ray satellites .
one of the important conclusions of this paper is the confirmation of a luminosity - dependent density evolution , in the sense that lower luminosity sources peak at lower redshifts , @xmath14 , while only the high luminosity sources are significantly more abundant at @xmath15 , as observed in optical quasar surveys ( e.g. , boyle et al .
additionally , using this x - ray luminosity function and evolution it was possible for ueda et al .
( 2003 ) to convincingly account for the observed properties of the extragalactic xrb .
extending the argument presented by soltan ( 1982 ) to the x - ray wave - band , agn activity can be used to trace the history of mass accretion onto supermassive black holes ( fabian & iwasawa 1999 ) .
marconi et al . ( 2004 ) and shankar et al . ( 2004 ) used the luminosity function of ueda et al .
( 2003 ) to calculate the spatial density of supermassive black holes inferred from agn activity and compared that with observations .
these authors reported in general a good agreement between observations and the density inferred from agn relics , suggesting that there is little or no room for further obscured accretion , once compton thick agn are properly accounted for .
a similar conclusion was also obtained by barger et al .
( 2005 ) from an independently determination of the luminosity function , thus confirming this result .
below we discuss plausible scenarios for forming these umbhs at low redshift .
there are 2 feasible channels for doing so : ( i ) expect extremely rare umbhs to form from the merging of black holes due to the merging of galaxies via the picture suggested by volonteri et al .
( 2003 ) ; ( ii ) form from accretion onto high redshift ` seeds ' with perhaps a brief period of super - eddington accretion , the descendants of the smbhs that power the most luminous quasars at @xmath16 as proposed recently by volonteri & rees ( 2005 ) ; begelman , volonteri & rees ( 2006 ) ; lodato & natarajan ( 2007 ) and volonteri , lodato & natarajan ( 2007 ) .
we discuss these two possible channels for growing umbhs in more detail below . following the merging dm hierarchy of halos starting with seed bhs at @xmath17 , populating the @xmath18 peaks , volonteri et al .
( 2003 ) are able to reproduce the mass function of local bhs as well as the abundance of the rare @xmath19 bhs that power the @xmath20 sdss quasars . proceeding to rarer peaks
say , @xmath21 at @xmath17 in this scheme yields the rarer @xmath22 local umbhs . and in fact , the formation of a very small number density of umbhs at @xmath1 is inevitable in the standard hierarchical merging @xmath23cdm paradigm . a massive dm halo with mass , @xmath24 at @xmath1 which is the likely host to an umbh , is likely to have experienced about 100 mergers between @xmath16 and @xmath1 , starting with @xmath25 at @xmath16 .
recently a numerical calculation of the merger scenario mentioned above has been performed in simulations by yoo et al .
focusing on the merger history of high mass cluster - scale halos ( @xmath26 ) .
they find that in ten realizations of halos on this mass scale , starting with the highest initial bh masses at @xmath27 of @xmath13 few times @xmath28 , 4 clusters contain umbhs at @xmath1 .
therefore , rare umbhs are expected in the local universe .
yoo et al . ( 2007 ) argue that black hole mergers can significantly augment the high end tail of the local bh mass function . similarly , using a model for quasar activity based on mergers of gas - rich galaxies , hopkins et al .
( 2006a ) showed that they could explain the observed local bh mass at low to intermediate bh masses ( 10@xmath29 - 10@xmath30@xmath31 ) .
however , at higher bh masses , their calculations overpredict the observed values even considering a possible change in the eddington fraction at higher masses .
conventional models of black hole formation and growth start with initial conditions at high redshift with seed bhs that are remnants of the first generation of stars in the universe . propagating these seeds via merger accompanied accretion events leading to mass growth for the bhs ( volonteri , haardt & madau 2003 ) it has been argued that in order to explain the masses of bhs powering the bright @xmath32 quasars by the sdss survey ( fan et al .
2004 ; 2006 ) that either a brief period of super - eddington accretion ( volonteri & rees 2005 ) or more massive seeds are needed ( begelman , volonteri & rees 2006 ; lodato & natarajan 2006 ; lodato & natarajan 2007 ) .
massive seeds can alleviate the problem of assembling @xmath33 bhs by @xmath16 which is roughly 1 gyr after the big bang in the concordance @xmath23cdm model .
the local relics of such super - grown black holes are expected to result in umbhs .
we note here that following the evolution of the massive black holes that power the @xmath34 quasars , in a cosmological simulation , di matteo et al .
( 2008 ) find that these do not necessarily remain the most massive black holes at subsequent times .
therefore , while umbhs might not be direct descendants of the smbhs that power the @xmath35 quasars , there is ample room for umbhs to form and grow .
below , we briefly present scenarios that provide the massive bh seeds in the first place that will eventually result in a small population of umbhs by @xmath1 .
these physically plausible mechanisms are critical to our prediction of umbhs at low redshift .
two models have been proposed , one that involves starting from the remnants of population iii stars with brief episodes of accretion onto them exceeding the eddington rate to bump up their masses ( volonteri & rees 2006 ) and the other that explains direct formation of massive bh seeds prior to the formation of the first stars ( lodato & natarajan 2006 ; 2007 ) .
volonteri & rees ( 2005 ) have proposed a scenario to explain the high bh masses @xmath36 needed to power the luminous quasars detected @xmath16 in the sdss .
this is accomplished they argue by populating the @xmath37 peaks in the dark matter density field at @xmath38 with seed bhs which arise from the remnants of population iii stars in the mass ranges @xmath39 and @xmath40 .
these remnant bhs then undergo an episode of super - eddington accretion from @xmath41 .
they argue that in these high redshift , metal - free dark matter halos @xmath42 gas can cool in the absence of @xmath43 via atomic hydrogen lines to about @xmath44 . as shown by oh & haiman ( 2002 ) the gas at this temperature settles into a rotationally supported ` fat ' disk at the center of the halo under the assumption that the dm and the baryons have the same specific angular momentum .
further , these disks are stable to fragmentation and therefore do not form stars and exclusively fuel the bh instead . the accretion is via stable super - critical accretion at rates well in excess of the eddington rate due to the formation of a thin , inner feeding disk . the accretion radius is comparable to the radiation trapping radius which implies that all the gas is likely to end up in the bh . any further cooling down to temperatures of @xmath45 for instance , halts the accretion , causes fragmentation of the disk which occurs when these regions of the universe have been enriched by metals .
this process enables the comfortable formation of @xmath19 bhs by @xmath16 or so to explain the observed sdss quasars . in a @xmath23cdm universe ,
the time available from @xmath16 to @xmath1 is @xmath46 gyr . to grow by an order of magnitude during this epoch
requires an accretion rate of @xmath47 which is well below the eddington rate ; however , it requires a gas rich environment . in recent work , lodato & natarajan ( 2007 )
have shown that an ab - initio prediction for the mass function of seed black holes at high redshift can be obtained in the context of the standard @xmath23cdm paradigm for structure formation combined with careful modeling of the formation , evolution and stability of pre - galactic disks .
they show that in dark matter halos at high redshifts @xmath48 , where zero metallicity pre - galactic disks assemble ( prior to the formation of the first stars ) , gravitational instabilities in these disks transfer angular momentum out and mass inwards efficiently . note that the only coolants available to the gas at this epoch are either atomic or molecular hydrogen . taking into account the stability of these disks , in particular the possibility of fragmentation
, the distribution of accumulated central masses in these halos can be computed .
the central mass concentrations are expected to form seed black holes .
the application of stability criteria to these disks leads to distinct regimes demarcated by the value of the @xmath49 where @xmath50 is the temperature of the gas and @xmath51 is the virial temperature of the halo .
the three regimes and consequences are as follows : ( i ) when @xmath52 the disk fragments and forms stars instead of a central mass concentration ; ( ii ) @xmath53 , when both central mass concentrations and stars form ; ( iii)@xmath54 , when only central mass concentrations form and the disks are stable against fragmentation . using the predicted mass function of seed black holes at @xmath48 , and propagating their growth in a merger driven accretion scenario
we find that the masses of black holes powering the @xmath35 optical quasars can be comfortably accommodated and consequently a small fraction of umbhs is predicted at @xmath1 . evolving and growing these seeds to @xmath1 , the abundance of umbhs can be estimated ( volonteri , lodato & natarajan 2008 ) .
the new evidence that we present in this work for the existence of a rare population of umbhs stems from using x - ray luminosity functions of agn and the implied accretion history of black holes .
hard x - rays have the advantage of tracing both obscured and unobscured agn , as the effects of obscuration are less important at these energies .
in particular , we use the hard x - ray luminosity function and luminosity - dependent density evolution presented by ueda et al .
( 2003 ) defined from @xmath55 out to @xmath56 .
we further assume that these agn are powered by bhs accreting at the eddington limit . in order to calculate bolometric luminosities starting from the hard x - ray luminosity the bolometric corrections derived from the agn spectral energy distribution library presented by treister et al .
( 2006 ) are used .
these are based mainly on observations of local agn and quasars and depend only on the intrinsic x - ray luminosity of the source , as they are based on the x - ray to optical ratios reported by steffen et al .
( 2006 ) . to account for the contribution of compton - thick agn to the black hole mass density missed in x - ray luminosity functions
, we use the column density distribution of treister & urry ( 2005 ) with the relative number of compton - thick agn adapted to match the spatial density of these sources observed by integral , obtained from the agn catalog of beckmann et al .
( 2006 ) . in order to account for sources with column densities @xmath57=10@xmath58 - 10@xmath59 @xmath60 which do not contribute much to the x - ray background , but
can make a significant contribution to the bh mass density ( e.g. , marconi et al .
2004 ) , we multiply the bh mass density due to compton - thick agn by a factor of 2 , i.e. , we assume that they exist in the same numbers as in the @xmath57=10@xmath61 - 10@xmath58 @xmath60 range , in agreement with the assumption of marconi et al .
( 2004 ) and consistent with the @xmath57 distribution derived from a sample of nearby agn by risaliti et al .
( 1999 ) . under this assumption ,
the contribution of sources with @xmath57@xmath6210@xmath58 @xmath60 to the total population of smbhs is @xmath137% .
we then convert these x - ray lf s to an equivalent bh mass function , and evolve these mass functions by assuming that accretion continues at the eddington rate down to @xmath55 .
the results of this procedure are shown in fig . 2 for three different values of the accretion efficiency @xmath63 .
note that we do not consider models in which the efficiency parameter varies with redshift or bh mass since such models merely add more unconstrained parameters . as can be seen clearly in fig . 2 , these simple models do not reproduce the observed local black hole mass function at the high mass end .
the functional form adopted for the x - ray luminosity function is a double power - law as proposed by ueda et al .
( 2003 ) : @xmath64^{-1}.\ ] ] and the evolution is best described by the luminosity dependent density evolution model ( ldde model ) , where the cut - off redshift @xmath65 is expressed by a power law of @xmath66 , consistent with observational constraints ( see ueda et al .
2003 for more details ) : @xmath67 where @xmath68 @xmath69^{p2}\,\,\,\,\,(z \geq z_{\rm c}(l_{\rm x})).\end{aligned}\ ] ] this simple and conservative analysis predicts a population of umbhs with a local abundance of @xmath133@xmath7010@xmath71 !
this is fairly robust as this population is predicted for a large range of efficiencies .
these lf s shown in fig . 2 also simultaneously account for the cosmic xrb , as shown by several authors ( for instance see treister & urry ( 2005 ) and gilli et al .
( 2007 ) and references therein ) , suggesting that the x - ray view presents a fairly complete picture of the accretion and growth of bhs .
note that our estimates of the black hole mass function are in general agreement with those of marconi et al .
( 2004 ) [ for a direct comparison see their fig . 2 , right - hand panel ]
, the very slight difference arises due to an alternate choice of bolometric correction factors and our prescription for including compton thick agn .
estimates by other authors are also in agreement with our treatment here out to masses of a few times @xmath72 . for bh masses
@xmath73 , there appears to be consistency between the optical and x - ray views of black hole growth .
however for @xmath74 , all models that assume eddington accretion with varying efficiencies systematically over - estimate the local abundance of high mass black holes . as can be seen in fig . 2 , for a reasonable value of the efficiency , @xmath75
, there is a good agreement between the bh mass density at @xmath55 , as obtained from the velocity dispersion of bulges , and the density inferred from agn relics , for bh masses smaller than @xmath76 .
however , for higher masses , in particular the umbh mass range , independent of the value of @xmath63 assumed , the bh mass density from agn relics is significantly higher than the observed value , indicating that umbhs should be more abundant than current observations suggest .
if there is a mass dependent efficiency factor for accretion such that higher mass bhs tend to accrete at higher efficiency and hence at lower rates , then our estimate of the high mass tail would be an over - estimate .
there is however no evidence for such a mass dependence at lower masses ( hopkins , narayan & hernquist 2006 ) .
the sdss first data release covers approximately 2000 square degrees ( abazajian et al .
2003 ) , yielding a comoving volume of a cone on the sky out to @xmath77 of @xmath78 mpc@xmath79 .
given our predicted abundance above , we expect @xmath13 1000 umbhs in the sdss volume , however only a few are detected .
* no combination of assumed accretion efficiency and eddington ratio coupled with the x - ray agn lf can reproduce the observed local abundance at the high mass end . * however , we find that modifying one of the key assumptions made above brings the predicted abundance of local umbhs into better agreement with current observations . in the modeling
we have extrapolated the observed x - ray agn lf slope to brighter luminosities .
we find that if this slope is steepened at the bright end , we can reproduce the observed umbh mass function at @xmath1 for @xmath8010@xmath81 as well . in order to reconcile the observationally derived local black hole mass function at the high mass end , the slope @xmath82 in eqn .
( 1 ) needs to be modified .
we find that the slope @xmath82 for black hole masses @xmath83 is @xmath84 , which however , does not provide a good - fit for higher masses .
a slope steeper than @xmath85 is required to fit bh masses in excess of @xmath86 , we find that formally the best - fit is found in reduced-@xmath87 terms for the value of @xmath88 . *
such a steepening simulates the cut - off of a self - regulation mechanism that limits black hole masses and sets in at every epoch*. in fig . 3 ,
the results of such a self - limiting growth model are plotted .
the predicted abundance of umbhs is now in much better agreement with observations at @xmath1 and is consistent with the number of umbhs detected by sdss . in a self - regulated mass growth model
we predict the abundance of umbhs at @xmath1 to be @xmath89 .
therefore , requiring consistency between the x - ray and optical views of black hole growth and assembly with the observed number of umbhs at @xmath1 , points to the existence of a self - regulation mechanism that limits bh masses .
the self - regulation is implemented as a steepening of the x - ray agn lf at the luminous end ( the value of @xmath82 needed is plotted in fig . 4 ) and does not have an important effect on the x - ray background , since this change in slope only affects sources with x - ray luminosities @xmath90 greater than 10@xmath91 erg s@xmath92 , while most of the x - ray background emission is produced by sources with luminosities of 10@xmath93 erg s@xmath92 , as found by treister & urry ( 2005 ) .
note that the use of a bolometric luminosity function , as reported by hopkins , richards & hernquist ( 2007 ) does not match the slope at the high mass end ( shown as the dashed curve in fig .
one key consequence of our model is that the slope of the @xmath94 relation at the high mass end likely evolves with redshift .
recent cosmological simulations find evidence for such a trend ( di matteo et al 2008 ) .
the functional form of this expected variation will depend on the specific model of self - regulation employed .
below , we explore physical processes that are likely to regulate the growth of bhs in galactic nuclei .
m@xmath95 , as shown by the _ black dashed line_.,width=340,height=340 ] converting the high end of the local black hole mass function into the equivalent velocity dispersions of the host spheroids we find values in excess of 350 kms@xmath92 . in the context of the currently popular hierarchical model for the assembly of structure ,
the most massive galaxies in the universe are expected to be the central galaxies in clusters .
high-@xmath3 peaks in the density fluctuation field at early times seed clusters that assemble at later times , and hence these are the preferred locations for the formation of the most massive galaxies in a cold dark matter dominated universe . for the index @xmath82 in the x - ray agn lf required to match the high mass end of the local black hole mass density .
in order to match the observed relation , the slope of the hard x - ray luminosity function was modified for masses higher than @xmath86m@xmath95.,width=340,height=340 ]
while we predict above that a few , rare umbhs are likely to exist at the centers of the brightest central galaxies in clusters , we further argue that there likely exists an upper limit to black hole masses .
evidence for this is presented using several plausible physical scenarios that attempt to explain the coeval formation of the black hole and the stellar component in galactic nuclei .
clearly the existence of umbhs is intricately related to the highest mass galaxies that can form in the universe .
given that star formation and black hole fueling appear to be coupled ( e.g. di matteo et al . 2005 and references therein ; silk & rees 1998 ) , it is likely that there is a self - limiting growth cycle for bhs and therefore a physical upper limit to their masses . here
we present several distinct arguments that can be used to estimate the final masses of bhs ( haehnelt , natarajan & rees 1998 ; silk & rees 1998 ; murray , quataert & thompson 2004 and king 2005 ) .
these involve self - limiting growth due to a momentum - driven wind , self - limiting growth due to the radiation pressure of a momentum - driven wind , and from an energy - driven superwind model .
murray , quataert & thompson ( 2004 ) argue that the feedback from momentum driven winds , limits the stellar luminosity , which in turn regulates the bh mass .
they argue for eddington limited star formation with a maximum stellar luminosity , @xmath96 where , @xmath97 is the gas fraction in the halo and @xmath3 the velocity dispersion of the host galaxy .
star formation in this scheme is unlikely to evacuate the gas at small radius in the galactic nucleus , therefore , all the gas in the inner - most regions fuel the bh .
the growing bh itself clears out this nuclear region with its accretion luminosity approaches @xmath98 . at this point
the fuel supply to the bh is shut - off and this may shut off the star formation as well . the final bh mass is then given by , @xmath99 where @xmath100 is the electron scattering opacity .
for the most massive , nearby early - type galaxies at the very tail of the measured sdss velocity dispersion function with velocity dispersions of @xmath13 350 - 400 kms@xmath92 ( bernardi et al . 2005 ) this gives a final bh mass of @xmath101 .
therefore , normal galaxies with large velocity dispersions are the presumptive hosts for umbhs .
furthermore , there appears to be a strong indication of the existence of an upper mass limit for accreting black holes derived from sdss dr3 by vestergard et al .
( 2008 ) in every redshift bin from @xmath102 .
an alternative upper limit can be obtained when the emitted energy from the accreting bh back reacts with the accretion flow itself ( haehnelt , natarajan & rees 1998 ) .
the final shut - down of accretion will depend on whether the emitted energy can back - react on the accretion flow prior to fuel exhaustion .
this arguement provides a limit , @xmath103 where @xmath104 is the fraction of the accretion luminosity which is deposited as kinetic energy into the accretion flow ( cf .
silk & rees 1998 ) , @xmath105 is the spin parameter of the dm halo , @xmath106 is the specific angular momentum of the disk , @xmath107 is the disk mass fraction .
the back - reaction timescale will be related to the dynamical timescale of the outer parts of the disk and/or the core of the dm halo and should set the duration of the optically bright phase .
it is interesting to note here that the accretion rate will change from super - eddington to sub - eddington without much gain in mass if the back - reaction timescale is shorter than the salpeter time .
the overall emission efficiency is then determined by the value of @xmath108 when the back - reaction sets in and is reduced by a factor @xmath109 compared to accretion at below the eddington rate . by substituting the value of the velocity dispersion of nearby cd s @xmath110 , we obtain a limiting value of the mass , if we assume that the bulk of the mass growth occurs in the optically bright quasar phase . due to the dependence on the spin parameter @xmath105 of the dm halo ,
the desired umbh mass range can arise preferentially in high velocity dispersion halos with low spin . king ( 2005 ) presents a model that exploits the observed agn - starburst connection to couple black hole growth and star formation . as the black hole grows , an outflow drives a shell into the surrounding gas which stalls after a dynamical time - scale at a radius determined by the bh mass .
the gas trapped inside this bubble cools , forms stars and is recycled as accretion and outflow .
once the bh reaches a critical mass , this region attains a size such that the gas can no longer cool efficiently .
the resulting energy - driven flow expels the remaining gas as a superwind , thereby fixing the observed @xmath4 relation as well as the total stellar mass of the bulge at values in good agreement with current observations .
the limiting bh mass is given by : @xmath111 where @xmath97 is the gas fraction ( @xmath112 , @xmath113 the electron scattering opacity and @xmath3 the velocity dispersion .
this model argues that black hole growth inevitably produces starburts and ultimately a superwind .
note that both the murray , quataert & thompson ( 2004 ) model and the king ( 2005 ) model predict @xmath114 while the haehnelt et al .
( 1998 ) and silk & rees ( 1998 ) predict a @xmath115 dependence .
the current error bars on the observational mass estimates for black holes preclude discrimination between these two possibilities .
shutdown of star formation above a critical halo mass effected by the growing agn has also been proposed as a self - limiting mechanism to cap bh growth and simultaneously explain the dichotomy in galaxy properties ( croton et al .
2006 ; cattaneo et al .
umbhs are expected to be rare in the local universe , from our analysis of the x - ray luminosity function of agn , we predict an abundance ranging from @xmath13 few times @xmath116 .
these estimates are in good agreement with those obtained from optical quasars in the sdss dr3 by vestergard et al .
the results of the first attempts to detect and measure masses for umbhs is promising .
dalla bonta et al . ( 2007 ) selected 3 brightest cluster galaxies ( bcgs ) in abell 1836 , abell 2052 and abell 3565 . using acs ( advanced camera for surveys ) aboard the hubble space telescope and the imaging spectrograph ( stis ) , they obtained high resolution spectroscopy of the @xmath117 and @xmath118ii emission lines to measure the kinematics of the central ionized gas .
they present bh mass estimates for 2 of these bcgs , @xmath119 and @xmath120 and an upper limit for the bh mass on the third candidate of @xmath121 .
it is interesting to note that bernardi et al .
( 2005 ) in a census of the most massive galaxies in the sdss survey do find candidates with large velocity dispersions @xmath122 .
the largest systems they find are claimed to be extremes of the early - type galaxy population , as they have the largest velocity dispersions . these @xmath123 systems ( see table 1 of bernardi et al .
( 2006 ) for details on these candidates ) are not distant outliers from the fundamental plane and the mass - to - light scaling relations defined by the bulk of the early - type galaxy population .
clear outliers from these scaling relations tend to be objects in superposition for which they have evidence from spectra and images .
we argue that these extreme early - type galaxies might harbour umbhs and likely their abundance offers key constraints on the physics of galaxy formation .
although the observations are challenging , a more comprehensive and systematic survey of nearby bcgs is likely to yield our first local umbh before long . as discussed above
, candidates from the sdss are promising targets for observational follow - up as they are extremely luminous . utilizing the hubble space telescope
, the light profile might show evidence for the existence of an umbh in the center ( e.g. lauer et al .
in fact , for @xmath124 and @xmath125 , it may be possible to measure spatially resolved velocity dispersion profiles even from ground - based facilities .
the interplay between the evolution of bhs and the hierarchical build - up of galaxies appears as scaling relations between the masses of bhs and global properties of their hosts such as the bh mass vs. bulge velocity dispersion - the @xmath126 relation and the bh mass vs. bulge luminosity @xmath127 relation .
the low bh mass end of this relation has recently been probed by ferrarese et al .
( 2006 ) in an acs survey of the virgo cluster galaxies .
they find that galaxies brighter than @xmath128 host a supermassive central bh whereas fainter galaxies host a central nucleus , referred to as a central massive object ( cmo ) .
ferrarese et al .
report that a common @xmath129 relation leads smoothly down from the scaling relations observed for more more massive galaxies .
extrapolating observed scaling relations to higher bh masses to the umbh range , we predict that these are likely hosted by the massive , high luminosity , central galaxies in clusters with large velocity dispersions .
the velocity dispersion function of early - type galaxies measured from the sdss points to the existence of a high velocity dispersion tail with @xmath130 ( bernardi et al .
if the observed scaling relations extend to the higher mass end as well , these early - types are the most likely hosts for umbhs .
recent simulation work that follows the merger history of cluster scale dark matter halos and the growth of bhs hosted in them by yoo et al.(2007 ) also predict the existence of a rare population of local umbhs . however , theoretical arguments suggest that there may be an upper limit to the mass of a bh that can grow in a given galactic nucleus hosted in a dark matter halo of a given spin . clearly the issue of the existence of umbhs is intimately linked to the efficiency of galaxy formation and the formation of the largest , most luminous and massive galaxies in the universe .
possible explanations for the tight correlation observed between the velocity dispersion of the spheroid and black hole mass involve a range of self - regulated feedback prescriptions .
an estimate of the upper limits on the black hole mass that can assemble in the most massive spheroids can be derived for all these models and they all point to the existence of umbhs . in this paper , we have argued that while rare umbhs likely exist , there is nevertheless an upper limit of @xmath131 for the mass of bhs that inhabit galactic nuclei in the universe .
we first show that our current understanding of the accretion history and mass build up of black holes allows and implies the existence of umbhs locally .
this is primarily driven by new work that predicts the formation of massive black hole seeds at high redshift ( lodato & natarajan 2007 ) and their subsequent evolution ( volonteri , lodato & natarajan 2008 ) . starting with massive seeds and following their build - up through hierarchical merging in the context of structure formation in a cold dark matter dominated universe
, we show that a viable pathway to the formation of umbhs exists .
there is also compelling evidence from the observed evolution of x - ray agn for the existence of a local umbh population . convolving the observed x - ray lf s of agn , with a simple accretion model
, the mass function of black holes at @xmath1 is estimated .
mimic - ing the effect of self - regulation processes that impose an upper limit to bh masses and incorporating this into the x - ray agn lf we find that the observed umbh mass function at @xmath1 is reproduced .
this self - regulation limited growth is implemented by steepening the high luminosity end of the agn lf at the bright end .
we estimate the abundance of umbhs to be @xmath132 at @xmath1 .
the key prediction of our model is that the slope of the @xmath4 relation likely evolves with redshift at the high mass end .
probing this is observationally challenging at the present time but there are several bright , massive early - type galaxies that are promising host candidates from the sdss survey as well as a survey of bright central galaxies of nearby clusters .
observational detection of umbhs will provide key insights into the physics of galaxy formation and black hole assembly in the universe .
we thank steinn sigurdsson and meg urry for useful discussions .
abazajian , k. , et al . , 2003 , aj , 126 , 2081 alexander , d. , smail , i. , bauer , f. , chapman , s. , blain , a. , brandt , w. , & ivison , r. , 2005 , nature , 434 , 738 barger , a. j. , et al .
2003 , aj , 126 , 632 barger , a. j. , cowie , l. l. , mushotzky , r. f. , yang , y. , wang , w .- h . , steffen , a. t. , & capak , p. 2005 , aj , 129 , 578 beckmann , v. , gehrels , n. , shrader , c. r. , & soldi , s. 2006 , apj , 638 , 642 begelman , m. , & meier , d. l. , 1982 , apj , 253 , 873 begelman , m. , & nath , b. , 2005 , mnras , 361 , 1387 begelman , m. , volonteri , m. & rees , m. j. , 2006 , mnras , 370 , 289 bernardi , m. , et al .
, 2005 , aj , 129 , 61 bernardi , m. , et al . , 2006 , aj , 131 , 2018 bernardi , m. , sheth , r. , tundo , e. , & hyde , j. , 2007 , apj , 660 , 267 bland
hawthorn , j. , wilson .
& tully , r.b . , 1991 , apjl , 371 , l19 boyle , b. j. , shanks , t. , croom , s. m. , smith , r. j. , miller , l. , loaring , n. , & heymans , c. 2000 , mnras , 317 , 1014 carilli , c. , et am . 2002 ,
aj , 123 , 1838 cattaneo , a. , dekel , a. , devriendt , j. , guiderdoni , b. & blaizot , j. , 2006 , mnras , 370 , 1651 cox , p. , et al . 2002 , a & , 387 , 406 .
corsini , e. m. , beifiori , a. , dalla bonta , e. , pizzella , a. , coccato , l. , sarzi , m. , bertola , f. , 2006 , proceedings of `` black holes : from stars to galaxies '' , iau symp .
238 , eds .
v. karas & g. matt , cambridge univ . press , astro - ph/0610624 croton , d. , et al . , 2006 , mnras , 365 , 11 dalla bonta , e. , ferrarese , a. , corsini , e. , miralda - escude , j. , coccato , l. , & pizella , a. , 2007 , mmsai , 78 , 745 , astro - ph/07061959 di matteo , t. , esin , a. , fabian , a. c. , & narayan , r. , 1999 , mnras , 305 , l1 di matteo , t. , springel , v. , & hernquist , l. , 2005 , nature , 433 , 604 di matteo , t. , colberg , j. , springel , v. , hernquist , l. , sijacki , d. , 2008 , apj , 676 , 33 fan , x. , et al .
2001 , aj , 122 , 2833 fan , x. , et al .
2003 , aj , 125 , 1649 fan , x. , et al .
2004 , aj , 128 , 515 fan , x. , et al .
2006 , aj , 132 , 117 fabian , a. c. 1999 , mnras , 308 , l39 fabian , a. c. , & iwasawa , k. , 1999 , mnras , 303 , l34 merrit , d. & ferrarese , l. , 2001 , apj , 547 , 140 ferrarese , l. , et al . , 2006 ,
apj , 644 , 21 gebhardt , k. , et al .
2000 , apj , 539 , l13 gilli , r. , risaliti , g. , & salvati , m. , 1999 , a&a , 347 , 424 gilli , r. , comastri , a. , & hasinger , g. 2007 , a&a , 463 , 79 haehnelt , m.g . &
rees , m.j . , 1993 , mnras , 263 , 168 haehnelt , m.g . , natarajan , p. & rees , m.j .
, 1998 , mnras,300 , 817 hopkins , p. , hernquist , l. , cox , t. , di matteo , t. , martini , p. , robertson , b. , & springel , v. , 2005 , apj , 630 , 705 hopkins , p. , hernquist , l. , cox , t. , di matteo , t. , robertson , b. , & springel , v. , 2006a , apjs , 163 , 1 hopkins , p. , robertson , b. , krause , e. , hernquist , l. , & cox , t. , 2006b , apj , 652 , 107 hopkins , p. , narayan , r. , & hernquist , l. , 2006 , apj , 643 , 641 hopkins , p. ; richards , g. t. ; hernquist , l. , 2007 , apj , 654 , 731 kauffmann , g. & haehnelt , m. , 2000 , mnras , 311 , 576 king , a. , 2003 , apj , 596 , l27 king , a. , 2005 , apj , 635 , l121 lehto , h.j . & valtonen , m.j . , 1996 ,
apj , 460 , 207 lauer , t. , gebhardt , k. , & richstone , d. , et al . , 2002 , aj , 124 , 1975 lauer , t. , tremaine , s. , richstone , d. , & faber , s , m. , 2007a , apj , 670 , 249 lauer , t. , et al .
, 2007b , apj , 662 , 808 lodato , g. , & natarajan , p. , 2006 ,
mnras , 371 , l1813 lodato , g. , & natarajan , p. , 2007
, mnras , 377 , l84 magorrian , j. , et al .
1998 , aj , 115 , 2285 marconi , a. , risaliti , g. , gilli , r. , hunt , l. , maiolino , r. & salvati , m. , 2004 , mnras , 351 , 169 martinez - sansigre , a. , et al . , 2005 ,
nature , 436 , 666 martnez - sansigre , a. , et al .
2007 , mnras , 379 , l6 mclure , r. , & dunlop , j. , 2004 , mnras , 352 , 1390 merloni , a. , rudnick , g. & di matteo , t. , 2004 , mnras , 354 , l37 murray , n. , quataert , e. , & thompson , t. , 2004 , apj , 618 , 569 mushotzky , r. f. , cowie , l. l. , barger , a. j. , & arnaud , k. a. 2000 , nature , 404 , 459 natarajan , p. & sigurdsson , s. , 1998 , mnras , 302 , 288 oh , p. , & haiman , z. , 2002 , apj , 569 , 558 omont , a. , cox , p. , bertoldi , f. , mcmahon , r. g. , carilli , c. , & isaak , k. g. 2001 , a&a , 374 , 371 reuland , m. , rottgering , h. , van breugel , w. , & de breuck , c. , 2004,mnras , 353 , 377 risaliti , g. , maiolino , r. , salvati , m. , 1999 , apj , 522 , 157 salucci , p. , szuszkiewicz , e. , monaco , p. , & danese , l. , 1999 , mnras , 307 , 637 sazonov , s. yu . ,
ostriker , j. p. , ciotti , l. , & sunyaev , r. , 2005 , mnras , 358 , 168 shankar , f. , salucci , p. , granato , g. l. , de zotti , g. , & danese , l. 2004 , mnras , 354 , 1020 silk , j. & rees , m.j . , 1998 , a & a , 331 , l1 sillanp , a. , haarala , s. , valtonen , m.j . , sundelius , b. & byrd , g.g . , 1988 , apj , 325 , 628 sotan , a. , 1982 , mnras , 200 , 115 steffen , a. t. , strateva , i. , brandt , w. n. , alexander , d. m. , koekemoer , a. m. , lehmer , b. d. , schneider , d. p. , & vignali , c. 2006 , aj , 131 , 2826 treister , e. , & urry , c. m. 2005 , apj , 630 , 115 treister , e. , et al . 2006 , apj , 640 , 603 tremaine , s. , et al .
2002 , apj , 574 , 740 tundo , e. , bernardi , m. , hyde , r. , sheth , r. , & pizzella , a. , 2007 , apj , 663 , 53 ueda , y. , akiyama , m. , ohta , k. , & miyaji , t. 2003 , apj , 598 , 886 valtonen , m.j .
, lehto , h. & pietil , h. , 1999 , a & a , 342 , l29 van der marel , r.p . , 1994 , mnras , 270 , 271 vestergaard , m. , fan , x. , tremonti , c. r. , osmer , p. & richards , g. t. , 2008 , apj , 674 , l1 volonteri , m. , haardt , f. & madau , p. , 2003 ,
apj , 582 , 559 volonteri , m. & rees , m. j. , 2005 , apj , 633 , 624 volonteri , m. & rees , m. j. , 2006 , apj , 650 , 669 volonteri , m. , lodato , g. & natarajan , p. , 2008 ,
mnras , 383 , 1079 walter , f. , et al .
2003 , nature , 424 , 406 worsley , m. a. , et al .
2005 , mnras , 357 , 1281 wyithe , s. , & loeb , a. , 2002 , apj , 581 , 886 yoo , j. , miralda - escude , j. , weinberg , d. , zheng , z. , & morgan , c. , 2007 , apj , 667 , 813 , astro - ph/0702199 yu , q. , & tremaine , s. , 2002 , 335 , 965 | we make a case for the existence for ultra - massive black holes ( umbhs ) in the universe , but argue that there exists a likely upper limit to black hole masses of the order of @xmath0 .
we show that there are three strong lines of argument that predicate the existence of umbhs : ( i ) expected as a natural extension of the observed black hole mass bulge luminosity relation , when extrapolated to the bulge luminosities of bright central galaxies in clusters ; ( ii ) new predictions for the mass function of seed black holes at high redshifts predict that growth via accretion or merger - induced accretion inevitably leads to the existence of rare umbhs at late times ; ( iii ) the local mass function of black holes computed from the observed x - ray luminosity functions of active galactic nuclei predict the existence of a high mass tail in the black hole mass function at @xmath1 .
consistency between the optical and x - ray census of the local black hole mass function requires an upper limit to black hole masses .
this consistent picture also predicts that the slope of the @xmath2-@xmath3 relation will evolve with redshift at the high mass end .
models of self - regulation that explain the co - evolution of the stellar component and nuclear black holes naturally provide such an upper limit .
the combination of multi - wavelength constraints predicts the existence of umbhs and simultaneously provides an upper limit to their masses .
the typical hosts for these local umbhs are likely the bright , central cluster galaxies in the nearby universe .
galaxies : evolution , active , nuclei .
x - rays : galaxies . |
many problems in science and engineering require the sorting , or ordering , of large amounts of multivariate data . since there is no canonical linear criterion for sorting data in dimensions greater than one , many different methods for sorting
have been proposed to address various problems ( see , e.g. , @xcite ) .
many of these algorithms abandon the idea of a linear ordering , and instead sort the data into layers according to some set of criteria .
we consider here _ nondominated sorting _ , which arranges a set of points in euclidean space into layers by repeatedly removing the set of minimal elements .
let @xmath0 denote the coordinatewise partial order on @xmath1 defined by @xmath2 the _ first nondominated layer _ , also called the _ first pareto front _ and denoted @xmath3 , is exactly the set of minimal elements of @xmath4 with respect to @xmath0 , and the deeper fronts are defined recursively as follows : @xmath5 this peeling process eventually exhausts the entire set @xmath4 , and the result is a partition of @xmath4 based on pareto front index , which is often called _ pareto depth _ or _ rank_. figure [ fig : demo ] gives an illustration of nondominated sorting of a random set @xmath4 .
nondominated sorting is widely used in multi - objective optimization , where it is the basis of the popular and effective genetic and evolutionary algorithms @xcite . of course , multi - objective optimization is ubiquitous in engineering and scientific contexts , such as control theory and path planning @xcite , gene selection and ranking @xcite , data clustering @xcite , database systems @xcite , image processing and computer vision @xcite , and some recent machine learning problems @xcite .
nondominated sorting is also equivalent to the longest chain problem , which has a long history in probability and combinatorics @xcite .
a _ chain _ in @xmath1 is a finite sequence of points that is totally ordered with respect to @xmath0 .
let @xmath6 be @xmath7 distinct points in @xmath1 and define @xmath8),\ ] ] where @xmath9 denotes the length of a longest chain in the finite set @xmath10 .
the notation @xmath11 $ ] is a special case of the more general interval notation @xmath12 = \big\{y \in { \mathbb{r}}^d \ , : x \leqq y \leqq z\big\ } = \prod_{i=1}^d [ x_i , z_i]\ ] ] that we shall use throughout the paper .
let @xmath13 denote the pareto fronts obtained by applying nondominated sorting to @xmath14 .
then @xmath15 if and only if there are no other points @xmath16 with @xmath17 , i.e. , @xmath18 . a point @xmath19 is on the second pareto front @xmath20 if and only if all points @xmath16 satisfying @xmath17 are on the first front , and one such point exists . for any such @xmath21 ,
@xmath22 is a chain of length @xmath23 in @xmath24 $ ] , and we see that @xmath25 . peeling off successive pareto fronts and repeating this argument yields @xmath26 hence , the pareto fronts @xmath13 are embedded into the level sets ( or jump sets ) of the longest chain function @xmath27 , as depicted in figure [ fig : demo ] . in @xcite , we proved the following continuum limit for nondominated sorting .
[ thm : prev ] let @xmath28 be a sequence of i.i.d . random variables in @xmath1 with density @xmath29 , and define @xmath27 by .
suppose there exists an open and bounded set @xmath30 with lipschitz boundary such that @xmath29 is uniformly continuous on @xmath31 , and @xmath32 for @xmath33 .
then @xmath34 uniformly on @xmath35 with probability one , where @xmath36 is the unique nondecreasing is nondecreasing if @xmath37 is nondecreasing for all @xmath38 . ]
viscosity solution of the hamilton - jacobi equation @xmath39 and @xmath40 are the universal constants given in @xcite .
this result shows that nondominated sorting of massive datasets , which are common in big data applications @xcite , can be approximated by solving a partial differential equation ( pde ) . in @xcite
, we exploited this idea to propose a fast approximate nondominated sorting algorithm based on estimating @xmath29 from the data @xmath6 and solving ( p ) numerically .
the proof of theorem [ thm : prev ] , given in @xcite , is based on the following continuum variational problem : @xmath41 where the supremum is over all @xmath42 curves @xmath43
\to [ 0,\infty)^d$ ] that are monotone nondecreasing in all coordinates and satisfy @xmath44 .
this can be viewed as a type of stochastic homogenization of the longest chain problem , where the curves @xmath45 are continuum versions of chains .
the variational problem first appeared in two dimensions in @xcite .
the pde ( p ) arises as the hamilton - jacobi - bellman equation associated with @xcite .
we used the same ideas to prove a similar hamilton - jacobi equation continuum limit for the directed last passage percolation model in statistical physics , which also has an obvious discrete variational formulation @xcite .
it is possible to view theorem [ thm : prev ] in the context of stochastic homogenization of hamilton - jacobi equations .
indeed , @xmath27 can be interpreted as a discontinuous viscosity solution of the hamilton - jacobi equation @xmath46 since the right hand side of is highly singular , it is worth discussing in what sense we interpret @xmath27 to be a solution of .
the obvious approach is to mollify the right hand side to obtain a sequence @xmath47 of approximate solutions .
it is possible so show that as @xmath48 , the sequence @xmath47 converges pointwise to @xmath49 , where the constant @xmath50 depends on the mollification kernel used . intriguing
as this observation is , it is unclear whether can be used directly to prove theorem [ thm : prev ] .
there is a growing literature on stochastic homogenization of hamilton - jacobi equations .
many of the proofs are based on applying subadditive ergodic theory to a representation formula for the solution , like the hopf - lax formula , or more generally , a control theoretic interpretation @xcite .
these techniques are similar in spirit to our original proof of theorem [ thm : prev ] that was based on the variational representation of @xmath36 given in .
recently , there has been significant interest in stochastic homogenization techniques that do not require a variational interpretation @xcite . in this paper
we give an alternative proof of theorem [ thm : prev ] that uses a direct verification argument , and completely avoids using the variational interpretation of ( p ) .
the argument is based on a heuristic derivation of ( p ) that originally appeared in @xcite .
we review this argument , which is reminiscent of dimensional analysis , in section [ sec : informal ] .
we believe this proof is new in the stochastic homogenization literature , and it seems to be more robust than arguments based on variational principles .
we present the proof in the simplest setting in this paper , but we believe it can be substantially generalized . in future work , we plan to apply this proof technique to other stochastic growth models that do not stem from underlying variational principles .
to simplify the presentation , we model the data here using a poisson point process . given a nonnegative function @xmath51 , we denote by @xmath52 a poisson point process with intensity function @xmath29 .
this means that @xmath52 is a random at most countable subset of @xmath1 , and for every bounded borel set @xmath53 , the cardinality of @xmath54 , denoted @xmath55 , is a poisson random variable with mean @xmath56 .
furthermore , for disjoint @xmath57 and @xmath58 , the random variables @xmath55 and @xmath59 are independent .
it is worthwhile to mention that in the special case where @xmath60 , one way to construct @xmath61 for any @xmath62 is by setting @xmath63 where @xmath28 is an _
_ sequence of random variables with density @xmath29 and @xmath64 is a poisson random variable with mean @xmath7 . for more details on poisson point processes ,
we refer the reader to kingman s book @xcite .
set @xmath65 for @xmath66 , @xmath67 and @xmath68 we define @xmath69\right).\ ] ] in this paper , we give a direct verification argument proof of the following result .
[ thm : cont ] let @xmath66 such that ( p ) has a unique nondecreasing viscosity solution @xmath36 .
then with probability one @xmath70 in @xcite , we showed that there exists a unique nondecreasing viscosity solution of ( p ) whenever @xmath29 is uniformly continuous on an open and bounded set @xmath30 with lipschitz boundary , and @xmath32 for @xmath71 .
in @xcite we extended this uniqueness result to the case where @xmath29 is continuous on @xmath35 , which allows the support of @xmath29 to be non - compact .
we expect uniqueness to hold under weaker regularity conditions on @xmath29 .
figure [ fig : demo2 ] gives a visual illustration of theorem [ thm : cont ] . in the proof of theorem [ thm :
cont ] , we assume the family of poisson point processes @xmath72 is defined on a common probability space and we apply deterministic pde arguments to the individual realizations @xmath73).\ ] ] in the literature on the longest chain problem , it is common to take a construction for which @xmath74 is monotone nondecreasing for all realizations @xmath75 . indeed , this was necessary to obtain almost sure convergence in some early works ( see @xcite ) .
due to the sharp concentration of measure results established by frieze @xcite , bollobs and brightwell @xcite , and more recently by talagrand @xcite , the monotonicity property is no longer required for almost sure convergence .
thus , our proof of theorem [ thm : cont ] does not assume any particular construction of @xmath72 ( see theorem [ thm : complete ] ) .
we should mention , however , that if we do construct a family of poisson point processes @xmath76 on a common probability space in such a way that @xmath77 then the function @xmath78)$ ] is monotone nondecreasing as a function of @xmath79 for _ every _ realization .
it then follows directly from theorem [ thm : cont ] that @xmath80 with probability one as @xmath81 .
one way to ensure that is satisfied is to construct @xmath82 via a generalization of @xcite .
as motivation , let us give an informal derivation of the hamilton - jacobi pde continuum limit ( p ) .
this derivation was originally given in a slightly different form in @xcite , and our proof of theorem [ thm : cont ] is based on these heuristics .
suppose @xmath83 is continuous and that @xmath84 uniformly with probability one for some @xmath85 $ ] .
the argument below not only derives ( p ) , but also suggests the order of growth @xmath86 .
let @xmath87 .
since @xmath27 is nondecreasing , the uniform limit @xmath36 must be nondecreasing and so @xmath88 for all @xmath38 .
let us assume that @xmath89 for all @xmath38 . fix @xmath90 with @xmath91 and consider the quantity @xmath92 .
this is approximately the number of pareto fronts passing between @xmath93 and @xmath94 when @xmath7 is large .
when counting these fronts , we may restrict ourselves to poisson points falling in the set @xmath95 this is because any points in @xmath96 will be on previous pareto fronts and only points that are coordinatewise less than @xmath94 can influence the pareto rank of @xmath94 .
see figure [ fig : derivation ] for a depiction of this region and some quantities from the derivation .
since @xmath97 for all @xmath38 , and @xmath36 is @xmath42 , @xmath57 is well approximated for small @xmath98 by a simplex with orthogonal corner , and the points within @xmath57 are approximately uniformly distributed .
let @xmath99 denote the number of poisson points falling in @xmath57 .
since we can scale the simplex into a standard simplex while preserving the pareto fronts within @xmath57 , it is reasonable to conjecture that the number of pareto fronts within @xmath57 ( or the length of a longest chain in @xmath57 ) is asymptotic to @xmath100 for some constant @xmath101 , independent of @xmath93 . by the law of large numbers
, we have @xmath102 .
hence when @xmath103 is small @xmath104 where @xmath105 denotes the lebesgue measure of @xmath57 .
let @xmath106 denote the side lengths of the simplex @xmath57 .
then @xmath107 for a constant @xmath108 .
since @xmath109 lies approximately on the tangent plane to the level set @xmath110 , we see that @xmath111 this gives @xmath112 , and hence latexmath:[\[\label{eq : measurea } for small @xmath98 , we can combine and to obtain @xmath114 therefore @xmath115 where @xmath116 . since the left hand side is independent of @xmath117 , this suggests that @xmath86 and @xmath118 although this derivation is informal , it is straightforward and conveys the essence of the result . inspecting the argument
, we see that the main ingredients are an invariance with respect to scaling along coordinate axes , and some basic properties of the coordinatewise partial order .
it is intriguing that we can derive the order of growth @xmath119 and the continuum limit ( p ) from such basic information .
this seems analogous to dimensional analysis arguments in applied mathematics , which use ideas like self - similarity and scaling to derive fundamental laws for natural phenomena @xcite .
a fundamental feature of dimensional analysis is that , like our argument above , it usually gives no information about the constants of proportionality .
contrary to dimensional analysis , which is typically nonrigorous , we show in this paper that the argument above can be made rigorous using the framework of viscosity solutions . our new proof of theorem [ thm : cont ] is a rigorous version of the heuristic dimensional analysis argument given in section [ sec : informal ] .
the main ideas used in our proof draw some inspiration from the barles - souganidis framework for convergence of numerical schemes @xcite .
the barles - souganidis framework establishes convergence of any numerical scheme that is monotone , consistent , and stable , provided the pde admits a strong comparison principle . in our proof of theorem [ thm : cont ] , we view @xmath27 as a numerical approximation of the viscostiy solution of ( p ) , and use ideas that parallel monotonicity , stability and consistency to prove convergence to the viscosity solution of ( p ) .
the longest chain problem is monotone in the following sense : for any finite sets @xmath120 , @xmath121 when proving convergence of numerical schemes to viscosity solutions , monotonicity of the scheme is used to replace the numerical solution by a smooth test function for which consistency is trivial .
the monotonicity property is used for precisely the same purpose in the proof of theorem [ thm : cont ] , though consistency is non - trivial in this setting .
stability of numerical schemes for pde usually refers to a uniform bound on the numerical solutions that is independent of the grid resolution . here
, we use a type of asymptotic equicontinuity .
namely , in theorem [ thm : holder2 ] we show that a bound of the form @xmath122 with probability one .
the stability estimate and the monotonicity of @xmath27 can be combined with the arzel - ascoli theorem to show that for every realization @xmath75 in a probability one event @xmath123 , there exists a subsequence @xmath124 converging locally uniformly on @xmath35 to a locally hlder - continuous function @xmath125 .
the final step in the proof is to use consistency and monotonicity to identify @xmath125 as the unique nondecreasing viscosity solution of ( p ) .
our consistency result , theorem [ thm : con ] , is a statement on the asymptotic length of a longest chain in sets resembling the approximate simplex used in the heuristic derivation .
namely , for a smooth test function @xmath126 with @xmath127 for all @xmath38 , we prove asymptotics of the form @xmath128 as @xmath129 , where @xmath130 notice the set @xmath131 is similar to the set @xmath57 used in the heuristic derivation and defined in .
the main difference is that @xmath131 is defined using a smooth test function , and more importantly , it is defined by looking `` backward '' from the point @xmath93 , instead of looking `` forward '' as in .
this respects the natural domain of dependence of the nondominated sorting problem , and allows us to use the monotonicity of @xmath132 in the proof of theorem [ thm : cont ] .
this is analogous to the idea of an upwind numerical scheme for hamilton - jacobi equations .
let us give a brief history of the longest chain problem .
let @xmath6 be _ i.i.d . _ random variables uniformly distributed on the unit hypercube @xmath133^d$ ] , and let @xmath134 be the length of a longest chain .
hammersley @xcite was the first to study the asymptotics of @xmath135 , motivated by a connection to ulam s famous problem of finding the longest monotone subsequence in a random sequence of numbers @xcite .
hammersley showed that in dimension @xmath136 , @xmath137 in probability as @xmath129 , and he conjectured that @xmath138 . the value @xmath138 was later established in @xcite .
hammersley s proof is now a classic application of subadditive ergodic theory .
bollobs and winkler @xcite extended hammersley s result to dimensions @xmath139 showing that @xmath140 in probability .
they also established the bounds @xmath141 which by stirling s formula shows that @xmath142 . aside from @xmath143 ,
the exact values of @xmath144 are still unknown . when the random variables @xmath145 are defined on a common probability space in such a way that @xmath146 is monotone nondecreasing along sample paths , the stronger almost sure convergence can be obtained via an observation of kesten @xcite . here
, we consider the longest chain among poisson points . for @xmath147 let @xmath148^d).\ ] ] then @xmath149 has the same distribution as @xmath150^d).\ ] ] notice that the family @xmath151 is defined on a common probability space and @xmath152 is monotone nondecreasing along sample paths .
subadditive ergodic theory @xcite can be employed to show that @xmath153 since @xmath149 and @xmath154 merely have the same distribution , this argument only gives convergence in probability for @xmath149 .
since we have made no assumptions about how the poisson point process @xmath61 is to be constructed , the proof of our main result ( theorem [ thm : cont ] ) relies on establishing almost sure convergence for sequences of random variables that have the same distribution as @xmath149 , but may not necessarily satisfy the monotonicity condition . using the concentration of measure results provided by talagrand s isoperimetric theory
@xcite we can establish the following result .
[ thm : complete ] for @xmath155 , let @xmath156^d)$ ] .
let @xmath157 be an increasing sequence of positive real numbers such that @xmath158 then @xmath159 recall that a sequence of random variables @xmath160 _ converges completely _ to a random variable @xmath161 if @xmath162 by the borel - cantelli lemma , complete convergence implies almost sure convergence .
the advantage of complete convergence is that it depends only on the distributions of @xmath163 and @xmath161 , and not on any particular construction of the random variables .
we believe theorem [ thm : complete ] is likely well - known in the probability literature , but to the best of our knowledge it has not been published . for the sake of completeness , we have sketched the proof of theorem [ thm : complete ] in the appendix . in this section ,
we prove in theorem [ thm : holder2 ] a stability result for the longest chain problem .
we first have a preliminary lemma .
[ lem : holder ] let @xmath164 , and define @xmath27 by .
then for any @xmath165 @xmath166^d)}^\frac{1}{d}|x - y|^\frac{1}{d}\ ] ] with probability one , where @xmath167 .
we first show that @xmath168)}^\frac{1}{d}h^\frac{1}{d},\ ] ] for all @xmath169 and @xmath170 . without loss of generality
we may assume that @xmath171 . if @xmath172 for some @xmath173 , then @xmath174 , and hence @xmath175 with probability one .
thus , we may assume that @xmath176 for all @xmath173 .
let @xmath177 and let @xmath178 be a chain in @xmath179 $ ] of length @xmath180 .
we can split this chain into two chains , @xmath181 and @xmath182 , such that @xmath181 lies in the rectangle @xmath11 $ ] , and @xmath182 belongs exclusively to @xmath183 $ ] . see figure [ fig : stability ] for a depiction of the setup . by the definition of @xmath27 and @xmath182 from the proof of lemma [ lem : holder].,scaledwidth=60.0% ] @xmath184
let @xmath185 and set @xmath186 let @xmath187 be a poisson point process with intensity @xmath188 and let @xmath189 .
then @xmath190 is a poisson point process with intensity @xmath191 , where @xmath192 let @xmath193 . by scaling
, @xmath194 has the same distribution as @xmath195^d)$ ] . by theorem [ thm : complete ] and the inclusion @xmath196 we have @xmath197 with probability one .
combining this with establishes . we can apply in each coordinate and note that @xmath198 to establish the desired result .
recall that @xmath199 and @xmath200 denote the upper and lower semicontinuous envelopes of @xmath29 , respectively , defined by @xmath201 and @xmath202 .
[ thm : holder2 ] let @xmath164 , and define @xmath27 by . then with probability one @xmath203^d)}^\frac{1}{d}|x - y|^\frac{1}{d}\ ] ] for all @xmath165 , where @xmath167 .
it is important to emphasize in theorem [ thm : holder2 ] that holds with probability one _ simultaneously _ for all @xmath165 .
this is a far stronger statement than lemma [ lem : holder ] , where it was shown that for fixed @xmath165 , holds with probability one .
this stronger result gives us a form of compactness that is used in the proof of theorem [ thm : cont ] . for @xmath165
let @xmath204 denote the event that holds for @xmath205 .
by lemma [ lem : holder ] , @xmath206 .
let @xmath207 being the countable intersection of probability one events , @xmath123 has probability one .
let @xmath208 .
let @xmath209 such that @xmath210 and @xmath211 , and set @xmath212 then since @xmath27 is nondecreasing @xmath213 and therefore @xmath214^d)}^\frac{1}{d}|{\widehat{x}}-{\widehat{y}}|^\frac{1}{d } \
\ \text{for all } \omega \in \omega.\ ] ] since we can choose @xmath215 and @xmath216 arbitrarily close to @xmath93 and @xmath217 , respectively , we obtain @xmath218^d)}^\frac{1}{d}|x - y|^\frac{1}{d }
\ \ \text{for all } \omega \in \omega,\ ] ] which completes the proof .
we now establish our main consistency result . for @xmath219 , we define @xmath220 the set @xmath221 is a simplex with orthogonal corner at @xmath93 and side lengths @xmath222
. see figure [ fig : simplex ] for an illustration of @xmath221 .
we also define @xmath223 .
.,scaledwidth=50.0% ] [ lem : gen - simplex ] let @xmath219 and let @xmath157 be any increasing sequence of positive real numbers satisfying .
then @xmath224 in the sense of complete convergence .
we first show that @xmath225 note that we can write @xmath226^d \ , : \ , x_1 + \cdots + x_d \geq d-1\big\}.\ ] ] let @xmath227 be a positive integer and define @xmath228^d \ , : \ , kx \in { \mathbb{z}}^d \text { and } x_1 + \cdots + x_d = d-1-\frac{d}{k}\right\}.\ ] ] it follows from this construction that _ every _ chain in @xmath229 is contained entirely in @xmath230 $ ] for some @xmath231 . since @xmath232 is a finite set @xmath233\right ) \ , : \ , x \in i\big\}.\ ] ] fix @xmath231 and let @xmath234 .
since @xmath235)$ ] has the same distribution as @xmath236^d)$ ] we have by theorem [ thm : complete ] that @xmath237 ) = c_d\prod_{i=1}^d(1-x_i)^\frac{1}{d}\leq \frac{c_d}{d}\sum_{i=1}^d(1-x_i ) = c_d\left(\frac{1}{d } + \frac{1}{k}\right),\ ] ] in the sense of complete convergence . combining this with and noting that @xmath227 was arbitrary yields @xmath238 completely . for the other direction ,
let @xmath239 . by theorem [ thm :
complete ] @xmath240 ) = c_d \prod_{i=1}^d ( 1-z_i)^\frac{1}{d } = \frac{c_d}{d}.\ ] ] in the sense of complete convergence .
this establishes .
we now establish with a simple scaling argument .
let @xmath219 and define @xmath241 by @xmath242 since @xmath243 for all @xmath38 , the mapping @xmath244 preserves the partial order @xmath0 and therefore @xmath245 let @xmath190 be the poisson point process induced by the mapping @xmath244 , i.e. , @xmath246 . then @xmath190 is a poisson point process on @xmath1 with constant intensity @xmath247 . since @xmath248
, we have by that @xmath249 and @xmath250 have the same distribution .
the result now follows directly from .
[ lem : simplex - f ] let @xmath164 .
then for any @xmath219 @xmath251 and @xmath252 with probability one .
let @xmath219 , @xmath253 , and define @xmath254 let @xmath187 be a poisson point process with intensity @xmath188 and let @xmath189 . then @xmath190 is a poisson point process with intensity @xmath191 , where @xmath255 by lemma [ lem : gen - simplex ] @xmath256 with probability one , and hence @xmath257 with probability one . since @xmath258 , @xmath259 , which establishes . to prove , note first that if @xmath260 , then the result is trivial .
hence , we may take @xmath261 . for each point @xmath161 from @xmath61 that falls in @xmath221 , color @xmath161 red with probability @xmath262 .
the red points form another poisson point process on @xmath221 with intensity @xmath263 @xcite .
let us denote this poisson point process by @xmath264 .
by lemma [ lem : gen - simplex ] , we have that @xmath265 since @xmath266 , @xmath267 .
combining this with completes the proof . [
lem : simplex - f - r ] let @xmath164 .
then with probability one @xmath268 and @xmath269 for all @xmath270 .
let us prove ; the proof of is similar . for each @xmath219
let @xmath271 denote the event that holds for @xmath272 .
by lemma [ lem : simplex - f ] , @xmath273 .
let us set @xmath274 being the countable intersection of probability one events , @xmath123 has probability one .
let @xmath219 and let @xmath275 .
let @xmath276 such that @xmath277 for all @xmath38 , and let @xmath278 such that @xmath279 and @xmath280 we claim that @xmath281 .
to see this , let @xmath282 and write @xmath283 where we used in the last step that @xmath284 .
therefore @xmath285 .
it follows that @xmath286 , and from we deduce @xmath287 sending @xmath288 in such a way that @xmath289 and @xmath290 yields @xmath291 the result follows by noting that the limit above holds for all @xmath276 with @xmath292 for all @xmath38 .
we denote by @xmath293 the open ball of radius @xmath294 centered at @xmath67 .
we now prove our main consistency result .
[ thm : con ] let @xmath66 .
then with probability one @xmath295 and @xmath296 for all @xmath297 and all @xmath298 such that @xmath299 for all @xmath38 , where @xmath300 let @xmath123 be the event that and hold for all @xmath219 . by lemma [ lem : simplex - f - r ] , @xmath123 has probability one .
let @xmath297 , @xmath301 and let @xmath302 such that @xmath299 for all @xmath38 .
. we will give the proof of ; the proof of is very similar .
since @xmath303 , there exists @xmath304 such that @xmath305 for all @xmath306 , @xmath170 , and all @xmath301 sufficiently small .
it follows that for any @xmath306 @xmath307 we therefore deduce @xmath308 , there exists a constant @xmath50 such that @xmath309 for all @xmath306 and @xmath301 sufficiently small .
it follows that @xmath310 setting @xmath311 we have @xmath312 , and hence @xmath313 by @xmath314 sending @xmath315 completes the proof .
we now have the proof of theorem [ thm : cont ] .
let @xmath123 denote the event that the conclusions of theorem [ thm : holder2 ] and theorem [ thm : con ] hold , and @xmath316 on @xmath317 for all @xmath7 .
then @xmath318 . let us fix a realization @xmath275 .
the remainder of the proof is split into several steps .
we first use stability ( theorem [ thm : holder2 ] ) to obtain a compactness result .
since @xmath319 for all @xmath7 , it follows from theorem [ thm : holder2 ] that the sequence @xmath320 is bounded for all @xmath321 . by a diagonal argument
, there exists a subsequence @xmath322 such that for all @xmath323 , @xmath324 is a convergent sequence , whose limit we denote by @xmath325 . by theorem [ thm : holder2 ] @xmath326 ,
where @xmath167 .
hence , we can extend @xmath125 uniquely to a function @xmath327 such that for every @xmath328 , @xmath329^d)$ ] and @xmath330_{\frac{1}{d};[0,m]^d } \leq c_d d m^\frac{d-1}{d } \|f^*\|_{l^\infty([0,m]^d)}^\frac{1}{d}.\ ] ] furthermore , @xmath125 is nondecreasing and @xmath331 on @xmath317 .
we claim that @xmath332 locally uniformly on @xmath35 .
the proof of this is similar to ( * ? ? ?
* theorem 1 ) .
we include it here for completeness .
fix @xmath328 and let @xmath301 .
let @xmath333 and for any multi - index @xmath334 , let @xmath335 .
let @xmath232 be the set of multi - indices @xmath334 for which @xmath336^d$ ] . since @xmath125 is continuous on @xmath337^d$ ]
, we can choose @xmath338 large enough so that @xmath339 since @xmath232 is a finite set and @xmath340 for all @xmath341 , we deduce @xmath342 let @xmath343^d$ ] and let @xmath341 such that @xmath344 .
since @xmath345 and @xmath125 are nondecreasing we have @xmath346 similarly , we deduce @xmath347 combining these inequalities yields @xmath348^d ) } < \max_{\alpha \in i } \left| n^{-\frac{1}{d } } u_{n}(x_\alpha ) - u^\omega(x_\alpha ) \right| + { \varepsilon}.\ ] ] invoking we see that @xmath349^d ) } < { \varepsilon}.\ ] ] sending @xmath350 establishes the claim . \2 .
we now show that @xmath125 is a viscosity subsolution of ( p ) . for simplicity ,
let us set @xmath351 fix @xmath328 and let @xmath352 .
let @xmath298 such that @xmath353 has a local maximum at @xmath354 .
since @xmath355 is nondecreasing , @xmath356 for all @xmath38 .
if @xmath357 for some @xmath38 , then the subsolution property is trivially satisfied . hence , we may assume that @xmath299 for all @xmath38 .
without loss of generality , we may also assume that @xmath353 has a strict maximum at @xmath354 , relative to the set @xmath337^d$ ] .
then there exists a sequence @xmath358 in @xmath337^d$ ] converging to @xmath354 such that @xmath359 has a maximum at @xmath360 , relative to @xmath337^d$ ] .
let @xmath361 such that @xmath362 and @xmath363 for all @xmath364 and all @xmath38 .
let @xmath365 such that @xmath366 since @xmath367 uniformly on @xmath337^d$ ] and @xmath368 , there exists @xmath369 such that for all @xmath370 @xmath371 let @xmath372 \ , : \ ,
v_k(z ) \geq v_k(x_0 ) - { \varepsilon}\big\}.\ ] ] we claim that @xmath373 for all @xmath370 . to see this , let @xmath374 such that @xmath375 and note that @xmath376 since @xmath359 has a maximum at @xmath360 , @xmath377^d.\ ] ] combining this with , and we deduce @xmath378 for all @xmath379 with @xmath375 , and @xmath370 . since @xmath380 is nondecreasing , @xmath373 , establishing the claim .
we now claim that @xmath381 to see this , since @xmath382 is integer valued , we can write @xmath383 as @xmath384 \ , : \ , u_{n_k}^\omega(z ) \geq u_{n_k}^\omega ( x_0 ) - \lfloor { \varepsilon}n_k^\frac{1}{d}\rfloor\big\},\ ] ] where @xmath385 is the largest integer less than or equal to @xmath386 .
since @xmath387 and @xmath388 for @xmath389 , we must have that @xmath390 .
let @xmath391 and let @xmath392 be a chain in @xmath393 $ ] of length @xmath180 .
for any @xmath394 , @xmath395 , and hence @xmath396 . hence the chain
@xmath397 belongs to @xmath383 for @xmath398 and @xmath399 which establishes the claim . by and @xmath400 for all @xmath370 and @xmath401^d$ ] . since @xmath402 it follows that @xmath403 for @xmath370 , where @xmath404 is defined as in theorem [ thm : con ] .
invoking and the monotonicity of @xmath132 we have @xmath405 since theorem [ thm : con ] holds for all @xmath275 and @xmath303 , we deduce @xmath406 and hence @xmath407 \3 .
we now show that @xmath355 is a viscosity supersolution of ( p ) .
let @xmath298 such that @xmath353 has a local minimum at @xmath354 .
without loss of generality , we may assume that @xmath354 is a minimum of @xmath408 relative to the set @xmath337^d$ ] .
since @xmath355 is nondecreasing , @xmath356 for all @xmath38 . for @xmath409 , set @xmath410 then @xmath411 has a strict minimum at @xmath354 , relative to the set @xmath412 $ ] , and @xmath413 for all @xmath38 .
therefore , there exists a sequence @xmath358 in @xmath412 $ ] converging to @xmath354 such that @xmath414 has a minimum at @xmath360 , relative to @xmath412 $ ] .
let @xmath361 such that @xmath362 and @xmath415 for all @xmath364 and all @xmath38 .
let @xmath301 and let @xmath416 be a chain in @xmath417 . then by the definition of @xmath383 @xmath418 and therefore @xmath419 hence @xmath420 and therefore @xmath421 since @xmath414 has a minimum at @xmath360 @xmath422 for @xmath423 $ ] .
since @xmath368 and @xmath367 locally uniformly , we can choose @xmath424 larger , if necessary , so that @xmath425.\ ] ] it follows that @xmath426 where @xmath427 is as defined in theorem [ thm : con ] . invoking and the monotonicity of @xmath132 @xmath428 and therefore @xmath429 since theorem [ thm : con ] holds for all @xmath275 and @xmath413 @xmath430 thus , we arrive at @xmath431 since @xmath409 was arbitrary , @xmath355 is a viscosity supersolution of ( p ) .
\4 . by uniqueness of nondecreasing viscosity solutions of ( p ) , we have @xmath432 .
since we can apply the same argument to any subsequence of @xmath433 , and extract a further subsequence converging locally uniformly to @xmath36 , we find that @xmath434 locally uniformly on @xmath35 for all @xmath275 , where @xmath123 has probability one .
we sketch the proof of theorem [ thm : complete ] .
let @xmath301 .
let @xmath28 be a sequence of independent and uniformly distributed random variables on @xmath133^d$ ] and let @xmath134 be the length of a longest chain .
let @xmath64 denote the cardinality of @xmath435^d$ ] . conditioned on @xmath436 , @xmath149 and @xmath135
have the same distribution , so by the usual tail bounds on poisson random variables we deduce @xmath437 by @xcite there exist constants @xmath438 such that @xmath439 \leq c_2n^\frac{1}{d}\ ] ] for all @xmath440 , and @xmath441 = c_d>0.\ ] ] since the longest chain function @xmath132 is a _ configuration function _ ( see ( * ? ? ?
* definition 7.1.7 ) ) , we have @xmath442-\lambda\right),p\left(\ell_n \geq { \mathbb{m}}[\ell_n]+\lambda\right ) \leq 2\exp\left(-\frac{\lambda^2}{4({\mathbb{m}}[\ell_n ] + \lambda ) } \right),\ ] ] for any @xmath409 , where @xmath443 denotes any median of @xmath135 . a short computation involving and yields @xmath444 \to c_d$ ] as @xmath129 .
fix @xmath338 large enough so that @xmath445 - { \mathbb{e}}[\ell_k]| \leq c|n - k|^\frac{1}{d}.\ ] ] combining this with we deduce @xmath446 for all @xmath447 sufficiently large .
the other inequality is similar , and the result follows from .
the author is grateful to lawrence c. evans and fraydoun rezakhanlou for valuable discussions about this work .
fortin , s. grenier , and m. parizeau .
generalizing the improved run - time complexity algorithm for non - dominated sorting . in _ proceeding of the fifteenth annual conference on genetic and evolutionary computation _ , pages 615622 .
acm , 2013 .
d. kossmann , f. ramsak , s. rost , et al .
shooting stars in the sky : an online algorithm for skyline queries . in _ proceedings of the 28th international conference on very large data bases
_ , pages 275286 , 2002 . | nondominated sorting is a combinatorial algorithm that sorts points in euclidean space into layers according to a partial order .
it was recently shown that nondominated sorting of random points has a hamilton - jacobi equation continuum limit .
the original proof , given in @xcite , relies on a continuum variational problem . in this paper
, we give a new proof using a direct verification argument that completely avoids the variational interpretation .
we believe this proof is new in the homogenization literature , and may be generalized to apply to other stochastic homogenization problems for which there is no obvious underlying variational principle . |
einstein proposed the idea of teleparallelism , or _ fernparallelismus _ ( aka _ absolute parallelism _ , _ distant parallelism _ ) , in order to unify gravity and electromagnetism @xcite .
unlike general relativity , spacetime in teleparallelism is equipped with a connection with zero curvature , but non - vanishing torsion . since the spacetime is _ flat
_ , the parallel transport of a vector is independent of path , and hence the name _ _
tele__parallel , meaning `` parallel at a distance '' .
einstein s quest for unification via teleparallel theory can be found in the interesting account of @xcite . despite einstein s failure to formulate a unified theory ,
teleparallelism was later revived and studied as a pure theory of gravity .
the main motivations were ( 1 ) that the teleparallel formulation was considered to have advantages with regard to the identification of the energy - momentum of gravitating systems , and ( 2 ) teleparallel theory can be regarded as a gauge theory of local translations ( see , e.g. , @xcite ) . in fact einstein s general theory of relativity ( gr ) can be rewritten in teleparallel language ( see , e.g. , @xcite ) .
the result is a theory equivalent to gr known variously as the _ teleparallel equivalent of general relativity _ ( tegr ) or gr@xmath0 .
recently tegr has been generalized to @xmath1 theory , a theory of modified gravity formed in the same spirit as generalizing general relativity to @xmath2 gravity @xcite .
it seems however that observational data puts rather tight constraints on @xmath1 gravity , effectively reducing the theory ( at least for some choices of @xmath1 ) back to tegr ( however , see also @xcite and @xcite ) .
this includes constraints from considering a varying fine structure constant @xcite , the formation of large - scale structures @xcite and a varying gravitational constant @xcite , as well as solar system constraints @xcite .
this could hint at problems with @xmath1 gravity at the theoretical level .
in this paper we will briefly review tegr and more general theories of teleparallel gravity , including @xmath1 gravity , in section [ review ] . in section [ degree ] , we discuss the degrees of freedom for @xmath1 theory .
we know that there are extra degrees of freedom in the theory beyond those of gr ; that the number of extra degrees of freedom is generically 3 was clearly revealed by the dirac constraint analysis carried out by li et al .
despite this , in a careful linear perturbative analysis up to second order on flat friendmann - lematre - robertson - walker ( flrw ) background , no extra mode appears @xcite . in the case of minkowski background , this had already been pointed out in @xcite .
disappearance of extra degrees of freedom is known to happen in some gravity theories ( e.g. , in massive gravity , see @xcite ) ; this can be caused by the accidental disappearance of kinetic terms at the linear level .
if so , as we shall argue , it is likely to be a signal of superluminal propagation .
we will also discuss the related danger of the _ nonlinear constraint effect _ , which plagued some other theories of gravity in the past .
our conclusion is that @xmath1 theory is very _
likely _ to give rise to unwanted problems due to the effects of nonlinear constraints , in addition to propagation outside of the null cone . as a consequence ,
the number and type of degrees of freedom could vary as a physical system evolves .
this is further explored via the method of characteristics in section [ method ] .
the method of characteristics is properly introduced in section [ method ] , where we also provide an illuminating example in an application to the nonlinear proca field .
the same example is given a hamiltonian analysis that clearly shows the close connection of superluminal propagating modes and nonlinear constraints .
although the characteristic equations allow one to obtain superluminal propagating degrees of freedom , one must be careful to distinguish between _ physical _ degrees of freedom and _ gauge _ degrees of freedom , as the example of the maxwell field illustrates . in section
[ f(t ) ] , we investigate the dynamical equations in @xmath1 theory and derive a system of partial differential equations that describe the characteristics of the theory .
this calculation reveals nonlinear features that are very prone to allow superluminal propagation outside of the metric null cone .. ] furthermore , we also show that the constant - time hypersurface in flrw spacetime , and as a special case minkowski spacetime , can be a cauchy _ horizon _ , which means that even the infinitesimal future time evolution from it is not unique .
we conclude with some discussion in section [ discuss ] .
let @xmath3 be a semi - riemannian manifold of signature @xmath4 with metric @xmath5 .
the teleparallel theories we wish to discuss use the vierbein field or tetrad @xmath6 as the fundamental dynamical variables .
we use greek letters for spacetime coordinate indices and latin letters to label the tetrad vectors .
suppose that the manifold is _ parallelizable _ , i.e. , there exist @xmath7 vector fields @xmath8 such that at _ any _ point @xmath9 the tangent vectors @xmath10 s provide a basis of the tangent space at @xmath11 . note that this _ particular _ set of vector fields must be able to span the tangent spaces at all points of the manifold , i.e. , it is a set of everywhere linearly independent _ globally non - vanishing vector fields_. then we can view the mapping between the bases of the tangent space in a coordinate frame @xmath12 to that of a non - coordinate frame @xmath13 as an isomorphism @xmath14 .
orthonormality of the tetrad is thus imposed by the minkowski metric on the tangent space @xmath15 , related to the metric on the manifold @xmath3 by @xmath16 .
one then equips the manifold @xmath3 with the metric - compatible weitzenbck connection @xcite @xmath17 defined by @xmath18 where @xmath19 .
equivalently , the connection coefficients are @xmath20 in local coordinates , the components of the torsion tensor are @xmath21 the weitzenbck connection can be shown to be curvature - free . works . ]
the weitzenbck connection differs from the levi - civita connection by the _ _ contortion tensor _ _
@xmath22 the einstein - hilbert action of general relativity is ( in units where @xmath23 ) @xmath24 up to a total divergence ( which does not affect the field equations ) this action is equivalent to @xmath25 where @xmath26 , @xmath27 and @xmath28 is a certain quadratic - in - torsion scalar : @xmath29 the latter is the so - called _ torsion scalar_. in this teleparallel language , ( [ tegraction ] )
is called the _
tegr action_. @xmath1 gravity simply promotes the `` torsion scalar '' in the tegr lagrangian to a function , i.e. the action of @xmath1 gravity is is absorbed into @xmath30 , i.e. , tegr corresponds to @xmath31 . ]
@xmath32 one can then obtain the field equation for @xmath1 theory of modified gravity by varying with respect to the frame @xmath33 , this leads to @xmath34 we remark that the ( quadratic - in-)torsion scalar defined above is a very special one .
we note that mathematically we can define a more general torsion `` scalar '' @xmath35 by relaxing the coefficients : @xmath36 in particular one can construct the one - parameter combination @xmath37 = \frac{1 + 2\lambda}{4}\overset{w}t{}^\rho{}_{\eta\mu}\overset{w}t_\rho{}^{\eta\mu } + \frac{1 - 2\lambda}{2}\overset{w}t{}^\rho{}_{\mu\eta}\overset{w}t{}^{\eta\mu}{}_\rho - \overset{w}t_{\rho\mu}{}^\rho \overset{w}t{}^{\nu\mu}{}_\nu,\ ] ] where @xmath38 .
years ago a teleparallel theory using this as the lagrangian was proposed as an alternative to gr .
this theory passed all observational ( as of , perhaps up to , 1990 ) @xcite and some theoretical tests @xcite .
but then certain theoretical problems were found , which we will discuss in the next section .
we summarize how teleparallel theories work : we start with choosing a good local orthonormal frame @xmath13 which is _ declared _ to be covariantly constant under parallel transport by imposing the weitzenbck connection .
this defines a global frame field that sometimes goes by the name _ orthoteleparallel frame _ or ot frame .
recall that there are infinitely many possible choices of frames that span the same tangent space at @xmath9 , related to each other by elements of the proper orthochronous lorentz group .
the torsion tensor and consequently the torsion scalar , can thus be expressed in terms of any one of these infinitely many bases .
once we get a parallelization , it is defined up to _ global _ lorentz transformation .
this defines a weitzenbck geometry .
the set of all possible parallelizations of @xmath3 is then partitioned into equivalence classes consisting of parallelizations that relate to each other in the same class by a global lorentz transformation ( thus , one _ can not _ pass from one parallelizations to another by a local lorentz transformation ) .
note that at this stage , nonzero torsion _ does not _
mean that we have some `` distinguished frames '' or preferred frames ; any equivalence class of frames are on the same footing , in the sense that they parallelized the spacetime equally well .
however , once we introduce the physics via the action , we get field equations . _ now _ we can only admit certain classes of parallelizations , just as in general relativity where we can only admit metrics that satisfy the einstein field equations .
we remark that due to the lack of local lorentz invariance ( see also the discussions in @xcite and @xcite ) , unlike in general relativity where we can change coordinate systems and use any frame fields , this is not so in generic teleparallel theories . as a consequence ,
this implies that we can not directly extract the tetrad from the metric in the straightforward way .
tegr is special since its action does not determine the admissible ot frame but only the metric . from now onwards , we will suppress all the explicit overscript @xmath39 s on the torsion scalar and connection coefficients etc .
we are reminded of the problem faced by the one - parameter teleparallel theory in which the teleparallel lagrangian has the form ( [ optp ] ) .
this theory has 4 extra degrees of freedom @xcite .
kopczyski @xcite first pointed out that this theory has predictability problems ; he argued that one is unable to determine uniquely the evolution of the teleparallel geometry .
however , further analysis @xcite showed that the problem is not generic , instead it only occurs for a special class of solutions .
cheng et al .
verified this behavior using the hamiltonian approach and identified the troubles as a certain effect of nonlinear constraints , there referred to as _ constraint bifurcation _
@xcite , in which the chain of constraints could `` bifurcate '' depending on the values of the fields . that is to say ,
the number or type of constraints would depend on the values of the phase space variables .
one could imagine that the evolution of some system as depicted by a curve in phase space could pass through regions with a different number or type of constraints
. this would be rather strange ( since this means that the number of gauges and _ physical _ degrees of freedom also change ) , and one will find it hard , if at all possible , to predict the evolution of the system . to put it in another way , the matrix with the poisson brackets of the constraints as entries may not have constant rank , due to the nonlinearity of the constraints . the number or type of constraints may change as one approaches a point in the phase space where the rank changes .
the more general poincar gauge theory also has similar problems @xcite .
we suspect that this might also be the case for @xmath1 theory ( and probably for many teleparallel theories ) .
it is of course conceivable that with a specific choice of @xmath1 , we could prevent the solutions from ever approaching such problematic regions in phase space , e.g. , if they would act as a dynamical _
repeller_. the number of degrees of freedom of a generic @xmath1 theory of gravity ( in @xmath40-dimensions ) has been found to be five by miao li et al .
@xcite using a hamiltonian approach .
that is , there are three extra degrees of freedom , which the authors suggest could correspond to one massive vector field or one massless vector field with one scalar field . to understand the properties of the degrees of freedom in a theory
, one should first try to analyze them by a perturbative approach .
however , it could happen that not all degrees of freedom show up in the perturbation level . for the case of @xmath1 gravity ,
as remarked by li et al . at the end of their paper @xcite , for some special backgrounds ( including minkowski space with @xmath41 ) , some of the extra degrees of freedom do not appear in the _ linear _ perturbation .
it is also found that @xcite , on the flat flrw background with a scalar field , linear perturbation up to second order does not reveal _ any _ extra degree of freedom . in other words ,
@xmath1 theory is _ highly nonlinear _ : the number and type of constraints in the linear theory is _ different _ from that of the full , nonlinear theory .
this is also one of the problematic features discussed in the context of poincar gauge theory @xcite .
the worry of such behavior is that linearized modes which are `` good '' may cease to be so in the full nonlinear theory , perhaps accompanied by anomalous characteristics , as we will further comment on in the later sections . the revealing hamiltonian discussion of li et al . for @xmath1 gravity in ( 3 + 1 ) dimensional
spacetime is rather complicated and some of the results are only inferred implicitly .
however they also worked out _ explicitly _ the analogous but somewhat simpler hamiltonian formulation for @xmath1 gravity in @xmath42-dimensions ; for this case the problem we wish to address is clearly evident . in this case
there are six first class constraints @xmath43 , where @xmath44 and @xmath45 , as well as four second class constraints @xmath46 , which leads to two degrees of freedom .
they define the quantities & y_i = \{h_0 , ^i } , y_3 = \{h_0 , ^12 } , + & x_0 = \{h_0 , } , x_i = \{^i , } , x_3 = \{^i , } , and @xmath47\partial_m\phi.\end{aligned}\ ] ] the self - consistency equation in matrix form is shown to be @xmath48 , where @xmath49 and @xmath50 satisfies @xmath51 .
this matrix generically is of rank 4 .
however , as previously mentioned , some of the poisson brackets could vanish in some cases , which would result in rank changes of the matrix as the system evolves .
miao li et al . found the explicit formulas for the lagrange multipliers : @xmath52 @xmath53 @xmath54 @xmath55 while they were interested in the most general case and thus only considered the case with @xmath56 , this type of special case is _ precisely _ what we are interested in in this work .
indeed one sees that in principle @xmath57 can vanish , yet the numerators of these expressions are not generally constrained to vanish at the same time , which allows one or more of the lagrange multipliers to become unbounded . as we will explain below , this is signaling a superluminal propagation mode .
a viable theory of gravity should satisfy certain theoretical criteria , including the lack of tachyonic modes , or modes that carry negative kinetic energy ( i.e. , ghosts ) .. ] the theory should also support a well - posed initial value problem , that is , satisfy the cauchy - kowalevski theorem @xcite .
moreover any propagation mode in the theory should also be described by hyperbolic quasi - linear partial differential equations with well - behaved characteristics , i.e. , the characteristic surfaces should be non - spacelike .
one should also consider a theory with a `` good '' minkowski limit as preferable to a theory that does not have such behavior .
the method for studying characteristics is well - known ; see for example the detailed classical work of courant and hilbert @xcite , or chapter 8 of the more recent text of @xcite .
a discussion of the method of characteristics in the case of general relativity can be found in lecture 14 of @xcite . for more rigorous treatment
see @xcite , as well as in @xcite .
see also @xcite for some useful discussions in the context of external fields in gauge theories .
here we only give a brief summary .
recall that if @xmath58 is a linear differential operator of order @xmath59 , then we can consider @xmath58 as a polynomial in the derivative @xmath60 . in multi - index notation
we can write this as @xmath61 we may ask in what directions it is really of order @xmath59 . for an ode @xmath62 for example , it is obvious that @xmath63 is the condition required for the equation to be of order @xmath59 everywhere . for pdes with multiple variables ,
we need to be careful .
consider for example the laplacian @xmath64 in @xmath65 .
this is clearly of order 2 in both the @xmath66- and @xmath67-direction , however for an operator which is mixed , e.g. , @xmath68 , it is _ not _ of order 2 in either the @xmath66- or @xmath67-direction .
nevertheless , it _ is _ a second - order operator , which can be revealed by introducing new variables , @xmath69 , which renders @xmath70 .
that is , this operator is of order 2 in the @xmath71 and @xmath72 directions . in general then , given any operator @xmath58 at @xmath66 , and a direction @xmath73 , we can make an orthogonal change of variable so that @xmath73 points along one of the new coordinate axes , say the @xmath74-axis .
now , if the coefficient of the partial derivative @xmath75 ( in the new coordinate system ) is nonzero at @xmath66 , then we can say that @xmath58 is of order @xmath59 at @xmath66 in the direction @xmath73 .
we refer to such situation as _
noncharacteristic_. that is , _ characteristic _ refers to the case in which the coefficient of @xmath75 vanishes at @xmath66 .
the _ total symbol _ or _ top - order symbol _ of @xmath58 is simply a polynomial obtained by replacing the derivative @xmath60 with a variable , say @xmath76 . the _ principal symbol _
, denoted @xmath77 , is the highest degree component of the total symbol .
that is , @xmath78 the principal symbol almost completely determines the qualitative behavior of the solutions of the system .
furthermore , it is well known that for hyperbolic ( as well as parabolic ) partial differential equations , the zeros of the principal symbol describe the characteristics of the system .
that is to say , _
the characteristic directions are exactly those for which the principal symbol vanishes_. indeed the `` standard procedure '' is to begin with the equation of motion or the field equation , keep only the highest derivative terms , and then replace the said derivatives @xmath79 with components of the normal vector to the characteristics @xmath80 and set the equation to zero .
this is called the _
characteristic equation_. we can study how the characteristics propagate by looking for what sorts of vectors are allowed as solutions to the characteristic equation .
note that for a characteristic equation that is a _ matrix _ equation , called the _
characteristic matrix _
, it suffices to consider the vanishing of the determinant of the characteristic equation ( known as the _ characteristic determinant _ ) instead of the vanishing of the equation itself ( the latter is of course a stronger statement ) . in this paper , we are interested in one ( or a few ) special characteristic direction that gives the signal of a superluminal mode . by analyzing the characteristic determinant
, we can examine all directions at once .
the _ characteristic surface _ which is orthogonal to the characteristic direction coincides with the edge of cauchy development , i.e. , the cauchy horizon .
this is because the higher order derivative term with respect to the characteristic direction disappears .
the disappearance of the higher order derivative term causes the evolution to be singular .
thus , the cauchy development of the characteristic surface is only the characteristic surface itself and the evolution from it is meaningless even if it is spacelike . note that for application in theories of gravity , we would like our characteristic directions to be null or spacelike , as a timelike characteristic direction is the signal of superluminal propagation ( and could violate causality , although this is not necessarily so even in theories with local lorentz invariance @xcite ) .
correspondingly , there should _ not _ be any non - trivial solution that corresponds to any _ timelike _ vector .
the characteristic method is best illustrated via an example or two .
consider the lagrangian of the proca field with a nonlinear term @xcite : @xmath81 where @xmath82 and @xmath83 the signature is @xmath4 .
the field equations are @xmath84 upon taking the divergence of this equation , because the first term vanishes by the antisymmetry of @xmath85 , we obtain an implicit constraint satisfied by the system : @xmath86 for @xmath87 , this gives @xmath88 , which upon substituting back into the field equation ( [ nlprocaeq ] ) gives @xmath89 which is the klein - gordon equation .
the characteristic equation for this case is just @xmath90 , i.e. , the characteristic is null . to find the characteristics in the general @xmath91 case ,
replace the highest derivative terms @xmath92 by @xmath93 . here
@xmath94 means it is not the value of the vector @xmath95 , but represents the change of the vector in a certain direction .
we then get the relations @xmath96 a linear combination of these gives @xmath97k\cdot \tilde a=0.\label{anomolous}\ ] ] now from eq .
( [ procafechar ] ) one can see that modes with @xmath98 propagate with null characteristics , but from eq .
( [ anomolous ] ) one can see that modes with @xmath99 generally have non - null characteristics .
the normal to the characteristic surface could be timelike , indicating a superluminal characteristic . in that case
there is a lorentz frame in which @xmath100 .
in such a frame the condition is @xmath101 we want to emphasize that this tachyonic mode of the nonlinear proca field can be detected using the hamiltonian formulation , as we now explain . from the lagrangian ,
the canonical momenta are @xmath102 that is , explicitly , @xmath103 and @xmath104 ; the latter is a _
primary constraint_. the hamiltonian density , constructed according to the dirac - bergmann constraint procedure @xcite , is @xcite & ^i_i + f^ijf_ij - a_0 _ c ^c + m^2(a^ia_i - a_0 ^ 2 ) + & + ( a^ia_i - a_0 ^ 2)^2+u^0 , where the primary constraint @xmath105 '' denotes dirac s weak equality , i.e. , it only holds on the constraint surface within the phase space . ] has been included with an unknown lagrange multiplier @xmath106 . from the hamiltonian evolution equation @xmath107one
finds the meaning of the multiplier , it is the missing `` velocity '' . preserving the primary constraint leads to the secondary constraint : @xmath108 the poisson bracket of the two constraints is @xmath109\delta^3(x - y).\ ] ] generically this is non - vanishing , so they make up a 2nd class pair
. however , there is an important exception when the rhs vanishes , _ this is exactly the same as the anomalous characteristic condition _ ( [ procachar ] ) .
the dynamical consequence shows up when we require preservation of the @xmath110 constraint , which is the relation that determines the `` unknown multiplier '' @xmath106 : @xmath111 + \mathcal{g},\ ] ] where @xmath112 denotes a collection of field dependent terms which , generically , are non - vanishing .
consequently , the field `` velocity '' @xmath113 , given by the lagrange multiplier , @xmath106 , becomes unbounded at any point(s ) where @xmath114 approaches 0 .
that is , @xmath115 this is an indication that , with respect to this constant time spacelike hypersurface , there is instantaneous propagation of the @xmath116 mode .
the nonlinear constraint has led to a field - dependent constraint poisson bracket value , a signal for the occurrence of superluminal propagation , in complete agreement with the previous analysis obtained from the characteristic equation .
one issue that we have to be careful with is the possible presence of gauge degrees of freedom in a theory ( which was _ not _ present in the previous nonlinear proca field example ) . in order to appreciate the issue
, first we shall look at the simple example of scalar fields .
we consider the lagrangian for the single scalar field defined by @xmath117 we know that the scalar field @xmath118 propagates in the null direction . in terms of characteristics
, we can understand it as follows .
the equation of motion is @xmath119 so the characteristic equation for @xmath118 is @xmath120 , which means the characteristic is null .
let us introduce another scalar field @xmath121 which does not appear in the lagrangian .
trivially , the lagrangian does not change under the transformation @xmath122 .
this property is a certain kind of gauge transformation .
the equation of motion which comes from the variation of the lagrangian ( [ lags ] ) with respect to @xmath121 is trivial , i.e. , @xmath123 .
thus the characteristic equation for @xmath118 and @xmath121 becomes @xmath124 , which is trivially satisfied for any @xmath125 .
the triviality comes from the gauge mode @xmath121 .
since @xmath121 does not appear in the lagrangian , we should not include it in the discussion of characteristics .
the situation for any gauge field theory is similar . for simplicity
, we consider the maxwell theory of the electromagnetic field ; the lagrangian is @xmath126 where @xmath127 . the equation of motion is obtained to be @xmath128 the principal symbol yields the characteristic equation @xmath129 where we want to seek some nontrivial @xmath130 .
if @xmath131 is null then @xmath132 gives a propagating mode on the null cone .
however , if @xmath131 is not null , then the characteristic equation can be satisfied by _ any _
@xmath133 for arbitrary constant @xmath134 .
in particular @xmath131 can be chosen to be timelike , which means that the characteristic direction will become timelike .
this superluminal mode however , as we shall see , actually corresponds to gauge mode and so it is not physical .
alternatively we can consider the determinant method . if we write down the characteristic equation without removing the gauge degree of freedom
, we will obtain characteristic matrix that satisfies @xmath135 = 0 . \label{eommax}\end{aligned}\ ] ] the left - hand side of this equation is @xmath136 for any @xmath137 , and so this equation becomes trivial .
the origin of the triviality is the gauge degree of freedom . if we operate @xmath138 to the left - hand of eq .
( [ eommax2 ] ) , it becomes algebraically @xmath136 .
therefore , eq . ( [ eommax2 ] ) has the information of only three independent equations .
we know the maxwell field has @xmath139-gauge , i.e. , the lagrangian is invariant under the @xmath139-gauge transformation @xmath140 .
since the gauge degree of freedom @xmath141 does not appear in the lagrangian , the variation of the lagrangian with respect to the gauge degree of freedom @xmath141 must be trivial , which is the origin of the above trivial equation . according to the above simple example of scalar fields
, we should eliminate the contribution from gauge degrees of freedom in the discussion of characteristics .
subtracting the contribution from the gauge degree of freedom , i.e. , considering the @xmath142 matrix whose basis can be anything independentof @xmath137 , we can obtain the exact characteristic equation , which is satisfied only when @xmath125 is null . a more rigorous treatment of the removal of gauge degrees of freedom and the characteristics of maxwell equations can be found in @xcite .
in this section , we discuss the cauchy problem for @xmath1 gravity by an analysis of the characteristics .
first , we shall determine the characteristic equations of @xmath1 gravity .
after that , we will show that a constant - time hypersurface of the flrw metric does not provide a good initial condition and that the time evolution of the flrw metric is not well behaved .
finally , we concretely show the problematic solution where the time evolution from a flrw initial condition is not unique . from eq .
( [ vfe ] ) , we see that the vacuum field equation in @xmath1 gravity is explicitly given by 0&=f_t - ee_a^f + & + ee_a^s_^ ( _ t)f_tt .
[ ftt ] this can be rewritten as @xmath143\partial_\mu \partial_{\alpha } e^b_{~\beta } - ee_a{}^\lambda t^\rho_{~\mu\lambda}s_\rho^{~\nu\mu}f_t - e\frac{1}{4}e_a{}^\nu f.\ ] ] the highest order derivative terms are just @xmath144\partial_\mu \partial_\alpha e^b_{~\beta}.\ ] ] we recall that @xmath145 is _ linear _ in @xmath146 , so we may write @xmath147 for some @xmath148 which is _ algebraic _ in the tetrad @xmath149 . the explicit form of @xmath148 is @xmath150 \nu } -\frac{1}{2}e_b^{\ [ \mu } g^{\nu ] [ \alpha}e_a^{\ \beta ] } \nonumber \\ & + & e_a^{\ [ \mu } g^{\nu ] [ \alpha}e_b^{\ \beta ] } , \end{aligned}\ ] ] where the bracket @xmath151 $ ] means antisymmetrization , i.e. , @xmath152\cdots } : = \frac{1}{2 } \left ( t^{a\cdots bc \cdots } -t^{a\cdots cb \cdots}\right).\end{aligned}\ ] ] consequently we have @xmath153 thus , we can re - write eq .
( [ highest ] ) in a more simplified form as @xmath154\partial_\mu\partial_\alpha e^b_{\beta}.\ ] ] following the standard procedure we can replace @xmath155 in eq .
( [ char ] ) to obtain the _
characteristic equations for @xmath1 theory of gravity _
@xmath156k_\mu k_\alpha \tilde{e}^b_{~\beta } = 0;\ ] ] where @xmath157 here the notation @xmath158 indicates that these are not the values of the frame , but rather represent the change of the frame in a certain direction .
we note that eq .
( [ chareqn ] ) thus provide @xmath159 quadratic polynomial conditions on the four @xmath137 s .
that is , the characteristic matrix becomes @xmath160k_\mu k_\alpha.\end{aligned}\ ] ] since the tetrad @xmath161 always has the inverse @xmath162 , the property of the matrix @xmath163 is the same as that of @xmath164k_\mu k_\alpha.\label{charm}\end{aligned}\ ] ] therefore for convenience , we shall analyze the matrix in eq .
( [ charm ] ) instead . before discussing the property of the characteristic matrix of eq .
( [ charm ] ) , we must first find out the gauge degrees of freedom .
we have 16 equations of motion ( [ vfe ] ) , but not all of them are independent due to gauge degrees of freedom .
we actually have four nontrivial identities .
indeed , let us consider the invariance of the action under an infinitesimal coordinate transformation : @xmath165 this leads to the transformation of the tetrad @xmath161 by @xmath166 and we know this is related to a gauge mode , i.e. , @xmath167 , where @xmath168 are four scalar fields with an internal index .
varying the action with respect to the component @xmath169 one obtains the nontrivial noether differential identities @xcite @xmath170 where @xmath171 is defined in eq .
( [ vfe ] ) .
we stress that eq .
( [ identity ] ) is an _ identity _
, i.e. , it is _ algebraically satisfied _ without using the equation of motion eq .
( [ vfe ] ) .
since the infinitesimal coordinate transformation also leads to a transformation of metric , the metric has the information about all the gauge degrees of freedom .
therefore , all the degrees of freedom of the gauge modes can be eliminated by gauge fixing of the metric components .
now , we move on to the discussion of the characteristic determinant which is constructed by the @xmath172 part of the characteristic matrix in eq .
( [ charm ] ) having removed the gauge degrees of freedom . here
, we analyze the characteristic matrix @xmath173 by the decomposition into symmetric and antisymmetric parts with respect to @xmath174 and @xmath91 and also with respect to @xmath175 and @xmath176 . without loss of generality we can choose an appropriate basis to write down the @xmath177 matrix such that the left half @xmath178 submatrix represents the symmetric part and the right half @xmath178 submatrix represents the antisymmetric part .
first , we consider the first term of the characteristic matrix in eq .
( [ charm ] ) .
the first term of the symmetrized characteristic matrix is well behaved because it is the only contribution in tegr .
the symmetric part of the characteristic equation is related to the metric component , which _ a priori _ has 10 degrees of freedom .
however we know that 4 of the degrees of freedom are gauge modes and we have only 6 physical modes .
we claim that its rank is generically 6 , and indeed after removing the gauge degrees of freedom the second order derivatives are associated with all components of the symmetrized tetrad .
we consider next the antisymmetrized component with respect to @xmath174 and @xmath91 .
the first term then gives zero contribution , which means that the rank of the first term is indeed @xmath179 for general @xmath137 .
to summarize , the characteristic matrix of @xmath1 gravity , after removing the gauge degrees of freedom , can be written as a @xmath177 matrix , with four @xmath180 block submatrices .
the generic rank of the upper left block is 6 , corresponding to the symmetrized tetrads .
next , we consider the antisymmetrized components of the second term in eq .
( [ charm ] ) .
the form of the second term is the product of two @xmath181 and its rank is one ( since one of the eigenvalue vanishes ) .
one can thus form a @xmath182 matrix at the very center of the @xmath177 matrix , in which each element belongs to one of the four blocks , respectively . together with the @xmath180 block corresponding to the symmetrized tetrad
, we thus get a @xmath183 matrix .
therefore we have found that the ranks of the first term and the second term of @xmath173 are 6 and 1 , respectively . by the fact that @xmath184\mu } k_\mu$ ] is generically nonzero
, we can prove that the rank of matrix @xmath173 is generically 7 .
this rests on the fact that the determinant of the @xmath183 matrix is generically nonzero , which follows from elementary linear algebra ( see appendix.[appendix ] ) .
it is important to note that the @xmath182 matrix at the center , being of rank 1 , corresponds to only _
one _ of the extra degrees of freedom in @xmath1 gravity .
however we know that the theory contains 3 extra degrees of freedom
so where are the remaining two ?
in fact the five antisymmetrized components corresponding to the tetrad @xmath185 ( @xmath186 ) , which satisfy @xmath187\mu } e_{a\lambda } \partial_\mu e^{(i)a}{}_\nu = 0,\end{aligned}\ ] ] become zero .
this means that in the equations for the five tetrads @xmath185 the second order derivative does not appear and that the characteristic equation should be at most a first order differential equation . in other words ,
the remaining two simply _ do not _ show up in the characteristic matrix !
thus , in order to obtain the complete characteristic we must derive the characteristic equation for @xmath185 separately .
this is what we will do next for the sake of completeness , although it is not a crucial part for our subsequent analysis ( since it is sufficient to show that the one degree of freedom that _ does _ show up in the characteristic matrix , leads to problematic behavior ) . in order to derive the characteristic equation for @xmath185 , we must go back to the equation of motion eq .
( [ vfe ] ) , which can be written as @xmath188 if we consider equation of motion from the variation of the action with respect to @xmath185 , we can easily show that the first term in eq .
( [ eqanti ] ) _ does not appear_. the equation of motion ( [ eqanti ] ) is obtained from variation of the action @xmath189\nonumber\\ & = & - \int d^4x \bigl [ \left ( \partial_\mu \delta e^a_{~\nu } \right)e s_a^{~\mu\nu}f_t \nonumber\\ & & \qquad + \delta e^a_{~\nu}\left(\frac{1}{4}e e_a^{~\nu}f+e e_a^{~\lambda}t^\rho_{~\mu\lambda}s_\rho^{~\nu\mu } f_t\right)\bigr],\end{aligned}\ ] ] and for @xmath190 the first term of the right - hand side becomes zero due to eq .
( [ 5part ] ) . the second term in eq .
( [ eqanti ] ) is also zero because of antisymmetrization .
eventually , only the last term in eq .
( [ eqanti ] ) gives a non - zero contribution .
although the last term in eq .
( [ eqanti ] ) seems to be a higher order equation of the first order derivative of @xmath185 ( which , if true , will render the characteristic method unworkable ) , we can show that it is actually linear .
indeed , we can multiply it by @xmath191 for convenience .
we then have the antisymmetrized equation with respect to @xmath192 and @xmath174 @xmath193 where @xmath194 this gives the two branches of the equation , @xmath195 and @xmath196 if we choose the first branch , i.e. , @xmath195 , the five equations become degenerate , and thus , time evolution can not be fixed uniquely .
this causes trouble regarding the well - posedness of the cauchy problem .
therefore , we choose the second branch . gravity . ] the second order terms of the first derivative in eq .
( [ antie ] ) can be written as @xmath197}e_b^{~\rho}+g^{\gamma[\alpha}e_b^{~\nu]}e_a^{~\rho}\right)\partial_{[\beta}e^a_{\rho]}\partial_{[\mu}e^b_{~\gamma]}=0 .
\label{2nd}\end{aligned}\ ] ] if both derivative factors are variations of @xmath185 , eq . ( [ 2nd ] ) becomes trivial .
thus at least one of the derivative term is _ not _ such a variation .
this means that eq .
( [ antie ] ) is not second order but first order with respect to @xmath198 and the characteristic equation for @xmath185 can be shown to be & ( g^^[t^]^ + g^^[t^]^ + t^[][g^ ] + t^g^ + & + t^[g^][g^ ] ) e_a k_e^(i)a_=0 , + [ charaanti ] where the notation @xmath199 is similar to @xmath200 , i.e. , it also represents the change of the frame in a certain direction .
( [ charaanti ] ) gives the characteristic matrix for @xmath185 : @xmath201\mu}}{{}^{\lambda } } \contraction { \biggl ( \frac{1}{4}g{}^{\beta}{}^{[\alpha}t^{\nu]\mu}{}^{\lambda } + g}{{}^{\alpha}}{{}^{[\beta}t^{\lambda]\mu}}{{}^{\nu } } \biggl ( \frac{1}{4}g{}^{\beta}{}^{[\alpha}t^{\nu]\mu}{}^{\lambda } + g{}^{\alpha}{}^{[\beta}t^{\lambda]\mu}{}^{\nu } \nonumber \\ & + & \frac{1}{2}t^{[\alpha\nu][\beta}g^{\lambda]\mu } + \frac{1}{2}t^\mu g^{\alpha[\lambda}g^{\beta]\nu}+\frac{1}{2}t^{[\beta}g^{\lambda][\nu}g^{\alpha]\mu}\biggr ) k_\mu .
\nonumber\\ { } \end{aligned}\ ] ] then , we have the exact characteristic matrix which is a combination of the @xmath202 matrix @xmath203 and the @xmath204 matrix @xmath205 . and the other 7 do not contribute to the characteristic matrix , since their components have only one derivative .
since the nonzero components of the @xmath202 matrix @xmath203 have two derivatives , the contribution of the cross terms with any smaller number of derivatives should not appear in the characteristic equation . ] to summarize , the remaining two extra degrees of freedom do not show up in the characteristic matrix , since their corresponding characteristic equation only contains _
one _ derivative term .
they can nevertheless be described by another @xmath206 characteristic matrix @xmath205 .
we remark that despite the single derivative term here , this _ does not _ necessary mean that the _ dynamics _ of the two extra degrees of freedom is governed by first order equations
. this may be analogous to the dirac equation , which is first order and linear .
however , together with its conjugate equation , one can derive a _ second order _
klein - gordon equation governing the dynamics of the system .
likewise , maxwell s equations are first - order but combining them gives the second order wave equation . in the next section
we will give concrete examples to demonstrate problematic behaviors that could arise in @xmath1 gravity due to bad characteristics . to appreciate what is likely to go wrong with the characteristic ,
let us look at the characteristic equation eq .
( [ chareqn ] ) again .
we see that the first term is good since it is the only term present in tegr , and we know that characteristics are null in general relativity .
in fact the first term gives dynamics to the 6 metrical components of @xmath158 .
the second term quadratic in the field strength @xmath207 governing _ one _ of the extra degrees of freedom however is almost certainly going to give a disaster if @xmath208 for timelike @xmath131 .
it suggests that with some field values the characteristics could leak outside the metric null cone , which in turn means that there might be problems with tachyonic modes .
this nonlinear behavior is reminiscent of the nonlinearity we encountered in the generalized proca field example .
note that the case for tegr is very different since the theory is locally lorentz invariant , which allows us to choose a gauge that simplifies a lot of the calculation . in any case
, tegr corresponds to @xmath209 for any value of @xmath210 and there is no problem with the characteristics . in this section , we will show that in @xmath1 gravity the time direction , i.e. , @xmath211 , in the flrw metric is the characteristic direction , and thus the time evolution of the flrw metric is not unique _ even in the infinitesimal future_. we analyze the characteristic equation in @xmath1 gravity which has been derived in the previous subsection .
the characteristic determinant can be decomposed into two parts , the @xmath202 matrix @xmath203 and the @xmath204 matrix @xmath205 . in our analysis
we look into the @xmath202 matrix @xmath203 , which has the information about all of the six symmetric components and one of the antisymmetric components .
the antisymmetric component appears only in the second term of the characteristic matrix in eq .
( [ charm ] ) , and thus for some @xmath131 , the determinant becomes zero if the value of its antisymmetric component is zero .
this means that @xmath131 is one of the characteristic directions . to see this ,
it is enough to solve @xmath212\mu}k^\mu=0 $ ] for @xmath131 . in the case of the flrw metric
, @xmath213 is a solution .
let us see this explicitly .
we consider the cotetrads @xmath214 these cotetrads give the flrw geometry with flat spatial section : @xmath215 given the cotetrads , we can calculate the torsion tensor @xmath216 and @xmath217 .
they are @xmath218 while the other components all vanish . here
@xmath219 is the hubble parameter , i.e. , @xmath220 .
then , it is trivial that @xmath213 is a solution of @xmath184\mu}k_\mu=0 $ ] .
we have shown that the flrw solution has a spacelike characteristic hypersurface , and thus we can not properly discuss the time evolution from a constant time hypersurface even into the infinitesimal future . here , we show a concrete solution where the time evolution from the flrw initial condition without torsion is _ not unique _ in @xmath1 gravity with or without matter . we introduce a matter action @xmath221
then , equation of motion ( [ vfe ] ) is modified as @xmath222 assume that the matter action is locally lorentz invariant , i.e. , it depends on the tetrad only through the metric .
then eq .
( [ mattereom ] ) can be rewritten as @xmath223 let us now consider the ansatz @xmath224 which also gives the flrw metric ( [ flrw ] ) . for @xmath225 ,
this ansatz is equal to eq .
( [ flrwtetrad ] ) .
we note that this is a good choice of tetrad despite the lack of local lorentz invariance in the theory .
in particular the vector fields are everywhere smooth and linearly independent and thus provide a parallelization of flrw spacetime .
this choice of tetrad gives a parallelization that belongs to a different equivalence class of parallelizations than the usual diagonal tetrad of flrw spacetime ( see the penultimate paragraph of section [ review ] ) , but it is nevertheless a solution to the field equation , as we will show below . note that of course with the choice @xmath226 , our analysis below is in particular applicable to minkowski spacetime .
the torsion tensor @xmath227 and @xmath228 are given by @xmath229 where the other components are zero .
here @xmath230 . from this
one readily finds that convention .
the torsion scalar is @xmath231 if one works in the often used convention @xmath232 in the literature of @xmath1 cosmology . ]
@xmath233 substituting the above frame @xmath161 , metric , torsion tensor , torsion scalar and @xmath228 into the equation of motion ( [ vfe2 ] ) , we can find that @xmath234 does not appear in eq .
( [ vfe2 ] ) .
this means that if ansatz ( [ flrwtetrad ] ) ( i.e. @xmath225 ) originally satisfies eq .
( [ vfe2 ] ) , for _ any _ form of @xmath234 the ansatz ( [ ansatz ] ) is indeed an _
exact _ solution of @xmath1 gravity .
now , we can choose the form of the function @xmath234 such that @xmath235 for @xmath236 and @xmath237 for @xmath238 .
in this solution , on the spacelike hypersurface @xmath239 torsion suddenly emerges seemingly from nothingness , and the value of @xmath237 for @xmath238 is arbitrary .
this means that even in the infinitesimal future time evolution is not uniquely determined by the initial data , in other words the cauchy problem is not well - defined .
note that this problem does not arise in the case of tegr due to local lorentz invariance of the theory .
to conclude , in this work we have derived the set of partial differential equations that govern the characteristics of @xmath1 theory of modified gravity , from which we see that there will generally be a real danger of superluminal propagation that leaks outside of the metric null cone unless @xmath209 for any value of @xmath210 , which is the case of tegr .
we also took a closer look at the degrees of freedom in the theory .
we find that @xmath1 gravity is a highly nonlinear theory in which many things could go wrong .
first of all , as commented by li et al . the linearized theory has a _ different _ number of degrees of freedom than the full theory , which has three extra degrees of freedom compared to tegr ( and thus gr ) .
such nonlinear behavior was further confirmed by careful linear perturbation on a flat flrw background @xcite , in which _ none _ of the three extra degrees of freedom appear even in the second order linear perturbation . in this work
we note that the characteristic equation of @xmath1 gravity , i.e. , eq .
( [ chareqn ] ) contains a highly nonlinear second term .
indeed , we see that generically it excites the six pieces of @xmath158 , essentially @xmath240}$].}$ ] but also @xmath241 can be dynamical .
] of course , a more careful analysis by li et al
. showed that there are only 3 extra modes instead of 6 .
regardless , we see that the antisymmetric modes _ are _ excited in a highly nonlinear fashion that is very prone to causing troubles .
in fact , these extra modes are certain to have a very singular weak field limit , since the linearized theory has only the metric degrees of freedom .
the number of degrees of freedom and the number of constraints thus seem to be likely to depend on the amplitude of the field .
as the tetrad fields approach some specific values , some of the poisson brackets are approaching zero , which in turn means that their corresponding `` velocity '' lagrange multipliers tend to become unbounded , which signals the occurrence of _ instantaneous _ ,
i.e. , faster - than - light propagation , much like the situation we have exhibited for the nonlinear proca field ( see also the similar discussion in @xcite . ) .
moreover these extra dynamic parts are likely to propagate outside the metric null cone and are also likely to allow for negative energy propagating waves . by further analyzing the characteristic matrix of @xmath1 gravity
, we have shown that there is indeed a _
physical _ superluminal mode that could arise from one of the three extra degrees of freedom . closely related to this issue
, we have also demonstrated that the cauchy problem is ill - posed in flat flrw spacetime and even in minkowski spacetime .
we feel that this is a very bad property for any theory of gravity .
this echoes the remark ( lesson 4 ) in @xcite that extra degrees of freedom in modified gravities are prone to give rise to complications that are very hard to control .
although that remark was made under the context of @xmath2 gravity , it apparently also holds true for @xmath1 gravity .
we recall that @xmath1 gravity was often thought of as being more well - behaved and easier to deal with compared to @xmath2 gravity , since its equation of motion is of second order instead of fourth order .
however it now seems that this advantage comes with many trade - offs .
for example , unlike its @xmath2 counterpart , @xmath1 gravity is not locally lorentz invariant , and thus it is much more difficult to impose an ansatz for a solution _ a priori _ except in the simplest situations , such as for a flat spacetime .
furthermore , as we pointed out in this work , many problems that do not arise in @xmath2 theory could arise in @xmath1 gravity due to extra degrees of freedom and nonlinearity of the constraints .
we recall that in @xcite , it is pointed out that the extra degrees of freedom do not appear at the linear perturbation level on a flrw background , and that this behavior is similar to that of nonlinear massive gravity @xcite in which , while nonlinear analysis shows that there are generically five gravitational degrees of freedom @xcite , in the second order action on open flrw background there are only two tensor degrees of freedom @xcite .
since the hidden degrees of freedom in fact cause nonlinear instability in the case of nonlinear massive gravity @xcite , it is a concern that @xmath1 theory could exhibit a similar pathology .
here we would like to point out another similarity between nonlinear massive gravity and @xmath1 gravity : they both probably exhibit superluminal propagation modes . the case for nonlinear massive gravity
was studied in @xcite , in which the authors investigated the effect of the helicity-0 mode of the theory and found that energy can probably be emitted superluminously on the self - accelerating background . in the case of the ghost - free wess - zumino
massive gravity @xcite , deser and waldron recently showed that it also exhibits superluminal propagation modes @xcite . here
we should also make some comments on the possibility of _ negative energy solutions _ in @xmath1 gravity . in @xcite it was shown , directly in terms of the teleparallel variables , that tegr only permits positive energy . in general relativity , this is a well - known theorem proved initially by schoen and yau @xcite and subsequently by witten @xcite . in the tegr formulation
the argument was that one can arrange for the hamiltonian to be dominated by positive terms . a key step in the argument
was using the local frame gauge freedom to remove or at least control the non - positive terms in the tegr hamiltonian .
but in @xmath1 gravity ( and teleparallel theory in general ) one does _ not _ have local frame gauge freedom .
therefore there is no way to kill or control some of the non - positive terms . in the positive energy test developed in @xcite ,
it is argued that while it is hard to prove that the energy for some theory is positive for every solution to the initial value constraints , it is not so hard to show that a negative energy solution exists .
indeed it is sufficient to show that the initial value constraints admit a non - trivial zero energy solution . if so the theory should be discarded . in principle
this is a very strong test of a theory ( i.e. , it could exclude many theories which pass other tests ) .
unfortunately in practice it was not so easy to find _ even one _ bad solution explicitly , so it is generally hard work getting results beyond those of the linearized theory . in any case , to us
it seems likely that @xmath1 gravity would be very vulnerable to having a nontrivial non - positive energy solution for most choices of @xmath30 , although it is not expected to be very easy to show this .
although our work is purely classical , we would also like to make a brief comment regarding strong coupling problem once we consider _ quantizing _ the theory : naively , the lack of extra degrees of freedom at the linear level is caused by accidental disappearance of certain kinetic terms at the linear level .
vanishing of kinetic terms at the linear level corresponds to the small limit of their coefficients , which in turn means the nonlinear coupling terms of the canonically normalized modes are becoming very strong .
consequently the perturbative approach can not be applied and the theory is out of control at the quantum level .
then there is no reason to trust the classical effective action any more @xcite .
this is expected to be a problem if one attempts to naively quantize @xmath1 gravity .
it is also worth mentioning that existence of ghosts in a classical theory of gravity may be acceptable if the ghost can be pushed to the planck scale , to be dealt with by quantum theory of gravity @xcite .
finally , it is worth emphasizing that due to the fact that @xmath1 theory is highly nonlinear , a full rigorous analysis is at best a very difficult task .
one could have a viable theory if one manages to find a particular choice of @xmath30 that could avoid the problems raised in this work .
in other words , these seemingly serious `` problems '' could in fact be a blessing in disguise , since they may provide a guide for narrowing down the viable forms of @xmath30 .
we conclude by the following remark : we agree that non - linear theories merit investigation , but , as far as we can see , @xmath1 gravity does not seem to have the right kind of non - linearity ; its non - linearity may be too simple . as commented in @xcite ,
it is possible that this could be an indication that an _ even more nonlinear theory _ is required .
recall that the linearized spin-2 theory of gravity has problems that are only cured by nonlinearities of the full theory of general relativity ( see , e.g. , route 5 in box 17.2 in mtw @xcite , or feynman s lectures on gravitation @xcite ) .
furthermore , in fierz - pauli massive gravity @xcite , the vaishtein mechanism @xcite ( introduced to avoid the vdvz @xcite discontinuity ) excites the boulware - deser ghost @xcite due to non - linearity .
however , by introducing additional non - linearity , this ghost mode can be killed ( this is the so - called " nonlinear massive gravity ) @xcite . not surprisingly , this newly introduced non - linearity seems to give rise to other problems @xcite . while inherently non - linear theories certainly merit serious consideration ; finding one that is free of problems , in particular of the kind we have discussed , seems not so easy .
yen chin ong would like to thank brett mcinnes and wu - hsing huang for fruitful discussions about various aspects of differential geometry .
yen chin ong and james nester would also like to thank friedrich hehl for providing comments on the early version of the draft .
keisuke izumi is supported by taiwan national science council ( tnsc ) under project no .
nsc101 - 2811-m-002 - 103 .
pisin chen is supported by tnsc under project no .
nsc 97 - 2112-m-002 - 026-my3 , by taiwan s national center for theoretical sciences ( ncts ) , and by the us department of energy under contract no .
de - ac03 - 76sf00515 .
james m. nester is also supported by tnsc under project no .
100 - 2119-m-008 - 018 and 101 - 2112-m-008 - 006 and in part by ncts .
yen chin ong is supported by the taiwan scholarship from taiwan s ministry of education .
let us write @xmath242 to denote a matrix @xmath3 , while @xmath243 denotes the matrix element at the @xmath244 row and @xmath245 column of @xmath3 .
let @xmath242 be an @xmath246 matrix with non - vanishing determinant .
let @xmath247 be a @xmath182 matrix with vanishing determinant .
let @xmath248 be the @xmath249 matrix constructed from @xmath3 and @xmath250 , such that the first term is @xmath260,\ ] ] where @xmath261 is the remaining terms in the laplace expansion .
the second term is @xmath262_{n \times n}=-a_{21}a_{12}\det{(m)_{(n-1)\times(n-1)}}.\ ] ] vanishing of determinant of @xmath250 means that @xmath263 .
therefore we see that eq .
( [ 1 ] ) simply gives it immediately follows that if @xmath264 , then @xmath265 since @xmath266 .
this establishes the claim in section [ characteristic matrix ] , that the determinant for the @xmath183 matrix is generically nonzero .
f. w. hehl , j. d. mccrea , e. w. mielke , y. neeman , _ metric affine gauge theory of gravity : field equations , noether identities , world spinors , and breaking of dilation invariance _ ,
* 258 * ( 1995 ) 1 , http://arxiv.org/abs/gr-qc/9403012[[gr-qc/9402012 ] ] . v. c. de andrade , l. c. t. guillen , j. g. pereira , _ gravitational energy momentum density in teleparallel gravity _ ,
* 84 * ( 2000 ) 4533 , http://arxiv.org/abs/gr-qc/0003100[[gr-qc/0003100 ] ] .
l. l. so , j. m. nester , _ on source coupling and the teleparallel equivalent to gr _ , proceedings of the tenth marcel grossman meeting on general relativity 2003 , p.1498 , http://arxiv.org/abs/gr-qc/0612062v1[[gr-qc/0612062v1 ] ] .
p. baekler , f. w. hehl , _ beyond einstein - cartan gravity : quadratic torsion and curvature invariants with even and odd parity including all boundary terms _ , class .
* 28 * ( 2011 ) 215017 , http://arxiv.org/abs/1105.3504[[1105.3504 [ gr - qc ] ] ] .
d. zwanziger , _ method of characteristics in the external field problem or how to recognize an acausal equation when you see one _ , invariant wave equations , lecture notes in physics * 73 * ( 1978 ) 143 , springer berlin / heidelberg .
bruneton , _ on causality and superluminal behavior in classical field theories : applications to k - essence theories and mond - like theories of gravity _ , phys .
d * 75 * ( 2007 ) 085013 , http://arxiv.org/abs/gr-qc/0607055[[gr-qc/0607055v2 ] ] .
n. afshordi , d. j. h. chung , g. geshnizjani , _ cuscuton : a causal field theory with an infinite speed of sound _
d * 75 * ( 2007 ) 083513 , http://arxiv.org/abs/hep-th/0609150v2[[hep-th/0609150v2 ] ] .
a. e. gumrukcuoglu , c. s. lin , s. mukohyama , _ anisotropic friedmann - robertson - walker universe from nonlinear massive gravity _ , phys .
b * 717 * ( 2012 ) 295 , http://arxiv.org/abs/1206.2723[[1206.2723v2 [ hep - th ] ] ] .
a. de felice , a. e. gumrukcuoglu , s. mukohyama , _ massive gravity : nonlinear instability of the homogeneous and isotropic universe , _ phys .
* 109 * ( 2012 ) 171101 , http://arxiv.org/abs/1206.2080[[1206.2080v2 [ hep - th ] ] ] . n. arkani - hamed , h. georgi , m. d. schwartz , _ effective field theory for massive gravitons and gravity in theory space _
, annals phys .
* 305 * ( 2003 ) 96 , http://arxiv.org/abs/hep-th/0210184[[hep-th/0210184 ] ] .
p. chen , _ gauge theory of gravity with de sitter symmetry as a solution to the cosmological constant problem and the dark energy puzzle _ , mod .
lett . a * 25 * ( 2010 ) 2795 , http://arxiv.org/abs/1002.4275v2[[1002.4275 [ gr - qc ] ] ] . | teleparallel theories of gravity have a long history .
they include a special case referred to as the _ teleparallel equivalent of general relativity _
( tegr , aka gr@xmath0 ) .
recently this theory has been generalized to @xmath1 gravity .
tight constraints from observations suggest that @xmath1 gravity is not as robust as initially hoped .
this might hint at hitherto undiscovered problems at the theoretical level . in this work ,
we point out that a generic @xmath1 theory can be expected to have certain problems including superluminal propagating modes , the presence of which can be revealed by using the characteristic equations that govern the dynamics in @xmath1 gravity and/or the hamiltonian structure of the theory via dirac constraint analysis .
we use several examples from simpler gauge field theories to explain how such superluminal modes could arise .
we also point out problems with the cauchy development of a constant time hypersurface in flrw spacetime in @xmath1 gravity . the time evolution from a flrw ( and as a special case , minkowski spacetime ) initial condition is not unique . |
artificial neural networks are now becoming a more popular tools for handling astronomical data .
it is very important to have some automatic means of analyzing large databases like the surveys and upcoming space missions which would release terabytes of data to the community .
that is why , anns are widely used as an automatic tools for analyzing astronomical data .
some of the previous attempts on anns in astronomy are : @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
artificial neural network techniques used in the field of astronomy have been mostly supervised algorithms . here
we introduce an application of unsupervised technique to identify different classes of objects and try to make a cluster of similar type of stars using this technique .
the @xcite library is used for classification using som algorithm as unsupervised ann .
this algorithm configures output into a topological presentation of the original multi - dimensional data , producing a som in which input vectors with similar features are mapped to the same map unit or nearby units . in this way at the output map we will have the clusters of similar objects at different positions of the map .
the expected clusters in this case are the stellar spectral type ranging from o to m. the self- organizing map algorithm is explained in section 2 . in section 3
we describe the input data and their preprocessing .
the result of classification and discussion are presented in sections 4 and 5 respectively .
self - organizing map neural net developed by @xcite is important from many point of view as it pays attention to spatial order unlike the traditional neural models and that is why this technique acquired a very special position in neural network theory .
the following section is the description of this technique in details .
neural network learning is not restricted to supervised learning , wherein training pairs are provided , that is , input and target output pairs .
a second major type of learning for neural network is unsupervised learning , in which the net seeks to find patterns or regularities in the input data .
self - organizing map ( som ) , developed by kohonen , groups the input data clusters , a common use for unsupervised learning .
the adaptive resonance theory networks , are also clustering type networks .
is the assigned weight vector of the unit on the map ] self - organizing map consists of two layers : a one dimensional input layer and a two dimensional competitive layer , organized as a 2d grid of units as shown in the figure [ somnetwork ] .
som lattice structure composed of @xmath0 neural units .
this layer can neither be called hidden nor an output layer .
each input is connected to all output neurons in the map .
a weight vector with the same dimensionality as the input vectors is attached to every neuron in the map .
the learning algorithm for the som accomplishes two important things : 1 . clustering the input data 2 .
spatial ordering of the map so that input patterns tend to produce a response in units that are close to each other in the grid . since every unit in the competitive layer represents a cluster , the number of clusters that can be formed , will be limited by the number of units in the competitive layer . the weight vector for a competitive layer unit in a clustering net serves as a representative , exemplar , or codebook vector for the input patterns , which the net has placed on that cluster .
while training , the net determines the output units that is the best match for the current input vector ; then the weight vector of the winner is adjusted according to net s learning algorithm .
the beauty of the self - organizing maps is their ability to find regularities and correlations in the input layer and group them into vectors without any external adjustment or prior knowledge of the expected outcomes .
kohonen s som is one of the most referenced mapping techniques because of its ability to flatten high dimensional input into two or three dimensional data .
one of the most important attribute of kohonen s som is that , it does data compression without loss to relative distance between data points .
the neurons are connected to adjacent neuron by a neighborhood relation , which decides the topology , or structure , of the map .
the figure [ lattice ] shows the rectangular and hexagonal lattice structures and their discrete neighborhoods ( size 0,1 and 2 ) of the centermost unit .
the som is trained iteratively . in each training step
one vector @xmath1 from the input data set is randomly chosen and then the distance between this vector and all the weight vectors of the map will be calculated and this distance measurement , typically is an euclidian distance .
the neuron in the map whose weight vector is closest to the input vector , @xmath1 , is called best - matching unit ( bmu ) , here denoted by c : @xmath2 after finding the bmu , the weight vector of the map are updated such that the bmu is moved closer to the input vector in the input space .
the topological neighbors of the bmu are treated similarly .
this process of adaptation of weight vectors intern moves the bmu and its topological neighbors towards the input sample vector @xmath1 .
the updating scheme aims at performing a stronger weight adaptation at the bmu location than in its neighborhood .
this update rule in som for the weight vector of _ i _ is : @xmath3 , \ ] ] where @xmath4 presents the time , @xmath5 is randomly chosen an input vector from the input data set at time @xmath4 , @xmath6 is the neighborhood kernel around the @xmath7 unit @xmath8 and @xmath9 is the learning rate at time @xmath4 .
the function @xmath6 has a very important role ; it is defined over the lattice points . to achieve convergence
it is necessary that @xmath10 as t@xmath11 .
classically , a gaussian function is used , leading to : @xmath12 here , the euclidian norm is chosen and @xmath13 is the 2d location for the @xmath14 neuron in the network . @xmath15 species the width of the neighborhood during time @xmath4 .
the learning rate @xmath16 must always be selected much less than 1 , typically at most 0.4 .
best results are usually achieved by slowly decreasing it as training progresses , the radius of the neighborhood around a cluster unit ( winner ) also decreases as the clustering process progresses , this decreasing of the learning rate and radius of neighborhood must be monotonic for better performance of the network .
up to a point , large learning rate will indeed make training process faster .
if the learning rate selected to be very large then convergence may never occur , and in this case weight vector may oscillate highly and it can lead to instability .
the @xcite library is used as a data set for the classification and input to the som network .
the jhc library covers the wavelength range of 3510 - 7427.2 for various o to m stellar type and luminosity classes v , iii , and i. this library contains 161 spectra of individual stars , we selected 158 of them as an input to the som classifier .
the jhc library has a resolution of 4.5 with one sample per 1.4 .
each spectrum of this library is having 2799 number of data points and they are normalized to the unity before feeding into the network .
the som technique is actually mapping 2799 dimensional data into two dimensional output map .
the 158 input spectra are labled as the jhc s atlas list , specifying their spectral , subspectral and luminosity types .
these lables are require while analysing the output map .
the lables are placed beside every node ( neuron ) on the output map which indicates , that specific neuron is activated by the corresponding input spectra .
hence we can point out the location of each input spectra in the map and study the behavior of the network in making clusters of similar spectral type .
in this work the classification of the library of stellar spectra , @xcite using unsupervised artificial neural network , self - organizing map algorithm ( som ) is presented .
the 158 spectra of the jhc library with 2799 data points each , are the input vectors of the network and mapped into two dimensional output grid .
a problem with large soms is that when the map size is increased , the time it takes to do any operations on the map increases linearly .
when the grid size becomes very large , finding the best matching unit takes longer and longer .
the size of the som determines the resolution of the visualization it produces . on a small som
, lots of data samples will be projected to each som unit , whereas on a large som the similarity relationships of the samples are more readily visible .
the user must select the map size in advance .
this may lead to many experiments with different sized maps , trying to obtain the optimal result .
different map size and itterations are tried out and their performances analysed some of which are given in table 1 .
the map size of the @xmath17 with highest performance is selected for detailed analysis .
.list of map sizes with number if itterations and their performances [ cols="^,^,^",options="header " , ]
as it is seen from table 2 and figure [ som3 ] misclassified patterns belong to their neighboring clusters . for example
12b0v which is misclassified as o spectral type , is at the neighburhood of its true class .
likewise if we check all the misclassified patterns , we see that they are misclassified to their neighbouring true classes .
the jhc libraray is very detailed and it also includes the subspectral types from 0 - 9 . the misclassified patterns could be looked in detail to consider the subspectral types as well .
it is found that the misclassified patterns are very close to their true class .
for example , 80f0iii is wrongly classified as spectral type a , but it is close to the neighbouring class , 30a9v .
this is true with most of the misclassified patterns and it shows that the classification accuracy can be higher than 92.4% .
this technique and algorithm can be used for even larger unknown set of stellar spectra .
specially where we have different astronomical objects and need a pre - classification to identify the object types first , and then do further analysis on each cluster .
the beauty of this algorithm is that we do nt need any exemplars to train the network but it finds the similarities by its own without any supervision and makes the clusters of similar objects .
the author wishes to thank lukasz wyrzykowski for his constructive comments . | we present an automatic , fast , accurate and robust method of classifying astronomical objects .
the self organizing map ( som ) as an unsupervised artificial neural network ( ann ) algorithm is used for classification of stellar spectra of stars .
the som is used to make clusters of different spectral classes of jacoby , hunter and christian ( jhc ) library .
this ann technique needs no training examples and the stellar spectral data sets are directly fed to the network for the classification .
the jhc library contains 161 spectra out of which , 158 spectra are selected for the classification .
these 158 spectra are input vectors to the network and mapped into a two dimensional output grid .
the input vectors close to each other are mapped into the same or neighboring neurons in the output space .
so , the similar objects are making clusters in the output map and making it easy to analyze high dimensional data . after running the som algorithm on 158 stellar spectra , with 2799 data points each ,
the output map is analyzed and found that , there are 7 clusters in the output map corresponding to o to m stellar type .
but , there are 12 misclassifications out of 158 and all of them are misclassified into the neighborhood of correct clusters which gives a success rate of about 92.4% . |
quantum discord ( qd ) @xcite has recently been identified as a general resource in quantum information protocols ( see , e.g. , refs .
@xcite ) . in quantum computation
, it has been conjectured as the origin of speed up in the deterministic quantum computation with one qubit ( dqc1 ) mixed - state model @xcite .
moreover , a fundamental role has been attributed to qd in tasks such as quantum locking @xcite and quantum state discrimination @xcite . besides quantum protocols , remarkable applications of qd have also been found in the characterization of quantum phase transitions @xcite and in the description of quantum dynamics under decoherence @xcite . in order to use quantum correlations ( as provided by qd ) as a resource , we are faced with the problem of their distribution throughout a multipartite state . in this context , a monogamous behavior may reveal important information about the structure of quantum correlations . for instance , monogamy has been found to be the essential feature allowing for security in quantum key distribution @xcite . in general grounds
, it has been investigated whether any given measure @xmath0 of bipartite quantum correlation can obey a monogamy bound in an arbitrary tripartite quantum system @xmath1 .
a measure @xmath0 has typically been defined as monogamous if it follows the inequality @xmath2 therefore , given a fixed value for the quantum correlation @xmath3 between subsystem @xmath4 and the remaining part @xmath5 of the system , then @xmath4 can not be freely correlated with the individual subsystems @xmath6 and @xmath7 .
such a monogamy bound is obeyed by certain entanglement measures , which is indeed the origin of _ tangle _
@xcite as a measure of genuine multipartite entanglement . whether or not monogamy can be obtained for a quantum correlation measure in arbitrary ( pure or mixed ) quantum states poses therefore as a further challenge . however , many attempts turned out to find a polygamous behavior for qd @xcite .
more generally , such a negative answer has been obtained for any sensible measure @xmath0 of quantum correlations @xcite ( even though a function of @xmath0 may be able to yield a state - dependent trade - off @xcite ) .
the violation of the monogamy inequality ( [ mon1 ] ) means that the _ single - site correlation _ @xmath3 is unable in general to set a limit for the sum of pairwise correlations .
however , it does _ not _ imply that subsystems can be freely correlated . more specifically
, it is possible to restore monogamy if we can find out a physical quantity that is able to provide a bound for the sum of pairwise correlations .
_ here , we will investigate this problem by exploring such a route , providing a monogamy bound that holds for either pure or mixed states in a multipartite system composed of a number @xmath8 of subsystems_. in order to achieve this aim , we will show that a limit for the sum of pairwise quantum correlations is provided by a multipartite extension of qd , named as _ global quantum discord _ ( gqd ) , which has been proposed in ref .
@xcite . in the case of tripartite states , a monogamy bound
will then be obtained by replacing @xmath3 for the gqd , denoted by @xmath9 , yielding @xmath10 a schematic description comparing the monogamy bounds for @xmath3 and @xmath9 is displayed in fig .
the validity of inequality ( [ mon2 ] ) will be shown for all quantum states whose bipartite quantum discords do not increase under discard of subsystems . moreover , we will show that gqd allows for the extension of eq .
( [ mon2 ] ) for the case of @xmath8 subsystems .
we observe that the maximum of gqd typically increases with @xmath8 , which accounts for the fact that pairwise correlations are less monogamous than entanglement . on the other hand , as will be illustrated in several examples , gqd is sufficiently limited to provide a relevant constraint for the sum of pairwise qd . as an application , we will use eq .
( [ mon2 ] ) to propose a natural measure of residual multipartite quantum correlations .
qd , which has been introduced in ref . @xcite and provides a measure of the quantumness of correlations , can be suitably defined as the minimal loss of total correlation after a non - selective measurement @xcite . indeed , consider a bipartite system @xmath11 composed of subsystems @xmath4 and @xmath6 . denoting by @xmath12 the density operator of @xmath11 and by @xmath13 and @xmath14 the density operators of parts @xmath4 and @xmath6 ,
respectively , the total correlation between @xmath4 and @xmath6 is measured by the quantum mutual information @xcite @xmath15 where @xmath16 is the von neumann entropy for @xmath4 and @xmath17 is the entropy of @xmath4 conditional on @xmath6 . by operating on subsystem @xmath6 , qd can then be expressed as @xmath18 $ ] , where @xmath19 denotes a non - selective measurement @xmath20 on part @xmath6 of @xmath12 , which yields @xmath21 .
note that this definition is asymmetric with respect to measurement on @xmath4 or @xmath6 .
in particular , a strictly classical state requires both @xmath22 and @xmath23 .
indeed , this corresponds to a density operator @xmath24 , where @xmath25 is a joint probability distribution and the sets @xmath26 and @xmath27 constitute orthonormal bases for the systems @xmath4 and @xmath6 , respectively .
such strictly classical states can also be identified by a single measure , which is the symmetrized version of qd @xmath28 , \label{bipartitediscord}\ ] ] where the measurement operator @xmath29 is given by @xmath30 eq .
( [ bipartitediscord ] ) provides the symmetric qd considered in ref .
@xcite and experimentally witnessed in refs .
the vanishing of @xmath31 occurs if and only if the state is classical .
in particular , the absence of @xmath31 is the key ingredient for local sharing of pre - established correlations ( local broadcasting ) @xcite .
generalizations of quantum discord to multipartite states have been considered in different scenarios @xcite , which intend to account for quantum correlations that may exist beyond pairwise subsystems in a composite system . in this direction ,
one possible approach to account multipartite quantum correlations is to start from the symmetrized qd and then to systematically extend it to the multipartite scenario .
this originates gqd as a measure of global quantum discord , as proposed in ref .
the gqd @xmath32 for an arbitrary multipartite state @xmath33 under a set of local measurements @xmath34 can suitably be expressed as @xcite @xmath35 , \label{gqd - def}\ ] ] where @xmath36 with @xmath37 and @xmath38 denoting the index string @xmath39 ) . in eq .
( [ gqd - def ] ) , the mutual information @xmath40 is defined following the natural generalization proposed in ref .
@xcite , which implies that @xmath40 and @xmath41 are given by @xmath42 where @xmath43 gqd is symmetric with respect to subsystem exchange and shown to be non - negative for arbitrary states @xcite .
moreover , it can be detected through a convenient ( with no extremization procedure ) witness operator @xcite .
concerning its applications , gqd has been shown to be useful in the characterization of quantum phase transitions @xcite . in terms of operational interpretation
, gqd may play a role in quantum communication , in the sense that its absence means that the quantum state simply describes a classical probability multidistribution @xmath44 ( with @xmath45 , @xmath46 ) and , therefore , allows for local broadcasting of correlations @xcite .
let us begin by defining the loss of correlation @xmath47 in a quantum state @xmath33 generated by a measurement @xmath48 , which is given by @xmath49 we can then show that @xmath47 can be decomposed in terms of loss of correlation for suitable bipartitions , as provided by the following theorem . _ given a non - selective measurement @xmath48 , the loss of correlation @xmath47 can be decomposed as _ @xmath50 [ t1 ] we rewrite @xmath47 by using eqs .
( [ imulti ] ) and ( [ dphi ] ) , which yields @xmath51 \nonumber \\ & & -s\left ( \hat{\rho}_{a_1 \cdots a_n } \right ) + s\left(\phi \left ( \hat{\rho}_{a_1 \cdots a_n}\right ) \right ) .
\label{dentropy}\end{aligned}\ ] ] we then add and subtract @xmath52 and @xmath53 in eq .
( [ dentropy ] ) .
after rearrangement of the terms , we obtain @xmath54 by recursively applying eq .
( [ d2 ] ) , we can reduce the term @xmath55 to a sum of bipartite contributions @xmath56 which , when added to @xmath57 in eq .
( [ d2 ] ) , leads to eq .
( [ dn ] ) . by taking a tripartite system @xmath1 , theorem [ t1 ] implies that @xmath58 note that eq .
( [ eqdphi ] ) provides a decomposition of @xmath59 in terms of bipartite contributions . as a further step to achieve a monogamy trade - off ,
we will now introduce gqd by minimizing @xmath47 over all measurements @xmath48 , i.e. , @xmath60 .\ ] ] then , a monogamy bound for quantum correlations in @xmath8-partite systems can be obtained , which is proved below . _ for an arbitrary density operator @xmath33 with an amount of gqd given by @xmath32 , pairwise qd obeys the monogamy bound _
@xmath61 _ provided that the bipartite qds @xmath62 , with @xmath63 ) , do not increase under discard of subsystems , i.e. , @xmath64 . _ [ t3 ] starting from theorem 1 , we minimize both sides of eq .
( [ dn ] ) with respect to @xmath48 , yielding @xmath65 however , we have that @xmath66 then , by inserting eq .
( [ minim ] ) into eq .
( [ discord - in1 ] ) , we obtain @xmath67 hence , by using the condition that @xmath64 we obtain the monogamy bound stated by inequality ( [ monogamybound ] ) . in particular , for tripartite states , the monogamy inequality ( [ monogamybound ] ) reduces to inequality ( [ mon2 ] ) , provided that the condition @xmath68 is satisfied .
we observe that this non - increasing behavior of bipartite qd under discard of subsystems is a sufficient condition for the validity of monogamy , but it is _ not _ a necessary requirement .
this means that monogamy may hold even in more general scenarios .
for example , consider a tripartite system composed of qubits abc in the mixed state @xmath69 , where @xmath70 denotes the computational basis and @xmath71 .
for this state , we have that qd vanishes for the bipartition a with bc , i.e. @xmath72 .
however , by discarding subsystem c , we obtain @xmath73 , which mixes nonorthogonal states for subsystem b. we then obtain a nonvanishing qd between a and b , which is given by @xmath74
. however , the monogamy trade - off @xmath75 keeps still valid , since gqd turns out to be equal to @xmath76 , namely , @xmath77 and @xmath78 .
hence , we obtain a monogamous behavior as given by inequality ( [ mon2 ] ) , with saturation achieved for this state .
let us illustrate the monogamy bound ( [ monogamybound ] ) in the case of multipartite pure states @xmath79 that admit schmidt decomposition , whose explicit conditions of existence are discussed in ref .
we will assume that the system is composed by a set of qubits .
in such a case , we can write @xmath80 , where @xmath81 are orthonormal bases , @xmath82 , and @xmath83 . for the density operator @xmath84
, we obtain @xmath85 since schmidt decomposition implies equal spectrum for all single - qubit reduced density operators @xmath86 , we obtain that @xmath87 , for any individual subsystem @xmath88 .
therefore , the mutual information is @xmath89 . in order to consider measurements @xmath48 over @xmath33
, it can be shown that , by adopting projective ( von neumann ) measurements , the minimization of the loss of correlation is obtained in schmidt basis , namely , @xmath90 .
this is a consequence of both the group homomorphism of @xmath91 to @xmath92 and the monotonicity of entropy under majorization ( see discussion for the state @xmath93 in ref .
. then , @xmath94 , which implies @xmath95 . moreover @xmath96 .
therefore , the mutual information after measurement is @xmath97 , which implies that @xmath98 .
therefore , the sum of pairwise qds is upper limited by the von neumann entropy @xmath99 of an individual subsystem , which holds for quantum systems composed of an arbitrary number @xmath8 of qubits . as an example , consider the ghz state for @xmath8 qubits , namely , @xmath100 . for this state
, we obtain @xmath101 ( independently of @xmath8 ) and vanishing pairwise correlations @xmath102 , which is in agreement with theorem [ t3 ] .
let us now consider a tripartite system @xmath1 described by the w - ghz state @xmath103 where @xmath104 and @xmath105 .
note that , for @xmath106 and @xmath107 , we have pure states , given by @xmath108 and @xmath109 states , respectively .
therefore , by adopting projective measurements , we will have that @xmath110 for @xmath111 .
as @xmath112 increases , we numerically find out a monotonic increase of gqd until @xmath113 for @xmath114 .
this can be seen as a consequence of the absence of schmidt decomposition for the @xmath115 state , which leaves gqd unconstrained by the entropy of an individual subsystem . for the complete range of @xmath112 , we plot gqd in fig .
[ f2 ] as well as the pairwise sum @xmath31 + @xmath116 .
we observe that monogamy as given by inequality ( [ mon2 ] ) is obeyed for any @xmath112 , with distinct minimizing bases @xmath117 for gqd and pairwise qd .
note also from the plot that the bound is considerably tight in the sense that , for any @xmath112 , the value of gqd is sufficiently limited to ensure that the subsystems are _ not _ freely correlated .
the monogamy bound ( [ mon2 ] ) allows for the definition of a measure @xmath118 for _ residual _ multipartite quantum correlations , namely , contributions to quantum correlations beyond pairwise qd .
this is similar to the definition of _ tangle _ as a measure of residual multipartite entanglement @xcite .
indeed , let us consider the particular case of tripartite permutation invariant states , such the w - ghz state .
in such a case , we can define the residual qd as @xmath119 note that , by considering the non - increasing behavior of qd under discard of subsystems ( such as in the w - ghz state ) , monogamy is ensured , implying that @xmath120 .
if permutation invariance is absent , the tripartite residual correlations can be defined by @xmath121 , where minimization is taken over all subsystem permutations ( see ref .
@xcite for a similar definition in the case of entanglement ) . applying the residual measure for the ghz state
, it follows that @xmath122 , since pairwise contributions vanish . for the w state ,
residual qd is lower than in the case of the ghz state , since pairwise qd is nonvanishing .
indeed , it can be be shown that @xmath123 ( with the minimizing measurement found in the @xmath124 eigenbasis ) .
this implies that @xmath125 .
this behavior is exhibited in fig .
[ f2 ] and is in agreement with the hierarchy found in ref .
in conclusion , we have introduced a monogamy bound for pairwise quantum correlations based on the amount of gqd available to a multipartite system .
remarkably , this monogamy inequality holds for general states whose bipartite quantum discord is non - increasing under discard of subsystems .
in particular , oppositely to the typical monogamy bound ( [ mon1 ] ) , it covers both ghz and w tripartite classes of states , providing therefore a promising setup for the investigation of measures for multipartite classical and quantum residual correlations .
we leave such topics for further research .
we thank kavan modi and jzsef pitrik for helpful discussions .
this work is supported by cnpq , capes , faperj , and the brazilian national institute for science and technology of quantum information ( inct - iq ) . | we introduce a monogamy inequality for quantum correlations , which implies that the sum of pairwise quantum correlations is upper limited by the amount of multipartite quantum correlations as measured by the global quantum discord .
this monogamy bound holds either for pure or mixed quantum states provided that bipartite quantum discord does not increase under discard of subsystems .
we illustrate the monogamy behavior for multipartite pure states with schmidt decomposition as well as for w - ghz mixed states . as a by - product
, we apply the monogamy bound to investigate residual multipartite correlations . |
attempts to describe granular media in terms of a more fundamental underlying statistical mechanics have met with considerable success . as a prototype for this approach , in the preceding paper @xcite standard linear response methods from normal fluids have been applied , _ mutatis mutandis _
, to the case of an impurity particle diffusing in an isolated one component fluid of smooth inelastic hard spheres ( @xmath0 ) or disks ( @xmath1 ) of diameter @xmath2 .
the collisions are characterized by a coefficient of normal restitution @xmath3 . the isolated system ( or with periodic boundary conditions ) , is not in the typical gibbs state as for elastic collisions , but rather in a time dependent homogeneous cooling state ( hcs ) .
it has been shown in @xcite , and will be elaborated further here , that this time dependent hcs can be exactly transformed to a stationary state description . in the present work ,
this stationary state description is evaluated by molecular dynamics ( md ) simulation to measure the mean square displacement of the impurity , its velocity autocorrelation function , and the resulting diffusion coefficient defined in terms of a formal einstein or green - kubo relation , respectively . for practical purposes ,
attention is restricted to the case of self - diffusion , where the mechanical properties of the impurity are the same as those of the fluid particles .
the velocity autocorrelation function was approximated in @xcite using a cumulant expansion and also by means of kinetic theory methods . in the usual first sonine polynomial expansion for the latter , this leads to a simple exponential decay in an appropriate dimensionless time scale defined below .
for the numerical results considered here , one further approximation is made in evaluating the decay rate , namely two - particle velocity correlations are neglected . in this case
, the results are the same as those obtained from the enskog kinetic theory . earlier studies of self - diffusion @xcite show excellent agreement of this theory with md results at very low number density @xmath4 , over a wide range of values for @xmath3 .
the present work extends that study to higher densities .
more specifically , three - dimensional systems with densities in the interval @xmath5 , for @xmath6 , will be considered .
the hcs is known to be unstable under long wavelength perturbations or fluctuations @xcite . to avoid this problem , the system size in most cases considered
is chosen to be smaller than the critical wavelength .
the latter is a function of the density and inelasticity , decreasing with increasing values of each .
consequently , at the highest densities and smallest @xmath3 a small system of @xmath7 particles is required . at moderate densities and inelasticities , a system size of @xmath8 particles
the agreement of theory and simulation is found to be quite good for all @xmath3 at @xmath9 and for all densities at @xmath10 . at higher densities and smaller @xmath3
, significant discrepancies occur , with the diffusion coefficient obtained from md being almost an order of magnitude greater than the theoretical estimate at @xmath11 and @xmath12 .
the failure of the enskog theory at high densities is well - known for the case of elastic collisions , due to cage effects and correlated binary collisions .
however , in that case the diffusion coefficient from md is smaller than the theoretical prediction .
therefore , it is clear that a quite different density mechanism is effective for inelastic collisions .
it is reasonable to expect that the underlying instability is responsible in part for the above discrepancies .
although the instability is , in principle , avoided by the use of small systems , the latter is prone to large fluctuations .
these are not quantitatively significant at @xmath13 but grow rapidly with decreasing @xmath3 , as can be seen in the noise level for the kinetic temperature .
this will be illustrated later on in fig .
a related issue is the role of the instability in affecting the usual mode coupling corrections to the enskog theory .
such mode coupling terms arise from correlations of spontaneous fluctuations in the long wavelength hydrodynamic modes , which are significantly modified by the potential instability .
enhanced fluctuations will be addressed in the qualitative analysis of the data below .
the plan of the paper is as follows . in sec . [
s2 ] , the relevant results in the preceding paper are shortly summarized , and a new scale transformation , useful for practical purposes , is introduced .
the equations of motion of the new phase - space variables lead in a direct way to a steady - state simulation method for the hcs . starting from an arbitrary initial condition ,
the system rapidly approaches a steady state , whose properties are simply related to those of the hcs .
it is important to stress that this relationship is not approximate , but an _ exact _ consequence of a change of variables . in particular , the method to measure the mean square displacement of a tagged particle as well as its velocity autocorrelation function is discussed in detail .
moreover , both quantities are shown to lead to equivalent results for the self - diffusion coefficients .
although in practice some discrepancies appear in the numerical results obtained by the two procedures , their origin is well understood .
comparison of theory and simulation is also started in sec . [ s2 ] , and completed in sec .
[ s3 ] . as mentioned above
, the agreement is fairly good at low densities , but relevant effects , which are not taken into account by the theory developed in @xcite , show up even at moderate density .
the peculiar nature of these effects for inelastic collisions is discussed .
numerical evidence is provided indicating that the underlying hydrodynamic instability associated to the shear mode plays an essential role in the `` anomalous '' behavior of the self - diffusion coefficient .
nevertheless , such changes in the diffusion constant do not compromise the existence of the diffusion process , which is confirmed for such conditions by simulation results for the mean square displacement .
finally , sec .
s4 contains a short summary of the results and also some indications of the possible extensions of the reported work .
to investigate the nature of diffusion in the hcs and to test the theoretical results presented in the preceding paper @xcite , a more fundamental description via md simulation will be considered .
the hcs distribution function for a fluid of @xmath14 particles has the scaling form @xmath15 ^{-d(n+1)}\rho _ { hcs}^{\ast } \left ( \{{\bf q}_{ij}/\ell , { \bf v}_{i}/v(t)\}\right ) .
\label{2.1}\]]here and in the following the notation is the same as in ref .
in particular , @xmath16 is the thermal velocity defined in the usual way ( with boltzmann s constant set equal to one ) and @xmath17 is proportional to the mean free path of the gas .
the time dependence of the above distribution is due to collisional cooling and is determined from @xmath18where @xmath19 is the cooling rate . a direct simulation of the cooling fluid , as described by the liouville dynamics in the actual phase variables , is difficult , since the rapid cooling of the fluid leads to numerical inaccuracies very soon .
one method of dealing with this would be to periodically redefine the time - scale of the simulation , so that the typical particle velocity remains of the order of unity .
however , if the rescaling is only used to control the numerical stability of the simulation , the state being simulated would nevertheless be time - dependent . for this reason ,
some simulation studies employ different type of thermostats , such as an externally imposed brownian force or thermal boundaries , in order to generate a steady state . while these methods provide more or less realistic models of various experimental procedures , they obviously probe a state which is in some way related to , but not identical with , the homogeneous cooling state .
soto _ et al . _
@xcite , noting the fact that there is no intrinsic time scale in the dissipative hard sphere model , have proposed to rescale all particle velocities after every collision , thus establishing a steady state similar to that described in @xcite . however , this procedure has the effect of replacing the binary collision dynamics of the dissipative hard sphere model by an @xmath20-body dynamics , since the result of a collision is to alter _ all _ atomic velocities and not just those of the colliding atoms .
it is reasonable to imagine that the binary dynamics is recovered in the infinite system limit , but the connection between the hcs and this dynamics is not clear for the case of the small systems considered in most of the md simulations .
instead , we follow the procedure used in ref .
lutskopre , according to which an _ exact _ mapping of the hcs onto a steady state is exploited as the basis of the simulation method . following the ideas developed in @xcite ,
all velocities are scaled relative to @xmath21 and the new dimensionless time is given by @xmath22this time scale is a measure of the average collision number .
the corresponding liouville equation in these variables supports a stationary hcs solution given by @xmath23 , where @xmath24 and @xmath25 .
moreover , the average value of any phase function @xmath26 is given by eq.(41 ) of reference @xcite.@xmath27the dynamics in the new phase space is obtained from@xmath28where @xmath29 is the dimensionless liouville operator , and @xmath30 is the dimensionless cooling rate .
this is the same as the usual dynamics for hard spheres , except that the term proportional to @xmath31 represents an acceleration between collisions which balances the energy lost during collisions , thus enabling a steady state .
although it is possible to relate average values of the relevant dynamical functions of the original variables with the average values of the same functions of the scaled variables for a general situation , attention will be restricted in the following to the hcs .
for homogeneous systems , @xmath32 becomes time independent . in principle , the scaled dynamics defined by eq .
( [ 2 ] ) can be simulated for the properties of interest , without the complications of continuous cooling in real time . in this formulation ,
the simulation is qualitatively similar to that for elastic collisions , in the sense that the system rapidly approaches a state for which the time - average of the instantaneous scaled temperature is constant , and subsequent determination of average properties is simpler and numerically more accurate .
a technical complication is the need to know the exact value of @xmath33 _ a priori _ , which in general is not possible , since it is determined by the original dynamics .
consequently , it is useful to make a second change of scale , @xmath34where @xmath35 is an arbitrary time independent dimensionless frequency .
equation ( [ 2 ] ) then becomes@xmath36with @xmath37 the same @xmath29 , but with the replacements @xmath38 .
this is the form most convenient for md simulation , with any reasonable choice for @xmath35 .
more explicitly , the particle dynamics implied by eq.([4 ] ) and implemented in our event - driven simulations consists of an accelerating streaming between collisions @xmath39while the effect of the collision of two particles is to alter their relative velocity according to @xmath40where @xmath41 , @xmath42 is the unit vector pointing from atom @xmath43 to atom @xmath44 and the center of mass velocity remains the same .
the relationship of @xmath33 to @xmath35 is determined by the steady state temperature obtained in the simulation . to see this note that while the instantaneous kinetic energy of the entire system is clearly not constant , a corresponding `` temperature '' defined as @xmath45 does approach a constant , as is shown in appendix [ appa ] , @xmath46 e^{-w^{\ast } s^{\ast } /2}\right\ } ^{-2 } \nonumber \\ & \rightarrow & \frac{1}{2}\left ( \frac{w^{\ast } } { \zeta ^{\ast } } \right ) ^{2}. \label{5}\end{aligned}\]]thus @xmath47 which relates quantities measured during the steady state simulation to those calculated in @xcite . in summary ,
simulation of eqs.([4 ] ) is expected to yield a steady state after a short transient period .
subsequently , ensemble averages of properties can be determined as time averages by making the usual assumption of ergodicity .
these properties are directly related to those of the hcs by a simple scale transformation as described above .
this steady - state simulation method removes any limitation on the time for which trajectories may be followed , but there are other limitations due to a long wavelength hydrodynamic instability@xcite for systems with dimension larger than a critical size @xmath48 where @xmath49 is the shear viscosity .
these instabilities occur when the decay rate of a shear mode fluctuation is less than the cooling rate of the thermal velocity so that their size grows relative to the temperature . in the scaled dynamics , the instabilities show up as ordinary unstable modes that grow exponentially with @xmath50 , and are therefore easily identified in the simulations @xcite .
since these growing modes represent a spontaneous breaking of the ( assumed ) spatial homogeneity of the system , the impurity particle no longer undergoes simple diffusion and , therefore , its motion is beyond the scope of this paper .
the critical size @xmath51 is a function of the density and coefficient of restitution , being smaller for high density and small restitution coefficient . in all the cases reported here , the system size is smaller than that required for the instability to show up . for a system of @xmath52 atoms and @xmath53 ,
this provides no limitation on the accessible densities and coefficients of restitution , otherwise only sufficiently low ( high ) values of the density ( coefficient of restitution ) can be studied .
a different limitation on the simulation is the `` inelastic collapse '' @xcite , which occurs for small values of @xmath3 thereby setting a lower limit on the values accessible to simulation . for the system sizes considered in this paper , the above mentioned limits
have been characterized in more detail elsewhere @xcite .
the three - dimensional simulations reported in the following begin in all cases with an equilibrated fluid subject to elastic collisions ( i.e. @xmath54 ) with @xmath55 .
the coefficient of restitution is then set to the desired value and the scaling parameter @xmath35 is set to @xmath56 using the enskog estimate for the cooling rate @xmath57 as given by eq .
( 74 ) in ref .
this choice ensures that the temperature remains of order @xmath58 for all values of @xmath3 , for optimal numerical stability , as well as provides a continuous path to the equilibrium system for which @xmath13 and @xmath59 .
the simulation is then continued for @xmath60 collisions to allow the system to reach the steady state at @xmath61 where @xmath33 is the true cooling rate generated by the simulation .
since the largest system considered consists of @xmath62 atoms , this corresponds to at least @xmath63 collisions per atom , and is more than sufficient to reach the steady state . a final simulation of @xmath60 additional collisions
is then performed , during which all statistical averages of interest are accumulated .
figure [ fig1 ] shows the behavior of the instantaneous kinetic energy ( or temperature ) during several typical simulations . in all cases
, the instantaneous temperature fluctuates around a stationary average value , with the size of the fluctuations increasing as the coefficient of restitution decreases .
the fact that the average value itself varies with @xmath3 is due to ( a ) the inadequacy of the simple estimate of @xmath33 we have used for decreasing values of @xmath3 and ( b ) to the presence of long - lived shear fluctuations , or vortices , which decay on a time scale of thousands of collisions near the instability .
as will be discussed below , the latter give rise to large fluctuations in all components of the kinetic contribution to the pressure tensor as well as raising the apparent temperature .
the system is expected to show a typical diffusive behavior under the scaled dynamics after a short transient time , i.e. for @xmath64 . then the diffusion coefficient can be obtained either from the mean squared displacement ( msd ) using the einstein relation , eq . ( 70 ) of ref .
@xcite , or from the green - kubo relation , eq .
( 69 ) of ref.@xcite .
the former is evaluated , for @xmath0 , from@xmath65figure [ fig2 ] shows some typical simulation data exhibiting the expected linearity of the mean - square displacement as a function of @xmath50 . the green - kubo relation expresses the diffusion coefficient in terms of the velocity autocorrelation function ( vacf ) as @xmath66the vacf is evaluated from@xmath67where @xmath68 , @xmath69 is the sampling period , and @xmath70 is a function of the amount of data available and the time at which the steady state is established .
notice that the stationarity of the scaled dynamics is explicitly used to evaluate the vacf for a given time separation @xmath50 , by averaging over many different samplings during the simulation . in the simulations
, @xmath69 is taken to be @xmath71 of the boltzmann mean free time at @xmath55 , i.e. @xmath72 . as discussed above , @xmath55 is not the exact value of the converged temperature , but in formulating @xmath73 only a reasonable order of magnitude is required . to evaluate the right side of eq .
( [ 2.14 ] ) , we follow the procedure of futrelle and mcginty futrelle ( see also the discussion in ref .
@xcite ) , according to which the components of the velocity for each atom sampled at intervals of @xmath73 are stored .
when the number of values stored reaches a threshold , @xmath74 , the convolution theorem is used to evaluate eq .
( [ 2.14 ] ) by means of the fast fourier transform , and so to obtain an estimate of the velocity autocorrelation function .
this procedure is then repeated until the end of the simulation and the various estimates , typically between tens and hundreds of samples , are averaged to obtain the final estimate of the correlation function .
figure [ fig3 ] shows some typical correlation functions obtained in this way , using @xmath75 .
the functions have been normalized in all cases by their initial values , i.e. the function @xmath76 , as defined in the preceding paper , has been plotted . to get the diffusion constant , the integral of the velocity correlation function
is computed using simpson s rule and the variance of the pool of estimates is used as the basis for the calculation of the standard error .
both methods , einstein relation and green - kubo relation , should yield the same values of the coefficient of self - diffusion .
figures 2 and 3 show that the mean square displacement becomes linear in @xmath50 and the velocity autocorrelation function decays to zero for @xmath77 . for larger @xmath50 both forms for the diffusion coefficient approach a constant , and all further discussion
is restricted to this _
diffusion constant_. the consistency between the green - kubo and einstein relations is confirmed in fig .
[ fig4 ] , where the estimates obtained by both procedures are compared for a system of 108 atoms and several values of the density and of the coefficient of restitution . while in qualitative agreement , the difference in values obtained by the two methods is in some cases as great as @xmath78 .
it is expected that the results based on the velocity autocorrelation function are more accurate , because the evaluation at each fixed time @xmath50 , is based on an average over all times included in the simulation . in contrast , the mean square displacement is obtained from a single evaluation for each @xmath50 . this qualitative difference in accuracy has been confirmed by running multiple simulations for a few systems , and observing variations in the coefficient of self - diffusion as large as @xmath78 when determined from the msd , whereas that determined from the vacf showed variations of less than @xmath79 .
figure [ fig5 ] shows the ratio of the observed diffusion constants to the value predicted by the enskog - level theory , as a function of @xmath3 for densities @xmath80 , and @xmath81 . for the lowest density , theory and simulation
are in excellent agreement even at strong dissipation , as expected from earlier comparisons of simulation results and predictions of the boltzmann - enskog equation @xcite . as the density increases ,
deviations from the enskog prediction are expected , since they are known to occur even at equilibrium ( @xmath13 ) . in this respect ,
our results for @xmath82 are quite consistent with previous studies of elastic hard spheres@xcite .
it is seen in the figure that deviations increase with decreasing @xmath3 .
for example , at @xmath83 the deviations are less than @xmath78 for @xmath84 , but increase rapidly to the order of @xmath85 for @xmath86 .
this behavior can also be observed in the normalized vacf itself which , in the enskog approximation ( here and below we refer to the first sonine approximation to the enskog theory ) , is given by @xmath87 , with @xcite @xmath88where @xmath89 is the pair correlation function for two particles at contact . in fig .
[ fig6 ] , the logarithm of the normalized velocity autocorrelation function , @xmath90 , is plotted as a function of @xmath91 , for the case of _ elastic _ collisions , @xmath13 , and the several densities we have been considering . here and below the choice has been made , somewhat arbitrarily but consistently , of truncating the vacf at @xmath92 in order to eliminate the tails which are dominated by noise .
the data confirm findings of earlier studies @xcite that the enskog theory gives a good description at low densities , but increasingly underestimates the diffusion constant , due to the neglect of correlated collisions and cage effects , at higher densities with a maximum deviation around @xmath93 . above this density ,
the neglected processes begin to cancel one another and near @xmath11 , the diffusion constant crosses the enskog prediction , and for higher densities is over - estimated by enskog theory . since the enskog theory is exact at short times@xcite , these effects appear as deviations from the simple exponential form of the vacf at longer times .
now consider @xmath94 .
a similar plot , fig .
[ fig7 ] , for an inelastic system with @xmath95 , again shows good agreement with the enskog prediction at the lowest density , but it exhibits much larger deviations than in the elastic case at higher densities , except at short times where the enskog theory is exact , as already pointed out .
interestingly , the data clearly indicate a crossover to a slower constant decay rate at longer times .
of course , this behavior increases the time integral of the velocity autocorrelation function , and it is responsible for the enhancement of the diffusion coefficient seen in fig .
[ fig5 ] , in contrast to the opposite behavior for high densities at @xmath13 . to characterize the deviation of the vacf from the simple exponential form , we start with an exact expression for the vacf based on the zwanzig - mori formalism ( see appendix [ appb]),@xmath96where @xmath97 is known as the memory function .
if @xmath97 is neglected the exponential decay of the above enskog approximation is recovered , so the memory function incorporates all of the effects neglected in that approximation .
a simple ansatz for this function as an exponential is qualitatively successful in modelling the vacf of fluids with elastic collisions ( see @xcite and references therein ) , and can be expected to work also for inelastic systems , when formulated in the dimensionless time @xmath98 .
if we substitute @xmath99 into eq.([2.16 ] ) , and solve for the vacf with the boundary condition @xmath100 , the resulting model is@xmath101where the constants @xmath102 and @xmath103 can be related to @xmath104 and @xmath105 .
figure [ fig7 ] includes the result of fitting these two parameters to the data , being evident that the above model is able to capture the crossover from the enskog behavior at short times to the slower relaxation for longer times .
while the memory function model provides a framework for describing the results of the simulations , it does not explain them , since any effect not captured in the enskog theory will give rise to a non - zero contribution to the memory function . in a previous study @xcite ,
similar deviations from the enskog theory were found for the pressure of the system , and evidence was provided there suggesting that the large discrepancies between theory and simulation for dense , dissipative states may be due to the underlying hydrodynamic instability .
additional support for this possibility comes from fig .
[ fig8 ] which shows the diffusion constants obtained from the simulation data for systems composed by 500 particles compared to those for systems composed by only 108 particles . for a density of @xmath83
, there is a significant further increase of the diffusion constant in the large system relative to the small system as @xmath106 decreases , with the diffusion constant of the 500 atom system growing to almost twice that of the 108 atom system at @xmath86 . at @xmath107 ,
the enhancement is even larger , but we know @xcite that the system is in the unstable regime , at least for @xmath108 and possibly for smaller values , so that most of the enhancement is undoubtedly due to the spontaneous formation of shear flow in the system .
this suggests that in the smaller systems , although they are stable , there are present large fluctuations characteristic of the instability , such as spontaneous vortices that form and breakup . in this case , the impurity would find itself in a fluctuating local flow field which could enhance the velocity correlations .
to test whether a local flow field plays a relevant role in the self - diffusion process , two different methods were used to calculate the autocorrelation of the impurity velocity _ relative to the instantaneous local flow field_. in the first method , the instantaneous local flow field was calculated by dividing the simulation cell into @xmath109 cubic subcells .
then each atom s velocity relative to the instantaneous average velocity of the fluid in the subcell containing it , was used when calculating both the temperature and the vacf .
for the 108 particle system , this procedure leads to a lowering of the measured value of the equilibrium diffusion constant by about @xmath78 , which we believe to be a finite - size effect due to the removal of a significant fraction of the degrees of freedom of the system .
for @xmath93 it also reduces the steady state temperature from nearly twice the initial temperature ( see fig .
[ fig1 ] ) to about @xmath110 thus demonstrating the size of the contribution of these fluctuations to the temperature .
when the same procedure is applied to a system of 500 atoms , the shift in the equilibrium diffusion constant is negligible .
the results obtained by this method are presented in figs .
[ fig9]-[fig11 ] . in the graphs ,
the @xmath3-dependant diffusion constant has been scaled by its measured equilibrium value .
the conclusion emerging from the figures is that the @xmath3-dependence of the new diffusion constant is substantially closer to the enskog form .
however , this method may be criticized on the grounds that the statistics of the local velocity field are poor , since the calculations only involve a few atoms in each cell .
we , therefore , also consider a second method which is specifically designed to eliminate only the part of the local flow due to the longest wavelength fluctuations in the system .
a local flow field @xmath111 is defined by summing over the smallest fourier components compatible with the size and shape of the simulation cell , @xmath112where @xmath113 is the volume of the cubic simulation cell .
the fourier components are determined from the instantaneous velocities of the bulk fluid via@xmath114this method is expected to give better statistics than the cell method discussed above .
we then use this flow field to define the relative velocity used in the calculation of the temperature and the vacf .
figures [ fig9]-[fig11 ] also show the resulting diffusion constant when this method of subtracting the instantaneous fluctuations is used
. from the figures follows that both methods are mutually consistent over the range @xmath115 , and that the agreement between the simulation results and the theoretical prediction from enskog theory is significantly improved .
the effects of subtracting the local flow field are much larger for the inelastic case due to the mechanism responsible for shear instability .
in this paper , we have continued the discussion of diffusion in a model granular system begun in ref .
the main motivation has been to show that transformation to the steady - state dynamics allows to carry over many standard methods of nonequilibrium statistical mechanics with relatively minor modifications . here , the exact correspondence between the usual formulation of the dissipative hard - sphere model of granular fluids and the steady - state dynamics was exploited to formulate a particularly convenient simulation method which eliminates the need for additional complications such as exothermic boundary conditions .
we have demonstrated that in the steady - state variables , the homogeneous cooling state exhibits standard diffusive behavior such as the linear increase of the mean - squared displacement with time , and the correspondence between the einstein and green - kubo methods of determining the diffusion constant .
in addition , the simplest approximation of enskog kinetic theory provides excellent agreement at low densities for the whole range of inelasticity @xmath116 .
however , although the formalism and concepts of the statistical mechanics of diffusion have been shown to give an adequate description of diffusion in the hcs , large quantitative deviations from the enskog kinetic theory predictions have been observed even at relatively moderate densities . at the level of the vacf
, it was shown that the deviations are principally due to a crossover from the enskog time dependence , which is exact for short times , to another , slower decay which nevertheless also appears to be exponential in form .
the analytic form of the vacf was shown to be well approximated by modeling the memory function of the diffusive process by a simple exponential as has been used in early studies of memory effects in equilibrium fluids @xcite .
physically , we have shown by means of constrained simulations that the deviations from the enskog model are primarily due to the effect of the longest wavelength velocity modes in the system .
for sufficiently large systems , the shear mode is linearly unstable ( in the steady - state picture ) in the classic sense that the time constant associated with its decay goes to zero as the critical value of @xmath3 is approached from above .
a linear stability analysis in the steady - state variables@xcite shows that the shear fluctuations of wavevector @xmath117 decay exponentially with a time constant of @xmath118,where @xmath119 is the shear viscosity and the stars indicate quantities expressed in the reduced units of @xcite .
the dominant @xmath3-dependence of this expression comes from @xmath120 and the size of the system enters through the fact that the smallest non - zero value of the wavevector the system can sample is @xmath121 where @xmath29 is the longest dimension of the simulation cell . for fixed values of @xmath94 , there will always be a critical value of @xmath29 above which the system is unstable to shear fluctuations . for values of @xmath3 above the critical value , the decay of these fluctuations is nevertheless slowed relative to equilibrium and we therefore reason that , even before the onset of the instability and even in systems too small to exhibit the instability , the slowing down of this mode means that long - lived , long - wavelength fluctuations are present .
these fluctuations give rise to the observed enhancement of the diffusion constant .
this effect is somewhat analogous to what is observed in turbulent systems since the removal of these modes in the calculation of the diffusion constant removed a large part of the discrepancy from the enskog prediction .
the size dependence of the deviations from the enskog results supports this conclusion , since larger systems have a higher critical value of @xmath3 because the wavevector can take on smaller values .
we conclude that a complete theoretical description of diffusion in hcs will require a model for the memory function taking into account the slow relaxation of the shear modes by means , e.g. , of a mode - coupling mechanism .
the research of jwd was supported by national science foundation grant phy 9722133 .
jjb acknowledges partial support from the direccin general de investigacin cientfica y tcnica ( spain ) through grant no .
pb98 - 1124 .
jl acknowledges support from the universit libre de bruxelles .
the average value of an observable @xmath122 for a general homogeneous state @xmath123 is given by eq .
( [ 1 ] ) above which can be written @xmath124and@xmath125the subscript on @xmath126 indicates explicitly that it is the dimensionless cooling rate associated with the hcs@xmath127where @xmath17 is the mean free path and @xmath128 is the hcs thermal velocity . the temperature @xmath129 obeys the equation @xmath130 now make the change of variables@xmath131to get@xmath132with the definition @xmath133 .
in particular , the kinetic energy is@xmath134the corresponding temperature is@xmath135where @xmath136 .
the time derivative of @xmath137 is then found to be@xmath138 this is still valid for a general homogeneous state .
now assume the existence of a scaling solution , which implies @xmath139 .
then @xmath140the solution is that given by eq .
( [ 5 ] ) of the text .
the dimensionless velocity auto correlation function is defined by using the notation in ref .
@xcite this can be written in terms of the generators of the dynamics@xmath142the detailed forms for the linear operators @xmath143 and @xmath144 are given by eqs .
( 36 ) and ( 42 ) of ref .
dbyl01 but will not be required here .
a projection operator is defined by @xmath145it follows then that @xmath146with the choice @xmath147the equation of motion for @xmath148 is@xmath149and a closed equation for @xmath150 is obtained by operating on this equation with @xmath151 and @xmath152 to get the pair of equations@xmath153solving formally for @xmath154 in the second equation and substituting into the first gives the desired closed equation for @xmath150@xcite@xmath155use of ( [ a4 ] ) gives the corresponding equation for @xmath156 @xmath157with the definitions @xmath158and@xmath159the definition for @xmath160 is the same as that for the enskog approximation discussed in @xcite and approximated in eq .
( [ 2.15 ] ) above . finally
, it is clear that structurally identical expressions would have been obtained had we performed this derivation using the scaled variables introduced in eq .
( [ 3 ] ) of the text thus justifying the expression in eq . ( [ 2.16 ] ) . as a function of
the dimensionless scaled time @xmath161 in the steady - state dynamics for @xmath162 .
the values of the coefficient of restitution are @xmath166 ( lower curve ) , @xmath164 ( middle curve ) , and @xmath165 ( upper curve).,width=480,height=480 ] ( circles ) and @xmath162 ( diamonds ) , as determined from the mean - squared - displacement ( open symbols ) and the velocity autocorrelation function ( full symbols ) .
the values are normalized with those obtained for @xmath163,width=480,height=480 ] as a function of time for equilibrium for a system in equilibrium ( @xmath166 ) .
symbols are as in fig .
the lines are a guide to the eye , except the full line , which indicates the enskog result.,width=480,height=480 ] as a function of dimensionless time for the steady - state dynamics with @xmath164 .
symbols are as in fig .
the lines are the results of fits of the memory function model with an exponential kernel as discussed in the main text .
the lowest density results are indistinguishable from the enskog prediction on this scale.,width=480,height=480 ] as a function of @xmath173 for 108 atoms , calculated from the green - kubo relation ( error bars and connecting line ) , by removing the longest fourier modes ( diamonds ) , and by using the cell - method of computing the local velocity field ( circles).the heavy line is the prediction from enskog theory.,width=480,height=480 ] | the linear response description for impurity diffusion in a granular fluid undergoing homogeneous cooling is developed in the preceeding paper .
the formally exact einstein and green - kubo expressions for the self - diffusion coefficient are evaluated there from an approximation to the velocity autocorrelation function .
these results are compared here to those from molecular dynamics simulations over a wide range of density and inelasticity , for the particular case of self - diffusion .
it is found that the approximate theory is in good agreement with simulation data up to moderate densities and degrees of inelasticity . at higher density ,
the effects of inelasticity are stronger , leading to a significant enhancement of the diffusion coefficient over its value for elastic collisions .
possible explanations associated with an unstable long wavelength shear mode are explored , including the effects of strong fluctuations and mode coupling . |
the dicke superradiance model @xcite as a test - bed for mean - field like phase transitions @xcite has received renewed attention recently , in particular due to its successful implementation in cold - atom experiments @xcite and optical setups of cavities and lasers @xcite .
much progress has been made towards a realistic description of non - equilibrium and dissipation @xcite , multi - mode effects @xcite , the interplay of the superradiant phase transitions and bose
einstein condensation @xcite , spin glasses @xcite , the analysis of interactions @xcite , inhomogeneous couplings @xcite , finite size effects @xcite , or the adiabatic limit @xcite
. further extensions of the model have appeared allowing for driving @xcite , the creation of goldstone modes @xcite , feedback control @xcite , or the transfer to other platforms such as solid - state systems @xcite . in its simplest version ,
the dicke model hosts a mean - field type ground state phase transition ( ` quantum bifurcation ' ) at zero temperature @xcite between a field - free ( normal ) phase with unpolarised atoms , and a superradiant phase with macroscopic occupation of the field mode and polarisation of the atoms . above a critical coupling strength , the superradiant phase also persists at finite temperatures below a critical temperature which defines the corresponding thermal second order phase transition @xcite .
thermal aspects of the dicke - hepp - lieb superradiance phase transition have found recent interest again in the analysis of thermodynamic aspects like work extraction @xcite , and the discussion of van - hove type singularities in the microcanonical density of states at large excitation energies @xcite ( excited state quantum phase transitions ) .
this is our motivation to extend our previous studies @xcite of an extended dicke model towards finite temperatures in this paper . in the original dicke model @xcite ,
the atoms are approximated by two - level systems and the light field by a single ( bosonic ) mode of a resonator .
naturally , the questions arises what happens to the phases and phase transitions when the dicke model is generalised by having more than just two ( atomic ) energy levels and one bosonic mode . in our previous study at zero temperature
, we found an additional superradiant phase as well as phase transitions of first and second order in a three - level dicke model interacting with two bosonic modes in lambda - configuration .
these finding where also of interest in the context of the discussion of the no - go theorem @xcite for superradiant phase transitions in this and other generalised dicke models @xcite .
our paper is organised as follows : first , we give a short review of the model and previous findings for zero temperature @xcite .
then , the partition sum of the model is calculated in the thermodynamic limit from which we identify the thermodynamic phases relevant expectation values .
we discuss the properties of the phases and the phase transitions and finally recover the zero - temperature results as a limiting case of our theory .
is coupled via the two bosonic modes with frequencies @xmath0 and @xmath1 to either of the two ground - states @xmath2 and @xmath3 , respectively .
the corresponding coupling strengths are given by @xmath4 and @xmath5.,width=226 ] we consider a system of @xmath6 three - level systems ( the _ particles _ ) in lambda - configuration ( cf .
[ fig : model ] ) : two ground - states @xmath2 , @xmath3 with energies @xmath7 , @xmath8 , respectively , are coupled via two bosonic modes to the excited state @xmath9 with energy @xmath10 .
the two bosonic modes have frequencies @xmath0 , @xmath1 , respectively .
the hamiltonian is given by ( cf .
@xcite ) @xmath11 here , @xmath12 , @xmath13 and @xmath14 is the coupling strength of @xmath15 bosonic mode .
the operators @xmath16 are collective particle operators and can be written in terms of single - particle operators @xmath17 , @xmath18 the operator @xmath17 acts on the degrees of freedom of the @xmath19 particle only and can be represented by @xmath20 , where @xmath21 is the @xmath22 state of the @xmath23 three - level system .
we call the single - particle energy levels @xmath2 ( @xmath3 ) and @xmath9 together with the first ( second ) bosonic mode , the left ( right ) branch of the lambda - model . in a previous work
@xcite , we have studied the ground - state properties of the hamiltonian , eq . , as well as collective excitations above the ground - state .
the model shows three phases : a normal phase and two superradiant phases .
both types of phases show their distinctive features as in the original dicke model @xcite .
the normal phase has zero occupation of the bosonic modes and all three - level systems occupy their respective single - particle ground - state @xmath2 .
in contrast , the superradiant phases are characterised by a macroscopic occupation of the bosonic modes and the three - level systems .
the superradiant phases are divided into a so - called blue and red superradiant phase .
for the blue superradiant phase , the left branch of the lambda - model is macroscopically occupied , whereas for the red superradiant phase it is the right branch which shows macroscopic occupation .
the phase diagram features phase transitions of different order ; the phase transition between the blue superradiant phase and the normal phase is continuous .
in contrast , for finite @xmath24 , the phase transition between the normal and the red superradiant phase is of first order .
eventually , in the limit @xmath25 this phase transition becomes continuous as well .
the two superradiant phases are separated by a first order phase transition , irrespective of the parameters of the model .
all thermodynamic information of the equilibrium system is contained in the partition sum @xcite , @xmath26 \}$ ] .
here , @xmath27 is the temperature and @xmath28 , with boltzmann s constant @xmath29 . in order to evaluate the trace
, we represent the bosonic degrees of freedom by coherent states @xcite with respect to the @xmath30 and the trace of the particle degrees of freedom are split into single - particle traces @xmath31 , @xmath32 both integrals extend over the complex plane , respectively . due to the coherent states ,
the bosonic part of the trace is easily computed . for the particle part of the trace , we observe that the @xmath6 single - particle traces are all identical . in addition ,
we decompose both @xmath33 and @xmath34 in its real and imaginary part and scale them with @xmath35 , @xmath36 considering the thermodynamic limit @xmath37 , the partition sum eventually reads @xmath38 here , the single - particle hamiltonian @xmath39 is given by @xmath40 since all particles are identical , we have omitted the superindex @xmath41 at the single - particle operators @xmath17 .
the remaining single - particle trace is evaluated in the eigenbasis of the single - particle hamiltonian , @xmath42 where we have chosen a convenient basis for the matrix representation of @xmath43 . by virtue of cardano s formula
, the eigenvalues of @xmath44 can be calculated exactly .
however , the discussion of whether there is a phase transition or not and the analysis of the phase transition , is not very transparent .
therefore , we will pass the general case to a numerical computation and first consider the special case with @xmath45 only , which is amenable to analytical calculations .
for @xmath46 , the partition sum can be written in compact form as ( again in the thermodynamic limit @xmath37 ) @xmath47 with @xmath48\end{gathered}\ ] ] and a corresponding @xmath49 given by @xmath50 the remaining integrals can not be done exactly .
however , we can approximate them for large @xmath6 by laplace s method @xcite . since large values of @xmath51 are exponential suppressed
, the main contribution to the integral is given by the global minimum of @xmath51 .
given the minimum , the partition sum is proportional to @xmath52 where @xmath53 with the position the minimum @xmath54 .
we anticipate that both @xmath55 are identical zero at the minimum .
this will be shown below .
hence , the leading contribution to the free energy @xmath56 @xcite is given by @xmath57 .
before we determine the minimum of @xmath51 , we will first compute expectation values of observables using the same approximations as above . in order to identify the phases and phase transitions , we discuss several observables . of interest
are the occupations of the bosonic modes , @xmath58 , and the three - level systems , @xmath59 .
in addition , the quantities @xmath60 and @xmath61 are considered .
the real part of these give the macroscopic polarisations of the three - level systems of the left and the right branch of the lambda - model , respectively .
they are generalisations of the polarisation in the dicke model . there , the polarisation is proportional to the expectation value of the @xmath62 component of the atomic pseudo spin operator . for functions @xmath63 of operators of the two bosonic modes ,
the expectation value is given by ( the calculation can be found in the appendix [ sec : appa ] ) @xmath64 with that , we obtain for the occupation of both modes @xmath65 [ sec : lm_expvalatoms ] expectation values of collective operators of the three - level systems can be calculated in a similar manner .
let @xmath66 be a collective operator and @xmath67 the corresponding single - particle operator for the @xmath22 three - level system , such that @xmath68 then , the mean value of @xmath66 is given by ( the calculation can be found in the appendix [ sec : appb ] ) @xmath69 here , @xmath70 is the single - particle expectation value for a thermal state with the single - particle hamiltonian @xmath44 evaluated at the minimum @xmath54 of @xmath51 .
thus , the expectation values for the collective atomic operators @xmath71 can be traced back to the single - particle operators @xmath43 , @xmath72 this derivation and the following calculation holds even for non - zero @xmath24 .
let @xmath73 be the eigenvalues and @xmath74 the corresponding eigenvectors of @xmath75 .
then we evaluate the traces in eq . in this eigenbasis and the expectation values can be written as @xmath76 here @xmath77 \}$ ] is the partition sum of the single - particle hamiltonian @xmath75 and we have used that the matrix elements of @xmath43 are given by zeros , except for the entry of the @xmath22 row and @xmath78 column which is one .
we continue with the calculation of the partition sum , eq . .
therefore , we need to find the minimum of @xmath51 , eq .
. this analysis will be done in the following .
the variable @xmath55 enters only quadratically in @xmath51 , such that upon minimising @xmath51 , both @xmath55 need to be zero . minimising @xmath51 with respect to @xmath79 yields the two equations ( @xmath80 ) @xmath81,\ ] ] with @xmath82}{1 + 2 \mathrm e^{-\beta
\delta / 2 } \cosh \bigl [ \frac{\beta \delta}{2 } \omega \bigr]},\ ] ] and @xmath83 of course , eqs
. are always solved by the trivial solutions @xmath84 . but do non - trivial solutions exist , and for which parameter values ? we first observe that the eqs . do not support solutions where both @xmath85 and @xmath86 are non - zero . for given @xmath85 and @xmath86 ,
the parameter @xmath87 is fixed .
then the squared bracket can not be zero for both equations .
hence , the non - trivial solutions are given by one @xmath79 being zero and the other being finite .
in the following , the non - zero solution will be called @xmath88 . to check whether a non - zero @xmath88 really exists
, we have to analyse the equation @xmath89 with @xmath90 the function @xmath91 is bounded by one ( see left panel of fig . [
fig : lmdiscusspt ] ) and @xmath49 itself is always greater or equal one .
therefore , for @xmath92 , eq . has no solution and @xmath79 has to be zero as well .
( left ) and @xmath93 ( right ) for a certain @xmath94 and fixed temperature .
the red straight line corresponds to the first term , @xmath95 , the other blue curved line corresponds to @xmath91 of eq . involving the hyperbolic functions . of physical relevance
is the region @xmath96 only .
thus , for @xmath97 no physical solution is possible , whereas for @xmath98 a physical solution might exist.,width=321 ] on the other hand , for @xmath98 , non - trivial solutions of eq . can exist . in the right panel of fig .
[ fig : lmdiscusspt ] , both terms of eq . are drawn .
we see that for every finite temperature , the two curves always intersect twice , so that eq . always has two solutions . of course , for a differentiable @xmath51
, the two solutions can not both correspond to minima of @xmath51 .
hence , one solution stems from a maximum and the other from a minimum . since the right side of eq
is the derivative of @xmath51 , its sign - change signals whether a maximum ( plus - minus sign change ) or a minimum ( minus - plus sign change ) is passed when @xmath49 is increased .
therefore , the first solution corresponds to a maximum and the second to a minimum .
we gain additional insight , if we directly analyse @xmath99 for different coupling strengths [ @xmath100 from eq . ,
@xmath101 . this
is shown in fig .
[ fig : lmf_plots ] .
we see that for small coupling strengths , @xmath99 has one local minimum only which is located at @xmath102 , the trivial solution . if the coupling strength is increased , a maximum - minimum pair forms at finite values of @xmath79 . in general , this local minimum at @xmath103 is energetically higher than the local minimum of the trivial solution at @xmath102 , cf .
[ fig : lmf_plots ] .
hence , the trivial solution still minimises @xmath99 globally .
however , if the coupling strength is increased even further , the local minimum at @xmath104 becomes the global minimum .
so we see that the position @xmath88 of the global minimum jumps at a certain value of the coupling strength from zero to a finite value . of eq . with @xmath105 , and @xmath106 for fixed temperature
the coupling strength @xmath107 increases from the upper to the lower curves .
the units are arbitrary and @xmath51 has been rescaled for comparison , such that @xmath108 for all coupling strengths @xmath107 . on increasing @xmath107 , a local minimum forms distant from the origin . for large @xmath107
this minimum eventually becomes a global minimum.,width=226 ] the above discussion refers to the case @xmath46 .
however , for finite @xmath24 , the results are qualitatively the same .
we discuss the properties of the phase transition for finite @xmath24 below .
in the above analysis , we have shown the existence of three different minima of @xmath51 appearing in the exponent of the integrand of the partition sum .
we have also shown that for a given temperature @xmath27 , we find coupling strengths @xmath109 , below which the trivial solution @xmath110 minimises @xmath51 . in eq .
, @xmath88 measures the macroscopic occupation of the bosonic modes and can therefore serve as an order parameter of the phase transition . hence , in the parameter regime where the trivial solution minimises @xmath51 , the system is in the normal phase .
in addition , above the coupling strengths @xmath109 , @xmath51 is minimised by non - zero values of either @xmath111 or @xmath112 .
the first corresponds to the red superradiant phase , the latter to the blue superradiant phase of ref .
the two superradiant phases at non - zero temperatures show only one macroscopically occupied bosonic mode as well ; mode one for the blue superradiant phase and mode two for the red superradiant phase , while the occupation of the other mode is zero . as we have seen in the previous section , the location of the global minimum @xmath88 jumps from the trivial solution to non - zero solutions .
a jump of the order parameter defines a first - order phase transition .
therefore , the finite - temperature superradiant phase transition in the lambda - model is a first - order phase transition .
hence , the continuous phase transition at zero temperature of ref .
@xcite transforms into a first order phase transition at finite temperatures .
so far , we have identified the phases and phase transitions using the occupation of the bosonic modes . in the following ,
we will further characterise the phases using observables of the three - level systems . in the normal phase with @xmath110 , the hamiltonian @xmath75 is diagonal and the eigenvectors @xmath113
are given by cartesian unit vectors .
hence , the expectation value of all collective operators @xmath71 with @xmath114 vanish .
conversely for the diagonal operators @xmath115 , the occupations ; their expectation values are given by @xmath116 here , we explicitly see that in the normal phase the expectation values are independent of the coupling strengths @xmath4 and @xmath5 .
furthermore we note that for finite temperatures , in addition to the single - particle ground state @xmath2 , the energetically higher lying single - particle states @xmath3 and @xmath9 are macroscopically excited as well . hence in contrast to the bosonic modes , the particle part of the system gets thermally excited . from that point of view
, i.e.concerning the populations of the single - particle energy levels , the particle system in the normal phase behaves like a _ normal _ thermodynamical system . for the superradiant phases
, we can not give explicit expressions for the expectation values , neither for finite or vanishing @xmath24 .
that is because we need to compute the minimum of @xmath51 numerically .
though for @xmath46 , we can say that some expectation values are exactly zero .
this will be done next , separately for the red and the blue superradiant phases .
[ [ red - superradiant - phase ] ] red superradiant phase + + + + + + + + + + + + + + + + + + + + + + first consider the red superradiant phase with @xmath117 and @xmath118 .
then , the single - particle hamiltonian @xmath119 reads [ cf .
. ] @xmath120 and its exponential has the form @xmath121 the matrix elements @xmath122 and @xmath123 are given by ( the matrix element @xmath124 is needed below ) .
@xmath125 \pm \frac{\sinh \bigl [ \frac{\beta \delta \omega}{2 } \bigr]}{\omega } \right ) , \label{eq : lm_apm } \\
b_n & = - \frac{4 g_n y_n}{\delta \omega } \mathrm e^{-\frac{\beta \delta}{2 } } \ , \sinh \bigl [ \frac{\beta \delta \omega}{2 } \bigr ] , \quad n=(1,2 )
. \label{eq : lm_b12}\end{aligned}\ ] ] the product of the exponential operator , eq .
, with matrices of the form @xmath126 is traceless .
therefore , the expectation values of the collective operators @xmath127 , @xmath128 and their hermitian conjugates are zero , i.e.there is no spontaneous polarisation between both the single - particle states @xmath2 and @xmath3 , and the single - particle states @xmath3 and @xmath9 .
contrary , the polarisation in the left branch of the lambda - model , i.e.between the states @xmath2 and @xmath129 , is finite and macroscopic .
[ [ blue - superradiant - phase ] ] blue superradiant phase + + + + + + + + + + + + + + + + + + + + + + + for the blue superradiant case , the discussion is similar .
here we have @xmath130 and @xmath131 , and the exponential of the single - particle hamiltonian reads @xmath132 the matrix elements are given above , eqs , .
now , the product of the exponential operator , eq . , with matrices of the form @xmath133 is traceless and thus expectation values of the collective operators @xmath127 , @xmath134 and their hermitian conjugates are zero .
on the other hand , the expectation value of the operators @xmath128 , @xmath135 is finite and macroscopic .
hence , only the transition in the right branch of the lambda - model is spontaneously polarised . in conclusion
, we find that in the superradiant phases at finite temperature , only the corresponding branch of the lambda - model shows spontaneous polarisation ; the left branch in the red superradiant phase and the right branch in the blue superradiant phase . in the normal phase ,
the polarisation is completely absent .
hence , in contrast to the populations of the atomic system , the polarisations are not thermally excited and show a genuine quantum character .
thus , both the polarisations and the occupations of the two resonator modes show a similar behaviour in the three phases .
therefore we have two sets of observables , the polarisations for the three - level systems and the occupations of the bosonic modes , to detect the superradiant phase transition at finite temperatures .
the above analysis for vanishing @xmath24 already shows that the phase transition in the lambda - model for finite temperatures is a first - order phase transition .
this fact renders the calculation of the exact location of the phase transition with our methods impossible .
this can be understood with the help of the free energy as follows . in the thermodynamic limit
, the global minimum of the free energy defines the thermodynamic phase of the system .
we explicitly saw this when we have computed the partition sum . in a phase transition ,
the system changes from one thermodynamic state to another thermodynamic state .
this new state corresponds to a different , now global minimum of the free energy . for continuous phase transitions ,
the new minimum evolves continuously from the first minimum and the first minimum changes its character to a maximum .
hence , the continuous phase transition is characterised by a sign - change of the curvature of the free energy at the position of the minimum of the state describing the normal phase . often , this is tractable analytically .
in contrast in the case of first - order phase transitions , the new global minimum of the free energy appears distant from the old global minimum of the free energy , cf .
[ fig : lmf_plots ]
. there are still two minima and we can not detect the phase transition by the curvature of the free energy .
thus to find the phase transition for first - order phase transitions , we first need to find all minima of the free energy and then find the global minima of these .
this has to be done numerically here .
of the two bosonic modes for @xmath136 ( left ) and @xmath137 ( right ) .
the parameters are set to @xmath138 , @xmath139 , @xmath140 , and @xmath141 .
the computation has been done in the thermodynamic limit using laplace s method.,width=321 ] for the numerical computation , we do not solve eq . , but we test for the minima of @xmath142 , eq . , directly .
therefore , we apply a brute - force method , i.e.we look for the smallest value of @xmath142 on a @xmath85@xmath86 grid . due to the reflection symmetry of @xmath143
, we can confine the grid to positive values for @xmath85 and @xmath86 .
this yields the position of the minimum @xmath144 .
then we compute the eigenvalues and eigenvectors of the hamiltonian @xmath44 , at this point and obtain via eq .
the expectation values for the operators of the bosonic modes , and via eq .
the corresponding expectation values for the three - level systems . of the single - particle energy levels of the three - level systems for @xmath145 ( left ) and @xmath137 ( right ) .
parameters as in fig .
[ fig : lm_modeoccts].,width=321 ] , @xmath146 of the three - level systems for @xmath145 ( left ) and @xmath137 ( right ) . parameters as in fig .
[ fig : lm_modeoccts].,width=321 ] fig .
[ fig : lm_modeoccts ] , [ fig : lm_atomoccts ] , [ fig : lm_atompolts ] show the occupation @xmath147 of the modes of the resonator , the occupation @xmath115 of the single - particle levels of the atoms , and the polarisations @xmath134 , @xmath128 of the atoms for low and high temperatures , respectively .
all plots have been generated numerically for finite values of @xmath24 .
these figures corroborate our findings from the analytical discussion of the partition sum for vanishing @xmath24 .
we see three phases : a _ normal _ phase for coupling strengths @xmath4 and @xmath5 below the critical coupling strengths @xmath148 and @xmath149 , a _ red superradiant _ phase for large coupling strengths @xmath4 above the critical coupling strength @xmath148 , and a _ blue superradiant _ phase for coupling strengths @xmath5 above the critical coupling strength @xmath149
. the normal phase is characterised by a zero occupation of both bosonic modes ( fig .
[ fig : lm_modeoccts ] ) in addition , the polarisation , or coherence , of the three - level systems is zero in the normal phase ( fig .
[ fig : lm_atompolts ] ) .
in contrast , to the normal phase , the two superradiant phases are characterised by a macroscopic occupation of only one of the two bosonic modes ; mode one in the red superradiant phase and mode two in the blue superradiant phase .
in addition , the red ( blue ) superradiant phase shows a spontaneous polarisation @xmath60 ( @xmath61 ) .
we see that these defining properties remain for increasing temperature ( right part of figs .
[ fig : lm_modeoccts]-[fig : lm_atompolts ] ) . as discussed in sec .
[ sec : normalphasee ] , we see that the population @xmath150 of the single - particle energy level @xmath3 increases for rising temperature .
the same is true for the occupation @xmath151 , though this is not visible in the right part of fig .
[ fig : lm_atomoccts ] due to the fact that the temperature is yet too small . from the figs .
[ fig : lm_modeoccts]-[fig : lm_atompolts ] we also see that the shape of the phase boundary remains a straight line between the normal and the two superradiant phases . between the red and the blue superradiant phases ,
the form of the phase boundary seems to persist as well .
the only effect of the rising temperature is a shift of the phase boundary towards higher values of the coupling strengths @xmath4 and @xmath5 .
this is visualised in fig .
[ fig : lm_a13_g1_t ] where the polarisation @xmath60 of the transition @xmath152 of the three - level systems is shown for variable coupling strength @xmath4 and temperature @xmath27 .
the coupling strength of the second mode is fixed to @xmath153 .
we see that for increasing temperature , the superradiant phase diminishes . of the transition @xmath152 of the three - level system .
parameters as in fig .
[ fig : lm_modeoccts ] and @xmath154.,width=283 ] of the first bosonic mode as a function of the coupling strength @xmath4 for fixed coupling strength @xmath155 and rising temperature ( left to right , blue to red , respectively ) .
the other parameters are set to @xmath138 , @xmath139 , @xmath140 , and @xmath141.,width=283 ] in addition to the shift of the phase boundary , the jump in the observables at this first - order phase transition increases .
this is shown in fig .
[ fig : lm_n1jumps ] for the occupation @xmath156 of the first bosonic mode .
of course , numerically , jumps are hard to detect since we get a discrete set of points as an output anyway
. however , the dotted lines in fig . [ fig : lm_n1jumps ] connect two largely separated points ; each of the lines consists of 1000 data points .
thus we can really speak of jumps in the observables and thus of a first - order phase transition .
lastly , we analyse the zero - temperature limit of the lambda - model for @xmath46 . for decreasing temperature , the function @xmath91 , eq .
, becomes more and more step function like .
indeed , for @xmath157 , @xmath91 can be written as a fermi function , and eventually in the limit @xmath158 , @xmath91 is given by @xmath159 hence at zero temperature , eq . has always a unique solution for coupling strengths @xmath98 .
in addition , since @xmath160 shows no additional maximum , we have a continuous phase transition in this quantum limit .
furthermore , we can also compute the position of the minimum of @xmath51 . in the limit @xmath161 , eq .
reads @xmath162 solving for @xmath79 , we obtain @xmath163 if we identify @xmath79 with the mean - fields @xmath164 of ref .
@xcite , we reproduce our results @xcite for the superradiant phases of the quantum phase transition of the lambda - model . the mean - fields @xmath165 of ref .
@xcite can be reproduced as well .
consider for instance @xmath166 which is related to @xmath151 through @xmath167 , except for a possible phase .
using the results for the boltzmann operator in the red superradiant phase , eq . , plus the above expression for @xmath85 , eq .
, and finally insert everything into the expectation value of eq . , we obtain @xmath168 \omega},\ ] ] which in the zero - temperature limit @xmath169 reduces to @xmath170 here , @xmath171 from eq . .
if we finally insert the position @xmath85 of the minimum of the free energy , eq . , the population of the third single - particle energy level in the zero - temperature limit is given by @xmath172\ ] ] which agrees with the findings of ref .
@xcite for @xmath166 .
applying the same technique , we can obtain the expectation values of all other collective atomic operators @xmath71 in both superradiant phases .
this again coincides with the results of ref .
we have analysed the lambda - model in the thermodynamic limit at finite temperatures using the partition sum of the hamiltonian .
compared to the quantum phase transition of this model @xcite , we found that at finite temperatures the properties of the phases and phase transition partially persist .
namely , we found three phases : a normal and two superradiant phases .
these have the same properties as in the quantum limit . for small couplings and/or at high temperatures ,
the system is in the normal phase where all particles are in their respective single - particle ground - state and both bosonic modes are in their vacuum state . if one of the coupling strengths is increased above a temperature - dependent critical coupling strength , the system undergoes a phase transition into a superradiant phase wich is characterised by a macroscopic occupation of one of the bosonic modes only and a spontaneaous polarisation of the corresponding branch of the three - level system .
a new characteristic of the phase transition at finite temperatures is the appearance of first - order phase transitions only . for the quantum phase transition we already found first - order phase transitions between the normal and the red superradiant phase and between the two superradiant phases .
here , at finite temperatures , the phase transition from the normal to the blue superradiant phase becomes a first - order phase transition as well .
this change of the order of the phase transition would appear for a single bosonic mode as well .
hence , it is due to the additional single - particle energy level that first - order phase transitions show up .
in addition , we emphasise that even in the degenerate limit , @xmath173 , the phase transitions from the normal to both superradiant phases are of first order .
it is remarkable that in contrast to the original dicke model , the mean - field phase transitions at finite temperatures are not continuous .
this facet becomes significant if _ real _ atoms and photons are considered . here
, for the original dicke model with its continuous phase transition , there exists a no - go theorem @xcite .
however for first - order mean - field quantum phase transitions , it is known @xcite that this no - go theorem does not apply and superradiant phase transitions occur .
we expect an identical conclusion for the lambda - model at finite temperatures .
this work was supported by the deutsche forschungsgemeinschaft within the sfb 910 , the grk 1558 , and the projects br 1528/82 and 1528/9 .
we calculate the expectation value of functions @xmath63 of operators of the two bosonic modes as , @xmath174 in the last step we used that the exponential dominates for large @xmath6 .
hence , the function @xmath63 can be considered constant and the remaining integral plus the prefactor is equal to the partition sum . | we analyse the thermodynamic properties of a generalised dicke model , i.e.a collection of three - level systems interacting with two bosonic modes .
we show that at finite temperatures the system undergoes first - order phase transitions only , which is in contrast to the zero - temperature case where a second - order phase transition exist as well .
we discuss the free energy and prominent expectation values .
the limit of vanishing temperature is discussed as well . |
gait optimization algorithms for locomoting systems must contend with a number of nonlinear effects , including : shape - dependent system dynamics ( and thus a control structure that depends on the state of the controller and not just the plant ) ; and history - dependent input - output mappings ( e.g. if one part of the gait rotates the system , future translational sections will propel the system in a new direction ) .
the geometric mechanics community has long recognized that lie brackets of the system dynamics can be used to provide an understanding of these nonlinearities , and thereby to identify shape oscillations that produce useful net displacements through the world @xcite . for kinematic locomoting systems with two shape variables ( a class which includes a number of informative systems in viscous , granular , inertial , and fluid environments @xcite )
, the lie bracket can be visualized as a set of `` height functions '' over the shape space , whose topographies identify good placements for gait trajectories . for example , gaits that move a system forward encircle positive peaks ( or negative valleys ) in the @xmath0 component of the lie bracket , whereas gaits that turn the system with minimal translation enclose sign - definite regions in the @xmath1 component and balanced positive and negative regions in the @xmath0 and @xmath2 components .
an enduring goal @xcite of this community has been to go beyond simply identifying the direction in which small shape oscillations will propel the system , and to use the lie bracket principle to identify optimal gaits that generate the most displacement , either per gait cycle or per energy expended . for example , a gait that completely encircles a sign - definite region of a height function ( by tracing out its zero - contour ) would maximize displacement in the corresponding direction ; similarly , a contraction of the zero contour that balances the enclosed area against the perimeter length ( effort ) required to capture it would maximize the efficiency with which this displacement is achieved @xcite . a key obstacle to reaching this goal was the noncommutativity in the equations of motion for most interesting locomoting systems .
because translations and rotations do not commute with each other , summing up lie brackets over a finite region provides only an approximation of the displacement induced by a gait , and the error in this approximation increases with the amplitude of the oscillation .
historically ( e.g. , @xcite ) , this error was regarded as growing too quickly to provide any meaningful information about the optimal gait cycles . in @xcite , however , we introduced a choice of coordinates that converts much of the system s noncommutativity into nonconservativity , which _ is _ amenable to finite - scale integration . by working in these new coordinates ,
we have gained insight into the optimal gaits of swimmers and crawlers in low- and high - viscosity fluids @xcite , and in granular media @xcite .
further , we have used this insight to easily `` hand - draw '' optimized gaits that would otherwise only have been found by a long numerical search of a high - dimensional space of trajectory parameters @xcite . in this paper , we build on our previous work by encoding the geometric insight described above into a variational gait optimizer .
the chief elements of this optimizer are : a gradient ascent / descent component that pushes the cycle to enclose a large sign - definite region of the lie bracket to maximize the net displacement it generates ; a cost component based on a riemannian distance metric that limits the growth of the gait cycle ; anda perimeter - balancing component that evenly - spaces the waypoints in the trajectory , stabilizing the solution and providing an efficiency - optimal parameterization of the resulting motion . as illustrated in fig .
[ fig : overview ] , these components together form a dynamical system analogous to a soap bubble , whose boundary ( here , the gait curve ) is pushed outward by internal pressure ( here , the lie bracket ) and constrained by surface tension ( here , the metric - weighted pathlength ) .
the perimeter - balancing term corresponds to the concentration gradient that evenly distributes soap across the surface of the bubble . as a demonstration of this approach ,
we use it to identify optimal gaits for a set of example systems .
these systems include purcell s three - link swimmer @xcite ( a standard benchmark for locomotion analysis ) and a continuous - curvature extension of purcell s swimmer @xcite ( to emphasize that `` two shape variables '' does not restrict the applicability of this work to systems with `` two joints '' ) .
we identify optimal gaits for these systems in both the forward and turning directions , which match those found previously in works such as @xcite .
when analyzing a locomoting system , it is convenient to separate its configuration space @xmath3 ( i.e. the space of its generalized coordinates @xmath4 ) into a position space @xmath5 and a shape space @xmath6 , such that the position @xmath7 locates the system in the world , and the shape @xmath8 gives the relative arrangement of the particles that compose it
. the structure of a ( trivial , principal ) _ fiber bundle _ , with @xmath5 the _ fiber space _ and @xmath6 the _ base space_. ] for example , the positions of both the three - link system and the serpenoid swimmer in fig . [ fig:3link ] are the locations and orientations of their centroids and mean orientation lines , @xmath9 .
the shape of the three - link system is parameterized by the two joint angles , @xmath10 , and the serpenoid s shape is described by a pair of modal amplitudes multiplied by the curvature modes @xmath11 and @xmath12 @xcite that together specify the amplitude and phase of its body wave . a useful model for locomotion in
_ kinematic _ regimes where no gliding can occur , and which we employ in this paper , is that at each shape , there exists a linear relationship between changes in the system s shape and changes in its position , = - ( ) , [ eq : kinrecon ] ( a ) and that for the serpenoid swimmer are shown in fig .
[ fig : connection3link](b).note that scale in vector fields has been chosen to emphasize structure , and that scales in different components or systems are not comparable . also note qualitative similarity ( modulo rotation ) across systems ] in which @xmath13 is the body velocity of the system ( i.e. , @xmath14 expressed in the system s local coordinates ) , and the _ local connection _
@xmath15 acts like the jacobian of a robotic manipulator , mapping from joint velocities to the corresponding body velocity .
each row of @xmath16 can be regarded as a body - coordinates gradient of one position component with respect to the system shape , as illustrated for the three - link and serpenoid swimmers in fig .
[ fig : connection3link ] . in a drag - dominated environment
, the effort required to change shape can be modeled as the pathlength @xmath17 of the trajectory through the shape space , weighted by a riemannian metric @xmath18 as @xmath19 as discussed in @xcite , the metric tensor @xmath18 is the same as the map from shape velocity to power dissipated into the surrounding medium , p = , and so can be readily calculated from the first - principles physics of the system in this paper , we generate the dynamics for our example systems from a resistive force model , in which each element of the body is subject to normal and tangential drag forces proportional to their velocities in those ( local ) directions . the normal drag coefficient is larger than the tangential component ( here , by a factor of @xmath20 ) , corresponding to the general principle that it is harder to move an object in a fluid or on a surface crosswise than it is to move it along its length .
we then impose a quasi - static equilibrium condition that the net drag force and moment on the system is zero at all times ( treating the system as heavily overdamped , with acceleration forces much smaller than drag forces ) ; because the drag forces are not isometric , the system can use the angle - of - attack of its body surfaces to generate net motion .
together , these conditions impose a _ _ pfaffian constraint__. ] on the system s generalized velocity , f^b_x + f^b_y + f^b _ = 0 + 0 + 0 = ( ) + , [ eq : forcepfaffian ] in which the matrix @xmath21 that maps the velocities to the net forces on the body frame is an expression of the system s internal kinematics and depends only on the shape @xmath22 . by separating @xmath21 into two sub - blocks , @xmath23 $ ] ,
it is straightforward to rearrange into = -(_^-1_),[eq : invomega ] revealing the local connection as @xmath24 . for a more detailed treatment this process , we refer the reader to @xcite .
once @xmath15 has been found , it can be used to calculate the riemannian metric @xmath18 as p = f^b = = _ .
we apply this physics models to two example geometries : purcell s three - link swimmer @xcite , and a _ serpenoid swimmer
_ @xcite .
the three - link swimmer is a useful and widely - adopted @xcite minimal example for locomotion , because the two degrees of freedom can be easily visualized .
the serpenoid swimmer , whose shape is defined by the amplitude of sine and cosine curvature modes @xcite , provides an example of a two - dof system that has been shown to closely model how snakes and other animals use undulatory locomotion to move through the world @xcite .
optimal gait design has a long history of research in the physics , mathematics , and engineering communities , as part of the broader field of optimal control @xcite .
for systems of the classes we consider here , notable contributions include those of purcell , who introduced the three - link swimmer as a minimal template for understanding locomotion , a series of works @xcite aimed at numerically optimizing the stroke pattern , and the observation in @xcite that the optimal pacing for the gait keeps the power dissipation constant over the cycle . ) for the three link swimmer are shown in fig.[fig : heightfun](a ) and that for the serpenoid swimmer are shown in fig.[fig : heightfun](b ) .
note the qualitative similarity ( modulo a rotation ) between the two sets of height functions . ]
several efforts in the geometric mechanics community ( including our own ) , have aimed to use the structure of the systems _ lie brackets _ ( a measure of how non - canceling " the system dynamics are over infinitesimal cyclic inputs ) to understand the structure of the optimal solutions to the system equations of motion @xcite .
the core principle in these works is that because the net displacement @xmath25 over a gait cycle @xmath26 is the line integral of along @xmath26 , the displacement can be approximated by an area integral of the curvature @xmath27 of the local connection ( its total lie bracket @xcite ) over a surface @xmath28 bounded by the cycle : @xmath29}}_{\text{$-d\mixedconn$ ( total lie bracket ) } } , \label{eq : lie}\end{aligned}\ ] ] in which @xmath30 is the exterior derivative of the local connection ( its generalized row - wise curl ) and the local lie bracket term evaluates ( on @xmath31 ) as _
1,_2 = ^y_1^_2- ^y_2^_1 + ^x_2^_1- ^x_1^_2 + 0 . plotting these curvature terms as height functions ( as in fig .
[ fig : heightfun ] ) reveals the attractors that influence the optimal gait cycles : gaits that produce net displacement in a given @xmath32 direction are located in strongly sign - definite regions of the corresponding @xmath27 height function . for example , @xmath0-translation gaits encircle the center of the shape space for both the three - link and serpenoid systems , whereas @xmath2-translations or @xmath1-rotations are produced by cycles in the corners or edges of the shape space . in @xcite , we identified coordinate choices that make the approximation in accurate for large - amplitude gaits , which allowed us to make two extensions to the area principles : first , that maximum - displacement gaits ( the `` longest strides '' that the systems can take ) follow the zero - contours of the height functions , completely enclosing a sign definite region .
second , that the most efficient gaits ( the systems `` comfortable strides '' ) are contractions / straightenings of these zero - contours ; by giving up the low - yield regions at the edges of the sign - definite regions ( or crossing slightly into opposite - sign regions ) the system travels a shorter distance in each cycle , but the consequently shorter perimeter length means the system can repeat the cycles more quickly at a given level of power consumption .
our aim with this paper is to encode the geometric principles described in [ sec : gaitoptimization ] into a gait optimization algorithm .
we start from the basic variational principle that functions reach their extrema when their derivatives go to zero . given a gait parameterization @xmath33
, maximum - displacement cycles therefore satisfy the condition that the gradient of net displacement with respect to the parameters is zero , [ eq : maxdispcond ] _ = and maximum - efficiency cycles ( normalizing displacement by the pathlength - effort required to generate it ) likewise satisfy the condition that the gradient of this efficiency ratio is zero , [ eq : maxeffcond ] _
= _ -_= . for suitable seed values @xmath34 ,
solutions to and can therefore be reached by finding the respective equilibria of the dynamical systems [ eq : gaitequilibria ] = _ = _ . as discussed in the following subsections
, we then incorporate the geometric principles in and to evaluate the @xmath35 and @xmath36 terms in as functions of the system s lie bracket @xmath27 and riemannian metric @xmath18 .
we also use the metric @xmath18 to introduce a third term , @xmath37 , to .
this term acts orthogonally to the optimization process , but serves to enforce an even sampling of points around the perimeter of the gait .
this even distribution promotes stability of the optimizer and additionally corresponds to optimal pacing ( internal timing ) of the gait , as discussed in @xcite .
collecting all three terms and factoring out a coefficient of @xmath38 from , our algorithm thus finds the maximum - efficiency gait as the equilibrium of [ eq : fullequation ] = _ -_+ _ . as illustrated in fig .
[ fig : overview ] , this differential equation is directly analogous to the equations governing the shape of a soap bubble : @xmath35 takes the lie bracket as an `` internal pressure '' seeking to expand the gait cycle , @xmath36 is the `` surface tension '' that constrains the growth of the bubble , and @xmath37 is the `` concentration gradient '' that spreads the soap over the bubble s surface . in the following subsections ,
we explore each of the terms in , discussing both their fundamental geometric definitions and how they can be implemented in a direct - transcription solver that parameterizes the gait as a sequence of waypoints .
the first term in , @xmath35 , guides the gait towards maximum - displacement cycles . by sustituting the approximation from into this expression as _ _ _ _ a(-d ) , and noting that variations in @xmath33 affect @xmath26 but not @xmath27 , we can convert @xmath35 into gradient of an area integral with respect to variations in its boundary .
we can then invoke a powerful geometric principle , which states _ the gradient of an integral with respect to variations of its boundary is equal to the gradient of the boundary with respect to these variations , multiplied by the integrand evaluated along the boundary . _ formally , this multiplication is the _ _ interior product _ _ of the boundary gradient with the integrand , [ eq : interiorproduct ] _ _ _ a(-d ) = _ ( _ ) ( -d ) , which contracts @xmath27 ( a _ differential two - form _
@xcite ) along @xmath39 to produce a differential one - form that can be integrated over @xmath26 .
this formalism will become important in future work when we explore these principles on systems with more than two shape variables ; for the two - dimensional shape spaces we consider in this paper , the interior product reduces to a simple multiplication between the outward component of @xmath39 and the scalar magnitude of the lie bracket , [ eq : boundarygradientsimple ] _ _ _ a(-d ) = _ ( _ _ ) ( -d ) . _
implementation of the internal pressure : _ each waypoint @xmath40 forms a triangle with its neighboring points , whose base defines a local tangent direction @xmath41 as _ i+1 - _
i-1 = _ and a local normal direction @xmath42 orthogonal to @xmath41 . as illustrated in fig .
[ fig : derivation ] , the gradient of the enclosed area with respect to variations in the position of @xmath40 in the @xmath41 and @xmath42 directions is the change in triangle s area as @xmath40 moves . because the triangle s area is always one half base times height ( regardless of its pitch or the ratio of its sidelengths )
, this gradient evaluates to _ _
i_a = _ & _ 0 + /2 .
note that this term matches the right - hand side of , with only normal motions of the boundary affecting the enclosed area . the second term in takes @xmath36 as a measure of how variations in the gait affect the cost of executing it , and scales this by a factor of @xmath43 to compare how the return on this investment compares to the efficiency of the gait in its present state .
the gradient component of this term can be related to the system s riemannian metric by first incorporating the arclength calculation from , _ = _ _ ( ^d^2)^ and then applying standard calculus operations to consolidate terms . ] to arrive at [ eq : surfacetensionfull ] _ = _ ( 2 d+ ( _ ) d ) , in which the two parts of the integrand respectively measure how changes in the relative positions of the boundary elements and changes in the metric at the underlying points affect the pathlength , and hence the cost of motion .
the @xmath43 factor ( which normalizes the scales of @xmath35 and @xmath36 ) can be calculated directly from and the integral of .
although the calculation of @xmath25 could in theory make use of the area approximation in , this would be inefficient and impractical : integration of surfaces with arbitrarily complex boundaries requires significantly more computational resources than are needed for line integration around the boundary . using the true line integral
also improves the accuracy of the solution ; by continuously recalibrating to the true net displacement , the algorithm avoids compounding any errors introduced by the approximation in . _
implementation of the surface tension : _ each waypoint @xmath44 is at the head of a vector extending from @xmath45 , such that the @xmath46 vector and its gradient in can be taken as d= _ i - _ i-1 _ _ id= 1 + 1 . for computational simplicity , we evaluate the metric at the center of each segment .
this point moves at the mean speed of the segment endpoints , and its gradient with respect to changes in @xmath40 is thus one half of its gradient over the underlying space , _
these gradient relationships are illustrated in fig .
[ fig : derivation](b ) , with the metric represented by its tissot indicatrix ellipse @xcite . the third term in , @xmath37 , is included to ensure an even sampling of points around the gait during numerical optimization .
this term does not appear in or because it is orthogonal to those optimization criteria : varying a continuous curve along itself does not change its length or the region it encloses . when the curve is represented with a discrete basis , however , that basis induces a length coordinate @xmath47 along the gait that is not necessarily compatible with the natural coordinate @xmath48 .
this coordinate may lead to undersampling of some portions of the curve or pathological behavior , such as crossovers in which the gait spontaneously forms internal loops that do not correspond to features in @xmath27 . to reduce the chance of such instabilities occurring ,
we first identify an internal strain energy @xmath49 that measures how much an even spacing according to @xmath47 differs from even spacing according to @xmath48 .
this strain energy , [ eq : intstress ] = _ ( - _ ) ^2 , compares the average and local relative rates at which the two coordinates are changing around the curve ; if these rates are not matched , then the two measures of length are imbalanced . including the gradient term of this energy with respect to variations in the parameters , @xmath37 , then directs the optimizer to use its degrees of freedom that are ( instantaneously ) in the kernel of the primary optimization criteria to reduce the imbalance .
the expansion of this term depends on the choice of parameters for the gait ( because they affect the construction of @xmath47 ) , so we do not provide a general formula for it . in the case of direct transcription , however , it works out to be very straightforward . _
implementation of the concentration gradient : _ in a direct - transcription parameterization , the strain energy at a given waypoint corresponds to the difference in the tangential distance from that waypoint to each of its neighbors , _
i = ( ( _ i+1-_i ) _ - ( _ i-_i-1)_)^2 , whose gradient with respect to the position of @xmath40 is proportional to the sum of the tangential displacements of the neighboring points relative to @xmath40 , _ _
i _ & _ ( _ i+1 ) _ + ( _ i-1 ) _ + 0 .
we implemented the optimizer described in the previous section in matlab , both directly solving the differential equation in using ` ode45 ` and by providing as the gradient for the standard ` fmincon ` optimizer using the interior - point algorithm . as expected , both implementations converged on the same solutions , with the ` fmincon ` implementation completing more quickly ( on the order of minutes on a modern desktop computer for a gait with @xmath50 waypoints ) due to its ability to take larger steps through the parameter space .
we ran the optimizer on the three link system twice , first to find the gait that maximizes the displacement in the @xmath0-direction over a single cycle and then to find the maximum - efficiency cycle .
the dashed line in fig .
[ fig : optimized gait](a ) shows the gait that optimizes the maximum displacement over the cycle for the three - link swimmer . as expected
, this gait follows the zero - contour of the height function .
the solid line fig .
[ fig : optimized gait](a ) shows the maximum efficiency gait for this system .
because the maximum - efficiency optimizer places a cost on pathlength , this curve gives up the low - yield regions at the ends of the cycle and crosses slightly outside the zero contour .
[ fig : optimized gait](b ) illustrates essentially - similar behavior for the serpenoid : the dashed maximum - displacement gait traces the zero contour , and the solid maximum - efficiency gait captures a more compact area within the sign - definite region .
[ fig : optimized gait](c ) shows the gait that produces the most cost - effective rotational motion of the three - link swimmer ( the most sign - definite area it could capture on the @xmath1 height function without over - stretching itself ) .
the results the optimizer yields for the three link swimmer match those obtained for maximum efficiency in @xcite and , along the @xmath0 direction , those in @xcite .
in this paper , we have incorporated recent geometric insights about optimal gaits for kinematic systems into a variational trajectory optimizer .
the dynamics of this optimizer are analogous to those of a soap bubble , with the lie bracket providing an `` inflating pressure '' to the trajectory and the riemannian metric on the shape space contributing `` surface tension '' that halts growth of the cycle in the face of diminishing returns .
together , these elements drive the gait cycle to a `` comfortable stride '' that converts shape change effort into net displacement with optimal efficiency .
we demonstrated this optimizer in operation on a number of test cases , including purcell s three - link swimmer ( a standard minimal template for locomotion modeling ) and a serpenoid swimmer ( a model widely used in studies of animals and snake robots ) .
the optimizer correctly found the maximum - displacement and maximum efficiency cycles for these systems , as corroborated by previous exhaustive optimizations of the gait cycles such as @xcite .
a line of future work that we are pursuing is to extend this optimizer to systems whose dynamics are dominated by inertial effects , including those subject to conservation of momentum whose momentum is directed by nonholonomic constraints .
the cost of a gait for these types of systems depends on the acceleration through the configuration shape during the gait rather than on the energy dissipated into the environment .
therefore we expect that a cost associated with the gait s curvature through the shape space ( as well as its length ) will play an important role in finding the optimal gait for these systems .
k. a. morgansen , b. i. triplett , and d. j. klein , `` geometric methods for modeling and control of free - swimming fin - actuated underwater vehicles , '' _ ieee transactions on robotics _ , vol .
23 , no . 6 , pp .
11841199 , jan 2007 .
[ online ] .
available : http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4399955 e. a. shammas , h. choset , and a. a. rizzi , `` geometric motion planning analysis for two classes of underactuated mechanical systems , '' _ int .
j. of robotics research _ , vol .
26 , no .
10 , pp . 10431073 , 2007 .
j. b. melli , c. w. rowley , and d. s. rufat , `` motion planning for an articulated body in a perfect planar fluid , '' _ siam journal of applied dynamical systems _ ,
vol . 5 , no . 4 , pp . 650669 , november 2006 .
j. e. radford and j. w. burdick , `` local motion planning for nonholonomic control systems evolving on principal bundles , '' in _ proceedings of the international symposium on mathematical theory of networks and systems _ ,
padova , italy , 1998 .
, `` geometric motion planning : the local connection , stokes theorem , and the importance of coordinate choice , '' _ international journal of robotics research _ , vol . 30 , no . 8 , pp .
9881014 , july 2011 .
r. l. hatton , l. j. burton , a. e. hosoi , and h. choset , `` geometric maneuverability , with applications to low reynolds number swimming , '' in _ proceedings of the ieee / rsj international conference on intelligent robots and systems _ , san francisco , ca usa , september 2011 .
r. l. hatton and h. choset , `` connection vector fields and optimized coordinates for swimming systems at low and high reynolds numbers , '' in _ proceedings of the asme dynamic systems and controls conference ( dscc ) _ , cambridge , massachusetts , usa , sep 2010 .
p. bettiol , b. bonnard , l. giraldi , p. martinon , and j. rouot , `` the purcell three - link swimmer : some geometric and numerical aspects related to periodic optimal controls , '' oct .
2015 , working paper or preprint .
[ online ] .
available : https://hal.inria.fr/hal-01143763 l. giraldi , p. martinon , and m. zoppello , `` controllability and optimal strokes for n - link microswimmer , '' in _ decision and control ( cdc ) , 2013 ieee 52nd annual conference on _ , dec 2013 , pp . 38703875 .
a. e. bryson and y .- c
. ho , _ applied optimal control , optimization , estimation , and control_.1em plus 0.5em minus 0.4emnew york - london - sydney - toronto : john wiley & sons , 1975 .
[ online ] .
available : http://dx.doi.org/10.1002/zamm.19790590826 | in this paper , we present a geometric variational algorithm for optimizing the gaits of kinematic locomoting systems .
the dynamics of this algorithm are analogous to the physics of a soap bubble , with the system s lie bracket supplying an `` inflation pressure '' that is balanced by a `` surface tension '' term derived from a riemannian metric on the system s shape space .
we demonstrate this optimizer on a variety of system geometries ( including purcell s swimmer ) and for optimization criteria that include maximizing displacement and efficiency of motion for both translation and turning motions . |
more than twenty years ago , brezin , le guillou and zinn - justin ( bgz ) studied the phase transition of a cubic anisotropic system by means of renormalization group equations @xcite . within a @xmath4-expansion
, they found that to lowest nontrivial order in @xmath5 , the only stable fixed point for @xmath6 is the @xmath7-symmetric one , where @xmath8 is the number of field components appearing in the cubic anisotropic model .
they interpreted this as an indication that the anisotropy is irrelevant as long as @xmath8 is smaller than four . for @xmath9 , the isotropic fixed point destabilized and the trajectories crossed over to the cubic fixed point .
recently , our knowledge of perturbation coefficients of the renormalization group functions of the anisotropic system was extended up to the five - loop level by kleinert and schulte - frohlinde @xcite .
since the perturbation expansions are badly divergent , they do not directly yield improved estimates for the crossover value @xmath10 where the isotropic fixed point destabilizes in favor of the cubic one .
an estimate using pad approximants @xcite indicates @xmath10 to lie below @xmath11 , thus permitting real crystals to exhibit critical exponents of the cubic universality class . for a simple @xmath12-theory
, the pad approximation is known to be inaccurate .
in fact , the most accurate renormalization group functions for that theory have been obtained by combining perturbation expansions with large - order estimates and using a resummation procedure based on borel - transformations @xcite@xcite .
it is the purpose of this paper to derive the large - order behavior of the renormalization group functions for the anisotropic @xmath13-theory . in a forthcoming paper
we will combine these results with the five - loop perturbation expansion of kleinert and schulte - frohlinde to derive the precise value for the crossover value @xmath10 . for the simple @xmath13-theory ,
the large - order behavior of perturbation coefficients has been derived by lipatov @xcite , bgz @xcite and others @xcite@xcite in a number of papers .
the generalization to the @xmath7-symmetric case was given in @xcite .
an equivalent method for calculating the large - order behavior is based on the observation that for a negative coupling constant green functions possess an exponentially small imaginary part due to the fact that the ground state is unstable @xcite .
the imaginary part is associated with the tunneling decay rate of the ground state .
it determines directly the large - order behavior of the perturbation coefficients via a dispersion relation in the complex coupling constant plane . in the semiclassical limit
, the imaginary part of all green functions can be calculated with the help of classical solutions called _
instantons_. for a massless @xmath13-theory in @xmath14 space dimensions , these instantons can be found analytically .
the imaginary part is a consequence of a negative frequency mode in the spectrum of the fluctuation operator , whose determinant enters the one - loop correction to the instanton contribution .
mckane , wallace and de alcantara bonfim @xcite found a way to continue the results of the @xmath13-theory in @xmath14 to a field theory in @xmath15 dimensions .
they proposed an extended dimensional regularization scheme for nonperturbative renormalizing the imaginary parts of vertex functions . in the present work we have to extend this scheme to the case of a @xmath16-theory with cubic anisotropy , where the energy functional has the following form : @xmath17 \,\ , .\ ] ] for @xmath18 , the corresponding model in quantum mechanics was first studied by banks , bender and wu @xcite who used multidimensional wkb - techniques to derive the large - order behavior of the perturbation expansion for the ground state energy . in @xmath19 janke
@xcite presented a more efficient calculation using a path integral approach . in the present work
, this approach will be generalized to quantum field theory and extended by a careful discussion of the region near the isotropic limit @xmath20 .
this is important , since the infrared - stable cubic fixed point is expected to appear very close to the @xmath7-symmetric one .
in fact , it will be sufficient to give the quantum - field theoretical generalization of @xcite in terms of an expansion about the isotropic case in powers of @xmath1 .
the paper is organized as follows .
the method is developed by treating first the case @xmath21 . in
we derive the feynman rules for the power series expansion of all green functions around the isotropic limit . in
we calculate the small - oscillation determinants for the transversal and longitudinal fluctuations . in
we use the extended renormalization scheme of @xcite to find the full ( real and imaginary ) vertex functions , and derive the renormalization constants to one loop . in
we calculate the imaginary parts of the renormalization - group functions and thus the large - order behavior of the perturbation coefficients .
in , finally , we extend the results to the physically relevant case @xmath22 .
for positive coupling constants @xmath23 and @xmath1 , the system defined by the energy functional ( [ hn ] ) is stable and the green functions are real . on the other hand , if both coupling constants are negative the system is unstable and the green functions acquire an imaginary part .
the corresponding functional integrals can be calculated by an analytical continuation from positive to negative coupling constants , keeping the factor @xmath24}$ ] real .
we perform this analytical continuation by means of a joint rotation in the two complex planes , substituting @xmath25 and @xmath26 , and rotating the azimuthal angle @xmath27 from @xmath28 to @xmath29 . at the same time
, we rotate the contour of integration in the field space . the convergence of the functional integrals is maintained by the field rotation @xmath30 .
a natural parameter for the anisotropy of the system is the ratio @xmath31 .
the isotropic limit corresponds to @xmath32 . during the joint rotation of @xmath23 and @xmath1 ,
the parameter @xmath33 remains constant .
thus , @xmath33 is a good parameter for the anisotropy at both positive and negative couplings @xmath23 and @xmath1 .
we shall use the coupling constants @xmath34 and @xmath33 for a calculation of the green functions .
these are given by the functional integrals @xmath35\ } } { \int d \phi \exp\ { -h[\phi]\ } } \ , , \nonumber \\\end{aligned}\ ] ] where the subscripts @xmath36 run through the @xmath8 components of the field @xmath37 .
as explained in the introduction , we shall first study the case @xmath21 with the free energy functional @xmath38 = } \nonumber \\ \nonumber \\ & & \!\!\!\!\!\!\!\ !
\int d^dx \left\{\frac{1}{2 } \left[\phi_1 ( -{\nabla}^2 ) \phi_1 + \phi_2 ( -{\nabla}^2 ) \phi_2 \right]+\frac{g}{4 } \left[\phi_1 ^ 4 + 2(1-\delta ) \phi_1 ^ 2 \phi_2 ^ 2 + \phi_2 ^ 4 \right ] \right\}.\end{aligned}\ ] ] when expressed in terms of the coupling constants @xmath39 and @xmath33 , the green functions possess an imaginary part for @xmath40 .
for the reasons explained above , we shall derive an expansion of the imaginary parts of the green functions ( [ g2 m ] ) around the isotropic case , i. e. , in powers of @xmath33 : @xmath41\ , .\ ] ] each power @xmath42 has its own @xmath43-dependent imaginary part .
given such an expansion , the large - order estimates for the coefficients of the powers @xmath44 follows from a dispersion relation in @xmath39 : @xmath45 for a general discussion of the relationship between imaginary parts and large - order behavior see , for example , ch .
@xmath46 of @xcite .
if the power - series expansion of @xmath47 is @xmath48 the coefficients @xmath49 have the asymptotic behavior @xmath50.\ ] ] these estimates apply to perturbation coefficients , in which the maximal power of @xmath39 is much larger than the power of @xmath33 .
being interested in the region close to the isotropic limit , this restricted large - order estimation will be sufficient .
an expansion of the green function ( [ g2 m ] ) around the instanton yields exponentially small imaginary parts in both numerator and denominator . in order to isolate the imaginary part of the numerator
, we simplify the denominator in ( [ g2 m ] ) and calculate first the imaginary part of the approximate green functions @xmath51\ } } { \int d \phi \exp\ { -h_0 [ \phi]\ } } \ , , \nonumber \\\end{aligned}\ ] ] where the denominator contains only the free energy functional @xmath52 . with the aim of calculating ( [ img ] ) ,
we expand the fields around the classical solution of the isotropic limit @xmath32 for the space dimension @xmath14 .
thus we write : @xmath53 where @xmath54 is the well - known @xmath13-instanton in four dimensions @xmath55 and @xmath56 is the rotation angle of the isotropic instanton in the @xmath57-plane . the fields @xmath58 and @xmath59 correspond to the degrees of freedom orthogonal to the rotation of that instanton . the parameters @xmath60 and @xmath61 are position and scale size of the instanton , respectively .
inserting the expansion ( [ exp ] ) in @xmath62 , we obtain the expression : @xmath63 where @xmath64 collects all terms in @xmath58 and @xmath59 of order @xmath65 .
they are given by : + ( 1 ) linear terms : @xmath66 ^ 2 } \xi \nonumber \\ & & + \frac{\delta}{(-g)^{1/2}}\ , \sin^2(2\vartheta)\,8(2)^{1/2 } \int d^dx \frac{\lambda^3}{[1+\lambda^2(x - x_0)^2]^3 } \xi \nonumber \\ & & + \frac{\delta}{(-g)^{1/2}}\ , \sin(4\vartheta ) \,4(2)^{1/2 } \int d^dx \frac{\lambda^3}{[1+\lambda^2(x - x_0)^2]^3 } \eta\end{aligned}\ ] ] ( 2 ) quadratic terms : @xmath67 ^ 2}(\xi^2-\eta^2 ) \nonumber \\ & & + \delta \ , \sin(4\vartheta)\ , 6 \int d^dx \frac{\lambda^2}{[1+\lambda^2(x - x_0)^2]^2 } \xi \eta \ , , \end{aligned}\ ] ] where @xmath68 and @xmath69 are the operators @xmath70 ^ 2}\ , , \,\ , m_t = -{\nabla}^2- \frac { 8(1-\delta ) \lambda^2}{[1+\lambda^2 ( x - x_0)^2]^2}\ ] ] ( 3 ) cubic terms : @xmath71 \nonumber \\ & & -\delta \ , \frac{\sin^2(2\vartheta)}{2 } \ , ( -8g)^{1/2 } \int d^dx \frac{\lambda}{1+\lambda^2(x - x_0)^2 } ( 3 \eta^2 \xi-\xi^3 ) \nonumber \\ & & -\delta \ ,
\frac{\sin(4\vartheta)}{4 } \ , ( -8g)^{1/2 } \int d^dx \frac{\lambda}{1+\lambda^2(x - x_0)^2 } ( \eta^3 - 3 \eta \xi^2)\end{aligned}\ ] ] ( 4 ) quartic terms : @xmath72 \nonumber \\ & & -\delta \ , g \ ,
\frac{\sin^2(2\vartheta)}{8 } \int d^dx ( \xi^4+\eta^4 - 6\xi^2 \eta^2 ) \nonumber \\ & & + \delta \ , g \
, \frac{\sin(4\vartheta)}{4 } \int d^dx ( \xi \eta^3-\eta \xi^3)\end{aligned}\ ] ] the linear terms ( [ h1 ] ) contain factors @xmath5 or @xmath33 , since the expansion of the fields around the isotropic instanton is extremal only in the four - dimensional isotropic limit .
the classical contribution of the instanton is @xmath73 -\frac{\lambda^{\varepsilon } \delta } { g } \ ,
\frac{8\pi^2}{3 } \frac{\sin^2(2\vartheta)}{2}+ { \cal o } \left(\frac { \delta } { g}\ , \varepsilon \right ) \ ! .
\nonumber\end{aligned}\ ] ] the first term in @xmath74 is evaluated up to the first order in @xmath5 , because the one - loop renormalization will require replacing @xmath75 $ ] , where @xmath76 is the renormalized version of the coupling constant @xmath39 .
a contribution of order @xmath5 is not needed in the second , @xmath33-dependent term , where it would produce a further factor @xmath33 , and thus be part of the neglected terms @xmath77 in ( [ img ] ) .
due to the anisotropy of the action , the fluctuation expansion ( [ h1])([hclas ] ) contains @xmath56-dependent parts . however , all these terms are of order @xmath33 and can therefore be handled by straightforward perturbation theory .
note that the angle @xmath56 disappears when the isotropic instanton is directed along the coordinate axis .
then the isotropic instanton coincides with the exact solution in @xmath14 for @xmath78 .
this is also seen by inspecting the potential in ( [ hn2 ] ) . for @xmath78 ,
the term @xmath79 is larger than @xmath80 , so that the `` tunneling - paths '' of extremal action are obviously straight lines along the coordinate axis . for the region @xmath81 ,
the exact instanton follows from the known fact that the action ( [ hn2 ] ) is invariant under the orthogonal transformation : @xmath82 with the new coupling constants : @xmath83 satisfying @xmath81 for @xmath84 .
in contrast to the method in @xcite , the treatment of the fluctuations in a power series in @xmath33 does not allow us to deal with all modes perturbatively .
near the @xmath85-symmetric case , the gaussian approximation for the rotation of the instanton becomes invalid , and the gaussian integral must be replaced by an exact angle integration .
the separation of the rotation mode must be done with the help of collective coordinates .
the jacobian of the relevant transformation can be deduced from the isotropic system .
the field @xmath59 of small oscillations must not contain modes of @xmath69 with eigenvalues of the order @xmath33 , since these would vanish for @xmath32 .
the discussion of the longitudinal part is given in @xcite .
in dimensional regularization , all eigenvalues of @xmath68 which would be zero for @xmath14 are of order @xmath5 in @xmath86 dimensions . to avoid eigenvalues of the order @xmath87 in the propagator for the longitudinal fluctuations , all collective coordinates of the four - dimensional case must be retained in @xmath88 dimensions .
therefore the field @xmath58 in ( [ exp ] ) contains no modes of @xmath68 with eigenvalues proportional to @xmath5 . in order to calculate the fluctuation factor and to separate the almost - zero modes from @xmath89 and @xmath90 ,
it is convenient to do a conformal transformation onto a sphere in @xmath91 dimensions leading to a discrete spectrum for the transformed differential operators @xmath68 and @xmath69 @xcite .
their products of eigenvalues can be given in terms of the riemann @xmath92-function which we easily expanded near @xmath93 .
simple @xmath87-poles , which are characteristic of dimensional regularization , appear directly from the known singularity of that function . in place of the fields @xmath58 and @xmath59 , we define the fields @xmath94 and @xmath95 on the unit sphere in @xmath91 dimensions : @xmath96 where @xmath97 .
the instanton is supposed to be centered at the origin and rescaled by @xmath61 .
the spherical operator corresponding to the differential operator @xmath98 is @xmath99 where @xmath100 is the total angular momentum operator on the @xmath101-dimensional sphere .
the eigenfunctions of @xmath102 are the spherical harmonics @xmath103 in @xmath91 dimensions @xcite .
they satisfy @xmath104 and have the degeneracy @xmath105 after the transformation , the angle - independent quadratic part of @xmath106 in ( [ h2 ] ) becomes @xmath107 \phi_2 \ , , \ ] ] where @xmath108 is the surface element of the unit sphere in @xmath91 dimensions . after rewriting the entire energy functional([hall ] ) in terms of the new fields
, we can summarize the feynman rules for the diagrammatic evaluation of the functional integrals as follows : + @xmath109 propagators ( from the @xmath56-independent part of @xmath106 ) : + + longitudinal propagator : @xmath110^{-1}\ ] ] transversal propagator : @xmath111^{'}\right\}^{-1}\ ] ] the prime indicates that the almost - zero modes are omitted when forming the inverse .
+ @xmath112 tadpoles
@xmath113 : @xmath114 the bold dot stands for the @xmath56-dependence of the vertex .
+ @xmath115 two - point vertices ( @xmath56-dependent part of @xmath106 ) : + @xmath116 @xmath117 cubic vertices @xmath118 : @xmath119 @xmath120 quartic vertices @xmath121 : @xmath122 in order to obtain the leading corrections we must consider all connected diagrams of order @xmath123 , @xmath124 , and @xmath125 .
the contributions of these diagrams are added to @xmath126 in ( [ hclas ] ) .
we consider first the contributions to @xmath123 , which arise from the connected tree diagrams @xmath127 and their @xmath128 insertions of zeroth order in @xmath39 : @xmath129 since each @xmath128-vertex produces an @xmath5-factor , the smallest order in @xmath5 is two .
however , terms of order @xmath130 are negligible for our calculation even after one - loop renormalization , since : @xmath131 $ ] and @xmath132 .
the diagrams of @xmath124 can be generated from those of @xmath123 by one of the substitutions : @xmath133 by inspection , we see that the only diagram with an @xmath5-power smaller than two is given by @xmath134 which is of order @xmath135 .
after the renormalization , a term of order @xmath136 appears .
but this term can be neglected in comparison with the @xmath137-term from @xmath126 in expression ( [ hclas ] ) .
hence all the tree diagrams would enter only in a higher - order calculation .
contributions to order @xmath138 can appear from one - loop diagrams .
the only possible candidates are @xmath139 where the @xmath87-pole from the loop integration and the @xmath5-factor of the @xmath128-vertex cancel . following the method of @xcite
, we have found that these diagrams contribute to the coefficient of the imaginary part of @xmath140 in ( [ tilg ] ) a factor @xmath141
the angle - independent quadratic form is decomposed into a longitudinal and a transverse part . for the longitudinal fluctuations we obtain the determinant @xmath142^{-\frac{1}{2 } } \!\!= \prod^{\infty}_{l=0 } \left[\frac{(l+\frac{1}{2 } d-3)(l+\frac{1}{2 } d+2 ) } { ( l+\frac{1}{2 } d-1)(l+\frac{1}{2 } d ) } \right]^{-\frac{1}{2 } \nu_l(d+1)}_{,}\ ] ] which coincides with that appearing in the one - component @xmath3 theory , and is therefore known @xcite .
it contains a bound state at @xmath143 , this being responsible for the expected imaginary part of @xmath140 , and @xmath91 almost - zero eigenmodes of order @xmath5 associated with the dilatation and translation degrees of freedom of the instanton in @xmath86 dimensions . extracting these modes from the product ( [ lprod ] ) in the framework of
collective coordinates @xcite , we obtain the well - known formal replacement rule for the determinant @xmath144 with @xmath145 \ , , \ ] ] and the jacobian of the collective coordinates transformation : @xmath146 \ , .\ ] ] the expression for @xmath147 contains the euler constant @xmath148 , and the derivative of the riemann - zeta function @xmath149 at the value @xmath150 . note the simple pole in @xmath5 , which is related to the ultraviolet divergence on the one - loop level .
the transverse fluctuations contain neither negative nor zero modes for @xmath33 larger than zero .
the corresponding fluctuation determinant is given by @xmath151 } { \det(-v_0 ) } \right\}^{-\frac{1}{2 } } \nonumber \\ \nonumber \\ & = & \prod^{\infty}_{l=0 } \left[\frac{(l+ \frac{1}{2 } d-1)(l+\frac{1}{2 } d)-2(1-\delta ) } { ( l+\frac{1}{2 } d-1)(l+\frac{1}{2 } d ) } \right]^{-\frac{1}{2 } \nu_l(d+1)}.\end{aligned}\ ] ] just one zero eigenmode appears for @xmath143 in the isotropic limit @xmath152 , due to the rotational invariance in that case . for @xmath78 , the numerator in ( [ tprod ] ) contributes a factor @xmath153 . to avoid this artificial zero - mode divergence for @xmath152
we separate , as announced earlier , the angular degree of freedom from the integral measure via the collective - coordinates method .
similar to the longitudinal case , we make the formal substitution @xmath154 with @xmath155 which coincides with the jacobian of the isotropic model .
since the integration interval for the angle @xmath56 is compact , no singularity appears in the limit @xmath152 .
it is useful to illustrate the appearance of a divergent factor @xmath153 in a careless use of gaussian integral .
assuming @xmath78 , we expand the angle - dependent classical action around @xmath156 up to quadratic order .
the ensuing gaussian integral can be evaluated after replacing the finite integration interval @xmath157 $ ] by the noncompact one @xmath158 $ ] . in this way , the integral @xmath159\ ] ] is evaluated to @xmath160 = \frac{1}{\sqrt{2 \delta}}\ ] ] showing the spurious would - be zero - mode singularity , as the consequence of false rotation - mode treatment . excluding the @xmath143 -mode in the numerator on the right - hand side of ( [ tprod ] ) , we obtain for @xmath161 : @xmath162 \nonumber \right . \\ \nonumber \\ & & \left .
-\frac{1}{2 } \sum_{l=1}^{\infty } \frac{(2l+d-1 ) \gamma(l+d-1 ) } { \gamma(d ) \gamma(l+1 ) } \ln \left [ 1+\frac{2\ , \delta}{(l+\frac{d}{2}+1 ) ( l+\frac{d}{2}-2 ) } \right ] \right\}. \nonumber \\\end{aligned}\ ] ] the first sum is the well - known contribution from the isotropic limit .
the second sum is to be expanded in powers of @xmath33 . for large @xmath163 ,
the sum diverges as @xmath5 tends to zero . in order to separate off the divergence as a simple @xmath87-pole
, we use the zeta - function regularization method described in @xcite .
a somewhat tedious but straightforward calculation leads to @xmath164\ , .\ ] ] the series in front of the @xmath5-singularity ends after the quadratic power of @xmath33 , thereby leading to the exact @xmath165-coefficient of the simple @xmath5-pole . collecting all contributions to the one - loop expression of @xmath166
, we obtain the imaginary parts @xmath167 \right\}\end{aligned}\ ] ] with @xmath168\ ] ] and @xmath169 .
the expression contains a standard factor @xmath170 , since for symmetry reasons each saddle point contributes only one half of the gaussian integral . it remains to perform the collective coordinates integration over the dilatation @xmath171 , translation @xmath172 , and the rotation @xmath173 degree of freedom .
having obtained the imaginary parts of the green functions , we go over to the bare imaginary part of the vertex functions . taking the fourier transform and excluding the @xmath174-factors , an amputation of the external legs of @xmath175 leads to : @xmath176 \right\ } [ 1+{\cal o}(\varepsilon , g)]\end{aligned}\ ] ] with @xmath177 where @xmath178 is the modified bessel - function . the coefficient @xmath179 is given by @xmath180.\end{aligned}\ ] ] an @xmath5-singularity results from the ultraviolet divergence of one - loop diagrams .
as discussed in the following section , this simple pole in @xmath5 is removed by a conventional coupling constant renormalization . the contribution from the imaginary part of the denominator in equation ( [ g2 m ] ) follows for @xmath181 in ( [ v2 m ] ) .
we observe that it is of higher order in @xmath39 in comparison with the imaginary part of the numerator , and can therefore be neglected .
so the imaginary part of ( [ tilg ] ) supplies the desired imaginary part of ( [ g2 m ] ) .
in the preceding section we have calculated the contribution of the quadratic fluctuations around the isotropic instanton .
the almost - zero eigenvalues of translation , dilatation and rotation have been extracted and treated separately .
of course , the resulting expressions are useless , if we are not able to renormalize the theory .
this is the most difficult additional problem which arises in the transition from quantum mechanics to higher - dimensional field theories . a systematic scheme to renormalize both the real and imaginary part of vertex functions for a @xmath3 theory in @xmath86 dimensions was introduced by mckane , wallace and de alcantara bonfim .
they calculated the full ( real and imaginary ) renormalization - group constants using an extended minimal subtraction scheme to one loop ( the conventional one is given in @xcite ) .
we have extended their method to the case of two coupling constants in view of applications to critical phenomena . for an illustration
we consider first the four - point vertex function .
the result ( [ v2 m ] ) of the one - loop calculation about the instanton reads : @xmath182^{2n } \nonumber \right . \\
\nonumber \\ & & \left . \mbox { } \times \left [ -c_{\rm b } \int\limits_0^{\infty } \frac{d\lambda}{\lambda } \lambda^{\varepsilon } \left(-\frac{\lambda^{\varepsilon}a}{g}\right)^{\frac{d+6 + 2n}{2 } } \!\!\ ! \exp \left(\frac{\lambda^{\varepsilon}a}{g } \right ) \prod^{4}_{\nu=1 } \left(\frac{q_{\nu}}{\lambda}\right)^2 \ !
\tilde{\phi } \left ( \frac{q_{\nu}}{\lambda } \right ) \right ] \right\}\end{aligned}\ ] ] with @xmath183 and @xmath184 .
it remains to evaluate the integral over the parameter @xmath61 .
this is a relict of the introduction of collective coordinates in the instanton calculation .
the integral converges for small @xmath61 due to the exponentially decreasing of modified bessel - function @xmath185 . for large @xmath61
, the product @xmath186 behaves like @xmath187 \nonumber \\ & & + \left(\frac{q_{\nu}}{2\lambda}\right)^2 \bigg [ \left(\frac{q_{\nu}}{2\lambda}\right)^{\varepsilon } \left(\frac{2}{\varepsilon}+\gamma-\ln \pi+{\cal o}(\varepsilon)\right ) \nonumber \\ & & \quad -\left(\frac{2}{\varepsilon}+1-\gamma-\ln \pi + { \cal o}(\varepsilon ) \right ) \bigg ] + { \cal o}\left(\frac{q_{\nu}}{2\lambda}\right)^4 \bigg\ } \ , , \end{aligned}\ ] ] so that the integral diverges logarithmically as @xmath5 goes to zero .
this divergence causes a @xmath87-pole in the imaginary part .
since after the above approximation of @xmath188 the integral diverges for @xmath189 , a small @xmath61 cutoff @xmath190 has to be introduced . in this way , the @xmath61-integral in ( [ pol ] ) takes the form @xmath191 \ , .\ ] ] the coupling constant @xmath39 is chosen to lie on top of the tip of the left - hand cut in the complex @xmath39-plane .
being integrated only near @xmath192 , the integral can be evaluated perturbatively in @xmath39 . using @xmath193= \frac{1}{\varepsilon}\frac{|g|}{a } \mu^{a\varepsilon } \exp\left[-\left(\frac{\mu^{\varepsilon } a}{|g|}\right)\right ] [ 1+{\cal o}(g)]\ , , \ ] ] we get for ( [ scal ] ) the expression @xmath194 after expanding the factor @xmath195 in powers of @xmath5 , we obtain the typical finite contribution @xmath196 .
since this term can be omitted by a minimal subtraction , we are left with @xmath197 \right\}\!,\end{aligned}\ ] ] where @xmath198 and @xmath199 the tensor structure arises explicitly upon doing the integrals @xmath200 for all combinations of indices in @xmath201 . the first term in ( [ gim ] )
contains only mixed index combinations , whereas the second term has only a contribution from equal indices .
now we can proceed to renormalize the bare results .
the real part of @xmath202 is the perturbative one , and it is easily calculated to one loop . in terms of our coupling constants , the real parts @xmath203
take the form @xmath204
g^2 \mu^{-\varepsilon } \right\ } \ { s_{ijkl}-\delta_{ijkl } \ } \nonumber \\
\nonumber \\ & & \!\!\!\!\ !
+ \left\{-6g+\left[\frac{3}{4\pi^2\varepsilon } \left(10 - 2 \delta+{\delta}^2\right ) + { \cal o}({\varepsilon}^0 ) \right ] g^2 \mu^{-\varepsilon } \right\ } \ , \delta_{ijkl}\ , , \end{aligned}\ ] ] where @xmath190 is the arbitrary momentum scale introduced above in ( [ scal ] ) . in the absence of an instanton
, the wave function renormalization has no one - loop contribution . in the consequence ,
the expression ( [ greal ] ) is rendered finite by a coupling constant renormalization only .
the subscript r of the coupling constants indicates renormalized expressions in the conventional perturbative manner , i. e. , @xmath205 inserting this into ( [ gim ] ) and ( [ greal ] ) we find the perturbatively renormalized vertex functions @xmath206\ ] ] with @xmath207 and @xmath208 the coefficient @xmath209 is given by @xmath210 \nonumber \\
\nonumber \\ & = & \!\!\ ! 3^{1/2 } 2^{-2/3 } \pi^{-2 } \exp \bigg[\frac{1}{3\varepsilon } \left(10 - 2 { \delta}_{\rm r}+{\delta}_{\rm r}^2 \right)- \frac{1}{3\varepsilon } \left(10 - 2 { \delta}_{\rm r}+{\delta}_{\rm r}^2 \right ) \nonumber \\ \nonumber \\ & & \qquad \qquad \qquad \quad + \frac{4\zeta^{'}(2)}{\pi^2}-\frac{7}{3}\gamma-\frac{2}{3 } + { \cal o}({\delta}_{\rm r } ) + { \cal o}({\delta}_{\rm r } \ , g_{\rm r } ) \bigg].\end{aligned}\ ] ] the @xmath87-pole terms cancel .
thus , @xmath209 remains finite for @xmath132 . for the leading expansion ( [ img ] )
we can take the perturbative renormalized expression ( [ cr ] ) at the position @xmath211 .
the only remaining singularity is the @xmath87-factor in @xmath212 .
it requires a further renormalization . in our special choice of coupling constants , the disastrous @xmath213-divergence of perturbation
series appears in the expansions coefficients of @xmath44 .
since @xmath33 is a dimensionless anisotropy measure for positive as well as negative @xmath39 , we have expanded the imaginary part of vertex functions around the isotropic case in a well defined manner . for the derivation of nonperturbative corrections to the renormalization - group functions and for later applications , however
, it is convenient to avoid ratios of coupling constants . therefore we proceed with @xmath214 and @xmath215 instead of @xmath136 and @xmath215 .
then @xmath216 receive the rule of @xmath136 .
now we apply the extended minimal subtraction and perform a second ( nonperturbative ) renormalization step to eliminate the @xmath87-pole term : @xmath217\ , , \nonumber \\ \nonumber \\ \nonumber \\
v_{\rm r}\!\!&=&\!\!v_{\rm r}-i 2 ^ 7 3 \pi^4 c_{\rm r } \frac{1}{\varepsilon } \displaystyle{\sum_{n=0}^{\infty } } \bigg [ \frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n
\left(\frac{4n}{2n+3}\right ) b \left(n+\frac{1}{2},n+\frac{5}{2 } \right ) \nonumber \\ & & \qquad \qquad \qquad \quad \times \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+4 + 4 n}{2 } } \exp \left(\frac{a}{g_{\rm r } } \right ) \bigg]\ , .\end{aligned}\ ] ] the subscript r denotes the fully renormalized coupling constants . the required wave function renormalization in nonperturbative terms give no leading contributions to the imaginary part ( see below ) .
therefore the @xmath87-singularity in ( [ gren ] ) is removed by a nonperturbative coupling - constant renormalization only .
we now proceed to investigate the two - step renormalization for the wave function .
there are two additional complications for the @xmath218-functions .
first , @xmath219 is quadratically divergent .
second , there is a momentum - dependent divergence in the imaginary part of @xmath218 . for the @xmath3-theory
it was shown in @xcite that the undesirable momentum - dependence disappears during the nonperturbative renormalization process .
this is comparable to the situation in conventional two - loop perturbation expansion .
we continue to show that this non - obvious cancellation still works in the more complex case of two coupling constants and more than one field components .
according to ( [ v2 m ] ) the imaginary part of the unrenormalized two - point vertex function reads @xmath220 ^ 2 \nonumber \right.\\ \nonumber \\ & & \left .
\qquad \qquad \times \frac{1}{2 } \int\limits_0^{2\pi } d \vartheta u_{l\,i}(\vartheta ) u_{l\,j}(\vartheta ) \exp\left[a \sin^2(2\vartheta)\right ] \right\}.\end{aligned}\ ] ] the integration over the angle @xmath56 can be done : @xmath221 } \nonumber \\ \nonumber \\ & & \!\!\!\!\!\!\ !
= \delta_{ij } \sum_{n=0}^{\infty } \frac{a^n}{n!}\ , 4^n \frac{\gamma\left(n+ \frac{1}{2 } \right ) \gamma\left(n+\frac{3}{2 } \right ) } { \gamma(2n+2 ) } = \delta_{ij } \sum_{n=0}^{\infty } \frac{a^n}{n!}\ , 4^n b \left(n+ \frac{1}{2},n+\frac{3}{2 } \right ) .
\nonumber \\\end{aligned}\ ] ] thus , we get the usual form : @xmath222 . the wave function renormalization is contained in @xmath223 . following a similar procedure as in the case of @xmath224
, we obtain the imaginary part @xmath225 \left(-\frac{\mu^{\varepsilon}a}{g } \right)^{\frac{d+2 + 2n}{2 } } \exp\left ( \frac{\mu^{\varepsilon}a}{g}\right ) \right\ } \ , .
\nonumber \\\end{aligned}\ ] ] apart from the appearance of the @xmath226 term , a very similar expression is found for @xmath227 . in order to show the cancellation of this term during the second stage of renormalization we need the two - loop expression for the real part of @xmath223 @xmath228 \ , , \ ] ] where p denotes some number which is the contribution of order @xmath229 . at @xmath230 , ( [ wfreal ] )
is commonly defined as the inverse wave function renormalization constant @xmath231 . in terms of the perturbative renormalized coupling constants @xmath232
is given by @xmath233 and together with ( [ wfim ] ) , we find after the first step of renormalization @xmath234 \nonumber \\ \nonumber \\ & + & \!\!i\,2 ^ 4 \pi^2 3 \,c_{\rm r}\ , \frac{1}{\varepsilon } \sum_{n=0}^{\infty } \left\{\frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{3}{2 } \right ) \nonumber \right .
\\ \nonumber \\ & & \left .
\mbox { } \times \left[\gamma-\frac{1}{2}+\ln \left(\frac{q}{2\mu}\right ) \right ] \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+2 + 4n}{2 } } \exp\left ( \frac{a}{g_{\rm r}}\right ) \right\ } \ , .\end{aligned}\ ] ] we have again introduced the coupling constants @xmath216 and @xmath215 which are more convenient for practical applications .
this expression reads in the fully renormalized couplings ( [ ren ] ) @xmath235 \nonumber \\ \nonumber \\ & & \times \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+2 + 4n}{2 } } \exp\left ( \frac{a}{g_{\rm r}}\right ) \bigg\ } \ , .\end{aligned}\ ] ] by using the definition of the @xmath236-function in terms of gamma - functions one can easily read off the cancellation of @xmath226 singularities for every power @xmath43 of the coupling constant @xmath237 .
therefore we are left with @xmath238 \nonumber \\ \nonumber \\ & + & \!\ ! i\,2 ^ 4 \pi^2 3 \,c_{\rm r}\ , \frac{1}{\varepsilon } \left(2 ^ 4 \pi^2 a\,\,{\rm p}+\gamma-\frac{1}{2}-\ln2 \right ) \nonumber \\ \nonumber \\ & \displaystyle{\times}&\!\!\ ! \sum_{n=0}^{\infty }
\bigg[\frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{3}{2 } \right ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+2 + 4n}{2 } } \exp\left ( \frac{a}{g_{\rm r}}\right ) \bigg ] \ , . \nonumber \\\end{aligned}\
] ] now we can define the nonperturbative minimally subtracted wave function renormalization to remove the @xmath87-pole : @xmath239 \ , .\end{aligned}\ ] ] the full renormalization constant is defined by @xmath240 and follows from the product of ( [ zp ] ) and ( [ znp ] ) . a further important green function is associated with the composite field @xmath241 : @xmath242 by a straightforward application of the techniques discussed in the preceding sections we obtain for the corresponding vertex function @xmath243 \hat{\phi } \left(\frac{q_3}{\lambda}\right ) \frac{1}{2 } \int\limits_0^{2\pi } d \vartheta u_{l\,i}(\vartheta ) u_{l\,j}(\vartheta ) \exp\left[a \sin^2(2\vartheta)\right ] \bigg\ } \ , , \nonumber \\\end{aligned}\ ] ] where @xmath244 now we can proceed as in the case of wave function renormalization . performing
the collective coordinates integration over @xmath61 and @xmath56 we obtain @xmath245 \left(-\frac{\mu^{\varepsilon}a}{g } \right)^{\frac{d+4 + 2n}{2 } } \exp\left ( \frac{\mu^{\varepsilon}a}{g}\right ) \right\ } , \nonumber \\\end{aligned}\ ] ] which agrees with the usual form @xmath246 . at @xmath247 , @xmath248 ,
the vertex function @xmath249 defines a third renormalization constant , called @xmath250 , via @xmath251 the perturbative @xmath250 follows from the real part of @xmath252 : @xmath253 where @xmath254 is a number of order @xmath229 . in terms of the perturbative renormalized coupling constants , the one - loop expression of @xmath255
is given by @xmath256 including @xmath257 and @xmath258 we obtain after the perturbative step of renormalization @xmath259 \nonumber \\
\nonumber \\ & + & \!\!i\,2 ^ 4 \pi^2 3 ^ 2 \,c_{\rm r}\ , \frac{1}{\varepsilon } \sum_{n=0}^{\infty } \left\{\frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{3}{2 } \right ) \nonumber \right .
\\ \nonumber \\ & & \left . \mbox { } \times \left[\gamma+\ln \left(\frac{q_3}{2\mu}\right ) \right ] \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+4 + 4n}{2 } } \exp\left ( \frac{a}{g_{\rm r}}\right ) \right\ } \ , .\end{aligned}\ ] ] in the second step of renormalization the @xmath260 singularities cancel , and we find @xmath261 \nonumber \\
\nonumber \\ & & \quad -i\,2 ^ 4 \pi^2 3 ^ 2 \,c_{\rm r}\ , \frac{1}{\varepsilon } \left(\frac{8\pi^2}{3 } { \rm q}-\gamma+\ln2 \right ) \nonumber \\ \nonumber \\ & & \quad \displaystyle{\times } \sum_{n=0}^{\infty } \bigg[\frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{3}{2 } \right ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+4 + 4n}{2 } } \exp\left ( \frac{a}{g_{\rm r}}\right ) \bigg ] \ , .
\nonumber \\\end{aligned}\ ] ] hence , the nonperturbative @xmath250 is @xmath262.\end{aligned}\ ] ] with respect to the leading expansion ( [ img ] ) this is equal to @xmath263 .
after having calculated the imaginary parts of the renormalized couplings and of the renormalization constants @xmath264 and @xmath250 , we are in the position to derive the nonperturbative renormalization group functions @xmath265 , @xmath266 and @xmath267 in the usual way . first we derive the @xmath268-functions from the expressions ( [ ren ] ) for @xmath269 and @xmath270 , respectively , and obtain : @xmath271\ , , \nonumber \\
\lefteqn{\beta_v(g_{\rm r},v_{\rm r})= } \nonumber \\ & & \mu \frac{\partial}{\partial \mu } v_{\rm r } = \mu \frac{\partial}{\partial \mu } v_{\rm r}- i 2 ^ 7 3 \pi^4 c_{\rm r } \frac{1}{\varepsilon } \displaystyle{\sum_{n=0}^{\infty } } \mu \frac{\partial}{\partial \mu } \bigg [ \frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{5}{2 } \right ) \nonumber \\ \nonumber \\ & & \qquad \qquad \qquad \qquad \qquad \times \left(\frac{4n}{2n+3}\right ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+4 + 4 n}{2 } } \exp \left(\frac{a}{g_{\rm r } } \right ) \bigg]\ , .\end{aligned}\ ] ] the derivatives are taken at fixed bare couplings @xmath39 and @xmath1 .
these functions will be finite as @xmath132 .
furthermore , they will have a left - hand cut in the complex @xmath39-plane . on top of the tip of this cut
we find @xmath272 \ , , \nonumber \\ \nonumber \\
\beta_v(g_{\rm r},v_{\rm r})= \beta_v^{\rm p}(g_{\rm r},v_{\rm r } ) \!\!\!\!&+&\!\!\!\!i 2 ^ 7 3 \pi^4 c_{\rm r } \displaystyle{\sum_{n=0}^{\infty } } \bigg [ \frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{5}{2 } \right ) \nonumber \\
\nonumber \\ & & \times \left(\frac{4n}{2n+3}\right ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+6 + 4 n}{2 } } \exp \left(\frac{a}{g_{\rm r } } \right ) \bigg ] \ , .\end{aligned}\ ] ] the real parts are the familiar perturbative @xmath268-functions @xmath273 in the leading contributions the pair of coupling constants @xmath274 may be replaced by @xmath275 in ( [ bet1 ] ) .
this can be proven using ( [ ren ] ) .
the result is @xmath276\ , , \nonumber \\
\beta_v(g_{\rm r},v_{\rm r})= \beta_v^{\rm p}(g_{\rm r},v_{\rm r } ) \!\!\!\!&+&\!\!\!\!i 2 ^ 7 3 \pi^4 c_{\rm r } \displaystyle{\sum_{n=0}^{\infty } } \bigg [ \frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n b \left(n+\frac{1}{2},n+\frac{5}{2 } \right ) \nonumber \\ \nonumber \\ & & \times \left(\frac{4n}{2n+3}\right ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+6 + 4 n}{2 } } \exp \left(\frac{a}{g_{\rm r } } \right ) \bigg ] \ , .\end{aligned}\ ] ] note that the @xmath268-functions have a form expected within a minimal subtracted scheme : @xmath277 and @xmath278 , where the @xmath279 denotes the @xmath268-functions in four dimensions .
the critical large - distance behavior of the correlation functions is defined by the anomalous dimension of the field @xmath37 , which is defined by @xmath280 with ( [ zp ] ) for @xmath232 and ( [ znp ] ) for @xmath281 we get the @xmath5-independent expression @xmath282 \ , .\end{aligned}\ ] ] the divergence of the length scale at critical temperature is governed by the anomalous dimension of the composite field @xmath283 : @xmath284 using ( [ ppp ] ) and ( [ nppp ] ) , we obtain @xmath285 .
\nonumber \\\end{aligned}\ ] ] for the leading contributions to the imaginary part , it was possible to replace @xmath274 by @xmath275 .
we remark that the expansions ( [ betar]),([gamr ] ) and ( [ npp2 ] ) have the correct isotropic limit for @xmath286 .
generalizing ( [ exp ] ) to @xmath22 we write : @xmath287 where @xmath27 and @xmath56 are the rotation angles of the isotropic instanton .
this change of variables yields the following expansion of the energy functional : @xmath288 ^ 2 } \xi \nonumber \\
\nonumber \\ & -&\!\!\ ! ( -8g)^{1/2 } \int d^dx \frac{\lambda}{1+\lambda^2(x - x_0)^2 } \left[\xi^3+(1-\delta)(\xi \eta^2+\xi \chi^2 ) \right ] \nonumber \\
\nonumber \\ & + & \!\!\ !
\delta \,\ , f(g,\vartheta,\theta,\xi,\eta,\chi)\end{aligned}\ ] ] with @xmath68 and @xmath289 of equation ( [ mlt ] ) . for brevity ,
we have written down explicitly only the terms responsible for the leading contributions in the expansion ( [ img ] ) .
the remaining terms are denoted collectively by @xmath290 .
the classical contribution to the energy functional is @xmath291 \nonumber \\ \nonumber \\ & & \!\!\!\!\ !
-\frac{\lambda^{\varepsilon } \delta } { 2 g } \ , \frac{8\pi^2}{3 }
\left[\sin^4{\theta } \sin^2(2\vartheta)+\sin^2(2\theta ) \right]+ { \cal o}\left(\frac { \delta } { g}\ , \varepsilon \right).\end{aligned}\ ] ] with similar steps as for @xmath21 we obtain the bare imaginary parts of the vertex functions : @xmath292 \bigg\ } \left[1+{\cal o}(\varepsilon , g ) \right ] \ , , \end{aligned}\ ] ] where @xmath293 \ , , \quad ( i_k\ , = \ , 1,2,3)\ , , \end{aligned}\ ] ] with @xmath294 , @xmath295 as in ( [ aa ] ) .
after evaluating the integrals over the collective coordinates @xmath61 , @xmath27 , and @xmath56 we obtain for the imaginary part of the four - point vertex function @xmath296 \bigg\}\ , .\end{aligned}\ ] ] similarly the bare imaginary parts of @xmath297 and @xmath298 are found to be @xmath299 } \nonumber \\ \nonumber \\ & & \times \frac{1}{\varepsilon } \sum_{n=0}^{\infty } \bigg[\frac{(-2\ , \delta)^n}{n ! } \ , i(n ) \ , \left(-\frac{\mu^{\varepsilon}a}{g }
\right)^{\frac{d+3 + 2n}{2 } } \exp\left ( \frac{\mu^{\varepsilon}a}{g}\right ) \bigg ] \
, , \end{aligned}\ ] ] and @xmath300 \left(-\frac{\mu^{\varepsilon}a}{g } \right)^{\frac{d+5 + 2n}{2 } } \exp\left ( \frac{\mu^{\varepsilon}a}{g}\right ) \right\ } , \nonumber \\\end{aligned}\ ] ] with @xmath301 now we apply the extended renormalization scheme .
using the coupling constants @xmath1 and @xmath39 , we obtain for the nonperturbative renormalized couplings @xmath302 \ , , \nonumber \\ \nonumber \\
v_{\rm r}\!\!\!&=&\!\!\!v_{\rm r } -i 2 ^ 7 3 \pi^4 c_{\rm r}^{(3 ) } \frac{1}{\varepsilon } \displaystyle{\sum_{n=0}^{\infty } } \bigg [ \frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n j_v(n ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+5 + 4 n}{2 } } \exp \left(\frac{a}{g_{\rm r } } \right ) \bigg]\ , , \nonumber \\\end{aligned}\ ] ] with @xmath303 \ , b \left(3 + 2n - p , p+\frac{1}{2}\right ) \nonumber \\ \nonumber \\ & & \qquad \qquad \qquad \qquad \quad \times b \left(n - p+\frac{1}{2},n - p+\frac{5}{2 } \right ) \bigg\ } \ , , \end{aligned}\ ] ] and @xmath304\ , .\ ] ] for the nonperturbative renormalization constants @xmath264 and @xmath250 we find @xmath305 \,\end{aligned}\ ] ] and @xmath306.\end{aligned}\ ] ] to calculate the nonperturbative renormalization group functions @xmath265 , @xmath266 and @xmath267 , we simply repeat the calculations of the case @xmath21 . using the expressions ( [ c3r ] ) for @xmath307 and @xmath237
yields the @xmath268-functions @xmath308\ , , \nonumber \\
\lefteqn{\beta_v(g_{\rm r},v_{\rm r})= } \nonumber \\ & & \beta_v^{\rm p}(g_{\rm r},v_{\rm r } ) + i 2 ^ 7 3 \pi^4 c_{\rm r}^{(3 ) } \displaystyle{\sum_{n=0}^{\infty } } \bigg [ \frac{v_{\rm r}^n}{n ! } \left(\frac{3}{4\pi^2}\right)^n j_v(n ) \left(-\frac{a}{g_{\rm r } } \right)^{\frac{d+7 + 4 n}{2 } } \exp \left(\frac{a}{g_{\rm r } } \right ) \bigg]\ , . \nonumber \\ \nonumber \\\end{aligned}\ ] ] the expressions for the nonperturbative anomalous dimension of the field @xmath37 and the composite field @xmath283 are given by @xmath309 \
, , \end{aligned}\ ] ] and @xmath310 \ , , \end{aligned}\ ] ] respectively .
our main results are the real and imaginary parts ( [ b3])([npp3 ] ) of the renormalization - group functions @xmath268 , @xmath148 and @xmath311 in terms of the nonperturbative renormalized couplings @xmath307 and @xmath237 . via the dispersion relation ( [ dr ] ) , we obtain the large - order behavior ( [ lo2 ] ) . it is useful to check the expansion terms proportional to powers of @xmath312 accompanying the @xmath313 in the imaginary part of each renormalization
- group function .
if the power of @xmath312 is denoted by @xmath314 , we note that @xmath314 is the same as for the corresponding vertex function before the integration over the dilatation parameter @xmath61 .
this integration with the measure @xmath315 reduces @xmath314 by @xmath316 [ see ( [ lambda ] ) ] .
however this effect is compensated by the differentiation with respect to the logarithm of the scale parameter @xmath190 in the definition of the renormalization group functions .
the power @xmath314 in the vertex functions can simply be explained : for a @xmath317-point vertex function with @xmath318 @xmath283-insertions there is first a power @xmath319 from the fields .
this is multiplied by a factor @xmath320 from the @xmath88 translational , @xmath321 rotational , and one dilatational would - be zero modes . the classical contribution to the energy functional yields a factor @xmath322 for each power of the anisotropic constant @xmath1 .
hence , for the renormalization group function which follows at the one - loop level from the vertex function @xmath323 , the value of @xmath314 is given by @xmath324 inserting the imaginary parts into a dispersion relation ( [ dr ] ) , we can estimate the large - order coefficients of the corresponding series expansions @xmath325 where @xmath326 .
the function @xmath327 stands for @xmath328 , @xmath329 , @xmath148 and @xmath311 , respectively .
the rules how to go from the imaginary parts of @xmath327 to @xmath330 are known from ( [ lo2 ] ) .
for instance , the large - order coefficients of the @xmath268-functions for @xmath22 are given by @xmath331 \
, , \nonumber \\ \nonumber
\\ \beta_{v , \,kn}\!\!&=&\!\!-c_{\rm r}^{(3)}2 ^ 7 3 \pi^3 \left(\frac{3}{4\pi^2}\right)^n \frac { j_v(n ) } { n ! } \nonumber \\ \nonumber \\ & & \qquad \times ( -1)^k \left(\frac{3}{8\pi^2}\right)^k k ! \ , k^{2n+(d+5)/2 } \left[1+{\cal o}(1/k)\right ] \ , , \end{aligned}\ ] ] with @xmath332 , @xmath333 from ( [ j ] ) and ( [ cc33 ] ) , respectively . the leading large-@xmath318 behavior can be checked by means of a simple combinatorial analysis .
we start with a series in terms of the standard couplings @xmath23 and @xmath1 [ see ( [ hn ] ) ] . for illustration
, we consider the contribution of the pure @xmath23-powers .
it is known from a theory with only one coupling constant , that the coefficients have the large - order behavior @xmath334 changing over to our couplings ( @xmath335 ) , the right - hand side of ( [ lao ] ) contributes @xmath336 with @xmath337 replacing @xmath163 by @xmath338 , the contribution of @xmath339 has the form @xmath340 now , for large @xmath318 , the @xmath341-functions can be approximated by @xmath342 thus in the region @xmath326 we obtain @xmath343 in agreement with ( [ concl ] ) .
as stated in the introduction , the stable cubic fixed point is expected to lie in the vicinity of the isotropic fixed point . since @xmath1 becomes very small in this region ,
reasonable results should be obtained by resumming the @xmath39-series accompanying each power @xmath313 .
this will be done in a forthcoming publication . | for the anisotropic @xmath0$]-theory with we calculate the imaginary parts of the renormalization - group functions in the form of a series expansion in @xmath1 , around the isotropic case .
dimensional regularization is used to evaluate the fluctuation determinants for the isotropic instanton near the space dimension @xmath2 .
the vertex functions in the presence of instantons are renormalized with the help of a nonperturbative procedure introduced for the simple @xmath3-theory by mckane et al . |
the production of lepton pairs in hadron collisions @xmath3 proceeds through an intermediate virtual photon via @xmath4 , and the subsequent leptonic decay of the virtual photon .
traditionally , interest in this drell - yan process has concentrated on lepton pairs with large mass @xmath0 which justifies the application of perturbative qcd and allows for the extraction of the antiquark density in hadrons @xcite .
prompt photon production @xmath5 can be calculated in perturbative qcd if the transverse momentum @xmath1 of the photon is sufficiently large . because the quark - gluon compton subprocess is dominant , @xmath6 , this reaction provides essential information on the gluon density in the proton at large @xmath7 @xcite .
unfortunately , the analysis suffers from fragmentation , isolation , and intrinsic transverse momentum uncertainties .
alternatively , the gluon density can be constrained from the production of jets with large transverse momentum at hadron colliders @xcite , but the information from different experiments and colliders is ambiguous . in this paper
we demonstrate that , like prompt photon production , lepton pair production is dominated by quark - gluon scattering in the region @xmath2 .
this realization means that new independent constraints on the gluon density may be derived from drell - yan data in kinematical regimes that are accessible in collider and fixed target experiments but without the theoretical and experimental uncertainties present in the prompt photon case . in sec .
[ sec:2 ] , we review the relationship between virtual and real photon production in hadron collisions in next - to - leading order qcd . in sec .
[ sec:3 ] we present our numerical results , and sec . [ sec:4 ] is a summary .
in leading order ( lo ) qcd , two partonic subprocesses contribute to the production of virtual and real photons with non - zero transverse momentum : @xmath8 and @xmath9 .
the cross section for lepton pair production is related to the cross section for virtual photon production through the leptonic branching ratio of the virtual photon @xmath10 .
the virtual photon cross section reduces to the real photon cross section in the limit @xmath11 .
the next - to - leading order ( nlo ) qcd corrections arise from virtual one - loop diagrams interfering with the lo diagrams and from real emission diagrams . at this order
@xmath12 partonic processes with incident gluon pairs @xmath13 , quark pairs @xmath14 , and non - factorizable quark - antiquark @xmath15 processes contribute also .
singular contributions are regulated in @xmath16=4 - 2@xmath17 dimensions and removed through @xmath18 renormalization , factorization , or cancellation between virtual and real contributions . an important difference between virtual and real photon production
arises when a quark emits a collinear photon . whereas the collinear emission of a real photon leads to a @xmath19 singularity that has to be factored into a fragmentation function , the collinear emission of a virtual photon yields a finite logarithmic contribution since it is regulated naturally by the photon virtuality @xmath0 . in the limit
@xmath11 the nlo virtual photon cross section reduces to the real photon cross section if this logarithm is replaced by a @xmath19 pole .
a more detailed discussion can be found in @xcite .
the situation is completely analogous to hard photoproduction where the photon participates in the scattering in the initial state instead of the final state .
for real photons , one encounters an initial - state singularity that is factored into a photon structure function . for virtual photons ,
this singularity is replaced by a logarithmic dependence on the photon virtuality @xmath0 @xcite .
a remark is in order concerning the interval in @xmath1 in which our analysis is appropriate . in general , in two - scale situations , a series of logarithmic contributions will arise with terms of the type @xmath20 .
thus , if either @xmath21 or @xmath22 , resummations of this series must be considered . for practical reasons , such as event rate , we do not venture into the domain @xmath21 , and our fixed - order calculation should be adequate . on the other hand , the cross section is large in the region @xmath22 . in previous papers @xcite , we compared our cross sections with available fixed - target and collider data on massive lepton - pair production , and we were able to establish that fixed - order perturbative calculations , without resummation , should be reliable for @xmath23 . at smaller values of @xmath1 ,
non - perturbative and matching complications introduce some level of phenomenological ambiguity . for the goal we have in mind ,
, contraints on the gluon density , it would appear best to restrict attention to the region @xmath24 , but below @xmath21 .
in this section we present numerical results for the production of lepton pairs in @xmath25 collisions at the tevatron with center - of mass energy @xmath26 and 2.0 tev .
we analyze the invariant cross section @xmath27 averaged over the rapidity interval -1.0 @xmath28 1.0 .
we integrate the cross section over various intervals of pair - mass @xmath0 and plot it as a function of the transverse momentum @xmath1 .
our predictions are based on a nlo qcd calculation @xcite and are evaluated in the @xmath18 renormalization scheme .
the renormalization and factorization scales are set to @xmath29 .
if not stated otherwise , we use the cteq4 m parton distributions @xcite and the corresponding value of @xmath30 in the two - loop expression of @xmath31 with four flavors ( five if @xmath32 ) .
the drell - yan factor @xmath10 for the decay of the virtual photon into a lepton pair is included in all numerical results . in fig .
[ fig:1 ] we display the nlo qcd cross section for lepton pair ( 7.6,10.5 ) production at the tevatron at @xmath26 tev as a function of @xmath1 for four regions of @xmath0 .
the regions of @xmath0 have been chosen to avoid resonances , _
i.e. _ between @xmath33 gev and the @xmath34 resonance , between the @xmath34 and the @xmath35 resonances , above the @xmath35 s , and a high mass region .
the cross section falls both with the mass of the lepton pair @xmath0 and , more steeply , with its transverse momentum @xmath1 .
no data are available yet from the cdf and d0 experiments . however , prompt photon production data exist to @xmath36 gev , where the cross section is about @xmath37 pb / gev@xmath38
. it should be possible to analyze run i data for lepton pair production to at least @xmath39 gev where one can probe the parton densities in the proton up to @xmath40 .
the ua1 collaboration measured the transverse momentum distribution of lepton pairs at @xmath41 gev up to @xmath42 @xcite , and their data agree well with our theoretical results @xcite . the fractional contributions from the @xmath43 and @xmath44 subprocesses through nlo are shown in fig .
[ fig:2 ] .
it is evident ( 7.6,10.5 ) that the @xmath43 subprocess is the most important subprocess as long as @xmath23 .
the dominance of the @xmath43 subprocess diminishes somewhat with @xmath0 , dropping from over 80 % for the lowest values of @xmath0 to about 70 % at its maximum for @xmath45 30 gev .
in addition , for very large @xmath1 , the significant luminosity associated with the valence dominated @xmath46 density in @xmath25 reactions begins to raise the fraction of the cross section attributed to the @xmath44 subprocesses .
subprocesses other than those initiated by the @xmath44 and @xmath47 initial channels are of negligible import .
we update the tevatron center - of - mass energy to run ii conditions ( @xmath48 tev ) and use the latest global fit by the cteq collaboration ( 5 m ) .
figure [ fig:7 ] ( 7.6,10.5 ) demonstrates that the larger center - of - mass energy increases the invariant cross section for the production of lepton pairs with mass 5 gev @xmath49 6 gev by 5 % at low @xmath50 gev and 20 % at high @xmath51 gev .
in addition , the expected luminosity for run ii of 2 fb@xmath52 should make the cross section accessible to @xmath36 gev or @xmath53 . this extension would constrain the gluon density in the same regions as prompt photon production in run i. next we present a study of the sensitivity of collider and fixed target experiments to the gluon density in the proton . the full uncertainty in the gluon density
is not known . here
we estimate this uncertainty from the variation of different recent parametrizations .
we choose the latest global fit by the cteq collaboration ( 5 m ) as our point of reference @xcite and compare results to those based on their preceding analysis ( 4m@xcite ) and on a fit with a higher gluon density ( 5hj ) intended to describe the cdf and d0 jet data at large transverse momentum .
we also compare to results based on global fits by mrst @xcite , who provide three different sets with a central , higher , and lower gluon density , and to grv98 @xcite . in fig .
[ fig:5 ] we plot the cross section for lepton pairs with mass between the ( 7.6,10.5 ) @xmath34 and @xmath35 resonances at run ii of the tevatron in the region between @xmath54 and 30 gev ( @xmath55 ) . for the cteq parametrizations
we find that the cross section increases from 4 m to 5 m by 2.5 % ( @xmath56 gev ) to 5 % ( @xmath54 gev ) and from 5 m to 5hj by 1 % in the whole @xmath1-range .
the largest differences from cteq5 m are obtained with grv98 at low @xmath1 ( minus 10 % ) and with mrst(g@xmath57 ) at large @xmath1 ( minus 7% ) .
the theoretical uncertainty in the cross section can be estimated by varying the renormalization and factorization scale @xmath58 around the central value @xmath59 .
figure [ fig:8 ] ( 9,8.08 ) shows this variation for @xmath60 at @xmath61 tev in the region between the @xmath34 and @xmath35 resonances . in the interval @xmath62 the dependence of the cross section on the scale @xmath58 drops from @xmath63 ( lo ) to the small value @xmath64 ( nlo ) .
the @xmath65-factor ratio ( nlo / lo ) is approximately 2 , as one might expect naively . a similar analysis for fermilab s fixed target experiment e772 @xcite is shown in fig .
[ fig:6 ] .
( 7.6,10.5 ) in this experiment , a deuterium target is bombarded with a proton beam of momentum @xmath66 800 gev , _
i.e. _ @xmath67 gev .
the cross section is averaged over the scaled longitudinal momentum interval 0.1 @xmath68 0.3 . in
fixed target experiments one probes substantially larger regions of @xmath69 than in collider experiments .
therefore one expects greater sensitivity to the gluon distribution in the proton .
we find that use of cteq5hj increases the cross section by 7 % ( 26 % ) w.r.t .
cteq5 m at @xmath70 gev ( @xmath71 gev ) and by 134 % at @xmath54 gev . with mrst(g@xmath72 )
the cross section drops relative to the cteq5m - based values by 17 % , 40 % , and 59 % for these three choices of @xmath1 .
figure [ fig:9 ] ( 9,7.68 ) shows the variation of the fixed target cross section on the renormalization and factorization scale @xmath58 . in the interval
@xmath73 the dependence decreases from @xmath74 ( lo ) to @xmath75 ( nlo ) .
an optimal scale choice might be @xmath76 , where the points of minimal sensitivity ( maximum of nlo ) and of fastest apparent convergence ( lo = nlo ) nearly coincide . at @xmath77 ,
the @xmath65-factor ratio is 2.6 .
the nlo cross section turns negative at the lowest scale shown @xmath78 gev , a value too low to guarantee perturbative stability .
the production of drell - yan pairs with low mass and large transverse momentum is dominated by gluon initiated subprocesses . in contrast to prompt photon production , uncertainties from fragmentation , isolation , and intrinsic transverse momentum are absent .
the hadroproduction of low mass lepton pairs is therefore an advantageous source of information on the parametrization and size of the gluon density .
the increase in luminosity of run ii increases the accessible region of @xmath69 from 0.03 to 0.1 .
the theoretical uncertainty has been estimated from the scale dependence of the cross sections and found to be very small for collider experiments .
it is a pleasure to thank l. e. gordon for his collaboration . | the hadroproduction of lepton pairs with mass @xmath0 and finite transverse momentum @xmath1 is described in perturbative qcd by the same partonic subprocesses as prompt photon production .
we demonstrate that , like prompt photon production , lepton pair production is dominated by quark - gluon scattering in the region @xmath2 .
this feature leads to sensitivity to the gluon density in kinematical regimes accessible in collider and fixed target experiments , and it provides a new independent method for constraining the gluon density .
anl - hep - cp-00 - 002 + desy
00 - 006 |
it has been one of the central goals of human knowledge to understand the position of the earth in the universe .
a lot of effort has been spent to derive the structure and the history of formation and evolution of our galaxy . doing
this is a particularly difficult task , simply because there is no good viewpoint , as opposed to the case for external galaxies whose structures can be observed directly .
the structure , formation and evolution of the galaxy are very important issues in contemporary astrophysics ; they are closely related to stellar quantities such as distance , age , metallicity , and kinematics of the individual stars , among others .
these properties can be obtained for large stellar samples using straightforward photometric and spectroscopic observations . in the present era , before the full exploitation of the planned huge spectral surveys of galactic stellar objects ( e.g. gaia , segue , 6df , lamost ) is possible , using star counts based on all - sky photometric surveys is a direct and one of the few accessible methods for the study of the structure of the galaxy . probing the galaxy using star counts has a history of about 200 years ( reid 1993 ) , with the modern epoch marked by the classic work of bahcall & soneira ( 1980 ) , based on several observations ( seares et al .
1925 , kron 1978 , tyson & jarvis 1979 , peterson , ellis & kibblewhite 1979 ) .
bahcall & soneira built a standard model , in which the galaxy was simplified and parameterized by an exponential disk and a spheroid characterized by a de vaucouleurs profile . assuming a density profile for each population , based on detailed studies of nearby spiral galaxies , and assuming a luminosity function as observed for stars in the local volume , their model predicts number density
magnitude colour distributions .
further studies on this subject provided certain modifications to the bahcall & soneira model .
theoretical models were further constrained using new observations were subsequently obtained . in 1984 ,
gilmore claimed for the first time that there is a thick disk in our galaxy .
the star counts approach to the study of the structure of the galaxy was extensively reviewed by reid ( 1993 ) ; he , and his colleagues , ( reid 1993 , reid , yan & majewski 1996 , majewski , siegel & kunkel et al .
1999 , siegel , majewski & reid 2002 ) carried out a series of studies appropriately termed `` star counts redivivus '' , dedicated respectively to : faint magnitude star counts , the halo luminosity function , and an exploration of the contamination of star counts by star streams .
when the sloan digital sky survey ( sdss ) made its first data release publicly available , it raised very promising prospects for the study of galactic structure thanks to the large sky coverage and the deep stellar photometry .
chen et al . ( 2001 ) analysed the sdss early data release ( edr ) ; they confirmed the existence of a galactic thick disk and obtained a set of structural parameters for the galaxy . with the increasing sky coverage in subsequent sdss data releases
, it is expected that one will be able to determine the galactic structure parameters ever more precisely .
except for constraining the density profile , the stellar luminosity function can also be determined with a higher accuracy . in gould ,
bahcall & flynnet ( 1996 ) , and zheng , flynn & gould ( 2001 , 2004 ) the disk m - dwarf and @xmath6-band luminosity functions are obtained using _
hubble space telescope _
star counts . in reid ( 2002 )
, the luminosity function in the solar neighborhood was studied based on the palomar / msu nearby star spectroscopic survey .
the galactic structural parameters derived from various studies show large divergence , which can not simply be attributed to statistical errors , although these are generally significant . in star counts from fields selected following a certain symmetry approach , there are deviations from the axisymmetric standard model which are obvious and can not be neglected .
spergel & blitz ( 1988 ) pointed out that this might be an indication of a triaxial structure of the galaxy , and they used a triaxial halo to explain the asymmetries detected from gas motions around @xmath1=180@xmath0 .
later , blitz & spergel ( 1991 ) showed that the observed hi gas is moving in an asymmetric gravitational potential . using areas selected from the asp catalogue , larsen & humphreys ( 1996 ) found deviations of star counts from the axisymmetric model
; they found that there are significant excesses of stars in quadrant i , at @xmath7 , compared to quadrant iv , around @xmath8 .
further work was published by parker , humphreys & larsen ( 2003 ) , who compare star counts in 120 poss i fields .
they claimed that the excess is not associated with any known stream such as those linked to the sagittarius dwarf , magellanic stream or fornax - leo - sculptor stream .
parker , humphreys & beers ( 2004 ) , however , concluded that the excess is due to thick disk stars and may be associated with the recently discovered canis major debris stream . in hartwick ( 2000 ) , possible reasons for causing triaxiality due to nearby spiral and dwarf galaxies are considered .
newberg & yanny ( 2005b ) built triaxial ellipsoidal stellar halo models and analysed f turn - off star samples from sdss dr3 .
to explain the observed excess of the stellar number density with respect to an axisymmetric halo , two competing scenarios have been proposed : the triaxialilty could be an inherent property related to the smooth structure of the galaxy , or it could be related to the remnants of some historical merger events .
recently , this debate has intensified with the reported `` virgo overdensity '' by juri et al.(2005 ) , which may be the cause of the large - scale asymmetry in the structure of the galaxy . by using the photometric parallax method , they measured stellar distances which allowed them to construct a map of the 3d structure of the sky area covered by current sdss survey .
their study has , for the first time , revealed large - scale patterns in the galaxy . in the same data that showed the virgo overdensity ,
an underdense area with respect to an axisymmetric halo model is also obvious in the direction of uma .
whereas the overdensity can be naturally interpreted as a large - scale star stream in the halo model , the uma underdensity can not be accounted for in the same way .
we believe that this observational fact is linked to the smooth structure of the halo in which the star streams ( including the virgo overdensity ) are embedded . to tackle this problem , a similar star count analysis of galactic structure will be followed in this paper .
the sloan digital sky survey is a five - colour ( @xmath9 ) broad - band photometric survey with a wavelength coverage from 3000 to 11,000 ( fukugita , ichikawa & gunn 1996 ) .
the dr4 imaging catalogue covers 6670 square degrees .
its detection repeatability is complete at a 95% level for point sources brighter than the limiting apparent magnitudes of 22.0 , 22.2 , 22.2 , 21.3 and 20.5 for @xmath10 and @xmath11 , respectively ( http://www.sdss.org ) .
it has facilitated great progress in the study of the structure of the galaxy , which was reviewed in the sdss - ii segue project .
for the study of galactic structure using star counts , sdss provides the most up - to - date and complete stellar sample in terms of both sky coverage and magnitude limit .
thus , the sdss is suitable to study the large - scale asymmetry of the galaxy . in this paper
, we build a model of the galaxy with a non - axisymmetric geometry , and constrain the parameters of the model using observations in a set of characteristic directions taken from sdss dr4 . in section 2 , the selected observational data are described . in section 3 ,
the model used to fit the observational data is introduced .
the results of triaxial halo model and the asymmetric structure of the galaxy are described and analysed in section 4 .
a discussion of the present results and our conclusions are given in the final section .
the sdss is based on observations obtained with a 2.5 m telescope with a 3@xmath0 field of view at the apache point observatory in new mexico , usa .
standard operations of the sdss were performed from 2000 to 2005 .
eventually , it will cover @xmath12 steradians centred on the northern galactic cap and three stripes near the southern galactic cap ( stoughton , lupton & bernardi 2002 ) .
the operating mode of observations is by means of drift scanning .
one continuous scan is designated by a `` run '' , and two such runs generate a `` stripe '' ; they have an overlap region of 1 arcmin on the edges of the two scans ( chen et al .
catalogues of objects classified as `` stars '' from dr4 were downloaded from the sky server for sky areas selected to meet our aims .
the retrieved data is free of overlap and has been deblended ; saturated stars and poor - quality objects near bright stars and the edges of the frames were all filtered out ( newberg & yanny 2005b ) .
if the galaxy is axisymmetric , as usually assumed in previous studies , the projected stellar number density at the same galactic latitude will increase towards the direction of @xmath13 , and decrease when approaching @xmath14 , and any sky area pair mirror - symmetrically selected on either side of the @xmath13 meridian plane should retrieve the same star number counts , within the uncertainties .
if the axisymmetric model is correct , such a picture should be reproduced by star counts across the whole sky area , and can in principle also be detected in symmetrically selected subsets of sky areas . limited by the current sdss sky coverage , whole - sky star counts are not possible , but a uniform data set covering all longitudes at a given latitude should suffice , as this would be a perfect probe to check the axisymmetric assumption . such a data set is now available from sdss dr4 , which will be addressed in the following . fig .
1 is a lambert projection of the northern galactic hemisphere , showing the selected sky areas used in this work ( marked by diamonds ) .
each of the selected sky areas has a size of about @xmath15 ; the galactic coordinates of the centres of the selected fields are given in table 1 .
the sky areas selected along the circle at @xmath16 form the first group of our working data set . to check the structure at latitudes different from @xmath16 , sky area pairs at different latitudes ( from @xmath17 , equally spaced by @xmath18 ) along our longitudinal grid ( in steps of @xmath19 from 0 to @xmath20 360@xmath0 ) are also selected . in order to minimise the effects of extinction , sky areas below @xmath21
were excluded . from the present sdss public data release ,
sky areas of 11 longitudinal strips are available , and selected field from these form the remaining 11 data groups in our sample . from the second to the twelfth group ,
the sky areas within each group have the same longitude , so that the structural properties reflected by the stellar number densities along a given longitude can be derived .
in addition , through comparison between sky areas having mirror symmetry with respect to the @xmath13 meridian deviations from the axisymmetric galaxy model can also be checked .
owing to the sdss observing strategy , stars brighter than @xmath22 will be saturated , and star counts will be not complete for magnitudes fainter than @xmath23 in the @xmath24 and @xmath25 bands , so that our work will be limited in the range of magnitudes from @xmath22 to @xmath26 in both the @xmath24 and @xmath25 bands .
to fully reveal the structure of the milky way , accurate measurements of the distances to individual objects are key ingredients .
such information is still lacking at the present time , however . as a consequence
, the structure of the galaxy can not be resolved satisfactorily .
very large - scale photometric surveys greatly improve our knowledge of galactic structure .
however , the problem remains even if we use the data generated by the sdss project , because of our inability to determine stellar distances accurately .
the major science drivers of the sdss are focused on extragalactic objects .
therefore , a wide - filter photometric system was designed to detect the faintest possible extragalactic objects .
the narrowest filter , @xmath27 , is 600 .
such a photometric system has an obvious disadvantage when one wants to measure absolute stellar magnitudes , compared with the classical @xmath28 system .
the latter system by itself is already characterised by a scatter of 0.3 mag for a and f - type stars ( crawford 1975 , 1979 ) .
the difficulty to distinguish between giant and main - sequence stars will affect counts of g and k - type stars even more seriously in the sdss system if one applies the colour - magnitude relation for main - sequence star to all stars .
furthermore , the large range of metallicity that is intrinsically present in the sample will inherently lead to large scatter in the colour - magnitude relation . to obtain distances to stars and to determine the structure of the galaxy directly from sdss data
is thus somewhat difficult . a preliminary approach to this problem using the photometric parallax method and its pitfalls are carefully analysed by juri et al .
( 2005 ) . to derive the structure of the galaxy from the integrated stellar light distribution along the line of sight , a theoretical model is needed .
our galactic disk models follows is based on that of bahcall & soneira ( 1980 ) , while we have also included a thick - disk component . because of large deviations from an axisymmetric structure ( described in section 4 ) , we have modified the standard axisymmetric halo ; these deviations will be accounted for by a triaxial halo instead .
the basic star count equation we use is : @xmath29 where @xmath30 is the projected surface density in the absolute magnitude bin from @xmath31 to @xmath32 in the direction of galactic longitude @xmath1 and galactic latitude @xmath33 ; @xmath34 is the heliocentric distance of a given star ; @xmath35 is the density profile of each stellar population as a function of @xmath36 , the distance to the galactic centre ; @xmath37 is the luminosity function of each contributing stellar population , which is a function of absolute magnitude and d@xmath38 is the solid angle element covered by the observations . for the thin and thick disk components ,
the density profile is assumed to be of the following exponential form , @xmath39 , \label{eq_diskdensity}\ ] ] where @xmath11 is the height above the galactic plane , @xmath40 the projected distance to the galactic centre in the galactic plane , and @xmath41 and @xmath42 are the scale height and scale length in @xmath11 and @xmath40 , respectively . as this study focuses on the structure of the halo , we used fixed models for both the thin and the thick disks , so that complications due to extra free parameters for these disks can be excluded .
the parameters of the disks are taken from chen et al .
for the thin disk , @xmath43 pc and @xmath44 kpc ; for the thick disk , @xmath45 pc and @xmath46 kpc .
the galactocentric distance of the sun is @xmath47 kpc , and the vertical distance of the sun from the galactic disk plane is @xmath48 pc . for the halo
, a triaxial model is adopted to fit the observational data .
a set of suitable coordinate frames is built following newberg & yanny ( 2005b ) , of which the schematic diagram is shown in fig .
2 . adopting the position of the sun as the origin of the coordinate frame , the direction from the sun to the galactic centre is taken as the @xmath40-axis .
the y - axis is the direction perpendicular to the @xmath40-axis in an anti - clockwise sense within the galactic plane , while the @xmath11-axis coincides with the direction towards the northern galactic pole .
the @xmath49 and @xmath50 directions are parallel to @xmath40 and @xmath51 , but their origin is at the galactic centre instead ; these axes also define a right - handed coordinate system .
furthermore , @xmath34 is the distance from a star to the sun , @xmath25 the distance from the star to the galactic centre and @xmath52 is the projection of @xmath25 on the galactic plane .
therefore , @xmath53 , @xmath48 pc , @xmath54 .
the galactic centre coordinate frame is rotated in such a way that its axes are settled right on the three axes of the triaxial ellipsoid .
the transformation between the cooordinate frames can be separated into three rotational steps : first , fix arbitrarily one of the three axes and rotate the other two .
the rotation angles of the three steps , denoted by @xmath55 , @xmath56 and @xmath57 , are defined when fixing the @xmath11 , @xmath40 and @xmath51 axes , respectively . in fig .
2 , for example , where we rotate @xmath49 and @xmath50 around the @xmath58 axis , the relevant transformations are , @xmath59 , @xmath60 , and @xmath61 .
the remaining two steps involve doing the same for the other two axes .
the new axes obtained after the third rotation are @xmath62 , and the three axes of the triaxial ellipsoid are @xmath63 and @xmath64 , where @xmath65 is the longest and @xmath64 the shortest axis . the axis ratios of the triaxial ellipsoid are @xmath66 , and @xmath67 , while @xmath68 .
the axisymmetric halo is described by @xmath69 .
a power - law ( reid 1993 ) is adopted for the halo density distribution , @xmath70 where @xmath71 is a normalisation constant .
we adopt the @xmath72-band luminosity function of robin & cr ( 1986 ) in our model , which provides sufficiently accurate two - dimensional distributions in both luminosity and spectral type down to @xmath73 mag .
a two - dimensional @xmath72-band luminosity function can easily be transformed to that in the @xmath24 band .
robin & crz ( 1986 ) compiled a number of luminosity functions available at the time , including those of wielen ( 1974 , 1983 ) and houk & cowley ( 1975 , 1978 , 1982 ) .
the work of wielen is widely used in star count studies to obtain number density - magnitude relations for @xmath74 mag , while houk & cowley s luminosity function is often used for the bright stars ( @xmath75 mag ) .
deutschman , davis & schild ( 1976 ) and hayes ( 1978 ) provided calibrations of stellar luminosity and spectral type . to deal with the data used for the present study , a further transformation relation ( chen et al
. 2001 ) from @xmath72 to @xmath24 , @xmath76 , was also adopted . by using the density profile and the luminosity function discussed and presented above
, the distribution in absolute @xmath24 magnitude of the stellar projected surface number density can be obtained from eq .
a colour - magnitude relation ( cmr ) is needed to obtain the number density distribution in the @xmath5 colour .
it is usual practice to use the cmr defined by colour - magnitude diagrams ( cmds ) of star clusters .
however , there are usually too few stars in cluster cmds with absolute magnitudes fainter than @xmath77 . therefore , at fainter magnitudes we do not have a precise cmr . as an alternative approach , we adopt the theoretical isochrones of girardi , grebel & odenkirchen ( 2004 ) to derive the cmr required here .
the mean ages of thin disk , thick disk and halo are assumed to be 4.5 , 11 and 13 gyr , and their metallicities are assumed to be , respectively , 0.019 , 0.004 and 0.0012 .
the theoretical cmr is preferred because it includes faint stars that are otherwise not available from observations .
however , we have to bear in mind that it can not mimic the scatter on the relation in real situations .
given a theoretical model for stellar distributions of different galactic components as described above , it is straightforward to calculate the projected surface number density in absolute magnitude bins in any given direction . from observations ,
the information that can be directly extracted is the projected surface number density in apparent magnitude bins .
the goal of the present study is to use this information to constrain the structure of the galaxy parameterised in the manner explained above .
a monte carlo method was adopted to simulate the observational statistics .
random points were generated based on several quantities , including the luminosity function , distances weighted by the input density profile , and a gaussian - type observational error , so that the observed star counts can be reproduced . for these simulations , we adopted the three - dimensional extinction model of the milky way derived from cobe observations ( drimmel , cabrera - lavers & lpez - corredoira 2003 ) . instead of correcting the observational data for the very complicated interstellar reddening
, we applied reddening corrections to the stellar distributions predicted by the model .
the extinction , @xmath78 , is in general small throughout the areas selected in this study , with a maximum of about 0.104 mag in the direction ( 0@xmath0 , 55@xmath0 ) and a minimum of 0.0385 mag in the anti - galactic centre direction ( @xmath79 ) , at a distance of about 600 pc along the line of sight .
@xmath78 is obtained through linear interpolation for distances less than 600 pc . for greater distances , @xmath78 is taken to be constant .
the @xmath24-to-@xmath72-band extinction ratios were taken from girardi et al .
( 2004 ) , and thus we can also obtain the apparent @xmath24 and @xmath25-band magnitudes at each theoretical point .
we can now caluclate the theoretical projected surface number density , theoretical star counts as a function of apparent @xmath24 and @xmath25 magnitude , and theoretical star counts as a function of @xmath5 colour down to the sdss limiting magnitudes , and hence we are now ready to compare these with the observational data .
the discrete small sky areas of group 1 , covering a complete circle at the galactic latitude of @xmath33=@xmath80 , form a unique probe of the structure of the galaxy in the current sdss data release .
direct measurements of the stellar number density distribution within this data set can tell us whether or not it is axisymmetryical .
we tested this idea by fitting the surface density distribution along the circle with an axisymmetric halo .
4 shows the projected surface number density distributions of the theoretical model and the observational distribution for stars of @xmath81 @xmath82 @xmath83 $ ] mag , where the diamonds represent the observational projected stellar surface number density . to account for observational fluctuations , each sky field
was divided into four subfields of @xmath84 ( except for the area [ @xmath85,@xmath80 ] , which is too small for this procedure ) , so that the statistical fluctuations can be measured .
the resulting fluctuations over the average of the four subfields are shown in the figure as error bars on the data points .
the dashed line is the theoretical prediction for an axisymmetric model .
the dashed theoretical line was normalised by taking the average values of the surface density at ( @xmath4,@xmath80 ) and ( @xmath2,@xmath80 ) from the observational data .
it is worthwhile to point out that the two parts of the dashed theoretical line , i.e. , that between @xmath87 and @xmath2 , and that between @xmath2 and @xmath3 , are not perfectly symmetrical because of extinction . in addition , due to the non - uniform extinction , the surface number density curves shown in fig
4 are not smooth .
it is evident that the axisymmetric model does not fit the observations very well .
comparing the observational data with the axisymmetric model prediction , it is obvious that the observational data between @xmath88 and @xmath2 on the one hand , and between @xmath89 and @xmath3 on the other are highly asymmetrical ; the projected surface density in the fields at @xmath1 @xmath90 is systematically higher than that in the fields mirror - symmetrically located on the other side of the meridian . the axisymmetric theoretical value is a little higher than derived from the observations for longitudes from @xmath4 to @xmath2 , and much lower than implied by the observational data for @xmath89 to @xmath3 .
the deviations from the axisymmetric model are rather smooth ; they seem to represent a systematic shift from the theoretical model . to find out what kind of stars are responsible for these asymmetric statistics , we performed a number of tests . in table 2
we list the star counts of stars with @xmath91 @xmath82 @xmath83 $ ] mag in the paired sky fields selected symmetrically with respect to the @xmath92 meridian plane , as well as the relative fluctuations ( defined as [ surface density of @xmath93 - surface density of @xmath94/[surface density of @xmath94 , where @xmath95 and @xmath93 are the galactic longitudes of the paired fields ) . in table 3 , faint stars with @xmath91 @xmath82 @xmath96 $ ] mag are considered in the same way .
they show the same trend as the previous data set , resulting in an uneven distribution with an even higher asymmetric ratio . in fig .
3 , star counts in four colours , @xmath97 , @xmath5 , @xmath98 and @xmath99 , in the fields towards the directions of ( @xmath100 , @xmath80 ) and ( @xmath101 , @xmath80 ) are compared .
the magnitude ranges are , respectively , @xmath102 , @xmath103 , @xmath104 , @xmath105 , and @xmath106 . in each of the colour distribution profiles there are two peaks , as shown in fig .
3 ( the blue tail at @xmath107 is mostly due to white dwarf stars ) .
the left - hand peak is dominated by halo stars , while the one on the right is dominated by thin disk stars , as inferred from the colours .
this is because most of halo stars are far more distant than the bulk of the disk stars , and therefore only the luminous stars in the halo ( predominantly main - sequence stars ) can be detected . for stars on the main sequence
, the intrinsically brighter ones have bluer colors .
in addition , the halo stars are generally more metal poor , and hence tend to populate the blue peak . based on a similar argument related to the stellar metallicity
, thin disk stars would form the red peak . therefore , it is clear that the excess of the stellar number density in the field at ( @xmath101 , @xmath80 ) with respect to that at ( @xmath100 , @xmath80 ) is due to the halo component .
the amplitude of the excess depends on colour , as shown in fig .
3 ; it is larger in @xmath108 and @xmath98 than in @xmath99 .
it may be caused by the different limiting magnitudes , or it might be an indication of the specific properties of the stars that causes the excess .
a triaxial halo model , as described in section 3 , is adopted to describe the asymmetry found in our data set .
given the same limiting magnitudes as those of the observations , the theoretical surface densities are calculated using a model of which the structural parameters are constrained by a comparison with the observations .
the three - component model ( including the thin and thick disks , and the halo ) is used to calculate projected surface number densities in the sky areas corresponding to those of the observations . as the main objective of the current work is to derive the structure of the halo , for the parameters of the thin and thick disk we adopt the results of chen et al .
the parameters of the halo are adjusted in such a way that the observational data are best fitted .
the integration of the observational surface density over all of the selected fields is used to constrain the surface density of the model .
there are a total of six parameters in our halo model , namely the power - law index @xmath109 , the two axis ratios @xmath110 and @xmath111 , and the three coordinate rotation angles @xmath56 , @xmath57 , @xmath55 ( based on which the @xmath112 axes of the triaxial ellipsoidal model can be translated to that of the galaxy ) .
the possible ranges of the parameters are estimated using initial tests over a preliminary grid , which is given in table 4 .
chi - squared minimisation tests comparing the theoretical and observational data sets were carried out using the algorithm of press et al .
( 1992 ) . for a non - gaussian distribution of discrete data points ,
pearson s @xmath113 can be expressed as , @xmath114 where @xmath115 is the theoretical frequency of data points in the @xmath116 bin , and @xmath117 is the equivalent frequency of the observations .
@xmath118 is the number of bins , @xmath119 the number of model parameters , and therefore @xmath120 is the number of degrees of freedom . assuming that the errors in the theoretical and observational frequencies are distributed in a poissonian fashion , we obtain @xmath121 . by changing the model parameters , the parameter combination that generates the minimum averaged @xmath113 of the 12 groups
can be obtained .
models for all possible combinations of the six parameters were calculated , and the corresponding @xmath113s were evaluated to find the parameter combination resulting in the minimum @xmath113 value .
it is obvious that the models with @xmath122 in the range of @xmath123 , and @xmath124 lead to better values for @xmath113 , regardless of any fine - tuning of the other four parameters .
the models provide only weak constraints for the power - law index @xmath109 . when @xmath109 is fine - tuned , the axial ratio should be adjusted accordingly , in order to obtain the optimum @xmath113 .
the overall trend is that for larger @xmath109 , larger @xmath110 or @xmath111 ratios are needed .
figs 4 and 5 show the observational surface densities in the selected regions and those of our theoretical models calculated by using a triaxial halo with the following parameter combination , @xmath125 , and @xmath126 , and the corresponding @xmath113 is 2.02 .
( this is the parameter set that leads to the minimum @xmath127 , see section 4.3 for more details ) .
4 shows the results for the group-1 fields , spanning a circle at @xmath128 . a comparison between the solid and dashed lines in fig .
4 shows that the triaxial halo model ( the solid line ) fits the observations much better than an axisymmetric halo ( the dashed line ) , with the triaxial halo clearly showing the degree of asymmetry as revealed by the observations .
the only sizable deviations of the triaxial model from the observations are found near @xmath129 ( where the theoretical prediction for the surface density is slightly higher ) , and near @xmath130 ( where the theoretical value is somewhat lower ) .
the six panels in fig . 5 , labelled a - f , show the surface densities of both the observations ( diamonds and triangles ) and the models ( dotted and dashed lines ) for groups 2 to 12 .
panel a shows the surface densities in the fields at @xmath1=@xmath4 and @xmath1=@xmath2 as a function of galactic latitude .
panels b through e show the surface densities in the sky - field pairs mirror - symmetrically selected on either side of the 0 degree galactic meridian plane .
panel f depicts that in the @xmath1=@xmath131 field , the companion field of which is not shown because the data has not yet been released . from these figures , it is clear that there are larger surface densities in the sky areas between @xmath89 and @xmath3 than in their corresponding mirror - symmetric fields .
these fields thus show the same trend as the sky areas in the first group , at @xmath16 .
the theoretical model fits the observations quite well in general , with several exceptions .
for the three groups at @xmath1=@xmath132 , and @xmath101 , the values are higher than those of the observations around a latitude of @xmath33=@xmath133 . for the group at @xmath134 and @xmath1=@xmath101 , the theoretical values are lower than the observational value around @xmath135 .
for the group at @xmath130 and @xmath1=@xmath136 , the theoretical values are lower than the observational values around @xmath137 and @xmath138 , respectively . in this paper ,
a power - law model is used to interpret the observations for the purpose of clarity .
models implemented with more complex physics , such as the hernquist model ( newberg & yanny 2005b ) , can fit the observations better
. this could be one of the reasons that our model does not fit the observations perfectly .
modifications to this problem will be considered in a future study . in the previous section
, the projected surface number densities were fitted using a triaxial halo model .
star counts combined with colour information can , however , reveal more information about the stellar populations in the data and constrain the model better .
therefore , we performed model fits in the @xmath5 colour .
the model described in section 3 is now used to calculate star counts in the @xmath5 colour . by using eq .
( 1 ) , the projected stellar number density in each absolute magnitude bin can be obtained .
monte carlo simulations were carried out for absolute magnitude and distance , so that artificial stars could be generated for a given model , and hence theoretical apparent magnitudes could be generated .
the @xmath5 colour as well as its statistical uncertainty at each point can then be obtained from the colour - magnitude relation .
comparisons between the theoretical and observational star counts for all the fields were also done following eq .
( 4 ) . subsequently
, the @xmath139 values of all fields in the 12 groups were averaged to obtain the mean value @xmath140 .
intensive and elaborate computational tests were done for all possible cases .
each set of parameters represents a model which can be evaluated by its @xmath113 and @xmath140 .
it was found that @xmath113 and @xmath140 exhibit similar trends .
however , when @xmath140 reaches its minimum value , @xmath113 may not be at its minimum . therefore , we have to sacrifice minimizing the latter in order to keep @xmath140 small .
table 5 lists the best - fit parameter combinations for power - law indices @xmath109 from 2 to 2.6 .
because of the six - dimensional parameter space , multiple solutions were derived .
fitting actual star counts in @xmath5 is much more difficult than fitting the projected surface number density .
the minimum @xmath113 value is about 1.735 , but that of @xmath140 is a few times higher . in our model ,
the power - law index @xmath109 still leaves much room for fine - tuning , from 2.0 to 2.6 ( beyond which @xmath113 and @xmath140 become too large ) .
if constrained only by the projected number density ( see section 4.2 ) , no firm constraint on the axial ratios @xmath122 can be reached . when supplemented with star counts in @xmath5 , each @xmath109 has only one best - fit @xmath122 combination . for @xmath141
, @xmath122 is @xmath142 .
this best - fit parameter set is different from that of the traditional axisymmetric model , even if @xmath143 . in this model ,
the y axis is equal to the z axis , and therefore resembles the rotationally symmetric model , but with an inclination and tilt with respect to the galactic plane .
we would like to call all these @xmath143 cases also triaxial models .
in such cases , good fits to the observational data can also be achieved .
however , the ellipticity seems too oblate .
when @xmath144 , @xmath122 becomes @xmath145 ; when @xmath146 or 2.6 , @xmath122 goes to @xmath147 .
the larger @xmath109 becomes , the higher the axial ratio will need to be . directly following from the power - law distribution , the number density decreases more rapidly for larger @xmath109 when going away from the galactic centre .
a higher axial ratio tends to enhance the number density in the directions of the axes .
hence , for larger @xmath109 , a larger axial ratio is required to balance the effect of the larger @xmath109 in order to fit the observational data .
such a correlation is the same as that for the direct surface number counts in section 4.2 .
figs 6 and 7 show star counts in @xmath5 for the @xmath33=@xmath80 fields .
the theoretical values were calculated with the same model as in figs 4 and 5 .
our purpose is to compare the mirror - symmetric pairs ; four pairs of fields at approximately this latitude are considered . because the sky field at ( @xmath148 ) is not included in the survey , we chose the field at ( @xmath149 ) to be paired with that at ( @xmath150 ) .
the thin - line histogram represents the observational data , the solid black histogram describes the theoretical results . in fig .
6 , panel a shows the results for the sky field at ( @xmath151 ) , and panel b for ( @xmath152 ) .
the panel pairs c - d , e - f , g - h and i - j are for four pairs of mirror - symmetric sky fields .
panel k in fig .
7 is for the field at ( @xmath153 ) , a single field without a companion on the other side of the galactic meridian plane . comparing the observational data in the paired sky fields , the histograms for @xmath1 from @xmath2 to @xmath3 exhibit higher halo star peaks , and similar peaks for the thin disk with respect to those on the other side of the meridian .
the theoretical model follows the same trend as the observations .
apart from the problem of the structural parameters , the stars in the galaxy are simplified as three populations ( thin and thick disks , and halo ) , and it is further assumed that each population internally has the same age and metallicity .
all of these assumptions will surely affect the distribution of the stars in the @xmath5 colour . a more precise @xmath24-band luminosity function and better partitioning of the populations
are needed to interpret the continuously updating large - scale survey data .
star counts based on the observational data of the milky way analysed in the present paper show an asymmetry in the large - scale areas of the sky surrounding the northern galactic pole , and therefore they reflect a certain non axisymmetric structure . these observations may be explained in one of two ways .
the present smooth galactic halo itself may have an asymmetric structure , which can be represented by a triaxial ellipsoid , possibly tracing the dark halo gravitational potential that might be triaxial ( jing & suto 2002 ) .
the triaxiality proposed as such is not unique ; it is considered common for dark haloes in extragalactic systems .
mazzei & curir ( 2001 ) analysed the effects of a triaxial dark halo on the bar - triggered star formation and feedback processes after the formation of the disk .
they indicated that the dark halo has important effects on the evolution of baryonic matter .
the star - formation rate is not only linked to the total mass of the dark halo , but also to the dynamical state of the dark halo .
is it then normal to form an asymmetric stellar halo in such an asymmetric gravitational field ? in the history of formation and evolution of the galaxy , the halo experienced accretion and merger processes .
wyse & gilmore ( 2005 ) claim that merger events and accretion of satellite galaxies in the past history of the galaxy do play some role in the evolution of the milky way . the outer halo , in particular ,
might be dominated by substructures that are likely the remnants of interactions .
if the halo were very sparse originally and were gradually built up from accretion and merger events , it would be expected that there have been accretion events happening throughout the lifetime of the milky way .
this would allow for any sizable features from early times to be smoothed out via dynamical evolution , leading to the formation of a smooth structure at the present time .
presumably , the star streams should have a random distribution due to the random infall angle and intensity , and therefore on average the overall features would have a smooth distribution after relaxation . newberg and yanny ( 2005b ) selected f - type turn - off stellar data in eight sdss stripes from the dr3 to map the distribution of stars in real space .
they selected fields that most likely contain only halo stars .
sky areas near any previously identified overdensities caused by star streams were also avoided in their studies , but their results can be altered by the virgo overdensity discovered later by juri et al .
after they compared the observational data with the results of five triaxial models , they obtained the best - fit parameter set for an oblate ellipsoid with its major axis pointed at 50@xmath154 from the line of sight towards the galactic centre from the sun , and with a 46@xmath0 inclination with respect to the galactic plane .
the minor axis , z , has a length equivalent to 65% of the major axis .
the intermediate axis , y , is about 75% of the major axis ( newberg & yanny 2005b ) .
an alternative cause of an asymmetric halo may be found in large - scale star streams that are embedded in the smooth background and that have not yet been melted into the halo completely . also using dr4 data , juri et al .
( 2005 ) analysed the structure of the galaxy using the photometric parallax method .
they found that there are significant overdensities in the range ( @xmath155 ) , with the overdensity peaking in the direction of virgo , and the entire overdensity region covers about one thousand square degrees on the sky .
they selected both a sky field in the overdensity region and a control sky field mirror - symmetrically located on the other side of the meridian with respect to the target field , as shown in figure 24 of juri et al .
( 2005 ) . by comparing stars with @xmath156 mag and @xmath157 mag in both fields , they estimated the surface brightness and luminosity of the overdensity in the target field as @xmath158 mag arcsec@xmath159 , and @xmath160 l@xmath161 .
based on this result , we can calculate the additional number of stars per square degree due to the overdensity .
the magnitude range used by juri et al .
( 2005 ) spanned from @xmath162 to 21.5 mag . adopting the distance to the overdensity as 10 kpc , as also assumed by juri et al .
( 2005 ) , the lower and upper limit of the @xmath25-band absolute magnitude of their sample stars is therefore from 6.5 to 3 mag . from the transformation relation @xmath163 ( chen et al .
2001 ) , the limiting magnitudes in the @xmath72-band can be obtained ( taking the @xmath72
@xmath164 relation for main - sequence stars from cox 1999 ) .
the mass @xmath72 magnitude relation from reid ( 2002 ) , @xmath165 is used to calculate the mass of the most luminous ( 1.45 m@xmath161 ) and the faintest stars ( 0.81 m@xmath161 ) corresponding to the respective observed magnitude thresholds . integrating the luminosity - mass ratio , weighed by @xmath166 imf ( reid 2002 ) , the additional stellar density number per square degree can be calculated , @xmath167 where @xmath64 is the normalisation constant of the imf . the derived number is 614 stars per square degree , which is indeed enough to match the number density excess with respect to a symmetric structure . however , this is still not enough to fully resolve the problem . by subtracting the cmd of a control field from that of the sky area characterised by the density excess ,
the flux of the virgo star stream was derived . however
, if a triaxiality does exist in the structure , there would be no proper criteria to distinguish the contribution due to the star stream from that of triaxiality . as demonstrated in fig . 4
, there is a large region of stellar underdensity with respect to an axisymmetric model , which is centred at about @xmath168 in the direction of ursa major ( uma ) with an amplitude less than the virgo overdensity .
this was also shown in figure 3 of newberg & yanny ( 2005a ) , where they quote a minimum at @xmath169 for f - type turn - off stars , very close to what we have found in this paper .
the size of the uma underdensity region is smaller than , but comparable to , that of the virgo overdensity .
the uma underdensity is also clearly visible from figure 24 of juri et al .
( 2005 ; the blue colour - coded region near @xmath168 ) , and more prominently shown on the colour map in a recent news release on the sdss web page .
such an underdensity region can also be found in the spatial distribution of m - giants from 2mass observations ( figure 27 of juri et al .
the virgo overdensity and the uma underdensity regions are well aligned .
if the overdensity is to be accounted for only by a star stream , what is then the cause for the observed underdensity aligned with it ?
this infers that there is some substructure within the halo that can not be accounted for only by a star stream such as that of the virgo overdensity . as shown in fig .
4 , the overdensity is observed in a very large region , from @xmath170 to @xmath171 , which is more than a quarter of sky at the latitude of @xmath172 .
it can not be fully understood by a stream covering only 1000 deg@xmath173 .
the triaxial halo model may provide a plausible explanation , because the distribution of the stellar surface number density at @xmath16 , including these over- and under - densities , can be fitted simultaneously by a triaxial halo model . to disentangle the effects of a triaxial halo and the large - scale star stream , kinematic and chemical information is needed .
the substructures detected in the sdss data are summarised in table 2 of newberg , yanny & rockosi ( 2002 ) ; two of their star streams pass through the sky fields selected in this study , namely s341 + [email protected] and s297 + [email protected] .
since s341 + [email protected] is below the magnitude limits of the survey , it can hardly alter our results .
s297 + [email protected] may have a minor influence on our results because of its limited size in the given direction .
however , it will not change the general large - scale picture of the current work . as the primary goal of the current work
, a study of the large - scale structure of the galactic halo using star counts is presented .
the star streams can not easily be corrected for from the sdss data .
therefore , we chose all - sky coverage to try to fit the average structure .
we have shown that the stellar projected surface density of the halo stars at @xmath1 from @xmath2 to @xmath3 is significantly higher than that at @xmath1 from @xmath4 to @xmath2 at a latitude of 60@xmath0 below the sdss magnitude limit .
the observed stellar projected surface density distributions in the sky fields selected for the current work can be well fitted by using a triaxial halo model .
the parameter combinations of the best - fit triaxial models are , power - law index @xmath109 from 2.0 to 2.6 , @xmath55 from @xmath133 to @xmath175 , @xmath57 , @xmath56 less than @xmath176 , @xmath177 either @xmath178 , @xmath145 , or @xmath147 .
these parameter sets agree with the one of newberg & yanny ( 2005b ) , except that their axial ratios are larger than ours .
the results of the current work can still be affected by lumps in the star streams that have not yet been smoothly dispersed into the halo .
we are left with a number of questions , such as : what are the bulk of the stars in the star streams ? what is the proportion of stars in the streams across the whole halo ?
do star streams contribute significantly to the formation of a triaxial halo , or do they simply act as minor disturbances to the overall profile ?
how many of the star streams have already dispersed into the halo and become a smooth background , and how many of them can still be picked up from the smooth halo background ? there is no way to have any clear answer for these questions from the present data .
work in this subject will leap forward with the coming on - line of future projects aimed at spectroscopic sky surveys such as segue , lamost , sdss ii , gaia , and further photometric surveys of the southern sky .
we would like to thank b. chen , y. c. liang , s. d. mao , z. g. deng , x. y. xia , c. h. du , y. b. yang , and y. g. wang for valuable suggestions and fruitful discussions .
we are also very grateful for email communications with h. j. newberg and a. robin .
we would like to thank our colleague , dr .
r. de grijs , for proof reading the paper .
this work is supported by the national natural science foundation of china under grant no .
10333060 , no . 10573022 and no .
10403006 .
bahcall , j. n. , soneira , r. m. , 1980 , apjs , 44,73 blitz , l. , spergel , d. n. , 1991 , apj , 370 , 205 chen , b. , stoughton , c. , allyn smith , j. , et .
al . , 2001 ,
apj , 553 , 184 crawford , d. l. , 1975 , aj , 80 , 955 crawford , d. l. , 1979 , aj , 84 , 1858 deutschman , w. a. , davis , r. j. , schild , r. e. , 1976 , apjs , 30 , 97 drimmel , r. , cabrera - lavers , a. , l@xmath179pez - corredoira , m. , 2003 , a&a , 409 , 205 fukugita , m. , ichikawa , t. , gunn , j. e. , et .
al . , 1996 , aj , 111 , 1748 gilmore , g. , 1984 , mnras , 207 , 223 girardi , l. , grebel , e. k. , odenkirchen , m. , et .
al . , 2004 ,
a&a , 422 , 205 gould , a. , bahcall , j. n. , flynn , c. , 1996 , apj , 465 , 759 hartwick , f. d. a. , 2000 , aj , 119 , 2248 hayes , d. s. , 1978 , proc .
iau symp . , 80 , the hr diagram : the 100th anniversary of henry norris russell , p. 65
houk , n. , cowley , a. p. , 1975
, michigan catalogue of two - dimensional spectral types for the hd stars vol .
1 , univ michigan , ann arbor houk , n. , cowley , a. p. , 1978
, michigan catalogue of two - dimensional spectral types for the hd stars vol .
2 , univ michigan , ann arbor houk , n. , cowley , a. p. , 1982
, michigan catalogue of two - dimensional spectral types for the hd stars vol .
3 , univ michigan , ann arbor jing , y. p. , suto , y. , 2002 , apj , 574 , 538 juri@xmath180 , m. , ivezi@xmath180 , @xmath181 . ,
brooks , a. , et al . , preprint ( astro - ph/0510520 ) kron , r. g. , 1978 , phd thesis , university of california , berkeley larsen , j. a. , humphreys , r. m. , 1996 , apj , 468 , 99 majewski , s. r. , siegel , m. h. , kunkel , w. e. , et .
al . , 1999 ,
aj , 118 , 1709 mazzei , p. , curir a. , 2001 , a&a 372 , 803 newberg , h. j. , yanny , b. , rockosi , c. , et .
al . , 2002 , apj , 569 , 245 newberg , h. j. , yanny , b. , preprint ( astro - ph/0502386 ) newberg , h. j. , yanny , b. , preprint ( astro - ph/0507671 ) parker , j. e. , humphreys , r. m. , larsen , j. a. , 2003 , aj , 126 , 1346 parker , j. e. , humphreys , r. m. , beers , t. c. , 2004 , aj , 127 , 1567 peterson , b. a. , ellis , r. s. , kibblewhite , e. j. , et .
al . , 1979 ,
apj , 233 , l109 press , w. h. , teukolsky , s. a. , vetterling , w. t. , et .
al . , 1992 , numerical recipes in fortran : the art of scientific computing , cambridge university press , p. 616
reid , n. , 1993 , aspc , 49 , 37 reid , i. n. , yan , l. , majewski , s. , et .
al . , 1996 ,
aj , 112 , 1472 reid , i. n. , 2002 , aj , 124 , 2721 robin , a. , crz , m. , 1986 , a&a , 157,71 seares , f. h. , 1925 , apj , 61 , 114 siegel , m. h. , majewski , s. r. , reid , i. n. , et .
al . , 2002 , apj , 578 , 151 spergel , d. n. , blitz , l. , 1988 , baas , 20r1017s stoughton , c. , lupton , r. h. , bernardi , m. , et al . , 2002 , aj , 123 , 485 tyson , j. a. , jarvis , j. f. , 1979 , apj , 230 , 153 wielen , r. , 1974 , highlights of astronomy , 3 , p. 395
g. contopoulos wielen , r. , jahreiss , h. , kruger , r. , 1983 , in davis phillip , a. g. , upgren , a. r. , eds , proc .
iau colloq .
76 , nearby stars and the stellar luminosity function , p. 163 wyse , r. f. g. , gilmore , g. , preprint ( astro - ph/0510025 ) zheng , z. , flynn , c. , gould , a. , et al . , 2001 ,
apj , 555 , 393 zheng , z. , flynn , c. , gould , a. , et .
al . , 2004 , apj , 601 , 500 | using the stellar photometry catalogue based on the latest data release ( dr4 ) of the sloan digital sky survey ( sdss ) , a study of the galactic structure using star counts is carried out for selected areas of the sky .
the sample areas are selected along a circle at a galactic latitude of + 60@xmath0 , and 10 strips of high galactic latitude along different longitudes .
direct statistics of the data show that the surface densities of @xmath1 from @xmath2 to @xmath3 are systematically higher than those of @xmath1 from @xmath4 to @xmath2 , defining a region of overdensity ( in the direction of virgo ) and another one of underdensity ( in the direction of ursa major ) with respect to an axisymmetric model .
it is shown by comparing the results from star counts in the @xmath5 colour that the density deviations are due to an asymmetry of the stellar density in the halo .
theoretical models for the surface density profile are built and star counts are performed using a triaxial halo of which the parameters are constrained by observational data .
two possible reasons for the asymmetric structure are discussed .
[ firstpage ] stars : statistics the galaxy : structure , fundamental parameters , halo |
the discovery of gliese 229b ushered in a new chapter in stellar astronomy by penetrating unambiguously below the main sequence edge ( nakajima et al . 1995 ; oppenheimer et al .
1995 ) . in gliese 229b ,
absorption in the far infrared due to collision - induced absorption ( cia ) by h@xmath1 and the lack of absorption in the near infrared at the classic steam opacity windows together result in a redistribution of gliese 229b s emergent flux from the mid- and far - ir to the @xmath2 ( @xmath01.05 ) , @xmath3 ( @xmath01.25 ) , @xmath4 ( @xmath01.6 ) , and @xmath5 ( @xmath02.2 ) bands in the near - infrared .
these band fluxes exceed the corresponding black body values for a given by as much as two to three orders of magnitude . given the importance of the near - infrared bands , it is only natural that they be used to help establish the associated new spectral types ( burgasser et al .
2001a ; geballe et al .
2002 ) . following on the heels of gliese 229b , in the last few years more than twenty
so - called t " dwarfs ( strauss et al .
1999 ; burgasser et al . 1999,2000a,2000c,2001a ; cuby et al . 1999 ; tsvetanov et al .
2000 ; leggett et al . 1999,2000 ; geballe et al .
2002 ) and @xmath0200 so - called
l " dwarfs ( ruiz , leggett , and allard 1997 ; delfosse et al .
1997 ; kirkpatrick et al . 1999,2000 ; martn et al .
1999 ) have been discovered ( most by the 2mass , sdss , and denis surveys ) , introducing the first new stellar " spectral types since the establishment and articulation of the original mkk system ( morgan , keenan , and kellman 1943 ; morgan et al .
this new progression of m@xmath6l@xmath6 t from stars to brown dwarfs is one of the most exciting recent developments in stellar " astronomy .
while the use of the near - infrared to characterize and type l and t dwarfs may be natural for low - temperature objects ( most of which are substellar ) , this ignores their many interesting spectral features and unique behavior shortward of 1.0 micron , the classic realm of stellar classification .
it has recently been shown that the potassium resonance doublet near 7700 and the sodium d line(s ) around 5890 , along with their broad wings , dominate t dwarf spectra between 0.5 and 1.0 microns ( burrows , marley , and sharp 2000 ( bms ) ; tsuji , ohnaka , and aoki 1999 ; liebert et al .
it was also shown by bms that the i - band magnitude of gliese 229b measured by golimowski et al .
( 1998 ) and by matthews et al .
( 1996 ) is fully consistent with the potassium wing hypothesis .
the resulting red / purple / magenta visual " appearance and the diagnostic ( and at times , counterintuitive ; see [ system ] ) gravity , , and metallicity dependences of the short - wavelength spectra make the optical ( loosely defined by the ccd " cutoff ) from @xmath00.4 to @xmath01.0 an intriguing subject of study and theoretical inquiry .
hence , in this paper we focus on the spectra of t dwarfs below 1.0 .
we have calculated a new series of self - consistent spectra , colors , and spectral indices for theoretical brown dwarf atmospheres with from 600 k to 1300 k. the low of these models and the fact that they do not incorporate clouds / grains in their atmospheres exclude the l dwarfs from this inquiry ( burrows et al . 2001 ; ackerman and marley 2001 ) .
however , although we concentrate here on the t dwarfs , we still find hints of the presence of clouds , particularly for the early ts . in [ data ]
, we summarize some of the recent t dwarf measurements shortward of 1.0 micron , which we interpret with the new theoretical models . in [ theory ] , we present a full suite of new theoretical spectra .
these spectra will be used later in [ modelfits][comparisons ] to extract information from the new t dwarf data at short wavelengths .
section [ uncertain ] summarizes the major techniques , assumptions , and uncertainties in modeling of t dwarf spectra and includes an aside on the treatment of the alkali line profiles . in [ profbright ] , we give example t / p profiles and representative plots of the @xmath7 temperature , t@xmath8 .
next , in [ define ] we define the spectral indices and colors we use throughout , particularly in [ comparisons ] .
note that [ profbright ] and [ define ] describe diagnostics by which one can better understand the observational trends . then in [ modelfits
] , we provide representative fits to the spectra of gliese 229b and gliese 570d and continue in [ rainout ] with what can be learned with the new data at short wavelengths concerning the rainout of refractories ( burrows and sharp 1999 ; lodders 1999 ) and the line profiles of the alkali metals .
after this , in [ system ] we describe the systematic dependences of the optical and near - infrared spectra on , gravity , and metallicity and in [ comparisons ] we compare the theoretical spectral indices with those obtained using t dwarf spectra . finally , in [ clouds ] we list several hints of the residual influence of clouds in early t dwarf atmospheres ( see also marley et al .
2001 ) and in [ conclusion ] we summarize our general results and conclusions . given the remaining ambiguities in both the gas - phase opacities ( e.g. , of methane , the alkali metals , water ) and in the proper treatment of clouds and given the possible effects of non - equilibrium chemistry ( griffith and yelle 1999,2000 ; saumon et al .
2000 ; lodders 1999 ) , we are less interested in obtaining detailed fits than we are in the overall systematics of the class of t dwarfs .
hence , our focus in this paper is on the generic behavior at short wavelengths of the t dwarf _ family _ as a whole .
this is the first paper in which theoretical models and optical data for a large collection of t dwarfs , not just individual t dwarfs such as gliese 229b ( marley et al .
1996 ; allard et al .
1996 ; tsuji et al . 1996,1999 ; saumon et al .
2001 ) or gliese 570d ( geballe et al . 2001 ) , are compared .
table 1 itemizes 13 t dwarfs ( and 2 late l dwarfs ) with recently - obtained optical data .
it lists objects in order of assigned spectral subtype ( kirkpatrick et al . 1999 ; burgasser et al . 2001a ) and provides the associated references and short - hand names we employ .
more detailed discussions on the keck lris ( oke et al .
1995 ) optical spectra of these t dwarfs and the associated signal - to - noise ratios can be found in kirkpatrick et al .
( 2001 ) and burgasser ( 2001a , b ) .
figures [ fig:1 ] and [ fig:2 ] depict spectra for the objects listed in table 1 ( except for 2mass-1237 ) from 0.6 to 1.0 .
to facilitate intercomparison , these spectra have been normalized to a value of one at 1.0 .
normalization also serves to emphasize that we still do nt know the parallaxes for most of these t dwarfs and , hence , can not assign absolute flux levels .
the data shortward of @xmath00.8 have a signal - to - noise ratio per pixel as low as @xmath01 ( burgasser 2001b ; kirkpatrick et al .
therefore , below 0.8 we have used boxcar smoothing for the noisiest spectra to help discriminate between the relative flux levels of these t dwarfs in the @xmath00.7 region . as is clear from figs .
[ fig:1 ] and [ fig:2 ] , the spectral slopes shortward of 1.0 cover a broad range of values and the spectra themselves fan out into a family that no doubt reflects variations in the underlying physical properties . ignoring the two l dwarfs , the salient features are the water feature near 0.93 , the spectral slope in the 0.8 to 0.95 region , and the relative height of the bump between the ki resonance lines ( @xmath07700 ) and the na d lines ( @xmath05890 ) .
these features inform our choice of the spectral indices we define for this study ( [ define ] ) . indeed , as was demonstrated by liebert et al .
( 2000 ) with their optical spectrum of sdss-1624 ( t6 ) , it is clear from figs .
[ fig:1 ] and [ fig:2 ] that the k i doublet at @xmath07700 and its wings dominate the region from 0.7 to 1.0 ( see [ theory ] ) . also , as is clearest for the early t dwarfs and 2mass-0559 ( t5 , green in fig .
[ fig:1 ] ) , absorption by the na d lines is a natural explanation for the suppression of flux shortward of 0.7 . in principle , the equivalent widths of the cs features at 8523 and 8946 and of the rb features at 7802 and 7949
are additional diagnostics ( basri et al . 1999 ; griffith and yelle 2000 ) .
a useful index in this spectral region is the i@xmath9-z@xmath9 ( sloan ab ; fukugita et al .
1996 ) color , where i@xmath9 peaks in the region from @xmath00.7 to @xmath00.8 and z@xmath9 extends from approximately @xmath00.8 to @xmath01.0 .
the fact that the behavior of the spectra depicted in figs .
[ fig:1 ] and [ fig:2 ] is not completely monotonic with the infrared - determined spectral subtypes indicates to us that the t dwarf spectral subtypes are not determined solely by .
gravity and , perhaps , metallicity also play a role .
in particular , the 2mass-0559 ( t5 ) ( green curve in fig . [ fig:1 ] ) spectrum shortward of 0.9 is redder " than that of sdss-1624 ( t6 ) ( blue curve in fig .
[ fig:1 ] ) , despite the former s earlier spectral subtype . as we argue in [ comparisons ] , this implies that the surface gravity of sdss-1624 ( t6 ) is lower than that of 2mass-0559 ( t5 ) .
in addition , even though the spectral subtype of sdss-1021 is t3 , while that of sdss-1254 is t2 , the relative flux level of sdss-1021 is generally above ( at both 0.73 and 0.83 ) that of sdss-1254 . since in [ theory ] and [ system ] we demonstrate in the likely range of sdss-1021 and sdss-1254 that decreases in and increases in gravity increase the redness of the spectrum shortward of 0.9 , we conclude that the gravity ( and presumably the mass ) of sdss-1021 is lower than that of sdss-1254 . though is the major determinant , a sufficiently large gravity difference can reverse the apparent dependence of subtype on changes in . the optical data depicted in figs .
[ fig:1 ] and [ fig:2 ] are state - of - the - art , but , in particular for the later t dwarfs and shortward of 0.8 , they are of fairly low resolution . nevertheless , these spectra can help guide us in this exploration of the dependence of optical spectra on t dwarf physical properties and in discerning the range of gravities and represented by the objects in table 1 .
the near absence of parallaxes is an impediment to detailed fits , as is the noise in the observed spectra at shorter optical wavelengths .
however , the optical colors and spectral indices ( [ define ] and [ comparisons ] ) that can be calculated for the growing list of t dwarfs with measured spectra shortward of 1.0 still provide useful physical diagnostics with which our suite of theoretical models can be used to determine the possible range of their , gravities , and metallicities .
as a prelude to the discussions in [ modelfits][comparisons ] , we present a collection of new theoretical spectra that span the t dwarf regime .
figure [ fig:3 ] portrays the absolute flux ( @xmath10 ) in millijanskys versus wavelength from 0.4 to 1.5 for solar - metallicity models for the two gravities and the full range of ( see [ uncertain ] ) . the higher curves in fig .
[ fig:3 ] are for models with the higher .
prominent are the na d and k i resonance doublets at @xmath05890 and @xmath07700 , the water bands around 0.93 , 1.15 , and 1.4 , the cs i lines at 8523 and 8946 , the li i line at 6708 , the rb i lines at 7802 and 7949 , and ( for the hottest models ) the tio and vo features near @xmath00.45 and [email protected] . at 1.2432/1.2522
the k i doublet is seen for at 700 k and above .
it was assumed in these models that clouds / grains / dust that may form at depth nevertheless have no effect on an object s t / p profile , nor on its spectrum . as we argue in [ clouds ] , this may not be true for the earliest t dwarfs . in determining molecular abundances , unless otherwise stated we employed the rainout prescription for the refractory silicates described in burrows and sharp ( 1999 ) . as explained in that reference and in burrows , marley , and sharp ( 2000 )
, rainout depletes the atmospheres of the refractory elements ca , al , mg , fe , and si .
this affects the abundance profiles of not only the nascent alkali metals , but tio , vo , feh , and crh .
in particular , rainout suppresses the formation of alkali feldspars and enables atomic na and k to survive to lower temperatures and pressures , at which point they form na@xmath1s(c ) and kcl(c ) ( lodders 1999 ) . as a consequence , their influence on the emergent spectrum in the optical " between 0.5 and 1.1 is enhanced .
rainout also restricts the range of lower temperatures and pressures where tio , vo , fe(l , c ) , crh , and feh can be found .
however , since the crh opacities are soon to be updated significantly , feh and crh bands were not incorporated into this model set and we defer a discussion of their role to a later work ( see also [ uncertain ] ) . the normalized observed spectra shortward of 1.0 were presented in figs .
[ fig:1 ] and [ fig:2 ] .
the corresponding normalized theoretical spectra at solar - metallicity and for gravities of 10@xmath11 cm s@xmath12 and 10@xmath13 cm s@xmath12 are displayed in fig .
[ normal ] .
a comparison of figs .
[ fig:1 ] , [ fig:2 ] , and [ normal ] is illuminating and reveals that the theoretical family nicely spans the observations .
furthermore , the general overall spectral shapes , particularly from 0.8 to 0.9 , are reproduced .
we compare these model spectra with data in [ modelfits]-[system ] .
however , we first discuss some of the major uncertainties in the models and a variety of physical diagnostics that are useful in interpreting measurements .
to construct new atmosphere / spectral models of brown dwarfs , we employ the complete linearization / accelerated-@xmath14-iteration method of hubeny and lanz ( 1995 ) , molecular and atomic opacities as described in burrows et al .
( 2001 ) , and the equation of state of saumon , chabrier , and van horn ( 1995 ) .
we explore both rainout , using the prescription of burrows and sharp ( 1999 ) ( see also lodders 1999 ) , and non - rainout composition assumptions , though when not explicitly stated rainout has been incorporated .
the t / p profiles at depth obtained using the lodders ( 1999 ) and the burrows and sharp ( 1999 ) rainout prescriptions are the same ( all else being equal ) to better than 2% ( m. marley , private communication ) . the depletion of refractories into silicate clouds is an intrinsically non - equilibrium and dynamical process .
hence , any prescription that purports to address this can do so only crudely .
fortunately , it is the clouds at depth into which settle / rain the refractory elements that would have been in the upper atmospheres that are the most problematically handled ; the upper atmospheric depletions of mg , si , fe , al , and ca themselves and the resultant metal - depleted molecular compositions are better handled , yielding compositions that are probably good to better than 10% .
since our modeling is focussed on t dwarfs in which clouds play no central role ( however , see [ clouds ] ) , our rainout prescriptions are as good as current practice allows .
however , at the higher in the l dwarf regime , the actual physics of clouds ( and their particle sizes , spatial extent , optical properties , and degree of patchiness ) assumes a greater importance .
furthermore , the possible effects of non - equilibrium chemistry and rapid transport of spectroscopically - active species may also be important ( lodders 1999 ; griffith and yelle 1999,2000 ; saumon et al .
2000 ) . in particular
, though co and nh@xmath15 have no strong bands short of 1.0 micron and play only a very minor role in setting the t / p profile of a t dwarf , the suggestion that their abundances may be out of equilibrium due to dynamic convective transport deserves further attention ( saumon et al .
2000 ; noll , geballe , and marley 1997 ) .
quite obviously , no current prescription for cloud opacities or for dynamic transport and their effects on emergent spectra can be beyond reproach or improvement . in this paper ,
given our focus on t dwarfs we ignore both .
our models are produced using the opacity sampling technique to arrive at converged and consistent temperature / pressure profiles , from which higher - resolution spectra are then calculated for a given spectral interval . for the t / p profile calculation ,
no more than 2000 wavelengths are needed , which range from 0.4 to 300 .
each model was converged in temperature and flux to better than 0.1% in typically a few minutes on a standard workstation .
the speed of our code is a direct consequence of the complete linearization approach .
marley et al . ( 2001 ) and burrows et al . (
1997 ) use the k - coefficient method with the equivalent of @xmath0800 wavelengths .
similar to the odf method , their implementation avoids its major pitfall by combining the abundance - weighted opacities of all the relevant molecules and atoms before creating the distribution function that is at the heart of the odf approach .
we use a mixing - length prescription to handle convection , with a mixing - length parameter of 1.0 .
marley et al . ( 2001 ) flatten the entropy gradient in convective regions .
saumon et al . ( 2000 ) use the t / p profiles generated by marley , as in marley et al .
allard et al . ( 2001 ) use mixing - length theory , with a mixing - length parameter of 1.0 .
at low , the actual mixing length parameter is irrelevant ; changing the mixing - length parameter from 0.5 to 2.0 changes the t / p profile by no more than @xmath020 k. however , above @xmath01500 k , the actual value of this parameter is an issue of modest interest .
nevertheless , the theoretical t / p profiles in grainless atmospheres below = 1500 k are similar , differing by no more than 10% ( usually less ) at atmospheric levels from 800 k to 2000 k ( cf . this work ; marley , unpublished ; allard et al .
2001 ) . for
the spectra we display here , 5000 wavelengths , logarithmically spaced from 0.4 to 1.5 , were used and then the spectrum was boxcar - smoothed to an effective resolution , r ( @xmath16 ) , of 1000 at each wavelength .
models were generated with from 600 k to 1300 k , in steps of 100k ( plus at a few extra , when needed ) , at two gravities ( @xmath17 = @xmath18 cm s@xmath12 and @xmath19 cm s@xmath12 ) , and for three metallicities ( z ) ( @xmath20solar , solar , and @xmath21solar ) .
the anders and grevesse ( 1989 ) abundance data were used to represent / define the solar pattern .
this -gravity - z parameter set was chosen to allow us to span the model space in which it is reasonable to assume the newly - discovered t dwarfs reside , as well as to gauge the systematic behavior of the observables with , gravity , and metallicity .
the brown dwarf radii ( @xmath22 ) assumed were derived from the analytic formula found in burrows et al .
( 2001 ) and marley et al .
( 1996 ) : @xmath23 where @xmath17 is in cm s@xmath12 and is in kelvin .
this procedure is not perfectly consistent , since the evolutionary models upon which this equation for the radius is based incorporate slightly different atmosphere models as boundary conditions . nevertheless , the error in the radius is small ( @xmath2410% ) and the derived colors and spectral indices are independent of radius .
the major differences between theoretical models arise from the different opacity databases employed . for substellar dwarfs ,
the major gas - phase opacities are due to h@xmath1o , h@xmath1 , ch@xmath25 , co , nh@xmath15 , the neutral alkali metals , tio , vo , feh , and crh .
a subsidiary role is played by ph@xmath15 , mgh , caoh , cah , sio , and h@xmath1s .
importantly , the _ equilibrium _ abundances of the dominant molecules ( in particular , h@xmath1o , h@xmath1 , ch@xmath25 , co , n@xmath1 , and nh@xmath15 ) are easily calculated ( burrows and sharp 1999 ) .
most researchers use the same h@xmath1 ( borysow and frommhold 1990 ; zheng and borysow 1995 ; borysow , jrgensen , and zheng 1997 ) and h@xmath1o ( partridge and schwenke 1997 ) opacities .
we calculate our tio and vo opacities using plez ( 1998,1999 ) and jrgensen ( 1997 ) , including the effects of isotope shifts , and believe them to be state - of - the - art . as stated in
[ theory ] , we omit feh and crh from this t dwarf model set .
their omission affects predominantly l dwarfs in the spectral regions around 0.86 and 0.99 ( feh , the wing - ford band ) , but does not change to any significant degree t - dwarf t / p profiles , nor the vast majority of the spectrum where feh / crh features do not contribute . note
that , as do most researchers in this field ( saumon et al .
2000 ; marley et al . 2001 ; allard et al .
2001 ) , we currently use the 52-million line subset of the 308-million line partridge and schwenke ( 1997 ) water database .
though recently questions have been raised concerning its accuracy in some wavelength regimes ( jones et al .
2002 ) , the partridge and schwenke compilation remains the best available .
methane opacities are an important wildcard , with the hot bands " and the red side of the @xmath4 photometric band being the most problematic ( burrows et al .
however , apart from a minor band near 0.89 , methane plays only a very modest role in optical " t dwarf spectra , though its effects in the infrared help to define the t dwarfs and the l@xmath6 t transition ( burgasser et al . 1999 ) .
the methane feature that has the most effect on a t / p profile is the strong @xmath26 band at 3.3 , but its opacity seems to be well in hand . not unexpectedly , most researchers use much the same methane opacity data ( cf .
burrows et al .
2001 ; saumon et al .
2000 ; marley et al .
2001 ) . nevertheless , it is in the different implementations of the extant line lists and opacity databases , in the different collisional broadening prescriptions , and in the inaccuracies in these various databases that one is likely to find the origin of most of the differences between the various theoretical models ( burrows et al .
1997 ; allard et al .
2001 ; tsuji et al . 1999 ; saumon et al 2000 ; geballe et al .
2001 ; marley et al .
2001 ) and between theory and data .
however , the alkali lines and their wing profiles are a special case to which we now turn .
as emphasized in burrows , marley , and sharp ( 2000 ) , it is not strictly correct to use lorentzian profiles for the na d ( @xmath05890 ) and k i ( @xmath07700 ) doublets in their far wings , as we do for this paper and as is the standard practice in most stellar atmospheres work .
in fact , the dominance of these doublets is a central feature of brown - dwarf / t - dwarf spectra and colors at wavelengths from 0.4 through 1.0 . while the line core behavior and oscillator strengths for these transitions are well in hand ( piskunov et al .
1995 ) , there remain ambiguities in the treatment of these lines at large detunings ( @xmath27 ) and the as - yet - unknown shapes of their wings will have a bearing on the viability of spectral fits .
for this study , we have used the default lorentzian out to a transition wavelength redward of the line cores ( 7700 for na d and 9800 for the k i doublet ) , after which we allow the strength to decay as a gaussian in wavelength with a width of 0.075@xmath28 .
the latter merely results in a smooth cutoff at a large detuning , but is otherwise arbitrary .
this prescription for the alkali wings is different from that suggested by burrows , marley , and sharp ( 2000 ) because in this paper we want to make only minimal alterations to the standard lorentzian until better estimates of the wing shapes are available .
the bms prescription has an ad hoc parameter @xmath29 whose value affects the far wing line shape and introduces an effective " cutoff , but is in fact unconstrained . a choice of @xmath30 for k and @xmath31 for na provides as good a fit to t dwarf spectra as the simpler prescription we employ here , given the noise in the optical data ( liebert et al .
however , the use of a bms @xmath29 parameter implies a certain functional form that is not motivated by the proper physical chemistry , only by the need for a cutoff ( [ rainout ] ) . though the general overall spectral shapes seen in fig .
[ normal ] , particularly from 0.8 to 0.9 , are reproduced ( liebert et al . 2000 ) with either the bms or our modified lorentzian prescriptions , there are two features that are problematic in the comparision of figs . 1 , 2 , and 4 .
one is that the theoretical peak between the k i and na d troughs is shifted by @xmath00.02 to shorter wavelengths than the observed average . this could easily be a consequence of the all - too - simple alkali line profile algorithms used .
the other is that for the earlier t dwarfs the k i trough at 7700 is a bit narrower than derived using either the bms or the current theory .
both prescriptions yield equally discrepant spectral shapes , though the fractional deviations in the flux densities or equivalent widths " derived using either prescription are at most a few tens of percent .
it should be noted that the prescriptions for the alkali line shapes used in the literature have included a pure lorentzian with a sharp cutoff at an arbitrary wavelength at a large detuning ( allard et al .
2001 ) , a pure lorentzian with a sharp cutoff at an arbitrary wavelength at a small detuning ( tsuji et al .
1999 ) , and the burrows , marley , and sharp ( 2000 ) prescription ( marley et al . 2001 ; saumon et al .
2000 ; geballe et al .
frequently , however , the alkali - line - shape prescription employed in a paper is unexplained and sometimes the na / k alkali metals have been omitted ( griffith , yelle , and marley 1998 ) .
the data shortward of 0.8 are indeed noisy , so this might be a major factor in any discrepancies between our theoretical spectra and observation , but as the comparison of figs .
[ fig:1 ] , [ fig:2 ] , and [ normal ] nevetheless demonstrates , the remaining @xmath010 - 50% ambiguity in the flux densities is much smaller than the factors of @xmath0210 deviations previously seen in the literature shortward of 0.8 ( cf .
tsuji et al . 1999 ) .
hence , our current approach to the na / k alkali - line shapes is as reasonable as any , but the reader should bear in mind the need for significant further improvement .
the dependence of the theoretical spectra on , gravity , and metallicity is a major focus of this paper .
effects such as the general reddening in the optical with decreasing ( at a given gravity ) , which reverses at lower effective temperatures ( particularly noticeable in fig .
[ normal ] around 0.75 ) , will be discussed in [ system ] . to facilitate that discussion
, we present here representative model temperature / pressure profiles in fig .
[ profiles ] and representative plots of the t@xmath8 temperature versus wavelength in fig .
[ bright ] .
t@xmath8 is defined in this paper as the temperature level in the brown dwarf atmosphere at which the total optical depth is @xmath32 . roughly , it is the temperature of the layer to which one is probing when measuring a spectrum at the corresponding wavelength and is a measure of the depth to which an observed spectrum is allowing us to peer ( see also saumon et al .
2000 , their figure 6 ) .
[ profiles ] can be used to gauge the dependence on , gravity , and metallicity of the optical depth above a given temperature level in an atmosphere . as fig .
[ profiles ] suggests , lower - metallicity models have higher pressures at a given temperature and higher - models have lower pressures at a given temperature .
moreover , as would be expected from hydrostatic equilibrium , higher - gravity models have higher pressures at a given temperature .
t@xmath8 is the decoupling " temperature at a given wavelength , or the wavelength - dependent temperature level to which one probes an atmosphere by measuring its spectrum .
[ bright ] indicates that shortward of 1.0 , higher gravities result in slightly lower t@xmath8s and higher always result in higher t@xmath8s .
in addition , fig .
[ bright ] shows that in the @xmath2 band ( @xmath01.05 ) , one is probing to @xmath01500 - 1600 k for the 1100 k models and that in the @xmath3 band ( @xmath01.25 ) , one is probing to @xmath01400 - 1500 k , for the same models .
note that below the na d line at @xmath05890 , t@xmath8 for the sample models is rising to between 1200 and 1600 k , while at the centers of the strong na d and k i resonance doublets , t@xmath8 is near 800 k.
color - color diagrams are traditional tools used to determine the physical properties of stars . in addition
, non - traditional , but diagnostic , spectral indices ( ratios of flux levels or flux averages at different wavelengths ) can perform the same function ( burgasser et al . 2000b ; burgasser 2001b ; geballe et al . 2002 ; tokunaga and kobayashi 1999 ) .
such indices are very useful for encapsulating and describing trends in , gravity , and metallicity and in isolating one class of objects from another .
this is especially true when spectral data are noisy , when subtle changes in plotted spectra are difficult to discern by eye , or when distances are not known ( as is currently the case for the majority of t dwarfs ) .
hence , we have calculated a set of spectral indices and colors that highlight various features of the family of t dwarf spectra , both theoretical and observed , and use them to derive physical facts about the t dwarfs listed in table 1 , as well as about the spectral class as a whole .
the indices we have defined are : @xmath33 @xmath34 @xmath35 and @xmath36 note that @xmath37 ( not @xmath38 ) is used , that a range of wavelengths implies an average in the indicated wavelength range , and that the indices do not have a prefactor of 2.5 . these indices were chosen to highlight important features in t dwarf spectra at shorter wavelengths .
both x97 and x98 were defined to avoid the water feature at @xmath00.93 .
x97 is a measure of the relative flux in the spectral bump between the na d and k i ( 7700 ) lines and the region around 0.9 and captures the effect of rainout .
x98 is a measure of the slope between 0.8 and 0.9 and , hence , is a measure of the shape of the red wing of the k i line at 7700 .
x23 is a measure of the depth of the water feature near 0.93 and reflects the influences of h@xmath1o abundance , gravity , and the residual effect of silicate clouds .
x126.105 is a measure of the ratio of the fluxes at 1.26 and 1.05 , near the traditional @xmath3 and @xmath2 bands at the prominent peaks in t dwarf spectra ( see fig .
[ fig:3 ] ) .
our indices short of 1.0 are similar to those employed by burgasser et al .
( 2001b ) and geballe et al .
( 2002 ) , but are a bit better tuned to capture physical effects , as opposed to observational trends . however , almost any of the optical " indices recently defined by those measuring l and t dwarfs could have sufficed .
( tokunaga and kobayashi ( 1999 ) do not define indices short of 1.0 , the spectral region on which we have focussed in this paper . )
we also use the the colors i@xmath9 - z@xmath9 and @xmath39 , where the former is in the sloan ab system and the latter is in the bessell ( or cousins ) system ( bessell and brett 1988 ) .
importantly , the choice of system or index for the purpose of deriving trends is arbitrary .
we chose bessell for the @xmath40 color to connect with the traditional color system in light of the continuing confusion in the use of ukirt , ks , and 2mass filters ; there is as yet no uniformity from telescope to telescope in the use of near - infrared filter sets .
hence , our colors are just like spectral indices ( e.g. , x97 @xmath41 ) and those calculated and discussed here should be viewed as such .
our colors are derived from spectra , not photometry , and they are used to discover trends , not to compare with standard - star - calibrated values .
nevertheless , our i@xmath9 - z@xmath9 and @xmath40 ( bessell ) colors are as good as any , given the differences between the underlying spectra of calibration standards and t dwarfs and given the accuracy of the flux calibration for the data ( burgasser et al .
note that @xmath40 has been included in this paper , despite our emphasis on the optical , to make at least one tie - in with the spectral range now being used to define the t dwarf spectral subtypes . in table 2
, we provide these indices and colors for from 600 k to 1300 k , gravities of 10@xmath11 cm s@xmath12 and 10@xmath13 cm s@xmath12 , and metallicities of 0.3@xmath42solar , solar , and 2.0@xmath42solar , where solar is defined by anders and grevesse ( 1989 ) .
we include no - rainout " models at 900 k and the two gravities and the observed t dwarfs listed in table 1 .
table 2 summarizes all the basic trends ( on which we focus in [ comparisons ] ) .
we use these indices in the following sections to facilitate the diagnosis of the observed spectral trends .
that our prescription for the alkali line shapes has some merit is demonstrated in fig . [ fittwo ] , where solar - metallicity spectral fits for the two t dwarfs , gliese 229b and gliese 570d , for which there are published parallaxes ( perryman et al . 1997 ; 5.8 and 5.9 parsecs , respectively )
are presented . for gliese 570d ( as for the other t dwarfs in this paper )
, ours is the first paper to compare data with theory below 0.8 .
( for gliese 229b , there are no spectral data as yet below 0.85 , only @xmath22 and @xmath43 band photometry . )
the spectra portrayed in fig .
[ fittwo ] are given in absolute flux units .
the shape of the spectrum from 1.0 to 1.1 , the slope of the spectra around 0.8 to 0.9 , the gliese 229b wfpc2 @xmath22 band measurement ( golimowski et al 1998 ) , the flux level at @xmath00.7 for gliese 570d , and the i@xmath9-z@xmath9 color for gliese 570d are all reproduced . as fig .
[ fittwo ] indicates , our theoretical fluxes for both gliese 229b and gliese 570d ( which span more than three orders of magnitude ) are good to 1030% , though the noise below 0.8 is problematic .
geballe et al .
( 2001 ) do not study gliese 570d shortward of 0.8 .
however , they obtain comparably good fits above 0.8 , though we reproduce the shapes and magnitudes in the @xmath2 and @xmath3 bands more closely .
focussed on gliese 229b and nh@xmath15 , saumon et al .
( 2000 ) do not discuss spectra shortward of 1.0 .
marley et al . (
2001 ) do not compare theoretical spectra with observational data . in their gliese 229b campaign , griffith , marley , and yelle ( 1998 ) and griffith and yelle ( 1999 )
do not include the na and k alkali lines in their opacity database .
instead , they posit the existence at altitude of a cloud of red grains whose imaginary index of refraction has a more extreme dependence on wavelength than that of titan tholins , red phorphorus , or polyacetylenes ( noy , podolak , and bar - nun 1981 ; khare and sagan 1984 ) . in their important gliese
229b study , tsuji et al .
( 1999 ) conclude that a combination of silicate dust with the k i doublet at 7700 can explain its spectrum shortward of 1.0 . however , as is still the standard practice in stellar atmospheres work , tsuji et al .
truncated the alkali lines at very small detunings , which necessitated the introduction of another component to explain the spectrum . as did bms
, we find that including the wings of the potassium doublet obviates the need for another component to explain the sharp declivity shortward of 1.0 in t dwarf spectra in general ( fig .
[ fig:1 ] and [ fig:2 ] ) and in gliese 229b s spectrum in particular . as fig .
[ fittwo ] indicates , the relative flux ratios in the @xmath2 and @xmath3 bands are well - modeled . in contrast , the best - fit model of gliese 229b by tsuji et al .
( 1999 ) in the @xmath2 and @xmath3 bands is off by a factor of 23 , reflecting the need to cutoff the alkali metal wings as discussed in [ alkali ] and [ rainout ] .
importantly , the corresponding no - rainout spectra below 0.8 ( not shown ) ( allard et al .
2001 ) are generically off by a factor greater than 5 ( [ rainout ] ) .
slightly more precise fits can in fact be achieved by adjustments in , metallicity , and gravity . however , our purpose here is not to obtain the ultimate fit , but to verify our general approach and to extract the essential conclusions about the class of t dwarfs as a whole using spectra at short wavelengths .
nevertheless , the depths of the water features near 0.93 , 1.15 and 1.4 ( not shown ) are not fit well .
leakage in the spectrometer of light from adjacent peaks into the troughs might explain the effect at 1.4 , and perhaps a fraction of the effect at 1.15 , but these are major discrepancies that may point to a residuum of dust in the atmosphere ( tsuji et al 1999 ; saumon et al .
2000 ) or to problems with the water opacity database used ( partridge and schwenke 1997 ) .
the models of all groups fail to explain these water troughs ( saumon et al .
2000 ; marley et al .
2001 ; allard et al .
lowering the metallicity enough , particularly for a majority of the t dwarfs in table 1 which all manifest this problem , would not seem to be a viable option generically ( allard et al .
2001 ; however , see griffith and yelle 1999 ) .
furthermore , a low - enough metallicity that would solve the problem at 0.93 would also shift the @xmath40 magnitudes away from the observed values for the t dwarfs ( burgasser et al .
2001a ) by as much as @xmath44 magnitudes at 900 k. ( however , a lower metallicity may be implicated for 2mass-0937 ; see [ comparisons ] . )
a for gliese 570d of @xmath0750 k at a low gravity of 10@xmath11 cm s@xmath12 ( mass @xmath030 ) is similar to that derived in geballe et al .
( 2001 ) , but at their low end .
equally good solar - metallicity fits to gliese 570d can be found from @xmath0750 k to @xmath0840 k , where the corresponding gravities range from @xmath18 cm s@xmath12 to @xmath0@xmath45 cm s@xmath12 .
our range for gliese 570d is a bit more conservative than that of geballe et al .
( 2001 ) due to the fact that we have not imposed an age constraint .
however , as demonstrated in fig .
[ fittwo ] for gliese 229b , at the current level of observational precision it is only a trajectory in -gravity space ( given roughly by /@xmath46 = const . ) that can be constrained .
gliese 229b could have a near 780 k and a gravity near 10@xmath47 cm s@xmath12 ( mass @xmath4815 ) .
what is shown in fig .
[ fittwo ] are just representative fits to the gliese 229b and gliese 570d data .
solar - metallicity models of gliese 229b that fit these spectra include those with from 780 k to 950 k ( higher would imply a higher gravity than is realistic for brown dwarfs ) .
very subsolar - metallicity models for gliese 229b would compromise the acceptable fits in @xmath40 . without parallaxes , we prefer not to provide spectral fits for the other t dwarfs in table 1 .
it is not that fits ca nt be obtained . that they can should be clear from figs .
[ fig:3 ] and [ normal ] and from table 2 .
rather , the problem is that too many models can be found to fit , and and gravity are only loosely constrained
. nevertheless , our models and the indices discussed in [ comparisons ] yield for all but the early t dwarfs a range of 900@xmath49150 k. with parallaxes , we should be able to tighten this constraint and determine the /@xmath17 trajectory ( and core entropy ; burrows , marley , and sharp 2000 ) for each observed t dwarf .
however , at this stage , we do nt want to claim more than is prudent .
there are a few conclusions of a qualitative nature , in particular concerning the truncation of the resonance lines of na and k and the seeming necessity of rainout , that deserve special mention .
figure [ figcompare ] demonstrates many of these with a collection of representative theoretical model spectra at = 900 k and solar metallicity from 0.4 to 1.5 .
all the spectral models have t / p profiles that are consistently derived for the opacities and abundance prescriptions used .
the solid red line depicts the [ 900 k/@xmath18 cm s@xmath12 ] model and the solid blue line the [ 900 k/@xmath19 cm s@xmath12 ] model .
the black line portrays a model with @xmath17 = @xmath19 cm s@xmath12 , but uses lorentzian line profiles without cutoffs for the na d and k i ( 7700 ) lines .
in addition , this spectrum does not incorporate the rainout of ti and v species , and , hence , leaves them suspended in the atmosphere in strict chemical equilibrium .
the dashed blue line is a [ 900 k/@xmath19 cm s@xmath12 ] model , but one that does not incorporate a prescription for the rainout of the silicates and the consequent survival of na and k to lower temperatures and pressures in the atmosphere .
the spectrum of 2mass-0559 ( t5 ) is included on fig .
[ figcompare ] in green and is the absolute flux spectrum under the assumption that its distance is 10 parsecs . that unmodified lorentzians produce spectra that do not fit is demonstrated in fig .
[ figcompare ] by the relative positions of the black line ( lorentzian model ) and the solid blue line .
the latter spectrum in the @xmath2 ( @xmath01.05 ) and @xmath3 ( @xmath01.25 ) bands is much closer to the data shown in fig .
[ fittwo ] and to those obtained for the objects in table 1 by burgasser ( 2001b ) ; the truncated models do not decapitate the @xmath2 and @xmath3 bands by as much as an order of magnitude in the way that the untruncated lorentzian does . as mentioned in
[ modelfits ] , the absence of such a cutoff is the reason for the discrepancy by a factor of 2 to 3 in the tsuji , ohnaka , and aoki ( 1999 ) model for gliese 229b in the @xmath2 and @xmath3 bands .
the difference at = 900 k and @xmath17 = @xmath19 cm s@xmath12 between our standard model result and that with untruncated lorentzians is @xmath01.2 magnitudes in i@xmath9-z@xmath9 ; the x98 and x97 indices are similarly incorrect ( table 2 ) .
note that , since na is @xmath020 times more abundant than k , it is na that would account for most of this anomalous suppression beyond 1.0 , despite the greater spectral distance of the na d line .
one concludes from fig .
[ figcompare ] that the true line profiles must be truncated and are not strictly lorentzians . as is clear from fig .
[ figcompare ] , the no - rainout ( dashed blue ) and rainout ( solid blue ) spectra differ by 0.5 - 1.0 dex from one another , particularly at 0.7 between the dominating na d and k i resonance doublets .
a comparison of these curves with the new t dwarf spectra provided in figs .
[ fig:1 ] and [ fig:2 ] leads to the strong conclusion that the no - rainout models do not fit the new optical data .
this is clear confirmation of the bms prediction ( see their figures 2 and 3 ) for the entire " family of t dwarfs , not just for sdss-1624 ( liebert et al .
2000 ) , and is consistent with the marley et al .
( 2001 ) figure 3 .
geballe et al .
( 2001 ) came to a similar conclusion for gliese 570d vis vis rainout , but did not have data nor theory shortward of 0.8 .
given the ambiguities in the alkali line profiles near 1.0 , spectra from 0.8 to 1.0 can not be used to draw clear conclusions about the occurrence of rainout ; it is only with the new data at shorter wavelengths ( @xmath51 ) that one can distinguish unambiguously between the effects of rainout and of the unknown alkali line profile shapes , given the well - known oscillator strengths .
as table 2 indicates , i@xmath9 - z@xmath9 for the no - rainout models is as much as 1.0 - 1.5 magnitudes discrepant .
the indices x98 and x97 tell a similar story . without rainout , the redness " of the optical spectra and the large flux contrast between the @xmath2 peak and the bump at 0.7 - 0.75 can not be reproduced .
rainout leads to an enhancement in the abundance of atomic na and k at lower temperatures ( and lower pressures ) in the atmosphere and to greater absorption shortward of 0.9 . a comparison on fig .
[ figcompare ] of the spectrum of 2mass-0559 ( t5 ) with the no - rainout / rainout theoretical curves serves to emphasize this point ( 2mass-0559 is here merely representative of the t dwarfs listed in table 1 ) .
hence , we have in the new t dwarf spectra at short wavelengths evidence for the influence of cloud formation at depth and element depletion ( si , mg , al ) at altitude ( burrows and sharp 1999 ; burrows , marley , and sharp 2000 ; lodders 1999 ) , leading to the suppression of alkali feldspar production at higher temperatures near @xmath01400 k and the consequent persistence of atomic na and k to lower atmospheric temperatures . in particular , the fit displayed in fig .
[ fittwo ] for gliese 570 d and the approximate fit to the gliese 229b datum in the @xmath22 band do not seem possible with any no - rainout models . as a test , for the lorentzian model portrayed on fig .
[ figcompare ] we suppressed the effects of rainout on the tio and vo abundances , thereby enhancing the abundances of those molecules at altitude .
as this = 900 k model demonstrates , around 0.45 ( near the @xmath52 band ) tio features would be visible , even for a model with such a low .
this is a consequence of the low gaseous opacities at the shortest wavelengths . the high t@xmath8s ( fig .
[ bright ] ) in this wavelength range say the same thing ; at short wavelengths one is probing deeply . if such features are not seen ( and this we suspect ) , it would be another indication of rainout .
moreover , cloud veiling could also be implicated .
hence , observations at wavelengths even shorter than the na d line(s ) can provide information concerning the deep atmosphere .
from the spectra presented in figs . [ fig:3 ] , [ normal ] , and [ figcompare ] , the representative t / p profiles and t@xmath8s shown in figs . [ profiles ] and [ bright ] , and the indices and colors tabulated in table 2 , we can determine the general theoretical trends for t dwarf spectra with , gravity , and metallicity .
just as fig .
[ profiles ] and all previous calculations indicate , as the effective temperature decreases for a given gravity the pressure in the atmosphere at a given temperature increases .
this implies that the column depth in gm cm@xmath12 to a given temperature level increases with decreasing ( in hydrostatic equilibrium , column depth = @xmath54 ) .
hence , for a ubiquitous species such as water , the @xmath55 surface that roughly determines the position of the photosphere " ( on average , or at a given wavelength ) is at progressively lower temperatures with decreasing ( see fig .
[ bright ] ) .
the result is that t@xmath8 in the water absorption troughs near 0.93 , and between the @xmath2 , @xmath3 , and @xmath5 bands ( for instance ) , as well as those at the emission peaks themselves , decreases .
however , since the near - ir is not on the rayleigh - jeans tail , these decreases in t@xmath8 lead to an increase in the flux contrast from peak to valley , as well as to an overall decrease in the fluxes .
hence , the water troughs deepen with decreasing and , as table 2 demonstrates , @xmath40 gets _ bluer _ with decreasing ( burrows et al . 1997 ; allard et al .
roughly , this may be what is seen for the t dwarfs in the near ir ( burgasser et al .
2001a ) and in figs .
[ fig:1 ] and [ fig:2 ] . at the higher ( in the ls and early to mid ts ) , the systematics with decreasing described above for water can also be seen in the spectral region shortward of 1.0 micron dominated by the alkali metal lines and wings .
in addition , increasing pressure with decreasing leads to greater pressure - broadening , enhancing the influence of the alkali metal wings .
both effects lead to a _ reddening _ of the spectrum below @xmath01.0 with decreasing which has been identified in previous theoretical explorations ( e.g. , bms ; marley et al .
2001 ; allard et al .
2001 ) . however , the na and k abundances are very small below a certain temperature ( 1050 k to 850 k , depending on element and pressure ) .
since there is such a top " to the atomic alkali region , as decreases further at low the region containing na and k atoms becomes more and more buried .
the result is a gradual diminution of the role of the na and k lines from 0.4 to 1.0 and a consequent bluing " of the spectrum below @xmath01.0 .
this effect below 1.0 was predicted by bms ; we now quantify it and in fig .
[ normal ] , which shows the incipient closing of the k i absorption trough at 7700 and the slight bluing of the optical spectrum at low .
the gravity- and metallicity - dependent reversal of the indices x97 , x98 , and i@xmath9-z@xmath9 at low is captured in table 2 .
recently , this prediction for the low behavior of t dwarfs , as seen in i@xmath9-z@xmath9 , has been verified by marley et al .
the turnaround with decreasing in the behavior of these spectral indices and of the i@xmath9-z@xmath9 color is even more clearly portrayed in figs .
[ fig:10 ] , [ fig:12 ] , and [ fig:13 ] ( discussed in [ comparisons ] ) . for the i@xmath9-z@xmath9 color , for this transition ranges from @xmath0700 k to @xmath0800 k. in this range , both i@xmath9-z@xmath9 and the overall spectral shape shortward of 1.0 micron are weak functions of . a t dwarf with a of @xmath241000 k and a gravity of 10@xmath13 cm s@xmath12 has a mass near 70 ( and an age @xmath4810@xmath56 years ) .
one with a gravity of 10@xmath11 cm s@xmath12 has a mass near 35 .
as we go down in gravity to 10@xmath47 cm s@xmath12 , we are exploring masses just above 15 . the observed t dwarfs
may possibly span this entire range .
figure [ profiles ] shows that as the gravity increases the pressure at a given temperature in the atmosphere increases as well .
what this figure does not show is that the column depth of material actually decreases with increasing gravity .
this means that for a given , the depths of the water troughs and the contrasts between the emission peaks and these troughs actually _ decrease _ with increasing gravity .
in particular , we discover that the relative depth of the water feature near 0.93 ( measured by x23 ) decreases with increasing gravity .
this is opposite to the effect of decreasing .
one consequence is that any indices constructed from the contrasts between peak and water trough are not a one - parameter family in alone .
hence , we find that the effect of decreasing can be compensated by an increase in gravity .
importantly , as table 2 demonstrates , increasing the gravity by 0.5 dex has the same effect on @xmath40 as decreasing by 100 - 200 k. hence , given the current method of t dwarf spectral subtyping in the near - ir ( burgasser et al . 2001a ; geballe et al .
2002 ) , effective temperature alone can not be construed to imply subtype . at a given , lowering the gravity can lead to a later subtype .
moreover , a later subtype does not necessarily imply a lower .
as table 2 indicates , @xmath40 gets bluer with increasing gravity .
this is predominantly a consequence of the pressure dependence of the cia opacity of h@xmath1 at 2.2 .
shortward of 1.0 micron , at higher increasing gravity reddens the i@xmath9-z@xmath9 color and increases ( steepens ) the x97 and x98 indices .
however , at lower an increase in gravity leads to a slight reversal and i@xmath9-z@xmath9 becomes bluer .
this newly - quantified behavior can be seen in fig .
[ fig:12 ] ( [ comparisons ] ) and table 2 and is a consequence of the same systematics in the atmospheric profiles ( fig . [ profiles ] ) discussed earlier that are responsible for the reversal in the dependence of the optical colors and slopes at lower .
figure [ figmetal ] portrays the metallicity dependence of t dwarf spectra from 0.4 to 1.5 , for a single representative of 900 k and a gravity of 10@xmath13 cm s@xmath12 .
metallicities at 0.3 , 1.0 , and 2.0 times solar are compared .
shortward of 0.9 , we find that the lower - metallicity model is redder and the higher - metallicity model is bluer than the solar model .
this newly - quantified effect may seem backwards , since one might have expected a lower metallicity to weaken the effects of the alkali metals .
however , as the metallicity is lowered , the abundances of the other metals ( such as oxygen in the form of water ) are also lowered , leading to a more transparent atmosphere . as a result
, one is peering more deeply into the atmosphere to higher pressures .
this is clearly apparent in the relative positions on fig .
[ profiles ] of the red ( low - metallicity ) lines . at higher pressures ,
the na d and k i resonance lines at @xmath05890 and @xmath07700 are stronger .
hence , i@xmath9-z@xmath9 reddens with decreasing metallicity .
this is quite similar to what happens in subdwarf m stars .
overall , the solar and 2.0@xmath42solar models are very similar shortward of 1.0 . unless the metallicity is significantly subsolar ( griffith and yelle 1999 ) , the metallicity dependence of t dwarf spectra in the optical is weak ( except perhaps shortward of 0.6 ) .
note that our lower - metallicity model has a more rounded @xmath2 band peak .
such a shape is not in evidence in the t dwarf data .
the differences in the near - ir at @xmath4 and @xmath5 are more pronounced , with lower - metallicity models being generically more blue in @xmath40 and @xmath57 .
for instance , from 600 to 1300 k , the differences between the solar - metallicity and the 0.3@xmath42solar - metallicity models vary from 0.1 to 0.5 magnitudes in @xmath40 ( cf . table 2 and fig .
[ fig:12 ] ) .
the higher pressures in the low - metallicity atmospheres increase the cia opacity and suppress the @xmath5 band flux .
we find that decreasing the metallicity by a factor of three has a greater effect on the @xmath5 band than increasing the gravity by the same factor .
this behavior with metallicity requires that if gliese 229b has a low metallicity ( griffith and yelle 1999 ) it must also have a low gravity .
high - gravity models for gliese 229b can not simultaneously be low - metallicity ( allard et al . 2001 ) .
it would be very useful for the determination of the physical parameters of the t dwarf gliese 229b to obtain its spectrum from @xmath00.6 to 0.8 and its x97 and i@xmath9-z@xmath9 indices ( consistently calibrated ) .
figures [ fig:10 ] , [ fig:11 ] , [ fig:12 ] , and [ fig:13 ] are color - color " diagrams that summarize the numerical data in table 2 .
the upper - case letters on the figures represent the positions in these index plots of the measured t dwarfs .
the letter that corresponds to a given t dwarf is indicated in table 2 and in the figure caption to fig .
[ fig:10 ] .
errors of @xmath480.1 dex in the x indices and @xmath00.25 in i@xmath9-z@xmath9 and @xmath40 ( bessell / cousins ) colors have been suppressed on these figures to avoid further clutter and to allow the basic collective gestalt to emerge .
the lines on the plots are families as a function of at given gravities and metallicities .
each point represents an from 600 k to 1300 k , in steps of 100 k. the basic success of the models is clear from the correspondence between the regions occupied by the observed t dwarfs and where the theory points reside . in principle , a comparison of the observed positions on these plots with those for the theoretical models would allow one to determine , gravity , and metallicity .
however , the errors in the indices and the overlaps of the model families make such a determination a bit less straightforward . in fact , such errors make it difficult to determine and gravity for most of the t dwarfs listed in table 1 to better than @xmath49150 k in and 0.25 dex in gravity .
fits to most of the measured spectra can be obtained , but without parallaxes and given the remaining errors in the data they tell us very little we can use .
hence , we prefer not to provide specific object - by - object and gravity estimates , unless parallaxes are available or the object is an outlier in its spectrum or indices .
nevertheless , from table 2 , fig .
[ normal ] , the new optical data , and our theoretical models we can infer that the of the listed t dwarfs , except for the three early ts , are 900@xmath49150 k. the corresponding gravity range could in principle span our model set from 10@xmath47 cm s@xmath12 to 10@xmath13 cm s@xmath12 .
note that values for the of the early t dwarfs are approximately bracketed by 1100 k and 1300 k ( see table 2 , figs .
[ fig:10 ] , [ fig:11 ] , [ fig:12 ] , and [ fig:13 ] ) .
parallaxes would enable a determination of the possible /@xmath17 line for each object ( as obtained in [ modelfits ] for gliese 229b ) and are eagerly awaited . from fig .
[ fig:10 ] , we see that our theory successfully provides the correct general overall x97/x98 slope and distinguishes the l dwarf region from the t dwarf region .
furthermore , it demonstrates that if the observed early t dwarfs ( green ) are not generically low - gravity objects ( dashed ) , that the t dwarf edge has an effective temperature near 1300 k. ( if these early t dwarfs were low - gravity objects , the t dwarf edge would be even lower in . )
this l / t boundary temperature is in contrast with the @xmath01600 k suggested by basri et al .
( 1999 ) . for higher gravity , sdss-1254 ( t2,d ) would have a near 1200 k. however , despite the inclusion of rainout , the models are still displaced downward by @xmath00.2 dex in x97 .
( note that the non - rainout models [ cf .
table 2 ] would be displaced downward in x97 by a further @xmath00.5 dex and to the left in x98 by @xmath00.2 dex ! ) the fact that 2mass-0937 ( t6p , i ) , gl 570d ( t8,n ) , and 2mass-0415 ( t8,o ) are outliers is consistent with their lower , higher gravity , or lower metallicity . in particular , 2mass-0937 , due to its high x97 ( and i@xmath9-z@xmath9 ; see fig .
[ fig:11 ] ) and to the fact that in the near - ir it was typed by burgasser et al .
( 2001a ) as an earlier subtype than the other two , seems clearly to have either a higher gravity or a lower metallicity than the average t dwarf observed to date ( burgasser et al .
the lower - metallicity solution is slightly favored when one looks at the other index figures and makes a consensus judgment . furthermore , the @xmath2-band spectrum of 2mass-0937 ( burgasser et al .
2001a ) is a bit more rounded than average , reminiscent of the behavior of the low - metallicity model depicted in fig .
[ figmetal ] .
we do not draw any conclusions from the lonely position of 2mass-1237 ( k ) on this figure and on figs .
[ fig:11 ] and [ fig:12 ] due to the noisiness of its spectrum between the na d line and the k i line at 7700 .
better data are clearly needed .
figure [ fig:11 ] demonstrates once again that our basic theoretical treatment shortward of 1.0 micron is not far wrong .
we also see that the of modest- to high - gravity models of the early t dwarfs must be @xmath581300 k and that 2mass-0937 ( i ) is best fit with a low - metallicity model , though not definitively so . however , the proximity of the model lines , their overlap , and the errors in the data together make it difficult to draw detailed conclusions concerning the central " t dwarfs .
figure [ fig:12 ] is a color - color diagram that demonstrates the non - monotonic behavior of i@xmath9-z@xmath9 with ( [ system ] ) and the outlier status of 2mass-0937 .
it also makes the case that the early t dwarfs are not fit well in the near infrared by this grainless model set .
no doubt , part of the problem lies with the incomplete methane opacity database in the @xmath4 and @xmath5 bands , notably the absence of the hot " bands of methane .
however , this can not account for the full discrepancy .
the theoretical models are too blue in @xmath40 and the best explanation for this is the influence of residual dust / grains / clouds even in the early t dwarfs sdss-0837 ( t1,c ) , sdss-1254 ( t2,d ) , and sdss-1021 ( t3,e ) .
the possible presence of grains is also suggested in fig .
[ fig:13 ] , which portrays both the behavior of x98 with x23 ( h@xmath1o jump ) and the non - monotonic behavior with discussed previously .
figure [ fig:13 ] shows that the depth of the water feature near 0.93 is not well reproduced by the models , unless many of the observed t dwarfs are of low metallicity ( griffith and yelle 1999,2000 ) .
we surmise that extra opacity due to dust may be shallowing the water trough near 0.93 by 0.1 to 0.15 dex .
such dust may also be contributing to the general reddening shortward of 1.0 micron ( tsuji , ohnaka , and aoki 1999 ) , still dominated by the wings of the k i ( 7700 ) doublet .
the index figures and table 2 allow one to draw some qualitative conclusions concerning specific t dwarfs . as indicated in fig .
[ fittwo ] , gliese 229b can be fit by a low - g / low - or a high - g / high - model .
however , its value of x23 ( j on fig .
[ fig:13 ] , one of the highest for the observed t dwarfs ) implies that the low - gravity solution is preferred .
it is probable from their values of i@xmath9-z@xmath9 , x97 , and x23 that sdss-1624 s gravity is lower than that of 2mass-0559 .
furthermore , a direct comparision of sdss-1624 s spectrum with that of gliese 229b leads one to conclude that its gravity is greater than that of gliese 229b .
in addition , a comparison of the i@xmath9-z@xmath9 , x98 , and x97 indices for sdss-1021 ( t3 ) and sdss-1254 ( t2 ) leads one to conclude that sdss-1021 has the lower gravity .
this is also the conclusion arrived at in
[ data ] directly from the spectra .
sdss-1021 could also have a higher .
the discussion in [ rainout ] points to a quite interesting fact . at near the early edge of the t dwarf family ( @xmath01300 k ) , t@xmath8 in the @xmath2 and @xmath3 bands reaches 1700 - 1800 k for atmospheres without the opacity effects of clouds .
this suggests that the silicate / iron clouds expected to reside above 1500 - 1700 k might significantly affect the shape and strength of the @xmath3 and @xmath2 emission peaks , even at such low .
a very similar conclusion was reached by marley et al .
this is not unreasonable , since the l to t transition is thought to coincide with the settling to depth of the clouds that dominate the l dwarf spectral type ( ackerman and marley 2001 ; burrows et al .
the spectra of the early t dwarfs should show the residual influence of such clouds .
indeed , as we argue in [ comparisons ] , the ( i@xmath9-z@xmath9)/(@xmath40 ) color - color diagram suggests that clouds are required to fit the early t dwarfs sdss-0837 ( t1 ) , sdss-1254 ( t2 ) , and sdss-1021 ( t3 ) .
as the t@xmath8 temperature argument suggests , it is in the @xmath2 and @xmath3 bands where the effects of clouds are primarily expected .
our spectral models at @xmath01300 k ( fig . [ fig:3 ] ) manifest tio and vo features in these bands because we have not included the effects of clouds .
clouds should not only mute the effects of tio and vo on the @xmath2 and @xmath3 bands of the = 1200/1300 k models in fig .
[ fig:3 ] , but should also help reduce the contrast between the emission peaks at @xmath2 , @xmath3 , and @xmath5 and the water troughs at 1.15 and 1.4 ( jones and tsuji 1997 ) . furthermore , because the subordinate k i absorptions at 1.2432/1.2522 are from a thermally - excited state , they are formed only at high atmospheric temperatures .
coincidently , they reside in the middle of the @xmath3 band where one is probing higher temperatures , as indicated by the higher t@xmath8s there ( fig . [ bright ] ) . however , the strengths we derive for these subordinate lines ( see fig .
[ fig:3 ] ) are generally greater than observed ( mclean et al .
2000 ; saumon et al .
this too may point to a mitigating role of veiling clouds / dust / grains . hence
, there are numerous spectral regions that speak to the potential influence of clouds on t dwarf spectra .
however , there remains a puzzle that may not be so easily explained by dust .
the classic li feature at 6708 is included in our model set and can be seen on the curves in fig .
[ fig:3 ] . observations ( kirkpatrick et al . 1999 ) suggest that the strength of this feature peaks in the mid - l spectral range and then drops monotonically .
currently , there is little indication of this line in the observed late l dwarfs , nor in the t dwarfs .
however , in both our rainout and no - rainout ( unpublished ) models the li line survives to below = 600 k. allard et al .
( 2001 ) also obtain this feature at low and pavlenko ( 2001 ) has an extensive theoretical discussion concerning its formation in late l dwarfs .
since lithium burning in objects that currently have t dwarf occurs only for a narrow range of masses and gravities ( @xmath59 cm s@xmath12 ) , nuclear burning seems to be excluded as a generic explanation .
chemistry would be the natural culprit , but both the burrows and sharp ( 1999 ) and lodders ( 1999 ) rainout prescriptions and the no - rainout prescription ( true equilibrium ) give qualitatively the same result .
since atomic li should exist in abundance to temperatures around 1300 k ( lodders 1999 ) and should trail off below that and since t@xmath8s around 6700 are 1000 k to 1350 k for from 700 to 1300 k ( see fig .
[ bright ] for 700@xmath61100 k ) , there is no obvious reason the 6708 li line should disappear before the l / t boundary .
therefore , we suspect there is an interesting story to be told about real brown dwarf atmospheres in the resolution of this li - line problem .
we have generated spectral models for the temperature range 600 k to 1300 k thought to encompass the t dwarfs .
comparing these models with the new t dwarf data shortward of 1.0 micron , we find that the models can explain in qualitative and semi - quantitative fashion , the new observations and their systematic trends .
furthermore , we demonstrate that untruncated lorentzian line profiles for the na d and k i ( 7700 ) resonance doublets are disfavored , that the atmospheric abundances of sodium and potassium have indeed been enhanced by the rainout of silicates ( bms ; lodders 1999 ) , that water bands weaken with increasing gravity , that modest decreases in metallicity enhance the effect of the alkali lines in the optical , and that at low the behavior of the optical spectra with reverses and becomes bluer with further decreases in ( as predicted by bms ) . moreover , we find that the upper edge of the t dwarf range is near @xmath01300 k. we determine that the optical range is rich in diagnostics that are complementary to those in the near - infrared now used for spectroscopic classification .
an important conclusion is that the t dwarf subtype is not a function of alone ; subtype is also a non - trivial function of gravity and metallicity .
even if the range of metallicities represented by the known t dwarfs is small , gravity will play an important role in the near - ir and shortward of 1.0 micron in determining spectral shape , colors , and spectral subtype . from the shallowness of the water feature near 0.93 , from the weaker than predicted k i features in the @xmath3 band , from the failure of the models to fit the @xmath40 colors of the early t dwarfs , and from the presence in the hotter models of tio and vo features in the @xmath2 and @xmath3 bands ( not seen in the observations of the earlier t dwarfs ) , we find evidence of residual dust / grains / clouds in early t dwarf atmospheres .
a similar conclusion was recently reached by marley et al .
the l dwarfs are known to be dominated by clouds ( burrows et al .
2001 , and references therein ) , and the transition from the l dwarfs to the t dwarfs is predominantly due to the depletion and rainout of heavy metals ( ackerman and marley 2001 ; burrows and sharp 1999 ; marley et al .
1996 ; allard et al .
however , the continuing effect of clouds in the early t dwarfs seems to be indicated by the data .
clouds in t dwarf atmospheres do not explain the redness of the spectrum shortward of 1.0 ( though they might contribute to it ) ; this is mostly explained by the k i line at 7700 and its wing ( burrows , marley , and sharp 2000 ) .
however , the l and t transition is not as abrupt as earlier inferred by the sharp swing to the blue of @xmath40 .
one can still see deeply into the atmosphere , even for low , into regions occupied by silicate ( and other ? ) clouds .
the opacity spectrum of clouds is much more continuous and featureless than that of gas . as a consequence ,
the influence of clouds is manifest by broad shape changes in the optical and near - ir spectra and by the suppression of gas - phase spectral features , not by specific and distinct spectral features , until one gets to @xmath010 and the classical silicate band appears . certainly , good models of the particle size spectrum , spatial extent , and composition of the grains or droplets are required to reproduce the measured spectra of the early t dwarfs in detail .
the later t dwarfs are easier to fit , but the residual effect of clouds at depth on the emergent spectra can not yet be discounted .
the t dwarfs are all brown dwarfs and the number of such substellar objects known to astronomy is growing fast .
their spectra hold the key to their physical characterization and shortward of 1.0 micron these spectra have great diagnostic potential .
we have attempted to make only a preliminary stab at comparing the new data at short wavelengths with theory and look forward to the creation of better opacity databases and cloud models , more precise spectral measurements , and the measurement of parallaxes that together will enable the final , detailed determination of the masses , gravities , , and compositions of the objects that inhabit this fascinating new spectral class .
we are happy to thank christopher sharp , curtis cooper , richard freedman , jonathan lunine , bill hubbard , didier saumon , mark marley , hugh harris , conard dahn , and phil pinto for insightful discussions and advice . for electronic versions of the synthetic spectra discussed in this paper
, please contact the first author .
this work was supported in part by nasa grants nag5 - 10760 , nag5 - 10629 , nag5 - 7499 , and nag5 - 7073 .
a.j.b . acknowledges support from the space telescope science institute through nasa / hst proposal # 8563 .
portions of the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california , and the national aeronautics and space administration .
ackermann , a. and marley , m.s .
2001 , , 556 , 872 allard , f. , hauschildt , p.h . ,
baraffe , i. & chabrier , g. 1996 , , 465 , l123 allard , f. , hauschildt , p.h . ,
alexander , d.r . ,
tamanai , a. , and schweitzer , a. 2001 , in press .
anders , e. and grevesse , n. 1989 , geochim .
acta , 53 , 197 basri , g. et al .
2000 , , 538 , 363 bessell , m.s . , and brett , j. m. 1988 , , 100 , 1134 borysow , a. and frommhold , l. 1990 , , 348 , l41 borysow , a. , jrgensen , u.g . , and
zheng , c. 1997 , , 324 , 185 burgasser , a.j . , et al . 1999 , , 522 , l65 burgasser , a.j . ,
et al . 2000a , , 531 , l57 burgasser , a.j . ,
kirkpatrick , j. d. , reid , i. n. , liebert , j. , gizis , j. e. , & brown , m. e. 2000b , , 120 , 473 burgasser , a.j . , et al .
2000c , , 120 , 1100 burgasser , a.j . ,
et al . 2001a , , 563 , in press burgasser , a.j .
2001b , ph.d .
thesis , california institute of technology burrows , a. , marley m. , hubbard , w.b .
lunine , j.i . ,
guillot , t. , saumon , d. freedman , r. , sudarsky , d. and sharp , c.m .
1997 , , 491 , 856 burrows , a. and sharp , c.m .
1999 , , 512 , 843 burrows , a. , marley , m. s. , and sharp , c. m. 2000 , , 531 , 438 burrows , a. , hubbard , w.b . ,
lunine , j.i . , and liebert , j. 2001 , rev .
phys . , 73 , 719 cuby , j.g . ,
saracco , p. , moorwood , a.f.m . ,
dodorico , s. , lidman , c. , comern , f. , and spyromilio , j. 1999 , , 349 , l41 delfosse , x. , tinney , c.g . ,
forveille , t. , epchtein , n. , bertin , e. , borsenberger , j. , copet , e. , de batz , b. , fouqu , p. , kimeswenger , s. , le bertre , t. , lacombe , f. , rouan , d. , and tiphne , d. 1997 , , 327 , l25 fan , x. , et al .
2000 , , 119 , 928 fukugita , m. et al .
1996 , , 111 , 1748 geballe , t.r .
, saumon , d. , leggett , s.k . ,
knapp , g.r . ,
marley , m.s . , and lodders , k. 2001 , , 556 , 373 geballe , t.r .
2002 , , 563 , in press golimowski , d. a. , 1998 , , 115 , 2579 griffith , c.a . ,
yelle , r.v . , and marley , m.s .
1998 , science , 282 , 2063 griffith , c.a . and
yelle , r.v .
1999 , , 519 , l85 griffith , c.a .
and yelle , r.v .
2000 , , 532 , l59 hubeny , i. and lanz , t. 1995 , , 439 , 875 jones , h.r.a . and tsuji , t. 1997 , , 480 , l39 jones , h.r.a . , pavlenko , y. , viti , s. , and tennyson , j. 2002 , mon . not
soc . , in press .
jrgensen , u.g . , 1997 , in _ molecules in astrophysics : probes and processes _ , ed .
van dishoeck , ( kluwer , iau symp .
178 ) , pp .
441 - 456 khare , b.n . and sagan , c. 1984 , icarus , 60 , 127 kirkpatrick , j.d . , reid , i.n . ,
liebert , j. , cutri , r.m . ,
nelson , b. , beichman , c.a .
, dahn , c.c . ,
monet , d.g .
, gizis , j. , and skrutskie , m.f .
1999 , , 519 , 802 kirkpatrick , j. d. , reid , i. n. , liebert , j. , gizis , j. e. , burgasser , a. j. , monet , d. g. , dahn , c. c. , nelson , b. , & williams , r. j. 2000 , , 120 , 447 kirkpatrick , j.d . , et al .
2001 , submitted to leggett , s. k. , toomey , d. w. , geballe , t. r. , and brown , r. h. 1999 , , 517 , l139 leggett , s. k. , et al .
2000 , , 536 , l35 liebert , j. , reid , i. n. , burrows , a. , burgasser , a. j. , kirkpatrick , j. d. , & gizis , j. e. 2000 , , 533 , l155 lodders , k. 1999 , , 519 , 793 marley , m.s .
, saumon , d. , guillot , t. , freedman , r.s . ,
hubbard , w.b . , burrows , a. & lunine , j.i .
1996 , science , 272 , 1919 marley , m.s . ,
seager , s. , saumon , d. , lodders , k. , ackerman , a.s . , and freedman , r. 2001 , , in press ( astro - ph/0105438 ) .
martn , e. l. , delfosse , x. , basri , g. , goldman , b. , forveille , t. , & zapatero osorio , m. r. 1999 , , 118 , 2466 matthews , k. , nakajima , t. , kulkarni , s.r . , & oppenheimer , b.r .
1996 , , 112 , 1678 mclean , i. , et al .
2000 , , 533 , l45 morgan , w. w. , keenan , p.c . , and kellman , e. 1943 , an atlas of stellar spectra , with an outline of spectral classification .
( chicago : univ .
chicago press ) morgan , d.h . , tritton , s.b . ,
savage , a. , hartley , m. , and cannon , r.d .
1992 , in digitised optical sky surveys , ed .
macgillivray and e.b .
thomson ( dordrecht : boston ) , p. 11
nakajima , t. , oppenheimer , b.r . ,
kulkarni , s.r .
, golimowski , d.a . ,
matthews , k. and durrance , s.t .
1995 , nature , 378 , 463 noll , k. , geballe , t.r .
, & marley , m.s .
1997 , , 489 , 87 noy , m. , podolak , m. , and bar - nun , a. 1981 , j. geophys .
research , 86 , 11985 oke , j.b .
1995 , pasp , 107 , 375 oppenheimer , b.r . ,
kulkarni , s.r .
, matthews , k. , and nakajima , t. 1995 , science , 270 , 1478 partridge , h. and schwenke , d.w .
1997 , j. chem .
106 , 4618 pavlenko , ya.v . , 2001 , astronomy reports , 45 , 144 perryman , m.a.c .
1997 , , 323 , l49 piskunov , n.e . ,
kupka , f. , ryabchikova , t.a . ,
weiss , w.w .
, & jeffery , c.s .
1995 , suppl . , 112 , 525 plez , b. , 1998 , , 337 , 495 plez , b. 1999 , private communication .
ruiz , m.t . ,
leggett , s.k . , and allard , f. 1997 , , 491 , l107 saumon , d. , chabrier , g. , and van horn , h.m .
1995 , , 99 , 713 saumon , d. , geballe , t.r . , leggett , s.k . ,
marley , m.s . ,
freedman , r.s . ,
lodders , k. , fegley , b. , and sengupta , s.k .
2000 , , 541 , 374 schultz , a.b .
1998 , , 492 , l181 strauss , m.a . , et al .
1999 , , 522 , l61 tokunaga , a.t . , and kobayashi , n. 1999 , , 117 , 1010 tsuji , t. , ohnaka , k. , aoki , w. , and nakajima , t. 1996 , , 308 , l29 tsuji , t. , ohnaka , k. , and aoki , w. 1999 , , 520 , l119 tsvetanov , z.i .
, et al . 2000 , , 531 , l61 zheng , c. and borysow , a. 1995 , icarus , 113 , 84 cccc 2massi j1507038 - 151648 & 2mass-1507 & l5 & 1 + 2massw j1632291 + 190441 & 2mass-1632 & l8 & 1 + sdssp [email protected] & sdss-0837 & t1 & 2 + sdssp [email protected] & sdss-1254 & t2 & 2 + sdssp [email protected] & sdss-1021 & t3 & 2 + 2massi j0559191@xmath60140448 & 2mass-0559 & t5 & 3 + sdssp j162414.37 + 002915.6 & sdss-1624 & t6 & 4 + 2massi j0937347 + 293142 & 2mass-0937 & t6p & 5 + sdssp [email protected] & sdss-1346 & t6 & 6 + 2massi j1237392 + 652615 & 2mass-1237 & t6.5 & 7 + gliese 229b & gl 229b & t6.5 & 8 + 2massi j0727182 + 171001 & 2mass-0727 & t7 & 5 + 2massi j1217110@xmath60031113 & 2mass-1217 & t7.5 & 5 + 2massw j1457150 - 212148 & gl 570d & t8 & 9 + 2massi j0415195@xmath60093506 & 2mass-0415 & t8 & 5 + ccccccc 2mass-1507 ( l5,a ) & @xmath61&@xmath62&@xmath63&@xmath64&@xmath65&@xmath65 + 2mass-1632 ( l8,b ) & @xmath66&@xmath67&@xmath68&@xmath69&@xmath65&@xmath65 + sdss-0837 ( t1,c ) & @xmath70&@xmath71&@xmath72&@xmath73&@xmath74&@xmath75 + sdss-1254 ( t2,d ) & @xmath76&@xmath77&@xmath78&@xmath79&@xmath80&@xmath81 + sdss-1021 ( t3,e ) & @xmath82&@xmath83&@xmath78&@xmath84&@xmath85&@xmath86 + 2mass-0559 ( t5,f ) & @xmath87&@xmath88&@xmath72&@xmath89&@xmath90&@xmath91 + sdss-1624 ( t6,g ) & @xmath92&@xmath93&@xmath94&@xmath79&@xmath65&@xmath65 + 2mass-0937 ( t6p , i ) & @xmath95&@xmath96&@xmath97&@xmath98&@xmath99&@xmath100 + sdss-1346 ( t6,h ) & @xmath101&@xmath102&@xmath103&@xmath104&@xmath105&@xmath106 + 2mass-1237 ( t6.5,k ) & @xmath107&@xmath108&@xmath94&@xmath109&@xmath110&@xmath111 + gl 229b ( t6.5,j ) & @xmath60 & @xmath60 & @xmath112&@xmath113&@xmath114&@xmath115 + 2mass-0727 ( t7,l ) & @xmath116&@xmath117&@xmath118&@xmath119&@xmath120&@xmath121 + 2mass-1217 ( t7.5,m ) & @xmath122&@xmath123&@xmath119&@xmath79&@xmath124&@xmath125 + gl 570d ( t8,n ) & @xmath126&@xmath127&@xmath128&@xmath129&@xmath130&@xmath131 + 2mass-0415 ( t8,o ) & @xmath132&@xmath133&@xmath97&@xmath134&@xmath135&@xmath136 + & & & & & & + & & & & & & + & & & & & & + & & & & & & + g = @xmath137 cm s@xmath12 : & & & & & & + & & & & & & + 1300 k & @xmath138&@xmath139&@xmath140&@xmath141&@xmath142&@xmath143 + 1200 & @xmath144&@xmath145&@xmath104&@xmath146&@xmath147&@xmath69 + 1100 & @xmath148&@xmath149&@xmath128&@xmath150&@xmath151&@xmath72 + 1000 & @xmath152&@xmath153&@xmath154&@xmath134&@xmath155&@xmath68 + 900 & @xmath126&@xmath156&@xmath157&@xmath158&@xmath159&@xmath160 + 800 & @xmath161&@xmath162&@xmath163&@xmath104&@xmath164&@xmath140 + 700 & @xmath165&@xmath166&@xmath167&@xmath168&@xmath169&@xmath97 + 600 & @xmath170&@xmath171&@xmath172&@xmath158&@xmath173&@xmath168 + & & & & & & + g = @xmath19 cm s@xmath12 : & & & & & & + & & & & & & + 1300 k & @xmath174&@xmath175&@xmath176&@xmath118&@xmath177&@xmath178 + 1200 & @xmath122&@xmath179&@xmath94&@xmath119&@xmath180&@xmath119 + 1100 & @xmath181&@xmath182&@xmath158&@xmath168&@xmath183&@xmath184 + 1000 & @xmath185&@xmath186&@xmath187&@xmath187&@xmath188&@xmath176 + 900 & @xmath189&@xmath117&@xmath113&@xmath190&@xmath191&@xmath140 + 800 & @xmath192&@xmath133&@xmath193&@xmath187&@xmath194&@xmath97 + 700 & @xmath195&@xmath166&@xmath196&@xmath190&@xmath197&@xmath103 + 600 & @xmath198&@xmath199&@xmath200&@xmath158&@xmath201&@xmath202 + & & & & & & + & & & & & & + no rainout : & & & & & & + & & & & & & + 900 k/[email protected] & @xmath204&@xmath205&@xmath206&@xmath178&@xmath207&@xmath208 + 900/[email protected] & @xmath209&@xmath210&@xmath211&@xmath178&@xmath212&@xmath213 + & & & & & & + & & & & & & + & & & & & & + & & & & & & + g = @xmath137 cm s@xmath12 : & & & & & & + & & & & & & + 1300 k & @xmath214&@xmath215&@xmath68&@xmath141&@xmath216&@xmath143 + 1200 & @xmath70&@xmath217&@xmath218&@xmath94&@xmath219&@xmath220 + 1100 & @xmath221&@xmath222&@xmath84&@xmath129&@xmath223&@xmath69 + 1000 & @xmath161&@xmath224&@xmath104&@xmath109&@xmath225&@xmath72 + 900 & @xmath226&@xmath227&@xmath190&@xmath98&@xmath228&@xmath184 + 800 & @xmath229&@xmath127&@xmath113&@xmath128&@xmath230&@xmath231 + 700 & @xmath232&@xmath127&@xmath233&@xmath234&@xmath235&@xmath220 + 600 & @xmath236&@xmath127&@xmath237&@xmath234&@xmath238&@xmath73 + & & & & & & + g = @xmath19 cm s@xmath12 : & & & & & & + & & & & & & + 1300 k & @xmath239&@xmath240&@xmath72&@xmath94&@xmath241&@xmath176 + 1200 & @xmath242&@xmath243&@xmath244&@xmath150&@xmath245&@xmath244 + 1100 & @xmath246&@xmath156&@xmath160&@xmath109&@xmath247&@xmath248 + 1000 & @xmath226&@xmath249&@xmath218&@xmath202&@xmath250&@xmath176 + 900 & @xmath232&@xmath251&@xmath118&@xmath234&@xmath252&@xmath244 + 800 & @xmath253&@xmath254&@xmath143&@xmath255&@xmath256&@xmath141 + 700 & @xmath257&@xmath258&@xmath97&@xmath233&@xmath259&@xmath218 + 600 & @xmath260&@xmath251&@xmath134&@xmath234&@xmath261&@xmath118 + & & & & & & + & & & & & & + & & & & & & + & & & & & & + g = @xmath137 cm s@xmath12 : & & & & & & + & & & & & & + 1300 k & @xmath204&@xmath262&@xmath158&@xmath220&@xmath263&@xmath264 + 1200 & @xmath265&@xmath266&@xmath187&@xmath218&@xmath267&@xmath141 + 1100 & @xmath268&@xmath269&@xmath270&@xmath150&@xmath271&@xmath184 + 1000 & @xmath272&@xmath273&@xmath274&@xmath150&@xmath275&@xmath78 + 900 & @xmath276&@xmath273&@xmath277&@xmath89&@xmath278&@xmath244 + 800 & @xmath279&@xmath224&@xmath280&@xmath150&@xmath281&@xmath218 + 700 & @xmath282&@xmath162&@xmath283&@xmath134&@xmath284&@xmath97 + 600 & @xmath285&@xmath286&@xmath287&@xmath119&@xmath288&@xmath168 + & & & & & & + g = @xmath19 cm s@xmath12 : & & & & & & + & & & & & & + 1300 k & @xmath289&@xmath290&@xmath97&@xmath84&@xmath291&@xmath292 + 1200 & @xmath82&@xmath83&@xmath104&@xmath134&@xmath293&@xmath294 + 1100 & @xmath116&@xmath71&@xmath98&@xmath129&@xmath295&@xmath72 + 1000 & @xmath285&@xmath296&@xmath233&@xmath129&@xmath297&@xmath160 + 900 & @xmath298&@xmath286&@xmath299&@xmath104&@xmath300&@xmath118 + 800 & @xmath301&@xmath302&@xmath200&@xmath103&@xmath303&@xmath89 + 700 & @xmath181&@xmath88&@xmath304&@xmath158&@xmath305&@xmath187 + 600 & @xmath306&@xmath307&@xmath308&@xmath84&@xmath309&@xmath233 + & & & & & & + | we have generated new , self - consistent spectral and atmosphere models for the effective temperature range 600 k to 1300 k thought to encompass the known t dwarfs .
for the first time , theoretical models are compared with a _ family _ of measured t dwarf spectra at wavelengths shortward of @xmath01.0 micron . by defining spectral indices and standard colors in the optical and very near - infrared
, we explore the theoretical systematics with , gravity , and metallicity .
we conclude that the short - wavelength range is rich in diagnostics that complement those in the near - infrared now used for spectral subtyping .
we also conclude that the wings of the na d and k i ( 7700 ) resonance lines and aggressive rainout of heavy metals ( with the resulting enhancement of the sodium and potassium abundances at altitude ) are required to fit the new data shortward of 1.0 .
furthermore , we find that the water bands weaken with increasing gravity , that modest decreases in metallicity enhance the effect in the optical of the sodium and potassium lines , and that at low , in a reversal of the normal pattern , optical spectra become bluer with further decreases in .
moreover , we conclude that t dwarf subtype is not a function of alone , but that it is a non - trivial function of gravity and metallicity as well . as do marley et al .
( 2001 ) , we see evidence in early t dwarf atmospheres of a residual effect of clouds . with cloudless models , we obtain spectral fits to the two late t
dwarfs with known parallaxes , but a residual effect of clouds on the emergent spectra of even late t dwarfs can not yet be discounted
. however , our focus is not on detailed fits to individual objects , but on the interpretation of the overall spectral and color trends of the entire class of t dwarfs , as seen at shorter wavelengths . |
consider a network @xmath0 , where @xmath1 denotes the set of vertices in @xmath2 , and @xmath3 denotes the set of edges in @xmath2 .
a vertex in @xmath2 is said to be a _ source _ if it is only incident with outgoing edges , and a _ sink _ if it is only incident with incoming edges .
often , a source or sink is referred to as a _ terminal _ vertex .
a non - terminal vertex is said to be a _ hub _ if its degree is greater than or equal to @xmath4 . in this paper , we are primarily concerned with the minimum number of hubs needed when certain constraints on the flow demand between multiple pairs of sources and sinks are imposed .
the flow demand constraints considered in this paper will be in terms of the vertex - cuts between pairs of sources and sinks .
this can be justified by a vertex version of the max - flow min - cut theorem @xcite , which states that for a network with infinite edge - capacity and unit vertex - capacity , the maximum flow between one source and one sink is equal to the minimum vertex - cut between them . here , we remark that with appropriately modified setup , our results can be stated in terms of edge - cuts as well .
more precisely , for given @xmath5 , let @xmath6 denote the set of all finite networks @xmath2 ( see figure [ illustrator ] for an example ) such that * there are @xmath7 sources @xmath8 and @xmath7 sinks @xmath9 in @xmath2 ; * all edges in @xmath2 , except those incident with a source or sink , are undirected ( alternatively , bi - directional ) ; * for each feasible @xmath10 , the minimum vertex - cut from @xmath11 to @xmath12 is @xmath13 .
now , we define @xmath14 where @xmath15 denotes the number of hubs in @xmath16 . the above definition can be roughly interpreted as follows : for a given @xmath2 , we try to find a subgraph @xmath16 that contains the minimum number of hubs required to satisfy the vertex - cut constraints , and @xmath17 gives us the minimum number corresponding to the worst - case scenarios among all possible @xmath2 . with @xmath18 hubs.,scaledwidth=30.0% ] at first glance , @xmath19 can be infinite .
one of the main ( and somewhat surprising ) results in this paper , theorem [ finiteness ] , however , states that for any given @xmath20 , @xmath19 is in fact finite .
with finiteness confirmed , we are primarily interested in computing the value of @xmath19 .
we say a graph @xmath2 is @xmath21 in @xmath22 , if @xmath23 , but for any @xmath24 , @xmath25 .
it then follows from the fact that every graph in @xmath6 has at least one minimal subgraph that @xmath26 the vertex - connectivity version of the classical menger s theorem @xcite states that for a network with one pair of source @xmath27 and sink @xmath28 with the minimum vertex - cut between them being @xmath29 , there exist @xmath29 vertex - disjoint paths connecting @xmath27 and @xmath28 , which immediately implies that @xmath30 for any given @xmath29 .
theorem [ finiteness ] states that for any given @xmath23 , one can always find a subgraph @xmath16 of @xmath2 such that @xmath15 is upper bounded by a constant , which is independent of the choice of @xmath2 . in some sense ,
theorem [ finiteness ] can be viewed as a generalization of the vertex - connectivity version of menger s theorem .
mathematically , the proposed problem of computing @xmath19 is a natural combinatorial optimization problem . on a more practical side ,
hubs in networks naturally correspond to more costly vertices .
for instance , in a transportation network , as opposed to `` relaying '' vertices with degree @xmath31 , hubs may have to be better equipped for traffic scheduling ; for this reason , when designing the route map , an airline may need to avoid running too many airline hubs to reduce the cost .
so , as might be expected , @xmath19 is of significance to cost - minimizing resource allocation in transportation networks . to the best of our knowledge ,
the proposed problem of computing or estimating @xmath32 has not yet been examined previously and there is little related work in the vast literature of graph theory . on the other hand , to a great extent , this work is motivated by the study of network encoding complexity ( see @xcite and references therein ) , where the number of encoding vertices in directed networks is of primary concern . moreover ,
our approaches to tackle the problem are influenced by those in network encoding complexity theory , particularly , to a greater extent , those in @xcite .
the remainder of the paper is organized as follows . in section [ minimalgraph ] ,
we give necessary and sufficient conditions ( theorem [ equivalent ] ) for a graph being minimal in @xmath33 . in section [ sectionminimalrepresentation ] ,
we introduce the notion of a representation of a graph in @xmath33 and we present the structural decomposition theorem ( theorem [ decomposition ] ) for representations of minimal graphs in @xmath33 .
we will introduce in section [ xu - algorithm ] a novel path - searching algorithm , the analysis of which will aptly produce an upper bound on @xmath34 for any given @xmath35 . in section [ sectionhmn ] , we derive the value of @xmath36 ( theorem [ nmn2mn ] ) , which is a main result in this paper .
another main result is theorem [ finiteness ] , which establishes the finiteness of @xmath19 , @xmath37 , through a recursive bounding argument .
the remaining part of section [ more - than - two ] will be devoted to the derivations of the values of @xmath32 for some special parameters ( theorem [ thm11mn ] , [ thmh222 ] ) .
we first introduce some notation and terminologies that will be used throughout the paper . a sequence of edges in @xmath2 , @xmath39 , @xmath40 , with @xmath41 , @xmath42 ,
can be linked to form a _ path _
@xmath43 , denoted by @xmath44 ; furthermore , @xmath43 is called a _ cycle _ if @xmath45 .
a path @xmath43 of this form is said to be _ directed _ each @xmath46 is oriented such that all of them concatenate . for a path @xmath47 and two vertices @xmath48 , @xmath49 of @xmath47 , let @xmath50 $ ] denote the subpath of @xmath47 between @xmath48 and @xmath49 . for a directed path @xmath51 , let @xmath52 , @xmath53 denote the head , tail of @xmath47 , respectively ; and we say @xmath46 is @xmath54 than @xmath55 on @xmath47 , denoted by @xmath56 , if @xmath57 is the head of the directed path @xmath58 $ ] ( or alternatively , @xmath55 is _ larger _ than @xmath46 , denoted by @xmath59 ) . for a graph @xmath23 , by the vertex - connectivity version of menger s theorem , for each @xmath10 , one can find a set @xmath60 of @xmath13 vertex - disjoint @xmath61-paths from @xmath11 to @xmath12 .
if the choice of @xmath60 is unique , @xmath60 is said to be @xmath62 , otherwise it is said to be _
reroutable_. @xmath2 is said to be _ non - reroutable _ if all @xmath60 s are non - reroutable , _ reroutable _ otherwise . for any index set @xmath63 , let @xmath64 denote the subgraph of @xmath2 induced on the edges of @xmath65-paths .
@xmath2 is said to be a _ @xmath66-graph _ if @xmath67 , that is , each edge in @xmath2 belongs to some @xmath60-path . in order to compute @xmath19 ,
it is enough to only consider all @xmath66-graphs , since @xmath68 is a subgraph of @xmath2 and also in @xmath22 .
sections [ minimalgraph ] to [ sectionhmn ] will be devoted to derive the value of @xmath34 . for notational convenience only
, we often rewrite @xmath69 , @xmath70 as @xmath61 , @xmath71 , respectively . for a @xmath72-graph @xmath2 , an edge in @xmath2
is said to be _ public _ if it is shared by a @xmath61-path and a @xmath71-path , _ private _ otherwise .
evidently , for each @xmath10 , from @xmath11 to @xmath12 , each @xmath61 or @xmath71-path in @xmath2 induces a natural orientation to all its edges .
we note that a public edge in @xmath2 may have opposite @xmath61-direction and @xmath71-direction ( such `` inconsistency '' will be dealt with in section [ sectionminimalrepresentation ] ) .
we say @xmath2 can be _ naturally orientable _ if each public edge in @xmath2 has the same natural @xmath61 and @xmath71-direction . a cycle @xmath73 in @xmath2 , where @xmath74 , @xmath41 for @xmath75 and @xmath45 , is said to be a _ @xmath61-consistent cycle _ , if it satisfies the following property : for any @xmath76 , if @xmath46 belongs to a @xmath61-path , then its natural @xmath61-direction is from @xmath77 to @xmath78 . and
we similarly define a @xmath71-consistent cycle .
the following theorem gives necessary and sufficient conditions for a @xmath72-graph being minimal .
[ equivalent ] the following three statements are equivalent for a @xmath72-graph @xmath2 : * @xmath2 is minimal ; * @xmath2 is non - reroutable ; * @xmath2 has no @xmath61 or @xmath71-consistent cycle .
* 1 . * * ( ii ) @xmath79 ( i ) * : any edge @xmath80 in @xmath2 must belong to some @xmath61 or @xmath71-path . after deleting @xmath80 from @xmath2 , we no longer find @xmath13 vertex - disjoint paths from @xmath11 to @xmath12 for some @xmath81 .
so @xmath82 , and therefore @xmath2 is minimal .
* 2 . * * ( ii ) @xmath79 ( iii ) * : suppose , by way of contradiction , that @xmath2 has a @xmath61-consistent cycle @xmath83 , which can be written in the following form @xmath84 where each @xmath85 is a subpath of some @xmath61-path , and each @xmath46 is a private @xmath71-edge . furthermore , we assume @xmath83 has the smallest @xmath86 ( the number of private @xmath71-edges ) among all @xmath61-consistent cycles
. then , each @xmath85 belongs to a different @xmath61-path ( since otherwise we can always find a @xmath61-consistent cycle with fewer private @xmath71-edges ) , which further implies that @xmath87 .
suppose @xmath88 $ ] for @xmath76 , @xmath89 for @xmath90 and @xmath91
. then one can find another group of @xmath92 vertex - disjoint paths @xmath93 from @xmath94 to @xmath95 in @xmath2 , where @xmath96 \circ ( u_1,v_d ) \circ \phi_d[u_d , r_1]&\textrm { for } i=1;\\ \phi_i[s_1,u_i ] \circ ( u_i , v_{i-1 } ) \circ \phi_{i-1}[v_{i-1},r_1]&\textrm { for } 2\le i\le d;\\ \phi_i&\textrm { for } d+1\le i\le c_1,\\ \end{cases}\ ] ] which contradicts the assumption that @xmath2 is non - reroutable . with a parallel argument
, we conclude that @xmath71-consistent cycles do not exist either .
* 3 . * * ( i ) or ( iii ) @xmath79 ( ii ) * : suppose , by contradiction , that @xmath2 is reroutable , and furthermore , by symmetry , that there exists another group of @xmath92 vertex - disjoint paths @xmath93 from @xmath94 to @xmath95 in @xmath2 with @xmath97 sharing the same outgoing edge from @xmath94 as @xmath98 for every @xmath10 . pick a @xmath99-path , say , @xmath100 , such that @xmath101 , and let @xmath102 denote the smallest vertex on @xmath103 ( under the natural @xmath61-direction ) where they leave each other . assume that , after @xmath102 , @xmath100 first meets some @xmath61-path , say , @xmath104 at the vertex @xmath105 .
denote by @xmath106 the smallest vertex where @xmath107 and @xmath104 leave each other .
assume that , after @xmath106 , @xmath107 first meets some @xmath61-path , say , @xmath108 at the vertex @xmath109 .
continue the procedure in a similar manner to obtain an index sequence @xmath110 , and similarly define @xmath111 s and @xmath112 s .
pick the smallest @xmath113 such that @xmath114 for some @xmath115 .
notice that @xmath116 is smaller than @xmath112 on @xmath117 , which easily follows from three facts : * @xmath118 first takes apart from @xmath117 at @xmath116 ; * @xmath119 meets @xmath117 at @xmath112 ; * @xmath119 and @xmath118 are vertex - disjoint .
hence , we conclude that @xmath120 \circ \phi_{i_{j+1}}[u_{i_j},v_{i_{j+1 } } ] \circ \hat{\phi}_{i_{j+1}}[v_{i_{j+1}},u_{i_{j+1 } } ] \circ \phi_{i_{j+2}}[u_{i_{j+1}},v_{i_{j+2 } } ] \circ \\ \cdots & \circ \hat{\phi}_{i_{k-2}}[v_{i_{k-2}},u_{i_{k-2 } } ] \circ \phi_{i_{k-1}}[u_{i_{k-2}},v_{i_{k-1 } } ] \circ \hat{\phi}_{i_{k-1}}[v_{i_{k-1}},u_{i_{k-1 } } ] \circ \phi_{i_{j}}[u_{i_{k-1}},v_{i_{j}}]\end{aligned}\ ] ] is a @xmath61-consistent cycle in @xmath2 .
since all @xmath71-paths are vertex - disjoint and the above cycle does not contain any terminal vertex , at least one edge @xmath80 in this cycle does not belong to any @xmath71-path ( that is , @xmath80 is a private @xmath61-edge ) . notice that each edge of @xmath121 $ ] , @xmath122 , is a private @xmath71-edge , and each edge of @xmath123 $ ] , @xmath122 , does not belong to any @xmath99-path .
so , @xmath80 must belong to @xmath123 $ ] for some @xmath124 with @xmath122 , and hence @xmath80 does not belong to any @xmath99-path . in the graph @xmath125
, we can find a set @xmath99 of @xmath92 vertex - disjoint paths from @xmath94 to @xmath95 , and a set @xmath71 of @xmath126 vertex - disjoint paths from @xmath127 to @xmath128 .
thus @xmath129 and @xmath2 is not minimal , a contradiction .
[ ifreroutable ] the third part of the proof of the theorem has actually proved that for a @xmath130-graph @xmath2 , if @xmath61 ( resp .
@xmath71 ) is reroutable , then there exists a private @xmath61 ( resp .
@xmath71)-edge @xmath80 such that @xmath131 .
this fact will be used later in the paper .
in this section , we will transform a minimal @xmath130-graph @xmath2 into @xmath132 , its representation , through the following steps : * step 1 [ remove relays ] * : in this step , we remove all non - terminal vertices in @xmath2 with degree @xmath31 . in more detail , for any non - terminal vertex @xmath49 with @xmath133 and @xmath134 , @xmath135 being the two edges incident with @xmath49 , where @xmath136 , delete edges @xmath137 , @xmath138 and vertex @xmath49 , and then add a new edge @xmath139 .
let @xmath140 denote the resulting graph .
* step 2 [ stretch crossings ] * : we say a @xmath61-path and a @xmath71-path _ cross _ at vertex @xmath49 if they share @xmath49 , but not any edges incident with @xmath49 . in this step ,
we convert each crossing into a pair of degree @xmath4 vertices . in more detail , for any vertex @xmath49 in @xmath140 with @xmath141 and @xmath134 , @xmath135 being the two @xmath61-edges incident with @xmath49 , @xmath142 , @xmath143 being the two @xmath71-edges incident with @xmath49 , where all @xmath77 are all distinct , delete edges
@xmath137 , @xmath138 , @xmath144 , @xmath145 and vertex @xmath49 , and then add two vertices @xmath146 , @xmath147 and edges @xmath148 , @xmath149 , @xmath150 , @xmath151 and @xmath152 .
let @xmath153 denote the resulting graph .
* step 3 [ match directions ] * : for any public edge @xmath80 in @xmath153 with inconsistent @xmath61 and @xmath71-direction , we will perform the following operations to obtain consistency : assume edge @xmath154 belongs to both path @xmath155 \circ(w_1,u)\circ ( u , v)\circ(v , w_2)\circ \phi_i[w_2,r_1]$ ] and path @xmath156\circ(w_3,v)\circ(v , u)\circ(u , w_4)\circ\psi_j[w_4,r_2]$ ] .
we delete edges @xmath157 , add edges @xmath158 , and then obtain a new @xmath71-path @xmath159\circ ( w_3,u)\circ ( u , v)\circ ( v , w_4)\circ \psi_j[w_4,r_2].\ ] ] let @xmath132 denote the resulting graph , which , evidently , is naturally orientable ; let @xmath160 denote the directed version of @xmath161 , equipped with the consistent natural orientation induced on all its @xmath61 and @xmath71-paths .
apparently , @xmath162 with its @xmath61 and @xmath71-paths determined by the original ones , and all the non - terminal vertices in @xmath161 are hubs .
the obtained @xmath132 after the above three steps is said to be a _ representation _ of @xmath2 . the following lemma states that @xmath161 must be a minimal @xmath72-graph as well .
[ minimalrepresentation ] the representation @xmath161 of a minimal @xmath72-graph @xmath2 is also minimal .
first , since @xmath2 is minimal , by theorem [ equivalent ] , @xmath2 is non - reroutable .
evidently , @xmath140 is also non - reroutable , and thus minimal . by way of contradiction
, suppose that @xmath132 is not minimal .
again , by theorem [ equivalent ] , @xmath132 is reroutable .
notice that in step 2 and 3 , both crossings or inconsistently oriented public edges are converted into consistently oriented public edges .
then , by remark [ ifreroutable ] , one can find a private edge @xmath80 such that @xmath163 , implying @xmath164 also belongs to @xmath38 , which contradicts the minimality of @xmath140 .
let @xmath165 denote the subset of @xmath38 consisting of all networks @xmath2 such that @xmath2 is minimal , naturally orientable and all the non - terminal vertices in @xmath2 are of degree 3 .
apparently , @xmath166 is the set of the representations of all minimal @xmath130-graphs .
the following theorem says that in order to compute @xmath34 , it is enough to only consider all graphs in @xmath165 .
@xmath167 the `` @xmath168 '' direction follows from the observation that @xmath169 and the `` @xmath170 '' direction immediately follows from lemma [ minimalrepresentation ] .
a path in @xmath171 is said to be an _ alternating _ path if all its edges are privates edges and the terminal pair of this path is one of the following : @xmath172 , @xmath173 , @xmath174 or @xmath175 .
an alternating path has the following properties : 1 .
each of its @xmath61-edges is only adjacent to @xmath71-edges , and each of its @xmath71-edges is only adjacent to @xmath61-edges .
2 . each of its @xmath61 ( resp .
@xmath71)-edge belongs to a different @xmath61 ( resp .
@xmath71)-path .
this follows from the fact that after step @xmath176 , vertices with degree @xmath31 have been removed and thus no two private @xmath61 ( or @xmath71)-edges are adjacent .
we show that in any alternating path , each @xmath61-edge belongs to a different @xmath61-path .
suppose , by contradiction , that for an alternating path path @xmath177 , where @xmath74 , two edges @xmath178 , @xmath179 , both belong to @xmath180 . if @xmath181 is smaller than @xmath182 on path @xmath180 in @xmath183 , then @xmath184\circ e_{k+1 } \circ e_{k+2 } \circ \cdots
\circ e_{l-1}\ ] ] is a @xmath61-consistent cycle in @xmath132 , which , by theorem [ equivalent ] , gives us a contradiction .
if @xmath182 is smaller than @xmath181 on path @xmath180 in @xmath183 , then the subpath @xmath185 $ ] in fact can be expressed as @xmath186 , where @xmath187 are public and @xmath188 are private .
then @xmath189 is a @xmath71-consistent cycle in @xmath132 , which , by theorem [ equivalent ] , gives us a contradiction . with a parallel argument
, we conclude that each private @xmath71-edge belongs to a different @xmath71-path .
we next present the structural decomposition theorem of a representation @xmath190 , which , roughly speaking , states that after deleting public edges in @xmath161 , each connected component in the resulting graph is an alternating path .
more precisely , letting @xmath191 denote the subgraph of @xmath132 induced on all private edges in @xmath161 , we have [ decomposition ] @xmath191 consists of @xmath192 alternating paths .
consider @xmath193 , the naturally oriented version of @xmath191 .
obviously , the degree of any non - terminal vertex in @xmath193 is @xmath31 .
now , starting from @xmath94 , traverse along an outgoing @xmath61-edge , say , @xmath137 , and then traverse against the @xmath71-edge adjacent to @xmath137 , say , @xmath138 , and then along a @xmath61-edge adjacent to @xmath138 , say , @xmath144 , and then against a @xmath71-edge adjacent to @xmath144 .
continue the procedure in this fashion , we will always reach @xmath95 or @xmath127 , since otherwise the set of edges that we have traversed will contain a cycle , which is both @xmath61-consistent and @xmath71-consistent .
evidently , a similar argument can be applied to the case when we start from an incoming edge incident with @xmath128 .
it then follows that one can find a set of @xmath192 vertex - disjoint paths in @xmath191 from @xmath194 to @xmath195 ; moreover , it can easily checked that each edge in @xmath191 belong to one of the above - mentioned vertex - disjoint paths .
[ head - tail ] let @xmath49 be a non - terminal vertex of an alternating path in @xmath191 and let @xmath196 be the two edges incident with @xmath49 .
it then follows from the proof of theorem [ decomposition ] that if @xmath49 is the head ( resp .
tail ) of @xmath80 in @xmath193 , then it is also the head ( resp .
tail ) of @xmath197 .
an alternating path in @xmath191 is said to be an @xmath198 , @xmath199 , @xmath200 , @xmath201-alternating path if its terminal pair is @xmath172 , @xmath173 , @xmath174 or @xmath175 , respectively .
a path is said to be an @xmath94-alternating path is it is either an @xmath198-alternating path or @xmath199-alternating path , similarly , a path is also referred to as an @xmath128-alternating path is it is either an @xmath200-alternating path or @xmath201-alternating path .
apparently , there are @xmath92 @xmath94-alternating paths and @xmath126 @xmath128-alternating paths . for two vertices @xmath202 of an @xmath94 ( resp .
@xmath128)-alternating path @xmath203 , we say @xmath48 is on the right of @xmath49 in @xmath203 if @xmath49 is `` nearer '' to @xmath94 ( resp .
@xmath128 ) than @xmath48 in @xmath203 ( more precisely , the number of edges between @xmath48 and @xmath94 ( resp .
@xmath128 ) in @xmath203 is more than that between @xmath49 and @xmath94 ( resp .
@xmath128 ) ) ( the word `` right '' arises since we will `` position '' the vertices of a path in a plane in section [ xu - algorithm ] for an easy illustration ) . for two edges @xmath204 in @xmath203
, we say @xmath80 is on the right of @xmath197 if one of two vertices incident with @xmath80 is on the right of the two ones incident with @xmath197 in @xmath203 .
figure [ illustrator ] shows a naturally oriented minimal @xmath205-graph with @xmath206 hubs . from @xmath94 to @xmath95 , there are two vertex - disjoint @xmath61-paths @xmath207 and @xmath208 ; and from @xmath127 to @xmath128 , there are two vertex - disjoint @xmath71-paths @xmath209 and @xmath210 .
the edges @xmath144 , @xmath211 , @xmath212 , @xmath213 are public since each of them is shared by some @xmath61-path and some @xmath71-path .
the others are all private
. then we can find that @xmath191 consists of four alternating paths : two @xmath94-alternating paths @xmath214 and two @xmath128-alternating paths @xmath215 moreover , @xmath216 is an @xmath217-alternating path , where edge @xmath218 is on the right of edge @xmath219 , vertex @xmath220 is on the right of vertex @xmath29 .
in this section , we introduce an algorithm , the analysis of which will be instrumental for deriving the value of @xmath34 . before rigorously describing the algorithm , we roughly illustrate its idea .
consider a representation @xmath161 of a minimal @xmath130-graph @xmath2 .
imagine on a two - dimensional plane , each alternating path @xmath73 in @xmath191 is `` positioned '' into a `` double deck '' ( see figure [ alternatingpath ] for an example ) , where @xmath221 is on the right of @xmath46 , and all the tails ( in @xmath222 ) are on the upper deck and heads on the lower deck ( see remark [ head - tail ] ) .
a vertex on the upper deck is referred to as an _ upper vertex _ if it is not a source , and a vertex on the lower deck is referred to as a @xmath223 if it is not a sink .
notice that there is one more lower vertices than upper vertices in an @xmath217-alternating path , while one less in an @xmath224-alternating path , and there are equally many of them in an @xmath225-alternating path or an @xmath226-alternating path .
-alternating path , @xmath227 , @xmath228 , @xmath229 are upper vertices and @xmath230 , @xmath231 are lower vertices.,scaledwidth=50.0% ] in each @xmath217-alternating ( resp . @xmath224-alternating ) path , one of the lower ( resp .
upper ) vertices will be labeled as a @xmath232 ( which could be a `` bottleneck '' for the path extending procedure in the algorithm ) , which is initially the rightmost lower ( resp .
rightmost upper ) vertex of the path and will be updated during the execution of the algorithm .
as summarized below , the algorithm will iteratively find a set of the so - called _ interconnecting _ paths in @xmath161 , which will `` link '' the double decks .
* in the beginning of each iteration , we initialize the interconnecting path as a public edge , whose tail is an unoccupied lower vertex .
see step 2 . *
we first traverse along edges in @xmath160 to extend the interconnecting paths to link the double decks .
when we reach a choke vertex , we check if it is possible to switch paths ( see figure [ switching ] ) and find new ways to extend the interconnecting paths .
see step 3(a ) , 3(b ) .
* then we traverse against edges in @xmath160 to extend the interconnecting paths in a parallel manner .
see step 5(a ) , 5(b ) .
* at the end of @xmath113-th iteration , a set of @xmath113 interconnecting paths is found , whose vertices are all labeled as occupied .
see step 6 .
as proven later , the algorithm will produce all interconnecting paths and upon termination of the algorithm , all hubs in @xmath161 are occupied .
let @xmath233 be the number of @xmath217-alternating paths , then @xmath234 , and moreover , the numbers of @xmath224-alternating paths , @xmath225-alternating paths and @xmath226-alternating paths are @xmath235 , respectively .
the following algorithm will find @xmath233 interconnecting paths in @xmath161 .
[ algorithm ] input : @xmath161 ; output : a set @xmath236 of @xmath233 interconnecting paths .
step 1 : initialize the algorithm .
: : orient each edge in the direction of @xmath61 and @xmath71-paths ; + label all hubs as `` unoccupied '' ; + * for * each @xmath217-alternating path * do * + its choke vertex @xmath237 its rightmost lower vertex ; + * for * each @xmath224-alternating path * do * + its choke vertex @xmath237 its rightmost upper vertex ; + @xmath238 ; + @xmath239 .
step 2 : initialize the forward extension .
: : @xmath240 an arbitrarily picked unoccupied lower vertex and label @xmath49 as `` occupied '' ; + @xmath241 the public edge whose tail is @xmath49 ; + @xmath242 and label @xmath48 as `` occupied '' ; + @xmath243 .
step 3(a ) : prepend a private edge .
: : @xmath244 the alternating path to which @xmath48 belongs ; + * if * @xmath203 is an @xmath224-alternating path * and * @xmath48 is the choke vertex of @xmath203 * then * * begin * + * if * there are no unoccupied upper vertices in @xmath203 * then * + go to * step 4 * ; + * else * * begin * + @xmath245 the rightmost unoccupied upper vertex of @xmath203 ; + @xmath246 the subpath of @xmath203 between @xmath247 and @xmath48 ; + * if * @xmath248 * then * * begin * + the choke vertex of @xmath249 ; + label @xmath250 as `` occupied '' ; + @xmath251 ; + @xmath252 ; + * end * + * else * ( _ viz . _
@xmath253 ) * begin * + * for * @xmath254 * to * @xmath86 * do * + @xmath255 the interconnecting path in @xmath236 containing @xmath256 ; + @xmath257 ; + the choke vertex of @xmath249 ; + label @xmath250 as `` occupied '' ; + @xmath258 ; + @xmath259 ; + @xmath260 \circ e$ ] ; + * for * @xmath254 * to * @xmath261 * do * + @xmath262 \circ ( x_{i+1},y_i ) \circ p_{i}[y_i , h(p_{i})]$ ] ; + @xmath263 \circ ( u , y_d ) \circ p_{d}[y_d , h(p_{d})]$ ] ; + @xmath264 ; + * end * + * end * + * end * + * else * * begin * + * if * @xmath265 an @xmath94-alternating path * then * + @xmath266 the private @xmath71-edge whose tail is @xmath48 ; + * else * ( _ viz . _
@xmath267 an @xmath128-alternating path ) + @xmath266 the private @xmath61-edge whose tail is @xmath48 ; + label @xmath268 as `` occupied '' ; + @xmath269 . + * end * step 3(b ) : prepend a public edge .
: : @xmath241 public edge whose tail is @xmath268 ; + @xmath242 and label @xmath48 as `` occupied '' ; + @xmath270 ; + go to * step 3(a)*. step 4 : initialize the backward extension .
: : @xmath271 ; + @xmath272 the interconnecting path in @xmath236 whose tail is @xmath49 ; + @xmath273 ; + @xmath274 .
step 5(a ) : append a private edge .
: : @xmath275 the alternating path to which @xmath276 belongs ; + * if * @xmath203 is an @xmath217-alternating path * and * @xmath276 is the choke vertex of @xmath203 * then * * begin * + * if * there is no unoccupied lower vertex in @xmath203 * then * + go to * step 6 * ; + * else * * begin * + @xmath277 the rightmost unoccupied lower vertex of @xmath203 ; + @xmath278 the subpath of @xmath203 between @xmath250 and @xmath276 ; + * if * @xmath248 * then * * begin * + the choke vertex of @xmath279 ; + label @xmath247 as `` occupied '' ; + @xmath280 ; + @xmath281 ; + * end * + * else * ( _ viz . _
@xmath253 ) * begin * + * for * @xmath254 * to * @xmath86 * do * + @xmath282 the interconnecting path in @xmath236 containing @xmath256 ; + @xmath257 ; + the choke vertex of @xmath279 ; + label @xmath247 as `` occupied '' ; + @xmath283 ; + @xmath259 ; + @xmath284 $ ] ; + * for * @xmath254 * to * @xmath261 * do * + @xmath285 \circ ( x_i , y_{i+1 } ) \circ p_{{i+1}}[y_{i+1 } , h(p_{{i+1}})]$ ] ; + @xmath286 \circ ( x_d , w ) \circ \widehat{p}[w , h(\widehat{p})]$ ] ; + @xmath264 ; + * end * + * end * + * end * + * else * * begin * + * if * @xmath287 an @xmath128-alternating path * then * + @xmath266 the private @xmath61-edge whose head is @xmath276 ; + * else * ( _ viz . _
@xmath287 an @xmath94-alternating path ) + @xmath266 the private @xmath71-edge whose head is @xmath276 ; + label @xmath288 as `` occupied '' ; + @xmath289 . + *
end * step 5(b ) : append a public edge .
: : @xmath241 public edge whose head is @xmath288 ; + @xmath290 and label @xmath276 as `` occupied '' ; + @xmath291 ; + go to * step 5(a)*. step 6 : terminate the iteration .
: : * if * @xmath292 * then * + terminate the algorithm ; + @xmath271 ; + @xmath293 ; + go to * step 2*. before switching paths 0.5 cm after switching paths * 1 .
* for step 2 in each iteration , an unoccupied lower vertex always exists since there is at least one @xmath217-alternating path with its chock vertex left unoccupied .
each iteration of algorithm [ algorithm ] consists of forward and backward extensions , and each interconnecting path starts from an @xmath217-alternating path and ends at an @xmath201-alternating path .
the algorithm can not be simplified into a `` one - direction extension '' version : for each iteration , the extending procedure will terminate at an @xmath217-alternating path with only one unoccupied lower vertex , but , in the beginning of each iteration , such an alternating path may not exist ( to see this , notice that more than two unoccupied vertices of an alternating path can become `` occupied '' , if we switch paths during one iteration ) .
we will need the following two lemmas in the next section .
[ distinctaasequence ] for a representation @xmath161 of a minimal @xmath130-graph , upon the termination of algorithm [ algorithm ] , each private edge of any interconnecting path belongs to a different alternating path . by contradiction , suppose that for an interconnecting path @xmath294 , two private edges @xmath295 and @xmath296 , @xmath297 , belong to the same alternating path @xmath203 .
let @xmath83 be the cycle @xmath298\circ l[t(e_{2l}),h(e_{2k})].\ ] ] if @xmath296 is on the right of @xmath295 in @xmath203 , then @xmath83 is a @xmath71-consistent cycle in @xmath161 .
if @xmath295 is on the right of @xmath296 in @xmath203 , then @xmath83 is a @xmath61-consistent cycle in @xmath161 . by theorem [ equivalent ] ,
both cases imply that @xmath161 is not minimal , which , by lemma [ minimalrepresentation ] , further implies that @xmath2 is not minimal , a contradiction .
[ atmostmconnectingpaths ] for a representation @xmath161 of a minimal @xmath130-graph @xmath2 , upon the termination of algorithm [ algorithm ] , @xmath236 consists of at most @xmath299 interconnecting paths , and each hub in @xmath161 is in exactly one of the interconnecting paths in @xmath236 .
each interconnecting path in @xmath236 starts from a lower vertex of a distinct @xmath217-alternating path and ends to an upper vertex of a distinct @xmath224-alternating path .
so the number of interconnecting paths is @xmath233 with @xmath234 .
now upon the termination of algorithm [ algorithm ] , pick an unoccupied lower vertex , and then execute the algorithm from step 3(b ) ; or pick an unoccupied upper vertex , and then execute the algorithm from step 3(a ) .
since all chock vertices have been unoccupied , the algorithm will fail to terminate , violating the fact that @xmath132 is finite .
[ nmn2mn ] @xmath300 * the `` @xmath301 '' direction : * for a representation @xmath161 of a minimal @xmath130-graph @xmath2 , apply algorithm [ algorithm ] to obtain a set @xmath236 of @xmath233 interconnecting paths . in the @xmath124-th iteration of the algorithm ,
the forward extension stops at the choke vertex of some @xmath224-alternating path .
let @xmath302 denote this alternating path .
note that after @xmath124 iterations , 1 ) all hubs of @xmath302 become occupied ; 2 ) by lemma [ distinctaasequence ] , each of the @xmath124 obtained interconnecting paths contains at most two hubs of @xmath302 ; 3 ) one obtained interconnecting path contains exactly one hub of @xmath302 .
hence , we deduce that the number of hubs of @xmath302 is at most @xmath303 , and therefore the total number of hubs of all @xmath224-alternating paths is at most @xmath304 . similarly , the total number of hubs of all @xmath217-alternating paths is at most @xmath305 as well .
again , by lemma [ distinctaasequence ] , the total number of hubs of all @xmath225 and @xmath226-alternating paths is at most @xmath306 . by lemma [ atmostmconnectingpaths ] , each hub in @xmath161 belongs to some interconnecting path in @xmath236 .
therefore , @xmath307 where the last inequality follows from @xmath234 . *
the `` @xmath170 '' direction : * we only need to construct a minimal @xmath130-graph @xmath2 with @xmath308 hubs ( see figure [ picn33 ] for an example ) .
the graph @xmath2 can be described as follows : @xmath309 is naturally oriented , and there is a set of @xmath92 vertex - disjoint paths @xmath310 from @xmath94 to @xmath95 and a set of @xmath126 vertex - disjoint paths @xmath311 from @xmath127 to @xmath128 .
furthermore , in @xmath312 , the directed version of @xmath2 , paths @xmath98 and @xmath313 meet at vertex @xmath314 and depart at vertex @xmath315 , for @xmath316 , @xmath317 , and * on path @xmath98 , @xmath318 * on path @xmath313 , @xmath319 -graph with @xmath320 hubs.,scaledwidth=40.0% ]
in this section , we are concerned with @xmath32 with more than two parameters , which turns out to be much more difficult to compute than @xmath32 with two parameters .
we will use an inductive argument on @xmath7 . notice that the case when @xmath323 has been established in theorem [ nmn2mn ] . by way of induction , we assume that the theorem is true for @xmath324 and proceed to prove it for @xmath325 . consider any minimal @xmath326-graph @xmath2 .
let @xmath60 denote the set of @xmath13 vertex - disjoint paths from @xmath11 to @xmath12 for @xmath327 .
after necessary rerouting of @xmath328 within @xmath329 , we can assume that @xmath329 is minimal and thus @xmath330 let @xmath16 denote the subgraph of @xmath329 induced on all the edges , each of which is simultaneously an @xmath328-edge , and let @xmath331 denote the number of connected components in @xmath16 .
obviously , each connected component of @xmath16 is in fact a path , and therefore we have @xmath332 . in this proof ,
we say a hub in @xmath2 is _ new _ if it is a hub in @xmath2 , however , not one in @xmath329 .
and we say a new hub is _ global _ , if this hub belongs to @xmath16 , _ local _ , if this hub is in @xmath333 . for each @xmath338-path ,
we `` cut '' it at each of its local new hubs and then `` divide '' the path into `` segments '' , each of which , evidently , is a subpath of the original @xmath339-path
. let @xmath340 denote the set of all obtained subpaths .
then we have @xmath341 . 1 .
add sources @xmath345 , sinks @xmath346 .
2 . for any @xmath347 and for any connected component of @xmath16 ,
whose natural @xmath60-direction is from one end vertex , say @xmath146 , to the other end vertex , say @xmath147 , add two directed edges @xmath348 and @xmath349 ( notice that a connected component may have opposite @xmath350-direction and @xmath351-direction for different @xmath352 , @xmath353 ) .
then , for any @xmath354 , we obtain a group @xmath355 of @xmath331 vertex - disjoint paths from @xmath356 to @xmath357 .
3 . for any @xmath340-path ,
whose natural @xmath339-direction is from one end vertex , say @xmath146 , to the other end vertex , say @xmath147 , add two directed edges @xmath358 and @xmath359 .
then , we obtain a group @xmath360 of @xmath361 vertex - disjoint paths from @xmath362 to @xmath363 .
obviously , the number of global new hubs in @xmath2 is just @xmath364 .
it follows from the minimality of @xmath2 and the observation that for any @xmath365 , any @xmath355-consistent cycle in @xmath344 naturally corresponds an @xmath60-consistent cycle in @xmath2 that @xmath344 is a minimal @xmath343-graph .
we then proceed to deduce that @xmath360 is non - reroutable in @xmath344 , since otherwise there exists an edge @xmath80 which is exclusively owned by @xmath360-paths ( this follows from a parallel argument leading to remark [ ifreroutable ] ) such that @xmath366 , violating the fact that @xmath344 is minimal .
now , after necessary rerouting of @xmath355 within @xmath344 , we assume that @xmath367 , the subgraph of @xmath344 consisting of all @xmath355-paths and @xmath360-paths , is non - reroutable and thus , by theorem [ equivalent ] , minimal .
notice that any hub @xmath49 in @xmath344 must be a hub of some @xmath367 , since otherwise @xmath344 contains an edge incident with @xmath49 that does not belong to any @xmath367 and therefore is not minimal .
hence , by theorem [ nmn2mn ] , we obtain that @xmath368 therefore , @xmath369 the proof is then complete . * the `` @xmath168 '' direction : * we will establish this direction using an inductive argument on @xmath7 .
suppose the inequality holds when @xmath375 .
consider a minimal @xmath376-graph @xmath2 with @xmath92 vertex - disjoint paths @xmath377 from @xmath94 to @xmath95 , @xmath126 vertex - disjoint paths @xmath378 from @xmath127 to @xmath128 , and a path @xmath379 from @xmath11 to @xmath12 for @xmath380 .
let @xmath140 be the subgraph of @xmath2 induced on @xmath381 now we split @xmath94 into @xmath92 copies @xmath382 ; @xmath95 into @xmath92 copies @xmath383 ; @xmath127 into @xmath126 copies @xmath384 ; @xmath128 into @xmath126 copies @xmath385 , such that @xmath98 has starting point @xmath386 and ending point @xmath387 for @xmath316 ; @xmath388 has starting point @xmath389 and ending point @xmath390 for @xmath391 .
let @xmath153 denote the resulting graph and @xmath331 be the number of weakly connected components ( which means connected components when disregarding the orientation ) in @xmath153 .
now for any @xmath392 , we identify @xmath386 and @xmath393 if they belong to the same component and we perform similar operations on @xmath387 s , @xmath389 s and @xmath390 s .
let @xmath16 denote the resulting graph , which consists of @xmath331 connected components .
note that the @xmath10-th component @xmath394 is in fact a minimal @xmath395-graph , where @xmath396 , @xmath397 and @xmath398 .
notice that for any @xmath10 , at least one of @xmath399 , @xmath400 and @xmath401 is nonzero , but some @xmath399 s , @xmath400 s , @xmath401 s can be zero , for which case , a @xmath395-graph can be interpreted as if all zero - valued parameters are simply dropped . for each component @xmath394 , by the induction hypothesis , we have a hub in @xmath2 is said to be _ new _ if it is not a hub in @xmath16 .
notice that the path @xmath406 `` meets '' each component at most once , yielding at most two new hubs ( more precisely , it meets one edge in @xmath394 , then departs from one edge in @xmath394 , and thereafter , it will never meet any edge in @xmath394 again ) , otherwise @xmath2 is not minimal .
hence , @xmath407\\ \le & \
2\omega+2(c_1+c_2+t-1-\omega)+2\left[\sum_{i=1}^\omega(c_{1,i}-1)^+\right]\cdot\left[\sum_{i=1}^\omega(c_{2,i}-1)^+\right]\\ \le & \ 2(c_1+c_2+t-1)+2(c_1 - 1)(c_2 - 1)\\ = & \ 2(c_1c_2+t).\end{aligned}\ ] ] it turns out the graph @xmath2 , described below in detail , is such a graph : @xmath2 is naturally orientable ; there is @xmath92 vertex - disjoint paths @xmath377 from @xmath94 to @xmath95 , @xmath126 vertex - disjoint paths @xmath378 from @xmath127 to @xmath128 , and a path @xmath379 from @xmath11 to @xmath12 for @xmath410 ; paths @xmath98 and @xmath313 meet at vertex @xmath314 and depart at vertex @xmath315 for @xmath316 and @xmath317 ; paths @xmath379 and @xmath411 meet at vertex @xmath412 and depart at vertex @xmath413 for @xmath410 .
furthermore , in @xmath312 , we have so we only need to prove the other direction . by contradiction , suppose that a minimal @xmath420-graph @xmath2 has at least @xmath422 hubs , and there exists a set of two vertex - disjoint path @xmath423 from @xmath94 to @xmath95 , a set of two vertex - disjoint paths @xmath424 from @xmath127 to @xmath128 , and a set of two vertex - disjoint paths @xmath425 from @xmath426 to @xmath427 . here , we note that when there are three pairs of sources and sinks , the equivalence statements as in theorem [ equivalent ] do not hold any more : it turns out that a minimal graph can be reroutable ( see figure [ n222minimalreroutable ] ) . *
we first consider the case @xmath2 is minimal and non - reroutable .
* let @xmath428 be the subgraph of @xmath2 induced on the edges of @xmath61-paths and @xmath71-paths ; similarly , we define @xmath429 and @xmath430 . by theorem [ equivalent ]
, these three subgraphs are all minimal , since they are all non - reroutable .
suppose @xmath428 has the most hubs among them .
every hub in @xmath2 belongs to at least one of these three subgraphs , so we have @xmath431 and hence @xmath432 .
now , we transform @xmath428 into a corresponding graph @xmath433 by shrinking each public edge into a vertex ( this can be regarded as the `` inverse '' operation of step 2 in section [ sectionminimalrepresentation ] ) . in more detail , for a public edge @xmath148 ,
say @xmath434 are two private edges incident with @xmath146 , and @xmath435 , @xmath436 are two private edges incident with @xmath147 .
then we delete edges @xmath137 , @xmath138 , @xmath144 , @xmath145 and vertices @xmath146 , @xmath147 , and then add a new vertex @xmath49 and edges @xmath437 , @xmath438 , @xmath439 , @xmath440 .
if @xmath428 has at least five hubs , then as shown in figure [ n222threecases ] , up to isomorphism , @xmath433 has three possibilities ( note that the first and second are different ) .
next , we examine in the ways one can add @xmath441-paths into @xmath442 to form @xmath2 such that @xmath2 is minimal . in the following
, a hub in @xmath2 is said to be @xmath443 if it is not a hub in @xmath428 , and an edge is said to be @xmath443 if it does not belong to @xmath428 and not incident with @xmath426 or @xmath427 .
* case 1 : * this case is shown in figure [ n222threecases](a ) , where each edge is labeled . since @xmath433 has four hubs , @xmath428 has at most eight hubs . and
since @xmath444 , @xmath445 , @xmath446 have to be added to generate at least @xmath447 new hubs . without loss of generality ,
say , @xmath445 contains at least three new hubs .
observe that each of these new hubs is incident with exactly one @xmath445-edge that does not belong to any @xmath61 or @xmath71-path , and @xmath426 and @xmath427 are also incident with one such @xmath445-edge .
so at least @xmath448 edges of @xmath445 are exclusively owned by @xmath445 , and thereby at least one of them , say @xmath449 , is a new edge ( not incident with @xmath426 or @xmath427 ) .
we next discuss the possible locations of @xmath449 ; here , notice that it is possible that either @xmath146 or @xmath147 is not a new hub .
suppose @xmath450 ( this means @xmath146 is one of vertices in @xmath451 , including its two end vertices ) .
then , one verifies that if @xmath452 or @xmath453 , then @xmath61 is reroutable ; if @xmath454 or @xmath455 , then @xmath71 is reroutable .
for example , if @xmath456 , then we can find another two vertex - disjoint @xmath71-paths @xmath457\circ \phi_2[a , v_1]\circ ( v_1,v_2)\circ \psi_2[v_2,r_2]\end{aligned}\ ] ] from @xmath127 to @xmath128 , and thus @xmath428 is reroutable . and
if @xmath458 , then we can find another two vertex - disjoint paths @xmath459\circ \psi_2[b , a]\circ \phi_2[a , r_1],\\ \phi_2'&=\phi_2[s_1,v_1]\circ ( v_1,v_2)\circ \phi_1[v_2,r_1]\end{aligned}\ ] ] from @xmath94 to @xmath95 , and thus @xmath428 is reroutable .
hence , we only need to check the last possible case : all new edges ( including @xmath148 ) are incident with public edges in @xmath428 . by symmetry ,
say , @xmath464 .
then , if @xmath465 or @xmath220 , @xmath71 is reroutable ; if @xmath466 , @xmath61 is reroutable ; if @xmath467 , there must be a new edge @xmath468 of @xmath446 with ( @xmath469 and @xmath470 ) or ( @xmath471 and @xmath472 ) ( see figure [ n222case1and2finalsentence](a)(b ) ) , since otherwise @xmath473 is still in @xmath474 , which contradicts the fact that @xmath2 is minimal .
however , in both cases , we can still find another two vertex - disjoint paths @xmath475\circ ( v_1',v_2')\circ \phi_2[v_2',r_1],\\ \phi_2'&=\phi_2[s_1,v_1]\circ ( v_1,v_2)\circ \phi_1[v_2,r_1]\end{aligned}\ ] ] from @xmath94 to @xmath95 , and thus @xmath428 is reroutable .
hence , we only need to check the last possible case : all new edges ( including @xmath468 ) are incident with public edges in @xmath428 . by symmetry ,
say @xmath464 or @xmath220 . if @xmath464 and @xmath483 or @xmath484 , then @xmath61 is reroutable ; if @xmath464 and @xmath485 , then @xmath71 is reroutable ; if @xmath486 and @xmath487 or @xmath484 , @xmath71 is reroutable ; if @xmath486 and @xmath466 , there must be a new edge @xmath468 of @xmath446 with @xmath469 and @xmath470 ( see figure [ n222case1and2finalsentence](c ) ) , since otherwise @xmath488 is still in @xmath474 , which contradicts the fact that @xmath2 is minimal .
however , we can still find another two vertex - disjoint paths @xmath489\circ ( v_1,v_2)\circ \phi_2[v_2,r_1],\\ \phi_2'&=\phi_2[s_1,v_2']\circ ( v_2',v_1')\circ \phi_1[v_1',r_1]\end{aligned}\ ] ] from @xmath94 to @xmath95 , and thus @xmath428 is reroutable .
* case 3 : * this case is shown in figure [ n222threecases](c ) , where each edge is also labeled . since @xmath433 has three hubs , @xmath428 has at most six hubs . and
since @xmath444 , @xmath445 , @xmath446 have to be added to generate at least @xmath490 new hubs .
so at least @xmath491 edges of @xmath445 and @xmath446 do not belong to any @xmath61 or @xmath71-path , and thereby at least two of them , say , @xmath148 and @xmath468 are two new edges .
we next discuss their possible locations .
let @xmath504 be the subgraph of @xmath2 induced on the edges of @xmath505 $ ] and @xmath506 $ ] and @xmath507 be the subgraph induced on the edges of @xmath508 $ ] and @xmath509 $ ] .
if @xmath510 , either @xmath61 or @xmath71 is reroutable , and hence @xmath511 or @xmath512 . similarly , we deduce that @xmath513 or @xmath514 .
then we just need to check the last possibility : @xmath515 and @xmath516 , for which we can also easily rule out the subcases when ( 1 ) @xmath517 and @xmath518 ( 2 ) @xmath519 and @xmath520 ( 3 ) @xmath521 ( 4 ) @xmath522 ( 5 ) @xmath523 and @xmath524 ( 6 ) @xmath525 and @xmath526 .
so , in the following , we examine the remaining two subcases : \(2 ) @xmath525 and @xmath524 . by symmetry , we assume @xmath527 on @xmath532 in @xmath529 . if @xmath530 , then @xmath441 is reroutable ; if @xmath531 , then @xmath71 is reroutable .
now , we are ready to conclude that any minimal and non - reroutable graph in @xmath474 has at most @xmath206 hubs .
* we next consider the case when @xmath2 is a minimal but reroutable @xmath420-graph*. suppose @xmath441 is reroutable in the subgraph @xmath430 . by theorem [ equivalent ] , there exists a @xmath441-consistent cycle @xmath84 where each @xmath85 is a subpath of some @xmath441-path and each @xmath46 is a private @xmath71-edge ( in @xmath533 ) .
we further assume that @xmath86 ( the number of private @xmath71-edges ) is the smallest among all possible @xmath441-consistent cycles in @xmath533 .
then , each @xmath46 belongs to a different @xmath71-path , which implies @xmath534 . since @xmath2 is minimal , each @xmath441-private edge ( in @xmath533 ) in @xmath535 must belong to @xmath61-paths as well .
so , each edge of the @xmath441-consistent cycle in @xmath2 belongs to either some @xmath61-path or @xmath71-path . note that for any edge @xmath80 of this cycle , @xmath536 does not belong to @xmath33 , since otherwise , together with the observation that @xmath441 can be rerouted using edges in @xmath537
, we deduce that @xmath538 , which contradicts the minimality of @xmath2 .
then we can find a minimal subgraph @xmath539 of @xmath428 , which is of either case 1 ( when @xmath540 ) or case 2 ( when @xmath541 ) .
let @xmath542 be the subgraph induced on the edges of @xmath441-paths .
then @xmath543 , by the minimality of @xmath2 .
notice that all contradictions in case 1 and case 2 arise from the rerouting of @xmath61 or @xmath71-paths .
so we can follow similar arguments and conclude @xmath544 .
the proof is then complete . | in this paper , a hub refers to a non - terminal vertex of degree at least three .
we study the minimum number of hubs needed in a network to guarantee certain flow demand constraints imposed between multiple pairs of sources and sinks .
we prove that under the constraints , regardless of the size or the topology of the network , such minimum number is always upper bounded and we derive tight upper bounds for some special parameters . in particular , for two pairs of sources and sinks , we present a novel path - searching algorithm ,
the analysis of which is instrumental for the derivations of the tight upper bounds . |
the electric @xmath6 and magnetic @xmath7 pion polarizabilities characterize the pion s deformation in an electromagnetic field , as occurs during @xmath8 compton scattering .
they depend on the rigidity of the pion s internal structure as a composite particle , and are therefore important dynamical quantities to test the validity of theoretical models . based on qcd chiral dynamics , the chiral perturbation theory effective lagrangian
, using data from radiative pion beta decay , predicts the pion electric and magnetic polarizabilities
@xmath9 = -@xmath10 = 2.7 @xmath11 0.4 , expressed in units of @xmath12 @xcite .
other theoretical predictions are also available @xcite .
the pion polarizabilities are usually investigated via their effect on the shape of the measured @xmath13 real compton scattering ( rcs ) angular distribution , as in ref . @xcite .
since pion targets are unavailable , pion rcs is approximated using different artifices , as shown in fig .
[ fig.polpos ] : the @xmath14 primakoff @xcite and @xmath15 radiative pion photoproduction reactions @xcite ; or by the crossing symmetry @xmath16 two - photon reaction @xcite . in the primakoff scattering ,
a high energy pion scatters from a ( virtual , practically real ) photon in the coulomb field of the target nucleus .
values of @xmath17 measured by these experiments are given in table [ tab : mespol ] .
.experimental values of @xmath17 .
[ cols="^,^,^",options="header " , ] the experimental and theoretical cdfs for s@xmath18 and r@xmath19 look similar . for s@xmath20 and q@xmath19 ,
some regions of q@xmath19 ( s@xmath20 ) have the experimental cdfs larger ( smaller ) than the theoretical cdfs .
the values of p(d ) ( see table [ tab : ks ] ) are sufficiently high for all five cdf s , as expected if the experimental data are consistent with theory . in addition
, the prediction of a total of 8 events is in agreement with the observed @xmath21 data events .
this further supports the conclusion from the k - s test that we observe pion vcs events . from the limited statistics and sensitivity of this first pion vcs experiment
, we can not determine the @xmath22 polarizability value , nor can we determine which value of @xmath22 is preferred . in a future experiment , the sensitivity to pion polarizability may be increased @xcite by achieving a data set in which the final @xmath3 ( @xmath23 ) has higher ( lower ) energies
. however , such data correspond to a lower cross section , and therefore require a high luminosity experiment .
the pion virtual compton scattering via the reaction @xmath24 is observed .
we developed and implemented a simulation with a vcs event generator .
we defined cuts that allow background reduction and vcs signal extraction .
the measured number of reconstructed pion vcs events , and their distributions with respect to the mandelstam invariants , are in reasonable agreement with theoretical expectations .
the corresponding @xmath23 vcs experimental cross section is @xmath4 , in agreement with the theoretical expectation @xmath5 .
the authors are indebted to the staffs of fermi national accelerator laboratory , the max planck
institut fr kernphysik , carnegie mellon university , petersburg nuclear physics institute and tel aviv university for invaluable technical support .
we thank drs .
c. unkmeir and s. scherer for the vcs matrix element calculation .
this project was supported in part by bundesministerium fr bildung , wissenschaft , forschung und technologie , consejo nacional de ciencia y tecnologa ( conacyt ) , conselho nacional de desenvolvimento cientfico e tecnolgico , fondo de apoyo a la investigacin ( uaslp ) , fundao de amparo pesquisa do estado de so paulo ( fapesp ) , the israel science foundation founded by the israel academy of sciences and humanities , istituto nazionale de fisica nucleare ( infn ) , the international science foundation ( isf ) , the national science foundation ( phy # 9602178 ) , nato ( grant cr6.941058 - 1360/94 ) , the russian academy of science , the russian ministry of science and technology , the turkish scientific and technological research board ( tbitak ) , the u.s .
department of energy ( doe grant de - fg02 - 91er40664 and doe contract number de - ac02 - 76cho3000 ) , and the u.s .- israel binational science foundation ( bsf ) .
the authors are indebted to the staffs of fermi national accelerator laboratory , the max planck
institut fr kernphysik , carnegie mellon university , petersburg nuclear physics institute and tel aviv university for invaluable technical support .
we thank drs .
c. unkmeir and s. scherer for the vcs matrix element calculation .
this project was supported in part by bundesministerium fr bildung , wissenschaft , forschung und technologie , consejo
nacional de ciencia y tecnologa ( conacyt ) , conselho nacional de desenvolvimento cientfico e tecnolgico , fondo de apoyo a la investigacin ( uaslp ) , fundao de amparo pesquisa do estado de so paulo ( fapesp ) , the israel science foundation founded by the israel academy of sciences and humanities , istituto nazionale de fisica nucleare ( infn ) , the international science foundation ( isf ) , the national science foundation ( phy # 9602178 ) , nato ( grant cr6.941058 - 1360/94 ) , the russian academy of science , the russian ministry of science and technology , the turkish scientific and technological research board ( tbitak ) , the u.s .
department of energy ( doe grant de - fg02 - 91er40664 and doe contract number de - ac02 - 76cho3000 ) , and the u.s .- israel binational science foundation ( bsf ) . deceased present address : infineon technologies ag , mnchen , germany now at imperial college , london sw7 2bz , u.k .
now at instituto de fsica da universidade estadual de campinas , unicamp , sp , brazil now at physik - department , technische universitt mnchen , 85748 garching , germany present address : the boston consulting group , mnchen , germany present address : fermilab , batavia , il now at instituto de fsica terica da universidade estadual paulista , so paulo , brazil present address : spss inc . ,
chicago , il now at university of alabama at birmingham , birmingham , al 35294 present address : medson ltd .
, rehovot 76702 , israel present address : doe , germantown , md present address : siemens medizintechnik , erlangen , germany present address : deutsche bank ag , eschborn , germany 0.25 cm j. portales , m. r. pennington , hep - ph/9407295 , + d. morgan and m. r. pennington , phys .
* b272 * ( 1991 ) 134 .
d. babusci , s. bellucci , g. giordano , g. matone , a.m. sandorfi , m.a .
moinester , phys . lett . * b277 * , 158 ( 1992 ) . b. r. holstein , comments nucl .
* 19 * ( 1990 ) 221 .
a. klein , phys . rev .
* 99 * ( 1955 ) 988 ; + a. m. baldin , nucl .
phys . * 18 * ( 1960 ) 310 ; + v. a. petrunkin , sov . jetf . * 40 * ( 1961 ) 1148 .
m. antipov et al . , phys .
let . * b121 * ( 1983 ) 445 ; + yu . m. antipov et al . , z.phys.c * 26 * ( 1985 ) 495
. t. a. aibergenov , p. s. baranov , o. d. beznisko , s. n. cherepniya , l. v. filkov , a. a. nafikov , a. i. osadchii , v. g. raevsky , l. n. shtarkov , e.i .
tamm , czech .
* 36 * , 948 ( 1986 ) .
j. boyer et al . , phys . rev .
* d42 * ( 1990 ) 1350 .
s. scherer , few body syst . suppl .
* 11 * ( 1999 ) 327 ; czech .
* 49 * ( 1999 ) 1307 . c. unkmeir , s. scherer , a.i .
lvov , d. drechsel , phys . rev . *
d61 * , 034002 ( 2000 ) .
lvov , s. scherer , b. pasquini , c. unkmeir , d. drechsel , phys . rev . *
c64 * , 015203 ( 2001 ) .
s. kerhoas et al .
phys . * a666 - 667 * ( 2000 ) 44 , + j. roche et al .
* 85 * ( 2000 ) 708 .
b. pasquini , s. scherer , d. drechsel , phys . rev . *
c63 * , 025205 ( 2001 ) ; b. pasquini , s. scherer , d. drechsel , nucl - th/0105074 ; + and references therein .
j. russ et al . , a proposal to construct selex , proposal pe781 , fermilab , 1987 , http://fn781a.fnal.gov/. c. unkmeir , a. ocherashvili , t. fuchs , m.a .
moinester , s. scherer , phys . rev . * c65 * ( 2002 ) 015201 .
+ a. ocherashvili , pion virtual compton scattering , ph.d .
thesis , sept .
2000 , tel aviv university , http://muon.tau.ac.il/@xmath25aharon/phd.html . c. unkmeir , pion polarizabilities in the reaction of radiative pion photoproduction on the proton , ph.d .
thesis , johannes gutenberg university , mainz , sept . 2000 ; + t. fuchs , b. pasquini , c. unkmeir , s. scherer , hep - ph/0010218 .
lepage , j. comp .
* 27 * ( 1978 ) 192 .
j. d. bjorken , s. d. drell , relativistic quantum mechanics , mcgraw - hill , new york , 1964 .
e. byckling k. kajantie , _ particle kinematics _ , state , johnwiley & son ltd , 1973 , p. 102
g. davidenko et al . , ge781 : a monte carlo package for fixed target experiments , in : r. shellard and t. d. nguyen , eds . , _ proceedings of the international conference on computing in high energy physics95 _ ( world scientific , singapore , 1996 ) p.832 , + g. dirkes , h. kr@xmath26ger , n. p. kuropatkin , j. simon , data and ge781 comparison for elastic scattering physics , mpi heidelberg , 1999 ( unpublished ) , + v. steiner , ge781 code development for hadron electron topics , selex collaboration meeting presentations and private communications , + i. giller , measurement of the pion charge radius , m.sc .
thesis , tel aviv university , 1999 , + i. eschrich et al .
* 86 * ( 2001 ) 5243 . c. caso et al .
( pdg ) , eur .
j. * c3 * ( 1998 ) 1 - 794 .
s. brandt , _ data analysis : statistical and computational methods for scientists and engineers _ , springer verlag , 1998 .
j. simon , messung des elektromagnetischen ladungsradius des sigma- bei @xmath1 ( in german ) , ph.d .
thesis , mpi heidelberg , 2000 , + g. dirkes , messung der elektromagnetischen formfaktoren von pionen i m selex experiment ( in german ) , m.sc .
thesis , mpi heidelberg , 1999 .
k. vorwalter , determination of the pion charged radius with a silicon microstrip detector system , ph.d .
thesis , mpi f. kernphysik / univ . heidelberg , 1998 , + h. kr@xmath26ger , untersuchung der elastischen hadron - elektron - streuung bei 540 gev / c zur messung des elektromagnetischen ladungsradius des protons ( in german ) , ph.d .
thesis , mpi heidelberg , 1999 .
p. mathew , construction and evaluation of a high resolution silicon microstrip tracking detector , and utilization to determine interaction vertices , ph.d .
thesis , carnegie mellon university , 1997 , + j. l. langland , hyperon and anti - hyperon production in @xmath27 interactions .
thesis , university of iowa , 1996 , umi-96 - 03058 , + j. l. langland , hyperon beam flux parameterization for e781 based on e497 data , h - note 693 , selex internal report , 1994 .
w. h. press , s. a. teukolsky , w. t. vetterling , b. p. flannery , _ numerical recipes in fortran : the art of scientific computing _ , second edition , university of cambridge , 1992 , p. 617 - 622 . | pion virtual compton scattering ( vcs ) via the reaction @xmath0 was observed in the fermilab e781 selex experiment .
selex used a @xmath1 @xmath2 beam incident on target atomic electrons , detecting the incident @xmath2 and the final state @xmath2 , electron and @xmath3 .
theoretical predictions based on chiral perturbation theory are incorporated into a monte carlo simulation of the experiment and are compared to the data .
the number of reconstructed events ( 9 ) and their distribution with respect to the kinematic variables ( for the kinematic region studied ) are in reasonable accord with the predictions .
the corresponding @xmath2 vcs experimental cross section is @xmath4 , in agreement with the theoretical expectation @xmath5 . |
a number of experimental facilities to significantly improve on present sensitivity to the connecting mixing @xmath3 in neutrino oscillations and have access to the cp - violating phase @xmath4 have been discussed in the literature : neutrino factories ( neutrino beams from boosted - muon decays ) , superbeams ( very intense conventional neutrino beams ) , improved reactor experiments and more recently @xmath5-beams @xcite .
the original standard scenario for beta beams with lower @xmath6 and short baseline @xmath7 km from cern to frejus with @xmath8 and @xmath9 ions has seen a variant by using an electron capture facility for monochromatic neutrino beams @xcite .
new proposals also include the high @xmath10 value @xmath11 and @xmath12 isotopes in a @xmath13 facility @xcite . in this paper
we discuss the physics reach that a high energy facility for ec beams may provide with the expected sps upgrade at cern .
such study has been made for the standard beta beam facility @xcite . in section 2
we discuss the virtues of the suppressed oscillation channel @xmath14 in order to have access to the parameters @xmath3 and @xmath4 .
the interest of energy dependence , as obtainable in the ec facility , is emphasized . in section 3
we present new results on the comparison between ( low energies , short baseline ) and ( high energies , long baseline ) configurations for an ec facility with a single ion .
section 4 gives our conclusions .
the observation of @xmath15 violation needs an experiment in which the emergence of another neutrino flavour is detected rather than the deficiency of the original flavour of the neutrinos . at the same time , the interference needed to generate cp - violating observables can be enhanced if both the atmospheric and solar components have a similar magnitude .
this happens in the suppressed @xmath16 transition .
the appearance probability @xmath17 as a function of the distance between source and detector @xmath18 is given by @xcite @xmath19 where @xmath20 .
the three terms of eq .
( [ prob ] ) correspond , respectively , to contributions from the atmospheric and solar sectors and their interference .
as seen , the @xmath15 violating contribution has to include all mixings and neutrino mass differences to become observable .
the four measured parameters @xmath21 and @xmath22 have been fixed throughout this paper to their mean values @xcite . from general arguments of cpt invariance and absence of absorptive parts
the cp - odd probability is odd in time and then odd in the baseline l ( formally ) .
vacuum oscillations are only a function of @xmath23 so that , at fixed @xmath24 , the cp - odd probability is odd in the energy ( formally ) .
this proves that the study of neutrino oscillations in terms of neutrino energy will be able to separate out the cp phase @xmath4 from the mixing parameters .
a control of this energy may be obtained from the choice of the boost in the ec facility with a single ion . in the beams with a continuous spectrum ,
the neutrino energy has to be reconstructed in the detector . in water - cerenkov detectors ,
this reconstruction is made from supposed quasielastic events by measuring both the energy and direction of the charged lepton .
this procedure suffers from non - quasielastic background , from kinematic deviations due to the nuclear fermi momentum and from dynamical suppression due to exclusion effects @xcite . for ec decay ,
the rare - earth nuclei above @xmath25 have a small enough half - life for electron capture processes .
this is in contrast with the proposal of ec beams with fully stripped long - lived ions @xcite .
we discuss the option of short - lived ions @xcite .
the neutrinos emerging from a boosted ion beam decaying by ec are concentrated inside a narrow cone around the forward direction . if the ions are kept in the decay ring longer than the half - life , the energy distribution of the neutrino flux arriving to the detector in absence of neutrino oscillations is given by the master formula @xmath26 with a dilation factor @xmath27 .
it is remarkable that the result is given only in terms of the branching ratio and the neutrino energy and independent of nuclear models . in eq .
( [ master ] ) , @xmath28 is the total number of ions decaying to neutrinos . as a result ,
such a facility will measure the neutrino oscillation parameters by changing the @xmath29 s of the decay ring ( energy dependent measurement ) and there is no need of energy reconstruction in the detector . for the study of the physics reach associated with such a facility , we combine two different energies for the same @xmath30 ion using two setups . in all cases we consider @xmath31 decaying ions / year , a water cerenkov detector with fiducial mass of @xmath32 kton and both appearance ( @xmath33 ) and disappearance ( @xmath34 ) events .
setup i is associated with a five year run at @xmath0 ( close to the minimum energy to avoid atmospheric neutrino background ) plus a five year run at @xmath1 ( the maximum energy achievable at present sps ) , with a baseline @xmath7 km from cern to frejus .
the results for setup i are going to be compared with those for setup ii , associated with a five year run at @xmath1 plus a five year run at @xmath2 ( the maximum achievable at the upgraded sps with proton energy of @xmath35 gev ) , with a baseline @xmath36 km from cern to canfranc . for the setup
i we generate the statistical distribution of events from assumed values of @xmath3 and @xmath4 .
the corresponding fit is shown in fig .
[ fit - setupi ] for selected values of @xmath3 from @xmath37 to @xmath38 and covering a few values of the cp phase @xmath4 .
as observed , the principle of an energy dependent measurement ( illustrated here with two energies ) is working to separate out the two parameters . with this configuration
the precision obtainable for the mixing ( even at 1 degree ) is much better than that for the cp phase .
the corresponding exclusion plot which defines the sensitivity to discover cp violation @xmath39 is presented in fig .
[ fit - ec - des ] .
for @xmath40 cl the sensitivity to see cp violation becomes significant for @xmath41 . in the case of setup
ii the longer baseline for @xmath1 leads to a value of @xmath23 well inside the second oscillation ( see fig . [ proba ] ) . in that case the associated strip in the ( @xmath3 , @xmath4 ) plane has a more pronounced curvature , so that the two parameters can be better disantangled .
the statistical distribution generated for some assumed values of ( @xmath3 , @xmath4 ) has been fitted and the @xmath42 values obtained .
the results are given in fig .
[ fit - ec ] .
qualitatively , one notices that the precision reachable for the cp phase is much better than that for setup i. one should emphasize that this improvement in the cp phase has been obtained with the neutrino channel only , using two appropriate different energies . the corresponding exclusion plot which defines the sensitivity to discover cp violation @xmath39
is presented in fig .
[ fit - ec - des ] .
for @xmath40 cl the sensitivity to see cp violation becomes now significant even for values of @xmath3 smaller than those for setup i. at the time of the operation of this proposed facility it could happen that the connecting mixing @xmath3 is already known from the approved experiments for second generation neutrino oscillations , like double chooz , t2k and nova . to illustrate the gain obtainable in the sensitivity to discover cp violation from the previous knowledge of @xmath3 we present in fig .
[ fit - ec - con ] the expected sensitivity with the distribution of events depending on a single parameter @xmath4 for a fixed known value of @xmath3 .
the result is impressive : even for a mixing angle of one degree , the cp violation sensitivity at @xmath40 cl reaches values around @xmath43 for setup ii .
the simulations of the physics output for ec beams indicate : 1 ) the upgrade to higher energy ( @xmath44 gev ) is crucial to have a better sensitivity to cp violation , which is the main objective of the next generation neutrino oscillation experiments , iff accompanied by a longer baseline ; 2 ) the best @xmath23 in order to have a higher sensitivity to the mixing @xmath45 is not the same than that for the cp phase . like the phase - shifts ,
the presence of @xmath4 is easier to observe when the energy of the neutrino beams enters into the region of the second oscillation .
the mixing is better seen around the first oscillation maximum , instead . in particular ,
setup ii in ec beams , i.e. , with @xmath46 between @xmath47 and @xmath32 and a baseline @xmath36 km ( canfranc ) , has an impressive sensitivity to cp violation , reaching a sensitivity around @xmath48 , for @xmath40 cl , or better ( if some knowledge on the value of @xmath3 is previously established ) .
besides the feasibility studies for the machine , most important for physics is the study of the optimal configuration by combining low energy with high energy neutrino beams , short baseline with long baseline and/or ec monochromatic neutrinos with @xmath8 @xmath49 antineutrinos . | the goal for future neutrino facilities is the determination of the u(e3 ) mixing and cp violation in neutrino oscillations .
this will require precision experiments with a very intense neutrino source and energy control . with this objective in mind ,
the creation of monochromatic neutrino beams from the electron capture decay of boosted ions by the sps of cern has been proposed .
we discuss the capabilities of such a facility as a function of the energy of the boost and the baseline for the detector .
we conclude that the sps upgrade to 1000 gev is crucial to reach a better sensitivity to cp violation iff it is accompanied by a longer baseline .
we compare the physics potential for two different configurations : i ) @xmath0 and @xmath1 ( maximum achievable at present sps ) to frejus ; ii ) @xmath1 and @xmath2 ( maximum achievable at upgraded sps ) to canfranc .
the main conclusion is that , whereas the gain in the determination of u(e3 ) is rather modest , setup ii provides much better sensitivity to cp violation . |
the great regularity of the tully - fisher relation @xcite has long been thought to originate from a strong mass - velocity relation and a near constancy of mass - to - light ratio .
the latter requires a fair but not unreasonable amount of regularity in stellar populations .
put simply , ... when it comes to pepper grinders '' ( van den bosch 2001 , private communication ) . ]
@xmath0 there have long been indications ( sancisi 1995 , private communication ) that this simple scaling may fail at low luminosities .
this has become more clear as data have improved @xcite,@xcite .
this breakdown of the tully - fisher relation might arise because of the chaotic star formation histories of low mass galaxies , or as a result of a breakdown in the underlying mass - velocity relation .
another possibility is that optical luminosity ceases to trace mass because stars cease to be the dominant mass component in these disks @xcite .
it has now become clear that this last possibility is in fact the case .
low mass galaxies are often dominated by gas rather than stars .
if instead of luminosity or stellar mass , we plot disk ( star + gas ) mass against the flat rotation velocity , a nice mass - velocity relation is recovered over many orders of magnitude ( fig . 1 ) .
this ` baryonic tully - fisher relation ' ( btf ) is @xcite @xmath1 for which the data in fig . 1
give a = 50_^-4 ^ 4 + b = 4.0 0.1 .
the normalization of the btf is rather uncertain : formally acceptable values fall in the range @xmath2 .
the precise value of the slope has been modestly controversial : @xmath3 was given by @xcite while @xmath4 was found by @xcite .
this difference can be traced to different assumptions about the ( rather goofy @xcite ) distance to the uma cluster for which some of the better rotation curve @xcite and photometric data @xcite exist . as the distance increases , the gas mass increases faster than the stellar mass ( as @xmath5 and as @xmath6 , respectively ) .
this boosts the total mass of gas dominated galaxies by a larger factor than star dominated galaxies . since these reside at opposite ends of the relation , the slope tips to shallower values with increasing @xmath6
nevertheless , the population models of @xcite are consistent with a slope of @xmath3 ( fig . 2 ) .
while the calibration of the btf can always be improved , it already provides an excellent indicator of disk mass .
moreover , continuity between gas - rich and star - rich galaxies constrains stellar population mass - to - light ratios .
the favored values are reasonable in terms of population synthesis models ( fig .
2 ) , but unpleasantly heavy for cuspy dark matter halos . -band and ( * b * ) the @xmath7-band predicted by a slope 4 btf for the uma galaxies @xcite , @xcite , @xcite . these are plotted as a function of @xmath8 color , together with the bruzual & charlot , salpeter imf model from @xcite ( the first model in their table 4 ) .
the population synthesis models are in good agreement with the btf , indicating that we have a good handle on @xmath9 and disk masses . ]
rotation curves , by themselves , can only give a lower limit on the total halo mass : that enclosed by the last measured point .
however , if the functional form of the halo were known , it might be possible to provide some constraint by fitting the observations to the known form .
the nfw halo paradigm @xcite,@xcite which has arisen from cosmological n - body simulations in principle gives a way to do this .
unfortunately , if not surprisingly , observed rotation curves never extend far enough to constrain the circular velocity at the virial radius , @xmath10 @xcite .
there is a great deal of degeneracy between the concentration @xmath11 and @xmath10 .
an example is given in fig .
3 , which shows how difficult it can be to distinguish between fits with nfw halos of rather different parameters . and
@xmath12 for @xmath13 .
another tolerable fit with @xmath14 and @xmath15 is also shown ( lower dotted curve ) to illustrate the degeneracy between parameters .
though many models sort of fit , their concentrations are implausibly low for @xmath16cdm . ]
matters are made worse by the general failure of the nfw form to provide a good description of the data .
the data just do nt look like nfw halos .
statistically , halos with constant density cores are almost always preferred over those with cusps @xcite .
this is most clear in the best resolved cases @xcite .
the most important systematic concern at this point is not observational .
resolution has improved by an order of magnitude @xcite , @xcite , @xcite over the original 21 cm data for lsb galaxies @xcite , @xcite .
the nfw shape has not become apparent as the data have improved .
instead , the systematics pointed out by @xcite,@xcite as problematic for cdm ( independent of the cusp issue ) have only become more clear .
concern over slit mispositioning @xcite are misplaced : independent observers reproduce one anothers results @xcite,@xcite . while there are certainly cases in which the error bars are large enough to allow an nfw fit , isothermal fits are inevitably better . simply changing the size of the error bars wo nt change this : a systematic change in the shapes of @xmath17 high resolution rotation curves is required .
one can certainly imagine ways in which this might happen @xcite , but it is extremely unlikely that any of these ideas apply to real data , let alone to * all * of the data from various independent sources .
the most serious issue is in the mass models : stars have mass . even in the limit of zero stellar mass , which is the most favorable to the nfw case , isothermal halos are statistically preferred @xcite,@xcite . the situation only becomes more grim if stars are allowed to have mass .
though lsb galaxies are dark matter dominated down to small radii , plausible @xmath9 models do require that _ some _ of the velocity be attributed to luminous mass .
this pulls the inferred dark matter distribution further away from the expected cusp slope .
we are hardly unique in reaching these conclusions , which are shared by * all * published analyses of high resolution long slit h@xmath18 data @xcite,@xcite,@xcite,@xcite,@xcite,@xcite,@xcite .
high resolution fabry - perot @xcite,@xcite,@xcite and co @xcite data are also inconsistent with cuspy halos , as are a variety of data for the milky way itself @xcite .
the only analyses which are favorable to nfw are those of low resolution data with large error bars @xcite,@xcite . when the error bars are large , any model can be driven through them .
though it has not been emphasized , constant density cores provide as good or better fits even in these cases . the isothermal halo form , while effective , is an extremely flexible fitting function which lacks the motivation of the nfw halo form .
so one might persists that the nfw fits are still more appropriate in that they can be related to cosmology .
standard @xmath16cdm makes a clear prediction @xcite for what the concentrations of dark matter halos should be : @xmath19 for @xmath20 .
scatter about this value should be modest the largest estimate @xcite finds a lognormal distribution with @xmath21 .
the median observed concentration is @xmath22 @xcite which is different from the standard @xmath16cdm prediction by many @xmath23 .
the problem with nfw halos is not just a matter of getting fits to individual galaxies , but also of understanding how the observed concentrations can be so low .
these low concentrations would be tolerable in a very low density universe with @xmath24 @xcite .
even then there exists a significant tail of very low concentration ( @xmath25 ) galaxies which simply should not exist for any plausible cosmology .
the debate over halo profiles , while contentious , misses the real point .
many halo profiles are nominally viable because they have lots of degenerate free parameters .
mass modeling is a bit like fitting a high order polynomial to a few data points : the line goes through the data , but means nothing .
one would prefer to have a minimal parameter description of the data .
such a prescription exists @xcite .
it has long been noted that there is a strong coupling between mass and light .
oddly , this coupling persists for dark matter dominated lsb galaxies .
one needs only a single parameter per galaxy , the stellar mass - to - light ratio , in order to fit the rotation curve in comparable or greater detail than can be matched by many - parameter halo models .
the mass - to - light ratio in the @xmath26-band is close enough to constant that one can make a good zero parameter prediction with such data @xcite . until we come to terms with this observed phenomenology , debating the cusp slope of dark matter halos
is rather akin to debating the number of angels that can dance on the head of a pin .
+ * acknowledgements : * i am most grateful to vera rubin , erwin de blok , and albert bosma for their work on the issues discussed here .
i would also like to thank renzo sancisi , marc verheijen , rob swaters , and frank van den bosch for many lively and stimulating conversations .
no doubt , i have yet to hear the end of it !
the work of ssm is supported in part by nsf grant ast9901663 .
freeman : ` historical introduction ' . in : _ the low surface brightness universe , iau colloquium 171 _ , ed .
davies , c. impey , s. phillipps ( astronomical society of the pacific , san francisco 1999 ) pp . 3 - 8
mcgaugh : ` dynamical constraints on disk galaxy formation ' . in : _
galaxy dynamics : from the early universe to the present _ , ed .
f. combes , g.a .
mamon , v. charmandaris ( astronomical society of the pacific , san francisco 2000 ) pp . | i review what we currently do and do not know about the masses of disk galaxies and their dark matter halos .
the prognosis for disks is good : the asymptotic rotation velocity provides a good indicator of total disk mass .
the prognosis for halos is bad : cuspy halos provide a poor description of the data , and the total mass of individual dark matter halos remains ill - constrained . |
in the 1970s works by penrose and ammann showed that the euclidean space can be filled gapless and non - overlapping by two or more tiles which are arranged in a nonperiodic way according to matching rules @xcite .
these quasiperiodic tilings showed a hierarchical structure and long - range oriental order .
the study of such tilings was strongly intensified after the discovery of quasicrystals by shechtman et al . in 1982
these materials are characterized by a perfect long range order without having a three - dimensional translational periodicity .
the former is manifested by the occurrence of sharp spots in the diffraction pattern and the latter in the occurrence of rotational symmetries forbidden for conventional crystals @xcite .
while the diffractions patterns of quasicrystals were not explainable by the structure of crystals with their periodic repetition of unit cells , the mathematical concept of quasiperiodic tilings was suitable to describe the structure of this new material class @xcite .
various experimental investigations revealed rather exotic physical properties .
for instance quasicrystalline surfaces are anti - adhesive in combination with a high level of hardness making them suitable for the production of coatings of cookware , engines , tools , etc . @xcite .
further , they possess a rather low thermal and electrical conductance although they contain a high amount of well - conducting elements @xcite . this motivated extensive research to obtain a better theoretical understanding of the structure and the physical properties of these materials .
today several exact results are known for one - dimensional quasicrystals @xcite .
however , the characteristics in two or three dimensions have been clarified to a much lesser degree and are mainly based on numerical studies . due to the lack of translation symmetry in quasicrystals , one has to study very large approximants to obtain an insight into the physical properties of macroscopic quasicrystals .
in addition many physical properties in quasiperiodic systems show a power - law behaviour with respect to time or system size , which indicates the absence of a typical time / length scale in the system . that behavior is in correspondence with the self - similarity of quasicrystalline structures , what allows scaling between the various hierarchic levels of the tilings .
the resulting scaling exponents provide an insight into the properties of the macroscopic systems .
such a scaling approach was also used for the numerical determination of the electronic properties for the penrose tiling , the ammann - kramer - neri tiling , or octagonal tiling models @xcite .
however , all these calculations were performed for relatively small systems , making it difficult to determine the exact scaling behaviour . to obtain a better insight into the properties of quantum - mechanical wave functions and the quantum diffusion in quasicrystals , we study here models of @xmath0-dimensional quasicrystals with a separable hamiltonian in a tight - binding approach @xcite .
this method is based on the fibonacci sequences , which describes the weak and strong couplings of atoms in a quasiperiodic chain .
the fibonacci chain is the one - dimensional analogue of the penrose tiling and the icosahedral quasicrystal @xcite and , hence , it is related to quasicrystals with five - fold and ten - fold rotational symmetries .
our higher - dimensional tilings are then constructed as a direct product of these chains .
also the eigenstates of these tilings can be directly calculated by multiplying the wave functions and by adding / multiplying the energies of the quasiperiodic chains , which yields the hypercubic tiling or the labyrinth tiling , respectively .
this approach allows us to study very large systems with up to @xmath1 sites in three dimensions based on the solutions in one dimension . the quantum diffusion in the fibonacci chain and the corresponding hypercubic tiling has been studied quite intensively in the past @xcite .
further , in our previous work we obtained more insight into the electronic eigenstates and the transport properties of the labyrinth tiling .
for instance , we showed that the product structure of the labyrinth tiling leads to a connection of the multifractal properties of the wave functions in different dimensions @xcite . in recent years also some new properties for one - dimensional quasiperiodic chains were found .
these include the step - like behaviour of the wave - packet dynamics of metallic - mean chains reported by us @xcite and the discovery of log - time - periodic oscillations of wave packets in the fibonacci chain @xcite .
in this paper we report new results for the multifractal properties of wave functions in quasiperiodic systems and for the return probability of the wave packets .
in particular , we study the participation numbers which provide an insight into the distribution of the probability density of electronic wave functions and allow us to distinguish extended and localized states .
this can be related to the electronic transport properties , because e.g. bloch functions indicate ballistic transport and localized wave functions are associated with insulators . to describe the electronic transport properties more quantitatively two quantities
are usually studied : the mean square displacement and the temporal autocorrelation function .
the scaling exponent @xmath2 of the mean square displacement is connected to the electrical conductivity of a quasicrystal according to the generalized drude equation @xcite . in our recent work
we showed that the scaling exponent @xmath2 can be determined analytically for the labyrinth tiling in the regime of strong quasiperiodic modulation by a renormalization group ( rg ) approach @xcite . in the present paper we focus on the temporal autocorrelation function , which is directly related to the properties of the ( local ) density of states .
further , it describes the decay of the wave packet which can be linked to the quantum diffusion of electrons in the system .
we also show that for systems with absolute continuous energy spectra an improved lower bound for the exponent @xmath2 can be obtained , which is not restricted to the systems studied here .
the paper is structured in the following way . in sec . [
sec : state - of - art ] we briefly describe the construction rules for the @xmath0-dimensional separable tilings .
we also introduce the participation numbers and the temporal autocorrelation function and give an overview of known results for these quantities in quasiperiodic systems . in section [ sec : rg - approach ] we briefly describe the rg approach , which is applied in the following two sections .
this is followed by the presentation of new results for the properties of the wave functions in these systems in sec .
[ sec : wavefunctions ] .
we also apply the rg approach to show that the wave functions do not become localized even for very small coupling strengths . in sec .
[ sec : returnprob ] we study the wave - packet dynamics with a special focus on the return probability with numerical methods and with the rg approach .
further , the lower bounds for the scaling exponent @xmath2 of the mean square displacement are discussed in sec .
[ subsec : lower - bound ] , and the results are briefly summarized in sec .
[ sec : conclusion ] .
the construction of the separable quasiperiodic tilings is based on the fibonacci chain defined by the inflation rule @xmath3 . starting with the symbol
@xmath4 we obtain after @xmath5 iterations the @xmath5th order approximant @xmath6 of the fibonacci chain .
the length @xmath7 of an approximant @xmath6 is given by the recursive rule @xmath8 with @xmath9 .
further , the ratio of the lengths of two successive approximants as well as the ratio of the numbers @xmath10 of occurrence of the symbols @xmath11 and @xmath4 in an approximant approach the golden mean @xmath12 for @xmath13 .
thus , we obtain with the continued fraction representation @xmath14 = [ 1,1,1, ... ]$ ] the relations @xmath15 our model describes an electron hopping from one vertex of a quasiperiodic chain to a neighboring one .
the aperiodicity is given by the underlying quasiperiodic sequence of couplings , where the symbols @xmath11 and @xmath4 denote the weak and strong bonds of the chain .
solving the corresponding time - independent schrdinger equation @xmath16 for the fibonacci chain with zero on - site potentials , we obtain discrete energy values @xmath17 and wave functions @xmath18 represented in the orthogonal basis states @xmath19 associated to a vertex @xmath20 .
if not stated differently we use free boundary conditions , which yields @xmath21 eigenstates .
the hopping strength @xmath22 in the schrdinger equation is given by the fibonacci sequence @xmath6 with @xmath23 for a strong bond and @xmath24 for a weak bond ( @xmath25 ) .
for @xmath26 only isolated clusters of single vertices or two strongly coupled vertices occur , and for @xmath27 the system is periodic .
the @xmath0-dimensional separable quasiperiodic tilings are then constructed from the product of @xmath0 quasiperiodic chains which are perpendicular to each other . for this setup
two special cases are known for which the systems become separable @xcite : _ hypercubic tiling @xmath28 _
this tiling corresponds to the usual euclidean product of @xmath0 linear quasiperiodic chains , i.e. , only vertices connected by vertical and horizontal bonds interact as shown in figure [ fig : tilings ] .
such a tiling is referred to as a square tiling in two dimensions and in general as a hypercubic tiling @xcite . in literature
this model has been studied mainly for the fibonacci sequence @xcite , and for reasons of comparison we will also present a few results of this model . _
labyrinth tiling @xmath29 _ in the other case only coupling terms to neighbours along the diagonal bonds are considered ( cf .
figure [ fig : tilings ] ) , where the bond strengths of this tiling equal the product of the corresponding bond strengths of the one - dimensional chains .
the resulting grid is denoted as labyrinth tiling .
( left ) and labyrinth tiling @xmath30 ( right ) constructed from two fibonacci chains @xmath31.,title="fig:",width=151 ] ( left ) and labyrinth tiling @xmath30 ( right ) constructed from two fibonacci chains @xmath31.,title="fig:",width=151 ] the hamiltonians of these tilings are separable .
thus , the eigenstates of the tilings in @xmath0 dimensions can be constructed from the eigenstates of @xmath0 one - dimensional chains . in both cases
the wave functions of the higher - dimensional tiling are constructed as the product of the one - dimensional wave functions , i.e. , @xmath32 the superscripts @xmath33 enumerate the eigenvalues @xmath34 , and the subscript @xmath35 represents the coordinates of the vertices in the quasiperiodic tiling .
however , the energies are treated in a different way for the two cases . in @xmath0 dimensions
the eigenenergies are given by @xmath36 for the hypercubic tiling and by @xmath37 for the labyrinth tiling .
while for the hypercubic tiling all combinations of the indices @xmath38 are allowed , for the labyrinth tiling only certain sets yield the correct solution due to symmetries of the eigenfunctions @xcite . for the one - dimensional quasiperiodic systems the energy spectrum is known to be singular continuous @xcite .
the properties of the energy spectrum for the hypercubic tiling and the labyrinth tiling can be obtained from studying the scaling behaviour of the width of the energy bands @xcite .
the numerical results for both tilings provided evidence for a transition from a singular continuous to an absolutely continuous energy spectrum upon increasing the coupling parameter @xmath11 including an intermediate region where both parts coexist @xcite .
the transition parameters depend on the dimension and the underlying quasiperiodic structure . in particular , we like to point out the threshold values @xmath39 in two dimensions and @xmath40 in three dimensions above which the spectrum of the hypercubic tiling and the labyrinth tiling becomes absolutely continuous @xcite .
these transition parameters @xmath41 seem to correspond to the threshold values @xmath42 for which the density of states becomes gapless in two and three dimensions @xcite .
for @xmath26 the isolated clusters lead to a pure point spectrum and the periodic systems for @xmath43 possess an absolutely continuous spectrum regardless of the dimension .
further , it was shown analytically that the energy spectrum of the square tiling is singular continuous in the regime of strong quasiperiodic modulation @xcite .
another interesting property is that the energy spectrum of the fibonacci chain splits into ( sub)bands , which can be described by an rg approach proposed by niu and nori @xcite .
we use this property later to derive a relation between the structure and the properties of the electronic wave function . in sec .
[ sec : wavefunctions ] we study the multifractal properties of the wave functions in fibonacci chains and the labyrinth tiling .
this includes the computation of the participation ratio @xmath44 , which reflects the fraction of the total number of sites for which the probability measure of the wave function @xmath45 is significantly different from zero .
hence , it provides information about the degree of localization of the wave function @xmath46 @xcite . for the discrete positions @xmath47 the participation ratio is defined via the inverse participation number @xmath48 of a wave function @xmath46 and the number of sites
@xmath49 in @xmath0 dimensions as @xmath50^{-1 } \right\rangle \propto v_a^{-\gamma}\;.\ ] ] the expression @xmath51 denotes the arithmetic average over all eigenstates in the spectrum or over a certain energy interval . in quasicrystalline systems
the participation ratio @xmath52 of a state @xmath53 was found to scale with a power law with respect to the number of sites @xmath49 , where the exponent @xmath54 is connected to the nature of this eigenstate @xcite .
the wave function of a localized state is characterized by a scaling exponent @xmath55 , and @xmath56 corresponds to an extended state .
intermediate values of @xmath57 ( @xmath58 ) indicate fractal eigenstates , which are neither extended over the whole system nor completely localized at a certain position and show self - similar patterns @xcite .
the participation ratios within a small energy interval usually fluctuate over a certain range , but the participation ratios and the inverse participation numbers as well as the corresponding scaling exponents do not show any significant trend over the whole energy spectrum for the systems considered here @xcite .
while the eigenstates of purely periodic systems can be characterized by a single scaling exponent @xmath59 , the wave functions of chaotic systems and quasicrystals are usually multifractals @xcite . the generalized dimensions @xmath60 describe the nature of the multifractality of the eigenstates , where @xmath60 is a monotonically decreasing function of @xmath61 .
the dimension @xmath62 is equal to the capacity dimension , which can be understood as the box counting dimension here , @xmath63 is equal to the information dimension , and @xmath64 to the correlation dimension .
the generalized dimensions can be determined either via the generalized inverse participation numbers or via a multifractal analysis .
for the first method one investigates the scaling behaviour of the generalized inverse participation numbers averaged over all wave functions of a system with the linear system size @xmath65 , i.e. , @xmath66 however , in sec .
[ subsec : dq - num ] we present numerical results obtained via a multifractal analysis , which shows less numerical problems for negative parameters @xmath61 in comparison to the generalized inverse participation numbers .
the multifractal analysis is based on a standard box - counting algorithm , i.e. , in @xmath0 dimensions the system of linear system size @xmath65 is divided into @xmath67 boxes of linear size @xmath68 @xcite . for each box @xmath69
we measure the probability to find the electron in this particular box , i.e. , we study the spatial distribution of the wave functions .
hence , for an eigenstate @xmath70 we determine the probability @xmath71 and then compute the mass exponent @xmath72 by a least squares fit of the quantity @xmath73 versus @xmath74 considering box sizes from @xmath75 @xcite .
the quantity @xmath76 denotes the ratio of the box size @xmath68 and the system size @xmath65 .
we only consider boxes with integer fractions of @xmath77 @xcite . in principle
it is possible to improve the quality of the results by adapting the multifractal analysis to work also for non - integer fractions @xmath77 and by averaging over different initial positions of the boxes @xcite .
however , we do not apply this improved method because it requires significantly more computing time .
we also found that the simple partitioning scheme yields sufficiently accurate results for the determination of the mass exponents .
the generalized dimensions @xmath60 are easily obtained from the mass exponents according to the relation @xmath78 . in the second part of the paper
we investigate the dynamics of the quasiperiodic systems by means of the time evolution of a wave packet @xmath79 , which is constructed from the solutions @xmath80 of the time - dependent schrdinger equation obtained by the separation approach .
the wave packet is then represented as a superposition of these orthonormal eigenstates @xmath81 @xcite , where we assume that it is initially localized at the position @xmath82 of the quasiperiodic tiling , i.e. , @xmath83 .
hence , with the completeness relation and normalized basis states the wave packet is defined by @xmath84 for a wave packet @xmath85 initially localized at the position @xmath82 the temporal autocorrelation function @xcite @xmath86 corresponds to the integrated probability to be at the position @xmath82 up to time @xmath87 , i.e. , @xmath88 here , we denote the integrand as return probability @xmath89 , i.e. , @xmath90 the autocorrelation function for an infinite system is expected to decay with a power law @xmath91 for @xmath92 @xcite , where the scaling exponent @xmath93 depends on the initial position @xmath82 of the wave packet .
further , it is known that the scaling exponent @xmath93 is equivalent to the correlation dimension @xmath94 of the local density of states @xmath95 of a system @xcite . due to this equivalence
we can distinguish two limit cases and their associated spectral properties .
for @xmath96 the autocorrelation function and the return probability remain constant .
a pure point spectrum implies @xmath96 , but the opposite statement may not be true @xcite . in one dimension
@xmath97 corresponds to the ballistic motion of an electron , where the electron moves freely on the chain without scattering @xcite . in two and three dimensions
@xmath97 usually does not indicate ballistic transport .
further , an absolutely continuous spectrum implies @xmath97 but again the reverse is not necessarily true @xcite . for quasiperiodic structures one often obtains intermediate values of @xmath98 , i.e. , @xmath99 , which is related to the singular continuous energy spectrum @xcite . to provide a better insight into the dynamical properties and to compare the properties of different systems , we average the results over different initial positions of the wave packet , i.e. , we investigate the scaling behaviour of @xmath100 .
previous studies mostly use non - averaged data , which do not adequately display the macroscopic transport properties @xcite . due to restrictions in computing time
we can not consider all possible initial positions @xmath82 .
in particular , for the fibonacci chain we have chosen the 500 positions nearest to the center of the chain and in two and three dimensions we have randomly selected 100 vertices in a square of linear size 25 and a cube of linear size 15 around the center of the tilings .
zhong and mosseri already studied the temporal autocorrelation function and the return probability for the square and cubic fibonacci tiling @xcite .
as for this model the higher - dimensional energy values are the sum of the one - dimensional energy values ( cf .
, the time - evolution operator becomes separable .
therefore , the dynamics of the wave packets in higher dimensions are much better understood than for the labyrinth tiling .
they were able to show that @xmath101 and that the scaling behaviour of the temporal autocorrelation function is given by @xcite @xmath102 further , they found that for the limit case @xmath27 one always obtains @xmath103 . with the same arguments ,
also the threshold values for the transition to an absolutely continuous energy spectrum can be directly obtained from the one - dimensional scaling exponents @xmath104 according to @xmath105 .
in the regime of strong quasiperiodic modulation it is possible to relate the energy spectra and the wave functions of different approximants of the fibonacci chain by an rg approach proposed by niu and nori @xcite .
we use this approach to show in the following two sections that the wave functions do not become localized even for @xmath106 . in sec .
[ sec : returnprob ] we also apply the rg approach to find a relation between the quasiperiodic structure of the fibonacci chain and its return probability .
the general idea is that a given system can be transformed ( renormalized ) by the application of an rg step in such a way that we obtain a new system with a reduced number of degrees of freedom ( i.e. sites / bonds ) . in particular , for the fibonacci chain for @xmath107 the isolated sites ( atoms ) and biatomic clusters ( molecules ) coupled by a strong bond are visualized in figure [ fig : rg - fib ] .
this yields three highly degenerate energy levels : @xmath108 for the atomic sites and @xmath109 for the bonding and antibonding states of the molecules @xcite . for non - zero coupling parameters
@xmath11 one finds a coupling between these isolated atoms and molecules , where the dominant contribution occurs between sites of the same type .
hence , for the fibonacci chain @xmath110 one can distinguish two possible rgs @xcite : : in the atomic rg all atomic sites survive an rg step , and in the molecular rg all molecular sites are replaced by a new site .
the thick ( thin ) lines denote the strong ( weak ) bonds.,title="fig:",width=302 ] : in the atomic rg all atomic sites survive an rg step , and in the molecular rg all molecular sites are replaced by a new site .
the thick ( thin ) lines denote the strong ( weak ) bonds.,title="fig:",width=302 ] _ atomic rg _
the renormalized grid after one rg step consists of the atomic sites of the original chain as shown in figure [ fig : rg - fib ] .
these new sites are then connected by new strong and weak bonds . in particular , atoms separated by one molecule in the original chain become connected by a new strong bond and atoms separated by two molecules in the original chain get connected by a new weak bond .
_ molecular rg _ in one rg step all molecular sites of the original chain are replaced by new atomic sites ( cf .
figure [ fig : rg - fib ] ) .
the atoms in the renormalized grid are connected by new strong bonds between the sites of originally neighboring molecules and by new weak bonds between the sites for molecules that were separated by an atomic site . for both rg approaches
we obtain a new fibonacci chain , which is scaled in length and energy .
the derivation of the length scalings @xmath111 and the scaling factors @xmath112 for the new bond strengths of an approximant under the application of one rg step is based on the brillouin - wigner perturbation theory . for the fibonacci chain
@xmath110 one obtains for the @xcite * _ atomic rg _ a length scaling of @xmath113 and an energy scaling of @xmath114 and for the * _ molecular rg _ a length scaling of @xmath115 and an energy scaling of @xmath116 .
in other words , the energy spectrum @xmath117 of the hamiltonian @xmath118 of the fibonacci chain @xmath6 can be related to the spectrum of three hamiltonians of smaller approximants @xcite . with the scaling relations given above
, it can be shown that the energy spectrum splits according to @xcite @xmath119 this means that the central main band of the energy spectrum corresponds to the spectrum @xmath120 of the hamiltonian @xmath121 comprising the @xmath122 atomic sites for @xmath107 scaled by a factor @xmath123 .
the left and right main bands at energies @xmath124 correspond to the bonding and antibonding states of the spectra of the hamiltonian @xmath125 for the @xmath126 molecular sites @xcite .
the recursive structure of the chains also has an influence on the properties of the individual wave functions . in particular , for the fibonacci chain @xmath6 with periodic boundary conditions pichon provided a quantitative result for the amplitudes of the wave functions for different approximants @xmath5 @xcite .
note that in this case the number of sites / eigenstates is @xmath127 .
for the energy levels @xmath128 ( @xmath129 ) in the central main band of the energy spectrum associated to the atomic sites , the relation is given by @xmath130{\tau^{-3 } } | \psi_{l^\prime , a-3}^i(e/\bar{z})| \;.\ ] ] likewise , one obtains the recursive relations for wave functions associated to the energy levels @xmath17 belonging to the left main band ( @xmath131 ) and to the right main band ( @xmath132 ) of the energy spectrum : @xmath133 the sites @xmath134 and @xmath135 are those linked by a bond of strength @xmath4 , which is replaced by the new site @xmath136 during one step of the molecular rg .
these are approximate results because the atomic ( molecular ) rg approach ignores the probability of being on a molecular ( atomic ) site . comparing this with the numerical results in figure [ fig : wave - functions - rg ] ,
we see that this nevertheless is a good approximation since the probability density on the other type of cluster is much lower in the regime of strong quasiperiodic modulation ( @xmath137 ) . for the fibonacci chain @xmath6 for the molecular rg ( top ) and the atomic rg ( bottom ) for @xmath138 and @xmath139.,width=302 ]
note that this rg approach is different from the rg approach used by wang and sanchez for the analysis of the electronic transport in fibonacci chains and the corresponding hypercubic tilings @xcite . with their renormalization rules they were able to derive an iterative procedure for the evaluation of kubo - greenwood formula at zero temperature .
while this method allows one to efficiently compute the conductivity for large quasiperiodic systems , the rg approach applied in this paper allows us to directly connect the structure of the systems to certain properties of the wave functions and the transport properties .
in this section we first present numerical results for the multifractal properties of the wave functions for the fibonacci systems .
we then show that the multifractality is directly related to the structure of the quasiperiodic systems for small coupling strengths @xmath11 and that even for @xmath106 the wave functions remain multifractals . in figure
[ fig : partratio ] we show numerical results for the scaling exponent @xmath57 of the average participation ratio according to eq . for the fibonacci chain and the labyrinth tilings .
the data indicate that the scaling exponents for the labyrinth tiling in two and three dimensions approach the one - dimensional results for large system sizes @xcite .
we recently showed analytically that this property is fulfilled in the limit of infinite systems @xcite .
additionally , for all systems the scaling exponents @xmath57 approach a constant value for @xmath106 .
this limit behaviour can be related to the hierarchical structure of the chains and is further discussed in sec .
[ subsec : limitdq ] . of the inverse participation number @xmath140 and
the scaling exponent @xmath57 of the participation ratio @xmath44 averaged over all wave functions of the systems for @xmath139 .
the largest considered systems are the fibonacci chain @xmath141 with @xmath142 sites and the labyrinth tilings @xmath143 with @xmath144 sites and @xmath145 with @xmath146 sites .
the dotted lines in the insets correspond to the analytically predicted values @xmath147 and @xmath148 .
, width=302 ] used in the multifractal analysis with the ratio @xmath149 for the fibonacci chain @xmath150 for @xmath138 and @xmath139.,width=302 ] the multifractal properties of the wave functions are obtained by a multifractal analysis described in sec .
[ subsec : def - part - ratio ] .
the scaling behaviour of the quantity @xmath151 with the box size @xmath68 is visualized for the fibonacci chain in figure [ fig : multanalysis ] . for most of the systems
the results can be well described by a power law .
of course , for large negative values of @xmath61 the accuracy deteriorates .
a similar scaling behaviour can be observed for other values of @xmath11 and for the higher - dimensional systems although we find some deviations from the power law for higher - dimensional labyrinth tilings and small coupling parameters @xmath11 .
the corresponding results for the generalized dimensions @xmath60 are shown in figure [ fig : dqphi ] .
the results yield a box counting dimension of @xmath152 , which corresponds to the spatial dimension @xmath0 as expected . with increasing coupling parameter
@xmath11 the range of the scaling exponents @xmath60 narrows , i.e. , the strength of the multifractality is stronger for smaller coupling parameters @xmath11 . while in the limit @xmath27 there is a single scaling exponent @xmath153 , we find that for @xmath106 the distribution approaches a limit function , which can be related to the quasiperiodic structure .
this is considered in more detail in the following section .
determined via a multifractal analysis for the fibonacci chain @xmath150 ( solid lines ) and the two - dimensional labyrinth tiling @xmath154 ( dashed lines ) for @xmath139.,width=302 ] further , the results in figure [ fig : dqphi ] indicate that the scaling exponents of different dimensions are related by @xmath155 .
this is not surprising as we have shown analytically that for the golden - mean labyrinth this relation is approached with increasing system size @xcite .
if the coupling strength @xmath11 of the weak bond approaches zero , one can observe that the scaling exponent @xmath57 of the participation ratio approaches a constant till it drops to zero for @xmath107 .
this behaviour was already reported before @xcite but no explanation for this phenomena was given .
a similar behaviour can also be found for the generalized dimensions @xmath60 .
this indicates that the wave functions do not become localized even for very small couplings @xmath11 and that they remain multifractals for @xmath106 regardless of the dimension .
thereby , this limit behaviour can be related to the hierarchical structure of the eigenstates of the quasiperiodic chains . in particular , for the fibonacci chain @xmath6 we can use the recursive structure of the energy spectrum according to eq . and of the wave functions according to eqs . and to derive the limit analytically .
we start by deriving the participation ratios for an atomic state ( @xmath129 ) of the fibonacci chain @xmath156^{-1 } \\ & \stackrel{{\rm rg}}{\simeq } & \frac{1}{f_a } \left [ \tau^{-3 } \sum_{l^\prime \in { \rm atom } } { \ensuremath{\left\vert\psi_{l^\prime}^i\right\vert}}^4 \right]^{-1}\\ & = & \frac{f_{a-3}}{f_a } \tau^3 p_{a-3 } ( \psi^i , f_{a-3 } ) \simeq p_{a-3 } ( \psi^i , f_{a-3 } ) \;.\end{aligned}\ ] ] analogously , we obtain for the molecular states of the left main band of the energy spectrum ( @xmath131 ) the expression @xmath157^{-1 } \\ & \stackrel{{\rm rg}}{\simeq } & \frac{1}{f_a } \left [ \tau^{-2 } \sum_{l^\prime \in { \rm mol } } { \ensuremath{\left\vert\psi_{l^\prime}^i\right\vert}}^4 \right]^{-1}\\ & = & \frac{f_{a-2}}{f_a } \tau^2 \frac{1}{f_{a-2 } } \left [ 2 \sum_{l^\prime=1}^{f_{a-2 } } { \ensuremath{\left\vert\psi_{l^\prime}^i\right\vert}}^4 \right]^{-1 } \\ & \stackrel{f_a \gg 1}{\simeq } & \frac{1}{2 } p_{a-2 } ( \psi^i , f_{a-2 } ) \;.\end{aligned}\ ] ] as these transformations do not directly depend on the energy we obtain the same expression also for the right main band of the energy spectrum .
the average participation ratio is then directly calculated from these results according to @xmath158 with the scaling behaviour of the average participation ratio @xmath159 according to eq .
, we obtain the relation @xmath160 this expression is independent of the coupling parameter @xmath11 and yields a scaling exponent @xmath161 , which corresponds very well to the numerical results for @xmath106 in figure [ fig : partratio ] . with the same approach we can also derive an analytical expression for the generalized dimensions @xmath60 . here , we start with the generalized inverse participation numbers for an atomic state according to eq . , i.e. , @xmath162 likewise , we obtain for the molecular states of the left main band of the energy spectrum @xmath163 the same expression is again valid for the right main band of the energy spectrum .
this yields the recursive equation for the average generalized inverse participation numbers @xmath164 with the scaling behaviour of the average generalized inverse participation numbers @xmath165 according to eq .
, we obtain the relation @xmath166 this expression is again independent of the coupling parameter @xmath11 and yields e.g. for @xmath167 the scaling exponent @xmath168 , which corresponds well to the numerical results for @xmath106 in figure [ fig : partratio ] .
while this analytical result shows an even better agreement with the numerical results for larger values of @xmath61 , there are significant differences for smaller values of @xmath61 as visualized in figure [ fig : dqphi ] .
this is caused by the fact that the measures for negative parameters @xmath61 are dominated by the boxes with the smallest probabilities .
however , these probabilities are not considered in the rg approach because we neglect the probability on the other type of cluster .
nevertheless , the rg approach allows us to relate the quasiperiodic structure to the multifractal properties of the system for @xmath169 .
in order to obtain a deeper understanding of the connections between the transport properties and the quasiperiodic structure of a system , we study in this section the wave - packet dynamics for the different tilings . in the limit of weak coupling ( @xmath170 )
it is again possible to connect the structure of the systems and the wave - packet dynamics by an rg approach .
some typical results for the temporal autocorrelation function @xmath171 introduced in sec . [ subsec : def - return - prob ] are shown in figure [ fig : autocorr ] for the fibonacci chain and the associated labyrinth tilings in two and three dimensions .
the corresponding scaling exponents @xmath98 obtained by least squares fits are compiled in figure [ fig : delta ] .
the power - law behaviour can be observed over several orders of magnitude in time , before @xmath171 approaches a constant due to finite size effects .
while we obtain @xmath172 for a system with localized wave functions , we expect to get @xmath173 for a periodic chain of finite size @xmath65 due to the completely extended wave functions .
a comparison with our numerical data for @xmath174 in figure [ fig : autocorr ] shows a good agreement with the latter result .
further , the quantity @xmath171 decreases more strongly for higher values of the coupling parameter @xmath11 in agreement with the more extended wave functions in these systems .
however , for a periodic chain in one dimension one does not obtain the expected result @xmath175 by a least squares fit due to sublogarithmic contributions @xcite .
additionally , for small values of @xmath11 we observe some oscillations .
this corresponds to a step - like behaviour of @xmath86 , which can be related to the hierarchical structure of the chains @xcite . comparing the scaling behaviour of the average temporal autocorrelation function @xmath171 in figure [ fig : autocorr ] with respect to the dimension
, we observe that for the one - dimensional chain the absolute value of the slope steadily increases with increasing coupling parameter @xmath11 , while in two and three dimensions @xmath171 is nearly identical for @xmath176 and @xmath177 .
these differences are also clearly visible for the corresponding scaling exponents @xmath98 shown in figure [ fig : delta ] , which approach one when the coupling parameter @xmath11 becomes larger than the threshold parameters @xmath178 and @xmath179 for which the spectrum becomes absolutely continuous .
it is expected that @xmath175 for infinite systems if @xmath180 .
this limit behaviour of the scaling exponent @xmath98 is indicated by the dashed lines in figure [ fig : delta ] .
similar results can be found for other metallic - mean systems @xcite .
although @xmath181 indicates ballistic transport for the one - dimensional periodic chains @xcite , the behaviour in two and three dimensions does not correspond to a transition towards ballistic spreading .
the latter result is not always correctly used in literature because e.g. zhong and mosseri wrongly state that there is a crossover from ballistic - like behaviour to a non - ballistic - like behaviour in the hypercubic tiling @xcite .
the results for the mean square displacement of the wave packet for the hypercubic or the labyrinth tiling clearly reveal anomalous transport for @xmath182 . only for @xmath43 ballistic transport
is observed @xcite . over 10 data points ( light colors ) and
the corresponding temporal autocorrelation function @xmath171 ( dark colors ) for the fibonacci chain @xmath141 .
both quantities are averaged over 500 initial positions of the wave packet.,width=302 ] in order to obtain information about transport properties instead of the properties of the energy spectrum we can make use of the fact that already the integrand @xmath183 in eq
. shows a power - law behaviour according to @xmath184 .
the integration itself is only used for smoothing the results because @xmath183 shows considerable oscillations ( cf .
e.g. figs .
[ fig : paca - smooth ] and [ fig : paca - cmp ] ) .
naturally , one would expect that the scaling exponents of @xmath171 and @xmath183 are equal , i.e. , @xmath185 .
however , the integration leads to two basic disadvantages .
for @xmath186 we obtain an additional logarithmic contribution @xmath187 from the integration @xcite .
hence , for the periodic one - dimensional systems and the commonly used system sizes one obtains exponents @xmath98 significantly lower than @xmath188 in agreement with our result @xmath189 .
further , in higher dimensions the scaling exponent @xmath190 of @xmath183 becomes greater than 1 for large coupling parameters @xmath11 ( cf .
figure [ fig : delta ] ) .
zhong and mosseri pointed out that in this case the integral in eq .
converges and yields a constant factor , which results in @xmath191 .
hence , the scaling exponent @xmath175 is obtained in the absolute continuous regime @xcite . from figure
[ fig : delta ] we find @xmath192 for @xmath193 in two dimensions and @xmath194 in three dimensions in reasonable agreement with @xmath195 and @xmath196 .
since the integral converges rather slowly for coupling parameters @xmath11 only slightly larger than @xmath41 , we do not observe @xmath175 for the finite systems considered here .
however , for an infinite system we expect to obtain a clear transition towards @xmath175 as indicated by the dashed lines in figure [ fig : delta ] . therefore , we lose information about the transport properties and @xmath175 is misleading because it often does not occur where it is expected ( e.g. for periodic one - dimensional systems ) but wrongly suggests a ballistic behaviour for @xmath197 in two and three dimensions .
hence , for studying the transport properties it is better to extract the scaling exponent @xmath190 via the calculation of @xmath183 , or the temporal autocorrelation has to be redefined so that the integrand does not converge anymore .
we choose the first option and use a moving average instead of the integration to smooth the results as shown in figure [ fig : paca - smooth ] . in figure
[ fig : delta ] the scaling exponent @xmath190 of the return probability @xmath183 divided by the spatial dimension @xmath0 is shown for the fibonacci chain and the labyrinth tilings .
the comparison of the scaling exponents for the temporal autocorrelation @xmath171 and its integrand @xmath183 shows that both results are quite close for the one - dimensional systems .
the main difference occurs for @xmath198 , where we obtain the expected value @xmath199 for @xmath43 while the evaluation of @xmath171 results in @xmath189 .
further , there are some differences for the fibonacci chain , especially , for small values of @xmath11 .
this probably originates from the fact that we have not considered large enough times in the calculations of @xmath171 .
actually , it is much more time - consuming to compute @xmath171 because the integration requires the consideration of much smaller time intervals .
this is another advantage of calculating only the quantity @xmath183 .
further , in two dimensions we observe that the scaling exponents @xmath200 are about twice as large as the one - dimensional scaling exponents @xmath201 .
this is in good agreement with the results for the hypercubic tiling in eq . .
however , the results for the three - dimensional labyrinth tiling do not entirely fit into this scheme . while for @xmath202 the expression @xmath203 is only slightly smaller than the exponent @xmath201 , the behaviour changes completely for large values of @xmath11 . in the latter case
the scaling exponent @xmath204 becomes almost constant and approaches @xmath205 .
the exact reason for this behaviour is unclear yet .
having a look at @xmath89 for different initial positions of the wave packet for @xmath43 , one can observe that after an unusually smooth decrease of the function @xmath89 according to a power law , there is an abrupt dip in the curves when finite size effects first occur ( cf . figure [ fig : paca - cmp ] ) .
for the bcc lattice with @xmath206 sites and the sc lattice with @xmath207 sites as well as the analytical results of the return probability @xmath89 for the bcc lattice according to eq .
.,width=302 ] since the square fibonacci tiling and the two - dimensional labyrinth tiling are also identical for @xmath43 , the scaling exponents @xmath98 and @xmath190 are identical . however , in three dimensions for @xmath208 the labyrinth tiling becomes a _ body centered cubic _ ( bcc ) lattice while the cubic tiling approaches a _ simple cubic _
( sc ) lattice .
therefore , these two grids have a different number of nearest neighbours independent of the coupling parameter @xmath11 , i.e. , eight in the labyrinth tiling and six in the cubic tiling . while the numerical results for the two - dimensional labyrinth tiling satisfy the relations in eqs . and , there are significant differences in the wave - packet dynamics in three dimensions , which are probably related to the different grid structure .
therefore , in figure [ fig : paca3d2 ] we plot for both systems the probability density of a three - dimensional wave packet along the @xmath209-@xmath112 plane through the center of the @xmath210 axis for different times @xmath22 and the corresponding return probability @xmath89 in figure [ fig : paca - cmp ] . while the probability density of the wave packet for the sc lattice is distributed in a cube around the origin , for the bcc lattice it spreads in the shape of an octahedron from the center with some prominent resonances along the @xmath210 , @xmath209 , and @xmath112 directions .
for the labyrinth tiling the wave packet already reached the boundaries for @xmath211 and was reflected due to the free boundary conditions .
this leads to constructive interference at two boundaries and to destructive interference at the opposite boundaries because the initial position of the wave packet is chosen from one of the four possible center positions of the labyrinth for even chain lengths , which results in a slight asymmetry of the wave - packet dynamics
. sites ( left column ) and the sc lattice with @xmath207 sites ( right column ) at the times @xmath212 ( top ) , @xmath213 ( center ) , and @xmath211 ( bottom ) .
the corresponding values of @xmath89 are shown in figure [ fig : paca - cmp].,title="fig:",width=151 ] sites ( left column ) and the sc lattice with @xmath207 sites ( right column ) at the times @xmath212 ( top ) , @xmath213 ( center ) , and @xmath211 ( bottom ) .
the corresponding values of @xmath89 are shown in figure [ fig : paca - cmp].,title="fig:",width=151 ] + sites ( left column ) and the sc lattice with @xmath207 sites ( right column ) at the times @xmath212 ( top ) , @xmath213 ( center ) , and @xmath211 ( bottom ) .
the corresponding values of @xmath89 are shown in figure [ fig : paca - cmp].,title="fig:",width=151 ] sites ( left column ) and the sc lattice with @xmath207 sites ( right column ) at the times @xmath212 ( top ) , @xmath213 ( center ) , and @xmath211 ( bottom ) .
the corresponding values of @xmath89 are shown in figure [ fig : paca - cmp].,title="fig:",width=151 ] + sites ( left column ) and the sc lattice with @xmath207 sites ( right column ) at the times @xmath212 ( top ) , @xmath213 ( center ) , and @xmath211 ( bottom ) .
the corresponding values of @xmath89 are shown in figure [ fig : paca - cmp].,title="fig:",width=151 ] sites ( left column ) and the sc lattice with @xmath207 sites ( right column ) at the times @xmath212 ( top ) , @xmath213 ( center ) , and @xmath211 ( bottom ) .
the corresponding values of @xmath89 are shown in figure [ fig : paca - cmp].,title="fig:",width=151 ] according to figure [ fig : paca - cmp ] the return probability for the sc lattice strongly oscillates and decreases with a scaling exponent of @xmath214 , while for the bcc lattice it decreases rather smoothly with a scaling exponent @xmath215 .
the dip for @xmath216 for the bcc structure is clearly due to finite size effects and thus has not to be considered for the fitting of the scaling exponent .
although the center of the wave packet decreases more slowly for the bcc structure than for the sc lattice , the overall wave packet spreads significantly faster for the bcc lattice .
this might be related to the higher number of nearest neighbours in the bcc lattice .
the data for the sc lattice are in good agreement with the results of zhong and mosseri ( cf .
[ subsec : def - return - prob ] ) .
their approach can also be used for the determination of the return probability @xmath217 for the bcc lattice .
the approach makes use of the energy spectrum of the bcc lattice in @xmath218 space @xcite @xmath219 the probability amplitude at the initial position @xmath82 of the wave packet can be determined by the fourier transform of this energy spectrum .
hence , we obtain @xmath220 we use the fact that the integration over the brillouin zone ( bz ) can be reduced to an integral over a cube from @xmath221 to @xmath222 due to the symmetries of the components of the @xmath218 vector @xcite .
the function @xmath223 denotes the bessel function of the first kind .
the integral over the bessel function is evaluated for various time steps by numerical integration .
the results agree exactly with the numerical data as shown in figure [ fig : paca - cmp ] .
unfortunately , it is not possible to evaluate this integral for arbitrary large times because the bessel function strongly oscillates , which leads to numerical problems . while it is possible to derive with this approach an analytical expression for the long - time behaviour of the return probability for the sc lattice , we were not able to obtain a corresponding result for the bcc lattice . while we considered the limit case of the periodic systems in the previous section , we can also find a connection between the structure of the fibonacci chain @xmath110 and the transport properties in the regime of strong quasiperiodic modulation .
in particular , we are able to derive an analytical expression for the scaling exponents @xmath190 of the return probability @xmath183 .
this is based on the results of the rg approach , which yields a scaling behaviour of the energy spectrum according to eq . and of the wave functions according to eqs . and .
first , we determine the return probability @xmath224 for an approximant @xmath5 for the two different kinds of initial positions @xmath225 . for @xmath225 being an atomic site
we obtain @xmath226 and for @xmath225 belonging to a molecular site we get @xmath227 in the last step we make use of the fact that for large times @xmath22 only the average behaviour of @xmath228 is relevant , which results in a factor of @xmath229 . the average return probability for an approximant @xmath5 is then given by @xmath230 for the fibonacci chain @xmath141.,width=302 ] with the scaling behaviour of the return probability @xmath184 , we obtain the equation @xmath231 which approaches the numerical results for decreasing values of @xmath11 as shown in figure [ fig : paca - cmp2 ] . note that pichon @xcite studied a similar quantity , which he also denoted as return probability .
however , that quantity is related to the spectral measure of the density of states .
in contrast , the return probability @xmath183 is related to the spectral measure of the local density of states for the initial position @xmath225 of the wave packet .
only the latter quantity is related to the wave - packet dynamics .
another quantity often considered for the description of the electronic transport properties is the mean square displacement ( also called width ) @xmath232^{\frac{1}{2}}\ ] ] of the wave packet .
the spreading of the width @xmath233 of the wave packet for an infinite system shows anomalous diffusion for @xmath92 according to @xmath234 , where the scaling exponent @xmath235 depends on the initial position @xmath82 of the wave packet @xcite .
the electronic transport properties are governed by the wave - packet dynamics averaged over different initial positions , i.e. , @xmath236 .
thereby , the scaling exponent @xmath2 is related to the conductivity @xmath237 via the generalized drude formula for zero - frequency conductivity @xmath238 @xcite .
the quantity @xmath239 denotes the elementary charge of an electron , @xmath240 the density of states at the fermi level @xmath241 , @xmath111 a constant , and @xmath242 a characteristic time beyond which propagation becomes diffusive due to scattering @xcite .
hence , @xmath243 corresponds to the absence of diffusion , @xmath244 to classical diffusion , and @xmath245 to ballistic spreading . for quasiperiodic structures
one often observes anomalous diffusion characterized by @xmath246 @xcite .
for @xmath170 the width of the wave packet can be related by an rg approach to the structure of the fibonacci chain @xcite and the labyrinth tiling @xcite .
the computation of the scaling exponent @xmath2 is computationally very expensive especially in higher dimensions because it includes an average over all sites of the system .
however , for the hypercubic tiling the scaling exponents are related according to @xmath247 due to the separability of the time - evolution operator .
numerical and analytical results suggest that this relation could be also valid for the labyrinth tilings @xcite . however , in general one can make use of several lower bounds for the scaling exponents @xmath2 , which have been proposed in literature .
we shortly introduce two of them here and compare them to our numerical data @xcite .
further , we present an improved lower bound for systems with absolute continuous energy spectra .
_ lower bound by guarneri : _ as a rule of thumb , the wave - packet propagation is faster for smoother spectral measures . in particular , guarneri proved for the scaling exponent @xmath2 that in one dimension the lower bound @xmath248 holds @xcite .
this can be generalized for arbitrary dimensions to @xmath249 this expression quantitatively relates the diffusive properties to the information dimension of the spectral measure of the local density of states .
a comparison with our numerical results in figure [ fig : bounds - beta ] shows that this inequality is clearly satisfied in one dimension .
the spectral dimensions @xmath250 are computed by a multifractal analysis with respect to the energy .
therefore , we divide the energy range into @xmath251 boxes of size @xmath252 and compute for a certain vertex @xmath82 of the tiling the function @xmath253 which shows a power - law behaviour with the generalized spectral dimension @xmath250 @xcite . with the inequality in eq .
it is possible to show that an absolutely continuous energy spectrum in one dimension implies ballistic transport @xcite and that in two dimensions this inequality requires a superdiffusive evolution of the wave packet ( @xmath254 ) in the presence of an absolutely continuous spectrum .
however , in three dimensions it allows the occurrence of subdiffusive wave - packet spreading ( @xmath255 ) although the energy spectrum is purely absolutely continuous , i.e. , @xmath256 @xcite .
this is a very interesting aspect because investigations on real quasicrystals revealed that they are usually accompanied by absolutely continuous spectra @xcite but the conductivity is well described by the generalized drude formula for the case @xmath255 @xcite .
since @xmath257 eventually becomes @xmath188 for an absolutely continuous spectrum , this bound strongly underestimates the actual scaling exponent @xmath2 in two and three dimensions for large values of @xmath11 as shown in figure [ fig : bounds - beta ] .
_ lower bound by ketzmerick et al . :
_ using the fact that the center of the wave packet is known @xcite to decay according to @xmath258 , ketzmerick et al . showed that due to the normalization condition the spreading of a wave packet is described by an exponent @xmath259 in @xmath0 dimensions as long as @xmath260 , i.e. , for singular continuous spectra @xcite . for wave packets which spread in a space of reduced dimension @xmath64 instead of @xmath0 dimensions they showed that the relation @xmath261 holds for all systems which are characterized by a single scaling exponent @xmath262 @xcite .
for the more general case of multiscaling in time ( @xmath263 ) they found that this yields a lower bound for positive moments , i.e. , @xmath264 for negative moments one obtains a respective upper bound . comparing this with our numerical results for @xmath167 in figure [ fig : bounds - beta ] for the one - dimensional case
, we find that the inequality in eq . is fulfilled and provides a good lower bound , which is significantly better than the result by guarneri in eq .
the largest differences occur for @xmath208 , which are partly caused by numerical issues because the relation @xmath265 is satisfied for a periodic chain .
the numerical results only fulfill the lower bound according to eq . because the wave - packet dynamics for the fibonacci chain show multiscaling in time .
this inequality also holds in two and three dimensions .
however , in the regime of an absolutely continuous spectrum , the spectral dimensions fulfill @xmath266 , and the inequality in eq .
is no longer a good lower bound as shown in figure [ fig : bounds - beta ] . addressing the latter case
we would like to point out that the decay of the center of the wave packet is only described by the correlation dimensions of the local density of states for @xmath267 because for @xmath197 we obtain @xmath265 and the integral of the autocorrelation function converges @xcite .
however , in this regime the scaling exponent @xmath190 , which describes the decay of the center of the wave packet , can become larger than 1 ( cf .
figure [ fig : delta ] ) .
hence , for @xmath0-dimensional systems the bound in eq . should be replaced by @xmath268 in the absolutely continuous regime . according to eq . for the hypercubic tiling this
bound is as good as for the one - dimensional chain .
this bound should hold also in general because is solely makes use of the normalization condition of the wave packet .
we also checked in figure [ fig : bounds - beta ] whether this relation is satisfied for the labyrinth tiling and found that this is a significantly better lower bound than that given in eq . for @xmath197
however , in three dimensions there are still large deviations compared to @xmath2 for coupling parameters @xmath208 .
this suggests that further studies are necessary in order to relate the width of the wave packet with the decay of its center .
the benefit of good lower bounds for @xmath2 is that the computation of the scaling exponents @xmath190 , @xmath94 , and @xmath64 is much less time consuming because for @xmath190 it is not necessary to average over all grid positions @xmath47 and the spatial and spectral dimensions , @xmath64 and @xmath94 , are even time - independent .
in this paper we presented the scaling exponents of the participation ratio and the inverse participation numbers for large approximants of the separable fibonacci tilings .
the results showed that the wave functions of the systems are multifractals and that the strength of the multifractality increases with decreasing coupling parameter @xmath11 in agreement with the literature . by applying an rg approach
we were able to show that the limit behaviour of the wave functions for @xmath106 is related to the structure of the fibonacci chain .
up to now this observation was only based on numerical results .
further , we presented numerical data for the quantum diffusion of electrons in these systems .
compared to previous results for the labyrinth tiling , we considered larger approximants and included an average over different initial positions of the wave packets , which allowed us to determine the scaling exponents with a higher precision .
our results also show that for the study of the quantum diffusion it is advisable to compute the return probability @xmath183 instead of the temporal autocorrelation function because for an absolutely continuous energy spectrum the integral over the return probability converges and no further information can be obtained . in higher - dimensional systems @xmath181
usually corresponds to the existence of an absolutely continuous part in the energy spectrum rather than to the occurrence of ballistic transport . in order to understand the diffusive properties of a system in detail it is also necessary to compute not only the width @xmath269 of the wave packet but also the return probability @xmath183
this can be easily understood from our results of the two tilings in three dimensions .
while for the labyrinth tiling and the cubic fibonacci tiling the scaling exponents @xmath2 hardly differ for the two models @xcite , the scaling exponents @xmath190 are rather different for large coupling parameters @xmath11 ( cf . figure [ fig : paca - cmp ] ) . in figure
[ fig : paca3d2 ] we have also seen that the wave packets spread in a quite different way for @xmath43 .
however , these differences are averaged out during the computations of the mean square displacement . for the fibonacci sequence
we were also able to derive an analytical expression of the scaling exponent @xmath190 of the return probability @xmath183 in the regime of strong quasiperiodic modulation , i.e. , we found a relation between the quasiperiodic structure of the chain and the emerging electronic transport properties .
further , we show that the exponent @xmath190 can be used to define a better lower bound for the scaling exponent @xmath2 of the width of the wave packet for systems with absolutely continuous energy spectra .
this result is not restricted to the quasiperiodic systems considered here .
10 enrose r 1974 _ bull .
inst . math .
* 10 * 266271 | we study the properties of wave functions and the wave - packet dynamics in quasiperiodic tight - binding models in one , two , and three dimensions .
the atoms in the one - dimensional quasiperiodic chains are coupled by weak and strong bonds aligned according to the fibonacci sequence .
the associated @xmath0-dimensional quasiperiodic tilings are constructed from the direct product of @xmath0 such chains , which yields either the hypercubic tiling or the labyrinth tiling .
this approach allows us to consider rather large systems numerically .
we show that the wave functions of the system are multifractal and that their properties can be related to the structure of the system in the regime of strong quasiperiodic modulation by a renormalization group ( rg ) approach .
we also study the dynamics of wave packets to get information about the electronic transport properties .
in particular , we investigate the scaling behaviour of the return probability of the wave packet with time . applying again the rg approach
we show that in the regime of strong quasiperiodic modulation the return probability is governed by the underlying quasiperiodic structure .
further , we also discuss lower bounds for the scaling exponent of the width of the wave packet and propose a modified lower bound for the absolute continuous regime . |
recently , the belle collaboration performed precise measurements of the @xmath0 reaction cross section for pion - pair invariant masses @xmath1 ranging from 0.8 to 1.5 gev @xcite . owing to the huge statistics and good energy resolution ,
a clear signal from the @xmath3 resonance was first discovered with the belle detector .
evidences for the @xmath3 production in @xmath5 collisions obtained in a series of previous measurements @xcite were essentially less conclusive .
the @xmath3 signal observed in the belle experiment turned out to be rather small .
this feature is in good qualitative agreement with the prediction of the four - quark model @xcite .
a detailed analysis of the preliminary belle data @xcite in the @xmath3 region was performed in ref .
in particular , it was found @xcite that the magnitude and shape of the @xmath3 peak observed in the @xmath0 reaction cross section excellently agree with the @xmath9 loop mechanism of the @xmath3 production , @xmath15 .
this result is one of many in favor of the four - quark nature of the @xmath3 state @xcite . in this paper
we clarify the current situation around the @xmath2 , @xmath3 , and @xmath4 resonances in the reactions @xmath10 by analyzing the final belle data @xcite on the @xmath0 cross section in the region @xmath16 gev , together with the crystal ball data @xcite on the reaction @xmath12 for @xmath17 gev . as the first step , in sec .
ii , the pure born cross sections and those involving the @xmath18 wave born contributions modified by strong final - state interactions are compared with the available data on the reactions @xmath10 @xcite .
such a comparison is very useful because it enables us to gain some idea of the scope and shape of other possible contributions to the cross sections .
the scheme taking into account the @xmath18 wave final - state interactions between pions and kaons , in which the @xmath2 and @xmath3 resonances take part , is presented in sec .
iii . in this scheme
we essentially use the results of the simultaneous analysis of the data on the @xmath19 decay , @xmath11 scattering , and reaction @xmath20 @xcite , as well as the results of the previous analyses of the @xmath10 reaction mechanisms @xcite . in secs .
iv and v , tentative values of the direct coupling constants of the @xmath2 and @xmath3 resonances to @xmath5 are determined from the belle and crystal ball data .
it is important that these constants turn out to be small , and consequently , the @xmath6 and @xmath7 decays are , in fact , the four - quark transitions caused by the @xmath8 and @xmath9 loop mechanisms , respectively .
the chiral shielding of the @xmath2 resonance takes place in the reactions @xmath10 @xcite as well as in @xmath11 scattering @xcite .
in addition , in sec .
v , we make some comments on the difficulties of interpreting the experimental measurements for @xmath12 production in the @xmath4 resonance region . here
we also point to a possible way out and announce some preliminary results of a simultaneous description of the @xmath0 data and the latest , very high statistics belle data on the @xmath12 reaction cross section .
the estimates for the @xmath13 decay width are also given .
conclusions are shortly formulated in sec .
the mark ii @xcite , cello @xcite , and belle @xcite data on the cross section for the reaction @xmath0 are shown in fig .
all the data correspond to a limited angular range of the registration of the charged pions , such that @xmath21 , where @xmath22 is the polar angle of the produced @xmath23 meson in the center - of - mass system of two initial photons .
the belle data are represented with statistical errors only , the relative values of which are approximately equal to 0.5%1.5% .
the @xmath1 bin size in the belle experiment has been chosen to be 5 mev , with the mass resolution of about 2 mev .
figure 1(a ) also represents the comparison between the data and theoretical curves corresponding to the pure born cross sections for the process @xmath0 for @xmath24 : the total integrated cross section @xmath25 and the integrated cross sections @xmath26 , where @xmath27 or 2 is the absolute value of the photon helicity difference . according the low theorem ,
the born contributions give the exact physical amplitude of the crossing reaction @xmath28 near its threshold .
if strong interactions of pions near the @xmath11 threshold are not too large ( this is the case owing to chiral symmetry ) , then the born contributions have to dominate near the threshold of the reaction @xmath0 as well . as is seen from fig .
1(a ) , this expectation does not contradict the data existing in the threshold region , though their errors are very large as yet .
moreover , the born amplitudes can be used as a quite reasonable approximation of the primary background ( nonresonance ) contributions in the reaction @xmath0 up to the @xmath4 resonance region , as well as a reasonable foundation for constructing the amplitudes involving the strong final - state interactions ; see , for example , @xcite .
the pure born contributions have the following features .
first , @xmath25 reaches its maximum at @xmath29 gev , in the vicinity of which @xmath30 .
however , @xmath31 involving the @xmath18 wave contribution decreases rapidly with increasing @xmath1 so that @xmath25 is fully dominated by the contribution from @xmath32 for @xmath33 gev ; see fig .
second , though @xmath32 is determined essentially by the @xmath34 wave contribution with @xmath35 , the differential cross section @xmath36 in the region @xmath24 changes very weakly ( see fig .
1(b ) ) , and it becomes more and more flat in this region of @xmath37 as @xmath1 increases .
therefore , in analyzing the data corresponding to this angular range , one should be very careful to make a definite conclusion about the partial wave structure of the smooth background .
in other words , the above example hints that the assumption of the @xmath18 wave dominance @xcite may be completely unjustified in reality .
figure 1(c ) shows only the belle data in the region @xmath38 to illustrate clearly the discovered signal from the @xmath3 resonance .
the visible height of the @xmath3 peak is of about 15 nb over the overall smooth background of the order of 100 nb , and its visible ( effective ) width is of about @xmath39 mev .
it is natural to consider that the large background under the @xmath3 resonance is mainly caused by the contributions from the born amplitude with @xmath35 and the @xmath4 resonance production amplitude also with @xmath35 .
figure 1(c ) shows the theoretical curves for the total integrated cross section @xmath40 and the integrated ones @xmath41 and @xmath32 . here
@xmath41 is the @xmath0 cross section with @xmath27 in which the contributions of the @xmath18 wave born amplitude are modified by the strong final - state interactions .
all the higher partial waves , @xmath34 , @xmath42 , etc .
, in @xmath41 are taken in the born approximation @xcite .
the above modification leads to a signal from the @xmath3 resonance in @xmath41 whose magnitude and shape are in surprising agreement with the belle data ; see fig .
an explicit form for @xmath41 will be given below .
the comparison of the curves in figs .
1(c ) and 1(a ) shows that the @xmath18 wave contribution to @xmath43 is small for @xmath33 in any case .
certainly , the @xmath4 resonance contribution is an important component needed for the description of the belle data in the whole region of @xmath1 from 0.8 to 1.5 gev . , and from crystal ball @xcite on the cross section for @xmath12 .
the theoretical curves correspond to the pure born cross sections , as well as the born cross sections modified for strong final - state interactions in the @xmath18 wave .
see the text for details . ]
the incorporation of final - state interactions in the @xmath18 wave born amplitude @xmath0 leads to a striking prediction for the @xmath18 wave amplitude @xmath12 @xcite .
figure 1(d ) demonstrates the comparison of the @xmath18 wave partial cross section for the reaction @xmath44 , @xmath45 , calculated in the way outlined above , with the crystal ball data @xcite . since @xmath45 does not contain any fitting parameter , the agreement with experiment in the region @xmath46 gev is quite reasonable .
note that , in the crystal ball experiments @xcite , the @xmath12 cross section was scanned with a 50-mev - wide step , and therefore the narrow @xmath3 structure could not be resolved .
it is also clear that in the reaction @xmath12 , as well as in @xmath0 , the @xmath4 resonance manifestation domain begins for @xmath47 gev .
the latest , high statistics belle data on the @xmath12 reaction cross section in the region @xmath48 gev are presented and discussed below in sec . v.
in the field theory approach , one has the following @xmath18 wave @xmath10 amplitudes , satisfying the unitarity condition and involving the electromagnetic born contributions `` dressed '' ( modified ) by strong final - state interactions @xcite
, @xmath49 \sin\delta^0_0(s)\right . & \nonumber\\ & \left.\pm12\alpha i_{k^+k^-}(s)|t_{k^+k^- \to\pi^+\pi^-}(s)|\right\ } & \nonumber \\ & + \frac{1}{3}e^{i\delta^2_0(s)}\left \{m^{\mbox{\scriptsize{born}}}_s(s)\cos\delta^2_0(s)+ \frac{8\alpha}{\rho_{\pi^+}(s)}\ , \mbox{re}[i_{\pi^+\pi^-}(s)]\sin\delta^2_0(s)\right\}\ , , & \end{aligned}\ ] ] @xmath50 \sin\delta^0_0(s)\right . &
\nonumber\\ & \left.\pm12\alpha i_{k^+k^-}(s)|t_{k^+k^- \to\pi^0\pi^0}(s)|\right\ } & \nonumber \\ & - \frac{2}{3}e^{i\delta^2_0(s)}\left \{m^{\mbox{\scriptsize{born}}}_s(s)\cos\delta^2_0(s)+ \frac{8\alpha}{\rho_{\pi^+}(s)}\ , \mbox{re}[i_{\pi^+\pi^-}(s)]\sin\delta^2_0(s)\right\}\ , , & \end{aligned}\ ] ] where @xmath51 is the @xmath18 wave born amplitude of the process @xmath5@xmath52@xmath8 ; for @xmath53 , @xmath54 @xmath55 , @xmath56 and @xmath57 are the phase shifts of the @xmath18 wave @xmath11 scattering amplitudes with isospin @xmath58 and 2 , respectively ( see below for details ) , and @xmath59 . the second equalities in eqs .
( [ mspm ] ) and ( [ ms00 ] ) are valid just in the elastic region , practically for @xmath60 , and clearly demonstrate the fulfillment of the watson theorem for the @xmath18 wave @xmath10 amplitudes with definite isospin . the function @xmath61 is given by @xmath62 ^ 2 - 1\,,\ \ & s\geq4m^2_{k^+}\ , , \\ \frac{m^2_{k^+}}{s}[\pi-2\arctan|\rho_{k^+}(s)|]^2 - 1\,,\ \ & 0\leq s\leq4m^2_{k^+}\ , , \end{array}\right.\end{aligned}\ ] ] @xmath63 for @xmath64 and @xmath65 if @xmath66 .
@xmath67 results from @xmath61 by replacing @xmath68 and @xmath69 by @xmath70 and @xmath71 , respectively .
finally , @xmath72 , @xmath73 , and @xmath74=@xmath75 are the @xmath18 wave amplitudes of hadronic reactions indicated in their subscripts . equations ( [ mspm ] ) and ( [ ms00 ] ) assume that the amplitudes @xmath76 and @xmath77 lie on the mass shell in the rescattering loops @xmath5@xmath52@xmath8@xmath52@xmath11 and @xmath5@xmath52@xmath9@xmath52@xmath11 .
the functions @xmath67 and @xmath61 are the attributes of the triangle diagrams @xmath5@xmath52@xmath8@xmath52@xmath78,@xmath79 and @xmath5@xmath52@xmath9@xmath52@xmath78,@xmath79 ( or any other scalars ) .
thus , the amplitudes in eqs .
( [ mspm ] ) and ( [ ms00 ] ) represent the @xmath18 wave @xmath5@xmath52@xmath8 and @xmath5@xmath52@xmath9 born contributions modified by the strong elastic and inelastic final - state interactions .
the helicity-0 cross section integrated over the range @xmath80 and involving the amplitude @xmath81 and the pure born higher partial wave amplitudes can be written in the form @xmath82 \biggr.\ ] ] @xmath83\right\}\,,\ ] ] where @xmath84 [ see eq .
( [ mspm ] ) ] and @xmath85 .
the cross section @xmath41 , shown , for example , in fig .
1(c ) by the dotted curve ( see also the solid curve in this figure and the next section for details ) , is given by latexmath:[\[\label{cs0}\sigma_0=\sigma_{\lambda=0}(\gamma\gamma\to\pi^+\pi^- , section for the reaction @xmath12 is given by latexmath:[\[\label{css00}\sigma_s(\gamma\gamma\to\pi^0\pi^0)= \frac{\rho_{\pi^+}(s)}{64\pi
s}\left cross section @xmath45 shown in fig .
1(d ) corresponds to @xmath88 . to construct @xmath41 and @xmath45 , the amplitudes @xmath72 , @xmath89 , and @xmath90
need to be known .
they are related to the @xmath18 wave @xmath11 scattering amplitude @xmath91 , the phase shift @xmath92 , and inelasticity @xmath93 with definite isospin @xmath94 in the conventional way : @xmath95/3,\qquad t_{\pi^+\pi^-\to\pi^0\pi^0}(s)=2[t^0_0(s)-t^2_0(s)]/3\,,\ ] ] @xmath96 - 1\}/[2i\rho_{\pi^+}(s)]\,,\ ] ] @xmath97 dots in @xmath98 denote the contributions from other inelastic channels @xmath99 , @xmath100 , etc . for @xmath101 ,
the amplitude @xmath102 for the @xmath103 channel contribution is small @xcite .
note that the plus sign is realized here according to ref .
we also set @xmath104 for all @xmath105 of interest .
a parametrization of the amplitudes @xmath106 and @xmath74 has been thoroughly described in ref .
it has been used for the simultaneous analysis of the data on the @xmath107 mass distribution in the @xmath19 decay , @xmath11 scattering in the region @xmath108 gev , and the reaction @xmath20 @xcite .
the key idea of this parametrization is that the amplitude @xmath106 incorporates the contributions from the mixed @xmath2 and @xmath3 resonances and from the nonresonant background having a large negative phase which hides the @xmath2 meson @xcite .
originally , the presence of such a background in @xmath11 scattering was established in the linear @xmath78 model @xcite .
it has been made clear that shielding of wide , lightest scalar mesons in chiral dynamics is very natural .
as for @xmath5 interactions , eqs .
( 1 ) and ( 2 ) transfer the chiral shielding effect of the @xmath2 from @xmath11 scattering to the @xmath10 reaction amplitudes @xcite .
the shielding of the @xmath78 meson takes place in the @xmath10 amplitudes for the strong destructive interference between the resonance and background contributions as in the @xmath109 amplitudes .
this was first demonstrated in the frame of the @xmath110 linear @xmath78 model in ref .
if such a shielding was absent , then the cross section @xmath45 would reach approximately 100 nb just above the @xmath107 threshold , owing to the @xmath8 loop mechanism of the @xmath6 decay @xcite , instead of about 10 nb as in experiment [ see fig .
1(d ) ] .
we now return to the parametrization of the strong amplitudes . in terms of the mixed @xmath2 and @xmath3 resonances and the necessary background contributions , the explicit form of the amplitudes @xmath106 and @xmath74
is given by @xcite @xmath111 @xmath112 } t^{k^+k^-\to\pi^+\pi^-}_{\mbox{\scriptsize{res}}}(s)\,,\ ] ] where @xmath113 and @xmath114 are the phases of the elastic background in the @xmath58 @xmath18 wave @xmath11 and @xmath115 channels , respectively , @xmath116 -1\}/[2i\rho_{\pi^+}(s)]$ ] is the @xmath58 @xmath18 wave @xmath11 background amplitude , the amplitudes of the @xmath117 resonance complex are @xmath118 @xmath119+g_{f_0\pi^+\pi^-}[d_{\sigma}(s)g_{f_0\pi^+\pi^-}+ \pi_{f_0\sigma}(s)g_{\sigma\pi^+ \pi^-}]}{d_\sigma(s)d_{f_0}(s)-\pi^2_{f_0\sigma}(s)}\,,\ ] ] and @xmath120 @xmath121+g_{f_0k^+k^-}[d_{\sigma}(s)g_{f_0\pi^+\pi^-}+ \pi_{f_0\sigma}(s)g_{\sigma\pi^+ \pi^-}]}{d_\sigma(s)d_{f_0}(s)-\pi^2_{f_0\sigma}(s)}\,,\ ] ] and the phase shift @xmath122 .
we use for @xmath123 , for propagators of the @xmath2 and @xmath3 resonances @xmath124 and @xmath125 , and for the nondiagonal matrix element of the polarization operator @xmath126 , the expressions presented in ref .
@xcite ( see also ref .
@xcite ) . in the accepted normalization the relation between the coupling constant @xmath127 and the corresponding partial decay width of the @xmath2
is given by @xmath128\,\rho_{\pi^+}(s)$ ] .
similar relations take place for the @xmath2 decays into @xmath115 , @xmath129 , @xmath130 , and @xmath131 , and for the @xmath3 decays into @xmath11 , @xmath132 , @xmath129 , @xmath130 , and @xmath131 .
remember that the @xmath2 couples mainly to @xmath11 , @xmath129 , @xmath130 , and @xmath131 , and the @xmath3 to @xmath115 , @xmath129 , @xmath130 , and @xmath131 @xcite .
the various fits corresponding to the different values of the parameters in the strong amplitudes have been considered in ref .
all of these fits give excellent descriptions of a large set of the data on the @xmath19 decay , @xmath133 , and @xmath98 .
the curves for @xmath41 and @xmath78 in fig .
1(c ) , and for @xmath45 in fig .
1(d ) , calculated with the use of eqs .
( [ mspm ] ) , ( [ ms00 ] ) , ( [ cs0pm ] ) and ( [ css00 ] ) , correspond to fit 1 from table i presented in ref .
@xcite ; see also ref .
for the phase shift @xmath57 , we take the parametrization of ref .
@xcite . according eqs .
( [ mspm ] ) and ( [ ms00 ] ) , the @xmath2@xmath52@xmath5 and @xmath3@xmath52@xmath5 decays are described by the @xmath8 and @xmath9 loop diagrams , resonances@xmath52(@xmath8,@xmath9)@xmath52@xmath134 [ @xmath67 , @xmath61 ] .
consequently , they are the four - quark transitions @xcite .
let us emphasize that there are no free parameters specific for the reactions @xmath10 in eqs .
( [ mspm ] ) and ( [ ms00 ] ) , and that the existing data are not indicative of the necessity of introducing such parameters . nevertheless ,
in the next section we shall supplement eqs .
( [ mspm ] ) and ( [ ms00 ] ) with the terms involving unknown direct coupling constants of the @xmath2 and @xmath3 resonances to @xmath5 , and we shall attempt to extract the values of these undoubtedly important physical characteristics from the data . note that some evidence for smallness of these constants was obtained previously in refs .
@xcite within the more simple models for the amplitudes @xmath10 .
later , a similar conclusion about the direct coupling of the @xmath2 to @xmath5 was also obtained in ref . @xcite .
we now add to the right - hand side of eq .
( [ mspm ] ) the amplitude @xmath136 caused by the contribution from the mixed @xmath2 and @xmath3 resonances @xcite with the direct coupling constants of the @xmath2 and @xmath3 to photons , @xmath137 and @xmath138 , @xmath139 @xmath140+g^{(0)}_{f_0\gamma \gamma}[d_{\sigma}(s)g_{f_0\pi^+\pi^-}+\pi_{f_0\sigma}(s)g_{\sigma\pi^+ \pi^-}]}{d_\sigma(s)d_{f_0}(s)-\pi^2_{f_0\sigma}(s)}\,.\ ] ] to the right - hand side of eq .
( [ ms00 ] ) , we also add the amplitude @xmath141=@xmath142 .
the factor @xmath105 in eq .
( [ mrd ] ) is due to gauge invariance .
the above amplitude also satisfies the unitarity condition . for @xmath143 ,
its phase coincides with the @xmath58 @xmath18 wave @xmath11 phase shift @xmath122 . of the existing data only the belle ones on the reaction @xmath0 in the vicinity of the @xmath3 [ fig .
1(c ) ] and the crystal ball data on the reaction @xmath12 for @xmath144 gev [ fig .
1(d ) ] may be sensitive to the coupling constants @xmath137 and @xmath138 .
therefore , to estimate @xmath137 and @xmath138 we perform a simultaneous fit to the crystal ball data in the above region of @xmath1 and the belle data for @xmath145 gev .
inclusion of the @xmath0 data from a sufficiently wide region around the narrow @xmath3 resonance in the fit is dictated by the following circumstance .
the @xmath3 peak is observed in the total cross section @xmath146 under the very large , noncoherent , smooth background caused by the contribution of the cross section with @xmath35 , i.e. , @xmath147 .
it is natural that the @xmath1 dependence of this background in the @xmath3 region can be fixed more or less reliably only with the use of the data outside this region .
certainly , @xmath148 is dominated by the born and @xmath4 resonance contributions
. however , first we use a purely phenomenological approximation of @xmath148 with a 4th order polynomial in @xmath1 in the region @xmath145 gev .
in addition , to obtain a correct fit to the belle data having the finest step in @xmath1 , we allow the @xmath3 resonance mass , @xmath149 , to be a free parameter .
we fix the values of the other parameters in the strong amplitudes in accordance with fit 1 from ref .
@xcite . .
they show that the direct couplings of the @xmath2 and @xmath3 resonances to @xmath5 are small .
the solid and dotted curves correspond to the cross sections @xmath146 and @xmath148 , respectively .
the right - hand plots emphasize the region of the @xmath3 peak .
see the text for details . ]
such a fit gives , in remarkable agreement with the prediction of ref .
@xcite , the negligible values of the direct coupling constants @xmath137 and
@xmath138 : @xmath150 kev and @xmath151 kev ; here @xmath152 gev @xcite . note , for comparison , that according to estimates presented in refs .
@xcite the @xmath153 decay width via the @xmath8 loop mechanism is of about @xmath154 kev for @xmath155 gev @xcite , and the @xmath7 decay width via the @xmath9 loop mechanism , averaged by the @xmath3 resonance mass distribution in the @xmath11 channel , is of about @xmath156 kev @xcite .
the results of fitting the belle data are shown in figs .
2(a ) and 2(b ) . for comparison , figs .
2(c ) and 2(d ) demonstrate the curves corresponding to the fit in which the crystal ball data are not taken into account and @xmath137=0 . in this case , @xmath157 kev and @xmath158 gev .
the fits to the belle data for @xmath137=@xmath138=0 are shown in figs .
2(e ) and 2(f ) ; here @xmath158 gev .
corresponding curves describing the crystal ball data in the region @xmath144 gev are not shown because , for the above three variants , they practically coincide with each other and with the curve in fig .
1(d ) .
signal in @xmath41 . the contributions from the @xmath159 and @xmath160 transition amplitudes and from the born and direct @xmath5 resonance decay amplitudes are presented .
the detailed description of the plotted curves is given in the text . ] thus , the available data on @xmath10 tell us that the direct couplings of the @xmath2 and @xmath3 to @xmath5 seem to be very small and that the @xmath6 and @xmath7 decays are in fact the four - quark transitions , because they are totally dominated by the @xmath8 and @xmath9 loop mechanisms , respectively .
to gain a complete understanding of the @xmath3 production mechanism in our model , we present in figs .
3(a)3(c ) all the main components shaping the @xmath3 signal in @xmath41 [ see eqs .
( [ cs0pm ] ) and ( [ cs0 ] ) ] using the fit shown in figs .
2(a ) and 2(b ) . above all
, we note that all the curves plotted in figs .
3(a ) and 3(b ) correspond to the different contributions to @xmath41 from the first term in the curly brackets in eq .
( [ cs0pm ] ) , @xmath161 with @xmath162 , that is , only from the contributions caused by the final - state interactions . the resulting picture involving the born and direct @xmath5 resonance decay contributions
( [ mspm ] ) , ( [ cs0pm ] ) , ( [ cs0 ] ) , and [ mrd ] ) ] is depicted in fig .
3(c ) . the crucial contribution from the @xmath163 transition amplitude , @xmath164 [ see eq .
( [ mspm ] ) , ( [ tkpi ] ) , and ( [ tkpir ] ) ] , to @xmath41 is shown by the solid curve in fig .
this contribution provides the natural scale of the resonance structure in @xmath41 in the 1 gev region .
the other curves in the figure represent its constituents .
the dashed ( dot - dashed ) curve corresponds to the contribution from the last ( first ) two terms in the numerator of eq .
( [ tkpir ] ) for @xmath165 , i.e. , only from the @xmath79 ( @xmath78 ) production in the @xmath9 channel ; see ref .
the dotted curve corresponds to the contributions from the last term in the numerator of eq .
( [ tkpir ] ) , i.e. , from the @xmath166 transition amplitude caused by the @xmath167 mixing . the dotted curve in fig .
3(b ) shows the contribution to @xmath41 from @xmath168 , i.e , from the @xmath18 wave resonance part of the @xmath160 transition amplitude [ see eqs .
( [ mspm ] ) , ( [ cs0pm ] ) , ( [ cs0 ] ) , ( [ tpipi ] ) , ( [ t00 ] ) , and ( [ tpipir ] ) ] . the dot - dashed curve in fig .
3(b ) corresponds to the contribution from @xmath169 , i.e , from the full @xmath18 wave @xmath160 transition amplitude [ see also eqs .
( [ mspm ] ) , ( [ cs0pm ] ) , ( [ cs0 ] ) , ( [ tpipi ] ) , ( [ t00 ] ) , and ( [ tpipir ] ) ] . as for the dashed curve in fig .
3(b ) , it is identical to the solid one in fig .
3(a ) and is shown for direct comparison of the @xmath159 and @xmath160 contributions . finally , the solid curve in fig .
3(b ) shows the total contribution to @xmath41 from the @xmath159 and @xmath160 rescattering amplitudes , @xmath170+@xmath171 .
a comparison of the dashed , dot - dashed , and solid curves in the figure gives a good idea of the important role of the interference between the background and resonance contributions . the dashed curve in fig .
3(c ) shows the total contribution to @xmath41 from the @xmath159 and @xmath160 rescattering amplitudes ; i.e. , it is identical to the solid curve in fig .
the dotted curve in fig .
3(c ) corresponds to the contribution from the above rescattering amplitudes plus the born contributions [ see eqs .
( [ mspm ] ) , ( [ cs0pm ] ) , and ( [ cs0 ] ) ] , and the solid curve in the figure represents the resulting picture of the @xmath3 resonance manifestation in @xmath0 , taking into account the contribution from the direct @xmath6 and @xmath7 decays , see eq .
( [ mrd ] ) and also ref .
@xcite .
we finish this section with a general comment . as already emphasized in ref .
@xcite , the complex residues of the @xmath78 pole in the @xmath109 and @xmath172 amplitudes , used to estimate the @xmath173 decay width @xcite , do not give us an idea about the nature of the @xmath78 meson . furthermore , as noted in ref .
@xcite , the majority of current investigations of the mass spectra in scalar channels do not study particle production mechanisms .
because of this , such investigations are essentially preprocessing experiments , and the derivable information is very relative . for example
, the very first estimate of the @xmath78 coupling to the photons via a two - pion intermediate state @xcite was restricted to the case of the `` bare '' ( without any background ) @xmath78 meson , which contradicts the low energy chiral dynamics , and the recent estimates of the @xmath7 decay width @xcite have not taken into account the rapidly changing @xmath9 loop production mechanism of the @xmath3 @xcite .
nevertheless , the progress in understanding the particle production mechanisms could essentially help us reveal the light scalar meson nature .
to estimate the @xmath13 decay width , @xmath174 , from the data on @xmath10 , it is usually assumed that the @xmath4 decay occurs mostly into @xmath5 states with @xmath35 @xcite .
for this , the specific models for the background amplitudes with @xmath175=2 and 0 are also needed @xcite . the large background under the @xmath4 in the @xmath0 channel [ see , for example , fig .
1(c ) ] is dominated by the born amplitude with @xmath175=2 .
the background situation in the @xmath12 channel is more pure . here , however , uncertainties in the data and bins of @xmath1 are still rather large .
different assumptions about the background amplitudes in the @xmath4 region have been used in the literature @xcite . in so doing ,
the central values of @xmath174 obtained in the independent experiments lie in the range from 2.3 to 3.6 kev @xcite . according to the particle data group ( pdg ) estimate @xcite , @xmath176 kev . in the recent work @xcite ,
the authors of the belle experiment presented `` a consistency check '' of their data in the @xmath4 region with this estimate .
they fixed the values of the @xmath4 resonance parameters as given by the pdg @xcite and , using a simple phenomenological parametrization of the background amplitudes , performed the fit to the data on @xmath177 , i.e. , for @xmath178 gev .
the resulting fit turned out to be very good , and they concluded that the `` consistency check is satisfactory . '' unfortunately , in ref .
@xcite the factor @xmath179 has been missed in the @xmath4 production amplitude in eq .
thus , the consistency is broken .
in fact , it follows from the belle data @xcite that @xmath180 kev , which is 1.5 times greater than the pdg estimate . at the same time , this value is in close agreement with our estimate , @xmath181 kev , which we obtained from the belle data , but in another way ( see below ) . to describe the belle data in the region @xmath16 gev
, we use the expression for the total cross section @xmath146 .
the cross section @xmath41 has been constructed in secs .
iii and iv , and the cross section @xmath148 involving the born and @xmath4 resonance contributions has the form @xcite : @xmath182 where @xmath183 $ ] is the born helicity-2 amplitude @xmath0 , @xmath184 , and the energy - dependent total width of the @xmath4 is given by @xmath185=@xmath186 .
the partial width @xmath187 is parameterized as @xcite @xmath188 where @xmath189=@xmath190 , @xmath191=@xmath192 , and @xmath193 is an interaction radius .
@xmath194 has the form similar to eq .
@xmath195 as a function of @xmath105 is approximated by the @xmath18 wave @xmath196 decay width ; see , for example , ref .
the branching ratios are @xmath197=0.847 , @xmath198=0.046 , and @xmath199=0.107 @xcite .
finally , @xmath200 by definition , @xmath201 . for @xmath202
we use the following parametrization : @xmath203 the second term in eq .
( [ g2 ] ) corresponds to the four - quark transition of the @xmath4 into photons via the @xmath8 real intermediate state , @xmath204 , where @xmath205\ ] ] is the born partial , helicity amplitude @xmath0 with @xmath206 .
this term ensures the fulfilment of the watson theorem requirement for the @xmath58 , @xmath206 amplitude @xmath10 in the elastic region .
it gives a rather small contribution to @xmath174 : @xmath207 it is generally accepted that the @xmath13 decay rate is dominated by the direct quark - antiquark transition @xmath208 , that is , by @xmath209 . as we saw above , for the lightest scalar mesons , the situation is reversed . in the fit , we use as free parameters @xmath210 , @xmath211 , @xmath212 , @xmath193 , @xmath213 , and also @xmath137 , @xmath138 , and @xmath149 .
the parameters @xmath193 and @xmath213 control the @xmath105 dependencies of the @xmath4 partial decay widths , and consequently , they are responsible for the shape of the @xmath4 line . the results of the fit to the belle data are shown in figs .
4(a)4(c ) .
figure 4(a ) demonstrates the overall picture of the most important contributions in the region @xmath214 gev ; here @xmath215 is the cross section corresponding to the @xmath4 resonance contribution and the dotted curve shows the contribution of the interference between the @xmath4 and background amplitudes in @xmath148 .
the descriptions of the belle data in the whole investigated region of @xmath1 and in the @xmath3 resonance region are presented in more detail in figs .
4(b ) and 4(c ) , respectively .
the parameters obtained are @xmath216 kev [ @xmath217 kev , see eq .
( [ wf2gg ] ) ] , @xmath211=1.279gev , @xmath218 gev , @xmath219 gev@xmath220 , @xmath213=3.52gev , @xmath137=0.482gev@xmath220 [ @xmath221=0.01kev ] , @xmath138=-0.867gev@xmath220 [ @xmath222=0.015kev ] , and @xmath149=0.975gev .
because of the smallness of statistical uncertainties in the belle data , the formally calculated errors of the above - listed parameters turn out to be negligible . in similar situations ,
the model dependence of the fitted parameter values is the most important source of their uncertainty .
cross section and the comparison with the crystal ball data on the @xmath12 cross section .
see the text for details . ]
the obtained description of the belle data as a whole seems to be quite satisfactory , except for minor details @xcite .
the more important result of the fit is that the values of the direct coupling constants of the @xmath2 and @xmath3 resonances to @xmath5 turn out to be small ( see the above - mentioned corresponding decay widths ) . of course ,
the obtained concrete values of these constants are , evidently , rather conditional ( compare , for example , the above fitting variant with that presented in sec .
iv ) . let us stress , however , that the very fact of the suppression of the direct @xmath2 and @xmath3 couplings to photons ( corresponding widths are much less than 1 kev ) can be considered to be well established .
in addition , it appears from the new belle data that @xmath223 is about a factor of 1.5 higher than the estimate quoted by the pdg @xcite .
we now construct the @xmath12 reaction cross section @xmath224 = @xmath225 , where @xmath226 is the @xmath4 production cross section in the @xmath107 channel ( an analog of @xmath215 for the @xmath8 one ) , and compare it with the crystal ball data @xcite .
the result is shown in fig .
as is seen , the agreement with the data is very poor in the whole region of the @xmath4 resonance influence , i.e. , for @xmath1 from 0.8 to 1.6 gev .
we verified that the parametrization of the @xmath4 contribution used by the belle collaboration @xcite leads to a similar resonance pattern . to improve the description of the available data on the reaction @xmath12 , it is necessary to raise the left wing of the @xmath4 resonance and to lower its right one @xcite , i.e. , to change the @xmath4 resonance shape in comparison with that established from the @xmath0 data .
as for the above difficulty , in fact , it arose as the first detailed experiments were carried out on the reactions @xmath12 @xcite and @xmath0 @xcite .
different parametrizations of the @xmath4 resonance shape have been used in refs .
@xcite . taking the corresponding formulas and fitting parameters from refs .
@xcite , we have made sure that in the early analyses @xcite the appreciably different shapes of the @xmath4 peak in the @xmath107 and @xmath8 channels were obtained .
the difference bears the above - mentioned character .
however , the existing uncertainties in the crystal ball @xcite , mark ii @xcite , and cello @xcite data [ see , for example , figs .
1(d ) and 1(a ) ] hamper any definite conclusions about their possible inconsistency or about the urgent need for searching the additional mechanisms to obtain a good simultaneous description of the @xmath107 and @xmath8 data in the @xmath4 region .
it is clear that the belle experiment on the reaction @xmath0 @xcite essentially aggravates the situation because its results are based on statistics which are about 3 orders of magnitude higher than those collected in the crystal ball experiments .
so , it is clear that , in the first place , very high quality data on the reaction @xmath12 and their partial wave analysis would be extremely useful to obtain reliable conclusions from the simultaneous description of the @xmath8 and @xmath227 channels @xcite . after this work was completed , very high statistics belle data on the reaction @xmath12 for @xmath228 gev
@xcite appeared , which are in close agreement with the crystal ball measurements @xcite .
probably , this implies that a damping form factor @xcite in the @xmath0 born amplitudes is really needed for the simultaneous description of the @xmath8 and @xmath107 production cross sections in the @xmath4 resonance region .
such an investigation requires considerable efforts , because the form factor influence should be taken into account in partial waves and in loop contributions .
we shall present a careful analysis of the compatibility of the new @xmath0 and @xmath12 data from belle elsewhere , together with comments on the choice of a suitable phenomenological form factor . here
we only announce some preliminary results of our analysis .
new high statistics results from belle on the @xmath12 reaction cross section @xcite are shown in fig .
5 . in spite of very small statistical errors @xcite
, these data can be quite satisfactorily described @xcite , separately from the data for the @xmath0 production ; see , as an example , the solid curve in fig .
5(b ) which corresponds to the parameters @xmath229 kev [ @xmath230 kev ] , @xmath211=1.272gev , @xmath231 gev , @xmath232 gev@xmath220 , @xmath213=0 , @xmath137=0.542gev@xmath220 , @xmath138=0.468gev@xmath220 , and @xmath149=0.969gev . however , such a fit to the @xmath12 cross section is in apparent contradiction with the @xmath0 data , see the solid curve for @xmath233 in fig .
this is caused mainly by the large born contributions to @xmath148 in the @xmath0 channel .
recall that such contributions are absent in @xmath12 .
thus , the situation can be essentially improved by multiplying the @xmath0 born amplitudes by some overall , damping form factor @xmath234 @xcite , where @xmath235 and @xmath236 are the usual mandelstam variables for the reaction @xmath0 . for this
we use here the expression proposed by poppe @xcite , @xmath237\,,\ ] ] where @xmath238 is a free parameter .
this ansatz is quite acceptable in the physical region of the reaction @xmath0 . replacing @xmath70 and @xmath238 by @xmath68 and @xmath239 , respectively
, we also obtain a form factor suitable for the @xmath240 born amplitude .
the solid curves for @xmath233 and @xmath225 in figs .
5(c ) and 5(d ) , respectively , show an example of the overall fit to the new @xmath0 and @xmath12 data , taking into account the form factors modifying the born contributions .
the obtained description is quite reasonable ( if not excellent ) , but only within systematic errors of the data , which are plotted in figs .
5(c ) and 5(d ) in the form of the shaded bands .
we think that such a treatment of the high statistics belle data is sufficiently justified .
statistical errors of the two belle measurements are so small that obtaining the formally acceptable @xmath241 for simultaneous fits to the data is practically impossible .
the curves in figs . 5(c ) and 5(d ) correspond to the parameters @xmath242 kev [ @xmath243 kev ; we consider this estimate for @xmath223 as the most preferable one ] , @xmath211=1.272gev , @xmath218 gev , @xmath244 gev@xmath220 , @xmath213=0 , @xmath245 , @xmath149=0.969gev , @xmath246 gev , and @xmath247 gev .
a comparison of figs .
5(b ) and 5(d ) shows that the impact of the form factor on the @xmath12 cross section turns out to be really small , in contrast to the case of the @xmath0 production ; see figs .
5(a ) and 5(c ) .
note also that all the above conclusions about the production mechanisms of the @xmath2 and @xmath3 resonances and a comment on the angular distribution for the nonresonance background , given in sec .
ii , remain valid .
@xcite and @xmath12 @xcite reaction cross sections .
the belle data in plots ( a ) , ( b ) , ( c ) , and ( d ) are represented by full squares with statistical error bars .
the shaded bands in ( c ) and ( d ) correspond to the belle data taking into account their systematic errors @xcite .
the curves in ( b ) and ( a ) correspond to the fit to the @xmath12 data and its consequence for the @xmath0 channel , respectively .
the curves in ( c ) and ( d ) correspond to the overall fit to the @xmath0 and @xmath12 data in the regions @xmath248 gev and @xmath249 gev , respectively , in the model with a form factor .
@xmath250 in ( c ) is the @xmath35 born cross section modified by a form factor .
see the text for details . ]
we have analyzed the new high statistics belle data on the reaction @xmath0 taking into account its main dynamical mechanisms .
the analysis has shown that the direct coupling constants of the @xmath2 and @xmath3 resonances to @xmath5 are small , which is typical @xcite for the four - quark nature @xcite of these states .
our main conclusion is that the @xmath6 and @xmath7 decays are the four - quark transitions that are dominated mainly by the @xmath8 and @xmath9 loop mechanisms , respectively .
in addition , we have presented some results of a simultaneous description of the @xmath0 data and the latest , very high statistics belle data on the reaction @xmath12 .
we have also estimated the @xmath13 decay width .
we intend to develop the above analysis to further understand the light scalar meson physics .
we thank y. watanabe very much for communication and for the table with the belle data for @xmath0 .
this work was supported in part by the presidential grant no .
nsh-5362.2006.2 for leading scientific schools and by the rffi grant no .
07 - 02 - 00093 from russian foundation for basic research .
99 t. mori _
et al . _ ( belle collaboration ) , in _ proceedings of the international simposium on hadron spectroscopy chiral symmetry and relativistic description of bound systems , tokyo , 2003 _ , edited by s. ishida , k. takamatsu , t. tsuru , s.y .
tsai , m. ishida , and t. komada ( kek proceedings 2003 - 7 ) ( kek , tsukuba , 2003 ) , p. 159 . t. mori _
et al . _ ( belle collaboration ) , phys .
d * 75 * , 051101 ( 2007 ) .
h. marsiske _
et al . , _ phys .
d * 41 * , 3324 ( 1990 ) .
et al . , _ phys .
d * 42 * , 1350 ( 1990 ) .
t. oest _ et al . , _ z. phys .
c * 47 * , 343 ( 1990 ) .
et al . , _ z. phys .
c * 56 * , 381 ( 1992 ) .
bienlein , in _ proceedings of the ix international workshop on photon - photon collisions , san diego , 1992 _ , edited by d. caldwell and h.p .
paar ( world scientific , 1992 ) , p. 241 .
r. barate _
et al . , _ phys .
b * 472 * , 189 ( 2000 ) .
achasov , s.a .
devyanin , and g.n .
shestakov , phys . lett . *
108b * , 134 ( 1982 ) ; z. phys .
c * 16 * , 55 ( 1982 ) .
n.n . achasov and g.n .
shestakov , phys .
d * 72 * , 013006 ( 2005 ) ; yad .
fiz . * 69 * , 1545 ( 2006 ) [ phys . at
. nucl . * 69 * , 1510 ( 2006 ) ] .
achasov , a.v .
kiselev , and g.n .
shestakov , nucl .
. suppl . * 162 * , 127 ( 2006 ) .
achasov , in _ proceedings of the 14th international seminar quarks-2006 , repino , st .
peterburg , 2006 _ , edited by s.v .
demidov , v.a .
matveev , v.a .
rubakov , and g.i .
rubtsov ( inr ras , moscow , 2007 ) , p. 37 ; arxiv : hep - ph/0609261 .
n.n . achasov and a.v .
kiselev , phys .
d * 73 * , 054029 ( 2006 ) .
n.n . achasov and g.n .
shestakov , phys .
lett . * 99 * , 072001 ( 2007 ) ; arxiv : 0704.2368 [ hep - ph ] .
n.n . achasov and g.n .
shestakov , phys .
d * 49 * , 5779 ( 1994 ) ; yad .
* 56 * , no . 9 , 206 ( 1993 ) [ phys . at
56 * , 1270 ( 1993 ) ] .
et al . _ ( belle collaboration ) , j. phys .
. jap . * 76 * , 074102 ( 2007 ) ; arxiv : 0704.3538 [ hep - ex ] . g. mennessier , z. phys .
c * 16 * , 241 ( 1983 ) .
lyth , j. phys .
phys . * 11 * , 459 ( 1985 ) .
johnson , ph.d .
thesis ( slac report no.-294 , 1986 ) .
d. morgan and m.r .
pennington , phys .
b * 192 * , 207 ( 1987 ) ; z. phys .
c * 37 * , 431 ( 1988 ) ; phys .
b * 272 * , 134 ( 1991 ) .
pennington , phys .
* 97 * , 011601 ( 2006 ) .
n.n . achasov and v.v .
gubin , phys .
d * 57 * , 1987 ( 1998 ) ; yad . fiz . * 61 * , 1473 ( 1998 ) [ phys . at
. nucl . * 61 * , 1367 ( 1998 ) ] .
the authors of refs . @xcite erroneously believe that the flatte work @xcite is related to the formula that they have use in their works for the @xmath3 propagator @xmath125 .
this formula for @xmath125 was first obtained in ref .
there is a misprint in this fit in the sign of the constant @xmath251 ; here we use @xmath252 .
n.n . achasov and g.n .
shestakov , phys .
d * 67 * , 114018 ( 2003 ) ; yad .
fiz . * 67 * , 1380 ( 2004 ) [ phys . at
. nucl . * 67 * , 1355 ( 2004 ) ] .
g. mennessier , p. minkowski , s. narison , and w. ochs , arxiv : 0707.4511 [ hep - ph ] .
this @xmath149 value is about 12 mev smaller than that obtained in fit 1 from ref .
@xcite , but in so doing the description of the @xmath11 phase shift @xmath133 essentially does not deteriorate .
this contribution from the @xmath78 production is negligible , and in fig .
3(a ) it is increased by a factor of 10 . the contribution from the direct @xmath5 resonance decay amplitude
@xmath136 is very small and that is why this is increased by a factor of 10 in fig .
3(c ) ( see the corresponding dot - dashed curve ) .
however , as is seen from this figure , the interference between the amplitude @xmath136 and the main ones is quite noticeable .
decelles and j.f .
goehl , phys . rev . *
184 * , 1617 ( 1969 ) .
( particle data group ) , j. phys .
g * 33 * , 1 ( 2006 ) . in principle , these details [ see figs .
4(b ) and 4(c ) ] can be artifacts associated with the event selection procedure employed .
it must be kept in mind that the systematic errors in the belle experiment are an order of magnitude higher than the statistical ones @xcite . here
we add yet an evident remark . in order to improve the data description ,
it is necessary to increase the flexibility of the phenomenological model in use . as a rule ,
this is accompanied both by new assumptions and by introducing additional free parameters .
formally , this is easy to implement , for example , by increasing @xmath193 and decreasing @xmath213 and @xmath211 relative to the @xmath8 case .
generally speaking , the consistent description of the crystal ball and belle data could be implemented by introducing a form factor into the born amplitudes .
such a form factor ( see refs .
@xcite for details ) suppresses the continuum in the @xmath4 region in the @xmath8 channel and , thus , permits the left wing of the @xmath4 resonance to be increased in agreement with the available @xmath8 and @xmath227 data .
m. poppe , int .
a * 1 * , 545 , ( 1986 ) .
n.n . achasov and g.n .
shestakov , mod .
lett . a * 9 * , 1351 ( 1994 ) .
( belle collaboration ) , arxiv : 0711.1926 [ hep - ex ] .
jaffe , phys .
d * 15 * , 267 , ( 1977 ) ; * 15 * , 281 ( 1977 ) .
flatte , phys .
b * 63 * , 224 ( 1976 ) .
achasov , s.a .
devyanin , and g.n .
shestakov , yad . fiz . * 32 * , 1098 ( 1980 ) [ sov .
. phys . * 32 * , 566 ( 1980 ) ] . | new high statistics belle data on the @xmath0 reaction cross section measured in the range of pion - pair invariant masses @xmath1 between 0.8 gev and 1.5 gev are analyzed to clarify the current situation around the @xmath2 , @xmath3 , and @xmath4 resonances in @xmath5 collisions .
the present analysis shows that the direct coupling constants of the @xmath2 and @xmath3 resonances to @xmath5 are small , and the @xmath6 and @xmath7 decays are four - quark transitions caused by the @xmath8 and @xmath9 loop mechanisms , respectively .
the chiral shielding of the @xmath2 resonance takes place in the reactions @xmath10 as well as in @xmath11 scattering .
some results of a simultaneous description of the @xmath0 and @xmath12 belle data are also presented . in particular , the following tentative estimate of the @xmath13 decay width is obtained : @xmath14 kev . |
lhc experiments provide the unique opportunity to explore some exotic states of heavy quarks . among them , particularly interesting states are those four - quark states consisting of double heavy quarks and double light antiquarks ( or their charge conjugator ) .
the existence of such states can be inferred from the heavy quark symmetry .
the double heavy quarks in the color anti - triplet state may form a diquark cluster by the attractive strong interactions . in the heavy quark limit ,
the double heavy quarks move slowly in a small relative velocity @xmath1 in the rest frame within a smaller distance ( @xmath2 ) , comparing to the size of light degree of freedom ( @xmath3 ) .
thus in the tetraquark states the double heavy quarks form a color anti - triplet diquark cluster .
it contributes the color interactions to double light antiquarks as a color source of a heavy antiquark .
the other two light antiquarks move around it with attractive interactions between them .
the dynamics of the light degrees of freedom of these tetraquark states are similar with those of the heavy baryons .
the picture is supported by scrutinizing the typical sizes of real hadrons .
the size of the heavy quarkonium is around @xmath4 fm while that of the light hadrons is around @xmath5 fm .
the masses of such states can be roughly estimated as the sum of two heavy quarks masses and @xmath6 .
for the tetraquark states containing the @xmath7 , @xmath8 , and @xmath9 quarks , their masses are around 3.4 gev , 6.8 gev , and 10.2 gev , respectively .
the flavor features of this sort of hadrons are very different from the conventional hadrons . once they are discovered in experiments
, it will be an undoubted evidence for the existence of the tetraquark states .
some theoretical studies on these states were presented in literatures@xcite .
the spin of a tetraquark state is a composition of the spins of the four quarks and the relative orbital angular momenta between them . for @xmath10waves , all orbital angular momenta
then the spin of the tetraquark state is just the composition of the spin of each quark or antiquark .
the composition of the spins of double heavy quarks may be 0 or 1 .
however , when two heavy quarks are identical , only spin 1 state is allowed due to the antisymmetry by exchanging identical fermions .
it is also the same case for the light antiquarks sector . in this paper
, we are interested in only the tetraquark states with all orbital angular momenta vanishing .
we denote states by @xmath11 ( @xmath12=0 , 1 ) , where @xmath13 and @xmath14 represent their heavy flavor indexes and @xmath12 is the spin of the double heavy quarks subsystem . as bound states with quite large masses they are difficult to be produced at usual high energy machines .
nevertheless , at lhc , they can be produced efficiently via a direct production process which involves effects happened at several distinct distance scales .
firstly , two heavy quark - antiquark pairs are produced via gluon - gluon fusion hard subprocess at distance scale @xmath15 or shorter .
secondly , for those produced two heavy quarks with small relative velocity @xmath1 there are certain probabilities to form a color anti - triplet diquark cluster at distance scale @xmath2 .
finally the diquark cluster evolves into the tetraquark state via the fragmentation process by picking up two light antiquarks from the vacuum at the distance scale @xmath3 . in this paper
, we calculate the hadronic production cross sections of @xmath16 , @xmath17 ( @xmath12=0 , 1 ) , and @xmath18 via gluon - gluon fusion process at lhc .
our results show that a number of these particles can be produced .
we also point out the signatures to detect those particles .
the rest of the paper are organized as follows . in sec .
[ sec2 ] , we present the calculation of the subprocess @xmath19 . sec .
[ sec3 ] devotes to the numerical results of the cross sections of the tetraquark states at lhc and at tevatron . in sec .
[ sec4 ] , we give some discussions about the results and the signals in the detectors .
as mentioned above , in the production process of the tetraquark states @xmath21 , there are three hierarchy distance scales , i.e. , @xmath22 .
accordingly , the cross sections of the subprocesses for the production of the @xmath10wave tetraquark states , @xmath23 , can be factored into three different parts accounting for physical effects happening at those distinct distance scales : @xmath24 where @xmath25 for @xmath26 and @xmath27 , and @xmath28 for @xmath18 and @xmath16 ; @xmath29 is the squared invariant mass of the double gluons ; @xmath30 is the lorentz invariant three - body phase space integral element ; @xmath31 is the short - distance coefficient describing the production rate of the color anti - triplet point - like @xmath32 state at the energy scale @xmath33 or higher ; @xmath34 is the wave function at the origin of the @xmath10wave diquark state ; @xmath35 is the fragmentation function of the diquark into the color - singlet tetraquark state @xmath23 .
the short - distance coefficient can be calculated by perturbative qcd and can be expanded in terms of @xmath36 at the short - distance energy scale @xmath33 or higher .
the rest parts are non - perturbative effects in nature .
to estimate the production cross sections , one needs to determine their numerical values .
the formation of the diquark cluster from the free double heavy quarks is described by the wave function at the origin .
their numerical values can be estimated by the potential model .
the diquark cluster provides an anti - triplet color source as a heavy antiquark .
thus , in the heavy quark limit , its fragmentation probability for forming the tetraquark states can then be approximately described by that for forming the heavy baryons by a heavy quark . in this paper
, we calculate the cross sections of hadronic production of the @xmath10wave states , @xmath16 , @xmath27 , @xmath37 , and @xmath18 , at lhc .
we compute the tree level short - distance coefficients @xmath38 in the leading order @xmath39 using perturbation qcd . by estimating the nonperturbative matrix elements , we carry out the numerical calculations of the total cross sections .
our results show that a number of these particles can be produced .
we first calculate the short - distance coefficients @xmath40 at the tree level .
they are proportional to the squared matrix elements @xmath41 .
to calculate the matrix elements , one needs to calculate the subprocess of @xmath42 , with @xmath43 and @xmath44 moving in the same @xmath45velocity and in the color anti - triplet state . at the tree level , the production processes involve 36 feynman diagrams for @xmath46 , and 72 ones for @xmath47 and @xmath48 .
the amplitudes can be classified as six gauge - invariant subsets in terms of six independent color bases . given this ,
the calculations of the amplitudes as well as their squares are straightforward .
@xmath34 s are the wave functions at the origin of the s - wave diquark clusters .
their precise values are difficult to be gained since the large range interaction potential between the double heavy quarks in the color anti - triplet state is not very clear while the short range one is dominated be coulomb potential . reasonably , we take values predicted by solving schdinger equation with coulomb potential , @xmath49 .
then the predicted @xmath50 s are 0.143 , 0.0382 , and 0.0198 gev@xmath51 for @xmath9 , @xmath52 , and @xmath7 diquark systems , respectively . including the confinement part ,
the wave function is squeezed to the central region and hence the wave function at the origin will be enhanced .
actually , we have done numerical calculations for the diquark state by solving the schrdinger equation using coulomb potential plus the linear potential fixed in the color - singlet case , the numerical value of wave function at the origin is enhanced by 10% - 30% . more reliable prediction for this nonperturbative number can be obtained by some nonperturbative method like lattice qcd .
we now turn to the fragmentation function of @xmath53-cluster to a tetraquark state .
as discussed above , the produced heavy diquark cluster in the color anti - triplet state provides the same color source as the heavy antiquark to the light antiquarks in the limit of the ratio of size of the diquark over that of the light antiquarks in the tetraquark state .
thus the fragmentation probabilities to produce the tetraquark states @xmath11 from the heavy diquarks are the same with that to produce the heavy baryons from the heavy quarks .
let s take a qed example to illustrate it .
imagine the hydrogen ion or deuterium ion passing material .
the probabilities forming the hydrogen atom or the deuterium one are the same if the velocities of both ions are the same since they possess the same electric charge . the fragmentation function is defined in the framework of the infinity momentum where the parton moves in the speed of light .
the fragmentation functions to produce the tetraquark states are nonperturbative in nature .
thus the shapes of them can only be described by certain phenomenological models @xcite .
one of the most commonly used models is the peterson model@xcite , in which the fragmentation function takes the following form : @xmath54 ^ 2 } , d_{\bar{3}}\to { t_{q_1q_2}^{i}}(x)=\frac{n}{x[1-(1/x)-\epsilon_{q_1q_2}/(1-x)]^2}\end{aligned}\ ] ] where @xmath55 is the only parameter determining the shape of the fragmentation function ; @xmath56 is the normalization constant .
once the fragmentation probability to produce the tetraquark state r is given , @xmath56 is fixed by the following condition : @xmath57 the fragmentation probabilities of @xmath58 and @xmath59 have been measured in @xmath60 collisions @xcite . according to pdg 2006 @xcite , @xmath61 , and @xmath62 .
these results are about 0.1 .
therefore as a good approximation we may take the fragmentation probability of @xmath63 , the value of @xmath64 in eq.([nor - sub ] ) , be 0.1 . from eqs.([frag - sub ] ) and ( [ nor - sub ] ) we know the normalization constant @xmath56 in the peterson model is dependent on the parameter @xmath65 .
the model suggested a scaling behavior for the parameter @xmath65 that is proportional to @xmath66 .
the @xmath67 determined by experiments is about @xmath68 @xcite . using the scaling behavior and taking @xmath67 to be 0.004
, we predict that @xmath69 , @xmath70 , @xmath71 and the corresponding normalization constants are 0.0075 , 0.0194 , 0.0047 . here
we take @xmath72 gev , @xmath73 gev and @xmath74 .
the fragmentation functions are energy scale dependent .
their convolution from one energy scale to another satisfies dglap equation .
however , for the hadronic production of the tetraquark states at lhc , the production is dominated by the smaller @xmath75 region of the produced particles , i.e. , compatible with the tetraquark masses . for smaller @xmath75 production of the tetraquark state
, there are no distinguishable energy scale differences .
thus the evolution of the fragmentation function used here is not significant .
therefore , in the numerical calculations we can take the fragmentation functions with the initial energy scale to be the tetraquark mass without evolution .
we now turn to the calculation of the total cross sections with @xmath76 of the subprocess given in eq .
( [ cs - sub ] ) . according to the parton model
, the cross sections of the processes @xmath77 , can be expressed as : @xmath78 where @xmath79 is the factorization energy scale , @xmath75 is the transverse momentum of the @xmath11 , and @xmath80 is the distribution function of the gluon in the proton .
here we use the cteq6l @xcite parton distribution function .
there is an uncertainty in the calculations arising from the choice of the factorization energy scale @xmath79 . here for comparison , we take two different values of @xmath79 , i.e. , @xmath81 and @xmath82/2 , where @xmath83 with @xmath84 being the transverse momentum of the diquark . with the choices of these factorization energy scales and
@xmath36 given in @xcite for the cteq6l @xcite parton distribution function , we calculate the total cross sections in cms , atlas and lhcb with the invariant mass of the @xmath85 system , @xmath86 tev and 14 tev , respectively .
the numerical results of the total production cross sections are listed in tables [ cross14tev ] and [ cross7tev ] .
the @xmath75 and the rapidity distributions are shown in figs.[cmscut]-[7ycut ] . from tables
[ cross14tev ] and [ cross7tev ] , we see that the total cross section to produce @xmath16 without @xmath87cut is much larger than the one with @xmath88 gev .
this can be explained with two reasons , i.e. , the small peak value in the @xmath75 distribution of the subprocess @xmath89 and the shape of fragmentation function of @xmath90cluster .
they also result in the lines of the @xmath87distributions across in figs.([cmscut ] ) , ( [ lhccut ] ) , ( [ 7cmscut ] ) , and ( [ 7lhccut ] ) . [ cols="^,^,^,^,^,^",options="header " , ] [ tevatron ] from table [ tevatron ] , we see that for @xmath91 integrated luminosity running at @xmath92 tev , for the production of @xmath93 , around @xmath94 events at tevatron can be accumulated with kinematic cuts @xmath95 gev and @xmath96 . for the production of @xmath97 , again by adding both @xmath98 and @xmath99 events together , we see that for @xmath91 integrated luminosity , for the production of @xmath97 , around @xmath100 events at tevatron can be accumulated with kinematic cuts @xmath95 gev and @xmath96 . for the production of @xmath101 , around @xmath102 events at tevatron can be accumulated with kinematic cuts @xmath95 gev and @xmath96 . there are some uncertainties in the predicted numerical results .
one arises from the ambiguity in choosing the factorization energy scale @xmath79 . from the tables
, we see that two different values of @xmath79 lead to around and less than factor 2 difference in the cross section .
the next leading order result will reduce this uncertainty .
however , it will be very hard calculation .
another one arises from the wave function at the origin , as discussed in sec .
[ sec2 ] , it will increase the cross sections by 20 - 70 % including the linear confinement potential between the heavy quark .
moreover , all excited heavy diquark states will decay into the ground diquark states by emitting photons or @xmath103 s . including those contributions ,
the production cross sections may be enhanced by 2 - 3 times . comparing the @xmath75 distributions at lhc and tevatron , we see that the differential cross sections decrease faster at tevatron than at lhc with the @xmath75 increase .
in calculating the accumulated events number , we take the integrated luminosity as the corresponding collider running at present or planning luminosity for about one year . from the above discussions , we see that the predicted event number of the tetraquark at the tevatron is about 3 - 4 orders of magnitude lower than that at lhc with @xmath104 tev for one year .
the events can be discovered at tevatron is much smaller than the ones can be discovered at lhc , especially when the energy and luminosity of lhc reach the maximum .
distributions of the tetraquark states , @xmath18 , @xmath16 , @xmath37 and @xmath27 , in cms and atlas with @xmath105=14 tev , with the pseudo - rapidity cut @xmath106 and @xmath79=@xmath107/2.,width=377 ] distributions of the tetraquark states , @xmath18 , @xmath16 , @xmath37 and @xmath27 , in lhcb at @xmath105=14 tev , with the pseudo - rapidity cut @xmath108 and @xmath79=@xmath107/2.,width=377 ] , @xmath16 , @xmath37 and @xmath27 , with @xmath79=@xmath107/2 , at lhc with @xmath105=14 tev.,width=377 ] distributions of the tetraquark states , @xmath18 , @xmath16 , @xmath37 and @xmath27 , in cms and atlas at @xmath105=7 tev , with the pseudo - rapidity cut @xmath109 and @xmath79=@xmath107/2.,width=377 ] distributions of the tetraquark states , @xmath18 , @xmath16 , @xmath37 and @xmath27 , in lhcb at @xmath105=7 tev , with the pseudo - rapidity cut @xmath108 and @xmath79=@xmath107/2.,width=377 ] , @xmath16 , @xmath37 and @xmath27 with @xmath79=@xmath107/2 , at lhc with @xmath105=7 tev.,width=377 ] distributions of the tetraquark states , @xmath18 , @xmath16 , @xmath37 and @xmath27 , at tevatron with @xmath105=1.96 tev , with the rapidity cut @xmath110 and @xmath79=@xmath107/2.,width=377 ] , @xmath16 , @xmath37 and @xmath27 with @xmath79=@xmath107/2 , at tevatron with @xmath105=1.96 tev.,width=377 ]
the decays of the tetraquark states with double heavy quarks possess unique signatures in the detectors . to analyze them in detail
is very useful for probing them in experiments . for the ground tetraquark states
, they can not decay by the strong and the electro - magnetic interactions .
however , they can decay through the weak interaction .
the inclusive decay can be regards as the cascade weak decays of the heavy quarks while the light antiquarks are spectators .
when the leptons can easily be identified in detectors , we focus on those semileptonic decays of the heavy quarks in the tetraquark states .
now we analyze the semileptonic decays of the @xmath111 states one by one . by weak interaction ,
the @xmath112 quark can decay to @xmath113 quark by emitting the @xmath114 or @xmath115 , and an invisible neutrino . in the @xmath16 , double @xmath112 quarks decay in this way independently , with the two antiquarks as spectators .
the emitted same plus sign double leptons and the @xmath75-distributions can be used to deduce the background . in the @xmath17 , there are two heavy quarks with different flavors . both the @xmath112 and the @xmath116 quarks have the semi - leptonic decay modes .
as discussed in the last paragraph , the @xmath112 quark can decay to @xmath113 quark by emitting , @xmath114 or @xmath115 , and an invisible neutrino .
the @xmath116 quark can decay to the @xmath112 quark by emitting the @xmath117 , @xmath118 , or @xmath119 lepton and the corresponding invisible anti - neutrino , followed by the @xmath112 quark decays as above . as a result , the two plus sign leptons and one lepton can be detected from those cascade decays , giving the special signature for identifying the @xmath17 .
the inclusive semi - leptonic decay of @xmath18 is somewhat more complicated .
each @xmath116 quark can decay to the @xmath112 quark with a lepton , @xmath117 , @xmath118 , or @xmath119 and the corresponding invisible anti - neutrino , followed by the @xmath112 quark decays as above .
these leptons , two leptons and two plus sign leptons , provide the unique signature for identifying the @xmath18 state .
the decay vertexes and the @xmath120tagging method will be very useful for identity those tetraquark states .
some other useful decays for identifying those tetraquark states are those 3-body or 4-body nonleptonic decays . considering the signatures and the efficiencies of the the detectors
, we see that lhc with high luminosity running at 14 tev provides best opportunity to discover doubly heavy tetraquark states . in conclusion ,
our results show that it is quite promising to discover these tetraquark states in lhc experiments both for large number events and for their unique signatures in detectors .
once they are discovered in experiments , it will be an undoubted evidence for the existence of the tetraquark states .
this will start a new stage in hadron physics .
+ * acknowledgments : * this work is partly supported by the nsfc with the contract 10875156 .
99 s. zouzou , b. silvestre - brac , c. gignoux and j. m. richard , z. phys .
c * 30 * , 457 ( 1986 ) .
j. vijande , a. valcarce , and j .-
richard , phys .
d * 76 * , 114013 ( 2007 ) .
l. heller and j. a. tjon , phys .
d * 35 * , 969 ( 1987 ) ; j. carlson , l. heller and j. a. tjon , phys .
d * 37 * , 744 ( 1988 ) ; a. v. manohar and m. b. wise , nucl .
b * 399 * , 17 ( 1993 ) ; s. pepin , fl .
stancu , m. genovese and j. m. richard , phys .
b * 393 * , 119 ( 1997 ) . j .-
ader , j .-
richard and p. taxil , phys .
d * 25 * , 2370 ( 1982 ) ; j. vijande , a. valcarce , j .-
richard and n. barnea , hep - ph/0902.1657 ( 2009 ) ; j. vijande , e. weissman , a. valcarce and n. barnea , phys .
d * 76 * , 094027 ( 2007 ) ; m. zhang , h. x. zhang and z. y. zhang , commun .
theor . phys . *
50 * , 437 ( 2008 ) .
ernest ma , james pantaleone and s. uma sankar , phys .
d * 46 * , 463 ( 1992 ) .
j. vijande , a. valcarce , and n. barnea , phys .
d * 79 * , 074010 ( 2009 ) .
c. peterson , d. schlatter , i. schmitt , and p. m. zerwas , phys .
d * 27 * , 105 ( 1983 ) . v. g. kartvelishvili , a. k. likhoded and v. a. petrov , phys . lett .
b * 78 * , 615 ( 1978 ) ; m. g. bowler , z. phys .
c * 11 * , 169 ( 1981 ) ; p. collins and t. spiller , j. phys .
g * 11 * , 1289 ( 1985 ) ; g. colangelo and p. nason , phys .
b * 285 * , 167 ( 1992 ) ; e. braaten , kingman cheung , sean fleming and tzu chiang yuan , phys . rev .
d * 51 * , 4819 ( 1995 ) .
particle data group , j. of phys .
g 33 ( 2006 ) , review of particle physics .
p200 , p770 .
sld collaboration , k. abe _
et al . _ , phys .
d * 65 * , 092006 ( 2002 ) ; delphi collaboration , j. abdallah _ et al .
_ , hep - ex/0311005v1 ( 2003 ) .
cdf collaboration , t. affolder _
et al . _ ,
* 84 * , 1663 ( 2000 ) ; anatoly adamov and gary r. goldstein , phys .
d * 56 * , 7381 ( 1997 ) .
sld collaboration , k. abe _
et al . _ ,
d * 53 * , 1023 ( 1996 ) ; aleph collaboration , a. heister _
et al . _ ,
b * 512 * , 30 ( 2001 ) ; the opal collaboration , g. abbiendi _
et al . _ , hep - ex/0210031v2 .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. nadolsky , and w. k. tung , hep - ph/0201195 .
@xmath121 , + @xmath122 , @xmath123 , @xmath124 , + @xmath125 , @xmath126 , @xmath127 , + @xmath128 and @xmath129 , when @xmath130 , + @xmath131 and @xmath132 , when @xmath133 . | we study the hadronic production of four - quark states with double heavy quarks and double light antiquarks at lhc .
the production mechanism is that a color anti - triplet diquark cluster consisting of double heavy quarks is formed first from the produced double heavy quark - antiquark pairs via @xmath0 fusion hard process , followed by the fragmentation of the diquark cluster into a four - quark ( tetraquark ) state .
predictions for the production cross sections and their differential distributions are presented .
our results show that it is quite promising to discover these tetraquark states in lhc experiments both for large number events and for their unique signatures in detectors .
keyword : tetraquark , diquark , heavy quark , fragmentation , production |
service - oriented architecture in wsns@xcite makes it possible to rapidly develop new applications . in a service - oriented wsn
, a typical application requires several different services , e.g. , data aggregation , data processing , decoding , which are provided by service providers that are also sensors .
the task of service composition is to assign each required service to an appropriate service provider according to certain criteria .
service composition with various performance metrics @xcite , e.g. , load balance , end - to - end delay and resource , have been well studied .
service composition in wsns has also recently been studied in @xcite .
@xcite studies the minimum - cost service placement based on service composition graphs with a tree structure .
@xcite considers the optimal placement of filters ( services ) with different selectivity rates .
habitat and environmental monitoring represent a class of wsn applications .
the queries in such applications in general are _ persistent _ ( or _ recurrent _ ) queries which need to be processed repetitively with a given frequency for a given duration @xcite , e.g. , an application requests receiving images in which the monitored area is dimly lit from 9:00am to 5:00pm@xcite .
three services are required in such a persistent query : service @xmath0 checking for dim images , service @xmath1 checking for `` sufficient '' motion between successive frames , and service @xmath2 fusing the identified motions(e.g .
, the appearance of a suspect ) . in a service - oriented wsn ,
such services are provided by sensor nodes in the network , thus , in - network processing is feasible , to reduce the possibly massive amounts of raw data . in wsns ,
energy consumption is a critical issue and sleep scheduling has been well studied as a conservative approach to save energy @xcite .
when a node is in sleep mode , all its provided services are not available , which may cause disruption to service composition .
@xcite studied a cross - layer sleep scheduling design in a service - oriented wsn which considers system requirement on the number of active service providers for each service at any time interval . in a service - oriented wsn ,
a query routing procedure which routes requesting services towards service providers is necessary . for a persistent query
, the query routing procedure might need to be conducted many times during its lifetime due to the sleep schedule in the mac layer , which might introduce more energy consumption .
take the query that starts at 09:00am and ends at 5:00pm with a frequency of 100s as an example . in fig .
[ example1](a ) , at 09:00am , a path is chosen to provide the requested services , while after 100s one of the sensors in this path switches into sleep mode , which results in unavailability of the service composition path .
it is necessary to conduct the query routing procedure again to find a new service composition path as shown in fig .
[ example1](b ) . in this paper
, we aim to use the minimum number of service composition solutions during a persistent query s lifetime such that the energy consumption caused by repetitively conducting query routing procedure is minimized .
once the minimum number of required service composition solutions is derived , we then select the service composition solutions with minimum transmission cost .
the contribution of the proposed work is summarized as follows : * we propose a service - oriented query routing protocol .
traditional routing in wsns only involves finding a path from source sensors to a sink .
service - oriented query routing protocol needs to ensure that the path from source sensors to the sink includes service providers , which imposes new challenges to routing in wsns . *
we propose an optimal greedy algorithm to minimize the number of required service composition solutions during a persistent query s lifetime , which can minimize the energy consumption caused by conducting the service - oriented query routing protocols .
* we propose a dynamic programming algorithm to minimize the total service composition cost which aims to reduce the transmission cost in executing a query . in a service - oriented wsn , width=240 ] the rest of the paper
is organized as follows . the network architecture and problem definition
are given in section [ network ] and [ prob ] respectively .
the algorithms and simulation results are presented in section [ algorithm ] and [ simulation ] respectively .
we conclude the paper in section [ conclusion ] .
in our network architecture , the service providers form a service provider overlay network as shown in fig .
[ model ] .
two service providers in the service provider overlay network may be multiple hops away from each other and the communication between them can be a multi - hop communication in the same wsn or through existing 802.11 wlan .
the service - oriented architecture at the sink has the following three layers : * _ service composition query layer_. this layer maps an application s query into a _ service composition query _ which specifies required services and their invocation order .
for example , the aforementioned query will be converted to a service composition query with services @xmath0 , @xmath1 and @xmath2 . * _ service layer_. this layer has the service information provided by the sensors in service provider overlay network .
we also assume that service layer has the sleep schedule information of service providers in service provider overlay network . * _ service composition layer_. this layer finds the service composition solutions for service composition queries , which is the problem to be studied in this paper . for a persistent query
, the service composition layer may find several service composition solutions during its lifetime since some service composition solutions may not always be feasible due to sleep schedule .
the service composition solutions are maintained in a _ service composition table _ as shown in fig .
[ model ] .
the service composition solution only specifies a service provider for each required service in a service composition query .
once the service composition solutions are identified , a routing protocol is invoked to find paths from source sensors to the first service provider in the service composition solution and find paths between any two adjacent service providers . in this paper , we propose the following service - oriented query routing protocol : * the sink broadcasts a _ service composition query routing _ message which includes service composition solution , duration , and interest .
such a message will reach all service providers in service provider overlay network . * upon receiving a _ service composition query routing _
message , if a service provider is the first service provider in the service composition solution , it will broadcast the interest to the sensor network .
source sensors can then send the data to the first service provider using any data - driven routing protocol in wsns .
thus , service composition is transparent to source sensors . * upon receiving a _ service composition query routing _
message , if a service provider is in the service composition solution but not the first service provider , it needs to find a path to its upstream service provider in the service composition solution .
this can be done by any routing protocol in wsns . during the lifetime of a persistent query
, it may be necessary to switch the service composition solutions due to the sleep schedule of service providers .
the service - oriented query routing protocol needs to be conducted again when the service composition solution changes , which consumes more energy .
the rest of the paper focuses on the service composition with minimum cost to avoid the frequent change of service composition solutions during a persistent query s lifetime .
notice that the service - oriented query routing protocol is a distributed routing protocol .
the sink only generates the service composition solutions which determines an appropriate service provider for each required service . to make such a decision
, the sink only needs to maintain the services availability and the sleep schedule information of each service provider . in a large - scale wsn ,
service providers are only a small portion of the whole network .
we believe that maintaining such information at the sink is well - paid when the duration of a persistent query is long .
let @xmath3 be a persistent service composition query and @xmath4 be service providers .
let @xmath5 be the set of services that sensor @xmath6 can provide and @xmath7 be the set of sensors that can provide service @xmath8 .
fig [ example2](a ) shows the service availability at the service layer .
given the duration @xmath9 and the frequency @xmath10 of a persistent query , the query should be executed for @xmath11 times during its duration @xmath9 and we assume that @xmath11 is an integer .
let @xmath12 be the start time of @xmath13-th execution of the persistent query where @xmath14 .
given the sleep schedule information of the service providers at the service layer , the sink can derive each service provider s availability at @xmath12 .
let @xmath15 be 1 if service provider @xmath6 is active at @xmath12 , otherwise , set @xmath15 be 0 .
figure [ example2](b ) gives the service provider availability at the service layer . with the service availability and the service provider availability information
, the service composition layer can derive a service composition solution at @xmath12 for @xmath16 . as shown in fig .
[ example2](c ) , the service composition solution @xmath17 is valid at @xmath18 and @xmath19 , @xmath20 is valid at @xmath21 and @xmath22 and so on . during this persistent query s lifetime ,
4 service composition solutions are required and thus the service - oriented query routing protocol needs to be conducted 4 times , which consumes energy .
this paper aims to minimize the number of service composition solutions for a persistent query . in a service - oriented wsn , width=268 ]
let @xmath23 be 1 if the service composition solution at @xmath12 is different from that at @xmath24 , otherwise let @xmath23 be 0
. then @xmath25 represents total number of service composition solutions during a persistent query s lifetime , which needs to be minimized .
under such an objective , a service composition solution may be used continuously in order to reduce the energy consumed by frequently invoking service - oriented query routing protocol .
although some service providers may be used continuously , this will not decrease the longevity of network . since if a service provider is to be active , it has to provide services for the system according to sleep scheduling .
though the service - oriented query routing procedure is the major source of energy consumption , the transmission of the data from the source sensor to the sink also consumes energy .
two service providers in the service provider overlay network may be multiple hops away and if the communication between them is through the same service - oriented wsn , relay sensors may also be in sleep mode .
thus , even a service composition solution can be used continuously over multiple executions , a local routing discovery procedure may be invoked between two service providers due to the sleep scheduling .
we use average transmission cost between two service providers to characterize such energy consumption caused by the local routing discovery between two service providers . besides minimizing the number of service composition solutions during a persistent query s lifetime , it is also important to minimize total transmission cost . in fig .
[ compositioncost ] , there are two sets of service composition solutions for a persistent query and both include 4 service composition solutions during the persistent query s lifetime .
thus , these two sets of service composition solutions consumes the same energy caused by service - oriented query routing procedure . in the first set ,
@xmath26 will be used for 3 times , twice , 3 times and twice respectively with a total cost of 184 . in the second set , @xmath26 will be used for twice , 3 times , 3 times and twice respectively with a total cost of 174 .
thus , the second set of service composition solutions will be more energy efficient . in this paper ,
firstly , we aim to minimize the number of service composition solutions during a persistent query s lifetime .
such a problem is referred to as problem * p1*. secondly , we need to minimize the total cost of the service composition solutions .
such a problem is referred to as problem * p2*.
[ algorithm ] in this section , we first approach problem * p1*. then based on the result of * p1 * , we approach the second problem * p2*. let @xmath27 be the number of executions that service provider @xmath6 can be continuously available from @xmath13-th execution ( including at @xmath13-th execution ) .
for example , if @xmath6 s availability at all execution instances of a persistent query is given as @xmath28 , @xmath29 is 2 since @xmath6 can be available at @xmath30st and @xmath31nd execution , @xmath32 is 0 as @xmath6 is not available at @xmath33rd execution .
the greedy algorithm which is shown in algorithm .
[ greedy ] is always to select the service provider with maximum @xmath27 for each @xmath8 in @xmath13-th execution such that the solution can be continuously used for the maximum number of times .
after the service composition solution is determined for @xmath13-th execution , the number of times that this solution can be used is determined by the minimum @xmath27 among all selected service providers .
let @xmath34 be the set of selected service providers for @xmath13-th execution and @xmath35 be the number of times that @xmath36-th service composition solution can be continuously used .
the worst case running time of this greedy algorithm is o(@xmath37 ) .
@xmath38 gives the minimum number of service composition solutions during a persistent query s lifetime .
we now prove the optimality of the greedy algorithm .
let @xmath25 be the solution obtained from the greedy algorithm where @xmath39 .
let @xmath40 be an optimal solution where @xmath41 .
[ lemma ] for any sequence @xmath42 and @xmath43 where @xmath44 , there must always exists @xmath45 .
we use induction to prove this lemma .
firstly , for @xmath46 , it is obvious that @xmath47 .
when @xmath48 , as greedy algorithm always selects the provider with maximum @xmath27 for each service , the value of @xmath49 must be no less than 0 , so @xmath50 .
assume that when @xmath51 we have @xmath52 .
for @xmath53 , in the given optimal solution , there is a service composition solution which can be continuously used from @xmath54 to @xmath55 . if @xmath56 , then we have @xmath57 ; if @xmath58 , then we must have @xmath59 since the greedy algorithm always selects the service providers which can provide longest continuous services . in both cases
, we have @xmath60 .
thus , lemma holds when @xmath53 .
[ theorem1]@xmath61 , the solution obtained from the greedy algorithm , must be optimal .
we prove it by contradiction .
assume that there exists @xmath62 , then @xmath63 .
according to lemma 1 , we also have @xmath64 .
the relationship among @xmath65 , @xmath66 , and @xmath67 is shown in fig . [ compare ] . and
@xmath68,height=57 ] [ compare ] fig .
[ compare ] denotes that there exists a service composition solution which can cover executions from @xmath67 to @xmath65 .
according to our greedy algorithm , we can find a solution which can be continuously used from @xmath67-th execution to the last execution .
thus @xmath69 , which conflicts the assumption . in the following , we approach problem * p2 * which minimizes the total routing cost based on the result of problem * p1*. let @xmath70 be the set of feasible service composition solutions at @xmath13-th execution which can be continuously used for the following @xmath71 executions . for any service composition solution @xmath72 , let @xmath73 be transmission cost if @xmath74 is selected to be executed once . let @xmath75 .
@xmath76 can be obtained by finding a shortest path in an auxiliary graph @xmath77 which is constructed as follows : * @xmath78 is the set of nodes consisting of @xmath79 layers @xmath80 and the @xmath81 layer @xmath82 contains all service providers which can continuously provide @xmath8 from @xmath13-th execution to @xmath83-th execution , e.g. , if @xmath6 can provide @xmath8 and it is available from @xmath13-th execution to @xmath83-th execution , node @xmath84 .
* let @xmath85 be the link set such that there is a direct link @xmath86 whenever @xmath87 and @xmath88 for @xmath89 .
the cost of @xmath90 is the shortest path cost from @xmath6 to @xmath91 in the physical network .
* add two special nodes @xmath92 and @xmath93 such that @xmath94 is the @xmath95 layer and @xmath96 is the @xmath97 layer .
link @xmath92 to each node in @xmath98 and link each node in @xmath99 to @xmath93 with cost 0 .
let @xmath100 be the minimum total cost from @xmath13-th execution to the last execution if @xmath36-th service composition solution starts at @xmath13-th execution .
then we have the following recursion : @xmath101 where @xmath102 .
we have the following boundary condition : @xmath103 for @xmath104 .
the dynamic programming is given in algorithm .
[ dynamic ] in which @xmath105 is the minimum total cost for the persistent query and @xmath106 $ ] stores @xmath36-th service composition solution .
the time complexity of the algorithm is @xmath107 .
in this section , we first introduce the design of our simulation . the number of service providers @xmath108 of each service is randomly generated between @xmath109 $ ] .
we then randomly generate @xmath108 service providers for @xmath8 from @xmath110 .
for each @xmath6 , we also randomly generate its availability at each execution . then we validate whether each @xmath8 can be provided by at least one active service provider at each execution . if infeasible , the instance is dropped from our simulation . to compare the performance of our algorithms
, we also introduce a baseline algorithm named _
min - cost - based _ algorithm which aims to select the service composition solution with minimum transmission cost for each execution .
we compare the number of service composition solutions during a persistent query s lifetime and the total transmission cost of our algorithms with _ min - cost - based _
algorithm respectively .
[ simulation ] , height=153 ] in the first set of experiments , we evaluate the performance of algorithms by varying @xmath11 in @xmath111 for @xmath112 .
the effectiveness of greedy algorithm for * p1 * is tested by comparing with _ min - cost - based _ algorithm . as shown in fig .
[ solutions ] , the number of service composition solutions during persistent query s lifetime in greedy algorithm is much less than that in _ min - cost - based _ algorithm . for example , with @xmath113 , the number of service composition solutions in our greedy algorithm is only 21 while it is 32 in _ min - cost - based _ algorithm .
the difference between two algorithms increases with the number of query s executions , which demonstrates the effectiveness and scalability of our work .
[ cost1 ] illustrates that total service composition cost obtained from dynamic programming based on the result of the greedy algorithm is higher than that obtained from _ min - cost - based _ algorithm . as we explained in section [ prob ] ,
the energy consumed in service - oriented query routing protocol is much higher than that in conducting service composition .
thus , though the solution obtained from our algorithms may consume more energy in the service composition phase , it consumes much less energy in service - oriented query routing phase which is the major energy consumption source in a persistent query . in the second set of our experiments
, we study in detail the impact of the number of required services on the total service composition cost and the impact of the number of service providers on the total service composition cost .
we have selected three scenarios @xmath114 , @xmath115 , @xmath116 by varying @xmath79 in @xmath117 $ ] . as shown in fig .
[ cost2 ] , the total service composition cost increases with the number of required services @xmath79 since more service providers may be involved in a service composition .
given @xmath79 , the service composition cost is lower in a network with more service providers . in a network with more service providers , more feasible service composition solutions are possible and our dynamic programming algorithm can find the service composition solution with minimum cost . ,
[ conclusion ] this paper studies service composition in service - oriented wsns with persistent queries .
we aim to provide service composition solutions during a persistent query s lifetime such that the involved routing update cost and transmission cost is minimized .
the optimality of greedy algorithm and dynamic programming provides the service composition solutions for persistent queries with the minimum energy consumption .
r. ha , p .- h . ho and x.s .
shen , cross - layer application - specific wireless sensor network design with single - channel csma mac over sense - sleep trees " , accepted by elsevier journal : computer communications . | service - oriented wireless sensor network(wsn ) has been recently proposed as an architecture to rapidly develop applications in wsns . in wsns
, a query task may require a set of services and may be carried out repetitively with a given frequency during its lifetime .
a service composition solution shall be provided for each execution of such a persistent query task . due to the energy saving strategy
, some sensors may be scheduled to be in sleep mode periodically .
thus , a service composition solution may not always be valid during the lifetime of a persistent query .
when a query task needs to be conducted over a new service composition solution , a routing update procedure is involved which consumes energy . in this paper , we study service composition design which minimizes the number of service composition solutions during the lifetime of a persistent query .
we also aim to minimize the total service composition cost when the minimum number of required service composition solutions is derived . a greedy algorithm and a dynamic programming algorithm are proposed to complete these two objectives respectively .
the optimality of both algorithms provides the service composition solutions for a persistent query with minimum energy consumption .
* keywords : * service composition , wireless sensor network , routing , query . |
clusters of 4 comprising several tens of atoms ( i.e. , of characteristic size of a few nm ) are perhaps the simplest quantum few - body system ; their investigation , both theoretical and experimental , has been pursued for a few decades now .
early theoretical studies yielded considerable information on their structure,@xcite bose condensation,@xcite excitations@xcite and their predicted superfluid behavior@xcite at temperatures of the order of 1 k. + probing experimentally their physical properties is obviously a difficult proposition , as pristine 4 clusters are very weakly bound , easily fragmented , e.g. , on impact with a scatterer .
thus , although some experiments were carried out , in which low energy 4 atoms were scattered off 4 clusters,@xcite most of what is now known about their physics , chiefly their superfluid properties , has come from spectroscopic studies of a single linear molecule ( e.g. , ocs ) embedded in them.@xcite + despite the remarkable microscopic insight afforded by this approach , the presence of the embedded impurity , whose linear size is typically that of a few 4 atoms , is expected to alter drastically the structure of a cluster of such a small size . specifically , a significant fraction of the 4 atoms are predicted to bind to the foreign molecule , while the rest arrange in shells around it.@xcite this is very different from the featureless structure of a pristine 4 cluster.@xcite the question therefore remains open as to whether one could probe experimentally the superfluid properties of a finite assembly of 4 atoms in a way that does not make use of a foreign impurity , and preserves at least the full rotational invariance of the cluster .
+ aside from possible experiments on free standing clusters , with the aforementioned difficulties that they entail , one possibility that does not seem to have been given much consideration so far is that of studying 4 clusters enclosed in nanometer size cavities , e.g. , those of a suitably chosen porous material .
for example , one could think of adsorbing 4 inside zeolites , whose pore diameter is typically of the order of a nm , and therefore may accommodate clusters of a few atoms .
@xcite one could therefore think of investigating the physics of the clusters , e.g. , by performing neutron scattering experiments on the system , on the assumption that the bulk of the signal should come from helium confined inside the ( relatively regularly shaped ) cages . obviously , the presence of a confining surface is likely to affect significantly the structure and physical behavior of the cluster , compared to those of a free one .
for example , one may expect superfluidity to be suppressed , with crystallization originating at the surface of the cavity , and then extending to adsorbed layers .
+ while a classical hard - sphere fluid in spherical confinement has been the subject of much theoretical investigation,@xcite surprisingly little is known about the effect of this type of confinement on the superfluid transition of 4 .
a number of studies of superfluidity of 4 in narrow cylindrical channels have been carried out , @xcite typically aimed at modeling 4 in porous media , or in the confines of carbon nanotubes.@xcite theoretical studies of adsorption of 4 on spherical substrates have mostly focused on layering on the outer surface of macromolecules such as fullerenes;@xcite however , the superfluid properties of 4 clusters confined inside spherical cavities of nanoscale size remain largely unexplored .
a recent theoretical study of parahydrogen clusters confined inside spherical cavities of nanoscale size has yielded surprising evidence of enhanced superfluid response , with respect to free clusters,@xcite showing that the curved surface can affect in nontrivial ways the behaviour of the system .
+ in this work , we carried out a theoretical study of structural and superfluid properties of clusters of 4 confined inside a spherical cavity of diameter 2 nm . specifically , we performed equilibrium quantum monte carlo simulation of a reasonably realistic model of the system at low temperature ( down to @xmath0 k ) .
we considered the two limiting cases or strong and weak substrate , in order to differentiate geometrical effects from those arising from the specific nature of the adsorbing medium .
our purpose is twofold : first , we aim at gaining some further insight in the effect of nanoscale confinement on the physical properties of 4 clusters ; second , we wish to furnish some definite theoretical predictions , helping in the design and interpretation of experiments on 4 adsorbed inside porous media whose open volume consists primarily of ( quasi)spherical cavities of characteristic size of few nm . +
as expected , confinement drastically alters the structure of the clusters , in comparison to free standing ones , imparting clusters a well - defined shell - like structure ; nevertheless , superfluidity is resilient , and still observable within such tight confines at temperatures of order 1 k , under relatively broad conditions .
the character of the superfluid response changes considerably , depending on the strength of the substrate . on a strong substrate , a quasi-2d superfluid film coats the surface ,
undergoes crystallization as the density increases , and no reentrant superfluid phase is seen as a second layer ( i.e. , a concentric shell ) forms . on a sufficiently weak substrate , on the other hand , the system forms two concentric superfluid shells , reminiscent of what seen for parahydrogen in ref . .
signatures of such different physics can be detected in the single - particle momentum distribution , probed by neutron scattering , a technique which has offered a great deal of insight into the physics of 4 and other quantum fluids ( like parahydrogen ) in the confines of porous media.@xcite + in the next section , we introduce the mathematical model of the system , and provide some basic details of the simulational technique utilized .
we then illustrate our results and outline our conclusion in the following two sections .
our system of interest is an ensemble of @xmath1 4 atoms , regarded as point particles of spin zero , enclosed in a spherical cavity .
the quantum - mechanical many - body hamiltonian is the following : @xmath2 here , @xmath3 , @xmath4 being the mass of a 4 atom , @xmath5 is the positions of the @xmath6th atom , measured with respect to the center of the cavity ( set as the origin ) , @xmath7 , @xmath8 describes the interaction of a pair of atoms , whereas @xmath9 describes the confinement of each atom inside the cavity .
for @xmath8 we used the accepted aziz potential,@xcite whereas for @xmath9 we used the expression@xcite @xmath10 where @xmath11 , @xmath12 , @xmath13 and @xmath14 and @xmath15 are two parameters that are adjusted to reproduce the main 4 adsorption features of specific substrates .
( [ potl ] ) is merely the extension to the case of a spherical cavity of the so - called 3 - 9 " potential , describing the interaction of a particle with an infinite , planar substrate .
@xmath15 is the depth of the attractive well of the potential experienced by a 4 atom in the vicinity of the substrate , whereas @xmath14 is essentially the distance of closest approach of a 4 atom to the substrate .
+ as a model of 4 in the confines of , e.g. , the cages of zeolites , eq .
( [ one ] ) clearly contains simplifications , chiefly the fact that the cavity is regarded as perfectly spherical and smooth .
however , it allows us to address the physical question that we wish to pose here , namely the effect of confinement on the superfluid response of a 4 cluster .
equivalent , or even simpler models ( e.g. , cavities with hard walls ) have been utilized to study structure of @xmath16he and classical fluids in confinement.@xcite + we obtained results for a cavity of radius @xmath17 , but with two distinct choices of the parameters @xmath15 , @xmath14 in eq .
( [ potl ] ) , corresponding to very different adsorption properties .
the first choice , henceforth labeled with li , has @xmath15=17.87 k and @xmath14=3.76 ; these values are apt to describe the interaction of a 4 atom with a li substrate , one of the most weakly attractive known , on which 4 is predicted to form a superfluid monolayer at low temperature.@xcite + the second choice , namely @xmath15=100 k and @xmath14=2.05 , is roughly in the ballpark of what one would expect for 4 atoms near a silica substrate;@xcite thus , we shall henceforth refer to the scenario described by this parameter set as glass .
the considerably deeper well , together with a much shorter range of the repulsive core , render the glass " much more attractive to 4 atoms than the li cavity .
obviously , in neither case do we aim at reproducing accurately any realistic interaction ( which would require more elaborate functional forms anyway ) .
rather , our aim is that of investigating opposite ends of the adsorption continuum .
+ we studied the low temperature physical properties of the system described by eqs .
( [ one ] ) and ( [ potl ] ) by means of first principle computer simulations based on the worm algorithm in the continuous - space path integral representation.@xcite specifically , we used a variant of the method which allows one to perform calculations in the canonical ensemble ( i.e. , fixed @xmath1).@xcite because this well - established computational methodology is thoroughly described elsewhere , we shall not review it here . the most important aspects to be emphasized here , are that it enables one to compute thermodynamic properties of bose systems at finite temperature , directly from the microscopic hamiltonian , in particular energetic , structural and superfluid properties of the confined 4 fluid , in practice with no approximation .
technical details of the simulation are standard , and we refer the interested reader to ref . .
all of the results reported here were obtained with a value of the imaginary time step @xmath18 k@xmath19 , with a high - temperature approximation for the many - particle propagator accurate up to order @xmath20 .
we begin the illustration of the results of the simulations by discussing the computed energetics .
we obtain ground state estimates by extrapolating to @xmath21 low temperature results . in pratice
, we find that energy values , as well as radial density profiles , remain unchanged , on the scales of the figures shown here , below @xmath22 1 k. + fig .
[ ene ] displays the energy per 4 atom ( in k ) as a function of the number of atoms in the cavity .
both curves feature minima at specific numbers of atoms , which correspond to the minimum filling of the cavity .
polynomial fits to the data shown in fig .
[ ene ] ( which , for a glass cavity include energy estimates up to @xmath1=110 , not shown for clarity in the figure ) , yield a minimum for a li ( glass ) cavity at approximately @xmath23 ( @xmath24 ) 4 atoms , at an energy close to @xmath25 k ( @xmath26 k ) , which can be compared to the value @xmath27 k ( @xmath28 k ) yielded by the corresponding 3 - 9 potential for a flat substrate.@xcite stronger atomic binding arises from the curved confinement , and it is interesting to note that the enhancement is proportionally greater for the weaker substrate .
+ the same fits show that the chemical potential , obtained as @xmath29 becomes comparable to that of bulk superfluid 4 ( @xmath30 k ) for @xmath31 for a li cavity , and @xmath32 for glass . this estimate is relevant to possible experiments in which a porous matrix in which 4 is adsorbed
is held in thermal contact with a bath of superfluid 4 . inside a glass substrate , at those physical conditions
, we estimate the kinetic energy per 4 atom to be @xmath33 k. this value is in the same ballpark as the most recent experimental estimates@xcite of the mean atomic kinetic energy of 4 adsorbed in cylindrical nanopores of diameter 24 , but the very large ( @xmath34 20% ) uncertainty quoted therein renders such a comparison scarcely meaningful . in general , one might expect that the kinetic energy to be mostly affected by the characteristic size of the confining medium , not so much by its specific geometry . for a cluster of 45 atoms inside a li cavity , we find the kinetic energy per 4 atom to be @xmath35 k at @xmath36 k. [ t ] ( in k ) versus number @xmath1 in the @xmath37 limit , inside a li ( upper panel ) and glass ( lower panel ) cavities of radius 10 . dashed lines are polynomial fits to the data .
statistical errors are at the most equal to symbol size . ]
[ h ] = 1 k for a cluster of 25 4 atoms inside a li cavity ( boxes ) and for 35 4 atoms inside a glass cavity ( circles ) .
the origin is at the center of the cavity .
statistical errors are at the most equal to symbol size . ]
[ h ] = 1 k for a cluster of 45 4 atoms inside a li cavity ( boxes ) and a glass cavity ( circles ) .
the origin is at the center of the cavity .
also shown is the result for a free 4 cluster ( diamonds ) , computed with respect to its center of mass .
statistical errors are at the most equal to symbol size . ]
[ rhoeq ] shows radial density profiles @xmath38 , computed with respect to the center of the cavity , for a cluster of 25 atoms inside the li cavity ( boxes ) , and for one of 35 inside the glass one ( circles ) , i.e. , in both cases close to the equilibrium filling .
clearly , the arrangement of the particles in the two cases is very different , reflecting the different strengths of the adsorption potentials .
inside the more attractive glass cavity , 4 forms a single thin film ( a spherical shell ) coating the surface , whose effective 2d density at the minimum ( @xmath39 ) can be estimated to be @xmath40 @xmath41 , based on the position of the peak of @xmath38 .
this value is only slightly greater than the 4 2d equilibrium density.@xcite the structure of the adsorbed film comprises a single shell . and
this remains the case for @xmath42 , at which point the second shell begins to appear .
this corresponds to a surface coverage of approximately 0.100 @xmath41 where second layer promotion begins to occur , quantitatively consistently with what observed on a flat substrate.@xcite + inside a li cavity , on the other hand , two largely overlapping , floppy concentric shells form , as atoms experience a repulsive core extending over 3 away from the surface .
thus , even though 4 does form a monolayer on a flat li substrate,@xcite inside a nm - radius spherical cavity the energetically favored arrangement is droplet - like .
this result mimics what observed for parahydrogen clusters in the same setup , on a weakly attractive substrate.@xcite fig .
[ hig ] shows density profiles for systems comprising @xmath43 4 atoms inside the two cavities , at @xmath36 k. the change is minimal on a glass substrate , as the 4 film coating the surface of the cavity becomes compressed , whereas the two shells inside the li cavity become noticeably more rigid , especially the inner one , and a clearer demarcation between them emerges , although substantial atomic overlap in the region between the two shells remains .
the structure of these clusters is clearly very different from that of a free cluster , as illustrated in fig .
a free 4 cluster is very losely bound , featureless and extends out all the way to approximately 12 from its center of mass.@xcite the absence of structure for the free cluster is in agreement with previous finite temperature simulations.@xcite the interesting issue , of course , is how confinement , which modifies so significantly their shape , affects the superfluid properties of these clusters , and whether one might be able to probe them experimentally .
[ t ] = 1 k of clusters of varying numbers of 4 atoms enclosed inside a li ( boxes ) and glass ( circles ) cavity of radius @xmath44 .
when not explicitly shown , statistical errors are at the most equal to symbol size . ] fig .
[ rhos ] shows the superfluid fraction @xmath45 computed using the well - knownarea " estimator@xcite at temperature @xmath36 k for clusters of varying number of atoms , starting from @xmath1=20 , at the lower limit of thermodynamic stability for a li cavity , to @xmath46 .
we also computed the same quantity for free 4 clusters comprising @xmath43 and 55 atoms , both yielding approximately 82% , i.e. , essentially the same value reported by sindzingre it et al .
for @xmath1=64 at this temperature . + the superfluid response is quantitatively different inside the two cavities . specifically , on the weakly adsorbing li substrate the two - shell structure that forms ( figs .
[ rhoeq],[hig ] ) displays a superfluid fraction in excess of 80% , remarkably constant ( within statistical uncertainties ) for @xmath47 , declining rapidly above @xmath48 and becoming negligible for @xmath49 , when the cluster acquires a solid - like structure with the appearance of atoms at the center of the cavity . inside a glass cavity ,
the single film coating the surface has a finite superfluid response at the equilibrium coverage , quantitatively similar to that inside a li cavity , for the same number of atoms . as the number of atoms is increased , the film is compressed and the superfluid response weakens , becoming negligible for @xmath50 , which corresponds to a 2d coverage around 0.072 @xmath41 , in quantitative agreement with the location of the @xmath21 melting density in 2d.@xcite indeed , even inside a glass cavity of such narrow diameter the physics of the adsorbed 4 is largely 2d .
[ h ] versus temperature for a cluster of @xmath43 4 atoms enclosed inside a li ( boxes ) and glass ( circles ) cavity of radius @xmath44 .
also shown is a result for a free 4 cluster ( diamond ) , computed with respect to its center of mass . when not explicitly shown , statistical errors are at the most equal to symbol size . ] + no reentrant superfluidity at higher density is seen in either cavity ; on the glass substrate , that means that inner shells are not superfluid , at least not at the lowest temperature considered here , namely @xmath51=0.5 k. this suggests that a glassy confining medium of the characteristic size considered here may be too tight for superfluidity to be observed , at least in an experiment in which a porous glass filled with helium is in equilibrium with bulk superfluid .
+ naturally , there is an objective limit to how closely the simple , spherical model of confinement considered here can reproduce results of experiments aimed at investigating superfluid properties of helium inside porous media , whose microscopic structure is typically one of interconnected cylindrical channels .
however , this result is consistent with experimental evidence indicating that a characteristic confining length of the order of 25 may correspond to the lower limit of existence of superfluidity of 4 in porous media.@xcite + fig .
[ rsoft ] shows our computed values of the superfluid fraction @xmath52 as a function of temperature , for a cluster comprising @xmath43 4 atoms , enclosed in the two spherical cavities considered here .
also shown for comparison is the corresponding value for a free 4 cluster.@xcite although the superfluid signal is weaker than that which one would observe in a finite cluster , nonetheless it remains robust , comparable values of the superfluid fraction being observed only at a fraction of a k lower temperature . within the statistical uncertainties of the calculation , the estimates obtained for this specific cluster are indistinguishable ,
this is peculiar , considering that the structure of the two clusters is very different , as shown in fig .
[ hig ] . on a glass substrate
, the character of the superfluid transition can be expected to be largely 2d ; in the li cavity , on the other hand , there are two shells , with a rather clear demarcation between the two , even though , as remarked above , some finite atomic overlap exists .
+ one might be inclined to think that in the less attractive li cavity the superfluid response is essentially that of two almost independent 2d shells , or radii @xmath53 and @xmath34 6 .
however , it should be noted that the numbers of particles in the two shells are 9 and 36 respectively ,
i.e. , the effective 2d density is approximately 0.08 @xmath41 for both , i.e. , well into the crystalline region of the bulk phase diagram.@xcite moreover , the rapid decrease of the superfluid response for @xmath54 concides with the sharp reduction of the overlap between the two shells , signaling that quantum - mechanical exchanges of atoms in different shells play a vital role in underlying the superfluid response of these clusters .
we computed the local superfluid density@xcite and consistently found it to be homogeneous throughout the cluster , i.e. , not concentrated mostly in some parts of it ( e.g. , one of the two shells ) .
this physical behaviour closely resembles that observed in simulations for parahydrogen clusters trapped inside nanoscale spherical cavities.@xcite [ h ] computed for a cluster of @xmath43 4 atoms enclosed inside a li ( lower panel ) and glass ( upper panel ) cavity of radius @xmath44 .
dashed lines refer to data at temperature @xmath51=2 k , solid lines at @xmath51=0.5 k. ] as mentioned in the introduction , one of the main reasons for studying the physics of 4 clusters in confinement is that one might be able to access their superfluid response more easily and directly than by performing spectroscopy of molecular probes embedded in free clusters .
for example , one may measure the momentum distribution , typically by scattering neutrons off a sample of 4 adsorbed inside some suitably chosen porous medium . on the assumption that the bulk of the signal should come from 4 inside the cavities , one may look for the appearance at low temperature of a peak at zero momentum,@xcite which signals the onset of bose - einsten condensation , intimately connected to superfluidity .
[ nq ] shows the spherically average single - particle momentum distribution @xmath55 , computed for a cluster of 45 4 atoms inside the two cavities that we considered in this work , at the two temperatures @xmath51=2 k and @xmath51=0.5 k. the appearance of a peak at short momenta as the temperature is lowered below @xmath56 k , i.e. , when the clusters turn superfluid , is clear .
the peak is not sharp as it would be in bulk superfluid , but rather broadened by the fact the system is confined over a length of @xmath34 1 nm . in contrast , no such a peak develops if crystallization occurs inside the cavities .
there is a quantitative difference between the signal observed in the glass cavity , inside which superfluidity is essentially 2d ( as mentioned above ) , and in the li one .
the peak inside the glass cavity is stronger , @xmath55 slightly more structured and falls off more slowly at high @xmath57 than inside the li cavity , due to the localization of atoms in the vicinity of the cavity surface .
the purpose of this study has been that of gaining insight into the physical behaviour of nanoscale 4 clusters inside a spherical cavity , with particular attention to the superfluid properties .
the idea is that of providing a possible , alternate avenue to the exploration of clusters , one that does not involve the use of a foreign probe embedded in the cluster . obviously , the presence of the confining surface has a significant effect on the structure of the clusters , but the main result of this work is that superfluidity can still be observed at temperatures comparable to those at which its onset is predicted for free clusters .
+ we have focused on spherical cavities of radius 1 nm , because they can accomodate clusters of a few tens of atoms , which is an interesting size because of the strong competition between surface and bulk energetics .
we have shown that in a cavity carved inside a medium like silica , which exerts a strong attraction on 4 atoms , superfluidity can only occur in the form of a film coating the surface of the cavity , at equilibrium density . in an experiment
in which the system is in thermal contact with a bath of superfluid 4 , our calculations show that cavities will be filled with frozen 4 , i.e. , no superfluid signal will be seen , at least down to @xmath0 k. on the other hand , in a weakly attractive cavity , especially one like that studied here in which the repulsive core of the interaction between atoms and cavity wall extends as far out as a third of the radius of the cavity itself , then one can expect the 4 fluid inside the cavity to display a significant superfluid response .
+ clearly , the mathematical model contains important simplifications , notably the assumption that the cavity may be regarded as perfectly spherical and smooth .
the expected minimum distance from the surface ( over 3 ) at which atoms sit in the case of a li substrate seems to justify at least in part the neglect of corrugation and surface defects , an assumption routinely made in numerical studies of adsorption of he or parahydrogen on alkali substrates.@xcite in the case of a strongly attractive substrate such as glass , such an assumption is clearly far more questionable .
it is our hope that this work will provide at least some general aid in the design of experiments aimed at probing the superfluid response of nanoscale clusters of helium , or other quantum fluids ( e.g. , parahydrogen ) in confinement .
this work was supported in part by the natural science and engineering research council of canada .
computing support from westgrid is gratefully acknowledged .
it should be noted that fabrication techniques with the potential of extending significantly the diameter of the cages are being developed .
see , for instance , w. roth _
et al . _ , nature chemistry * 5 * , 628 ( 2013 ) .
m. boninsegni , n. prokofev and b. svistunov , phys .
lett . * 96 * , 070601 ( 2006 ) .
m. boninsegni , n. prokofev and b. svistunov , phys .
e * 74 * , 036701 ( 2006 ) .
f. mezzacapo and m. boninsegni , phys .
lett . * 97 * , 045301 ( 2006 ) .
f. mezzacapo and m. boninsegni , phys .
a * 75 * , 033201 ( 2007 ) .
the values @xmath15=100 k and @xmath14=2.05
have been utilized in some studies to characterize the interaction of a helium atom with a glass substrate .
see , for instance , l. pricaupenko and j. treiner , phys .
lett . * 74 * , 430 ( 1995 ) .
the calculation for the specific free 4 clusters mentioned in this work was carried out in this work separately .
because such a small cluster is very loosely bound , an artificial confining potential was utilized , of the form @xmath58 , with @xmath59=20 and @xmath59=25 , yielding consistent results .
gordillo and d. m. ceperley , phys . rev .
b * 58 * , 6447 ( 1998 ) .
k. yamamoto , h. nakashima , y. shibayama , and k. shirahama , phys .
lett . * 93 * , 075302 ( 2004 ) .
see , for instance , f. mezzacapo and m. boninsegni , phys .
lett . * 100 * , 145301 ( 2008 ) .
see , for instance , r. t. azuah , w. g. stirling , h. r. glyde , m. boninsegni , p. e. sokol and s. m. bennington , phys .
b * 56 * , 14620 ( 1997 ) . | structure and superfluid response of nanoscale size 4 clusters enclosed in spherical cavities are studied by computer simulations .
the curved surface causes the formation of well - defined concentric shells , thus imparting to the system a very different structure from that of free standing clusters . on a strongly attractive substrate ,
superfluidity is only observed at low density , in the single layer coating the inner surface of the cavity .
if the substrate is very weak ( e.g. , li ) , on the other hand , a superfluid two - shell structure can form , whose physical properties interpolate between two and three dimensions .
it is shown how experimental signatures of this physical behavior can be detected through measurements of the momentum distribution . |
99 t. hayakawa , n. iwamoto , t. shimizu , t. kajino , h. umeda , and k. nomoto , phys .
lett . * 93 * , 161102 ( 2004 ) .
a. heger , e. kolbe , w.c .
haxton , k. langanke , g. martinez - pinedo and s.e .
woosley , phys .
b * 606 * , 258 ( 2005 ) .
myung - ki cheoun , eunja ha , t. hayakawa , s. chiba and t. kajino , phys .
c * 82 * , 035504 ( 2010 ) .
myung - ki cheoun , eunja ha , t. hayakawa , satoshi chiba , ko nakamura , toshitaka kajino , grant j. mathews , arxiv:1108.4229 [ nucl - th ] , ( 2011 ) .
k. yako , _ et al_. , phys .
b * 615 * , 193 ( 2005 ) .
f. simkovic , l. pacearescu , and a. faessler , * a733 * , 321 ( 2004 ) .
m. s. yousef , v. rodin , a. faessler , and f. simkovic , phys .
c * 79 * , 014314 ( 2009 ) .
m. k. cheoun , a. bobyk , amand faessler , f. simcovic and g. teneva , * a561 * , 74 ( 1993 ) ; * a564 * , 329 ( 1993 ) ; m. k. cheoun , g. teneva and amand faessler , * a587 * , 301 ( 1995 ) .
myung - ki cheoun , eunja ha , k. s. kim , and toshitaka kajino , j. phys .
g * 37 * , 055101 ( 2010 ) .
eunja ha and myung - ki cheoun , arxiv:1205.4561 [ nucl - th ] , ( 2012 ) .
i. hamamoto , phys . rev .
* c 69 * , 041306(r ) ( 2004 ) ; phys .
c * 76 * , 054319 ( 2007 ) . t. otsuka _
et al_. , phys .
lett . * 95 * , 232502 ( 2005 ) .
t. otsuka _
et al_. , phys .
lett . * 104 * , 012501 ( 2010 ) .
s. g. nilsson and i. ragnarsson , _ shapes and shells in nuclear structure _ ( cambridge university press , cambridge , uk , 1995 ) r. madey _
et al_. , phys .
c * 40 * , 540 ( 1989 ) .
g. a. lalazissis , s. raman , p. ring , at .
data and nucl .
data tables * 71 * , 1 - 40 ( 1999 ) . | with the advent of high analysis technology in detecting the gamow - teller ( gt ) excited states beyond one nucleon emission threshold , the quenching of the gt strength to the ikeda sum rule seems to be recovered by the high - lying gt states .
moreover , in some nuclei , the stronger gt peaks than any other peaks appear explicitly in the high - lying excited states .
we address that these high - lying gt excited states stems from the smearing of the fermi surface by the increase of the chemical potential owing to the deformation within a framework of the deformed quasi - particle random phase approximation ( dqrpa ) .
detailed mechanism leading to the smearing is discussed , and comparisons to the available experimental data are shown to explain the strong peaks on the high - lying gt excited states in a satisfactory manner .
the deformation in nuclei becomes more important than last decades with the recent development of rare isotope ( ri ) accelerator facilities , from which one may perform lots of challenging experiments related to the ri nuclei .
most of them are thought to be deformed more or less .
they are usually produced in the successive nuclear capture reactions in the cosmos , _
i.e. _ slow- and rapid- process , presumed to be occurred , respectively , at the initial and explosive stage of stellar evolution , such as supernovae ( sne ) explosion .
although they decay fractions of a second , they imprint their existence on the nuclear abundances of stars @xcite . another interesting nuclear process is the neutrino ( @xmath0)-process in the sne explosion which is treated as one of important nuclear processes for the nucleosynthesis , in particular , of the neutron - deficient nuclei , such as @xmath1la and @xmath2ta @xcite . here
@xmath2ta is a well known deformed nucleus .
since the neutrinos emitted from a proto - neutron star may have tens of mev energy high enough to excite the deformed nuclei , one needs to understand more precisely the high - lying excited states beyond one nucleon threshold . among them ,
the gt excitations are of great importance because most of the charge exchange reactions are dominated by the gt transition .
these high - lying gt excited states are also closely related to the nuclear structure . for example
, the gt quenching problem says that the difference of total running sums @xmath3 for empirical gt ( @xmath4 ) transitions , @xmath5 , is usually quenched compared to the ikeda sum rule ( isr ) , @xmath6 .
however , recent experimental gt data on the high - lying states deduced by the more energetic projectiles shed a new light on the gt states located above one nucleon threshold , whose contributions are thought to enable us to explain the quenching problem with the high - lying gt excited states through the multi - particle and multi - hole configuration mixing @xcite .
of course , the contributions from the @xmath7 excitation and from the two - body current may be still effective for the quenching problem . conventional approach to understand the nuclear structure
is based on the spherical symmetry . in order to describe neutron - rich nuclei and
their relevant nuclear reactions occurred in the nuclear processes , one needs to develop theoretical formalism including explicitly the deformation @xcite .
ref .
@xcite exploited the nilsson basis to the deformed quasi - particle random phase approximation ( dqrpa ) .
but the two - body interaction was derived from the effective separable force .
the realistic two - body interaction is firstly considered at ref . @xcite only with neutron - neutron ( nn ) and proton - proton ( pp ) pairing correlations which have only isospin t = 1 and j = 0 interaction .
but , to properly describe the deformed nuclei , the t= 0 and j = 1 pairing should be also taken into account because it may easily lead to the deformation by the j = 1 pairing . in this work ,
we extend our previous qrpa based on the spherical symmetry @xcite , which has been exploited as a useful framework for describing the nuclear reactions sensitive on the nuclear structure of medium - heavy and heavy nuclei @xcite .
since our dqrpa formalism is fully discussed at ref .
@xcite , we briefly summarize the dqrpa used in this work . starting from the deformed wood saxon potential @xcite
, we transform a physical state given by the diagonalization of total hamiltonian in the nilsson basis into the spherical basis , in which one may perform more easily theoretical calculations . in a cylindrical coordinate ,
eigenfunctions of a single particle state and its time - reversed state denoted as @xmath8 and @xmath9 in deformed woods - saxon potential are expressed as follows @xmath10,\end{aligned}\ ] ] where @xmath11 ( @xmath12 ) is a major shell number , and @xmath13 and @xmath14 are numbers of nodes of the physical state on the deformed harmonic oscillator wave functions in @xmath15 and @xmath16 directions , respectively .
@xmath17 is the projection of the orbital ( total ) angular momentum onto the nuclear symmetric axis @xmath15 .
the coefficients @xmath18 are obtained by the eigenvalue equation of the total hamiltonian in the nilsson basis . the deformed harmonic oscillator wave function , @xmath19 in eq .
( [ eq : nil ] ) can be expanded in terms of the spherical harmonic oscillator wave function @xmath20 as follows @xmath21 and the spatial overlap integral @xmath22 numerically calculated in the spherical coordinate system .
therefore , the expansion can be simply written as @xmath23 with the expansion coefficient summarized as @xmath24 .
the deformed bcs equation is solved by using _
ab initio _ brueckner g - matrix calculated from the realistic bonn cd potential for the nucleon - nucleon interaction .
for example , the pairing potentials @xmath25 is calculated as @xmath26 where @xmath27 is designed for the transformation to the deformed basis of the g - matrix . here
@xmath28 , which is a projection number of the total angular momentum @xmath29 onto the @xmath15 axis , is selected @xmath30 at the bcs stage because we consider the pairings of the quasi - particles at @xmath8 and @xmath31 states . in order to renormalize the g - matrix , strength parameters , @xmath32 , @xmath33 and @xmath34
are multiplied to the g - matrix to reproduce empirical pairing gaps @xcite .
the @xmath35 decay operator , @xmath36 , is defined in the intrinsic frame as @xmath37 the @xmath38 transition amplitudes from the ground state of an initial nucleus to the excited state are expressed by @xmath39 ~,\end{aligned}\ ] ] where @xmath40 denotes the correlated qrpa ground state in the intrinsic frame with the nomalization factor given as @xmath41 the wigner functions disappeared by using the orthogonality of two wigner functions from the operator and the excited state , respectively . to compare to the experimental data , the gt(@xmath4 ) strength functions @xmath42 and their running sums @xmath43 for @xmath44ge and @xmath45se
are enumerated as a ratio of the @xmath46 to the isr value @xmath47 calculated with the closure relation ( isr i ) and without it ( isr ii ) .
.deformation parameters @xmath48 , pairing strength parameters @xmath49 , and theoretical pairing gaps @xmath50 used in this work .
the isr i and ii in the last column are given as @xmath51 .
the particle - particle ( particle - hole ) strength parameters are taken as @xmath52 for two nuclei .
empirical pairing gaps , @xmath53 ( @xmath54)=1.561 ( 1.535 ) and 1.409 ( 1.544 ) mev , are exploited for @xmath55 and @xmath56 , respectively .
[ cols="^,^,^,^,^,^",options="header " , ] [ tab : result1 ] the single particle states are used up to @xmath57 for two nuclei , in the spherical limit .
since the gt strength distribution turns out to rely on the deformation parameter , @xmath48 @xcite , we exploited the deformation parameter , @xmath58 , as the default values .
corresponding values of the deformation parameter @xmath48 , the renormalized strengths @xmath33 and @xmath32 , and theoretical and empirical pairing gaps are tabulated in table i. the isr i and ii , which are deviated 8% maximally from the isr , seems to be reasonable more or less , if we remind that the particle model spaces spanned by the deformed basis are finite less than 4 @xmath59 and two - body currents including @xmath7-excitation are not taken into account in this work .
an interesting and important point to be noticed is that the @xmath60 values are abnormally deviated from 1 for some prolate @xmath48 values .
this turns out to results from the wide smearing of the fermi surface in deformed nuclei , which makes the @xmath61 coefficients for the paring gaps in eq.(3 ) small around the fermi surface and enlarges the @xmath62 to fit the empirical pairing gap .
this tendency is explicitly revealed in the occupation probabilities of @xmath63se in fig.1 , which shows a wide smearing , _ i.e. _ significant change of the occupation probabilities of the particles around the fermi surface by the increase of the deformation parameter @xmath48 . since
the increase of the prolate deformation usually raises the fermi surface energy , some states above the fermi surface burrowed below or around the fermi surface .
for example , ( 4 3 1 1/2 ) , ( 4 2 2 3/2 ) and ( 4 1 1 3/2 ) states in @xmath56 are such cases .
this gives rise to the wide smearing .
such wide smearing of the fermi surface is also found for the protons and neutrons in @xmath44ge , as shown in fig.2 .
se as a function of the single particle energy given by nilsson basis for two deformation parameters @xmath48 . with the increase of @xmath48 from 0.1 to 0.2 ,
the fermi energy @xmath64 is increased with the wider smearing of the fermi surface . ]
ge . for comparison ,
results for protons are shown in the left panel . ]
detailed changes of single particle states by the @xmath48 deformation leading to the smearing are presented in figs.3 and 4 .
some states above the fermi surface are reallocated below or around the surface , because the single particle state energies adopted from the deformed woods saxon potential depend on the parameter @xmath48 .
the deformation of nuclei may be conjectured to come from macroscopic phenomena , for example , the core polarization , the high spin states and so on .
microscopic reasons may be traced to the tensor force in the nucleon - nucleon interaction , which is known to account for the shell evolution according to the recent shell model calculations @xcite .
for example , t = 0 , j = 1 pairing , which is associated with the @xmath65 tensor force , may lead to the deformation compared to the t = 1 , j = 0 pairing .
therefore , the deformation parameter adopted in this work may include implicitly and effectively such effects , because the single particle states from the deformed wood saxon potential show strong dependence on the @xmath48 , as shown in figs .
3 and 4 @xcite .
se as a function of the different deformation parameter @xmath66 .
red dotted lines denote the fermi energy . ]
ge ] we have taken @xmath67 for the particle - hole ( particle - particle ) strength parameters for the two nuclei . since the nuclei considered here are expected to have large energy gaps between proton and neutron spaces , in this work , we considered only the @xmath68 and @xmath69 pairing correlation although the formalism is presented generally .
for example , in the neutron - rich nuclei of importance in the r - process , the _ np _ pairing may not contribute so much .
but for the p - process , the _ np _ pairing could be more important than the neutron - rich nuclei because of the adjacent energy gaps of protons and neutrons .
the calculations for the neutron - deficient nuclei in p - process are in progress . in the following ,
we show the gt strength distributions in terms of the @xmath48 parameter , as a function of excited energy of parent nuclei .
therefore , experimental data are presented by adding the empirical q values from the measured data . in fig 5 , the gt strength distribution , b(gt@xmath70 ) , on @xmath45se is presented as a function of the excitation energy @xmath71 w.r.t .
the ground state of @xmath45se for @xmath72 .
uppermost panel represents the experimental data deduced from the @xmath45se(p , n)@xmath45br reaction at 134.4mev @xcite . with the redistribution of the gt strengths ,
the gt data around @xmath73mev are well reproduced at @xmath48 = 0.2 ( see ( c ) panel in fig.5 ) .
for a reference , the @xmath48 value from the relativistic mean field ( rmf ) is 0.133 @xcite . ) on @xmath63se as a function of the excitation energy @xmath71 w.r.t .
the ground state of @xmath45se .
experimental data denoted as filled ( red ) points in uppermost panel are deduced from the @xmath45se(p , n ) reaction at 134.4mev @xcite . in each panel , we indicate @xmath48 . ]
ge as a function of the excitation energy @xmath71 w.r.t .
the ground state of @xmath44ge .
experimental data denoted as filled ( red ) points in uppermost panel are deduced from the @xmath44ge(p , n ) reaction at 134.4mev @xcite . ]
fig.6 shows b(gt@xmath70 ) of @xmath44ge as a function of the excitation energy @xmath71 with prolate shapes , @xmath74 .
uppermost panel represents the experimental data from the @xmath44ge(p , n)@xmath44as reaction at 134.4mev @xcite , which show a strong gt state peak around 12 mev .
the gt strength distributions are widely scattered due to deformation similarly to fig.5 . in particular , results for the @xmath48 = 0.35 , in which the isr i and ii are almost satisfied , nicely reproduce the peak on the high - lying excited states .
the value calculated by rmf is 0.157 @xcite .
both nuclei , @xmath44ge and @xmath45se , show a strong peak on high - lying gt states around 12 mev .
our @xmath48 values , @xmath48=0.35 and 0.2 for @xmath44ge and @xmath45se , customized to the isr values , are shown to reproduce neatly the gt states .
actually , these high - lying gt states turn out to be intimately associated with the smearing of some physical states around the fermi surface in deformed nuclei . for two dominant high - lying gt states ( @xmath71 = 10.74 and 11.51 mev ) w.r.t . the ground state of @xmath45se as a function of the configuration state for the panel ( c ) in fig.5 . ]
ge ( @xmath71 = 11.86 mev ) corresponding to the panel ( d ) in fig.6 ] figs.7 and 8 show the forward amplitudes @xmath75 of two excitation energies and one excitation energy , having the large b(gt ) values in ( c ) and ( d ) of figs.5 and 6 , respectively .
for @xmath45se , the main high - lying gt transition come from ( 4 3 1 3/2)@xmath76 ( 4 4 0 1/2 ) , ( 3 0 1 3/2)@xmath76 ( 3 0 1 1/2 ) and ( 4 1 1 3/2)@xmath76 ( 4 1 1 1/2 ) state transitions , while ( 3 0 3 7/2)@xmath76(3 0 3 5/2 ) , ( 3 0 1 3/2)@xmath76 ( 3 0 1 1/2 ) , ( 4 1 3 7/2)@xmath76 ( 4 1 3 5/2 ) and ( 3 1 2 3/2)@xmath76 ( 3 1 0 1/2 ) turn out to be main transitions for @xmath44ge . in summary ,
to describe the single particle state in deformed basis , we used the deformed axially symmetric woods - saxon potential .
we performed the deformed bcs and deformed qrpa with a realistic two - body interaction calculated by _ ab initio _ g - matrix based on bonn potential .
results of the gamow - teller strength , b(gt ) , for @xmath44ge and @xmath45se show that the deformation effect leads to a fragmentation of the gt strength and reproduces the high - lying gt excitation deduced by higher energy projectiles , which result from the wide smearing by the increase of the fermi surface energy due to the deformation .
this work was supported by the national research foundation of korea ( c00020 , 2012r1a1a3009733 , 2011 - 0015467 ) . |
the theory of polarons , the quasiparticles describing electrons in a polarizable medium , dates back to the early days of quantum theory @xcite , and it has been an active field of research ever since @xcite . in this paper
we will develop a microscopic theory of intersubband polarons , that is , a theory of intersubband transitions coupled to longitudinal optical ( lo ) phonons in semiconductor quantum wells .
the coupling between intersubband transitions and lo - phonons is relevant for a number of optoelectronic applications , as it determines the lifetime of carriers in excited subbands @xcite . in particular a precise knowledge of lo - phonons intersubband scattering rates
is important in the engineering of heterostructures for quantum cascade lasers @xcite .
normally optoelectronic devices are designed to avoid being in resonance with optical phonon transitions , due to the high absorption between transverse and longitudinal optical phonon frequencies ( restrahlen band ) .
a notable exception is provided by quantum cascade lasers operating near such optical resonances@xcite , in which instead the transitions between different subbands are almost resonant with lo - phonon modes .
even if the coupling between intersubband transitions and lo - phonons in semiconductor quantum wells has indeed received some attention @xcite and intersubband polaron resonances have been clearly and unambiguously observed @xcite , to the best of our knowledge , there is no microscopic theory of such excitations , as the spectra of intersubband polarons are normally calculated with indirect methods . while such methods allow to calculate , at least qualitatively , the polaron dispersions , missing a microscopic description makes it difficult to study more complex phenomena as nonequilibrium physics , quantum vacuum effects or quantum phase transitions . using a second quantization formalism , we will reduce the full electron - phonon hamiltonian to a quadratic , bosonic form , from which we will then calculate the polaron dispersions . in order to accomplish this task ,
we show that coulomb interaction between electrons in conduction subbands naturally separates into a dominant and a perturbative part , accordingly to the number of electrons that can participate to each transition .
moreover we will show how , thanks to the tight confinement of lo - phonons , the coupling between intersubband transitions and phonons can easily be in the ultrastrong coupling regime , a regime characterized by a coupling strength comparable to the bare frequency of the excitations @xcite
. such fact can have interesting observable consequences , as a whole new range of physics is _ a priori _ observable in this regime : spectral deformations @xcite , quantum vacuum emission phenomena @xcite , electroluminescence enhancement @xcite and even quantum phase transitions @xcite .
this article is organized as follows : in sec .
[ theory ] we will develop the general theory of the coupling between intersubband transitions and lo - phonons , that we will then apply in sec . [ numres ] to the case of gaas quantum wells , showing how the ultrastrong coupling regime can be reached even with such relatively weakly polar material . in sec .
[ appendix ] we will compare the calculated dispersions with the ones obtained using an homogeneous dielectric function approach . finally a few considerations on the impact of our results and on possible future developments will be drawn in sec .
in 1954 dicke @xcite noticed that a set of coherently excited identical dipoles relaxes radiatively much faster than a single , isolated one .
this is due to the phenomenon of superradiance : @xmath0 identical dipoles behave as a single collective dipole @xmath1 times bigger .
the concept of superradiance has been thoroughly applied to the study of intersubband polaritons @xcite in microcavity embedded quantum wells . in such systems the light couples to a collective electronic excitation and , as expected @xcite , the strength of the coupling between light and matter is proportional to the square root of the number of electrons involved .
we will study the coupling of intersubband transitions with longitudinal optical phonons , considering also the role of coulomb electron - electron interaction .
such couplings are extremely rich and , in order to limit the complexity of our investigation , we will need to determine which scattering channels are dominant and which are negligible .
in general , if @xmath0 electrons undergo a certain transition in a coherent way , the strength of the coupling is enhanced by a factor @xmath1 .
transitions involving a macroscopic number of electrons will thus be strongly enhanced and , for this reason , they will be treated exactly within an hamiltonian formalism , while the others ( involving only few electrons ) will be treated perturbatively ( or ignored altogether ) .
the degree of collective enhancement of a scattering process will be evaluated looking at the number of electrons that can undergo the transition given fixed amounts of transferred impulsion and energy . in fig .
[ cohex ] we show a few illustrative examples of collective and non - collective transitions . in the case of the coulomb interaction , as we will see , only the intersubband terms , responsible for the depolarization shift , are collective .
we will thus treat them exactly in the hamiltonian , while considering all the other terms in a rpa linear response approach .
a ) an example of collective transition : an intersubband transition with small transferred momentum .
all the electrons can undergo a transition resonant approximately at the same energy .
b ) an example of non - collective transition : an intrasubband transition with transferred momentum much smaller than the fermi wavevector @xmath2 .
the majority of electrons , being pauli blocked , do not participate to the process .
c ) another example of non - collective transition : an intrasubband transition with transferred momentum of the order of @xmath3 , that is the minimum to allow all the electrons to undergo the transition avoiding pauli blocking .
anyway there is a large energy spread between the different single electron transitions . for an infinite potential
well , electrons on the two opposite borders of the fermi sea , with initial momenta @xmath4 parallel to the transferred momentum , have initially the same energy @xmath5 , where @xmath6 is the electron effective mass and @xmath7 the fermi energy . after the transition they will end up with final momenta @xmath2 and @xmath8 , corresponding to final energies @xmath5 and @xmath9 respectively .
this implies that , even if the transition is not blocked , only a small fraction of the single electron transitions can be resonant at the same time . in both cases , given that only few electrons can participate to the collective transition , the superradiant enhancement factor will be small.,width=321 ] we will consider a symmetric quantum well of length @xmath10 in a bulk of height @xmath11 , @xmath12 will be the surface of the sample . for the moment , we will limit ourselves to the case of a single quantum well , the general case of multiple wells will be addressed later in this section . the quantum well
is supposed to be doped in such a way that its fermi level is between the first and the second conduction subbands , separated between them by the intersubband gap energy @xmath13 .
we will develop our theory using a zero temperature formalism ( @xmath14 ) , anyway our results will remain quantitatively accurate while the thermal population of the second subband remains negligeable . depending on material parameters and doping level , this could imply the necessity to perform experiments in different kinds of cryogenic environments .
electron states will be indexed by the subband index @xmath15 and by the value of the in - plane wavevector @xmath16 .
their wavefunctions will be given by @xmath17 where , for simplicity , we will choose @xmath18 to be real and , due to the symmetry of the quantum well , the @xmath18 have well defined and opposite symmetry .
wavefunctions in eq .
( [ electronicwf ] ) are chosen as basis for second quantization , the creation operator for an electron in the state described by eq .
( [ electronicwf ] ) will be denoted as @xmath19 .
the free hamiltonian of the electron gas in the two considered subbands thus reads @xmath20 where @xmath21 and @xmath22 . in eq .
( [ hfreec ] ) , as well as in the rest of this paper , we will omit the electron spin index .
this is justified by the fact that all interactions we consider are spin conserving .
given that we will consider only in - plane wavevector exchanges @xmath23 much smaller than the typical electron wavevector @xmath16 , we can make the approximation @xmath24 , and introduce the operators describing intersubband transitions with a well defined and dispersionless energy @xmath13
@xcite @xmath25 where @xmath0 is the number of electrons in the quantum well . in the dilute regime , that is
if the number of excitations is much smaller than @xmath0 , the @xmath26 operators are bosonic @xmath27 at higher excitation densities , that are out of the scope of the present work , saturation effects start to appear and corrections to eq .
( [ bos ] ) have to be taken into account . for a detailed analysis of nonbosonicity effects in intersubband transitions
we invite the interested readers to refer to ref .
[ ] . using eq .
( [ bdef ] ) we can rewrite the hamiltonian of the free electron gas in eq .
( [ hfreec ] ) in terms of bosonic intersubband excitations @xmath28 in this work we are interested in the resonant case in which @xmath29 is equal , or close , to the lo - phonon frequency @xmath30 .
we can thus neglect confinement effects on the phonons and consider bulk values for their frequencies @xcite .
we will thus describe lo - phonons by means of the three dimensional boson operators @xmath31 @xmath32 indexed by their in - plane and out - of - plane wavevectors . while we know that lo - phonon modes are confined inside the quantum well , we do not need to impose this constraint in the mode definition because , as we will see , intersubband transitions end up coupling with linear superpositions of phonon modes that are anyway confined inside the quantum well .
we will consider only the case of one single longitudinal optical branch , the expansion to the case of multiple branches not presenting any fundamental difficulty .
moreover , we are interested only in phonons with small in - plane wavevectors ( in order to couple with coherent intersubband excitations ) , we can thus ignore phonon dispersion and write the free phonon hamiltonian as @xmath33 interaction between electrons and lo - phonons can be described using the frhlich hamiltonian @xcite @xmath34 where @xmath35 where @xmath36 and @xmath37 are respectively the static and high frequency dielectric constants @xcite .
the hamiltonian in eq .
( [ hffq ] ) can be written in second quantization ( neglecting incoherent intrasubband scattering @xcite ) as @xmath38 where we have defined @xmath39 from eq .
( [ hfsq ] ) we see that , due to the three dimensional character of the lo - phonons @xcite , each electronic transition couples to multiple phonon modes , indexed by different values of the wavevector along the growth direction .
it is thus convenient to introduce second quantized operators corresponding to the particular linear superpositions of phonon modes that are coupled to electronic transitions @xmath40 whose spatial wavefunctions along the @xmath41 axis are @xmath42 from eqs .
( [ phi ] ) and ( [ fdef ] ) we see that the intersubband transitions naturally couple to phonon modes localized inside the quantum well ( it is easy to verify that @xmath43 vanishes to the first order in @xmath44 if @xmath41 is outside the common support of @xmath45 and @xmath46 ) .
the normalization factor @xmath47 can be fixed imposing bosonic commutation relations for the @xmath48 operators @xmath49 and thus @xmath50 where we have defined @xmath51 we can thus write hamiltonians in eq .
( [ hfreed ] ) and eq .
( [ hfsq ] ) in terms of coherent @xmath48 and @xmath26 operators as @xmath52 and @xmath53 where @xmath54 is the density of the two dimensional electron gas .
in order to pass from eq .
( [ hfreed ] ) to eq .
( [ hfint ] ) , we are neglecting higher order phonon modes confined inside the quantum well .
this is justified by the fact that we limit ourselves to long - wavelength modes . in order to treat the coulomb electron - electron interaction we start by the second quantized form of the hamiltonian describing the coulomb interaction @xcite ( see fig .
[ cimage ] ( a ) for a graphical representation of the interaction coefficients ) a ) index convention of the matrix element @xmath55 .
two electrons in subbands @xmath15 and @xmath56 , with momenta @xmath16 and @xmath57 are scattered into subbands @xmath58 and @xmath59 , with momenta @xmath60 and @xmath61 respectively .
b ) graphical representation of the four qualitatively different kinds of scattering processes from eq .
( [ elem]).,width=321 ] @xmath62 where @xmath63 is the two dimensional coulomb matrix element .
it is important to notice that in eq .
( [ v ] ) we used the high frequency dielectric constant @xmath37 instead of the static one .
this is due to the fact that @xmath36 includes the effect of the coupling to lo - phonons , that are already treated exactly in the hamiltonian . due to the symmetry of the wavefunctions , a certain number of matrix elements in eq .
( [ v ] ) can be seen to be zero , in particular all the matrix elements with an odd number of @xmath64 and @xmath65 indices @xmath66 the other elements can be evaluated as @xmath67 where @xmath68 , defined in eq .
( [ iq ] ) , is the same integral we encountered studying the electron - phonon frhlich interaction .
the four distinct nonzero possible values of the matrix elements correspond to different kinds of scattering processes . in fig .
[ cimage ] ( b ) a graphical representation for each of these processes is shown .
it is important at this point to notice a major difference between the elements in the first line of eq .
( [ elem ] ) and the others .
the elements in the first line ( upper left subpanel in fig .
[ cimage ] ( b ) ) represent intersubband excitations : each electron is scattered from one subband to the other .
such processes , responsible for the depolarization shift @xcite , describe a superradiant process in the sense defined above , that is , at least for small values of @xmath23 , a great number of electrons can coherently undergo the same transition , approximately at the same energy .
this is not the case for the interactions described in the other lines of eq .
( [ elem ] ) , that instead describe intrasubband excitations that , either due to pauli blocking or to the non - flat energy dispersions , involve only few electrons ( see fig .
[ cohex ] for a graphical visualization of this crucial point ) .
our previous discussion on superradiant processes thus implies that the terms in the first line of eq .
( [ elem ] ) strongly dominate over the others due to their superradiant enhancement .
for this reason we have to treat them exactly in an hamiltonian formalism , while we can limit ourselves to treat the others within a perturbative approach . here
we will thus construct an exact , hamiltonian approach , to treat the effect of the depolarization shift terms , neglecting the others .
we will analyze later the effect of the intrasubband terms .
let us start to rewrite the depolarization shift part of eq .
( [ hc1 ] ) in a more useful form @xmath69 we see from eq .
( [ hcnew ] ) that , thanks to its collective , superradiant nature , the depolarization shift can be naturally written in terms of the bosonic intersubband excitations defined in eq .
( [ bdef ] ) as @xmath70 where we have neglected the last constant term , that simply shifts the ground state energy .
we treated exactly the depolarization shift terms of the coulomb interaction in a bosonic excitation formalism .
moreover , we showed how the terms other that the ones responsible for the depolarization shift are strongly suppressed , due to their lack of collective enhancement and can thus be treated perturbatively . here
we will study the perturbative effect of such residual coulomb contributions , due to the intrasubband terms in the last three lines of eq .
( [ elem ] ) ( schematized in the last three panels of fig .
[ cimage ] ( b ) ) . an important result due to lee and galbraith @xcite , is that such intrasubband terms do not contribute to the screening of the intersubband ones at the level of the random phase approximation ( rpa ) .
this can be seen writing the dyson equation for the dynamically screened coulomb potential @xcite @xmath71 @xmath72 where @xmath73 is the rpa polarization function . in the case of an intersubband contribution ( e.g. , @xmath74 ) ,
( [ elem0 ] ) and ( [ elem ] ) imply that @xmath75 we have thus @xmath76 from which we see that the intrasubband coulomb terms ( @xmath77,@xmath78 , @xmath79 and @xmath80 ) do not intervene in the renormalization of the intersubband terms .
an analogous reasoning can be done for the phonon - electron interaction .
calling @xmath81 and @xmath82 the bare and screened version of the potential defined in eq .
( [ hfsq ] ) , we have the dyson equation @xmath83 and thus the formula for the screened potential is @xmath84 being the rpa screening only due to terms that are exactly treated in the hamiltonian , we can thus neglect the screening due to the two dimensional electron gas .
putting together eqs .
( [ hfreeb ] ) , ( [ hint ] ) and ( [ hc2 ] ) we arrive to the full hamiltonian for the intersubband transitions - lo - phonons system @xmath85 the hamiltonian in eq . ( [ hfull0 ] ) can be rewritten in a more compact form by introducing the intersubband transitions - lo - phonons coupling coefficient @xmath86 and the coulomb coefficient @xmath87 @xmath88 where we have dropped the dependences over the wavevector as we are interested in the long wavelength limit ( from eq .
( [ iq ] ) we can verify that @xmath89 tends to a constant value ) .
( [ omega ] ) , eq .
( [ hfull0 ] ) can be written as @xmath90 that can be cast in matrix form as @xmath91 where the column vector of operators @xmath92 is defined as @xmath93 @xmath94 is the diagonal metric @xmath95 , \end{aligned}\ ] ] and the hopfield - bogoliubov@xcite matrix @xmath96 is defined as @xmath97 diagonalizing the matrix in eq .
( [ hbm ] ) will yield the frequencies of the normal modes of the system @xmath98 , that are usually called polarons @xcite . in our case
we will name them more properly intersubband polarons , because the electronic part of the mixed excitations is an intersubband transition .
hamiltonian in eq .
( [ hfull ] ) can thus be put in the diagonal form @xmath99 where the @xmath100 are the annihilation operators for the two polaronic branches , given by a linear superposition of @xmath101 , @xmath102 , @xmath103 and @xmath104 operators and @xmath105 is the energy of the new ground state relative to the one of the uncoupled system .
the coupling between the intersubband transitions and the lo - phonons does not modify only the system s resonances but it also qualitatively modifies the nature of its ground state .
it is easy to verify that , if @xmath106 is the ground state for the uncoupled phonons and intersubband excitations , defined in the usual way as @xmath107 then @xmath108 that is , @xmath106 is not the ground state for the coupled system .
the real ground state of the hopfield matrix in eq .
( [ hbm ] ) , that has been thoroughly studied in ref .
[ ] , has the form of a two modes squeezed vacuum .
still , thanks to the bosonicity of the system , such new ground state does not influence the response of the system , that can be described as a gas of free bosonic excitations ( from eq .
( [ hd ] ) ) .
a notable exception is the case in which the parameters of the system are nonadiabatically modulated in time . in this case
the sudden change in the ground state @xcite can have observable effects , like the emission of quantum vacuum radiation @xcite .
it is also interesting to notice that , from eq .
( [ omega ] ) , we can write the coulomb coefficient @xmath87 as @xmath109 as it has recently been shown in ref . , eq . ( [ do ] ) implies that the ground state of the system will not undergo a dicke phase transition , regardless of the strength of the coupling .
until now we considered the case of a single quantum well .
this choice has been motivated by the fact that , as we will show , the presence of multiple wells does not modify our results .
given that we are considering rather large quantum wells ( in order for the transition to be resonant with the lo - phonon mode ) , the optical phonon spectrum is not modified @xcite and the optical phonon modes we consider are confined in each quantum well . this is a rather important difference between the intersubband polaron case we consider in this paper and the physics of intersubband polaritons . for intersubband polaritons ,
the electromagnetic mode coupled to the intersubband transitions extends over all the structure .
it thus couples to all the electrons ,
regardless of the quantum well they are in .
this means that the only meaningful parameter for intersubband polaritons is the total density of electrons , and the light - matter coupling thus scales as @xmath110 , where @xmath111 is the number of quantum wells inside the microcavity . in the present case instead , being the phonon modes confined inside each quantum well , electrons in different wells are completely decoupled .
this can also be inferred from the coupling integral in eq .
( [ v ] ) .
this integral does vanish , at least in the long wavelength limit ( first order in @xmath44 ) , if the wavefunctions for the two integration variables @xmath41 and @xmath112 do not have a common support , i.e. , if the two interacting electrons are in different quantum wells .
this means that , contrary to the intersubband polariton case , the intersubband polaron interaction scales only as @xmath113 and growing multiple quantum wells in the same sample will not increase the coupling .
in order to obtain some numerical predictions from hamiltonian in eq .
( [ hfull ] ) , we need to fix a few parameters concerning the material and the quantum well . for sake of simplicity we will consider the quantum well to be correctly approximated by a rectangular , infinite potential well of length @xmath10 .
we thus have @xmath114 and the electronic and phononic modes profiles are given by @xmath115 inside the well and zero outside .
as explained in sec . [ theory ] , we see here explicitly that the intersubband transitions couple to a linear superposition of phonon modes that is localized inside the quantum well ( the cubic sinus in the third line of eq .
( [ inf ] ) comes from the integral of the first two , as can be verified performing the integral in eq .
( [ phi ] ) ) . inserting eq .
( [ inf ] ) into eq .
( [ iq ] ) and performing the integral we have @xmath116 in fig .
[ coupling ] we plot the normalized coupling @xmath117 as a function of the density of the two dimensional electron gas , for a gaas quantum well . in the inset of fig .
[ coupling ] we instead present a comparison of the values of @xmath117 , at room temperature , for different semiconductors of the iii - v and ii - vi groups@xcite , as a function of the respective lo - phonon energies , for a reference doping @xmath118@xmath119 . in fig .
[ dispersion ] there is a plot of the intersubband polaron frequencies @xmath98 as a function of the intersubband frequency @xmath29 , in gaas , for @xmath118@xmath119 .
notice that , due to the effect of coulomb interaction , the resonant anticrossing is not at @xmath120 but at a lower frequency . in the inset of the same figure we plot the same quantity as a function of the electron density .
the length @xmath10 has been chosen in this case to have the two uncoupled modes at resonance ( @xmath120 , that is @xmath121 nm ) .
it is clear from the figures that intersubband polarons are not only strongly coupled , having coupling constants much larger than their linewidth ( usual linewidths being not bigger than a few mev ) , but they are indeed in the ultrastrong coupling regime , with values of the normalized coupling @xmath117 comparable or larger than the best ones reported in the literature . for physically realizable levels of doping ,
coupling values of a few tenths of the bare frequency of the excitation @xmath30 are predicted in gaas , and it seems that values much larger can be obtained using more polar materials .
the reason of such large coupling can be found in the superradiant nature of intersubband excitations and in the natural confinement of the phonons inside the quantum well , that gives an extremely small mode volume , when compared with what can be obtained with photonic microcavities .
the consequences of our results can be multiple , both for fundamental and applied research .
on the fundamental side , intersubband polarons could become a new laboratory to test quantum vacuum physics , typical of the ultrastrong coupling regime @xcite . on the applied side
our theory can be naturally exploited in the study of quantum cascade lasers working in or near the restrahlen band .
it can , for example , help explaining the anticrossing observed in ref .
, near the lo - phonon frequency .
moreover the capability to strongly modify the lo - phonon spectrum could have an impact on the performances of optoelectronic devices , as the electron - lo - phonon scattering rate determines the lifetime of carriers in excited subbands @xcite .
[ bc][b][1.2]@xmath122@xmath123 [ bc][b][1]@xmath124 [ b][b][1.2]@xmath125 [ b][b][1]@xmath125 normalized coupling @xmath117 in gaas as a function of the doping density @xmath126 .
inset : the same quantity as a function of the lo - phonon frequency @xmath30 for different materials , for @xmath118@xmath119.,width=340 ] [ bc][b][1.2]@xmath127 [ bc][b][1]@xmath128@xmath129 [ b][b][1.2]@xmath130 [ b][b][1]@xmath130 intersubband polaron frequencies @xmath98 as a function of the intersubband frequency @xmath29 in gaas for @xmath131
. inset : the same quantity as a function of the doping for @xmath120.,width=340 ]
in the previous sections we have developed a detailed microscopic theory for the intersubband transitions coupled to lo - phonons . here
we will compare the dispersions obtained from the microscopic theory with the ones obtained with an homogeneous dielectric function theory , as the one used in ref . .
the propagation of an electromagnetic wave in a dispersive , homogeneous medium obeys the equation @xmath132 this implies that it is possible to have propagating longitudinal waves , like polarons , only at frequencies for which @xmath133 where @xmath134 indicates the real part .
the @xmath41 component of the dielectric function of an homogeneous medium filled with quantum wells is given by @xcite @xmath135 where @xmath136 is the plasma frequency of the two dimensional electron gas and @xmath137 is the intersubband dipole @xmath138 the equation @xmath139 thus reads @xmath140 as eq .
( [ aeps ] ) neglects both the dielectric response in the @xmath141 plane and the non - homogeneity in the @xmath41 direction , in order to recover the same result from our microscopic approach , we will have to consider only phonon modes with @xmath142 and @xmath143 . from eqs .
( [ fdef ] ) and ( [ d12 ] ) we thus have @xmath144 following exactly the same procedure of sec . [ theory ] , but with the @xmath145 defined in eq .
( [ afdef ] ) and considering only the @xmath143 mode , we get @xmath146 and thus , from eq .
( [ do ] ) @xmath147 in order to obtain the polaronic eigenfrequencies we have to diagonalize the matrix in eq .
( [ hbm ] ) using the coupling coefficients for the homogeneous limit defined in eqs .
( [ aomega ] ) and ( [ ad ] ) .
we thus obtain the secular equation @xmath148 that , using eqs .
( [ ad ] ) and ( [ omegap ] ) can be put into the form @xmath149 equating the coefficients of eqs .
( [ anumeps ] ) and ( [ anumdiag2 ] ) , we obtain @xmath150 that is the well known lyddane - sachs - teller relation @xcite .
we have thus proved that the homogeneous version of our theory gives the same results as the homogeneous dielectric function approach .
it is anyway important to notice that the homogeneous limit in not exact , as a quantum well is , by definition , spatially inhomogeneous .
ignoring the higher @xmath151 modes leads to underestimate the intersubband dipole of a factor roughly equal to @xmath152 .
in this paper we have developed a microscopic theory of intersubband polarons , mixed excitations resulting from the coupling between intersubband transitions in doped quantum wells and lo - phonons .
we took into account the electron - electron coulomb interaction and we were able to treat exactly the resulting depolarization shift .
we proved that intersubband polarons can be in the ultrastrong coupling regime , reaching extremely high values of the coupling constant .
we critically discussed the relevance of our results both for fundamental and applied research .
we would like to thanks d. hagenmller , m. zaluzny , p. nataf , l. nguyen , j. restrepo , c. sirtori and y. todorov for useful discussions and comments .
c. c. is member of institut universitaire de france .
we acknowledge support from the anr grant qpol j. faist , f. capasso , d. l. sivco , c. sirtori , a. l. hutchinson and a. y. cho , science * 264 * , 5158 ( 1994 ) . f. castellano , a. bismuto , m. i. amanti , r. terazzi , m. beck , s. blaser , a. bchle and j. faist , j. app . phys . * 109 * , 102407 ( 2011 ) .
r. colombelli , f. capasso , c. gmachl , a. l. hutchinson , d. l. sivco , a. tredicucci , m. c. wanke , a. m. sergent and a. y. cho , appl .
lett . , * 78 * , 2620 ( 2001 ) .
s. butscher , j. forstner , i. waldmuller and a. knorr , phys .
b * 241 * , 11 ( 2004 ) .
s. butscher and a. knorr , phys .
lett . * 97 * , 197401 ( 2006 ) .
j. c. cao , y. l. chen and h. c. liu , superlattices and microstructures * 40 * , 119 ( 2006 ) .
h. c. liu , c. y. song , z. r. wasilewski , a. j. springthorpe , j. c. cao , c. dharma - wardana , g. c. aers , j. lockwood and j. a. gupta , phys .
* 90 * , 077402 ( 2003 ) . c. ciuti , g. bastard and i. carusotto , phys .
b * 72 * , 115303 ( 2005 ) .
a. a. anappara , s. de liberato , a. tredicucci , c. ciuti , g. biasiol , l. sorba and f. beltram , phys .
b * 79 * , 201303 ( 2009 ) .
y. todorov , a. m. andrews , r. colombelli , s. de liberato , c. ciuti , p. klang , g. strasser and c. sirtori , phys .
lett . * 105 * , 196402 ( 2010 ) .
d. hagenmller , s. de liberato and c. ciuti , phys .
b * 81 * , 235303 ( 2010 ) . s. de liberato , c. ciuti and i. carusotto , phys .
lett . * 98 * , 103602 ( 2007 ) . s. de liberato , d. gerace , i. carusotto and c. ciuti , phys .
a * 80 * , 053810 ( 2009 ) .
m. kardar and r. golestanian , rev .
phys . * 71 * , 1233 ( 1999 ) . s. de liberato and c. ciuti , phys .
b * 77 * , 155321 ( 2008 ) . s. de liberato and c. ciuti , phys .
b * 79 * , 075317 ( 2009 ) . c. emary and t. brandes , phys . rev .
lett . * 90 * , 044101 ( 2003 ) .
n. lambert , c. emary and t. brandes , phys .
lett . * 92 * , 073602 ( 2004 ) .
p. nataf and c. ciuti , phys .
lett . * 104 * , 023601 ( 2010 ) .
r. h. dicke , phys . rev . * 93 * , 99 ( 1954 ) .
d. dini , r. kohler , a. tredicucci , g. biasiol and l. sorba , phys .
lett . * 90 * , 116401 ( 2003 ) .
l. sapienza , a. vasanelli , c. ciuti , c. manquest , c. sirtori , r. colombelli and u. gennser , app .
. lett . * 90 * , 201101 ( 2007 ) .
l. sapienza , a. vasanelli , r. colombelli , c. ciuti , y. chassagneux , c. manquest , u. gennser and c. sirtori , phys .
* 100 * , 136806 ( 2008 ) .
a. anappara , a. tredicucci , f. beltram , g. biasiol , l. sorba , s. de liberato and c. ciuti , app .
. lett . * 91 * , 231118 ( 2007 ) . s. de liberato and c. ciuti , phys . rev
. lett . * 102 * , 136403 ( 2009 ) .
a. k. arora , e .- k .
suh , a. k. ramdas , f. a. chambers and a. l. moretti , phys . rev .
b * 36 * , 6142 ( 1987 ) . m. a. stroscio and m. dutta , _ phonons in nanostructures _ , cambridge university press , ( 2001 ) .
h. frhlich , adv . phys . * 3 * , 325 ( 1954 ) .
d. e. nikonov , a. imamolu , l. v. butov and h. schmidt , phys .
lett . * 79 * , 4633 ( 1997 ) . s .-
c . lee and i. galbraith , phys .
b * 59 * , 15796 ( 1999 ) . s .-
c . lee and i. galbraith , phys .
b * 62 * , 15327 ( 2000 ) .
p. sotirelis , p. von allmen and k. hess , phys .
b * 47 * , 12744 ( 1993 ) .
j. j. hopfield , phys . rev . * 112 * , 1555 ( 1958 ) .
o. verzelen , r. ferreira and g. bastard , phys .
lett . * 88 * , 146803 ( 2002 ) .
s. hameau , y. guldner , o. verzelen , r. ferreira , g. bastard , j. zeman , a. lematre and j. m. grard , phys .
* 83 * , 4152 ( 1999 ) .
p. nataf and c. ciuti , nat
. comm . * 1 * , 72 ( 2010 ) .
parameter for the iii - v group were taken from : _
www.ioffe.ru / sva / nsm / semicond_. for znse we used data presented in : s. adachi and t. taguchi , phys .
b * 43 * , 9569 ( 1991 ) . for cdte we used instead :
a. manabe , a. mitsuishi and h. yoshinaga , jpn .
* 6 * , 593 ( 1967 ) and d. t. f. marple , phys
* 129 * , 2466 ( 1963 ) .
g. gnter , a. a. anappara , j. hees , a. sell , g. biasiol , l. sorba , s. de liberato , c. ciuti , a. tredicucci , a. leitenstorfer and r. huber , nature * 458 * , 178 ( 2009 ) . | we present a microscopic quantum theory of intersubband polarons , quasiparticles originated from the coupling between intersubband transitions and longitudinal optical phonons . to this aim we develop a second quantized theory taking into account both the frhlich interaction between phonons and intersubband transitions and the coulomb interaction between the intersubband transitions themselves .
our results show that the coupling between the phonons and the intersubband transitions is extremely intense , thanks both to the collective nature of the intersubband excitations and to the natural tight confinement of optical phonons .
not only the coupling is strong enough to spectroscopically resolve the resonant splitting between the modes ( strong coupling regime ) , but it can become comparable to the bare frequency of the excitations ( ultrastrong coupling regime ) .
we thus predict the possibility to exploit intersubband polarons both for applied optoelectronic research , where a precise control of the phonon resonances is needed , and also to observe fundamental quantum vacuum physics , typical of the ultrastrong coupling regime . |
recently , there has been an upsurge of research interests in radio frequency ( rf ) energy harvesting / scavenging technique ( see @xcite and references therein ) , which is the capability of converting the received rf signals into electricity .
this technique has become a promising solution to power energy - constrained wireless networks .
conventionally , the energy - constrained wireless networks , such as wireless sensor networks , have a limited lifetime , which significantly confines the network performance . in contrast
, an rf energy harvesting network ( rf - ehn ) can have power supply from a radio environment .
consequently , rf - ehns have found their applications quickly in various forms , such as wireless sensor networks @xcite , wireless body networks @xcite , and wireless charging systems .
for example , @xcite presents a prototype implementation of sensor nodes powered by ambient rf energy . in @xcite
, the authors design an rf - powered integrated circuit with a work - on - demand protocol for wireless body networks in medical applications . with the increasingly emerging applications of rf energy harvesting
/ charging , the wireless power consortium ( www.wirelesspowerconsortium.com ) is also making efforts of establishing an international standard for the rf energy harvesting and transfer technology .
note that the rf energy harvesting typically refers to the capability of the wireless devices to harvest energy from rf signals .
the rf energy transfer refers to as the method and mechanism of an rf source to transmit rf energy to the wireless devices . in rf energy harvesting ,
radio signals with frequency range from @xmath0 ghz to as low as @xmath1 khz are used as a medium to carry energy in a form of electromagnetic radiation .
wireless information is modulated on the amplitude and phase of rf waves , while wireless energy transfer is carried out through far - field rf radiation .
rf energy transfer is characterized by low - power and long - distance transfer , and thus is suitable for powering a large number of devices with low energy consumption , dispersed in a relatively wide area .
due to the specific nature of the rf energy harvesting and wireless communication requirements , wireless networks have to be re - designed to achieve maximal efficiency for rf energy harvesting and transfer .
in particular , the resource allocation for wireless networks has to be optimized considering the tradeoff among network performance , energy efficiency , and rf energy supply .
this article presents recent advances in rf - ehns .
we first provide an overview of the rf - ehns .
then , we introduce and discuss about different resource allocation issues .
furthermore , we highlight the importance of a receiver operation policy by showing a case study in a general rf - ehn . realizing that none of previous works in the literature considers the receiver operation problem with service differentiation
, we aim to fill the gap .
specifically , we devise an optimal operation policy that provides service differentiation among different types of traffic , i.e. , low priority ( lp ) and high priority ( hp ) data , as well as to meet their quality - of - service ( qos ) requirements .
we formulate an optimization model and obtain the optimal operation policy that maximizes the weighted sum of throughput of lp and hp data under the constraints of energy availability and maximum packet loss probability .
finally , we also outline some open research directions in the rf - ehns .
in this section , we first describe a general architecture of an rf - ehn and the circuit design of an rf energy harvester .
then , we introduce the rf energy harvesting technique . a typical centralized architecture of an rf - ehn , as shown in fig .
[ network_architecture ] , has three major components , i.e. , information gateways , the rf energy sources and the network nodes / devices .
the information gateways are generally known as base stations , wireless routers and relays .
the rf energy sources can be either dedicated rf energy transmitters or ambient rf sources ( e.g. , a tv tower ) .
the network nodes are the user equipments that communicate with information gateways .
typically , the information gateways and rf energy sources have continuous and fixed electric supply , while the network nodes harvest energy from rf sources to support their operations . in some cases , the information gateway and rf energy source can be the same .
note that the decentralized rf - ehn also has the similar architecture as shown in fig .
[ network_architecture ] except that the network nodes can communicate among each other .
figure [ network_architecture ] also shows the block diagram of a network node with rf energy harvesting capability .
an rf energy harvesting node consists of the following major components : * the application , to perform some network functions ; * a low - power microcontroller , to process data from the application ; * a low - power rf transceiver , for information transmission or reception ; * an energy harvester , composed of an rf antenna module , an impedance matching , a voltage multiplier and a capacitor , to collect rf signals and convert them into electricity ; * a power management module , which decides whether to store the electricity obtained from the rf energy harvester or to use it for information transmission immediately ; and * an energy storage battery , to reserve the harvested rf energy for future use .
a typical rf energy harvester consists of an antenna module , impedance matching , voltage multiplier and capacitor .
figure [ network_architecture ] also illustrates the block diagram of the rf energy harvester . *
the antenna module can be designed to work on either single frequency or multiple frequency bands , in which the network node can harvest from a single or multiple sources simultaneously , respectively .
nevertheless , the rf energy harvester typically operates over a range of frequencies since energy density of rf signals is diverse in frequency . *
the impedance matching is a resonator circuit operating at the designed frequency to maximize the power transfer between the antenna module and the multiplier .
the efficiency of the impedance matching is high at the designed frequency . *
the main component of the voltage multiplier is a diode of the rectifying circuit which converts rf signals into dc voltage .
generally , higher conversion efficiency can be achieved by diodes with lower built - in voltage .
the capacitor ensures to deliver power smoothly to the load .
in addition , when rf energy is instantaneously unavailable , the capacitor can also serve as a reserve for a short duration . for the general node architecture introduced above
, the network node has the separated rf energy harvester and rf transceiver .
therefore , the node can perform energy harvesting and information transmission simultaneously .
in other words , this architecture supports both _ in - band _ and _ out - of - band _ rf energy harvesting . in the in - band rf energy harvesting ,
the network node can harvest rf energy from the same frequency band as information transmission . on the other hand , in the out - of - hand rf energy harvesting
, the network node harvests rf energy from the different frequency band from that used for information transmission .
since rf signals can carry energy as well as information , theoretically rf energy harvesting and information reception can be performed from the same rf signal input .
this is referred to as simultaneous wireless information and power transfer ( swipt ) @xcite concept .
this concept allows the information receiver and rf energy harvester to operate on the same antenna module . because an existing circuit is not capable of directly extracting energy from the same rf signals for information decoding @xcite , the authors in @xcite introduce practical implementations of a co - located information receiver and energy harvester
specifically , two receiver architectures , namely , _ time switching _ and _ power splitting _ , are presented .
the _ time switching _ architecture allows the network node to switch and use either the information receiver or the rf energy harvester for the received rf signals at a time . on the other hand , in the _
power splitting _ architecture , the received rf signals are split into two streams for the information receiver and rf energy harvester with different power levels . it has been recognized that in theory _ power splitting
_ achieves better information rate and amount of rf energy harvested than those of _
_ @xcite .
however , in practice , an implementation of _ power splitting _ has higher hardware complexity than that of the _ time splitting_. note that the _ power splitting _ architecture allows only in - band rf energy harvesting , while the _ time switching _ architecture can additionally support out - of - band rf energy harvesting . in rf energy harvesting ,
the amount of harvested energy depends on the transmit power , wavelength of the rf signals and the distance between an rf energy source and the harvesting node .
the amount of harvested rf energy can be calculated based on the friis equation @xcite . unlike energy harvesting from other sources , such as solar , wind , geothermal and vibrations , rf energy harvesting has the following characteristics : * rf sources can provide controllable and constant energy transfer over distance for rf energy harvesters , especially for a fixed rf - ehn .
* rf energy harvesting technique is suitable for mobile ( handheld ) devices .
* since the amount of harvested rf energy depends on the distance from the rf source , the network nodes in the different locations can have significant difference in the amount of harvested rf energy . with rf energy harvesting and transfer , proactive replenishment at mobile devices , rather than passive adoption to the environmental resources ,
can be achieved which is more suitable for applications with qos requirements .
the rf sources can mainly be classified into two types : dedicated rf sources and ambient rf sources .
dedicated rf sources can be deployed to provide energy to network nodes when more predictable energy supply is needed .
ambient rf sources are the rf transmitters that are not intended for rf energy transfer ( e.g. , tv and radio tower ) .
this rf energy is essentially free .
ambient rf sources can be static or dynamic .
the study in @xcite is an example of energy harvesting from dynamic ambient rf sources in a cognitive radio network .
a secondary user can harvest rf energy from nearby transmitting primary users , and it can transmit data when it is sufficiently far from primary users or when the nearby primary users are idle . it is reported in @xcite that , in between @xmath2 and @xmath3 meters from a base station , the aggregated power density over the gsm900 downlink frequency band ranges from @xmath4 to @xmath5 @xmath6 indoors everywhere or outdoors on a high level . while the power density received from gsm1800 downlink is in the same order of magnitude as those received from gsm900 frequency band .
also , a state - of - the - art prototype implementation in @xcite is shown to achieve an information rate of @xmath7 kbps between two wireless devices powered by ambient rf signals , at distances of up to @xmath8 feet and @xmath9 feet for outdoors and indoors , respectively
. tested at a variety of locations , the implemented end - to - end system is able to operate battery - free at distances of up to @xmath10 miles from the tv tower .
wireless sensor networks have become the most widely used applications of rf - ehns .
an rf energy harvester can be used in a sensor node to supply energy .
for example , @xcite presents a prototype implementation of sensor nodes powered by ambient rf energy .
the rf - ehn also has attractive healthcare and medical applications such as wireless body network .
benefiting from rf energy harvesting , low - power medical body sensors can achieve real - time work - on - demand power , which further enables a battery - free circuit , reducing the size of the nodes . in @xcite , the authors design the rf - powered energy - efficient application - specific integrated circuit , fabricated in standard @xmath11 cmos technology and featured with a work - on - demand protocol . the integrated circuit is for wireless body networks in medical applications .
furthermore , rf energy harvesting can be used to provide charging capability for a wide variety of low - power mobile devices such as electronic watches , hearing aids , and mp3 players , wireless keyboard and mouse , as most of them consume only micro - watts to milli - watts range of power .
rf - ehns introduce rf energy harvesting as a new function for wireless devices . as a consequence ,
resource allocation in rf - ehns has to take into account two different objectives , i.e. , information transmission / reception and rf energy harvesting .
these issues are receiver operation policy , beamforming , medium access control ( mac ) protocol , cooperative relaying and routing protocol .
moreover , we review the state - of - art design approaches attempted to address these issues .
a receiver operation policy is required for wireless nodes sharing the same antenna or antenna array for information reception and rf energy harvesting .
the policy can be designed to deal with various tradeoffs in the physical layer and mac layer to meet certain performance goals .
most of the existing policies are either based on _ time switching _ or _
power splitting _ architecture .
the focus of the _ time switching _ architecture is to coordinate the time for information reception and rf energy harvesting . on the other hand , for the _ power splitting _ architecture ,
the operation policy is to find an optimal ratio to split the received rf signals .
the authors in @xcite study a simple greedy switching policy based on the _ time switching _ architecture .
the idea of the policy is to let the relay node transmit when its remained energy can support information transmission . through simulation ,
the greedy switching policy is shown to achieve a near optimal performance over a wide range of signal - to - noise ratio ( snr ) . in @xcite ,
the authors consider a three - node amplify - and - forward network with an rf energy harvesting relay .
two relaying protocols for the relay node are proposed based on the _ time switching _ and _ power splitting _ architectures .
specifically , the authors derive the optimal rf energy harvesting time for the _ time switching _ based relaying protocol and the optimal value of power splitting ratio for the _ power splitting _ based relaying protocol .
the evaluation results indicate that the _ time switching _ based relaying protocol is superior in terms of throughput at relatively low snr and high transmission rates .
however , as the transmit power considered is variable , it incurs considerable hardware complexity .
@xcite studies receiver operation policy in a multiple - channel cognitive radio network , in which the secondary user select channels not only for information transmission but also for energy harvesting . in the context of complete csi at the secondary user , the optimal policy for the secondary user to maximize throughput
is determined , based on the remaining energy level and the number of waiting packets in data queue , by applying an mdp - based optimization .
a key concern for rf information and energy transfer is the decay in energy transfer efficiency with the increase of transmission distance due to propagation path loss .
multi - antenna techniques can be used to achieve spatial multiplexing .
furthermore , beamforming techniques employing multiple antennas can be applied to achieve improved efficiency of rf energy transfer @xcite as well as swipt @xcite , without additional bandwidth or increased transmit power .
a problem arising in beamforming is channel state estimation feedback . designing a feedback mechanism is challenging in rf - ehns because existing channel training and feedback mechanisms used for an information receiver are not applicable for an energy harvester due to the hardware limitation .
beamforming is first explored in a three - node multiple - input multiple - output ( mimo ) network @xcite with one transmitter , one energy harvester and one information receiver .
the authors in @xcite study the optimal transmission strategies to achieve tradeoff between information rate and amount of rf energy transferred under the assumption of perfect knowledge of channel state information at the multi - antenna transmitter employing beamforming .
more recently , @xcite considers to utilize energy beamforming in a large - scale mimo system to improve energy efficiency in long - distance power transfer . a resource allocation scheme is proposed to jointly optimize the transmit power and duration of rf energy transfer for maximizing energy efficiency under some qos requirement .
beamforming has also been advocated to provide secure communication in swipt systems with eavesdroppers .
for example , in @xcite , the authors aim to protect transmitted information to the intended receiver by jointly generating artificial noise to the eavesdroppers through beamforming . a non - convex optimization problem for beamforming design
is formulated to minimize the total transmit power , under the requirements for both information transmission and artificial noise generation . to achieve qos support and fairness for information transmissions , a mac protocol specifically designed for rf - ehns
is needed to coordinate the network nodes transmissions .
in addition to the channel access for information transmission , the network nodes also need to spend some time for rf energy harvesting .
the challenge is that the time taken to harvest enough energy is different for different nodes due to various factors such as types of the available rf energy sources and distance .
the mac protocols coordinate network nodes either in a contention - free approach ( e.g. , polling ) or a contention - based approach ( e.g. , csma / ca ) .
the contention - free mac protocol needs to take the node - specific rf energy harvesting process into account to achieve high throughput and fairness . with the contention - based mac protocol ,
each node contends for radio resources for information transmission .
if the rf energy harvesting duration is not optimally decided , an extended delay of resource contention due to communication outage may incur .
in @xcite , the authors present an energy adaptive mac protocol for rf - ehns .
two energy adaptive methods , i.e. , energy adaptive duty cycle and energy adaptive contention algorithms , are proposed to use the node energy harvesting status ( e.g. , rf energy harvesting rate ) as a control variable to manage the node s duty cycle and backoff time , respectively .
however , the energy adaptive mac protocol requires centralized control and out - of - band rf energy supply .
in contrast , the authors of @xcite consider in - band rf energy supply .
a csma / ca - based mac protocol called rf - mac is designed to optimize rf energy delivery rate to meet the energy requirement of sensor nodes while minimizing disruption to data communication .
the rf - mac also incorporates the methods to select rf energy sources .
also , the energy and data rate tradeoff is analyzed .
cooperative relaying can improve the network performance in terms of efficiency and reliability by using intermediate relay nodes .
hence , it is particularly suitable to be applied in energy constrained networks like rf - ehns .
relay selection is a decision factor in the performance of cooperative relaying .
the main challenge lies in that the preferable relay for information transmission does not necessarily coincide with the relay has the strongest channel for energy harvesting .
thus , as a tradeoff , relay selection has to leverage between the efficiency of information and energy transfer .
to this end , other than taking channel state information into account , information about energy status ( e.g. , internal energy reserve and potential external rf energy arrival ) must also be regarded , which makes relay selection more complex . in @xcite ,
the authors study two relay selection schemes , i.e. , the time - sharing selection and the threshold - checking selection . in the time - sharing selection , the source node switches among the relays with the maximum snr . in the threshold - checking selection ,
the source node chooses the relay with the highest rf energy harvesting rate .
it is demonstrated that the threshold - checking selection has better performance in terms of achieved capacity for the given rf energy harvesting requirement . on the other hand ,
the time - sharing selection has better performance in terms of outage probability when the normalized average snr per link is larger than @xmath12db .
@xcite investigates relay selection from a system perspective .
specifically , the authors examine a random relay selection policy based on a sectorized area with central angle at the direction of each receiver .
a geometry approach is adopted to study the impact of cooperative density and relay selection in a large - scale network with swipt .
an rf - ehn can be based on multihop transmission , where routing is a crucial issue .
unlike the energy - aware routing developed in conventional wireless networks , the routing protocols in the rf - ehn must take the rf energy propagation and the circuit design of network nodes ( e.g. , rf energy harvester s sensitivity ) into account .
this is due to the fact that the amount of harvested rf energy available at each node can be different .
in addition , the routing metric has to be jointly defined based on rf energy harvesting parameters ( e.g. , rf signal density , energy conversion rate , and distance from rf sources ) and network parameters ( e.g. , link quality and number of hops ) . in @xcite , the authors consider the routing problem in a wireless sensor network where the sensor nodes are charged wirelessly with in - band rf energy .
it is shown that simple metrics such as a hop count may not be suitable for routing in such networks .
therefore , a new routing metric based on the charging time of the sensor nodes is introduced .
then , the modified ad hoc on - demand distance vector ( aodv ) routing protocol considering the new routing metric is proposed . in this protocol ,
the sensors choose the path with the lowest value of maximum charging time .
in this section , we show a case study of a receiver operation problem with joint qos support and service differentiation in an rf - ehn , which , to the best of our knowledge , has not been addressed before .
we consider a node having two types of data , i.e. , lp and hp data .
the arriving packets ( e.g. , from different applications at higher layers ) are stored in two separate queues . both of the queues have finite sizes .
there is an access point ( ap ) which performs as a information gateway and dedicated rf energy source .
the node in a coverage area of the ap can decide to request for rf energy transfer from the ap or to transmit a packet of lp or hp data to the ap .
specifically , the node adopts the _ time switching _
receiver architecture as proposed in @xcite .
that is , the node works either in an rf energy harvesting mode or an information transmission mode . besides
, a battery , which has a finite capacity , is equipped with the node to store energy harvested from the ap .
we consider the packet loss requirement for each type of data . in this case , a packet is dropped ( i.e. , loss ) if the corresponding queue is full or the battery is empty upon its arrival .
the lp and hp data could have different maximum packet loss probability requirements .
when the node is in the coverage area of an ap , the node is facing a decision making problem ( i.e. , receiver operation ) whether to request and harvest rf energy or to transmit a packet from the queue of lp or hp data to the ap , based on the _ time switching _ architecture .
the decision making problem must be solved to achieve the objective ( i.e. , maximizing weighted sum of throughput of lp and hp data ) and meet the qos requirements taking the following factors into account .
* _ packet arrival _ : the node has independent packet arrivals for hp and lp data .
the probabilities of @xmath13 packets arriving at the node for the lp and hp data are denoted as @xmath14 and @xmath15 , respectively . *
_ packet transmission _ :
if the node decides to transmit a packet retrieved from the queue of either lp or hp data to the ap , the successful packet transmission probability is denoted as @xmath16 .
the transmission of one data packet by the node consumes @xmath17 units of energy from the battery .
if the battery has energy less than @xmath17 units , the node can not transmit the packet . *
_ rf energy harvesting and transfer _ : if the node decides to request for rf energy from the ap , the node can harvest @xmath18 units of energy ( i.e. , the energy level of the battery increases by @xmath18 units ) successfully with the probability @xmath19 .
note that this parameter can be adopted from experiments . to achieve the objective in terms of the throughput and to meet the qos requirement in terms of packet loss probability
, we formulate an optimization model to obtain the optimal operation policy for the node based on constrained markov decision process .
the optimal operation policy determines the action to be taken by the node given its current state .
the state space of the node is defined by the possible energy level in the battery , and the numbers of packets in the queues for lp and hp data . for the node , the possible action ( i.e. , the action space ) will be to transmit a packet from the queue of lp or hp data to the ap , or to request for rf energy from the ap .
the state transition happens in the following events : * _ data transmission : _ the energy level of the battery will reduce by @xmath17 units and the number of packets in the queue of hp or lp data decreases by one ( i.e. , depending on which type of data that the node decides to transmit ) with probability @xmath16 .
these state transitions happen given that the queue of the hp or lp data selected by the node is not empty .
* _ request for rf energy transfer : _ the energy level of the battery increases by @xmath18 units with probability @xmath19 . * _ packet arrival : _ the numbers of packets in the queues of hp and lp data increase by @xmath13 with probabilities @xmath14 and @xmath15 , respectively .
the mapping of a state to an action taken by the node is referred to as the policy denoted by @xmath20 .
the optimal operation policy is defined to achieve the maximum long - term average weighted sum of throughput of the lp and hp data , while the packet loss requirements of lp and hp data are maintained below the thresholds .
the objective function of the optimization model is expressed as follows : @xmath21 where @xmath22 is the function of weighted sum of throughput , @xmath23 and @xmath24 are the weights of lp and hp data , respectively . @xmath25 and @xmath26 are the successful packet transmission probabilities for lp and hp data at time @xmath27 , respectively .
we have @xmath28 , if the node is transmitting a packet from the queue of lp data ( i.e. , the queue of lp data is not empty ) and there is sufficient energy in the battery for transmitting the packet ( i.e. , the energy level of the battery is greater than or equal to @xmath17 ) .
otherwise , we have @xmath29 .
similarly , we have @xmath30 , if the node is transmitting a packet from the queue of hp data ( i.e. , the queue of hp data is not empty ) and there is sufficient energy in the battery . otherwise , we have @xmath31 .
it is worth noting that the data transmission is constrained by rf energy harvesting , i.e. , data transmission is successful provided that there is sufficient harvested energy in the battery .
let us consider lp data .
the constraint of a packet loss probability requirement is expressed as follows : @xmath32 where @xmath33 is the packet loss requirement for the lp data , and @xmath34 is the immediate packet loss probability .
the immediate packet loss probability is @xmath35 , if there is not enough space in the queue for lp data , whose maximum capacity is denoted by @xmath36 .
@xmath37 is the current number of packets in the queue .
@xmath38 is the maximum number of arriving packets .
@xmath39 is the average packet arrival rate of lp data .
the immediate packet loss probability for the hp data can be obtained in a similar way . due to the space limit ,
we omit the derivation of the transition probability matrix .
the detailed derivation of the markov decision process based optimization problem is similar to the approach in @xcite . to obtain the optimal operation policy of the node
, we can apply a standard method to solve the constrained markov decision process @xcite .
the node has the battery with the size of @xmath40 units of energy .
the maximum queue sizes for lp and hp data are @xmath41 packets . unless otherwise stated , the packet arrival probabilities for lp and hp are 0.15 . the successful packet transmission probability of the node to the ap is @xmath42 .
the probability of successful rf energy transfer and harvesting is @xmath43 .
if the rf energy harvesting is successful , the node will receive @xmath41 units of energy .
there is no packet loss probability requirement for the lp data , but it is @xmath5 for the hp data . for the comparison purpose ,
we consider a static policy in which the node chooses three actions with equal probabilities .
@xmath44 ( a ) & ( b ) & ( c ) \end{array}$ ] we first examine the optimal operation policy of the node obtained from solving the optimization model , when the energy level of the battery of the node is low ( i.e. , 5 units ) .
the results shown in fig . [ figure5 ] are obtained by solving the optimization problem ( [ primary ] ) under the constraints of packet lost probability requirement for both lp and hp data .
figure [ figure5](a ) shows that when the number of packets in the queue is small , the node tends to request for rf energy . on the other hand ,
when the number of packets in both queues is high , the node also tends to request for rf energy because of the high demand of energy for data transmission .
as the numbers of packets waiting in the queues grow , the node will be likely to transmit a packet especially for hp data to avoid violating the packet loss requirement .
figures [ figure5](b ) and ( c ) show such a policy for packet transmission . in fig .
[ weight ] , we vary the weight of the hp data and show the corresponding throughput and packet delay .
we observe that when the weight of hp data is large , the optimal operation policy yields more opportunity for the node to transmit a packet from the queue of hp data . as a result , the throughput of the hp data increases , while that of the lp data decreases .
correspondingly , when the weight of hp data increases , the delay of the lp data increases , while that of the hp data decreases .
to achieve the performance goal , the weights of the optimization model can be adjusted .
it is interesting to observe the unbalanced performance improvement and degradation of the hp and lp data , respectively , when the weight is adjusted .
specifically , while the throughput and delay of the hp data improve slightly , those of the lp data degrade significantly .
this is due to the fact that the node has to reserve energy resource by not transmitting a packet from the queue of lp data too much so that the energy can be used for the packet transmission of the hp data in the future .
@xmath45 \end{array}$ ] furthermore , we study the packet loss behavior of the proposed operation policy . for the comparison purpose
, we consider a static policy in which the node chooses three actions with equal probabilities .
figure [ fig : vary_arr ] shows the packet loss probability when the packet arrival rates of both the hp and lp data are varied . as expected ,
when the packet arrival probabilities increase , the packet loss probabilities increase .
however , at a certain point , the optimal operation policy successfully maintains the packet loss probability of the hp data at the defined threshold , which is 0.1 , while that of lp data increases and becomes unbounded .
in addition , we also show the results from the static policy , which fails to achieve acceptable performance , specially for the hp data .
the studied model can be extended to the case of multiple nodes where a scheduling policy is required to coordinate the uplink data transmission and downlink energy transfer .
the operation policy of nodes has to take the scheduling policy into account .
in this section , we discuss about some open research issues as followings . * _ distributed energy beamforming _ : distributed energy beamforming enables a cluster of distributed energy sources to cooperatively emulate an antenna array by transmitting rf energy simultaneously in the same direction to an intended energy receiver for better diversity gains .
the potential energy gains at the receiver from distributed energy beamforming are expected to be the same as that from the well - known information beamforming .
however , challenges arise in the implementation , e.g. , time synchronization among energy sources and coordination of distributed carriers in phase and frequency so that rf signals can be combined constructively at the receiver . * _ cooperative sensing and spectrum sharing _ : in contrast to a conventional cognitive radio network , spectrum occupancy by primary users is not necessarily undesirable in rf - ehns , as it results in rf energy harvesting opportunities . the cooperative spectrum sensing and sharing techniques for cognitive radio networks can be directly adopted to help secondary users with rf energy harvesting capability to identify the occupied spectrum bands and rf energy harvesting opportunity . however , because of the dispersed geographic locations of the secondary users , they may experience different spectrum conditions due to different activities and locations of primary users .
high utilization of frequency band and high efficiency in detecting frequency usage require information exchange and fusion among secondary users , which can be a challenging task in network design . *
_ interference management _ : existing interference management techniques , e.g. , interference alignment and interference cancellation , attempt to avoid or mitigate interference through spectrum scheduling
. however , with rf energy harvesting , harmful interference can be turned to useful energy through a scheduling policy .
in addition , the scheduling policy can be combined or integrated with power management schemes for further improvement in energy efficiency . * _ energy trading _ : in rf - ehns , rf energy becomes a valuable resource .
the rf energy market can be established to economically manage this energy resource jointly with radio resource .
for example , wireless charging service providers may act as rf energy suppliers to meet the energy demand from network nodes .
the wireless energy service providers can decide on pricing and guarantee the quality of charging service .
the key feature of this market is on - demand trading where the wireless charging service providers offer charging service to the network nodes on a real - time and on - demand basis .
one of the efficient approaches in this dynamic market is to develop demand side management , which allows the service providers and network nodes to interact like in smart grid , to guarantee energy - efficiency and reliability .
however , the issues related to the amount of rf energy and price at which they are willing to trade while optimizing the tradeoff between the revenue and cost must be investigated . *
_ wireless machine - to - machine ( m2 m ) communications _ deploying massive and unmanned wireless m2 m devices introduces the new challenge : how could such a huge amount of devices be powered ?
rf energy harvesting provides an alternative solution .
for example , the `` last meter '' technologies , e.g. , wifi , ieee 802.15 , zigbee and uwb , can be potentially used for rf energy supply for wireless m2 m devices . * _ vehicular communications _
vehicular transmitters in vehicular networks ( i.e. , vehicle - to - vehicle and vehicle - to - infrastructure communications ) can be energy sources for wireless devices belonged to passengers or even pedestrians .
for example , the passengers devices can harvest rf energy from an on - board unit deployed in a bus .
alternatively , when the passengers are at the bus stop , their devices can also harvest rf energy from roadside units . *
_ smart automation _
: in automation systems , rf energy harvesting can eliminate the wired connection for power supply for sensor and actuator devices , especially for the devices installed on the moving components , e.g. , wheel and rotating assemblies .
for example , the turbine blade sensors on mechanical engines can be steadily provisioned by dedicated rf sources to maintain the function of monitoring on the blades status . * _ device - to - device ( d2d ) communications _ : d2d communications , underlaid or overlaid with cellular networks , allows user equipment to access the same spectrum band for cellular communication with interference constraints .
the occupied spectrum provides rf energy harvesting opportunities , especially when network density is high .
the user equipment can harvest and use rf energy for their local direct d2d communications .
radio frequency ( rf ) energy harvesting and transfer techniques play an important role in powering the next generation of wireless networks . in this article , we have presented an overview of the rf energy harvesting networks ( rf - ehns ) , including the network architecture and the enabling techniques .
we have introduced the major design issues in resource allocation of the rf - ehns , and reviewed some up - to - date research progresses .
moreover , we have shown a case study on how to design a qos - aware receiver operation policy with service differentiation in a general rf - ehn .
we obtain an optimal operation policy to maximize the throughput of a mobile node with rf energy harvesting capability , as well as to provide service differentiation among two different types of data , under the constraints of packet loss probability .
in addition , we foresee the future research directions of the rf - ehns .
h. j. visser , and r. j. vullers , rf energy harvesting and transport for wireless sensor network applications : principles and requirements , " in _ proceedings of the ieee _ ,
vol . 101 , no . 6 , pp .
1410 - 1423 , june 2013 .
x. zhang , h. jiang , l. zhang , c. zhang , z. wang , and x. chen , an energy - efficient asic for wireless body sensor networks in medical applications , " _ ieee transactions on biomedical circuits and systems _ , vol .
11 - 18 , february 2010 .
v. liu , a. parks , v. talla , s. gollakota , d. wetherall , and j. r. smith , ambient backscatter : wireless communication out of thin air , " in _ proceedings of the acm conference on sigcomm _ , pp .
39 - 50 , ny , usa , aug . 2013 .
i. krikidis , s. timotheou , and s. sasaki , rf energy transfer for cooperative networks : data relaying or energy harvesting ? " _ ieee communications letters _ , vol .
11 , pp . 1772 - 1775 ,
november 2012 .
a. a. nasir , x. zhou , s. durrani , and r. a. kennedy , relaying protocols for wireless energy harvesting and information processing , " _ ieee transactions on wireless communications _
12 , no . 7 , pp . 3622 - 3636 , july 2013 .
d. niyato , p. wang , w .- l .
yeow , and h. p. tan , qos - aware data transmission and wireless energy transfer : performance modeling and optimization , " in _ proc . of ieee wireless communication @xmath46 networking conference ( wcnc ) _ ,
istanbul , turkey , 6 - 9 april 2014 .
x. chen , x. wang , and x. chen , energy - efficient optimization for wireless information and power transfer in large - scale mimo systems employing energy beamforming , " _ ieee wireless communications letters _ , vol .
2 , no . 6 , pp .
667 - 670 , dec . 2013 .
d. w. k. ng , e. s. lo , and r. schober , robust beamforming for secure communication in systems with wireless information and power transfer . " to appear in _ ieee transactions on wireless communications_. j. kim and j .- w .
lee , `` energy adaptive mac protocol for wireless sensor networks with rf energy transfer , '' in _ proceedings of ieee third international conference on ubiquitous and future networks ( icufn ) _ , pp .
89 - 94 , dalian , china , june 2011 .
p. nintanavongsa , m. y. naderi , and k. r. chowdhury , `` medium access control protocol design for sensors powered by wireless energy transfer , '' in _ proceedings of ieee infocom - mini conference _ , pp .
150 - 154 , turin , italy , april 2013 .
d. s. michalopoulos , h. a. suraweera , and r. schober , relay selection for simultaneous information transmission and wireless energy transfer : a tradeoff perspective . "
( available on - line at arxiv:1303.1647 ) .
r. doost , k. r. chowdhury , and m. di felice , `` routing and link layer protocol design for sensor networks with wireless energy transfer , '' in _ proceedings of global telecommunications conference _ ,
miami , florida , usa , december 2010 . | radio frequency ( rf ) energy harvesting and transfer techniques have recently become alternative methods to power the next generation of wireless networks .
as this emerging technology enables proactive replenishment of wireless devices , it is advantageous in supporting applications with quality - of - service ( qos ) requirement .
this article focuses on the resource allocation issues in wireless networks with rf energy harvesting capability , referred to as rf energy harvesting networks ( rf - ehns ) .
first , we present an overview of the rf - ehns , followed by a review of a variety of issues regarding resource allocation .
then , we present a case study of designing in the receiver operation policy , which is of paramount importance in the rf - ehns .
we focus on qos support and service differentiation , which have not been addressed by previous literatures .
furthermore , we outline some open research directions . |
the construction and study of area preserving mappings has led to a deeper understanding of apparently complex dynamics , especially of hamiltonian chaos .
the maps range from abstract models @xcite such as the cat maps and the baker map to more generic ones .
one of the most well studied of such generic mappings is the standard map @xcite , which has also been investigated extensively in its quantum version @xcite .
the classical map on the cylinder displays a range of dynamical behaviours , and when completely chaotic , diffusive random walks in momentum take place .
recent experiments using trapped ultra - cold sodium atoms in pulsed laser fields have probed this model , as a kicked rotator , and verified the central phenomenon of quantum localization of momentum diffusion @xcite . in this letter
we study the classical dynamics of a simple generalization of the standard map that naturally arises from the dynamics of a particle trapped in a one dimensional infinitely deep well .
the virtue of this generalization is that on the introduction of a competing length scale novel dynamical behaviour manifests , leading to large scale diffusive processes even for small external field strengths .
this places the standard map within a family of generally discontinuous area preserving mappings ; the points where there is continuity corresponding to the usual standard map .
we study the novel dynamical changes by simple methods exploiting separation of time scales in non - autonomous equations .
the failure of the poincar - birkhoff theorem leads to the formation of entirely stable islands and cantori or entirely unstable and chaotic orbits .
it is also to be pointed out that this can be one of the methods of controlling or enhancing chaos .
the classical motion of a particle in an infinite square well potential in the presence of a uniform monochromatic external field has been studied previously @xcite . in this case diffusion
was observed over a limited range of energies and there is no issue of competing length scales .
also , no analytic form of the stroboscopic mapping is derivable .
we have considered below a space variation of the external field that is also pulsed in time , the pulsing leading to analytically derivable mappings .
more recently a special case of this system was studied @xcite and quantized , where ( in the chaotic regime ) an important effect was found , namely delocalization of eigenstates as opposed to the well known exponentially localized states of the kicked rotator .
our independent and parallel studies are a generalization and in this letter we present the classical aspects as we feel that these deserve a more complete understanding .
recent advances in semiconductor physics , makes it possible to construct wells on atomic scales .
also , developments in investigating the quantum nature of electrons in a finite quantum well have been achieved , using tunnel - current spectroscopy , when the corresponding classical system shows chaotic behaviour @xcite ( see references therein ) .
we hope that the study of novel dynamical behaviours in simple models , as the generalization of the standard map , will reflect on some aspects of such experimentally realizable systems .
we consider a particle of mass @xmath0 , trapped in an one dimensional infinite square well potential @xmath1 of width @xmath2 .
there is an external field @xmath3 that is periodically pulsed with period @xmath4 .
the hamiltonian is @xmath5 where @xmath6 are the position and momentum of the particle respectively .
we consider below @xmath7 . here
@xmath8 and @xmath9 are the amplitude and wavelength of the external field respectively .
the effect of the pulsed field in @xmath10 is the same as that of an infinite number of travelling waves with identical amplitudes and frequencies which are multiples of @xmath11 ( the pulse frequency ) .
the stroboscopic map , relating the dynamical variables immediately after successive kicks , can be derived in a standard manner and results in the following dimensionless form : @xmath12 p_{n+1}&=&{(-1)}^{m_n } p_n + ( k/2 \pi ) \sin(2 \pi rx_{n+1 } ) . \end{array } \label{map}\ ] ] where @xmath13 $ ] , is the number of bounces of the particle at the walls during the interval between the @xmath14th kick and the @xmath15th kick . here
sgn ( .. ) and @xmath16 $ ] stand for sign and integer part of the argument respectively .
the dimensionless quantities are defined as @xmath17 while the dynamical variables @xmath18 , @xmath19 are the position and momentum of the particle just after the @xmath14th kick .
although sgn(@xmath20 ) is discontinuous and undefined for @xmath21 , one can see easily from the map that either of the values ( @xmath22 ) can be taken for @xmath23 as this does not alter the dynamics .
we refer to the mapping ( [ map ] ) as the _
well map_. note that @xmath24 and @xmath25 are the only two effective parameters of the well map . here
@xmath24 is related to the strength of the external field ; @xmath25 is the ratio of the width of the well to the wavelength of the external field .
we do not impose any constraints on @xmath25 , for example those while considering standing waves in a cavity , but allow it all values .
evidently , @xmath26 or @xmath27 as the particle motion is confined between two rigid walls .
we now relate the well map to a generalized standard map .
the motion of a free particle inside the potential well @xmath1 is related to the motion of a free rotator , the difference being one of boundary conditions .
both free motions are same unless the particle hits the walls .
the correspondence between them is made explicit in the following way .
let us consider the dynamics of a free rotator in discrete unit time steps , which is the map relating successive angles @xmath28 and angular momenta @xmath29 .
@xmath30 here the motion is confined to a cylinder @xmath31 .
similarly the discretized free motion in a well of unit width is @xmath32 let us denote @xmath33 , and @xmath34 .
if we have the same initial condition for both the maps ( [ rotor ] ) and ( [ particle ] ) i.e. , @xmath35 , we have the following relation after the @xmath14th iterate : @xmath36 where @xmath37 , is the _ total _ number of bounces of the particle after @xmath38 iterates . in the rest of the paper we consider the mapping @xmath39
{ \theta}_{n+1}&=&{\theta}_{n}+j_{n+1 } \hspace*{0.6 cm } ( \mbox{mod}\;\;1 ) . \end{array } \label{gsm}\ ] ] it is easy to see that we have the same relation ( [ state ] ) between the well map ( [ map ] ) and the map defined in ( [ gsm ] ) ( more precisely the time reversal of this map ) .
we refer to the map defined in ( [ gsm ] ) as the _ generalized standard map _ ( gsm ) . from the above description
it is clear that quantitative dynamical features are same for the well map and the gsm . in attempting to analyze the dynamics of the well map
, it is sufficient to understand the dynamics of the gsm .
also the map ( [ gsm ] ) is easier to handle than the well map , as it is only a slightly generalized version of the well studied standard map ( @xmath40 ) @xcite .
when it is stated that the standard map is a one parameter system the implicit assumption is that there is only one length scale .
here there are naturally _ two _ length scales whose ratio is the cause of many interesting effects .
both the maps ( [ map ] ) and ( [ gsm ] ) have reflection symmetry about their origin .
below we discuss various dynamical behaviour of the gsm in more detail . throughout this paper
we consider @xmath25 to assume positive real values only .
for the standard map ( gsm with @xmath40 ) , when @xmath41 , the dynamics is quasi - regular and the phase space is filled with a large number of kam tori which are _ rotationally invariant circles _ @xcite extending across the phase plane .
these closed loops @xmath42 encircle the cylinder and are invariant , meaning that @xmath43 , where @xmath44 is an area preserving _ continuous _ transformation on the cylinder ( for instance eqn.([gsm ] ) with @xmath40 ) .
kam tori are the principal barriers for the unstable orbits to diffuse in the phase space .
as @xmath24 increases there is a _ smooth _
transition from regular to chaotic behaviour with a reduction in the number of kam tori .
dynamical changes of kam tori with increase of @xmath24 are discussed in @xcite . at @xmath45
all the kam tori disappear from the phase space @xcite ; the dynamics is chaotic and diffusive for @xmath46 .
note that the gsm is periodic in @xmath47 and @xmath48 with unit period .
for integer values of @xmath25 , it is continuous and takes the form of standard map with @xmath49 . in this case
dynamical transitions are identical with that of the standard map .
as the effective perturbation here is @xmath50 , higher integer values of @xmath25 also make the dynamics more chaotic .
when @xmath25 assumes non - integer values the situation is entirely different from the earlier case , particularly when @xmath24 is small . shown in fig .
[ xp ] are typical phase space structures of the gsm for @xmath51 and @xmath52 at small @xmath24 .
it is seen that no kam tori exist even when @xmath25 is slightly away from unity . in general , no kam tori appear in the phase space when @xmath25 departs from integer values .
this is of course due to the discontinuity in the map when @xmath25 assumes non - integer values .
in other words , there are no closed loops @xmath42 encircling the cylinder such that @xmath53 , where @xmath54 is the transformation defined by gsm with non - integer values of @xmath25 .
as can be seen easily the discontinuity in @xmath54 arises at @xmath55 .
this discontinuous dynamics implies the abrupt breakup of kam tori and a failure of the poincar- birkhoff scenario .
the modifications that may occur are little understood , and chirikov in @xcite points out that the piecewise linear sawtooth map @xmath56 when @xmath24 is a non - integer such that @xmath57 is one such system where extremely complicated locally stable motions occur and `` it is not at all clear what could be a meaningful description , if any , of this apparently trivial model '' .
the particle in a well naturally gives rise to a non - linear generalization of the sawtooth map and below we show how a simple method may give us significant local stability information that explains the dramatic differences observed in fig .
[ xp ] .
in addition fig .
[ xp ] distinguishes the dynamics between the cases @xmath58 and @xmath59 .
the former case is more chaotic and the phase space is filled with unstable orbits , while many regular and chaotic regions are seen in the latter . thus the entire standard map may be regarded as being poised at a point of bifurcation when regarded as a function of the parameter @xmath25 .
however , it is seen that when @xmath24 is large , there is no such apparent difference in dynamics between the above two cases since the phase space becomes chaotic and diffusive in both the cases . in the absence of the general kam theorem for irrational tori and the poincar - birkhoff theorem for rational tori
, we expect to apply some kind of local analysis to study deviations from integer values of @xmath25 .
indeed we will see that the integer values of @xmath25 could be thought of as bifurcation points for entire orbits .
we noted earlier that even for small values of @xmath24 no kam tori exist in the phase space of the gsm when @xmath25 departs from unity . to explain the stability changes we look at the behaviour of a single irrational kam torus , which exists at @xmath40 , as @xmath25 varies .
a typical observation ( fig . [ kam ] ) shows that kam torus becomes unstable and chaotic for @xmath58 while it is replaced by a chain of stable islands for @xmath59 .
we begin by performing a linear stability analysis for a kam tori with respect to @xmath25 .
we assume below that @xmath24 is small enough for such an orbit to exist .
let us denote a gsm ( with @xmath60 , positive integer ) kam orbit as @xmath61 .
this satisfies the following mapping : also consider an orbit @xmath63 of the gsm with @xmath64 and @xmath65 , with the same initial conditions as the above kam orbit .
let us introduce the differences @xmath66 ; @xmath67 .
then it is possible to write an exact mapping equation for @xmath68 , @xmath69 as expanding in terms of @xmath71 and retaining first order terms leads to a time dependent force and the non - autonomous linear set of equations : @xmath72 { \delta}\theta_{n+1}&=&{\delta}\theta_n + { \delta}j_{n+1 } \end{array } \label{del1}\ ] ] with @xmath73 the behaviour of such linear non - autonomous equations can be quite complex ( compare for instance the mathieu differential equation which also arises in linear stability analysis ) . for our analysis on the stability of the tori we wish be intuitive and derive rough but useful estimates .
although we never expand @xmath25 about an integer , here we assume that this excursion is small , so that we can treat the non - autonomous stability equations perturbatively .
appealing to the method of averaging we simply replace the time dependent force by its time average .
the limitation of this is pointed out further ahead .
eqn.([lmap ] ) is a simple linear map and its stability can be seen from the corresponding jacobian matrix which gives the stability condition @xmath77 .
this leads to the following window of stability : @xmath78 the above stability window implies that if @xmath79 is odd kam tori which exist at @xmath60 are stable for @xmath80 and unstable for @xmath81 ; converse is the case if @xmath79 is even .
although @xmath82 is small in the above analysis , our numerical observations satisfy the stability condition ( [ win ] ) even for large values of @xmath82 ( see fig .
[ kam ] , which is with @xmath83 ) .
this explains large phase space regions in fig .
[ xp ] . before dwelling more in this region of phase space
we divert attention to those where this procedure has failed . the chaotic regions around @xmath84 in fig .
[ xp ] for instance presents such a case . in the interior of the original @xmath85 resonance ( at @xmath40 )
even a reversal of the stability criteria is observed .
the hyperbolic fixed point at @xmath40 persists for @xmath86 and is the genesis of the chaos .
the smooth stable and unstable manifolds of the hyperbolic fixed point that exist for @xmath40 and small @xmath24 can not exist for @xmath87 as they would then represent rotationally invariant curves .
therefore the presence of the hyperbolic points necessarily implies chaos . within the framework of the stability analysis done earlier what has failed around the hyperbolic fixed points
is the assumption of uniform measure .
if we think of the simple pendulum we are moving from energetic rotational motions ( kam ) to slow oscillations of large amplitude and clearly the time is spent preferentially around the turning points . also the replacement with time averages will fail as the time scales involving fast @xmath48 motions and slow @xmath88 motions become comparable .
the latter effect becomes important as we move into the @xmath85 resonance region .
we note that the question of the existence of two different time scales is initial value dependent and is present in the standard map as well , _
i.e. , _ it does not arise as a result of the presence of the additional parameter @xmath25 .
the modified measure can be approximated well by considering the simple pendulum .
we now take @xmath83 for simplicity and consider only excursions of @xmath25 from unity . the term @xmath89 in eq .
( [ del1 ] ) is replaced with the ergodic average : @xmath90 here @xmath91 is the scaled energy and @xmath92 is the turning point given by @xmath93 .
this is shown in fig .
[ avg2 ] and the point where @xmath94 crosses zero from above ought to be a point where stability is recovered .
this then explains the uniform chaos around the hyperbolic fixed point for @xmath95 .
this also explains the stability around the wall @xmath96 in all cases .
however the interior of the resonance is not accessible to the simple theory .
remarkably fig .
[ avg2 ] predicts a recovery of stability at @xmath97 which is seen when @xmath98 while stability is recovered at a much higher value of @xmath92 for @xmath99 and is not seen in fig .
[ avg2 ] . this fact we attribute to the failure of the averaging procedure in analyzing non - autonomous equations and have no other recourse than eq .
( [ del1 ] ) itself .
as we have noted , hyperbolic fixed points generate chaos when @xmath100 .
the lack of chaos in the `` meat '' of the phase space for @xmath101 where there were kam tori are therefore regions where there is a total absence of hyperbolic points .
the poincar - birkhoff theorem concerning the breakup of rational tori into an equal number of hyperbolic and elliptic periodic points is clearly violated .
in fact the modified scenario seems to be the creation of only elliptic fixed points for @xmath59 and only hyperbolic points for @xmath58 . to illustrate this
we have shown in fig .
[ gm ] the fate of points on the lines @xmath102 and @xmath103 .
while in the former case three elliptic islands emerge prominently , in the latter one can see chains of elliptic islands with the clear hierarchy of the number of islands @xmath104 deriving from the fibonacci sequences generating closer approximations to the golden mean .
these island chains alternate as the ratios approach the golden mean . as a limiting case we may think of a golden mean chain , but we have of course not rigorously proved its existence .
there appear to be two types of orbits in the stable regions : those that form an island chain and those that meander in the interstitial spaces between island chains on presumably a fractal set .
these would then be examples of strange non - chaotic sets in hamiltonian mechanics .
more investigations into these simple systems seems warranted .
we have thus acquired a rather detailed understanding , using simple methods , of the dramatic stability changes that accompanies the introduction of an additional length scale or discontinuity .
we may also speculate that this may prove a good way of controlling or enhancing chaos . while our study has been for hamiltonian systems dissipative systems may also display such a behaviour .
the lyapunov exponents can be calculated using the jacobian matrix of the gsm , which at @xmath14th iterate we denote as @xmath105 .
the stability changes induced by @xmath25 ought to reflect on the exponent somewhat like a phase transition effect on an order parameter . shown in fig .
[ le ] is a typical calculation of the positive lyapunov exponent @xmath106 for @xmath107 .
the orbit chosen is a kam torus for @xmath40 and we , from our earlier discussion , expect it to be unstable for @xmath108 and stable for @xmath109 but not too large .
it is seen that @xmath110 for @xmath59 in the range shown , which corresponds to the regular motion due to the stable islands formed from the orbit ; for @xmath108 , @xmath111 which indicates the chaotic behaviour of the orbit . also the orbit has a maximum value of the lyapunov exponent at @xmath112 , and is hence maximally unstable at this point .
we now explore this analytically .
the jacobian has trace @xmath113 and hence @xmath114 for @xmath115 as @xmath116 .
we remind the readers that this range of @xmath117 is the _ entire _ configuration space and @xmath118 represent the wall boundaries . since the jacobian is only a function of the angle at a given phase space point ( and not also the action )
, we conclude that the jacobian has real eigenvalues _ throughout _ the phase space .
in other words , the system is hyperbolic for @xmath115 .
this implies that there are contracting and expanding real directions or alternatively stable and unstable manifolds throughout phase space .
it is worth remarking that very few dynamical models are known exactly to be completely chaotic or hyperbolic .
these include the sawtooth maps @xcite , the baker maps , flows on surfaces of constant negative curvature .
the standard map even for large values of the parameter @xmath24 is not proven to be completely hyperbolic .
thus in this context , we place the fact that gsm has a parameter range for which it is completely hyperbolic for all positive values of the parameter @xmath24 .
the similarity of the gsm , when @xmath115 , with the piecewise linear sawtooth maps is made clear by the following linear approximation .
we approximate the force @xmath119 in the gsm by a monotonic linear function @xmath120 .
this leads to @xmath121 .
the lyapunov exponents in this approximation are given by @xmath122 where @xmath123 .
it is clear from fig .
[ le ] that the approximated exponent fits fairly well .
similar agreement was seen for a wide range of initial conditions and different values of @xmath24 as well .
this shows the validity of the above linear approximation for @xmath124 when @xmath115 , which enables us to understand the gross behaviour of the lyapunov exponents in this regime .
we note that the usual linear approximation @xmath125 near the origin was found not to be as good as the above approximation .
although this study has been essentially that of another area - preserving map , we feel that these novel effects are worth emphasizing , especially since the standard map has been realized in the laboratory in the form of kicked cold sodium atoms , and the transitions to chaos in the quantum regime have been studied as noted above .
implementation of the competing length scales in this experiment could lead to novel observable transport effects .
theoretical studies of the quantum dynamics of this system are currently underway .
arnold , and a. avez , ergodic problems of classical mechanics , wa benjamin , new york ( 1968 ) .
chirikov , phys .
52 ( 1979 ) 263 .
izrailev , phys .
196 ( 1990 ) 299 .
moore , j.c .
robinson , c.f .
bharucha , bala sundaram and m.g .
raizen , phys .
lett . 75 ( 1995 ) 4598 .
lin and l.e .
reichl , physica d 19 ( 1986 ) 145 .
bambi hu , baowen li , jie liu and yan gu , phys .
( 1999 ) 4224 .
wilkinson , t.m .
fromhold , l. eaves , f.w .
sheard , n. miura and t. takamasu , nature 380 ( 1996 ) 608 .
64 ( 1992 ) 795 .
s.j . shenker and l.p .
kadanoff , j. stat .
27 ( 4 ) ( 1982 ) 631 .
greene , j. math .
( 1979 ) 1183 .
giannoni , a. voros , and j. zinn - justin ( eds . ) , chaos and quantum physics , les houches lii ( north - holland , amsterdam 1991 ) .
q. chen , i. dana , j. d. meiss , and i. c. percival , physica d 46 ( 1990 ) 217 . + fig 1 : phase space of the gsm ( with @xmath126 ) at @xmath127 .
no kam tori are seen in any of the cases .
+ fig 2 : shown is an orbit with initial condition @xmath128 , at @xmath107 , which corresponds to a kam torus when @xmath40 . for @xmath58 ,
the orbit becomes unstable and chaotic , while for @xmath59 it becomes chain of stable islands .
+ fig 3 : the function @xmath129 which when positive determines the instability of some gsm orbits . + fig 4 : orbits with initial conditions on the line @xmath102 ( top ) and @xmath103 ( bottom ) for @xmath130 and @xmath99 . + fig 5 : positive lyapunov exponent ( dots ) of the orbit shown in fig .
the exponent is positive for @xmath108 with maximum at @xmath112 and zero for @xmath131 .
solid curve is the approximation ( [ le_approx ] ) for @xmath115 . | a discontinuous generalization of the standard map , which arises naturally as the dynamics of a periodically kicked particle in a one dimensional infinite square well potential , is examined .
existence of competing length scales , namely the width of the well and the wavelength of the external field , introduce novel dynamical behaviour .
deterministic chaos induced diffusion is observed for weak field strengths as the length scales do not match .
this is related to an abrupt breakdown of rotationally invariant curves and in particular kam tori .
an approximate stability theory is derived wherein the usual standard map is a point of `` bifurcation '' .
+ pacs number(s ) : 05.45ac , 45.05.+x , 47.52.+j + keywords : chaos |
metastable helium has been the subject of many experimental investigations at cold and ultracold temperatures .
these include various methods of laser cooling and trapping , production of an intense beam @xcite , optical collisions in magneto - optical traps and measurements of two - body trap loss rates , including that due to penning ionization , photoassociation spectroscopy @xcite and magnetostatic trapping @xcite .
there have also been several theoretical studies .
much of this interest has been stimulated by the prospect of obtaining a bose - einstein condensate with spin - polarized metastable helium @xmath3 atoms @xcite ; a quest successfully realized very recently @xcite .
the penning ionization ( pi ) and associative ionization ( ai ) processes , @xmath4 have high threshold rates in an unpolarized gas and limit the achievable density of trapped atoms . here
we denote a @xmath5 ) atom by the symbol @xmath6 .
however , these autoionization processes are spin - forbidden and suppressed from the spin - polarized state and only via the weak spin - dipole interaction can such processes occur .
consequently , a sufficient number of spin - polarized metastable atoms should remain trapped . in addition , the scattering length associated with the quintet potential , which controls the collision dynamics of spin - polarized metastable helium atoms , is predicted to be large and positive , a necessary requirement for a stable bose - einstein condensate . although some theoretical studies @xcite have estimated the scattering length associated with the quintet potential to be large and positive , no detailed study of the scattering lengths for metastable helium has been previously undertaken . in this present investigation
we calculate not only the possible ranges of values for the scattering lengths directly associated with the molecular potentials , but also experimentally observable scattering lengths , elastic cross sections and inelastic rate constants over a range of scattering lengths and temperatures for collisions of metastable helium atoms in the presence of a magnetic field .
measurement of cross sections or rates should then provide information on the scattering lengths and hence the potentials . for this theoretical investigation
we have chosen to simulate the penning and associative ionization processes that occur at small internuclear separations from the singlet and triplet molecular states by a complex optical potential .
the complex interaction potentials then have the form @xmath7 , where @xmath8 is the internuclear separation of the two atoms , @xmath9 is the usual adiabatic molecular potential for the molecular state @xmath10 with total spin @xmath11 , and @xmath12 is the corresponding total autoionization width representing flux loss due to the ionization processes .
since the penning and associative ionization processes are spin - forbidden from the quintet state , @xmath13 . in the absence of spectroscopic data which could be used to obtain high accuracy potentials , the adiabatic molecular potentials required in this investigation for the @xmath0 , @xmath1 and @xmath2 molecular states
were constructed using data from various sources .
the long - range interaction potential was described by a multipole expansion of the form @xmath14 using the most accurate dispersion coefficients available for the he@xmath15he@xmath15 interaction ( @xmath16 au ) . the short - range @xmath0 and @xmath1 molecular potentials and their corresponding autoionization widths for penning and associative ionization were obtained from mller _ et al . _
@xcite , while the short - range @xmath2 molecular potential was taken from strck and meyer .
the @xmath2 potential was reported with an uncertainty of 0.5% in the repulsive part of the potential and 1% in the attractive part of the potential .
the molecular potentials for metastable helium were constructed by fitting the three short - range potentials smoothly onto the long - range dispersion interaction around @xmath17 by interpolation through this region using an akima spline fitted to @xmath18 .
the uncertainties in the short - range potentials , the procedure used to connect these to the long - range potential , and the form used for the autoionization widths lead to uncertainties in the scattering lengths for the @xmath0 , @xmath1 and @xmath2 potentials and subsequently to the ultracold scattering properties of metastable helium atoms . to determine the extent of these uncertainties we vary the short - range potentials by @xmath19% for five different interaction potentials which use different methods to connect the long - range and short - range potentials or have different forms of the autoionization widths . since there is no available experimental data that can be used to determine the level of accuracy of these short - range potentials , we have chosen to vary them by more than their stated uncertainty to ensure that we obtain conservative estimates for the range of scattering lengths .
the first of these potentials , labeled ( a ) , uses the analytic short - range @xmath2 potential fitted smoothly onto the long - range potential at @xmath20 .
the numerical @xmath0 and @xmath1 molecular potentials of ref .
@xcite are used for @xmath21 and for larger @xmath8 , where the electronic structure calculations become inaccurate , we replace the potentials by @xmath22 .
the exchange term has the form @xcite @xmath23 , where @xcite @xmath24 , @xmath25 , @xmath26 and @xmath27 .
the autoionization widths @xmath28 of ref .
@xcite were used to represent the penning and associative ionization processes .
potential ( b ) is identical to ( a ) except that the short - range @xmath2 form is fitted to the long - range potential at @xmath29 .
potential ( c ) is identical to ( a ) except that the exchange term has the form @xmath30 with @xmath31 , @xmath32 and @xmath33 .
potentials ( d ) and ( e ) are identical to ( a ) but employ different forms for the autoionization widths . the autoionization width @xmath34 , given by garrison _ et al . _
@xcite , is used in ( d ) .
this autoionization width has a steeper exponential form which does nt dampen at small internuclear separations like @xmath35 or @xmath36 .
potential ( e ) uses another alternative form of the autoionization widths which was arbitrarily constructed to assess the sensitivity of the calculated results to the form of @xmath37 and is given by : @xmath38 all the molecular potentials considered have the same long - range form since the uncertainties in the long - range multipole potential were found to have a negligible effect on the scattering lengths .
the real parts of the potentials ( a ) to ( e ) with unmodified short - range forms possess the same number of bound states , calculated to be 28 for @xmath0 , 27 for @xmath1 and 15 for @xmath2 .
the scattering lengths for the @xmath0 , @xmath39 and @xmath2 molecular potentials were obtained by solving a single channel radial schrdinger equation of the form , @xmath40 + k^2 \biggr\ } u_{s , l } ( k , r ) = 0 \label{eq : sf1}\end{aligned}\ ] ] in the limit where @xmath41 . here
@xmath42 , @xmath43 is the reduced mass of the atomic system , @xmath44 is the total energy of the system and @xmath45 is the relative rotational angular momentum . as a result of the complex interaction potential , the scattering equation ( [ eq : sf1 ] ) and the wavefunctions @xmath46 are complex .
solution of this equation allowing for the non - unitarity of the hamiltonian , and subsequent fitting to free - field boundary conditions , provides a complex k - matrix and corresponding non - unitary s - matrix ( @xmath47 ) , as described previously @xcite .
the complex phase shift @xmath48 , defined by @xmath49 , can then be used to calculate the complex scattering lengths @xmath50 for each molecular state @xmath10 : @xmath51 where the scattering lengths are defined by @xmath52 and the superscripts ` @xmath53 ' and ` @xmath54 ' denote real and imaginary components , respectively .
this definition means that @xmath55 represents a loss process .
the scattering lengths for the three molecular states were calculated as a function of the percentage variation in the corresponding short - range molecular potential for the five potential cases ( a ) to ( e ) and are displayed in fig .
[ a5a1a3 ] . for the @xmath56 scattering length only the results for potential ( a )
are plotted because the @xmath57 potentials are identical for potential cases ( a ) , ( c ) , ( d ) and ( e ) and the results obtained with potential ( b ) differed by less than 5% .
the @xmath56 scattering length has no imaginary component since the penning process is spin - forbidden from the @xmath58 molecular state . of particular interest is the resonance in @xmath56 at a percentage variation of @xmath59 where the short - range potential is made sufficiently shallow that a bound state is removed from the @xmath2 potential . for
percentage variations @xmath60 it is found that @xmath56 is negative in contradiction to recent experimental evidence that @xmath56 is large and positive @xcite . with potentials ( a ) and ( b )
the scattering lengths @xmath61 and @xmath62 were nearly identical and are denoted by a single solid curve .
the scattering lengths associated with the molecular potentials are not observable experimentally , with the exception of @xmath56 , which is approximately equal to the scattering length for the spin - polarized state .
however , these scattering lengths provide unique parameterization of the @xmath0 , @xmath1 and @xmath2 potentials , from which the threshold scattering properties of metastable helium atoms can be obtained . of more practical interest
are the scattering lengths for collisions between atoms in given atomic states in the presence of a magnetic field .
to study collisions in the presence of a magnetic field a full multichannel scattering calculation must be undertaken in which the total hamiltonian describing the two - body collision includes the spin - orbit , zeeman and spin - dipole interactions in addition to the usual radial and rotational kinetic energy operators of the two atoms and the electronic hamiltonian of the quasimolecule formed during the collision .
the details of such a quantum - mechanical multichannel scattering model for metastable helium is described elsewhere @xcite . in brief , we perform the present calculations for atoms in initial atomic states @xmath63 and @xmath64 , including both @xmath65 and @xmath66-waves , and calculate the full non - unitary s - matrix which has elements @xmath67 . here
we let @xmath63 and @xmath64 denote the atomic states @xmath68 , where @xmath69 is the space - fixed projection of the spin @xmath65 for an individual atom . for collision energies up to @xmath70k the contributions of entrance @xmath71 and @xmath66-waves are negligible ( note that due to symmetrization
@xmath71-waves only contribute in collisions between atoms in different atomic states ) , so that only the @xmath65-wave entrance channel @xmath72,l=0 $ ] needs to be considered . the elastic cross section @xmath73 and inelastic rate @xmath74 are then given by @xcite @xmath75 where @xmath76 is the relative atomic velocity . in the wigner threshold region ( @xmath77 )
one can define the scattering lengths using @xmath78 and obtain expressions for the observable scattering lengths @xmath79 by replacing @xmath47 with the matrix element @xmath80 in eq .
( [ leneqn ] ) . the elastic cross sections and inelastic rates
can then be obtained using @xmath81 \nonumber \\
k^{\mathrm{inel}}_{\alpha,\beta } & = & 4 \pi\ , a^{\mathrm{im}}_{\alpha,\beta } /k \label{rateeqnl}.\end{aligned}\ ] ] the inelastic rate @xmath74 includes both contributions from the flux loss due to penning ionization and that due to the atoms exiting in different atomic states . since we calculate the full s - matrix , the contributions of these two processes can be easily separated .
we note that for @xmath82 or @xmath83 collisions , where the total projection of the spin ( @xmath84 ) is non - zero , the collision is dominated by the @xmath2 potential .
this is because parity considerations associate the @xmath1 potential with odd partial waves , and cold collisions are dominated by @xmath65 wave collisions , and the @xmath0 potential can only contribute when @xmath85 .
hence inelastic processes can only occur via the weak relativistic spin - dipole interaction .
the scattering lengths for these states are then almost identical to @xmath56 but with a small imaginary component .
the properties of @xmath86 collisions were investigated in detail in a previous paper @xcite .
the inelastic rates for @xmath87 and @xmath88 collisions , from which ionization can occur directly via strong exchange forces , are much larger and dominate the total inelastic rate for an unpolarized gas .
the @xmath87 and @xmath88 inelastic rates contain two different contributions .
the first is due to exothermic inelastic processes which includes the penning rate @xmath89 and the much smaller collision rate for exothermic fine - structure changing collisions @xmath90 .
the second is the rate for degenerate fine - structure changing collisions @xmath91 .
for example , in ultracold @xmath87 collisions , the entrance channel @xmath92 , l=0 $ ] can decay exothermically to the three channels @xmath93,l=2 $ ] ; @xmath94,l=2 $ ] and the penning channel , and to the two degenerate channels @xmath95,l=0 $ ] and @xmath95,l=2 $ ] .
the flux loss to the degenerate @xmath66-wave exit channels or exothermic @xmath66-wave exit channels ( ie @xmath96 ) only occurs via weak spin - dipole forces and is at least 3 orders of magnitude smaller than that lost to the penning channel or to degenerate @xmath97 exit channels that occurs through strong exchange forces .
importantly , exothermic and degenerate inelastic processes exhibit different threshold properties .
exothermic inelastic rates tend to a constant in the wigner threshold region whereas degenerate inelastic rates fall off as @xmath98 since , as for elastic processes , the incident and final wave number are identical . to represent these separate threshold behaviors in the inelastic rates , we write @xmath99 .
the slope and intercept of @xmath100 vs @xmath101 , for @xmath101 in the wigner threshold region , then gives the degenerate and exothermic scattering lengths @xmath102 and @xmath103 , respectively .
we have calculated these imaginary and the real scattering lengths for all possible collision processes in spin - polarized metastable helium for the five different potentials under investigation . from these calculated scattering lengths
one can use eq .
( [ rateeqnl ] ) to calculate the partial rates or the total rates in an unpolarized gas at temperatures where the wigner threshold approximation is valid .
the scattering lengths are calculated assuming a magnetic field of 10 gauss , however we find only a weak dependence on the magnetic field and results for fields in the range 0 to 20 gauss differ by less than 1% .
the scattering lengths can be used to calculate the rates and cross sections up to typically @xmath104k , except where the scattering lengths become @xmath105 .
scattering lengths for @xmath87 and @xmath88 collisions ( with @xmath65 and @xmath66-waves ) depend on both the @xmath2 and @xmath0 potentials and so their scattering lengths are a function of both the percentage variation of the short - range @xmath2 and @xmath0 potentials for potential cases ( a ) to ( e ) . however , we find that for a fixed percentage variation of the @xmath2 potential , the uncertainty in the scattering lengths induced by varying the short - range @xmath0 potential by @xmath106 is similar to that calculated by fixing the percentage variation in the @xmath0 potential to zero and using the five different potential cases ( a ) to ( e ) . in all instances
the percentage variation in the @xmath2 potential has the largest effect on the scattering lengths and resulting rates .
the @xmath107 or @xmath83 interactions depend only weakly on the singlet potential via the weak relativistic spin - dipole interaction and we find that varying the short - range @xmath0 potential for these collisions produces negligible changes in the scattering length .
therefore we only report scattering lengths as a function of the percentage variation in the @xmath2 potential for potential cases ( a ) to ( e ) , with the understanding that similar uncertainties result in the @xmath87 and @xmath88 scattering lengths by varying the short - range singlet potential .
figures [ a5eim ] , [ ars ] and [ ais ] show the real and imaginary components of the scattering lengths for @xmath108 and @xmath88 collisions .
the real scattering lengths all possess a resonance in the region where a bound state is removed from the @xmath2 potential and @xmath56 goes through @xmath109 .
similar plots exist for @xmath110 and @xmath111 collisions but are almost identical to that shown in fig .
[ a5eim ] for @xmath86 since all are dominated by the @xmath2 potential .
the underlying @xmath2 potentials are identical for potential cases ( a ) , ( c ) , ( d ) and ( e ) and we found that @xmath112 calculated with these potentials differed from those obtained using potential ( b ) by less than 2% .
these small differences are not observable on the scale used in fig .
[ a5eim ] and so for clarity a single solid curve is used to represent @xmath112 for the five potential cases . similarly for @xmath113 the results were identical except for cases ( d ) and ( e ) where different forms of the autoionization widths were used , and so we show only results for ( a ) , ( d ) and ( e ) potential cases .
we note that imaginary scattering lengths for collisions where the total spin projection is non - zero possess no degenerate component and the exothermic contributions are negligible when compared to those for @xmath87 and @xmath88 collisions where penning ionization can occur via exchange forces . for @xmath87 and @xmath88 collisions the values of @xmath114 in fig .
[ ars ] were almost identical for the five potential cases and are represented by a single solid curve for @xmath87 and a dashed curve for @xmath88 . in fig .
[ ais ] we show @xmath115 and @xmath116 for these collisions .
the scattering lengths @xmath115 , which measure @xmath117 , are independent of the percentage variation in the @xmath2 potential and thus @xmath56 , except very near the @xmath56 resonance where the contribution from @xmath90 is no longer negligible and a small increase in @xmath103 is observable .
therefore , the measurement of the ionization signal from trapped metastable helium atoms does not provide information about @xmath56 , the parameter which is required to make predictions of the formation or properties of a bose condensate of spin - polarized metastable helium atoms .
if @xmath90 is neglected then a simple examination of the hamiltonian shows that @xmath118 .
we have verified that this relation is valid to better than 1% and so in fig .
[ ais ] we plot results for @xmath119 for the five potential cases with the understanding that @xmath120 . the curves labeled @xmath116 in fig .
[ ais ] provide the degenerate temperature - dependent inelastic rates for either @xmath121 or @xmath122 .
these equal , exchange - dominated rates strongly mix the @xmath123 , @xmath124 and @xmath125 atoms and are equal to , or larger than , @xmath96 at temperatures greater than @xmath126k or when the quintet potential is near resonance . of the potentials
tested only those with very different exchange terms provided significantly different results and consequently @xmath116 for potentials ( a ) , ( b ) , ( d ) and ( e ) were nearly identical . for convenience
only @xmath127 for potentials ( a ) and ( c ) have been plotted in fig.[ais ] .
the elastic cross section depends on the real and imaginary scattering lengths and its measurement in a spin - polarized or unpolarized gas may provide useful information on @xmath56 . in figs .
[ totalcce][totalin ] we provide the thermally - averaged total elastic cross sections and penning ionization rates for @xmath86 collisions and for an unpolarized gas calculated using potential ( a ) . also shown
are results for @xmath128k and @xmath129k calculated from the scattering lengths using eq.([rateeqnl ] ) . in general the results obtained using eq.([rateeqnl ] ) for temperatures up to @xmath130k are identical to the thermally - averaged results whereas at higher temperatures , outside the wigner regime
, the use of scattering lengths is inappropriate and thermal averaging is required .
the rate equations @xmath131 define our partial rates for , respectively , identical and non - identical atom collisions , where @xmath132 is the number of colliding atoms in state @xmath133 and the superscript ` @xmath134 ' denotes ` @xmath135 ' , ` @xmath136 ' or ` @xmath137 ' .
the total thermally - averaged penning rates and cross sections for an unpolarized gas are obtained assuming an equal population of the @xmath138 magnetic substates so that @xmath139 and hence @xmath140 where @xmath141 denotes the thermal average . in this case
the assumption that the magnetic substates are evenly populated in an unpolarized gas is well justified on collisional grounds . at temperatures above
@xmath129k the degenerate rates @xmath142 evenly mix @xmath123 , @xmath124 and @xmath125 atoms . at lower temperatures the penning rates @xmath143 and @xmath144 , which dominate the exothermic inelastic rates , deplete the three different atomic populations @xmath145 equally since @xmath146 and the collision of @xmath87 results in the loss of two @xmath124 atoms . here
we have neglected the small contribution from spin - dipole processes , that is @xmath147 , and assume that any initial asymmetry in the populations @xmath145 due to preparation of the atoms in a light field for instance is small or has become small once the measurement of the collisional rate in the absence of light is performed .
the thermally - averaged results were calculated by averaging over a maxwell - boltzmann distribution of atomic velocity using 71 velocity nodes which correspond to collision energies in the range @xmath148k to @xmath149k .
since the results are for the case ( a ) potentials , with the percentage variation in the singlet potential set to zero , we estimate from the uncertainties in the scattering lengths that the errors in the elastic cross sections and total inelastic rates are of the order of 10% and 40% respectively .
the penning rates possess a larger uncertainty to account for the percentage variation of the @xmath150 potential whereas the unpolarized elastic rates , which are dominated by the real scattering lengths belonging to collisions with @xmath151 or 1 , are controlled only by @xmath152 . in an unpolarized gas
the @xmath71-waves can contribute in @xmath88 , @xmath153 and @xmath83 collisions .
these contributions to the total thermally - averaged penning rates were found to be negligible at 1 @xmath43k .
however the @xmath71-wave contributions increased the total penning rate ( compared to that obtained using only @xmath65-waves ) by approximately 7% at 500 @xmath43k and 12% at 1 mk . the @xmath71-waves modified the total elastic cross sections by less than 1% at all temperatures . for @xmath154 collisions ( and
similarly for @xmath155 or @xmath83 ) we observe a resonance in the inelastic rates at a percentage variation of + 1.875 due to the resonant enhancement of the exothermic rates .
we find that @xmath156 , indicating that most but not all of the flux leaving the @xmath157,l=0 $ ] entrance channel is subsequently lost through ionization .
these rates are much smaller than those from the @xmath87 and @xmath88 collisions and the total contribution to @xmath89 from @xmath82 and @xmath83 collisions is only observable in fig .
[ totalin ] as a small peak at + 1.875 in the unpolarized ionization rate .
the total elastic cross sections of an unpolarized or a polarized gas show strong dependences on the form of the @xmath2 potential and provide possible measures of @xmath56 .
the scattering lengths associated with the three molecular potentials relevant to collisions of metastable helium atoms have been reported .
the uncertainties in the molecular potentials and autoionization widths have been considered and probable ranges of values given for the scattering lengths for each molecular state . scattering lengths for collisions involving the various atomic states
have also been calculated and related to the elastic cross sections and inelastic collision rates for temperatures in the wigner threshold region , with the aim of providing a correspondence with experimentally measurable quantities .
in particular , it has been shown that measurement of the total elastic cross section in a polarized or unpolarized gas should provide a means of experimentally determining the @xmath56 scattering length , which is of importance in the attainment of a bose - einstein condensate in a gas of spin - polarized metastable helium atoms . in fig .
[ totalin ] we compare the total penning rates for an unpolarized gas calculated here with those from experiment . not shown are the theoretical uncertainties of @xmath158 40% which arise from uncertainties in the molecular potentials and in the form of the ionization widths
. the total elastic cross sections and penning rates are consistent with those reported in ref . where sightly different molecular potentials and ionization widths were used .
the experimental results possess uncertainties on the order of 50% which are not shown in fig .
[ totalin ] .
the experimental results correspond to the case of zero magnetic field whereas the theoretical predictions are made for @xmath159 g. however the scattering lengths were found to vary by less than 1% over the range 020 g which is negligible when compared to these uncertainties that arise from the form of the ionization width .
the comparison between theoretical and experimental data is satisfactory given these uncertainties , however the experimental results are consistently higher than the theoretical predictions .
prior to the submission of this paper , no experimental results or theoretical predictions existed for the scattering lengths , cross sections and rates calculated here for incident atoms in specific states .
the @xmath86 spin - polarized system has been investigated by shlyapnikov _
_ @xcite and @xcite but no ( quantitative ) @xmath160 scattering lengths were reported . however , during the review of this paper , two measurements of scattering length have been announced ; @xmath161 by @xcite and @xmath162 by @xcite .
these measurements , together with measured suppression by a factor of @xmath163 for the penning ionization rate for a spin polarized gas compared to that of an unpolarized gas , are consistent with the current predictions . finally ,
using the scattering lengths reported in this investigation , one can estimate the scattering lengths for the other isotopes of helium by mass scaling the vibrational defect .
this is related to the scattering length by @xcite @xmath164 , \ ] ] where @xmath165 is the vibrational defect and @xmath166 is defined by the leading term @xmath167 in the long - range potential ( @xmath168 for he ) .
the term @xmath169 is an asymptotic property which depends only on the long - range potential and can be approximated by @xmath170 for he @xcite .
to mass scale the vibrational defect we first calculate @xmath165 for @xmath171he for a given potential .
since the trigonometric function is periodic , this only gives the fractional part of the vibrational defect and one must include the multiple of @xmath172 where @xmath173 is the number of bound states supported by that potential , ie @xmath174 .
this vibrational defect can then be scaled using @xmath175 , to determine the vibrational defect for isotope @xmath176 . here
@xmath177 and @xmath178 are the reduced masses of @xmath179 and @xmath180 respectively .
vv acknowledges partial support from the engineering and physical sciences research council .
the institute for theoretical atomic and molecular physics is supported by a grant from the nsf to the harvard college observatory and the smithsonian astrophysical observatory .
c. westbrook _
tenicols 91 , papers presented at tenth international conference on laser spectroscopy , france , june 1991 _ , edited by e. g. m. ducloy and g. camy ( world scientific , singapore , 1991 ) , pp . | we report scattering lengths for the @xmath0 , @xmath1 and @xmath2 adiabatic molecular potentials relevant to collisions of two metastable @xmath3 helium atoms as a function of the uncertainty in these potentials .
these scattering lengths are used to calculate experimentally observable scattering lengths , elastic cross sections and inelastic rates for any combination of states of the colliding atoms , at temperatures where the wigner threshold approximation is valid .
# 1#2 # 1#2#3 # 1#2#3 # 1#2 # 1#2#3 # 1#2 |
for the theoretical calculations , we used a large scale atomistic tight binding approach with spin resolved sp@xmath77d@xmath78s * atomic orbitals with nearest neighbor interactions@xcite .
typical simulation domains comprise of 1.5 - 2 million atoms to capture realistic sized dots .
spin - orbit interactions are directly included in the hamiltonian as a matrix element between p - orbitals following the prescription of chadi @xcite .
the advantage of this approach is that no additional fitting parameters are needed to capture various types of soi such as rashba and dresselhaus soi in contrast to k.p theory .
we introduce monoatomic steps as a source of non - ideality consistent with other works@xcite .
the si interface was modeled with hydrogen passivation@xcite , without using sige .
this interface model is sufficient to capture the soi effects of a si / sige interface discussed in refs .
- @xcite .
we have used the methodology of ref .
- @xcite to model the micro - magnetic fields [ supplementary section s4 ] .
full magnetization of the micro - magnet is assumed . this causes the value of the magnetization of the micro - magnet to be saturated and makes it independent of @xmath8 . however , a change in the direction of @xmath8 changes the magnetization .
we include the effect of inhomogeneous magnetic field perturbatively , with the perturbation matrix elements , @xmath79 . here ,
@xmath80 and @xmath81 are atomistic wave - functions calculated with homogeneous magnetic field . for further details about the numerical techniques ,
see nemo3d references - @xcite .
method details about the experiment can be found in ref .
. 100 datta , s. & das , b. electronic analog of the electro - optic modulator .
lett . _ * 56 * , 665(r ) ( 1990 ) .
wolf , s. _ et al . _
spintronics : a spin based electronics vision for the future .
_ science _ * 294 * , 1488 - 1495 ( 2001 ) .
i. uti , fabian , j. & das sarma , s. spintronics : fundamentals and applications .
phys . _ * 76 * , 323 ( 2004 ) .
loss , d. & vincenzo , d. p. quantum computation with quantum dots .
a _ * 57 * , 120 ( 1998 ) .
petta , j. _ et al .
_ coherent manipulation of coupled electron spins in semiconductor quantum dots . _
science _ * 309 * , 2180 - 2184 ( 2005 ) .
koppens , f. _ et al .
_ driven coherent oscillations of a single electron spin in a quantum dot .
_ nature _ * 442 * , 766 - 771 ( 2006 ) .
hanson , r. , kouwenhoven , l. p. , petta , j. r. , tarucha , s. , vandersypen & l. m. k. spins in few electron quantum dots . _ rev .
_ * 79 * , 1217 ( 2007 ) .
xiao , d. , liu , g .- b . , feng , w. , xu , x. & yao , w. coupled spin and valley physics in monolayers of mos@xmath57 and other group vi dichalcogenides .
lett . _ * 108 * , 196802 ( 2012 ) .
gong , z. _ et al .
_ magnetoelectric effects and valley - controlled spin quantum gates in transition metal dichalcogenide bilayers .
_ nature communications _ * 4 * , 2053 ( 2013 ) .
xu , x. , yao , w. , xiao , d. & heinz , t. spins and pseudospins in layered transition metal dichalcogenides
. _ nature physics _ * 10 * , 343350 ( 2014 ) .
renard , v. t. _ et al .
_ valley polarization assisted spin polarization in two dimensions .
_ nature communications _ * 6 * , 7230 ( 2015 ) yang , c. _ et al . _
spin - valley lifetimes in a silicon quantum dot with tunable valley splitting .
_ nature communications _ * 4 * , 2069 ( 2013 ) .
hao , x. , ruskov , r. , xiao , m. , tahan , c. & jiang , h. electron spin resonance and spin - valley physics in a silicon double quantum dot .
_ nature communications _ * 5 * , 3860 ( 2014 ) .
kawakami , e. _ et al .
_ electrical control of a long - lived spin qubit in a si / sige quantum dot .
_ nature nanotechnology _ * 9 * , 666670 ( 2014 ) .
veldhorst , m. _ et al . _ an addressable quantum dot qubit with fault tolerant control fidelity .
_ nature nanotechnology _ * 9 * , 981985 ( 2014 ) .
maune , b. m. _ et al .
_ coherent singlet - triplet oscillations in a silicon - based double quantum dot .
_ nature _ * 481 * , 344347 ( 2011 ) .
wu , x. _ et al .
_ two - axis control of a singlet - triplet qubit with integrated micromagnet .
usa _ * 111 * , 11938 ( 2014 ) .
eng , k. _ et al .
_ isotopically enhanced triple quantum dot qubit .
_ science advances _ * 1 * , no .
4 , e1500214 ( 2015 ) .
kim , d. _ et al .
quantum control and process tomography of a semiconductor quantum dot hybrid qubit .
_ nature _ * 511 * , 70 - 74 ( 2014 ) .
kim , d. _ et al .
_ high fidelity resonant gating of a silicon - based quantum dot hybrid qubit .
_ npj quantum information _ * 1 * 15004 ( 2015 ) .
culcer , d. , sariava , a. l. , koiller , b. , hu , x. & das sarma , s. valley - based noise resistant quantum computation using si quantum dots .
lett . _ * 108 * , 126804 ( 2012 ) . rohling , n. & burkard , g. universal quantum computing with spin and valley states .
_ new j. phys .
_ * 14 * , 083008 ( 2012 ) .
rohling , n. , russ , m. & burkard , g. hybrid spin and valley quantum computing with singlet - triplet qubits .
* 113 * , 176801 ( 2014 ) .
veldhorst , m. _ et al .
_ spin - orbit coupling and operation of multivalley spin qubits .
b _ * 92 * , 201401(r ) ( 2015 ) .
scarlino , p. _ et al .
_ second - harmonic coherent driving of a spin qubit in sige / si quantum dot .
lett . _ * 115 * , 106802 ( 2015 ) .
scarlino , p. _ et al .
_ , dressed photon - orbital states in a quantum dot : inter - valley spin resonance , _ arxiv : _ 1608.06538 ( 2016 ) .
tokura , y. , van der wiel , w. g. , obata , t. & tarucha , s. coherent single electron spin control in a slanting zeeman field .
_ * 96 * , 047202 ( 2006 ) .
goldman , j. r. , ladd , t. d. , yamaguchi , f. & yamamoto , y. magnet designs for a crystal lattice quantum computer .
phys . a _ * 71 * , 11 ( 2000 ) .
golub , l. e. , & ivchenko , e. l. spin splitting in symmetrical sige quantum wells .
b _ * 69 * , 115333 ( 2004 ) nestoklon , m. o. , golub , l. e. & ivchenko , e. l. spin and valley - orbit splittings in sige / si heterostructures . _ phys . rev . b _ * 73 * , 235334 ( 2006 ) .
nestoklon , m. o. , ivchenko , e. l. , jancu , j. -m . & voisin , p. electric field effect on electron spin splitting in sige / si quantum wells .
b _ * 77 , * 155328 ( 2008 ) .
zandvliet , h. j. w. & elswijk , h. b. morphology of monatomic step edges on vicinal si(001 ) .
b _ * 48 , * 14269 ( 1993 )
. gamble , j. k. , eriksson , m. a. , coppersmith , s. n. & friesen , m. disorder - induced valley - orbit hybrid states in si quantum dots .
b _ * 88 , * 035310 ( 2013 ) . friesen , m. , & coppersmith , s. n. theory of valley - orbit coupling in a si / sige quantum dot .
b _ * 81 , * 115324 ( 2010 ) . friesen , m. , eriksson , m. a. & coppersmith , s. n. magnetic field dependence of valley splitting in realistic si / sige quantum wells , _ appl
lett . _ * 89 * , 202106 ( 2006 ) .
klimeck , g. _ et al .
_ atomistic simulation of realistically sized nanodevices using nemo 3d : part imodels and benchmarks .
_ ieee trans .
electron dev .
_ * 54 , * 20792089 ( 2007 ) .
klimeck , g. , oyafuso , f. , boykin , t. , bowen , r. , & allmen , p. development of a nanoelectronic 3d ( nemo-3d ) simulator for multimillion atom simulations and its application to alloyed quantum dots , _ computer modeling in engineering and sciences ( cmes ) _ * 3 * , no . 5 , 601 - 642 ( 2002 ) .
chadi , d. j. spin - orbit splitting in crystalline and compositionally disordered semiconductors .
_ phys . rev .
b _ * 16 , * 790 ( 1977 ) .
kharche , n. , prada , m. , boykin , t. b. & klimeck , g. valley splitting in strained silicon quantum wells modeled with 2 miscuts , step disorder , and alloy disorder .
_ * 90 , * 092109 ( 2007 ) .
lee , s. , oyafuso , f. , allmen , p. v. & klimeck , g. boundary conditions for the electronic structure of finite - extent embedded semiconductor nanostructures .
b _ * 69 , * 045316 ( 2004 ) .
this work was supported in part by aro ( w911nf-12 - 0607 ) ; development and maintenance of the growth facilities used for fabricating samples is supported by doe ( de - fg02 - 03er46028 ) .
this research utilized nsf - supported shared facilities at the university of wisconsin - madison .
computational resources on nanohub.org , funded by the nsf grant eec-0228390 , were used .
m.p.n . acknowledges support from erc synergy grant .
r.f . and r.r .
acknowledge discussions with r. ruskov , c. tahan , and a. dzurak .
r.f . performed the g - factor calculations , explained the underlying physics and developed the theory with guidance from r.r .
r.f . , r.r . ,
, p.s . , and m.p.n .
analyzed the simulation results and compared with experimental data in consultation with l.m.k.v .
, s.n.c . and
m.a.e .. e.k . and p.s .
performed the experiment and analyzed the measured data .
d.r.w . fabricated the sample .
d.e.s . and m.g.l .
grew the heterostructure .
r.f . and r.r .
wrote the manuscript with feedback from all the authors .
r.r . and l.m.k.v .
initiated the project , and supervised the work with s.n.c , m.f . and m.a.e .
the authors declare that they have no competing financial interests .
* s1 . analytic effective mass model to explain the effect of spin - orbit interaction ( soi ) on the valley dependent g - factor anisotropy in a silicon quantum dot * to explain our atomistic tight - binding ( tb ) results of the valley dependent g - factor and its anisotropy in a silicon quantum dot ( qd ) , we have also performed analytic effective mass calculations as shown here .
the electron hamiltonian can be written as , here , @xmath83 is the electron effective mass , which assumes a value of @xmath84 and @xmath85 for the transverse ( x or [ 100 ] , y or [ 010 ] ) and longitudinal ( z or [ 001 ] ) components in si respectively ( @xmath86 being the free electron mass ) .
@xmath87 is the potential defining the quantum dot , and has been discussed in the main text .
@xmath88 is the zeeman term , with @xmath17 the electron @xmath17-factor , @xmath18 the bohr magneton , and @xmath89 the applied magnetic field at the dot location .
@xmath90 is the soi term , where @xmath59 ( @xmath91 ) is the strength of dresselhaus ( rashba ) interaction , @xmath92 , @xmath93 are the pauli spin matrices , and @xmath94 ( @xmath95 ) is electron canonical momentum along x ( y ) direction .
@xmath96 , where @xmath97 is the vector potential and @xmath98 .
now , we treat @xmath99 as a perturbation to @xmath100 .
the unperturbed hamiltonian @xmath101 yields the spin degenerate eigenstates of a si qd .
typically , in these devices , orbital splitting ( @xmath102 ) is much larger than valley ( @xmath103 ) and spin ( @xmath104 ) splittings .
so , we only consider the four lowest energy states @xmath105 , @xmath106 , @xmath107 , @xmath108 as the basis for the perturbation calculation
. we can write the perturbation hamiltonian as , @xmath109 since @xmath115 .
for a magnetic field in the x - y plane , we assume , @xmath116 , @xmath117 and @xmath118 , where @xmath119 and @xmath120 are the x and y components of the magnetic field respectively . then , here , @xmath40 is the dipole moment along [ 001 ] crystal direction .
we assumed @xmath122 .
as the electrons in the @xmath0 and @xmath1 valley states have only @xmath123 valley component , we can replace @xmath17 here with @xmath124 , the g - factor perpendicular to the valley axis@xcite .
so , for an in - plane magnetic field , in the presence of only external magnetic field , @xmath5 , @xmath126 and @xmath127 , with @xmath43 being the b - field angle relative to the [ 100 ] crystal orientation for a counter clockwise rotation of @xmath8 .
however , to be consistent with the angle @xmath15 defined in the experiment ( i.e. an angle relative to the [ 110 ] crystal orientaiton with clockwise rotation of @xmath5 ) , we use the relation @xmath128 .
now , using the definition of @xmath119 and @xmath120 from above , we can obtain the angular dependence of the spin splitting for different valley states , so , from this equation we can see that , * without the dresselhaus contribution , there is no angular dependence or anisotropy in the spin splitting ( or g - factor ) for the different valley states*. now , for @xmath5 along the @xmath133 $ ] and @xmath134 $ ] crystal orientations , we get , @xmath135}-{{g}_{\bot } } \mu { b_{\textup{ext}}}=2\left ( { { \beta } _ { \pm } } -{{\alpha } _ { \pm } } \right)\left| e \right|\frac{\left\langle z \right\rangle } { \hbar } b_{\textup{ext}}\\ & e_{{\textup{zs}}(\pm ) } ^{[1\overline{1}0]}-{{g}_{\bot } } \mu { b_{\textup{ext}}}=2\left ( -{{\beta } _ { \pm } } -{{\alpha } _ { \pm } } \right)\left| e \right|\frac{\left\langle z \right\rangle } { \hbar } { b_{\textup{ext } } } \label{equ10}\end{aligned}\ ] ] @xmath136}+e_{{\textup{zs}}(\pm ) } ^{[1\bar{1}0]}-2{{g}_{\bot } } \mu { b_{\textup{ext } } } \right)}{4\left| e \right|\frac{\left\langle z \right\rangle } { \hbar } { b_{\textup{ext } } } } \\ & { { \beta } _ { \pm } } = \frac{e_{{\textup{zs}}(\pm ) } ^{[1\bar{1}0]}-e_{{\textup{zs}}(\pm ) } ^{[1\bar{1}0]}}{4\left| e \right|\frac{\left\langle z \right\rangle } { \hbar } { b_{\textup{ext } } } } \label{equ11}\end{aligned}\ ] ] for @xmath137800mt
. the angle of @xmath5 is defined clockwise with respect to the @xmath133 $ ] crystal orientation ) * b , * magnetic field dependence of @xmath4 for @xmath138 along @xmath133 $ ] and @xmath134 $ ] crystal orientations .
the tight - binding calculations correspond to the ideal ( interface ) case of figs .
3e and 3f of the main text . ] in fig .
s1 , we compared @xmath4 calculated from this analytic model with the atomistic tight - binding results for the ideal interface case , shown in figs .
3e and 3f of the main text .
both the anisotropic behavior of @xmath4 with respect to the angle of @xmath5 ( fig .
s1a ) , and the dependence on the magnitude @xmath139 along @xmath133 $ ] and @xmath134 $ ] crystal orientations ( fig .
s1b ) show perfect agreement between analytic and atomistic calculations . from the atomistic calculations of fig .
3a and equation ( s14 ) and ( s15 ) , we extracted @xmath140ev@xmath47 m , @xmath141ev@xmath47 m , @xmath142ev@xmath47 m , @xmath143ev@xmath47 m for @xmath124=1.9937 and @xmath144 nm , calculated using @xmath145 , where , @xmath146 @xcite . here , @xmath147 mvm@xmath32 is the vertical electric field . to include the homogeneous fields from the micro - magnets , in equation ( s3 ) , we use , @xmath148 , @xmath149 and @xmath150 . as , @xmath151 and @xmath152
, we can still use equation ( s6 ) and ( s7 ) for @xmath153 and @xmath154 . to develop an analytic model to capture the effect of the inhomogeneous magnetic field on the spin - splittings for different valley states
, we assume the hamiltonian in equation ( s1 ) as our unperturbed hamiltonian and include the magnetic field gradient as a perturbation , here , @xmath156 are the magnetic field gradients along different directions , where @xmath157 correspond to x , y , z co - ordinates . now
the lowest 4 eigen - states of the unperturbed hamiltonian are @xmath158 , @xmath159 , @xmath160 and @xmath161 .
the @xmath162 order correction due to @xmath163 to @xmath158 is given by , to simplify our analytic calculation , we assume that the spin mixing due to soi is very small and we can separate the spin part from the spatial part of the wavefunction . using this approximation in [ equ13
] , we get , here , @xmath166 is the dipole moment along the @xmath167 direction , @xmath168 . now , for spins in an in - plane magnetic field , @xmath169 , @xmath170 and @xmath171 .
the external magnetic field fully magnetizes the micro - magnets .
so , the inhomogeneous magnetic field from the micro - magnets ( @xmath172 ) depends on the direction ( @xmath43 ) of the external magnetic field . using these relations in equation ( s19 ) we get , for an electron in an ideal ( smooth ) interface , @xmath181 .
but the presence of interface steps in realistic devices makes @xmath182 and/or @xmath183 non zero .
we can ignore the 2nd order corrections because @xmath184 .
due to interface steps , @xmath185 and/or @xmath186 , but @xmath187 , for the electric field used in the experiment .
so , in the presence of interface steps , the magnetic field gradient adds to the difference in esr frequencies between the valley states , and from equation ( s21 ) , we get equation ( 3 ) of the main text . :
comparison between analytic effective mass and atomistic tight - binding calculations . for the latter , the contribution of the soi
is subtracted to obtain the change in @xmath4 due to @xmath26 only from fig .
1c of the main text ( difference between green dashed line and black solid line ) . ] in fig .
s2 , we compared @xmath188 calculated using equation ( 2 ) of the main text with atomistic tight - binding calculations .
we used @xmath189 nm and @xmath190 in equation ( 2 ) .
the atomistic results shown in fig .
s2 is the difference between the black curve ( circular marker ) and the green curve ( square markers ) shown in fig .
the analytic calculation qualitatively captures the tight - binding results .
the little mismatch between the analytic and atomistic calculation is due to ignoring the spin mixing from soi in our analytic model . the vertical electric field
, @xmath12 affects @xmath4 , because it affects the soi parameters@xcite and also the dipole moment parameters .
the effect of @xmath12 on @xmath4 due to the soi and gradient magnetic field , from atomistic tight - binding calculations , are shown in supplementary figs .
s3 and s4 respectively . here
@xmath12 only changes the magnitude of @xmath4 , but not its sign .
@xmath12 also effect the @xmath8 dependence of @xmath4 through soi , as shown in supplementary fig .
s3b . here also @xmath12 only changes the magnitude of the slope , @xmath191 but not its sign . on the change in @xmath4 due to soi . *
a , * increase in @xmath192 with increasing @xmath12 for @xmath8=0.8 t. * b , * change in @xmath193 with changing @xmath8 for different @xmath12 .
@xmath12 changes the magnitude of the slope , @xmath194 .
bottom panels in both figs .
a and b corresponds to @xmath5 along @xmath133 $ ] ( @xmath195 ) and in top panels @xmath5 is along @xmath134 $ ] ( @xmath21 ) .
interface condition and parabolic confinement for the dot used in these simulations are the same as that used to match experimental data in figs . 1 and 2 of main text .
] on the change in @xmath4 due to the gradient magnetic field @xmath26 , for @xmath5 along @xmath133 $ ] @xmath195 ( bottom panel ) and @xmath134 $ ] ( @xmath21 ) ( top panel ) .
interface condition and parabolic confinement for the dot used in these simulations are the same as that used to match experimental data in figs . 1 and 2 of the main text .
increasing @xmath12 increases @xmath196 . ]
we calculated the local magnetic field created by the micro - magnets when the external magnetic field is applied along the y@xmath197 ( [ 110 ] ) ( @xmath28 ) or along the x@xmath197 ( [ 1@xmath1980 ] ) axis ( @xmath29 ) of the device picture in fig .
s5 , assuming that the micro - magnets are fully magnetized @xcite .
the shape of the micro - magnets is shown in ref .
- @xcite .
[ fig : simu : bextz ] and fig .
[ fig : simu : bextx ] show the results of the numerical calculation of the total magnetic field gradient when the external magnetic field is applied along the y@xmath197 axis and along the x@xmath197 axis respectively , and the pink circle shows the estimated dot position .
we calculated the stray magnetic field created by the micro - magnets when the external magnetic field is applied between the y@xmath197 axis and the x@xmath197 axis ( @xmath199 ) according to the following approximation , we ignore @xmath202 , with @xmath201x@xmath197,y@xmath197,z . for small electric field like @xmath12=6.77 mvm@xmath32 used to match the experiment ,
the electron wavefunctions of the two valley states have similar spread along z , so @xmath203 .
thus the effects of @xmath202 on @xmath204 should be small .
however , the effects of these @xmath202 on @xmath205 or @xmath206 can be larger , but as shown in the figs .
1d and 2b of the main text , the effect of the gradient magnetic field on @xmath52 is negligible compared to that of the homogeneous magnetic fields . ,
y@xmath197 and z components of the magnetic field and their gradients along the x@xmath197 and y@xmath197 axes induced by the micro - magnets when the external magnetic field is applied along the y@xmath197 axis in the plane of the si quantum well , for fully magnetized micro - magnets .
the black solid lines indicate the edges of the micro - magnet as simulated . ] on the total homogeneous magnetic field .
* a , * anisotropic total magnetic field , due to the presence of the homogeneous magnetic field from the micro - magnet , @xmath213 .
@xmath213 adds vectorially to @xmath5 .
also the magnetization of the micro - magnet depends on the direction of @xmath5 or @xmath15 .
so @xmath213 itself is anisotropic , hence the superscript @xmath15 .
* b , * total magnetic field ( shown in green dashed line with square markers ) along along @xmath133 $ ] ( @xmath195 ) ( bottom panel ) and @xmath134 $ ] ( @xmath21 ) ( top panel ) with changing @xmath8 .
the black line represents @xmath214 . ]
veldhorst , m. _ et al .
_ spin - orbit coupling and operation of multivalley spin qubits .
rev . b _ * 92 * , 201401(r ) ( 2015 ) .
nestoklon , m. o. , ivchenko , e. l. , jancu , j. -m . , & voisin , p. electric field effect on electron spin splitting in sige / si quantum wells .
b _ * 77 , * 155328 ( 2008 ) .
nestoklon , m. o. , golub , l. e. , & ivchenko , e. l. spin and valley - orbit splittings in sige / si heterostructures .
b _ * 73 * , 235334 ( 2006 ) . roth , l. g factor and donor spin - lattice relaxation for electrons in germanium and silicon .
rev . _ * 118 , * 1534 ( 1960 ) goldman , j. r. , ladd , t. d. , yamaguchi , f. , & yamamoto , y. magnet designs for a crystal lattice quantum computer .
a _ * 71 * , 11 ( 2000 ) .
pioro - ladrire , m. , tokura , y. , obata , t. , kubo , t. & tarucha , s. micromagnets for coherent control of spin - charge qubit in lateral quantum dots .
_ applied physics letters _ * 90 * , 024105 ( 2007 ) .
kawakami , e. _ et al .
_ electrical control of a long - lived spin qubit in a si / sige quantum dot .
_ nature nanotechnology _ * 9 * , 666670 ( 2014 ) . | * spin qubits hosted in silicon ( si ) quantum dots ( qd ) are attractive due to their exceptionally long coherence times and compatibility with the silicon transistor platform . to achieve electrical control of spins for qubit scalability
, recent experiments have utilized gradient magnetic fields from integrated micro - magnets to produce an extrinsic coupling between spin and charge , thereby electrically driving electron spin resonance ( esr ) .
however , spins in silicon qds experience a complex interplay between spin , charge , and valley degrees of freedom , influenced by the atomic scale details of the confining interface . here , we report experimental observation of a valley dependent anisotropic spin splitting in a si qd with an integrated micro - magnet and an external magnetic field .
we show by atomistic calculations that the spin - orbit interaction ( soi ) , which is often ignored in bulk silicon , plays a major role in the measured anisotropy .
moreover , inhomogeneities such as interface steps strongly affect the spin splittings and their valley dependence .
this atomic - scale understanding of the intrinsic and extrinsic factors controlling the valley dependent spin properties is a key requirement for successful manipulation of quantum information in si qds . *
[ fi1 ] how microscopic electronic spins in solids are affected by the crystal and interfacial symmetries has been a topic of great interest over the past few decades and has found potential applications in spin - based electronics and computation @xcite . while the coupling between spin and orbital degrees of freedom has been extensively studied ,
the interplay between spin and the momentum space valley degree of freedom is a topic of recent interest .
this spin - valley interaction is observed both in the exotic class of newly found two - dimensional materials @xcite and in silicon @xcite the old friend of the electronics industry .
progress in silicon qubits in the last few years has come with the demonstrations of various types of qubits with exceptionally long coherence times , such as single spin up / down qubits @xcite , two - electron singlet - triplet qubits @xcite , three - electron exchange - only @xcite and hybrid spin - charge qubits @xcite realized in si qds .
the presence of the valley degree of freedom has also enabled valley based qubit proposals @xcite , which have potential for noise immunity . to harness the advantages of different qubit schemes , quantum gates for information encoded in different bases are required@xcite .
a controlled coherent interaction between multiple degrees of freedom , like valley and spin , might offer a building block for promising hybrid systems .
although bulk silicon has six - fold degenerate conduction band minima , in quantum wells or dots , electric fields and often in - plane strain in addition to vertical confinement results in only two low lying valley states ( labeled as @xmath0 and @xmath1 in figure 1b ) split by an energy gap known as the valley splitting . an interesting interplay between spin and valley degrees of freedom , which gives rise to a valley dependent spin splitting , has been observed in recent experiments@xcite .
soi enables the control of spin resonance frequencies by gate voltage , an effect measured in refs .
- @xcite .
however , the esr frequencies and their stark shifts were found to be different for the two valley states @xcite . in another work , an inhomogeneous magnetic field , created by integrated micro - magnets in a si / sige quantum dot device ,
was used to electrically drive esr@xcite .
magnetic field gradients generated in this way act as an extrinsic spin - orbit coupling and thus can affect the esr frequency@xcite .
remarkably , although soi is a fundamental effect arising from the crystalline structure , the esr frequency differences between the valley states observed in refs . - @xcite and - @xcite have different signs when the external fields are oriented in the same direction with respect to the crystal axes . to understand and achieve control over the coupled behavior between spin and valley degrees of freedom , several questions need to be addressed , such as 1 ) what causes the device - to - device variability ?
, 2 ) can an artificial source of interaction , like inhomogeneous b - field , completely overpower the soi effects of the intrinsic material ? ,
3 ) what knobs and device designs can be utilized to engineer the valley dependent spin splittings ? here we report experimentally measured anisotropy in the esr frequencies of the valley states @xmath2 and @xmath3 and their differences @xmath4 , as a function of the direction of the external magnetic field ( @xmath5 ) in a quantum dot formed at a si / sige heterostructure with integrated micro - magnets . at specific angles of the external b - field , we also measure the spin splittings of the two valley states as a function of the b - field magnitude . by performing spin - resolved atomistic tight binding ( tb ) calculations of the quantum dots confined at ideal versus non - ideal interfaces ,
we evaluate the contribution of the intrinsic soi with and without the spatially varying b - fields from the micro - magnets to the spin splittings , thereby relating these quantities to the microscopic nature of the interface and elucidating how spin , orbital and valley degrees of freedom are intertwined in these devices .
finally , by combining all the effects together , we explain the experimental measurements and address the questions raised in the previous paragraph . fig .
1 shows the experimental device , energy levels of interest and measured anisotropic spin splittings compared with the final theoretical results .
details of the device shown in fig .
1a can be found in ref .
- @xcite .
the external magnetic field is swept from the @xmath6}$ ] to @xmath7}$ ] crystal orientation .
a schematic of the energy level structure is shown in fig .
1b depicting the @xmath0 and @xmath1 valley states with different spin splittings , where @xmath0 is defined as the ground state . in the experiment ,
the lowest valley - orbit excitation is well below the next excitation , justifying this four - level schematic in the energy range of interest . ) and their differences for the two valley states as a function of the external b - field magnitude @xmath8 along two crystal directions , and comparison with theoretical calculations . *
* a , * @xmath4 and * b , * @xmath2 with @xmath9 along @xmath6 } \left(\theta=0^{\circ}\right)$ ] ( bottom panel ) and @xmath7 } \left(\theta=90^{\circ}\right)$ ] ( top panel ) . as in figs .
1c and 1d , the calculations progressively include soi ( labeled ` @xmath5 ( tb ) ' ) , homogeneous ( labeled ` @xmath10 ( tb ) ' ) , and gradient ( labeled ` @xmath11 ( tb ) ' ) b - field of the micro - magnet .
the cyan solid lines are the effective mass calculations and the red circles are the experimental data .
the dependence ( slope ) of @xmath4 on @xmath8 in ( a ) comes from the soi , while the micro - magnetic fields provide a shift independent of @xmath8 . ] [ fi2 ] [ fi3 ] the atomistic calculation with soi alone ( labeled ` @xmath5 ( tb ) ' ) for a qd at a specifically chosen non - ideal interface and vertical electric field ( @xmath12 ) qualitatively captures the experimental trend of @xmath4 in fig . 1c , but fails to reproduce the anisotropy of the measured @xmath2 in fig .
1d in the larger ghz scale .
the differences between the experimental data and the soi - only calculations in both figures arise from the micro - magnets present in the experiment .
we can separate the contribution from the micro - magnet into two parts , a homogeneous ( spatial average , @xmath13 ) and an inhomogeneous ( spatially varying , @xmath14 ) magnetic field .
the superscript @xmath15 here indicates that the micro - magnetic fields depend on the direction of @xmath5 ( supplementary section s4 ) .
the inclusion of the homogeneous part of the micro - magnetic field creates an anisotropy in the total magnetic field ( supplementary fig .
s7 ) , which captures the anisotropy of @xmath2 in fig .
1d very well ( @xmath16 , where @xmath17 is the land g - factor , @xmath18 is the bohr magneton and @xmath19 is the planck constant ) , but quantitative match with the experimental data in fig .
1c is not obtained .
next , we also incorporate the inhomogeneous part of the micro - magnetic field , and witness a close quantitative agreement in the anisotropy of @xmath4 , while the anisotropy of @xmath2 is unaffected .
this experiment - theory agreement of fig .
1c is achieved for a specific choice of interface condition and @xmath12 , whose influence will be discussed later . here , we conclude that mainly the intrinsic soi and the extrinsic inhomogeneous b - field govern the anisotropy of @xmath4 on the mhz scale , while the anisotropy in the total homogeneous magnetic field introduced by the micro - magnet dictates the anisotropy of @xmath2 ( and @xmath3 ) on the larger ghz scale . in fig . 2
, we show the measurements of the spin splittings as a function of the magnitude of @xmath5 ( @xmath8 ) , together with the theoretical calculations .
the bottom panels show @xmath4 ( fig .
2a ) and @xmath2 ( fig .
2b ) for @xmath5 along @xmath6}$ ] ( @xmath20 ) , whereas the top panels correspond to the b - field along @xmath7}$ ] ( @xmath21 ) . in fig .
2b , @xmath2 depends on @xmath8 through @xmath22 , with @xmath23 .
the addition of @xmath24 causes a change in @xmath25 and shifts @xmath2 to coincide with the experimental data .
the contributions of @xmath26 and soi are negligible here in the ghz scale . on the other hand , comparing the calculated @xmath4 from soi alone ( labeled ` @xmath5 ( tb ) ' ) for the chosen @xmath12 and interface condition , with experimental data , in both the top and bottom panels of fig . 2a ,
it is clear that the experimental b - field dependence of @xmath4 ( the slope , @xmath27 ) is captured from the effect of intrinsic soi , except for a shift between the soi curve and the experimental data ( different shift for @xmath28 and @xmath29 ) .
the addition of @xmath24 alone does not result in the necessary shift to match the experiment .
only after adding @xmath26 can a quantitative match with the experiment be achieved .
again the experiment - theory agreement is conditional on the interface condition and @xmath12 .
moreover , we see that the addition of @xmath26 does not change the dependency on @xmath8 . therefore , to properly explain the observed experimental behavior , we can ignore neither the soi , which is responsible for the change in @xmath4 with @xmath8 , nor the inhomogeneous b - field which shifts @xmath4 regardless of @xmath8 . note that to obtain a quantitative agreement between the experiment and the atomistic tb calculations , simultaneously in the anisotropy ( fig .
1c ) and the @xmath8 ( fig .
2a ) dependence of @xmath30 , the only knobs we have to adjust are @xmath12 and interfacial geometry i.e. how many atomic steps at the interface lie inside the dot and where they are located relative to the dot center .
these adjustments have to be done iteratively since the steps and @xmath12 not only affect the intrinsic soi but also the influence of the inhomogeneous b - field .
it is easy to separate out the contribution of the soi from the micro - magnet in the @xmath8 dependence of @xmath30 .
it will be shown in figs . 3 and 4 that , the slope , @xmath27 originates from the soi , while the micro - magnetic field shifts @xmath30 independent of @xmath8 .
first we individually match the experimental slope " from the soi and the shift " from the contribution of the micro - magnet for some combinations of the two knobs .
finally both effects together quantitatively match the experiment for @xmath31 mvm@xmath32 , and an interface with four evenly spaced monoatomic steps at -24.7 nm , -2.9 nm , 18.7 nm , 40.4 nm from the dot center along the x ( @xmath33}$ ] ) direction .
this combination also predicts a valley splitting of 34.4 @xmath34 in close agreement with the experimental value , given by 29 @xmath34 @xcite . to describe the qd , a 2d simple harmonic ( parabolic confinement )
potential was used with orbital energy splittings of 0.55 mev and 9.4 mev characterizing the x and y ( @xmath35}$ ] ) confinement respectively .
as the interface steps are parallel to y direction , the orbital energy splitting along y has negligible effects , but the strong y confinement significantly reduces simulation time .
to further our understanding , we have complemented the atomistic calculations with an effective mass ( em ) based analytic model with rashba and dresselhaus - like soi terms [ supplementary section s1 ] , as used in earlier works @xcite .
we have also developed an analytic model to capture the effects of the inhomogeneous magnetic field [ supplementary section s2 ] .
the contributions of the soi and @xmath26 on @xmath4 obtained from these models are shown in equations [ equ1 ] and [ equ2 ] respectively .
@xmath36 @xmath37 here , @xmath38 and @xmath39 are the rashba and dresselhaus - like coefficients respectively , @xmath40 is the spread of the electron wavefunction along z , @xmath41 and @xmath42 are the intra - valley dipole matrix elements , @xmath43 is the angle of the external magnetic field with respect to the @xmath33}$ ] crystal orientation and @xmath44 are the magnetic field gradients along different directions ( @xmath45 ) for a specific angle @xmath43 .
it is clear from these expressions that to match @xmath4 the difference in soi and dipole moment parameters between the valley states are relevant ( but not their absolute values ) .
the parameters used to match the experiment are @xmath46 ev@xmath47 m , @xmath48 ev@xmath47 m , @xmath49 nm , @xmath50 nm and @xmath51 nm . these fitting parameters and the separable effects of soi and @xmath26 in the em calculations enable us to obtain an even better match with the experimental data compared to tb in figs .
1c and 2a ( cyan solid lines ) .
in contrast , in the atomistic tb calculations , we include no new fitting parameters and the effects of soi and @xmath26 are intertwined by the effects of interface steps , electric field , and dot size .
as shown in figs . 1 and 2 , three physical attributes play a key role in explaining the experimental data , 1 ) soi , 2 ) @xmath24 , and 3 ) @xmath26 .
each of these contribute to @xmath52 , and only their sum can accurately reproduce the experimental data for a specific interface condition and vertical electric field , the two knobs mentioned in earlier paragraph . in figs . 3 and 4 ,
we show separately the effects of 1 ) and 3 ) respectively .
we show how the contributions of soi and @xmath26 are modulated by the nature of the confining interface ( knob 2 ) .
the influence of @xmath12 ( knob 1 ) on the effects of soi and @xmath26 are shown in the supplementary figs .
s3 and s4 respectively .
we also show how @xmath24 modifies the total homogeneous b - field in the supplementary fig .
s7 .
fig .
3a shows the angular dependence of @xmath52 for a si qd with a smooth interface calculated from tb .
both @xmath2 and @xmath3 show a 180@xmath53 periodicity but they are 90@xmath53 out of phase . from analytic effective mass study [ supplementary equation s11 ] , we understand that the anisotropic contribution from the dresselhaus - like interaction , caused by interface inversion asymmetry@xcite , results in this angular dependence in @xmath52 . moreover , the different signs of the dresselhaus coefficients @xmath39 for the valley states , give rise to a 90@xmath53 phase shift between @xmath2 and @xmath3 .
it is important to notice that the change in @xmath52 is in mhz range .
so , in ghz scale , like the blue curve ( diamond markers ) in fig . 1 d , this change is not visible
. however , if we compare @xmath2 and @xmath3 for this ideal interface case , we see @xmath54 at @xmath550@xmath53 and @xmath56 at @xmath5590@xmath53 , which does not explain the experimentally measured anisotropy .
we now discuss the remaining physical parameters needed to obtain a complete understanding of the experiment .
it is well - known that the interface between si / sige or si / sio@xmath57 has atomic - scale disorder , with monolayer atomic steps being a common form of disorder @xcite . to understand how such non - ideal interfaces can affect soi
, we first introduce a monolayer atomic step as shown in fig .
3b and vary the dot position laterally relative to the step , as defined by the variable @xmath58 . by fitting the em solutions to the tb results [ supplementary equation s15 ] , we have extracted
the dresselhaus - like coefficient @xmath39 and plotted it in fig .
3c as a function of @xmath58 .
it is seen that @xmath39 changes sign as the dot moves from the left to the right of the step edge .
both the sign and magnitude of @xmath39 depends on the distribution of the wavefunction between the neighboring regions with one atomic layer shift between them , as shown in fig .
3b . a monoatomic shift of the vertical position of the interface results in a 90@xmath53 rotation of its atomic arrangements about the [ 001 ] axis , which results in a sign inversion of the dresselhaus coefficient of that region@xcite .
a dot wavefunction spread over a monoatomic step therefore samples out a weighted average of two @xmath59s with opposite signs@xcite . .
* * a , * 1d cut of the wavefunctions of the two valley states close to a step edge , highlighting their spatial differences . a large vertical e - field ,
@xmath6030 mvm@xmath32 is used here to show a magnified effect .
* b , * the change in @xmath61 due to the inhomogeneous b - field ( @xmath26 ) alone as a function of the distance @xmath58 between the dot center and a step edge , as defined in fig . 3b .
* c , * angular dependence of @xmath62 for the various step configurations of fig .
3d ( same color code ) computed from atomistic tb .
* d , * @xmath62 as a function of @xmath8 .
@xmath62 shows negligible dependence on @xmath8 .
@xmath62 for c3 ( red lines with circular marker ) , in figs .
4c and 4d , corresponds to the contribution of @xmath26 ( the difference between the black solid curve / lines with circular markers and green dashed curve / lines with square markers ) of figs .
1c and 2a .
the fields used in the simulations of c and d are the same as that of figs . 1 and 2 , whereas the fields used for b are the same as that of fig . 3c . ]
[ fi4 ] next , we investigate the anisotropy of @xmath63 ( fig .
3e ) with various step configurations shown in fig .
3d .
@xmath63 in fig .
3e exhibits a 180@xmath53 periodicity , with extrema at the @xmath6}$ ] , @xmath7}$ ] , @xmath64}$ ] , @xmath65}$ ] crystal orientations .
both the sign and magnitude of @xmath63 depends on the interface condition .
since @xmath39 decreases when a qd wavefunction is spread over a step edge , the smooth interface case ( green curve ) has the highest amplitude .
fig .
3f shows that the slope of @xmath63 with @xmath8 changes sign for a 90@xmath53 rotation of @xmath5 and is strongly dependent on the step configuration . the step configuration labeled c3 in fig .
3d is used to match the experiment in figs . 1 and 2 .
so the curves for c3 in both figs .
3e and 3f correspond to the soi results of figs .
1c and 2a .
it is key to note here that , as @xmath12 also influences @xmath66 and @xmath67 , shown in supplementary fig .
s3 , a different combination of interface steps and @xmath12 can also produce these same soi results of figs .
1c and 2a , but might not result in the necessary contribution from micro - magnet to match the experiment .
now the dependence of @xmath4 on the interface condition will cause device - to - device variability , while the dependence on the direction and magnitude of @xmath5 can provide control over the difference in spin splittings .
these results thus give us answers to the questions 1 and 3 asked in paragraph 3 .
fig .
4 illustrates how the inhomogeneous magnetic field alone changes @xmath4 ( denoted as @xmath68 ) . since @xmath69 vectorially adds to @xmath5 , an anisotropic @xmath68
is seen in fig .
4c with and without the various step configurations portrayed in fig .
3d .
we also see that @xmath68 in fig .
4c is negligible for a flat interface , but is significant when interface steps are present .
this can be understood from figs .
4a and 4b , and/or equation [ equ2 ] .
interface steps generate strong valley - orbit hybridization@xcite causing the valley states to have non - identical wavefunctions , and hence different dipole moments , @xmath70 and/or @xmath71 , as opposed to a flat interface case , which has @xmath72 .
thus the spatially varying magnetic field has a different effect on the two wavefunctions , thereby contributing to the difference in esr frequencies between the valley states .
fig .
4b shows @xmath68 as a function of the dot location relative to a step edge , @xmath58 ( as in fig .
3b ) and illustrates that @xmath26 has the largest contribution to @xmath4 when the step is in the vicinity of the dot . also , @xmath68 is almost independent of @xmath8 , as shown in fig .
4d .
the curves labeled c3 in both figs .
4c and 4d correspond to the contribution of @xmath26 in figs .
1c and 2a .
now @xmath12 also influences @xmath73 , as shown in supplementary fig .
s4 .
thus a different combination of interface steps and @xmath12 can also produce these same @xmath26 results of figs .
1c and 2a , but might not result in the necessary soi contribution to match the experiment .
a comparison between figs .
3f and 4d ( also between equations [ equ1 ] and [ equ2 ] ) reveals that any dependence of @xmath74 on @xmath8 can only come from the soi .
this indicates that the experimental b - field dependency in fig .
2a can not be explained without the soi .
so the effect of the soi can not be ignored even in the presence of a micro - magnet and this answers question 2 raised in the third paragraph .
however , engineering the micro - magnetic field will allow us to engineer the anisotropy of @xmath4 ( question 3 , paragraph 3 )
. also , the influence of interface steps will cause additional device - to - device variability ( question 1 , paragraph 3 ) .
the coupled spin and valley behavior observed in this work may in principle enable us to simultaneously use the quantum information stored in both spin and valley degrees of freedom of a single electron .
for example , a valley controlled not gate@xcite can be designed in which the spin basis can be the target qubit , while the valley information can work as a control qubit . if we choose such a direction of the external magnetic field , where the valley states have different spin splittings , an applied microwave pulse in resonance with the spin splitting of @xmath75 , will rotate the spin only if the electron is in @xmath75 .
so we get a not operation of the spin quantum information controlled by the valley quantum information .
spin transitions conditional to valley degrees of freedom are also shown in ref .
- @xcite and an inter - valley spin transition , which can entangle spin and valley degrees of freedom , is observed in ref .
- @xcite .
to conclude , we experimentally observe anisotropic behavior in the electron spin resonance frequencies for different valley states in a si qd with integrated micro - magnets .
we analyze this behavior theoretically and find that intrinsic soi introduces 180@xmath76 periodicity in the difference in the esr frequencies between the valley states , but the inhomogeneous b - field of the micro - magnet also modifies this anisotropy .
interfacial non - idealities like steps control both the sign and magnitude of this difference through both soi and inhomogeneous b - field .
we also measure the external magnetic field dependence of the resonance frequencies .
we show that the measured magnetic field dependence of the difference in resonance frequencies originates only from the soi .
we conclude that even though the soi in bulk silicon has been typically ignored as being small , it still plays a major role in determining the valley dependent spin properties in interfacially confined si qds .
these understandings help us answer the questions raised in paragraph 3 , which are crucial for proper operation of various qubit schemes based on silicon quantum dots . |
electromagnetic processes have always provided invaluable probes of the strong interactions , the most prominent example being the deep inelastic scattering experiments in the early 1970s which eventually established qcd as the fundamental theory of the strong interactions . among the current challenges of qcd
is the description of the plethora of charmoniumlike states discovered during the past years , the so - called xyz " ( for reviews see , e.g. , refs .
@xcite ) , that lie above the @xmath1 threshold , and do not fit potential model expectations .
it is widely accepted that the effects of the coupling of a core charmonium , namely , a mainly @xmath8 bound state , to a @xmath1 meson pair , the so - called coupled - channel effects , are an important ingredient to understand those states .
we shall call the latter continuum states and the former valence states
. we will argue below that two - photon transitions between heavy quarkonium states may provide important experimental information on the continuum - valence coupling .
historically , it has been recognized for a long time that the continuum states may shift the mass spectrum of a pure @xmath9 state considerably @xcite . recently
, exploratory investigations on the mixing between discrete and continuum states in charmonium have been carried out in lattice qcd @xcite .
lattice qcd has also provided a detailed study of the so - called string breaking " in the static approximation @xcite , which may be used to extract information on the continuum - valence coupling @xcite .
if the heavy quark mass @xmath10 is much larger than the remaining scales in the system , then the heavy quarks move slowly in their center - of - mass frame , say with a typical velocity @xmath11 , that is , with typical three momentum @xmath12 and hence with typical binding energy @xmath13 .
nonrelativistic qcd ( nrqcd ) @xcite can be used to factorize the contributions from energies larger or of the order of @xmath10 , and provides a good starting point .
recall that the heavy - light meson pair threshold lies at a typical nonrelativistic energy @xmath14 , according to heavy quark effective theory ( hqet ) counting rules @xcite .
then , if the binding energy of a heavy quarkonium state lies much below open flavor threshold , namely , @xmath15 one can integrate out energies of order @xmath14 .
this leads to potential nrqcd ( pnrqcd ) in the strong coupling regime @xcite .
the dynamics of this effective theory reduces to a heavy quark and a heavy antiquark interacting through a potential@xcite ( see @xcite for a review ) . hence , in order to understand the properties of these states , there is no need to introduce explicitly the continuum states if the potential is chosen appropriately .
however , for states close to or above the open flavor threshold , the coupling of continuum to valence states needs to be addressed , and so far it is not known how to proceed in a model independent way . , which includes coupled channel effects @xcite .
see also @xcite for an even more recent proposal .
] hence , most of the analysis has been done by using different models .
recently , some general features of the coupled - channel effects have been obtained in the quark model @xcite under the assumptions that valence - continuum coupling is described by the @xmath16 model @xcite and the interaction between two mesons is negligible .
these results suggest that for the low - lying states the effect of continuum channels is hidden in the parameters in the potential model . in ref .
@xcite , it is mentioned that the radiative transition process may be sensitive to the continuum components .
however , previous works , based on cornell coupled - channel formalism , indicate that the relativistic corrections @xcite may be more important than the continuum contributions @xcite .
therefore , it is not straightforward to disentangle the continuum contributions in the one - photon transition process . here , we propose a new process , namely , @xmath17 which can provide an additional opportunity to pin down the valence - continuum coupling . from the theoretical point of view ,
electromagnetic transitions allow a cleaner analysis than hadronic decay processes . at the amplitude level
, the contribution to this two - photon transition can be divided into two parts .
we refer to the first one as the discrete part , which involves charmonium states only .
the discrete part is dominated by the following process : the @xmath18 state decays into a real or virtual @xmath2 state by radiating one photon , and then , the real or virtual @xmath2 state decays into @xmath7 plus another photon .
this process not only includes the cascade decay process ( on - shell region ) , but also the off - shell region .
we will study the whole phase space .
the second part is referred to as the continuum part , in which at least one photon is emitted by an intermediate charmed meson . in the discrete part
, the @xmath19 states can be on - shell , so the total contribution of the discrete subprocess should be much larger than that of the continuum part
. however , when the invariant mass of @xmath7 and one of the photons is far away from the resonance regions of the @xmath19 states , the discrete part contribution will drop down very fast . thus , the contribution of the continuum part may be important and measurable in the off - shell region .
let us mention that in ref .
@xcite , a similar decay process of @xmath20 is suggested to determine the values of the strong couplings @xmath21 and @xmath22 . on the experimental side ,
such a two - photon transition process has already been studied in the 1970s and 1980s @xcite . in recent years
, more precise measurements were carried out by the besiii @xcite and cleo @xcite collaborations .
however , they focused on the investigation of the cascade decay @xmath23 followed by @xmath24 and on the study of the properties of the @xmath25 states . only in ref.@xcite ,
was a discussion made on the possible amplitude of the two - photon transition in treating the backgrounds of the @xmath25 states .
hence , so far , no one has ever used it to study coupled - channel effects .
recently , the besiii @xcite collaboration reported the significant data excess from the known cascade backgrounds , which was interpreted as the nonresonance decay of @xmath26 .
we remark that this measurement is very sensitive to the line shapes of the @xmath25 states , especially in the data selection region .
because of the above reason , the significance of an eventual experimental determination of the continuous process depends very much on our knowledge on the discrete states . as far as we know ,
the discrete contribution has not been fully studied yet , and only the individual contribution of each @xmath25 state is known by using the non relativistic breit - wigner formula together with the dynamical factors to describe the @xmath25 line shape in the cascade decay of @xmath23 and @xmath24 @xcite .
so , in this paper we restrict ourselves to analyze the discrete contribution to this decay assuming that the coupling of @xmath27 and @xmath7 to @xmath28 is zero , and leave for a future work , the detailed evaluation of @xmath1 meson pair loops effects . in this work
, we will use effective field theory methods to calculate the decay width , the photon spectrum , and the @xmath7 polarization in the discrete subprocess . a complete study of the whole contribution of the discrete @xmath29 states together with some higher radial excitations , @xmath4 states , where @xmath30 , are ignored .
] , will be carried out .
the rest of this paper is organized as followings : in sec.ii , we will briefly introduce the effective lagrangian we use and determine the value of the effective couplings ; in sec.iii , we will calculate the discrete contribution to the @xmath0 process and show the results , and the discussion and summary will be given in the last section .
the heavy quarkonium states are mainly constituted by a @xmath31 pair and classified according to the spectroscopic notation @xmath32 , where @xmath33 is the radial quantum number , @xmath34 is the total spin of the heavy quark pair , @xmath35 ( or @xmath36 ) is the orbital angular momentum , and @xmath37 is the total angular momentum .
they have parity @xmath38 and charge conjugation @xmath39 . as mentioned in the introduction
, nrqcd is a good starting point to describe this system .
the lo nrqcd lagrangian is invariant under s=@xmath40 spin - symmetry group , an approximate symmetry of the heavy quarkonium states , that is inherited in the subsequent effective theories .
we assume that the entire dynamics of these states can be described by a ( nonperturbative ) potential .
that is the case if they are in the strong coupling regime of pnrqcd .
this is a reasonable assumption for the @xmath29 states @xcite , and to lesser extend for @xmath7 may well be better described by the weak coupling regime of pnrqcd @xcite . at leading order in this regime , the dynamics
is described by a ( perturbative ) potential , and the multipole expansion also holds . ] .
the remaining states are close to or above threshold and are subject to the uncertainties due to the influence of @xmath1 pairs , that we plan to analyze in a separate work . in any case
, we will only use the fact that the typical energy of the emitted photons is at the ultrasoft scale @xmath13 , and the typical size of the system at the soft scale @xmath41 .
this allows hs to carry out the multipole expansion of the photon field about the center - of - mass coordinate , which means that the photons see the heavy quarkonium states as pointlike particles .
hence , it is most convenient to introduce hadronic spin - symmetry multiplets , in an analogous way as it was initially done in hqet @xcite . for heavy quarkonium states ,
this formalism was developed in ref.@xcite .
the states have the same radial number @xmath42 and the same orbital momentum @xmath43 can also be expressed by means of a single multiplet @xmath44 @xcite , @xmath45 where @xmath46 is the four velocity associated to the multiplet @xmath44 ( not to be mistaken by @xmath47 , the typical velocity of the heavy quark in the heavy quarkonium rest frame ) , @xmath48 represents the spin - singlet effective field , and @xmath49,@xmath50 and @xmath51 represent the three spin - triplet effective fields with @xmath52 , and @xmath53 , respectively .
the four tensors are all completely symmetric and traceless and satisfy the transverse condition @xmath54 @xmath55 , @xmath56 .
the properties of @xmath57 and @xmath58 under parity , charge conjugation and heavy quark spin transformations can be easily obtained by assuming that the corresponding transformation rules of the multiplet @xmath44 follow as : @xmath59 @xmath60^{t}c,\end{aligned}\ ] ] @xmath61 where @xmath62 is the charge conjugation matrix ( @xmath63 in the dirac representation ) , and @xmath64 and @xmath65 correspond to the heavy quark and heavy antiquark spin - symmetry groups ( @xmath66=[s^{\prime},\slashed{v}]=0 $ ] ) .
since we are going to consider the two - photon decay of @xmath18 into @xmath67 via the intermediate states @xmath2 , it will be helpful to give the explicit expressions of the @xmath68- and @xmath69-wave multiplets that follow from eq.(1 ) . for the @xmath70 case , we have @xmath71 and for the @xmath72 case , @xmath73 the radiative transitions between @xmath74 and @xmath75 charmonium states in the nonrelativistic limit is given by the lagrangian : @xmath76v_{\nu}f^{\mu\nu}+\mathrm{h.c . } , \label{vertex}\ ] ] where @xmath77 is the coupling constant , and @xmath78 is the electromagnetic tensor .
the lagrangian in eq.(6 ) preserves parity , charge conjugation , gauge invariance , and heavy quark and antiquark spin symmetry . using the effective lagrangian , it is straightforward to calculate the e1 transition decay widths : [ e1 ] @xmath79 @xmath80 [ decays ] where @xmath81 is the energy of the emitted photon .
the results are slightly different from those given in refs.@xcite .
it is because the initial and final charmonium states can not be static simultaneously , and we choose different values of @xmath82 for them to maintain the transverse condition in eq.([trans ] ) .
namely , in the vertex ( [ vertex ] ) we have substituted the velocity @xmath83 in the current that produces the outgoing particle by @xmath84 , its four velocity .
the remaining velocities ( i.e. the explicit one and the one in the current that annihilates the incoming particle ) are chosen as @xmath85 , the four velocity of the incoming particle .
the explicit calculations in ( [ decays ] ) show that these changes maintain the spin - symmetry ratios between the decay widths of the different states ( provided that the same mass is used for each spin multiplet ) .
note that @xmath86 , and the results of ref.@xcite differ from ours by @xmath87 contributions . in ( [ decays ] )
we have also allowed the coupling constant @xmath77 of the spin - symmetry multiplet to depend on the total angular momentum @xmath37 , namely , @xmath88 . in this way we are accounting for spin - symmetry breaking terms due to the spin - orbit , spin - spin and tensor potentials in pnrqcd ( or in potential models ) .
note that although spin - breaking terms look like an @xmath89 in nrqcd ( like in hqet ) , this is actually the case only when ultrasoft particles are emitted , for instance , in magnetic transitions , or in processes involving pseudo - goldstone bosons @xcite .
soft emissions can only be virtual and lead to the above mentioned @xmath90 spin - breaking potentials , which make the coupling constants different at this order . in conventional potential models , the differences between @xmath91 are usually taken to be of order @xmath92 , although on general grounds they could be as large as order @xmath47 @xcite .
@xmath77 can , in principle , be obtained by calculating the matrix element of the electromagnetic current between the wave functions of the @xmath75 and @xmath74 states in pnrqcd ( or in any potential model , see ref.@xcite for a recent review ) .
nowadays they can also be obtained from lattice qcd @xcite .
here we will determine their values using experimental data , except for those related to the @xmath4 for which experimental data is not available . in the last case the values will be estimated using a potential model . for the @xmath5 case ,
the @xmath93 and @xmath94 processes have been measured with a very high precision @xcite , so we can determine the values of the relevant coupling constants by comparing with the experimental results . using the pdg data @xcite , @xmath95gev , @xmath96gev , @xmath97 , @xmath98 , @xmath99 and @xmath100
@xmath101 @xmath102 we obtain the absolute values of @xmath103 and @xmath104 listed in table i. for the @xmath6 case , only the @xmath105 [ formerly called @xmath106@xcite , ] has been included in the pdg , so we estimate the relevant parameters with the help of a potential model .
the spectrum of the @xmath6 states and their radiative decay into the lower @xmath68-wave states have been calculated by numerous groups ( for reviews see , e.g. , ref.@xcite ) . by using the screened potential ,
the updated results of the charmonium spectrum and the e1 transition rates are given in ref.@xcite . in this work , besides setting @xmath107 , @xmath108 is assigned to the @xmath109 state and @xmath110 is predicted . from the nonrelativistic results of @xmath111 presented in table iv of ref.@xcite
, we obtain the absolute values of the corresponding coupling constants and give them in table i. because the mass differences between the @xmath6 states are much smaller than the mass gap between the @xmath6 and @xmath112 ( @xmath113 ) states , in practice we use the center of gravity mass @xmath114 and the center of gravity coupling constant @xmath115 rather than calculating the contribution of the individual @xmath6 states .
.the numerical values of the coupling constants @xmath116 are shown . for the @xmath117 case ,
the results are obtained by fitting the experimental data , and for @xmath118 , the results are determined by comparing with the potential model predictions @xcite . [ cols="^,^,^,^,^,^,^",options="header " , ]
now , we proceed to calculate the decay rate of the process @xmath120 via intermediate states @xmath2 .
such @xmath121 process can be described by the following dimensionless variables : @xmath122 in terms of @xmath123 , the three - body phase space @xmath124 can be written as @xmath125 for each intermediate @xmath2 , there are two feynman diagrams , which are shown in fig.1 .
the corresponding feynman amplitude is denoted by @xmath126 .
putting the contributions of the three @xmath6 states together , the total feynman amplitude is then divided into four parts : @xmath127 each of @xmath128 on the right - hand side of the above equation can be obtained from the lagrangian in eq.(6 ) upon making the same replacements as discussed after ( [ vertex ] ) .
since the propagator of the @xmath5 fields may become on - shell , self - energy corrections must be considered .
we approximate them by introducing a constant decay width , which , parametrically , is @xmath129 .
being that these figures are much smaller than @xmath130 , @xmath131 corrections to the static propagator must be considered in order to match its size . as an alternative
, we use relativistic propagators , that include them .
for example , the feynman amplitude of the first diagram in [ fig : feynman ] for @xmath132 , which is the most complicated one , is @xmath133 \times\frac{(\pi_{\alpha\alpha_1}\pi_{\mu\mu_1}+\pi_{\alpha\mu_1}\pi_{\mu\alpha_1})/2- \pi_{\alpha\mu}\pi_{\alpha_1\mu_1}/3}{v_p^2 - 1+\mathrm{i}*\gamma_{\chi_{c2}(1p)}/m_{\chi_{c2}(1p ) } } \nonumber\\ & \times & \mathrm{tr}[\frac{1+\slashed{v}_{p}}{2}\gamma^{\alpha_1}\slashed{\epsilon}^{\ast}_{1s}\frac{1+\slashed{v}_{1s}}{2 } ] \times v_{\nu,2s}v_{\nu , p}\bar{f}_{2}^{\mu\nu}\bar{f}_{3}^{\mu_1\nu_1}. \label{amplitude}\end{aligned}\ ] ] where @xmath134 , @xmath135 , @xmath136 , and @xmath137 is the total width of @xmath138 . in the above expression
, we have omitted the imaginary unit i , which is a global factor and has no influence on the final result . for the convenience of further discussion
, we also divided the decay width of the discrete part into four parts : @xmath139 where @xmath140 is the sum of the three individual contributions of the @xmath25 states , which is proportional to @xmath141 , @xmath142 is the interference between the @xmath5 states @xmath143 , @xmath144 is the contribution involving the @xmath6 states only , and @xmath145 is the interference contribution between the @xmath5 and @xmath6 states . the @xmath146 " sign before the last term corresponds to the two possible relative phase angles @xmath147 @xmath148 between the @xmath5 and @xmath6 states .
decay into @xmath149 via intermediate states @xmath2 . ]
we have computed @xmath150 analytically , but the outcome is too lengthy to be presented here . after doing the phase space integrals numerically , we obtain that @xmath151 the numerical results yield that on the total decay width level the effects of the interference among the @xmath29 states as well as the effect of the @xmath6 states are so small that they can be neglected .
we have also calculated the photon spectrum @xmath152 , and display the figures for each part separately in [ fig : chic1p ] and [ fig : chic2p ] , respectively . since we have chosen different @xmath153 for the initial and final states in eq.([amplitude ] ) , we may also have a nontrivial @xmath7 polarization , which should be zero in the single @xmath153 case .
similar to the production case @xcite , here , we define the polarization parameter @xmath154 as @xmath155 where @xmath156 and @xmath157 are the decay widths for @xmath7 in the transverse and longitudinal polarization states , respectively .
we calculate @xmath154 in the rest frame of @xmath18 and get @xmath158 .
this is slightly different from zero , and we interpret it as a purely kinematic effect .
we find that @xmath154 is mainly determined by the resonance contribution , being the influence of the interference and the @xmath6 terms tiny .
the experimental measurement of the @xmath0 decay can be implemented in the off - mass shell region of @xmath25 , i.e. , the experimentally sensitive region in daltiz plot @xmath159 where @xmath160 is the invariant mass of the two photons and @xmath161 is the invariant mass of @xmath7 and the higher energy photon .
these cut can mostly exclude the contribution from the highly yielded @xmath29 states . from the photon spectrum in the cut region indicated in [ fig : chic_cut ] , the different contributions to the decay width ( [ decay ] ) read @xmath162 we have also computed the branching ratio , photon spectrum , and the polarization parameter in the cut region .
the result of the branching ratio is @xmath163 the result in eqs.([cut1],[cut2 ] ) shows that , contrary to the case of the total decay width ( [ total ] ) , in the cut region the effect of interference among @xmath5 states as well as the contribution from the @xmath6 states can not be ignored . it is more than @xmath164 of the sum of the separated contributions of the @xmath29 states .
the @xmath7 produced in the cut region tend to be unpolarized and the polarization parameter in the cut region becomes @xmath165 and @xmath166 for @xmath167 and @xmath168 , respectively .
if we only include the three individual contributions of @xmath29 , the value of @xmath169 turns to be @xmath170
. finally , the different contributions to the photon spectrum in the cut region are shown [ fig : chic_cut ] . as mentioned before
, the coupling constant @xmath77 is related to the spatial matrix element @xmath171 in potential models .
at least three different potential models , including the cornell potential @xcite and the screened potential @xcite , give the phase angle @xmath172 to be @xmath173 @xcite . hence , the @xmath168 option in our calculation appears to be favored .
we have estimated the discrete contribution to the @xmath174 due to electric dipole transitions in the whole phase space and , in particular , in the cut region used at besiii .
higher multipole electric transitions and the magnetic ones are suppressed by at least @xmath175 in the amplitude .
this is also the case for contributions arising from two - photon vertices .
the largest uncertainty in our calculation comes from the fact that we neglect the contribution of @xmath75 states for @xmath176 .
this contribution is not parametrically suppressed by powers of @xmath47 , although we expect it to be small for the following two reasons : ( i ) the propagator of the @xmath75 state is increasingly off - shell when @xmath42 increases , and ( ii ) the coupling constants @xmath177 and @xmath178 are proportional to the overlap of the radial wave functions of the corresponding states , which decreases with @xmath42 .
apart from this truncation , the estimate is reliable at leading order in @xmath92 .
in fact , the modifications we have made in ( [ amplitude ] ) , which introduce terms of higher order in @xmath47 , take into account relativistic effects in the kinematics , and hence help in providing a better estimate . from the point of view of potential models ,
the leading corrections to the @xmath179 transition processes , @xmath180 and @xmath181 , mainly arise from the three following sources : ( i ) relativistic modification of the nonrelativistic wave functions , ( ii ) finite size effects , and ( iii ) contribution of high @xmath92 order electromagnetic operators @xcite . in this work ,
we use the effective lagrangian , eq.([amplitude ] ) , to describe the @xmath179 transition process . for the @xmath5 case , we determine the values of the corresponding coupling constants from the experimental data .
thus , the effects of ( i ) , which are dominant , are taken into account in our results , as far as the vertices involving the @xmath5 states is concerned .
therefore , our results for the two - photon decay width in the off - shell region are more accurate than a potential model calculation for the one - photon @xmath179 transition process .
both the corrections due to ( ii ) and to ( iii ) could be taken into account by including higher dimensional operators in the vertices ( [ vertex ] ) .
as mentioned before these operators are suppressed by at least order @xmath92 .
let us next discuss how our results compare with the usual inputs in the monte - carlo ( mc ) codes that are used to analyze the experimental data . in the experimental treatment of the four contributions in eq.([decay ] ) in the peaking region , usually only the first one , @xmath140 , is taken into account , and is often modeled using the nonrelativistic breit - wigner line shape of @xmath25 and @xmath7 .
the other three components are negligible and generally omitted .
but in the off - shell region , as argued above , the other three components will be sizable and have to be considered in the data treatment .
furthermore , in the off - shell cut region ( [ eq : cut ] ) , even this naive nonrelativistic breit - wigner line shape description of the individual @xmath25 states contribution needs improvement .
the reason is that a ( single ) nonrelativistic breit - wigner is only a good approximation to the line shape of @xmath182 in the resonance peaking region , as shown in [ fig : res_a ] . in the cut region equation ( [ eq : cut ] ) , which lies between the @xmath183 and @xmath184 resonance peaks ,
there is no guarantee that the nonrelativistic breit - wigners will provide a correct description of data .
let us next point out the main ingredients it misses . in the @xmath179 transition process
, the decay rate is proportional to the factor @xmath185 as shown in eq.([e1 ] ) .
we may then improve on the nonrelativistic breit - wigner approximation by just introducing the correct photon energy dependence in each vertex i.e. the full @xmath186 scale factor from the two @xmath179 transitions .
note that if one only includes the @xmath185 correction due to the first @xmath179 transition , @xmath187 in the cut region will be overestimated , because the energies of the two photons are negatively correlated . in [ fig : mc_sim ]
, the mc simulation results of the line shapes in the cut region are shown , which are implemented in two ways .
one is done by including the @xmath188 correction from a double @xmath179 transition , and the other one is not .
the difference between these two mc simulation results can be understood as follows : the @xmath179 transition enhances the right tail of photon energy peak and depresses the left tail of the nonrelativistic breit - wigner . as a whole , the effect of the correlated emitted photons in the double @xmath179 transition increases the @xmath25 contribution in the cut region . for comparison , we also plot the effective lagrangian result of the sum of the three individual @xmath25 states contribution in [ fig : mc_sim ] .
it is clear that neither of the simulation results agree with that of the effective lagrangian calculation , although they are qualitatively similar . from the amplitude in eq.([amplitude ] ) , it can be verified that the @xmath186 factor and the nonrelativistic breit - wigner is only the leading - order nonrelativistic approximation of the effective lagrangian calculation in the off - shell region .
hence , even for a delicate description of the individual contribution in the off - shell region , including the double @xmath179 transition correction in the naive nonrelativistic breit - wigner only may not be enough . in summary
, we have estimated the discrete contribution to the @xmath189 due to electric dipole transitions in the whole phase space , and in particular in the cut region in the experimental measurement .
we find that for the full decay width the interference contribution and the contributions of higher excited states can be safely neglected .
however , in the regions of the phase space off the resonance peaks , their contributions are considerable and important for a delicate experimental measurement .
as argued in the introduction , a large deviation of our results from an experimental observation , would indicate that the effects of the @xmath1 threshold are significant . in the cut region , where the branching fractions are from pdg @xcite .
the dotted line denotes the naive nonrelativistic breit - wigner simulation , the solid line is the simulation including the @xmath190 factor , and the dashed line is the contribution of the three individual @xmath29 states , corresponding to @xmath140 in ( [ decay ] ) , calculated in this paper .
, scaledwidth=48.0% ]
we would like to thank bai - qing li for correspondence about the relative phase between @xmath5 and @xmath6 contribution .
he and lu also thank de - shan yang and qiang zhao for helpful discussions .
he and j.s . are supported by the csd2007 - 00042 consolider - ingenio 2010 program ( he under contract no.cpan08-pd14 ) , and by the fpa2007 - 66665-c02 - 01/ and fpa2010 - 16963 projects ( spain ) .
j.s . has also been supported by the rtn flavianet mrtn - ct-2006 - 035482 ( eu ) , the ecri hadronphysics2 ( grant agreement no
. 227431 ) ( eu ) , grant no .
fpa2007 - 60275/mec ( spain ) and the cur grant no .
2009sgr502 ( catalonia ) .
this work is also supported in part by national natural science foundation of china ( nsfc ) 10905091 , 100 talents program of cas , ministry of science and technology of china ( most ) , foundation b of president of gucas , srf for rocs of sem , and china postdoctoral science foundation .
g. s. bali , h. neff , t. duessel , t. lippert and k. schilling [ sesam collaboration ] , phys .
d * 71 * , 114513 ( 2005 ) . c. downum , in ix international conference on quark confinement and hadron spectrum : qchs-09 , aip conf .
1343 ( aip , new york , 2010 ) .
w. e. caswell and g. p. lepage , phys .
b * 167 * , 437 ( 1986 ) .
b. a. thacker and g. p. lepage , phys .
d * 43 * , 196 ( 1991 ) .
g. t. bodwin , e. braaten and g. p. lepage , phys .
d * 51 * , 1125 ( 1995 ) [ erratum - ibid .
d * 55 * , 5853 ( 1997 ) ] .
m. neubert , phys .
rept . * 245 * , 259 ( 1994 ) .
n. brambilla , a. pineda , j. soto and a. vairo , nucl .
b * 566 * , 275 ( 2000 ) .
n. brambilla , a. pineda , j. soto and a. vairo , phys .
d * 63 * , 014023 ( 2000 ) .
a. pineda and a. vairo , phys .
d * 63 * , 054007 ( 2001 ) [ erratum - ibid .
d * 64 * , 039902 ( 2001 ) ] .
n. brambilla , a. pineda , j. soto and a. vairo , phys .
b * 580 * , 60 ( 2004 ) .
n. brambilla , a. pineda , j. soto and a. vairo , rev . mod .
phys . * 77 * , 1423 ( 2005 ) .
t. himel _ et al .
_ , phys .
lett . * 44 * , 920 ( 1980 ) ; m. oreglia _ et al .
_ , phys .
* 45 * , 959 ( 1980 ) ; m. oreglia _ et al .
_ , phys .
d * 25 * , 2259 ( 1982 ) ; j. gaiser _ et al .
_ , phys .
d * 34 * , 711 ( 1986 ) .
n. brambilla , d. eiras , a. pineda , j. soto and a. vairo , phys . rev
. lett .
* 88 * , 012003 ( 2001 ) .
a. a. penin , a. pineda , v. a. smirnov and m. steinhauser , nucl .
b * 699 * , 183 ( 2004 ) [ erratum - ibid . *
829 * , 398 ( 2010 ) ] .
n. brambilla , y. jia and a. vairo , phys .
d * 73 * , 054005 ( 2006 ) .
y. kiyo , a. pineda and a. signer , nucl .
b * 841 * , 231 ( 2010 ) .
r. casalbuoni , a. deandrea , n. di bartolomeo , r. gatto , f. feruglio and g. nardulli , phys .
b * 302 * , 95 ( 1993 ) .
f. de fazio , phys .
d * 79 * , 054015 ( 2009 ) . | the decay mode @xmath0 is proposed in order to experimentally identify the effects of the coupling of charmonium states to the continuum @xmath1 states . to have a better understanding of such a two - photon decay process , in this work
we restrict ourselves to investigate the contribution of the discrete part , in which the photons are mainly produced via the intermediate states @xmath2 . besides calculating the resonance contributions of @xmath3
, we also take into account the contributions of the higher excited states @xmath4 and the interference effect among the @xmath5 and @xmath6 states .
we find that the contribution of the @xmath6 states and the interference terms to the total decay width is very tiny .
however , for specific regions of the dalitz plot , off the resonance peaks , we find that these contributions are sizable and should also be accounted for .
we also provide the photon spectrum and study the polarization of @xmath7 . |
transition metal dichalcogenides ( tmds ) @xmath15 , where @xmath3 is a transition metal such as w and mo , and @xmath16 is a chalcogen such as s , se , and te , receive considerable attention due to their important mechanical and electronic properties@xcite .
molybdenum disulfide mos@xmath0 , a prototypical example of tmds , is a layered system where mo atoms form hexagonal layers@xcite .
each of the mo hexagonal layer is sandwiched between two similar lattices of s atoms , forming a trilayer@xcite .
the atoms within each trilayer are held together by strong covalent bonds , while the trilayers of mos@xmath0 interact primarily through weak van der waals interactions .
it is this sandwiched structure that endows mos@xmath0 with the important mechanical properties for solid lubricants@xcite .
the electronic , optical , and lattice dynamical properties have been under intense investigations@xcite .
the research on multilayers of mos@xmath0 , among many other multilayers of tmds , have been fueled by their novel properties intrinsic to 2d materials .
for example , successes of mos@xmath0 multilayers have been demonstrated for the purposes of energy - efficient field - effect transistor@xcite , advanced electrocatalysts@xcite , thermoelectric devices@xcite with a large and tunable seebeck coefficient , phototransistors@xcite , superconductivity@xcite , etc .
mos@xmath0 is joining the rank of other low - dimensional companions , demanding both efficient and accurate treatment of a first - principles approach@xcite . even though the mechanical , electronic , and lattice dynamical properties of the equilibrium structure of mos@xmath0 have been studied extensively@xcite , there are relatively few first - principles studies of the anharmonic effects@xcite that contribute to the thermal properties such as thermal conductivity and thermal expansion coefficients ( tecs )
. the linear tecs of 2h - mos@xmath0 have been measured @xcite where it was found that the tec along the @xmath2 direction is larger than that along the @xmath1 direction . on the theoretical side
, tecs may be calculated by solving the vibrational self - consistent - field equations@xcite or the nonequilibrium green s function method@xcite .
tecs may also be determined from a quasiharmonic approximation ( qha ) calculation where a set of calculations is to be carried out over a grid or mesh of lattice - parameter points , where the dimensionality of the grid depends on the number of independent lattice parameters@xcite . recently ding _
et al._@xcite chose six volumes to perform phonon calculations to first obtain the volumetric tec .
another relation involving the linear tecs for @xmath1 and @xmath2 ( eq . 15 of ref . [ ] ) was set up and the values of tecs were solved . in this work , we develop a direct approach based on the grneisen formalism to calculate the tecs in the @xmath1 and @xmath2 directions .
our tec results are then compared with experiment .
the outline of this paper is as follows : section [ sec : method ] discusses the methodology used to efficiently calculate the thermal expansion coefficients of a general hexagonal system . section [ sec : results ] reports the results and discussions on the application of the method to mos@xmath0 .
section [ sec : conclusions ] contains the conclusions .
we shall first present the expressions for tecs for a general hexagonal system obtained from the grneisen formalism.@xcite results specific to the hexagonal mos@xmath0 will be presented later .
the linear tecs of the crystal along the @xmath9 , @xmath10 and @xmath8 directions , denoted by @xmath17 , @xmath18 , and @xmath19 , at temperature @xmath3 can be described by a matrix equation @xmath20 where @xmath21 is the equilibrium volume of the primitive cell , @xmath22 is the elastic compliance matrix@xcite .
the values @xmath23 are the matrix elements of the elastic constant matrix @xmath24 that corresponds to a hexagonal system@xcite where @xmath25 the integrated quantities in eq .
[ eq : alpha ] are given by @xmath26 where the integral is over the first brillouin zone ( bz ) .
the frequency @xmath27 of a phonon mode depends on the mode index @xmath28 and wavevector @xmath29 .
the heat capacity contributed by a phonon mode with frequency @xmath6 at temperature @xmath3 is @xmath30 with @xmath31 , @xmath32 and @xmath33 are the planck and boltzmann constants , respectively .
the grneisen parameter @xmath34 measures the relative change of a phonon frequency @xmath35 as a result of an @xmath7 type deformation with strain size @xmath36 applied to the crystal .
for example , if a uniaxial strain is applied in the @xmath9 direction then the strain parameters are @xmath37 ( in the voigt notation@xcite ) , i.e. , @xmath38 , and @xmath39 , for @xmath40 .
we apply uniaxial strains in the @xmath9 , @xmath10 , and @xmath8 directions to give @xmath41 , @xmath42 , and @xmath43 , respectively .
grneisen parameters are evaluated using a central - difference scheme , where a change in the dynamical matrices before and after deformation is used in the perturbation theory to deduce the changes in eigenfrequencies@xcite . by a proper sampling in the @xmath44-space
, we may calculate the phonon density of states as @xmath45 next we introduce a related quantity , @xmath46 , the phonon density of states weighted by the grneisen parameters as @xmath47 the usefulness of @xmath5 is that we may obtain @xmath4 in eq .
[ eq : integ ] from another relation @xmath48 to calculate the linear tecs , it appears that a set of three uniaxial deformations in the @xmath9 , @xmath10 , and @xmath8 directions are needed .
however , due to the symmetry of the hexagonal system , we should have @xmath49 on physical ground and hence @xmath50 , so that the tec eq .
[ eq : alpha ] reduces to @xmath51 & c_{13 } \\
2c_{13 } & c_{33 } \\
\end{pmatrix}^{-1 } \begin{pmatrix } i_{1}\\ i_{3 } \end{pmatrix } \label{eq:2d1}\ ] ] or @xmath52 \\ \end{pmatrix } \begin{pmatrix } i_{1}\\ i_{3 } \end{pmatrix}\ ] ] where @xmath53 .
therefore , for a hexagonal system two uniaxial strains , the first one either in the @xmath9 or @xmath10 direction , and a second one in the @xmath8 direction are sufficient to determine the linear tecs . for mos@xmath0 , the symmetry of the hexagonal system is not altered ( the space group remains as @xmath54 @xmath55 ) when a uniaxial strain is applied in the @xmath8 direction .
however , the symmetry is significantly lowered from hexagonal with a space group of @xmath54 @xmath55 to base - centered orthorhombic with a space group of @xmath56 @xmath57 after a uniaxial strain is applied in the @xmath9 or @xmath10 direction .
this will result in an increase of the computational cost compared to that which preserves the hexagonal symmetry where a phonon calculation is to be performed .
for example , after applying a uniaxial strain in the @xmath9 direction , a @xmath58 @xmath59-mesh required in a phonon calculation will result in @xmath60 irreducible @xmath59-points and @xmath61 irreducible representations or @xmath61 self - consistent - field calculations .
this is to be compared with the symmetry - preserving deformations ( e.g. , a uniaxial strain in the @xmath8 direction ) where the number of irreducible @xmath59-points is @xmath62 and the number of irreducible representations is @xmath63 , which clearly shows a substantial computational saving .
more savings are expected when complicated crystal structures are treated .
we propose to use a computationally efficient , symmetry - preserving biaxial strain in the @xmath11 plane ( hereafter it shall be called an @xmath11 biaxial strain ) where @xmath64 to evaluate the grneisen parameters @xmath65 and use eq .
[ eq : integ ] or eq .
[ eq : integ2 ] to obtain @xmath66 . due to the underlying symmetry , @xmath67 .
however it should be noted that grneisen parameters due to an @xmath11 biaxial strain may not the same as the average of the grneisen parameters due to @xmath9 and @xmath10 uniaxial strains .
these points will be elaborated later .
the phonon spectra of mos@xmath0 are calculated with density functional perturbation theory ( dfpt)@xcite . for the unstrained structure , a @xmath59-mesh of @xmath68 is used for the phonon calculations , which is equivalent to evaluating the force constants@xcite using a @xmath68 supercell .
the phonon calculations proceed by evaluating dynamical matrices at a number of irreducible @xmath59 points . from the the dynamical matrices the inter - atomic force constants in the real space
are obtained by an inverse fourier transform , and these force constants are used to construct dynamical matrices at any @xmath69 to calculate the phonon eigenfrequencies @xmath27 .
for the strained structures , a @xmath59-mesh of @xmath58 is also used . for the unstrained structure , a larger @xmath59-mesh of @xmath70
is used to confirm that a @xmath59-mesh of @xmath68 is sufficient for the purpose of tec calculations .
the bulk mos@xmath0 belongs to @xmath71 nonsymmorphic space group ( # @xmath72 ) , with two inequivalent atoms where a mo atom occupies a @xmath73 site and a s atom occupies @xmath74 site , @xmath75 .
this gives a total of six atoms in the hexagonal primitive cell .
density - functional theory ( dft ) calculations are carried out using a plane - wave basis code quantum espresso@xcite .
the orientation of the crystal adopted in this work is dictated by the choice of the primitive lattice vectors where @xmath76 , @xmath77 , and @xmath78 , @xmath1 and @xmath2 are the hexagonal lattice parameters .
we use @xmath79 ry as the cutoff energy for the plane - wave basis set . the local density approximation ( lda )
is used to describe the exchange and correlation .
pseudopotentials for mo and s are generated from pslibrary.1.0.0 based on the rappe - rabe - kaxiras - joannopoulos@xcite scheme .
a @xmath80 monkhorst - pack @xmath44-point mesh is used .
the hexagonal lattice parameters and the atomic positions are fully relaxed . the force tolerance is taken to be @xmath81 ev / .
we obtain @xmath82 and @xmath83 , which is in good agreement with the experimental result@xcite of @xmath84 , and @xmath85 .
this is also consistent with the fact that lda tends to overbind in crystals .
we perform elastic constants calculations@xcite to obtain the elastic constants @xmath86 gpa .
these results are in very good agreement with other computational results@xcite where the values of @xmath87 gpa .
the agreement of our results with the experimental results@xcite is rather good except @xmath88 and @xmath89 ( where the experimental values of @xmath90 gpa and @xmath91 gpa ) that may be due to the fact these two values are not directly determined in experiment , as discussed in ref .
along the high - symmetry directions for hexagonal mos@xmath0 , for ( a ) @xmath9 uniaxial , ( b ) @xmath10 uniaxial , ( c ) @xmath8 uniaxial , and ( d ) @xmath11 biaxial strains .
the corresponding densities of grneisen parameters , @xmath92 , are shown on the right .
a mesh of @xmath93 for the @xmath44-point sampling is used to calculate @xmath92 .
, width=302 ] we perform uniaxial deformations in @xmath9 , @xmath10 , and @xmath8 directions with strains set to @xmath94 . for the @xmath11 biaxial deformations , @xmath95 .
the grneisen parameters @xmath96 along the high symmetry directions due to these deformations are shown in fig .
[ fig : gp](a)@xmath97(d ) . to quantify more clearly the distribution of grneisen parameters in the brillouin zone
, we calculate the density of grneisen parameters according to @xmath98 which is shown on the right panels of fig . [
fig : gp ] .
the density of grneisen parameters due to @xmath9 uniaxial , @xmath10 uniaxial , and @xmath11 biaxial strains are also displayed in the inset of fig .
[ fig : gamma](b ) for a direct comparison . from figs .
[ fig : gp](a ) and ( d ) , some negative grneisen parameters are observed near the @xmath99 point , which correspond to the lowest transverse acoustic ( za ) modes
. however , the plots for densities of grneisen parameters , @xmath92 , show most grneisen parameters are populated between a small range of @xmath100 to @xmath101 , and negative grneisen parameters are completely suppressed .
the @xmath102 plots also show that large grneisen parameters , say , @xmath103 , are totally negligible when sampling is taken . from fig .
[ fig : gp](c ) we note that most grneisen parameters are very small for @xmath8 uniaxial deformation , which is consistent with the fact weak van der waals interactions exist between mos@xmath0 trilayers . for the unstrained mos@xmath0 structure .
( b ) the phonon density of states weighted by the grneisen parameters , @xmath5 .
the indices @xmath104 correspond to @xmath9 uniaxial , @xmath10 uniaxial , @xmath8 uniaxial , and @xmath11 biaxial strains , respectively .
the insert shows the densities of grneisen parameters , @xmath105 , for @xmath106 .
, width=302 ] the phonon density of states of the unstrained structure is shown in fig .
[ fig : gamma](a ) where there is a frequency gap from @xmath107 to @xmath108 @xmath109 .
the density of states weighted by the grneisen parameters , @xmath5 , are shown in fig .
[ fig : gamma](b ) , where a gap is inherited from fig .
[ fig : gamma](a ) .
we note that , @xmath110 , which is due to a @xmath8 uniaxial strain , has a broad peak near @xmath111 @xmath109 due to fact these frequencies are associated with more significant grneisen parameters ( see eq .
[ eq : gnu ] ) .
it is interesting to see that , while the density of grneisen parameters due to the @xmath11 biaxial strain , @xmath112 , is quite different from that due to the @xmath9 or @xmath10 uniaxial strains as shown in the inset of fig .
[ fig : gamma](b ) ( or even the average of grneisen parameters due to @xmath9 and @xmath10 uniaxial strains ) , @xmath113 , @xmath114 , and @xmath115 are essentially the same numerically , as shown in fig .
[ fig : gamma](b ) .
this justifies the proposal to use an @xmath11 biaxial strain to replace an @xmath9 uniaxial or @xmath10 uniaxial strain to calculate the integrated quantity @xmath116 . along @xmath117 path .
( b ) shows two doubly degenerate phonon modes for the unstrained structure ( at @xmath99 , the frequencies are @xmath118 and @xmath119 @xmath109 for @xmath120 and @xmath121 , respectively ) .
( c ) , ( d ) , and ( e ) show the detailed variations of the phonon frequencies for the @xmath9 uniaxial , @xmath10 uniaxial , and @xmath11 biaxial strained structures , respectively .
the strain values for uniaxial and biaxial strains are @xmath122 and @xmath123 , respectively .
, width=302 ] we now provide two evidences that explain the difference between grneisen parameters obtained with an @xmath11 biaxial strain and the average of grneisen parameters obtained with @xmath9 and @xmath10 uniaxial strains .
[ fig : ph - ga - brgo - zoom.pdf.pdf ] shows the phonon dispersions of the equilibrium structure and the strained systems along that @xmath124 path .
[ fig : ph - ga - brgo - zoom.pdf.pdf](b)-(e ) we have focused on the change of frequencies around @xmath125 @xmath109 .
it can be seen from figs .
[ fig : ph - ga - brgo - zoom.pdf.pdf](b ) and ( e ) that a @xmath11 biaxial strain can not destroy the degeneracies of two doubly degenerate @xmath120 and @xmath121 phonon modes . however , as seen in fig .
[ fig : ph - ga - brgo - zoom.pdf.pdf](c ) , under an @xmath9 uniaxial deformation , the two doubly degenerate @xmath120 and @xmath121 phonon modes around @xmath125 @xmath109 split into four nondegenerate phonon modes . fig .
[ fig : ph - ga - brgo - zoom.pdf.pdf](d ) shows the same splittings for a @xmath10 uniaxial strain .
this explains the average of grneisen parameters due to @xmath9 and @xmath10 strains is not the same as that due to an @xmath11 biaxial strain .
however , the integrations according to eq .
[ eq : gnu ] associated with @xmath9 uniaxial , @xmath10 uniaxial , and @xmath11 biaxial strains give rise to the same phonon density of states weighted by the gruneisen parameters , a fact which is expected on physical ground .
uniaxial strains .
the chosen paths are ( a ) @xmath126 and @xmath127 , and ( b ) @xmath126 and @xmath128 .
the paths are shown in the insets .
, width=302 ] fig .
[ fig : kgm ] furnishes another evidence that under an @xmath9 uniaxial strain , the planar brillouin zone now has a 4-fold rotation symmetry , in contrast to the 6-fold rotation symmetry for the case of biaxial strain ( where the hexagonal symmetry is preserved ) .
the agreement of grneisen parameters along @xmath129 and @xmath130 paths in fig .
[ fig : kgm](a ) shows that there is a reflection symmetry around @xmath10 axis .
similarly , the agreement of grneisen parameters between that along @xmath131 and @xmath132 paths shows that there is a reflection symmetry around @xmath9 axis . under the 6-fold rotation symmetry of the planar brillouin zone
, we expect grneisen parameters to be the same along @xmath129 and @xmath133 , or along @xmath131 and @xmath134 paths
. however , fig .
[ fig : kgm](b ) shows that there is no agreement of grneisen parameters along @xmath129 and @xmath133 paths , or along @xmath131 and @xmath134 paths . as a function of temperature .
( b ) the linear tecs of mos@xmath0 along the @xmath1 and @xmath2 directions denoted by @xmath135 and @xmath136 , respectively , as a function of temperature .
the volumetric tec is denoted by @xmath137 .
the insert shows the comparison of @xmath135 with the in - plane tec of a single - trilayer mos@xmath0 , obtained using a qha approach@xcite .
, width=302 ] the integrated quantities @xmath41 , @xmath42 , and @xmath116 shown in fig . [ fig : ltec](a ) are essentially identical , and they are much larger than @xmath138 , which is due to a @xmath8 uniaxial strain .
this may be traced to the fact that interactions between mos@xmath0 trilayers are weak compared to in - plane interactions that result in smaller grneisen parameters ( i.e. , smaller frequency changes ) for @xmath8 strains .
the linear tecs in @xmath1 and @xmath2 directions are shown in fig .
[ fig : ltec](b ) . even though @xmath139 is about three times smaller than @xmath66 for , say , @xmath140 k , @xmath141 is larger than @xmath142 . at @xmath143
k , @xmath144 k@xmath145 , @xmath146 k@xmath145 , and the volumetric tec @xmath147 k@xmath145 .
the main reason for this is that the value of @xmath148 ( @xmath149 gpa ) is much smaller than the value of @xmath150 ( @xmath151 gpa ) , therefore , according to eq . [ eq:2d1 ] it is possible that @xmath136 is larger than @xmath135 .
the result that @xmath136 is indeed larger than @xmath135 is consistent with the physical fact that it is easier to perform a deformation in the @xmath8 direction than in the in - plane direction , which is again , attributed by the weak interactions in the out - of - plane direction .
this is also confirmed by the results shown in the inset of fig .
[ fig : ltec](b ) where there is a striking similarity between the temperature dependence of in - plane tecs for both the bulk mos@xmath0 and a single - trilayer mos@xmath0 , which is obtained from a qha - lda treatment@xcite .
el - mahalawy and evans@xcite measured the linear tecs of 2h - mos@xmath0 between @xmath152 and @xmath153 k and reported @xmath154 k@xmath145 and @xmath155 k@xmath145 , which are consistently lower than our values .
the same group@xcite again measured the tec of 2h - mos@xmath0 between @xmath156 and @xmath157 k and found larger @xmath158 k@xmath145 , which brings a better agreement with our result , and @xmath159 k@xmath145 , which is somewhat larger .
we find the behavior of the rate of change of @xmath136 with @xmath3 to be different from that of @xmath135 where the former increases rapidly at low @xmath3 while the latter increases more gradually with @xmath3 .
et al._@xcite pointed out that the lattice constant @xmath1 ( @xmath2 ) increases linearly ( nonlinearly ) with temperature , which is consistent with our tec results at low @xmath3 .
in summary , we have proposed a direct way to calculate the linear thermal expansion coefficients ( tecs ) of a hexagonal system based on the grneisen formalism .
we have also proposed a way to replace the inefficient symmetry - lowering uniaxial strains by the efficient symmetry - preserving biaxial strains .
we successfully implemented the computational schemes and applied them to a technologically important material mos@xmath0 .
we found that mos@xmath0 has a large tec anisotropy where the thermal expansion coefficient in the @xmath2 direction is @xmath13 times larger than that in the @xmath1 direction at high temperatures .
we highlighted that even though the integrated quantities @xmath4 required by the tec calculations can be obtained via a symmetry - preserving biaxial strain , the grneisen parameters from a biaxial strain may not be a simple average of the grneisen parameters from uniaxial @xmath9 and @xmath10 strains .
we demonstrated that we only need a minimum of two symmetry - preserving deformations to directly calculate the tecs of a general hexagonal system .
in contrast , the quasiharmonic approximation , when dealing with a two - parameter system , may require an expensive search in the two - dimensional search space .
therefore , we expect the strategies adopted in this paper to treat a general two lattice - parameter hexagonal system can be similarly applied to treat other two lattice - parameter systems such as trigonal or tetragonal systems , thus opening the door for a truly predictive tec calculations for many important materials .
we also expect the tec calculations based on the gruneisen formalism via symmetry - preserving deformations may be readily incorporated in any phonon related codes such as phonopy@xcite .
we acknowledge stimulating discussions with ching hua lee .
we thank the national supercomputing computer , singapore ( https://www.nscc.sg ) for computing resources .
y.l . acknowledges support from the singapore national science scholarship . | using density - functional perturbation theory and the grneisen formalism , we directly calculate the linear thermal expansion coefficients ( tecs ) of a hexagonal bulk system mos@xmath0 in the crystallographic @xmath1 and @xmath2 directions .
the tec calculation depends critically on the evaluation of a temperature ( @xmath3 ) dependent quantity @xmath4 , which is the integral of the product of heat capacity and @xmath5 , of frequency @xmath6 and strain type @xmath7 , where @xmath5 is the phonon density of states weighted by the grneisen parameters .
we show that to determine the linear tecs we may use minimally two uniaxial strains in the @xmath8 , and either @xmath9 or @xmath10 direction .
however , a uniaxial strain in either @xmath9 and @xmath10 direction drastically reduces the symmetry of the crystal from a hexagonal one to a base - centered orthorhombic one .
we propose to use an efficient and accurate symmetry - preserving biaxial strain in the @xmath11 plane to derive the same result for @xmath12 .
we highlight that the grneisen parameter associated with a biaxial strain may not be the same as the average of grneisen parameters associated with two separate uniaxial strains in the @xmath9 and @xmath10 directions due to possible preservation of degeneracies of the phonon modes under a biaxial deformation .
large anisotropy of tecs is observed where the linear tec in the @xmath2 direction is about @xmath13 times larger than that in the @xmath1 or @xmath14 direction at high temperatures .
our theoretical tec results are compared with experiment . the symmetry - preserving approach adopted here
may be applied to a broad class of two lattice - parameter systems such as hexagonal , trigonal , and tetragonal systems , which allows many complicated systems to be treated on a first - principles level . |
clusters of galaxies are the largest gravitationally bound and collapsed systems in the universe .
they are filled with x - ray emitted hot gas with the temperature of @xmath010 kev .
high - resolution x - ray observations have revealed that the hot gas in many cluster cores is not smoothly distributed ( e.g. * ? ? ?
* ; * ? ? ?
, we define the cluster cores as the dense regions within @xmath1300 kpc from the cluster centers .
the complicated structures are often attributed to activities of active galactic nuclei ( agns ) .
in fact , bubbles of high energy particles have been found in some clusters ( e.g. * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , which is the evidence that agns affect the surrounding hot gas .
however , the complicated structures have also been observed in clusters in which agns are not active at the centers ( e.g. * ? ? ? * ) . in the outer regions of the cores ,
edge - shaped discontinuities in the gas density and temperature ( ` cold fronts ' ) are often found ( e.g. * ? ? ?
they may be attributed to the ` sloshing ' of the gas in the cluster gravitational potential well , although the origin remains an open question @xcite . from the above observations
, one may think of complex gas motion , that is , turbulence in the cores .
actually , turbulence is expected to be prevailing in cluster cores .
although it has not yet directly been observed in x - rays ( but see * ? ? ?
* ) , cold gas ( @xmath2 k ) moving with the velocity of 100@xmath3 has been observed .
if the ambient x - ray gas did not move with the cold gas , the latter would immediately mix with the former @xcite .
moreover , the turbulence may be playing an important role in heating of the cluster cores .
it had been expected that the gas in cluster cores cools by radiating away its thermal energy , which induces gas motion toward the cluster centers ( cooling flows ; * ? ? ?
however , x - ray spectra have revealed that the mass of the cooling gas is far smaller than that expected from this model @xcite .
the turbulence may transport thermal energy from the outside of the cluster cores , and may balance the radiative cooling of the cores @xcite .
however , the actual mechanism that creates the turbulence has not been understood .
the scenario we newly propose here is that the core turbulence is created by bulk gas motions , which are naturally produced in the x - ray hot gas in clusters .
cosmological numerical simulations have shown that clusters are knots of larger - scale filaments in the universe @xcite . along the direction of the filaments , small clusters ( or galaxies ) successively fall into a cluster , which increases the cluster mass .
the simulations have also shown that the infall velocities of the smaller clusters are more than @xmath3 @xcite .
the interactions of the hot gas and dark matter between the clusters should produce a large - scale gas motion ; the velocity is typically @xmath4 and sometimes reaches @xmath5 @xcite .
the cluster cores are exposed to these violent gas motions ; this would create possible turbulence and the complicated structures observed in the cores even if there are no agn activities . in order to test this scenario ,
we performed two - dimensional high - resolution hydrodynamic simulations to follow the long - term evolution of the core in ` a stormy cluster ' for the first time .
since we are interested in the cluster core and we need resolution high enough to resolve turbulence , we limited the calculations to the central region of a cluster ( within @xmath6 kpc from the cluster center ) .
these simulations were performed using a nested grid code @xcite , and the coordinates are represented by @xmath7 . while the resolution near the outer boundary is @xmath8 kpc , that at the cluster center is @xmath9 pc .
free boundaries were chosen .
thermal conduction , viscosity , magnetic fields , and the self - gravity of gas were ignored .
radiative cooling is included .
we adopted a cooling function for the metal abundance of 0.3 solar .
the gas is isothermal and the temperature is 7 kev at @xmath10 . the fixed gravitational potential and the initial gas distribution are the same as those in @xcite .
we approximated the bulk gas motions in a cluster by plane wave - like velocity perturbations represented by @xmath11 at @xmath12 kpc , where @xmath13 is the initial sound velocity , and @xmath14 is the wave length .
this assumption comes from the idea that continuous infall of matter and small clusters along a large - scale filament should create velocity perturbations in a particular direction .
the factor @xmath15 is a free parameter and @xmath16 is the cluster center .
we studied the perturbations with @xmath15 and @xmath14 shown in table 1 .
note that cosmological numerical simulations suggested that the velocity of the hot gas is about 0.2@xmath17 or larger even when a cluster is relatively relaxed @xcite .
moreover , it is expected that the scale of bulk gas motion in a cluster is @xmath18 kpc or larger @xcite . therefore , we think that the parameters we took are reasonable . however , since the large bulk motions in a cluster do not always generate large amplitude acoustic - gravity waves , our model may be appropriate for clusters undergoing hierarchical structure formation or similar violent events . since the energy input rate through waves per unit area is given by @xmath19 , where @xmath20 is the gas density , the integrated input rate for @xmath21 kpc at @xmath12 kpc is @xmath22 for our model cluster .
on the other hand , at @xmath10 , the x - ray luminosity of the gas for @xmath23 kpc and @xmath24 kpc is @xmath25 and is comparable to the wave energy input rate .
in figures [ fig1 ] and [ fig2 ] , we present the temperature distributions for @xmath26 and @xmath27 kpc ; the velocity perturbations propagate upwards as acoustic - gravity waves , which were called ` tsunamis ' in @xcite .
since the velocity amplitude is relatively large , the waves steepen and become weak shocks as shown in one - dimensional simulations @xcite . because of the pressure coming from the momentum of the waves , the coolest and densest gas noticeably shifts from the cluster center at @xmath28 gyr and the shift
is clearly seen at @xmath29 gyr ( fig .
[ fig1 ] ) . while the position of the coolest gas is at @xmath30 kpc ( fig .
[ fig1 ] ) , that of the densest gas is @xmath31 kpc .
the difference of the positions is made by the wave just passed ( fig .
[ fig1 ] ) ; the gas slightly deviates from pressure equilibrium there . comparing the result of @xmath26 with that of @xmath32
, we found that the gravitational energy of the gas for the former is larger than that for the latter by @xmath33 ergs for @xmath21 kpc and @xmath34 kpc at @xmath29 gyr .
the energy increase is smaller than the energy input by waves during the first one gyr for @xmath21 kpc ( @xmath35 ergs ) , suggesting that a part of the injected energy is used to produce the shift of dense gas .
after that , radiative cooling proceeds and the core region becomes cooler and denser . since waves no longer sustain the cooling core , the core falls in the potential well of the cluster . during the fall ,
rayleigh - taylor ( rt ) and kelvin - helmholtz ( kh ) instabilities develop around the core .
they non - linearly develop , and turbulent motion is eventually formed in and around the core as seen in figure [ fig2 ] , which shows the temperature distribution at @xmath36 gyr .
the cool core oscillates around the cluster center , and temperature jumps are formed ( fig .
[ fig2]a ) . associated with these temperature jumps ,
gas density also has discontinuities .
small cool and dense blobs randomly moving in the core cause new rt and kh instabilities around them , and smaller eddies are generated ( fig .
[ fig2]b ) . for the models of @xmath26 and 0.5
, the turbulence is maintained until our calculations are stopped at @xmath37 gyr .
it is interesting that the turbulence presented here is naturally caused even by the regular , wave - like , linear perturbations .
no irregular triggers , which are usually essential for initiating or maintaining turbulence , are necessary . in table 1
, we present the time when the gas temperature in any of the numerical grid points reaches zero ( @xmath38 ) .
the gas cooling is suppressed by heat transport through turbulent mixing , especially when @xmath15 is large . in other words ,
in addition to the wave input energy ( see [ sec : models ] ) , the thermal energy transported from the outside of the core is used to prevent the core from cooling .
we note that although there were several studies treating core heating by acoustic - gravity waves , the mechanism of the heating presented here is completely different from that proposed in the previous studies . in those previous studies , based on the analytical approach @xcite or one - dimensional numerical simulations @xcite ,
it was predicted that weak shocks evolved from acoustic - gravity waves directly heat the cluster core , while in this study the turbulence is the major player of the heating . since our simulations are multi - dimensional , the results should be much more realistic than those of the previous studies .
in fact , since turbulence is essentially multi - dimensional , those previous studies were not able to directly treat the turbulence and complicated structures of the cores . even with multi - dimensional simulations , the turbulent heating could not be found if the resolution were not high enough .
the present results show that the turbulence starts to develop after the core becomes dense through cooling . before that
, waves pass the cluster center without much changing the gas structure .
thus , we predict that this mechanism of turbulence generation works only for clusters with dense and cool cores , which had been called ` cooling flow clusters ' .
since the turbulence is spatially limited to the cool core , it does not totally mix the hot gas in a cluster .
thus , as long as violent mergers of clusters with comparable masses , which completely destroy the central gas structures of the clusters , do not happen , the metal abundance excess observed in cluster cores ( e.g. * ? ? ?
* ) would not completely be erased .
because of the turbulent motion , the fine structures of the core are not steady .
our simulation results sometimes show fine structures similar to the peculiar structures observed in clusters such as a1795 , centaurus , and 2a @xmath39 @xcite .
the temperature jumps seen in figure [ fig2]a may correspond to the ` cold fronts ' observed in some clusters .
the bulk gas motion in clusters would result in the formation of acoustic - gravity waves and the weak shocks as shown above .
direct observations of the waves may be difficult unless the wave fronts are almost parallel to the line of sight .
in a133 , however , a weak shock just passing through the core has been observed @xcite .
the maximum velocity of the turbulence in a cluster core is @xmath40 for @xmath26 . with a high spectral resolution detector like the x - ray satellite _ astro - e2 _
, the turbulence in cluster cores could be detected in the near future .
if turbulence is being developed , the metal lines in the x - ray spectra would have very complicated features owing to the gas motion @xcite . on the other hand , turbulence
could also be created by agn activities , especially by the buoyant motion of agn - origin bubbles .
the lifetime of the eddies associated with the bubble motion is @xmath41 , where @xmath42 and @xmath43 are the size and velocity of a bubble , respectively . for the virgo cluster , for example ,
the size of the observed bubbles is @xmath44 kpc , and the predicted velocity of a bubble is @xmath45 @xcite .
thus , the lifetime of the eddies is @xmath46 yr , which is much shorter than the lifetime of the bubble itself ( @xmath47 yr ; * ? ? ?
this means that turbulence is unlikely to exist in a cluster core without agn - origin bubbles .
thus , if turbulence is detected in such a core , it could be associated with the bulk gas motions outside of the core .
our model predicts that cooling of a cluster core is more suppressed for larger velocity amplitude .
the suppression should also work in smaller objects such as groups of galaxies and elliptical galaxies because their overall structures are similar to those of clusters of galaxies @xcite , and we expect that the bulk gas motions and turbulence are excited in them by the same mechanism presented here .
this is in contrast with the suppression by thermal conduction that does not work in the smaller objects because of their low temperatures @xcite . finally , we note the limitations of our model .
first , we approximated the bulk gas motions by waves with constant amplitude and length , and direction for propagation is also assumed to be fixed , but in reality all these conditions should be changed during formation and evolution of clusters . even bulk gas motions that are not represented by regular waves considered here could cause the motion of the cool cores and produce turbulence through rh and kt instabilities . moreover , it is likely that waves come from several directions corresponding to large - scale filaments .
these effects should be ultimately studied by fully three - dimensional , ultra - high - resolution cosmological simulations , although the basic mechanism initiating the turbulence would not be different from the one in the two - dimensional case here .
second , we did not include the heating by turbulent dissipation , agns , and thermal conduction .
if they are effective as thermal energy sources , the wave energy required to suppress the cooling of a core would be smaller than that we predicted .
note that the turbulence may tangle magnetic fields and reduce an effective conduction rate , however , @xcite indicated that chaotic magnetic fields do not much reduce the conduction rate .
thus , we may need to follow the evolution of magnetic field lines when we include the effect of thermal conduction .
third , we fixed the gravitational potential of a cluster . a change in the potential
may also shift the cool gas core from the gravitational center of the cluster , which may lead to the motion of the core and the development of turbulence @xcite .
fourth , we assumed a two - dimensional axi - symmetric geometry .
this might affect development and structure of turbulence .
it would be expected that cascade processes of eddies is different in full three - dimensional turbulence .
interaction between the turbulence and plane - waves in three - dimensions should be also clarified in future simulations .
we thank the anonymous referee for useful suggestions .
the authors were supported in part by a grant - in - aid from the ministry of education , culture , sports , science , and technology of japan ( y. f. : 14740175 ; t. m. : 14740134 ; k. w. : 15684003 ) .
all simulations were run on fujitsu vpp5000 at naoj . | based on high - resolution two - dimensional hydrodynamic simulations , we show that the bulk gas motions in a cluster of galaxies , which are naturally expected during the process of hierarchical structure formation of the universe , have a serous impact on the core .
we found that the bulk gas motions represented by acoustic - gravity waves create local but strong turbulence , which reproduces the complicated x - ray structures recently observed in cluster cores .
moreover , if the wave amplitude is large enough , they can suppress the radiative cooling of the cores .
contrary to the previous studies , the heating is operated by the turbulence , not weak shocks .
the turbulence could be detected in near - future space x - ray missions such as _ astro - e2_. |
it is well known that gamma - ray burst ( grb ) afterglow decays as a power - law @xmath1 .
the temporal decay index @xmath2 , together with the spectral index , provides us precious information about grb jets and their environment . in the pre - swift era
afterglow observations start typically a few hours after a burst . in such a late phase ,
the decay index is insensitive to the choice of the reference time @xmath0 , and grb trigger time is often used in afterglow modelings .
the multi - wavelength observatory swift was launched in nov 2004 .
thanks to its fast pointing capabilities swift is disclosing the early afterglow phase .
one of unexpected finds by swift is that early x - ray afterglows show a canonical behavior , where light curves include three components : ( 1 ) a steep decay component , ( 2 ) a shallow decay component and ( 3 ) a `` normal '' decay component . on top of this canonical behavior , many events have superimposed x - ray flares ( zhang et al . 2006a ; nousek et al .
2006 ; chincarini et al 2005 ; obrien et al . 2006 ) .
the transition from the early steep decay to the shallow decay typically occurs at several hundred seconds , and the timescale is comparable to the duration of rather long grbs .
when discussing the early afterglow and its connection to the prompt emission component , the decay index is very sensitive to the reference time @xmath0 one chooses . correctly choosing @xmath0 is therefore essential to derive the right index as well as to interpret each component in the canonical light curve ( piro et al .
2005 ; tagliaferri et al . 2005 ; quimby et al .
2006 ) .
tagliaferri et al .
( 2005 ) investigated the first two bursts grb 050126 and grb 050219a which have an x - ray light curve well sampled by the x - ray telescope on board swift .
they sought for a possible delay of the afterglow onset by fitting the early x - ray light curves ( the components 1 and 2 which we have discussed above ) with a single power - law model . in both cases ,
the decaying light curves can be fitted if the onset of the afterglow is shifted to @xmath3 s after the burst trigger with a single power - law . however , while in the case of grb 050126 the light curve does not allow us to clearly state whether a broken power - law modeling is better than a single power - law model , for grb 050219a a broken power - law definitively provides a better fit . in the standard grb model
the time shift between the grb trigger and the reference time @xmath0 is expected to be `` small '' .
the early steep decay should not be an artifact due to a wrong choice of @xmath0 .
lazzati and begelman ( 2006 ) studied forward shock emission , based on a simple energy injection model .
their numerical light curves show that measuring times from the beginning of the prompt phase is a good approximation .
the early steepening and x - ray flares are likely to be produced by another mechanism ( e.g. internal shocks ; burrows et al . 2005 ;
falcone et al .
2006 ; zhang et al . 2006a ; nousek et al . 2006
; ioka , kobayashi & zhang 2005 ; fan & wei 2005 ) .
recently long - lasting soft emission is reported in a short burst grb 050724 ( barthelmy et al .
such a soft component was hinted at in the sum of multiple short batste grbs , and it might be the onset of short burst afterglow ( lazzati , ramirez - ruiz & ghisellini 2001 ) .
it is therefore of interest to quantitatively examine the @xmath0 issue . in this paper
, we study the physics and timescales involved in early afterglow stage , and give more direct and clear arguments for the afterglow slopes . in
2 we study a simple analytic model . in
3 we evaluate light curves of reverse shock emission as well as that of forward shock emission , based on full hydrodynamic calculations . in
4 we address how inhomogeneity of a fireball affects an early afterglow light curve .
conclusions and discussion are given in 5 .
let @xmath4 be a radius of a forward shock expanding with a lorentz factor @xmath5 into homogeneous ambient medium .
since the shock is moving toward us at almost the speed of light @xmath6 , the difference of the the observed times between a photon emitted at @xmath4 and another emitted at @xmath7 is @xmath8 .
although the origin of observed time is arbitrary , a natural definition of observed time is given by the delay of photons emitted from a shock front at a lab time @xmath9 with respect to the photons emitted from the `` explosion '' at @xmath10 and @xmath11 .
the dashed line in figure [ fig1 ] depicts the trajectory of the photon from the explosion . in this paper ,
@xmath9 and @xmath12 denote the lab and observed time since the explosion , respectively .
distance , velocity and the corresponding lorentz factors are measured in the lab frame .
thermodynamic quantities ( pressure and density ) are measured in the local fluid frame .
the observed time is given by @xmath13 the shock radius is almost proportional to the lab time , @xmath14 . considering that internal shocks occur and produce gamma - rays at radii much smaller than the deceleration radius of the fireball , the prompt gamma - rays associated with the outermost element of the fireball should propagate practically on the dashed line in figure [ fig1 ] .
the grb trigger almost coincides with the explosion @xmath15 .
the difference is small , and order of the variability timescale of the prompt emission , because the variability timescale directly reflects the inhomogeneity scale in a fireball ( kobayashi , piran & sari 1997 ) .
the dashed line in figure [ fig1 ] will be called the gamma - ray front in the following sections . a fireball with an initial lorentz factor @xmath16 decelerates when it collects a large volume of ambient material with a mass density @xmath17 .
equalizing energy of the shocked ambient material @xmath18 and fireball energy @xmath19 , we obtain the deceleration radius @xmath20 where @xmath21 is the sedov length .
since the fireball density decreases as it expands , there is a possibility that a reverse shock evolves from newtonian to relativistic during the propagation .
a relativistic reverse shock reduces considerably the lorentz factor of the fireball material which it crosses . in such a case ,
the energy of the shocked ambient medium is still negligible at @xmath22 , because we have assumed @xmath23 at the deceleration to estimate @xmath24 .
the reverse shock crosses the fireball shell at @xmath25 and all the fireball material decelerates where @xmath26 is a fireball shell width ( see sari & piran 1995 and kobayashi , piran & sari 1999 for the details ) . in summary , the evolution of fireballs are classified into two cases depending on the value of @xmath26 relative to a critical value @xmath27 ( sari & piran 1995 ) . if @xmath26 is smaller than the critical value , it is called the thin shell case . , and it is longer than the duration of the prompt emission @xmath28 ( sari 1997 ) . the deceleration time @xmath29 approaches @xmath28 if we take a larger @xmath16 , and around @xmath30 ( or equivalently @xmath31 ) a reverse shock becomes relativistic during the propagation .
the lorentz factor of the shocked material at the crossing time becomes independent from the initial value , and it is given by @xmath32 .
the deceleration observed time is about @xmath33 . ]
the reverse shock is always in the newtonian regime , and it is too weak to slow down the fireball effectively .
the deceleration radius is @xmath34 .
if @xmath35 ( the thick shell case ) , we define the deceleration radius as the shock crossing radius @xmath36 .
the deceleration lab time is given by @xmath37 . if the blandford - mckee ( bm ) blast wave scaling @xmath38 ( blandford and mckee 1977 ) was valid for the whole fireball evolution ( the thin solid line in figure [ fig1 ] ) , integrating eq .
( [ eq : obstime ] ) , we obtain @xmath39 .
since the spectral characteristics of forward shock synchrotron emission are given by products of @xmath12 , @xmath40 , @xmath41 and constant parameters ( sari , piran & narayan 1998 ) , the light curve should be described by a power - law with the reference time @xmath42 .
however , the bm blast wave scaling is applicable only after the fireball is decelerated ( e.g. kobayashi , piran & sari 1999 ) . at earlier times the fireball ( the thick solid line )
is in the coasting phase , and it is slower than evaluated from the bm blast wave scaling .
the delay of photons from the shock front at the deceleration lab time is larger by @xmath43 than in the case that the bm blast wave scaling is applicable to the whole evolution ( see figure [ fig1 ] ) where @xmath44 and @xmath45 are the separations at the deceleration lab time between the gamma - ray front ( dashed line ) and the reference bm blast wave ( the thin solid line ) , and between the gamma - ray front and the fireball forward shock ( the thick solid line ) , respectively .
the reference time for the afterglow modeling is given by @xmath46 in the thick shell case , a forward shock keeps being energized for a longer time , and the deceleration phase starts at a later time .
it still overestimates the shift of the reference time if @xmath0 is set at the end of the energy injection or equivalently at the peak time of afterglow . in figure [ fig2 ]
the dot indicates the point at which the energy injection stops and the fireball turns into a bm solution . measuring times from the peak of afterglow corresponds to defining a reference null geodesic ( the dashed line ) as it goes through the dot .
imagine the reference bm blast wave ( the thin solid line ) associated with the new reference null geodesic .
the decay of the emission from this reference blast wave should be characterized by a power - law using the new reference time ( measuring times from the peak ) .
the evolution of a real fireball ( the thick solid line ) is also described by a bm blast wave after the deceleration .
however , it is a different solution ( a different curve on the diagram ) . applying the new reference time to
the afterglow modeling leads to a wrong estimate of the decay index especially in the early phase .
the decay index should become shallower than the real value .
using eq.([eq : obstime ] ) and the evolution of a bm blast wave @xmath47 , we obtain @xmath48 where @xmath49 is the lorentz factor at the deceleration radius @xmath24 .
since the lorentz factor of a fireball is constant @xmath50 at @xmath51 , we get @xmath52 for the thin shell . in the thick shell case ,
a reverse shock becomes relativistic before it crosses the fireball shell , and it begins to reduce considerably the lorentz factor of the shell s matter which it crosses .
the lorentz factor of the forward shock is also significantly reduced as @xmath53 during the reverse shock crossing ( sari & piran 1995 ; kobayashi , piran & sari 1999 ) , and we get @xmath54 for the thick shell . the reference time is given by @xmath55 where @xmath56 is the deceleration observed time .
we have considered a simple broken power - law model for the evolution of a fireball lorentz factor , and we obtained eq .
( [ eq : t0 ] ) . in reality
the fireball should decelerate gradually around the deceleration radius .
an artificial early steepening could happen only when @xmath57 , and in such a case @xmath0 should be set at @xmath58 ( the afterglow peak time ) to get the correct @xmath2 .
however , in both of the thin shell and thick shell cases , we have found that @xmath44 and @xmath45 are comparable .
therefore , measuring times from the beginning of the prompt phase @xmath42 should not induce an overestimate of afterglow decay right after the peak .
( [ eq : obstime ] ) gives the delay of photons emitted from a point on the shock front on the line of sight , while most photons suffer longer delays , since they are emitted from a shocked region of finite thickness behind the shock , and from positions off the line of sight .
although these effects could make both separations @xmath44 and @xmath45 larger by a factor of a few ( waxman 1997 ) , they are still comparable and our arguments are valid .
we employ a spherical relativistic lagrangian code based on the godunov method with an exact riemann solver to evaluate the hydrodynamic evolution of a relativistic fireball ( kobayashi et al .
1999 ; kobayashi & sari 2000 ) . using the einstein summation convention the equations describing the motion of a relativistic fluid
are given by the five conservation laws @xmath59 where @xmath60 is four velocity and @xmath61 is the stress - energy tensor ( @xmath62 ) , which for a perfect fluid can be written as @xmath63 . here , @xmath64 is the metric tensor , @xmath65 the fluid pressure , and @xmath66 the heat function per unit volume .
shocked material is extremely hot and pressure is related to internal energy @xmath67 and mass density @xmath68 as @xmath69 . in the godunov scheme ,
conservative variables are considered as piecewise constants over the mesh cells at each time step and the time evolution is determined by the solution of the riemann problem ( shock tube ) at the inter - cell boundaries ( e.g. mart & mller 1999 and references therein ) .
because of the relativistic beaming effect , the radiation from a jet before the jet break can be described by a spherical model with an isotropic energy .
the initial configuration for our simulation is a static uniform fireball surrounded by uniform cold ambient material ( ism ) .
it is determined by four parameters : an isotropic energy @xmath19 , a dimensionless entropy @xmath16 , an initial radius @xmath70 and the ism mass density @xmath17 .
@xmath19 and @xmath17 always appear as the ration @xmath71 in the hydrodynamics computation , the system is actually determined by three parameters , the initial radius @xmath70 , the entropy @xmath16 and the sedov length @xmath72 . for convenience ,
we set the initial lab time as @xmath73 rather than zero in numerical calculations , and the observed time is determined by the delay of photons from an emitter with respect to the photons emitted from the initial fireball surface at @xmath74 and @xmath75 . since the fireball immediately accelerates to a relativistic velocity , this is practically equivalent to the observed time used in figure [ fig1 ] .
we carry out numerical calculations for the total number and [ thicklc ] for @xmath76 . in the log - log space ,
the resulting light curves are almost identical .
the differences are less than several percent .
when we remove numerical oscillations , they are in a few percent agreement . ] of mesh cells @xmath77 .
a third of the cells are for the fireball , while the other cells describe the ambient medium within @xmath78 cm .
although the initial configuration is @xmath79 cells per decade , the nature of lagrangian method gives a much higher resolution for shocks which sweeps cells and compresses them by a factor of @xmath80 .
a reflection boundary and free boundary condition are imposed at the center @xmath10 and at @xmath81 , respectively .
we consider two cases the thin and thick shell cases .
we first consider the thin shell case : @xmath82 ergs , @xmath83 , @xmath84 cm , and @xmath85 @xmath86 @xmath87 where @xmath86 is the proton mass .
initially , as the fireball expands into a surrounding medium , a narrow shell with a radial width @xmath88 is formed .
the lorentz factor of the shell increases linearly with the radius during the free acceleration stage .
then , the fireball shell uses up all its internal energy , and it coasts with the lorentz factor of @xmath16 .
the coasting ends once the ism begins to influence the shell . after the deceleration radius @xmath24 ,
the profile of the shocked ism medium begins to approach the bm solution .
the evolution of a fireball is fully discussed by kobayashi et al .
( 1999 ) .
the interaction between the shell and the ism is described by two shocks : a forward shock propagating into the ism and a reverse shock propagating into the shell .
figure [ profnrs ] shows the propagation of the shocks .
initially , the unshocked fireball shell has all the energy of the system .
as the shell expands , the reverse shock decelerates the ejecta while the forward shock accelerates the ism .
the energy is transfered from the unshocked shell to the ism via the shocks , finally the shocked ism carries essentially all the energy of the system . in the intermediate stage , around the deceleration time @xmath89 sec , the shocked shell has comparable energy to the shocked ism .
the evolution of the energies in three regions , inside of the reverse shock ( unshocked shell ) , between the reverse shock and the contact discontinuity ( shocked shell ) , and between the contact discontinuity and the forward shock ( shocked ism ) are shown in figure [ energy ] .
the observed times of photons from the reverse shock front , the contact discontinuity and the forward shock front are used to describe the evolution of the energies . even after the deceleration time
, the reverse shocked shell ( the thin dashed - dotted line ) carries a significant fraction of the system energy for a long time , @xmath90 at @xmath91 and @xmath92 at @xmath93 . in the thin shell case ,
the reverse shock is newtonian or sub - relativistic in the frame of the unshocked shell ( the deceleration by the reverse shock is not significant as we can see in the top panel of figure [ profnrs ] ) .
it does not heat the shocked region well .
the reverse shocked region is already cold at the shock crossing time ( the deceleration time ) , and the shocked shell carries the energy mainly in the form of the kinetic energy @xmath94 . assuming a power - law decay @xmath95
, we obtain @xmath96 for @xmath97 ( kobayashi & sari 2000 ) .
this slow energy transfer should lead to the round - off of an afterglow peak . to evaluate the afterglow light curve ,
we consider here a simple case in which the energy of the magnetic field remains a constant fraction of the internal energy @xmath98 .
the electron random lorentz factor evolves as @xmath99 after the shock heating where @xmath65 and @xmath68 are the pressure and density of a fluid element .
the typical synchrotron frequency in the observer frame is @xmath100 .
since a relativistic shock totally ionizes material which it crosses , the total number of electrons in a fluid element ( mesh cell ) is @xmath101 where @xmath102 is a cell width in the lab frame . the spectral power at the typical frequency from a cell is given by @xmath103 .
assuming a power - law distribution of the electron random lorentz factor with index @xmath104 , the observed flux is @xmath105 where we have assumed that the observational band is located between the typical frequency @xmath106 and the cooling frequency @xmath107 .
in early afterglow phase , with the typical parameters , x - ray and optical band satisfy this condition and @xmath108 , the numerical results shown in this paper actually satisfy this condition in the time ranges of the plots ( figs [ thinlc ] and [ thicklc ] ) where @xmath109 and @xmath110 are the fractions of shock energy given to magnetic field and electrons at a shock , respectively . choosing a larger @xmath110 could produce an additional break at @xmath111s especially in x - ray light curves due to the passage of the forward - shock cooling frequency in x - ray band .
however , the steepening is small @xmath112=1/4 and it happens well after the deceleration of the fireball .
the cooling break does not affect our discussions about the onset of afterglows . ] for the forward shock and reverse shock emission , respectively .
we evaluate the flux and observed time of photons from each shocked fluid element , and superimpose the emission to construct the forward and reverse shock light curves .
the results are shown in figure [ thinlc ] .
although at later times the light curves approach the expected power - law decays @xmath113 ( forward shock ; the thick solid line ) and @xmath114 ( reverse shock ; the thin solid line ) , the peaks are rounded as we expected . at the deceleration time ( and after that ) , the light crossing time of the forward shocked region is comparable to the observed time @xmath115 . to evaluate the effect of the thickness of the forward shocked region
, we assume that all the shocked electrons emit photons at the forward shock front .
the resulting light curve peaks at an earlier time is obtained based on a simple model , and it could have an error of a factor of a few . ] @xmath116 , but the shape around the peak itself is similar . we conclude that the round - off of the afterglow peak is mainly due to the slow energy transfer from a fireball shell to a blast wave .
the internal shock model requires a highly irregular outflow from the grb central engine .
since the hydrodynamic interaction inside the flow smooths the velocity and pressure profiles , but not the density profile , fireball ejecta might have an irregular density profile at the deceleration time .
emission from the ejecta during a reverse shock crossing could reflect the light curve of the prompt emission produced by internal shocks ( nakar & piran 2004 ) .
the peak time of the reverse shock emission also depends on the density profile .
as a fireball expands , a reverse shock accelerates more and more electrons in the fireball while the lorentz factor and pressure of the shocked region decreases .
the balance determines the peak time . in the thin shell case
, a reverse shock does not effectively decelerate the shell material that it crosses .
this leads to large pressure gradient across the shocked region .
the contribution of emission from the inner part of the shell becomes negligible , and the reverse shock emission peaks slightly before the shock crossing time ( the thin solid line ) .
we consider the thick shell case in which a reverse shock becomes relativistic during the propagation .
the initial condition is @xmath117 ergs , @xmath118 , @xmath84 cm , and @xmath119 @xmath86 @xmath87 .
the fireball is decelerated around @xmath120 cm , which corresponds to the shock crossing time @xmath121 sec .
the energy transfer from the fireball to a blast wave is similar to that described in the thin shell case . at the shock crossing time , the forward - shocked ism and reverse - shocked shell have comparable energies ( the thick lines in figure [ energy ] ) .
the main difference is that in the thick shell case the shocked shell carries the energy in the form of the internal energy , instead of the kinetic energy , because the reverse shock significantly decelerate the shell .
figure [ profrrs ] depicts the profiles of @xmath122 , @xmath68 and @xmath65 .
when the reverse shock crosses the shell with a width of @xmath123 , a rarefaction wave begins to propagate towards the contact discontinuity , and it quickly transfers the shell s internal energy to the shocked ism in a timescale comparable to the deceleration time ( the shock crossing time )
. the steep decay of the shocked shell energy right after the peak ( the thick dashed - dotted line ) is due to the rarefaction wave propagation . since for our parameters
the reverse shock is mildly relativistic in the frame of the unshocked shell @xmath124 several ( sari & piran 1995 ) , the shocked shell becomes cold after the propagation , and the energy of the shocked shell begins to decay in the same rate as the thin shell does . in the thick shell case ,
the energy of a shocked shell is swiftly relayed to a forward shock by a rarefaction wave .
a broken power - law @xmath125 @xmath126 and @xmath127 @xmath128 describes the evolution of the forward shock well .
the round - off of an afterglow peak is expected to be small .
( [ eq : fnu ] ) and numerical results , we evaluate the light curves of the forward shock and reverse shock emission , which are shown in figure [ thicklc ] .
initially the luminosity of the both shocked region slowly increases as @xmath129 ( kobayashi 2000 ) , and they peak around the shock crossing time .
the forward shock light curve ( the thick solid curve ) is slightly steeper right after the peak compared to at late times .
although the decay is described well by a single power law if the reference time @xmath0 is set at the middle between the grb trigger and the afterglow peak as we discussed in section 2 , measuring times from the grb trigger also provides a reasonable estimate for the afterglow decay .
the reverse shock emission ( the thick dashed curve ) drops sharply @xmath130 during the rarefaction wave propagation . the high latitude ( off - axis )
emission could contribute to the early flux ( kumar & panaitescu 2000 ) .
the numerical fireball shell has a long tail in the density profile , and the shocked tail contributes to the light curve at late times .
the emission from the outer part containing 80% of the shell mass gives a steeper light curve ( the thin solid curve ) .
if a burst has a precursor , we might need a shift of the reference time to a later time .
a simple example is that a precursor with a negligible energy is followed by the main component .
the quiescent period could be due to a dormant period in the central engine or due to the property of the outflow ( e.g. homogeneous outflow in which prompt gamma - rays are not produced by internal shocks ) . in the former case ,
@xmath0 should be set at the explosion time of the main component .
if the separation between the precursor and the main component is much larger than the deceleration time of the main component , measuring times from the precursor introduces an artificial steeping in the early afterglow . assuming a precursor and measuring times from it
, we have replotted figures [ thinlc ] and [ thicklc ] .
if the precursor is located at @xmath131sec @xmath132 or @xmath133sec @xmath134 before the main component , the decay index of the early forward shock emission could be overestimated as @xmath135 or @xmath136 in the thin shell case . in the thick shell case , @xmath137 and @xmath138
lead to steep decay indexes of @xmath139 and @xmath140 , respectively .
since afterglow light curves are usually plotted with @xmath0 setting at the grb trigger time , precursors , although they might be energetically small , need to be strong enough to trigger grb detectors ( e.g. bat ) in order to cause an artificial steep decay .
x - ray flares were originally reported from bepposax observations , grb 011121 and grb 011211 ( piro et al 2005 ) .
grb 011121 was the second - brightest grb observed by bepposax in gamma - rays ( after grb 990123 ) and in x - rays ( after grb 010222 ) .
the fluence in the 2 - 700 kev range corresponds to an isotropic energy of @xmath141 ergs at the redshift of the burst @xmath142 .
the gamma - ray light curve shows a main peak starting at @xmath143 sec and ending at @xmath144 sec with minor substructures .
an x - ray flare took place at @xmath145 sec .
the fluence of grb 011211 in the 2 - 700 kev range gives an isotropic energy of @xmath146 ergs at the redshift of the burst @xmath147 .
the gamma - ray prompt emission has a long duration @xmath148 sec , and an x - ray flare is detected from 600 to 700 sec . in the source frame , the two flares occurred at a similar time @xmath149 sec and they have a width @xmath150 sec ( see piro et al .
2005 and references therein for the basic parameters of these bursts ) .
since the decay part of the x - ray flares and the following shallower part are described with a single power - law when the time is measured starting from the flare peak , the flares were suggested as the beginning of the afterglow caused by a thick shell ( piro et al .
the thick shell interpretation requires a long energy release from the central engine .
the burst should be long and extend all the way to the flare peak .
however , the observed prompt emission ended well before the onset of the flares . in order to suggest the thick shell case
, one must assume that the efficiency of conversion into gamma - ray varies dramatically with time .
furthermore , the engine should release most fireball energy at the last moment ( major reenergization at the flare ) , because measuring times from the grb trigger does not lead to an overestimate of the decay index even in the thick shell case , as long as the shell is homogeneous . in principle
, the low conversion efficiency could originate from ( 1 ) a small lorentz factor of late ejecta which does not allow gamma - ray radiation or ( 2 ) a small dispersion of the lorentz factor of late ejecta in the internal shock scenario .
the critical lorentz factor is given by @xmath151 where @xmath152 in the cgs unit and @xmath153 . in the scenario ( 1 ) , the low lorentz factor condition @xmath154 requires that the last energetic ejecta with an energy @xmath155 ergs and a width @xmath156 sec should be in the thin shell regime .
the deceleration observed time @xmath157 sec becomes much larger than the width of the flares @xmath158 sec , and one finds that the scenario ( 1 ) is inconsistent .
sec is smaller than the deceleration time @xmath159 sec .
as we have shown at the beginning of this section , a significant artificial steepening happens only when the ratio @xmath160 is larger than @xmath161 .
] to examine the scenario ( 2 ) , we consider an inhomogeneous shell ( two components ) expanding into ism .
corresponding to the major energy release from the central engine at the last moment , the inner edge of the shell with a width @xmath162 sec is assumed to have an energy @xmath163 larger than the energy @xmath164 ergs in the preceding outer part with a width @xmath165 sec .
the lorentz factors of the two components at the end of the internal shock phase are the same value of @xmath122 .
both components should be in the thick shell regime , otherwise the deceleration time of the shell becomes larger than the flare occurrence time @xmath149 sec .
the deceleration radius of the shell should be larger than the radius @xmath166 at which a reverse shock crosses the outer component where @xmath167 . the separation at the deceleration time between the gamma - ray front and the reference bm blast wave satisfies @xmath168 .
the separation at that time between the gamma - ray front and the fireball is @xmath169 where we have assumed the fireball evolution @xmath170 before the deceleration .
note that a broke power law description of @xmath122 around the deceleration radius is a good approximation in the thick shell case .
then , we obtain @xmath171 .
the precursor discussion at the beginning of this section corresponds to the two component model with @xmath172 .
we can show that the separation ratio at the deceleration time is @xmath173 in this case .
since replotting figure [ thicklc ] we have found that @xmath174 to produce an early steep decay , in the scenario ( 2 ) @xmath163 should be at least ten times larger than @xmath175 .
this requires a large energy budget for the central engine @xmath176 ergs . even in the limit of @xmath177 ,
the small ratio @xmath178 does not lead to a very steep decay of @xmath179 .
therefore , we conclude that even an inhomogeneous fireball can not produce x - ray flares in early afterglow via external shock emission process .
we have studied the reference time @xmath0 for the afterglow modeling . although measuring times from the beginning of the prompt emission ( grb trigger ) might cause a slight overestimate of the early afterglow slope in the thick shell case .
this choice of @xmath0 gives a reasonable approximation , and it does not induce a very steep decay ( @xmath180 ) like the early steep decay or x - ray flares in the canonical swift x - ray light curve . the leading model to explain the rapid decay and flares in early x - ray afterglow is the internal shock emission .
a clear , testable prediction of this model is that the temporal decay index @xmath2 of the tail part should be related to the spectral index @xmath181 by an equation @xmath182 ( kumar & panaitescu 2000 ) .
when evaluating the emission decay @xmath183 in the internal shock model , an important difference is that the grb trigger time is no longer special .
the reference time @xmath0 should correspond to the onset of each particular spike in the prompt emission or in afterglow ( kobayashi et al .
1997 ; zhang et al . 2006a
; nousek et al . 2006 ; fan & wei 2005 ) .
every time when the central engine is re - stared to eject sub - shells , the reference time @xmath0 should be re - set to the reactivation time of the engine .
although obrien et al . ( 2006 ) have found that @xmath2 appears to be largely independent of @xmath181 when the bat trigger time is used as @xmath0 , liang et al ( 2006 ) have shown that the relation @xmath182 is more or less satisfied in most cases if @xmath0 is set near the beginning of rising segment of the last pulse of the prompt emission or a corresponding x - ray flare , and if the underlying forward shock emission component is subtracted .
swift observations support the internal shock model
. the self - consistent internal shock interpretation should be more favorable than the beginning - of - the - afterglow interpretation .
the latter can not explain multiple x - ray flares in a single event .
such behavior is observed in many swift bursts ( burrows et al .
2005 ; falcone et al .
2006 ; romano et al . 2006 ; obrien et al 2006 ) .
usually x - ray flux already begins to decay before x - ray flares appear , and it suggests that the onset of afterglow is prior to the flares . if a large amount of energy is impulsively injected to a fireball during the deceleration , @xmath0 might be re - set to the injection time .
however , an afterglow baseline also should shift after a flare ( energy injection ) .
this clearly contradicts with observations in which after a flare peak , afteglow decays back to its pre - flare flux level .
we can not explain x - ray flares by the shift of @xmath0 associated with large energy injections .
swift discovered that a large fraction of x - ray afterglows have a slow decay phase , and it is suggested that energy injection into a blast wave takes place several hundred seconds after the burst .
this implies that right after the burst the kinetic energy of a blast wave is very low and in turn the efficiency of internal shock process is extremely high ( zhang et al .
2006a ; nousek et al .
2006 ; ioka et al . 2005 ; granot , knigl & piran 2006 ; zhang et al.2006b ; however see also fan & piran 2006 ) .
the round - off forward shock peak in the thin shell case might be a good candidate for the shallow decay phase .
however , if we interpret the observed shallow decay as the round - shape of the deceleration phase , the model light curve is shallower than some of the observed ones .
it may be because the observed curve is the combination of this round phase and the rapid decay from the grb tail emission ( high latitude ) . equalizing the deceleration time and the shallow phase timescale we obtain the initial lorentz factor @xmath184 where the time dilation effect is taken into account @xmath185 . for a wind environment
, we can discuss the @xmath0 issue in a very similar way .
the lorentz factor of a shocked shell is constant in both of the thin and thick shell cases during the shock crossing , while the bm blast wave decelerates as @xmath125 ( kobayashi & zhang 2003 ) .
the separations are comparable @xmath186 at the deceleration time .
measuring times from the beginning of the prompt phase should be a good approximation for events in a wind environment also . since most bursts in a wind environment fall in the thick shell case ( kobayashi , & zhang 2004 ) , and since the deceleration time is comparable to grb duration in the thick shell case , another process ( e.g. refreshed shocks ) rather than the afterglow peak is necessary to explain the shallow decay phase .
barthelmy , s.d .
2005 , nature , 438 , 994 blandford , r.d . & mckee , c.f .
1976 , phys , fluids , 19 , 1130 burrows , d .
2005 , science , 309 , 1833 chincarini , g .
2005 , submitted to apj , astro - ph/0506453 falcone , a.d .
2006 , apj , 641 , 1010 fan , y . & piran , t .
2006 , mnras , 369 , 197 fan , y . &
wei , d.m .
2005 , mnras , 364 , l42 granot , j.,knigl , a .
& piran , t .
2006 , mnras , 370 , 1946 ioka , k.,kobayashi , s.&zhang , b.2005 , apj , 631 , 429 ioka , k.,toma , k.,yamazaki , r.&nakamura , t.2005 , a&a in press , astro - ph/0511749 kobayashi , s .
2000 , apj , 545 , 807 kobayashi , s.,,p.&zhang , b .
2004 , apj , 601 , l13 kobayashi , s.,piran , t.&sari , r .
1997 , apj , 490 , 92 kobayashi , s.,piran , t.&sari , r . 1999 ,
apj , 513 , 669 kobayashi , s .
& sari , r .
2000 , apj , 542 , 819 kobayashi , s . &
zhang , b .
2003 , apj , 597 , 455 kumar , p. & panaitescu , a .
2000 , apj , 541 , l51 lazzati , d . &
begelman , m.c .
2006 , apj , 641 , 972 lazzati , d.,ramirez - ruiz , e . & ghisellini , g .
2001 , a&a , 379 , l39 liang , e.w .
et al . 2006 , apj , 646 , 351 mart , j.m .. & mller , e .
1999 , living rev .
relativity , 2 , 3 nakar , e .
& piran , t .
2004 , mnras , 353 , 647 nousek , j.a .
2006 , apj , 642 , 389 obrien , p.t .
2006 , apj , 647 , 1213 piro et al .
2005 , apj , 623 , 314 quimby , r.m .
2006 , apj , 640 , 402 romano , p .
2006 , a&a , 450 , 59 sari , r .
1997 , apj , 489 , l37 sari , r . & piran , t.1995 , apj , 455 , l143 sari , r.,piran , t . & narayan , r .
1998 , apj , 497 , l17 tagliaferri , g .
2005 , nature , 436 , 985 waxman , e .
1997 , apj , 491 , l19 zhang , b .
2006a , apj , 642 , 354 zhang , b .
2006b , submitted to apj | we discuss the reference time @xmath0 of afterglow light curves in the context of the standard internal - external shock model .
the decay index of early afterglow is very sensitive to the reference time one chooses . in order to understand the nature of early afterglow , it is essential to take a correct reference time .
our simple analytic model provides a framework to understand special relativistic effects involved in early afterglow phase .
we evaluate light curves of reverse shock emission as well as those of forward shock emission , based on full hydrodynamic calculations .
we show that the reference time does not shift significantly even in the thick shell case . for external shock emission components ,
measuring times from the beginning of the prompt emission is a good approximation and it does not cause an early steep decay . in the thin shell case , the energy transfer time from fireball ejecta to ambient medium typically extends to thousands of seconds .
this might be related to the shallow decay phases observed in early x - ray afterglow at least for some bursts . |
the internet can be viewed as an autonomous system in which the nodes are heterogeneously connected without any central control . in the internet ,
the unit of information is a packet , and many researchers have been investigating the statistical properties of packet density fluctuations . in 1994
leland et .
al analyzed the time series of packet flow density in the internet , and showed the existence of the @xmath0 type fluctuation@xcite .
similar to the packet flow fluctuation , csabai reported that the time series of round trip time ( rtt ) exhibits the self - similarity in certain path of the internet @xcite . following these pioneering works ,
many observations clearly demonstrate that internet traffic are characterized by the long - range dependency , and the assumption of poisson process , which had been a major traffic model in the traditional traffic theory , has clearly lost its validity @xcite .
takayasu et al . pointed out that the internet traffic can be viewed as a phase transition phenomenon between congested and sparse phases by both the rtt experiment @xcite and the packet flow density analysis@xcite . only at the critical point they found the @xmath0 type fluctuation consistent with the above results .
namely , they clarified that the self - similarity can not always be observed in the internet traffic , and the phase transition view is more general and adequate .
although there are many observational evidences for this phase transition , we still do not fully understand the physical explanation of the observed phenomena .
takayasu et al .
reported that a simple queue itself plays an important role in phase transition phenomena in general @xcite .
we believe that the theory is qualitatively correct , however , there are some quantitative differences between the observation and the simple queueing theory . for example , the congestion duration time is known to be characterized by the power law distribution whose exponent is @xmath1 at the critical point , while the simple queue model can only reproduce the traffic characterized by the power law distribution whose exponent is @xmath2 at the critical point .
thus , we need to find a mechanism that causes such difference in exponents . in this paper
, we focus on the effect of the ethernet ( csma / cd ) mechanism in order to give a more sophisticated physical explanation for the observation facts .
ethernet has been mostly used in the local area network in the internet . ethernet itself , however , has very complicated mechanisms to achieve efficient communication , so it is difficult to clarify the role of each mechanism in the phase transition .
for this reason , we perform a simulation based on the minimal mechanism of the ethernet algorithm , especially focusing on the two effects ; the competition among nodes that intend to send their packets to the shared media , and the exponential back - off algorithm in collision detection .
we review the observation facts in the following section . in section 3
we introduce a model of ethernet traffic with simplified algorithm for numerical simulation .
simulation results are described in section 4 which is divided into four sub - sections .
we discuss the cause of 1/f type fluctuations in section 5 .
the final section is devoted to the summary .
in order to collect the raw data trace , we set up a freebsd pc connected to the 10mbps ethernet link between the wide ( widely distributed environment project ) backbone and keio university in japan by a non - intelligent hub .
there are no other hosts in this link , so the tapping host can capture all packets and the time stamps through this link by using the tcpdump command . from the measurements
, we obtained 12 data traces , each 14440 seconds ( about 4 hours ) long between nov .
1997 and feb . 1998 .
each trace is categorized into the three typical time periods , i.e. early morning , business hour , and evening .
we reconstruct each original tcpdump trace into seven time sequences of the flow density fluctuation in bytes , whose sample size is @xmath3 seconds .
figure 1 shows an example of packet flow density fluctuation of a 4-hour measurement .
the abscissa shows the time in seconds and the ordinate is the corresponding flow density in bytes .
the periods denoted by ( a ) , ( b ) , and ( c ) in the figure indicates three typical periods , namely , sparse , moderately - congested , and congested periods , respectively .
the mean flow densities of the three periods are 150 kbyte / sec , 550 kbyte / sec , and 850 kbyte / sec , respectively . from the figure ,
it is clear that the flow density highly fluctuates in the moderately - congested periods ( b ) in figure 1 , though it roughly stays in lower or higher states in the other cases , ( a ) and ( c ) , respectively .
next we explain the data analysis method called as congestion duration length analysis@xcite .
congestion duration length is a major index to characterize the statistical nature of a given time fluctuation .
we first define a congestion state that the flow density has larger value than a threshold value .
then , the congestion duration length is defined as the sequential number of the congestion state multiplied by the bin size of the original time series .
we are interested in the cumulative distribution of this congestion duration length . at the critical point
, it is known that the congestion duration length distribution follows the power - law distribution with exponent @xmath1 .
this slope is theoretically corresponding to the so - called @xmath0 power spectrum in the original time series .
figure 2 shows the congestion duration length distribution in a log - log plot .
the flow data traces of these three plots correspond to the three periods shown in ( a ) , ( b ) , and ( c ) in figure 1 .
the plotted line clearly decays exponentially in the sparse period , and the distribution is approximated by a power law distribution with exponent @xmath1 in the moderately - congested period .
( the straight line in this figure represents the power law distribution with slope @xmath1 . )
thus , this observed power law distribution shows that the mean flow density in this period is close to the critical point .
moreover , above the critical point , the distribution deviates from the power law again due to large clusters of congestion . from these figures we confirm the existence of the typical two phases , namely , the sparse and congested phases . particularly in the intervals , which are close to this critical mean flow density
, the congestion duration length distribution clearly shows the self - similarity .
also , it is reported that the tendency of divergence of the autocorrelation time the whole behaviors can be confirmed near this critical mean flow density in the same data traces@xcite .
namely , the actual ethernet traffic flow generally changes its statistical property , which can be fully modeled by the phase transition view .
in this section , we focus on the dynamic aspect of the ethernet mechanism , in order to clarify the physical mechanism of the observed phase transition phenomenon .
since the actual ethernet dynamics consists of many complicated rules , it is difficult to extract the essence of the phase transition phenomenon directly from a real system .
therefore , we need to focus on a few basic properties of ethernet rules by introducing a minimal network architecture model to check the occurrence of the phase transition . for this purpose , we consider the two most basic effects in the ethernet mechanism ; the competition of the nodes at packet transmission , and the binary exponential back - off algorithm .
also , to simplify the simulation , we assume a topology consisting of two nodes ( to be called as nodes 1 and 2 ) sending packets randomly and one shared medium to connect them .
figure 3 illustrates our simplified discrete simulation algorithm . in every discrete time step ,
each node probabilistically generates a fixed - size packet to be sent to the medium .
each node has an assigned input rate , and a packet is created randomly with this rate and injected into the finite - size output queue of this node .
this input rate is given initially and it does not change during the simulation .
as known directly from this rule of packet creation , there is no temporal correlation among the input traffic to the queue .
the maximum size of the output queue is 16384 ( @xmath4 ) packets in this simulation .
when a generated packet finds no room in the queue , it is simply discarded .
if more than one packets are waiting in the queue , then the node tries to transmit a packet from the queue to the shared network medium .
when only one node intends to send a packet to this network , this node successfully transmits one packet from the queue , and it takes one time step to finish its transmission .
however , when more than one node try to send packets simultaneously , both nodes fail to transmit them , and the packets remain in their queues .
then , the nodes increase their back - off counters , which indicate the level of the back - off , by 1 .
furthermore , they set their own waiting time @xmath5 chosen randomly from [ 1 , @xmath6 where @xmath7 denotes the back - off counter .
each waiting time decreases one time step in every time step , and a waiting node tries to send a packet when its waiting time becomes zero .
the successful node , conversely , resets its back - off counter to zero after transmission . in our simulation the maximum level of the back - off counter
is set to @xmath8 , i.e. the waiting time is chosen from a random number smaller than @xmath9 ( @xmath4 ) .
also , the back - off counter remains the maximum number even after the node fails to send a packet more than @xmath8 times successively , while the actual ethernet algorithm drops the packet when the back - off limit is larger than the maximum back - off number .
it should be emphasized that new packets are stored in the queue corresponding to the input rate even when the node is in the waiting status .
here , we focus on the property of the total amount of packet traffic flow passing through the link based on our simplified ethernet algorithm . in figure 4
we plot macroscopic performance metrics ( packet dropping rate , throughput , and reliability ) in our simulation . the abscissa denotes the input rate of the nodes , where we set the same input rate for both nodes . from the definition of input rate
, the input rate 50% statistically corresponds to the maximum link capacity . here
, the packet dropping rate is given by the ratio of the number of packets overflowing at either of the nodes divided by the total number of input packets .
the throughput is defined as the number of packets normalized by the link capacity .
the reliability shows the rate of successfully transmitted packets , which is given by 100% minus the packet dropping rate .
obviously , packet dropping rate suddenly takes non - zero values for input rate above about 30% , which is to be called as the critical input rate .
figures 5 ( a ) , ( b ) , and ( c ) show three typical packet flow fluctuations for input rates 15 , 29.5 , and 45 % , respectively .
each figure consists of three sub figures , namely , the output flow fluctuations from node 1 , from 2 , and from 1 @xmath10 2 ( the total traffic in this link ) .
the abscissa shows the time step , and the ordinate indicates the number of output packets per every 16 time steps . at the low input rate
the two nodes randomly transmit their packets , and the total traffic seems to have spontaneous small bursts . near the critical input rate , the total packet fluctuation consists of two typical types of fluctuations .
one is a highly variable fluctuation observed around 28000 time steps , where the two nodes mutually send packets to the network with fine granularity , and the total traffic is nearly a superposition of such individual traffic flows .
the other type is observed around 5000 time steps in which one node continuously sends packets until its buffer becomes empty .
thus , we can expect the sizes of the congestion to be distributed widely at the critical mean flow density . in the case
that the input rate is high , each node alternately sends bursts of roughly same size depending on the buffer size .
the short fluctuations followed by the full bursts observed around 20000 time steps in node 1 are likely due to the imbalance of the buffer size and the maximum back - off waiting time . comparing the figures
, we found that the simulated packet flow captures the qualitative characteristics of the actual traffic flow .
next , we show the result of the congestion duration length distribution in this two - nodes simulation in figure 6 .
the original data trace is the total flow density fluctuations from the two nodes per 16 time steps , and the threshold value of congestion level is set to be two packets .
the three lines in this figure represent the low ( 15% ) , the medium ( 29.5% ) , and the high input rates ( 40% ) in log - log scale .
when the packet input rate is low , the congestion duration length follows the exponential distribution .
namely , the congestion duration has only short - range dependency . in the case of high input rate
the distribution deviates from the power - law distribution due to the large clustered congestion .
the decay of this plotted line at large @xmath11 is due to system size limitation ( e.g. , buffer size , simulation time , etc . ) . in the medium input rate case
the plotted line is approximated by a power law distribution with exponent @xmath1 .
this input rate is the same as the critical value at which a node begins dropping packets due to the buffer overflow in figure 4 . from these results , we confirm that the simplified packet traffic model can reproduce the two typical phases and the critical point characteristics observed in real systems .
the three types of distributions of the congestion duration lengths are consistent with the distributions obtained from the actual traffic measurement@xcite .
it should be noted that the packet input to the queue is based on purely random events of poisson process , so , the self - similarity observed in the output fluctuation at the critical point is caused by the evolution dynamics based on the competition and the binary exponential back - off algorithm . the observed network introduced in section 2 simply consists of only two nodes . in an ordinary network ,
however , there are generally many nodes connected to the shared media . therefore , we need to consider the situation in which more than two nodes participate . here
, we show the traffic fluctuations passing through a link connected to three nodes .
figure 7 represents the congestion duration length distributions in the aggregated traffic from the three nodes following the same simulation algorithm as in the preceding sub - section .
though the critical input rate ( 23% ) is lower than the case of the two - node simulation ( 29.5% ) , the same three typical types of distributions are clearly observed also in this figure .
similarly , we found that the three performance metrics ( packet dropping rate , reliability , and throughput ) bend at this critical point .
we also observed the same type of the phase transition in the four node simulation . the cases with more than four nodes give the same result in general .
these results clearly show that the occurrence of the phase transition behavior is considered to be independent of the number of nodes connected to the link . in the previous simulations
, we have assumed that all nodes had the same input rate . here ,
we analyze an asymmetric case consisting of two nodes having different input rates .
figure 8 depicts the phase diagram of the estimated critical point when the combination of the input rates of the two nodes changes .
the abscissa and the ordinate represent the input rates of nodes 1 and 2 , respectively .
the point ( 30% , 30% ) corresponds to the critical point in the previous symmetric case .
the plotted line represents the set of critical points estimated by the simulation , where self - similar traffic flows can be observed .
we can divide this diagram into two regions separated by the critical line .
the aggregated traffic belongs to the non - congested phase if the combination of input rates of the two nodes is in the lower - left region of this figure , and to the congested phase if it is in the upper - right region above the critical line .
their aggregated traffic fluctuations do not show self - similarity , and their congestion duration length distributions follow approximately an exponential distribution and a distribution with a plateau , respectively , as we saw in figure 6 .
it is found that the phase transition behavior disappears when the nodes connected to the link have extremely different input rates .
this can be understood as follows . in the extremely asymmetric case ,
the probability of collision , which is proportional to the multiplication of the input rates of the two nodes , is relatively small . as a result
, the back - off becomes less effective and the statistical property of the input process is likely to be preserved . in order to clarify the effect of the functional form of the back - off function
, we performed the simulation with a different back - off rule . with the normal ethernet algorithm , when collision is detected , the waiting time
is chosen randomly in the interval between 1 and @xmath12 , where @xmath7 is the back - off counter .
for comparison with this basic algorithm , we check the case of a linear back - off algorithm in this simulation , namely , the back - off time is selected from a random time between @xmath13 and @xmath14 , where the coefficient @xmath5 is set to be @xmath15 in this simulation .
we have confirmed that the following results are the same for @xmath5 between @xmath16 and @xmath17 .
figure 9 shows the results of the congestion duration length analysis for the linear back - off algorithm .
the plotted lines clearly follow exponential functions independent of the value of the input rate . thus
, neither the phase transition phenomenon nor the self - similarity can be observed in the resulting traffic .
it is now clear that the binary exponential back - off algorithm plays an important role in generating self - similar traffics .
similarly , we also check the case of ternary exponential back - off algorithm , that is , the waiting time @xmath5 is chosen randomly from [ @xmath13 , @xmath18 with @xmath7 being the back - off counter . by this simulation we find the power law distribution with the same slope @xmath1 at the critical point while the critical point itself shifts from the value in the binary exponential back - off case .
therefore , the exponential type back - off function is expected to be the essence of the observed phase transition behaviors .
we have found that our simplified ethernet simulation based on both competition in transmission and the exponential back - off algorithm can reproduce essentially the same traffic behaviors as observed in real ethernet traffic , when the total traffic flow rate is low , there are few collisions in the link , and the traffic statistics are dominated by the random input . the aggregated traffic is approximately consisted of superposition of the input traffic from the two nodes .
however , at the critical input rate , the number of collisions becomes not negligible and the back - off mechanism works effectively , and the resulting output traffic becomes correlated in long time scales .
the traffic is hovering about the two phases rather randomly causing large fluctuations showing self - similarity . above the critical input rate
the traffic is dominated by clustered congestion whose size depends both on the buffer size of the nodes and on the maximum number of the back - off count . in this phase
the traffic loses the self - similarity . in the single queue model the power law exponent for the congestion duration length distribution
is known to be @xmath2 at the critical point@xcite .
our results show that the network traffic behavior from more than two nodes following the simplified back - off algorithm exhibits a power law distribution with exponent @xmath1 .
thus , we conclude that the back - off delay due to competition among the nodes is a key factor in generating the @xmath0 type fluctuation . in particular
, it is very important that the phase transition and resulting self - similarity appear in the output traffic behavior even when the input traffic is purely random , having no temporal correlation .
moreover , our simulation demonstrates that the linearly incremental back - off algorithm fails in realizing a phase transition .
thus , it is clear that the multiplicative form of the back - off function is an essential factor in characterizing the statistical property of the ethernet traffic behavior .
finally , our simulation demonstrates that the critical point does not always exist when the nodes have extremely different input rates . in the highly asymmetric case packets have less chance to collide with other packets , and the back - off value
can not become very large .
for this reason the output traffic is not much modified from superposition of input traffics .
we can expect that the critical behavior can be found in a wider region in the phase diagram if the input traffic is temporally correlated because correlated packets generally have more chance of collision with other packets than the case of independent random inputs .
our simulation is based on only two effects in the complicated ethernet mechanism ; the competition among the nodes at transmission process and the exponential back - off .
the simplicity of the simulation algorithm enables us to investigate detailed statistical properties of packet traffics .
our results can be summarized as follows : ( 1 ) the simplified simulation can reproduce the basic properties of the real traffic fluctuations especially in terms of the phase transition between the sparse or mixed transmission phase and congested alternating transmission phase . ( 2 ) at the critical flow density the output flow shows a self - similarity in its fluctuation even though the input is totally random without having any temporal correlation .
( 3 ) phase transition does not appear when the input rates of the nodes are extremely asymmetric .
the approximate phase diagram in the parameter space is drawn with a critical curve .
( 4 ) the occurrence of the phase transition is expected to be independent of the number of the nodes . ( 5 ) the functional form of the back - off algorithm plays an essential role .
the linear back - off function can not realize a phase transition , while the exponential back - off always reproduce the @xmath0 type fluctuation at the critical point .
we wish to thank k. cho and y. takeshima for helpful discussion . | we perform a simplified ethernet traffic simulation in order to clarify the physical mechanism of the phase transition behavior which has been experimentally observed in the flow density fluctuation of internet traffic . in one
phase traffics from nodes connected with an ethernet cable are mixed , and in the other phase , the nodes alternately send bursts of packets .
the competition of sending packets among nodes and the binary exponential back - off algorithm are revealed to play important roles in producing @xmath0 fluctuations at the critical point .
phase transition ; internet traffic ; numerical simulation , |
an interesting feature of a non @xmath16-wave fermi condensate is that it may have plural superfluid phases , originating from active orbital and/or spin degrees of freedom . indeed
, this possibility has experimentally been confirmed in various fermi superfluid systems , such as heavy fermion superconductor upt@xmath17 @xcite , as well as superfluid liquid @xmath18he @xcite .
however , the pairing symmetry that has already been realized is still only the simplest @xmath16-wave one in cold fermi gas physics @xcite .
thus , going beyond this situation is a crucial challenge in this research field . in this regard ,
a @xmath0-wave superfluid fermi gas is a strong candidate , and the possibility of this spin - triplet pairing state has extensively been discussed both experimentally @xcite and theoretically @xcite .
several experimental groups have discovered a @xmath0-wave feshbach resonance in @xmath5k @xcite and @xmath19li @xcite fermi gases , so that we can now tune the strength of a @xmath0-wave interaction from the weak - coupling regime to the strong - coupling regime , by adjusting an external magnetic field .
this experimental development has realized @xmath0-wave feshbach molecules @xcite .
thus , although one still needs to overcome some difficulties , such as the three - body loss @xcite , as well as the dipolar relaxation @xcite ( that destroy @xmath0-wave molecules @xcite ) , the @xmath0-wave superfluid state seems a very promising non @xmath16-wave pairing state in an ultracold fermi gas .
it has been predicted @xcite that , when a @xmath0-wave interaction has a uniaxial anisotropy , a @xmath0-wave superfluid fermi gas may have two superfluid phases with different @xmath0-wave pairing symmetries .
such an anisotropic @xmath0-wave pairing interaction is considered to be realized in a @xmath5k fermi gas @xcite , because the split of a @xmath0-wave feshbach resonance into a @xmath7-wave channel and degenerate @xmath14-wave and @xmath15-wave channels by a magnetic dipole - dipole interaction has been observed , when an external magnetic field is applied in the @xmath20 direction .
since the observed resonance field of a @xmath7-wave feshbach resonance is higher than that of the other degenerate channels @xcite , a @xmath0-wave pairing interaction associated with this @xmath0-wave feshbach resonance has the uniaxial anisotropy , that is , a @xmath7-wave pairing interaction @xmath21 is stronger than @xmath14-wave ( @xmath22 ) and @xmath15-wave ( @xmath23 ) interactions . as a result , the @xmath7-wave superfluid state has the highest superfluid phase transition temperature @xmath24 . in this case , refs .
@xcite have pointed out that the system experiences the other phase transition from the @xmath7-wave state to the @xmath11-wave state at @xmath9 ( @xmath10 ) , when the uniaxial anisotropy satisfies a certain condition .
thus , the realization of a @xmath0-wave superfluid fermi gas would enable us to study physics of plural superfluid phases , from the weak - coupling regime to the strong - coupling limit in a systematic manner .
when the above - mentioned @xmath0-wave superfluid fermi gas is realized , the existence of strong pairing fluctuations near @xmath9 is an interesting research topic . in the @xmath7-wave superfluid phase ,
since single - particle excitations are gapless in the nodal direction ( @xmath25 ) of the @xmath7-wave superfluid order parameter @xmath26 , @xmath14-wave and @xmath15-wave pairing fluctuations can continue developing , to be the strongest at @xmath9 .
thus , even far below @xmath24 , an anisotropic pseudogap phenomenon is expected in the nodal direction of the @xmath7-wave superfluid order parameter near @xmath9 . in an isotropic @xmath16-wave superfluid fermi gas , since the bcs gap opens in all the momentum direction of single - particle excitations , pairing fluctuations are soon suppressed below the superfluid phase transition temperature @xmath27 .
indeed , it has been shown that a pseudogap in the density of states above @xmath27 @xcite soon changes to the @xmath16-wave bcs superfluid gap below @xmath27 @xcite . even in the @xmath7-wave case
, the enhancement of pairing fluctuations around the node would not occur in the absence of a @xmath14-wave and a @xmath15-wave interactions , because @xmath7-wave pairing fluctuations are soon suppressed by the @xmath7-wave superfluid order below @xmath24 .
thus , the above - mentioned pseudogap phenomenon in the @xmath7-wave state near @xmath9 is peculiar to an unconventional fermi superfluid with a nodal superfluid order parameter , as well as with plural superfluid phases . in this paper
, we theoretically investigate strong - coupling properties of a one component @xmath0-wave superfluid fermi gas with a uniaxially anisotropic @xmath0-wave pairing interaction ( @xmath1 ) . including @xmath0-wave pairing fluctuations within the framework of a strong coupling @xmath6-matrix approximation , we determine @xmath24 and @xmath9 . in the @xmath7-wave superfluid phase
, we calculate the angle - resolved single - particle density of states , to clarify that pairing fluctuations in the nodal direction ( @xmath25 ) of the @xmath7-wave superfluid order parameter continue to develop , leading to a pseudogapped single - particle excitation spectrum in the nodal direction .
the paper is organized as follows . in sec.ii
, we explain our strong - coupling formalism for a @xmath0-wave superfluid fermi gas . in sec.iii
, we show our numerical results on @xmath24 and @xmath9 . here
, we clarify the condition for the appearance of @xmath7-wave and @xmath11-wave superfluid phases . in sec.iv
, we examine the angle - resolved single - particle density of states ( ardos ) .
we show that a pseudogap anisotropically appears in ardos , not only in the normal state near @xmath24 , but also in the @xmath7-wave superfluid phase near @xmath9 . to characterize this anisotropic many - body phenomenon
, we introduce the characteristic temperature @xmath28 as the temperature below which a dip structure appears in ardos . throughout this paper , we take @xmath29 , and the volume of the system @xmath30 is taken to be unity , for simplicity .
we consider a one - component fermi gas with a uniaxially anisotropic @xmath0-wave interaction , described by the hamiltonian , @xmath31 here , @xmath32 is the creation operator of a fermi atom with the kinetic energy @xmath33 , measured from the fermi chemical potential @xmath34 ( where @xmath35 is the atomic mass ) .
@xmath36 ( @xmath37 ) is a pairing interaction in the @xmath4-wave cooper channel ( @xmath38 ) , having the uniaxial anisotropy @xmath39 @xcite . in this paper
, we do not deal with details of a @xmath0-wave feshbach resonance , but simply treat @xmath40 as a tunable parameter set . to eliminate the ultraviolet divergence , the last term in eq .
( [ eq.1 ] ) involves the cutoff function @xcite , @xmath41 where @xmath42 is a cutoff momentum . for simplicity
, we use the same cutoff function @xmath43 for all the @xmath0-wave cooper channels .
equation ( [ eq.2 ] ) gives a lorentzian cutoff when @xmath44 @xcite , and gives a sharp cutoff when @xmath45 @xcite .
we will discuss the cutoff dependence of our results in sec .
iii . as usual @xcite , we conveniently measure the strength of a @xmath0-wave interaction in terms of the @xmath4-wave scattering volume @xmath46 ( @xmath38 ) , as well as the effective range @xmath47 , that are given by , respectively , @xmath48 @xmath49 since we take the same cutoff function @xmath43 for all the @xmath4-wave interaction channels , the effective range @xmath47 are channel - independent .
we also introduce the anisotropy parameter , @xmath50 then , the @xmath0-wave interaction can be specified by the parameter set @xmath51 .
the weak - coupling side and the strong - coupling side are characterized as @xmath52 and @xmath53 , respectively , where @xmath54 is the fermi momentum . to deal with strong - coupling phenomena in the superfluid phase , it is convenient to rewrite the model hamiltonian in eq .
( [ eq.1 ] ) into the sum of the mean - field bcs part and the term describing fluctuation corrections @xcite . under the nambu representation @xcite , we have , @xmath55 \psi_{\bm p } -\frac{1}{2}\sum_{{\bm q},i = x , y , z } u_i\rho_i^+({\bm q})\rho_i^-(-{\bm q } ) , \label{eq.7}\end{aligned}\ ] ] where @xmath56 is the two - component nambu field , and @xmath57 ( @xmath38 ) are pauli matrices acting on particle - hole space @xcite . the first term in eq .
( [ eq.7 ] ) is just the mean - field bcs hamiltonian , where @xmath58 is a @xmath59 matrix @xmath0-wave superfluid order parameter . here , @xmath60 where @xmath61 has the form , @xmath62 the second term in eq .
( [ eq.7 ] ) gives fluctuation corrections to the mean - field hamiltonian , where the so - called generalized density operator @xcite , @xmath63 physically describes @xmath4-wave superfluid fluctuations ( where @xmath64 ) . in the @xmath6-matrix approximation ( tma ) .
the wavy line is the particle - particle scattering matrix @xmath65 given in ( b ) .
the solid line and the dashed line describe the mean field single - particle green s function @xmath66 , and the @xmath0-wave interaction , respectively . in ( a ) , the factor @xmath67 is assigned to each vertex ( solid circle ) . in ( b ) , we assign @xmath68 to each vertex . in this figure , @xmath69 means the opposite sign to @xmath70.,width=377 ] strong - coupling corrections to single - particle excitations can be conveniently described by the self - energy @xmath71 in the single - particle thermal green s function , @xmath72 where @xmath73 is the fermion matsubara frequency . in eq .
( [ eq.10 ] ) , @xmath74 is the single - particle green s function in the mean - field level . treating the last term in eq .
( [ eq.7 ] ) within the @xmath6-matrix approximation ( tma ) @xcite , we have the diagrammatic expression for the self - energy @xmath71 shown in fig .
[ fig1](a ) @xcite , which gives @xmath75 \left[p_j-{q_j \over 2}\right ] \nonumber \\ & \times & \left ( \begin{array}{cc } g_0^{22}({\bm p}-{\bm q},i\omega_n - i\nu_n ) \gamma^{-+}_{i , j}({\bm q},i\nu_n ) & g_0^{21}({\bm p}-{\bm q},i\omega_n - i\nu_n ) \gamma^{++}_{i , j}({\bm q},i\nu_n ) \\ g_0^{12}({\bm p}-{\bm q},i\omega_n - i\nu_n ) \gamma^{--}_{i , j}({\bm q},i\nu_n ) & g_0^{11}({\bm p}-{\bm q},i\omega_n - i\nu_n ) \gamma^{+-}_{i , j}({\bm q},i\nu_n ) \end{array } \right ) . \nonumber
\\ \label{eq.12}\end{aligned}\ ] ] here , @xmath76 is the boson matsubara frequency .
the tma particle - particle scattering matrix @xmath77 in eq .
( [ eq.12 ] ) obeys the equation , @xmath78 where @xmath69 means the opposite sign to @xmath70 , and @xmath79 \label{eq.15}\end{aligned}\ ] ] is the pair correlation function .
in particular , @xmath80 describes fluctuations in the @xmath4-wave cooper channel , and @xmath81 ( @xmath82 ) describes coupling between @xmath4-wave and @xmath83-wave pairing fluctuations . in the present anisotropic case ( @xmath1 ) , the highest superfluid phase transition temperature is obtained in the @xmath7-wave cooper channel . as in the @xmath16-wave case @xcite ,
the tma gap equation for the @xmath7-wave superfluid order parameter @xmath84 is obtained from the thouless criterion @xmath85^{-1}=0 $ ] in the @xmath7-wave cooper channel , where @xmath86 , \label{eq.14_2}\end{aligned}\ ] ] describes the phase fluctuations of the @xmath7-wave superfluid order parameter @xmath87 .
physically , this guarantees the existence of a gapless goldstone mode associated with the broken @xmath88 gauge symmetry .
noting that @xmath89 in the @xmath7-wave superfluid phase , we obtain the gap equation in the @xmath7-wave superfluid state as @xmath90 . \label{eq.18q}\end{aligned}\ ] ] the equation for @xmath24 is obtained from eq .
( [ eq.18q ] ) by setting @xmath91 as @xmath92 .
\label{eq.18}\end{aligned}\ ] ] in the case of uniaxially anisotropic @xmath0-wave interaction , ref .
@xcite pointed out that , without loss of generality , one may restrict the structure of the @xmath0-wave superfluid order parameter to the form , @xmath93 where @xmath94 and @xmath95 are real quantities .
thus , in the @xmath7-wave superfluid phase below @xmath24 ( where @xmath96 and @xmath97 ) , the other possible superfluid instability is only associated with the @xmath11-wave one , having the superfluid order parameter , @xmath98f_n({\bm p } ) .
\label{eq.18b}\end{aligned}\ ] ] since @xmath94 is already present below @xmath24 , the superfluid phase transition temperature @xmath9 is determined from the thouless criterion @xcite in the @xmath14-wave cooper channel @xmath99^{-1}=0 $ ] , @xmath100 , \label{eq.18c}\end{aligned}\ ] ] where the @xmath7-wave superfluid order parameter @xmath84 obeys the gap equation ( [ eq.18q ] ) .
we numerically solve eqs .
( [ eq.18q ] ) , ( [ eq.18 ] ) , and ( [ eq.18c ] ) , to self - consistently determine @xmath24 , @xmath9 , and @xmath87 . in this procedure , we also solve the equation for the total number @xmath101 of fermi atoms , @xmath102 , \label{eq.18d}\ ] ] to include strong - coupling corrections to the fermi chemical potential @xmath34 .
we examine the anisotropic pseudogap phenomenon by calculating the angle - resolved single - particle density of states ( ardos ) , @xmath103 , \label{eq.27}\ ] ] where @xmath104 , and @xmath105 is the ( 1,1 ) component of the analytic continued tma green s function in eq .
( [ eq.10 ] ) .
ardos in eq .
( [ eq.27 ] ) is related to the ordinary density of states @xmath106 as @xmath107 here , we choose the @xmath7 axis as the polar axis ( @xmath108 ) .
we note that , since the anisotropy of the @xmath7-wave superfluid order parameter @xmath26 lowers the symmetry of the system , we need much time to compute the number equation ( [ eq.18d ] ) below @xmath24 , compared to the symmetric @xmath16-wave case .
to avoid this difficulty , in this paper , we approximate the tma green s function @xmath109 in the number equation ( [ eq.18d ] ) to @xmath110 equation ( [ eq.18e ] ) is just the same form as the green s function in the strong - coupling theory developed by nozires and schmitt - rink ( nsr ) @xcite . the nsr theory has extensively been used in the @xmath16-wave case , to successfully explain the bcs - bec crossover behavior of the superfluid phase transition temperature @xcite , as well as the superfluid order parameter in the crossover region @xcite .
the nsr theory has also been extended to the @xmath0-wave case with @xmath111 @xcite .
thus , we expect that the nsr green s function in eq .
( [ eq.18e ] ) also works in determining @xmath24 , @xmath9 , @xmath34 and @xmath87 . on the other hand
, it is also known that the nsr theory unphysically gives negative density of states in the bcs - bec crossover region @xcite . since this serious problem
is absent in tma , we use the tma green s function in eq .
( [ eq.10 ] ) , in considering single - particle properties of a @xmath7-wave fermi superfluid . here
, we summarize our detailed parameter settings . for the effective range , we take @xmath112 , following the experimental result on a @xmath5k fermi gas @xcite . for the cutoff function @xmath43 in eq .
( [ eq.2 ] ) , we set @xmath113 .
the cutoff momentum @xmath42 in @xmath43 is determined so as to reproduce @xmath112 , which gives @xmath114 .
since we only deal with the normal state , as well as the @xmath7-wave superfluid state , ardos in eq .
( [ eq.27 ] ) is actually independent of the angle @xmath115 around the @xmath7 axis .
thus , the anisotropy can be simply specified by the polar angle @xmath116 measured from the @xmath7 axis . noting this
, we write eq .
( [ eq.27 ] ) as @xmath117 in what follows .
figure [ fig2 ] shows the phase diagram of a one component ultracold fermi gas in terms of the @xmath0-wave interaction strength , @xmath118 , and the temperature . when @xmath119 ( @xmath120 , fig .
[ fig2](a ) shows that the superfluid phase is dominated by the @xmath11-wave pairing state .
this superfluid region gradually shrinks with increasing the uniaxial anisotropy @xmath121 , as shown in figs . [ fig2](b ) and ( c ) .
since the present anisotropy ( @xmath1 ) favors the @xmath7-wave symmetry , the region of the @xmath11-wave state eventually vanishes , as shown in fig .
[ fig2](d ) .
we briefly note that the overall structure of this phase diagram is consistent with the previous work based on mean - field analyses @xcite .
-wave pairing interaction .
the solid line and the dashed line are @xmath24 and @xmath9 , respectively .
@xmath122 is the fermi temperature.,width=377 ] in addition to @xmath24 , the coupled equations ( [ eq.18 ] ) and ( [ eq.18d ] ) also give the fermi chemical potential @xmath123 shown in fig .
[ fig3 ] , which exhibits the typical bcs - bec crossover behavior @xcite .
that is , with increasing the interaction strength , @xmath123 gradually deviates from the fermi energy @xmath124 , to be negative in the strong - coupling regime , when @xmath53 . at @xmath9 ,
the coupled equations ( [ eq.18q ] ) , ( [ eq.18c ] ) and ( [ eq.18d ] ) also give @xmath125 , as well as the @xmath7-wave superfluid order parameter @xmath126 .
figure [ fig3 ] shows that @xmath127 in the whole interaction regime , indicating that the chemical potential is almost @xmath6-independent in the @xmath7-wave superfluid phase . for the @xmath7-wave superfluid order parameter @xmath126 , of course
, this quantity has already existed above @xmath9 , as shown in fig .
although the pairing symmetry changes from the @xmath7-wave one to the @xmath11-wave one at @xmath9 , it is known @xcite that this symmetry change occurs smoothly , in the sense that @xmath95 in eq .
( [ eq.18b ] ) continuously grows from zero below @xmath9 .
thus , the second order phase transition is expected at @xmath9 ( unless the superfluid order parameter exhibits an unexpected discontinuity at @xmath9 ) . at the two superfluid phase transition temperatures , @xmath24 and @xmath9 .
we set @xmath128.,width=377 ] in the @xmath7-wave superfluid order parameter @xmath129 , when @xmath128 .
each result ends at @xmath9.,width=377 ] in fig .
[ fig4 ] , we see that @xmath87 has a discontinuity at @xmath24 , which is , however , an artifact of tma we are using in this paper .
the same problem has already been known in the @xmath16-wave case @xcite . in the latter case , it has been pointed out @xcite that one needs to correctly include an effective repulsive interaction between cooper pairs beyond tma , in order to recover the expected second order phase transition .
although this improvement is also crucial in the @xmath0-wave case , we leave this problem as a future problem , and examine strong - coupling effects in the @xmath7-wave superfluid phase within tma . and
@xmath130 , as functions of the anisotropy parameter @xmath131 . in panel
( c ) , the solid circles shows the bec phase transition temperature @xmath132 obtained from eqs .
( [ eq.24a])-([eq.25 ] ) .
the upper dashed line in this panel shows @xmath133 .
the lower dashed line shows @xmath134.,width=566 ] in the weak - coupling bcs limit , the number equation ( [ eq.18d ] ) simply gives @xmath135 , so that the superfluid phase transition temperature @xmath24 is determined from eq .
( [ eq.18 ] ) with @xmath135 .
the resulting @xmath24 does not depend on the anisotropy parameter @xmath121 , for a fixed @xmath7-wave interaction strength @xmath136 . on the other hand , @xmath24 gradually comes to depend on @xmath121 , as one approaches the strong - coupling regime , as shown in fig .
[ fig5 ] . to explain the anisotropy dependence of @xmath24 in the strong - coupling regime shown in fig .
[ fig5](c ) , we first note that the system in the strong - coupling limit @xcite may be viewed as an ideal gas mixture of three kinds of tightly - bound molecules ( with the molecular mass @xmath137 ) that are formed by the three @xmath4-wave interactions ( @xmath38 ) . indeed , in the strong - coupling limit , the number equation ( [ eq.18d ] ) at @xmath24 is reduced to @xmath138 where @xmath139 is the number of molecules in the @xmath4-wave cooper channel , with @xmath140 being the bose distribution function .
the @xmath24-equation ( [ eq.18 ] ) gives the bose chemical potential @xmath141 in eq .
( [ eq.24b ] ) as @xmath142 in the absence of uniaxial anisotropy ( @xmath121=0 ) , all the three components simultaneously satisfy the bec condition , @xmath143 .
thus , the phase transition temperature ( @xmath144 ) is determined from the equation , @xmath145 , which gives @xmath146 where @xmath147 is the zeta function .
in contrast , when @xmath148 , eq .
( [ eq.25 ] ) shows that the @xmath14- and @xmath15-wave components no longer satisfy the bec condition . in the extreme case
when @xmath149 , the bose chemical potentials @xmath150 and @xmath151 in eq .
( [ eq.25 ] ) are much lower than zero , so that one can ignore the contributions of these components to the number equation ( [ eq.24a ] ) , as @xmath152 .
this means that most atoms form bound molecules in the @xmath7-wave cooper channel , which is quite different from the case of @xmath119 , where only one third of fermi atoms contribute to @xmath7-wave molecules . because of this , the bec phase transition temperature @xmath153 ) in this extreme case is higher than of @xmath119 in eq .
( [ eq.25b ] ) as , @xmath154 figure [ fig5](c ) shows that the bec phase transition temperature @xmath155 calculated from eqs .
( [ eq.24a])-([eq.25 ] ) monotonically increases from @xmath134 to @xmath133 , with increasing the anisotropy parameter @xmath121 .
the well agreement of @xmath24 with @xmath132 shown in this figure indicates that the anisotropy dependence of @xmath24 in this regime comes from the increase of the molecular bosons in the @xmath7-wave cooper channel , with increasing the uniaxial anisotropy of the @xmath0-wave interaction . of the anisotropy parameter at which @xmath9 vanishes ( solid line ) .
the dashed line shows the result within the bcs - leggett theory . in obtaining the solid line ,
we have taken a small but finite value of @xmath6 ( @xmath156 ) because of computational problems.,width=302 ] in contrast to @xmath24 , we see in fig . [ fig5 ] that the @xmath11-wave superfluid phase transition temperature @xmath9 decreases with increasing @xmath121 , to eventually vanish at a critical value @xmath157 .
( note that this vanishing @xmath9 has already been expected in fig .
[ fig2 ] . ) evaluating this critical value @xmath157 in the whole interaction strength , we obtain fig .
this figure shows that @xmath157 is not so sensitive to the interaction strength , to always lie in the narrow range , @xmath158 . to understand this behavior of @xmath157 , since thermal fluctuations are absent at @xmath159 , it is convenient to employ the bcs - leggett theory @xcite , which consists of the coupled eq .
( [ eq.18c ] ) at @xmath160 with the mean - field number equation at @xmath159 , @xmath161 = { 1 \over 2 } \sum_{\bm p } \left [ 1- { \xi_{\bm p } \over \sqrt{\xi_{\bm p}^2+|\delta_{p_x}({\bm p})|^2 } } \right ] .
\label{eq.25d}\ ] ] as shown in fig .
[ fig6 ] , the bcs - leggett theory semi - quantitatively reproduces the tma result for @xmath157 . in the weak - coupling bcs regime ( @xmath162 ) , the number equation ( [ eq.25d ] ) simply gives @xmath135 . substituting this into eq .
( [ eq.18c ] ) with @xmath160 , one obtains the upper bound of @xmath157 in the bcs - leggett theory as @xmath163 in the strong coupling regime where the chemical potential is negative and @xmath164 , the bcs - leggett theory gives the lower bound of @xmath157 as @xmath165 strictly speaking , although the tma result for @xmath157 coincides with eq .
( [ eq.20 ] ) in the weak - coupling limit , it is still different from eq .
( [ eq.21 ] ) even in the strong - coupling limit .
this is because the finite value of the effective range ( @xmath112 ) causes an effective repulsive interaction between bound molecules in the latter limit , leading to the so - called quantum depletion @xcite .
indeed , in the strong - coupling bec regime , the last term in eq .
( [ eq.18e ] ) , which describes fluctuation corrections to the mean - field green s function @xmath166 , modifies the mean - field number equation ( [ eq.25d ] ) as @xmath167 + { 8 \over 3\sqrt{\pi}}\sum_{i = x , y , z } \left ( { n_{\rm f } \over 2}a_{{\rm b},i}^3 \right)^{1 \over 2 } , \label{eq.22}\ ] ] where @xmath168 , and @xmath169 .
the last term in eq .
( [ eq.22 ] ) has the same form as the quantum depletion in a bose superfluid with @xmath170 bosons , when we interpret @xmath171 as an effective repulsive interaction between tightly bound @xmath4-wave molecules .
when we include this quantum depletion , the lower bound in eq .
( [ eq.21 ] ) is improved as @xmath172 which agrees well with the tma result ( solid line in fig .
[ fig6 ] ) in the strong - coupling regime . and
@xmath9 for various values of @xmath173 in the cutoff function @xmath43 .
we take @xmath174.,width=302 ] before ending this section , we comment on the the cutoff function @xmath175 we are using . as shown in fig .
[ fig7 ] , while the phase transition temperature @xmath24 is almost independent of @xmath173 , @xmath9 depends on this parameter .
generalizing eq .
( [ eq.21 ] ) to the case with an arbitrary @xmath173 , one finds that the lower bound ( in the bcs - leggett theory ) explicitly depends on @xmath173 as @xmath176 these @xmath173-dependences of @xmath9 and @xmath157 are because the factor @xmath7 in @xmath177 enhances this superfluid order parameter in the high momentum region , so that physical quantities in the @xmath7-wave superfluid phase depend on how this enhancement is suppressed by the cutoff function @xmath43 .
this implies that , in addition to the observable parameter set @xmath178 , one need one more experimental information about the high momentum regime of a real @xmath0-wave interaction , in order to unambiguously predict the phase boundary between the @xmath7-wave and @xmath11-wave superfluid phases .
we briefly note that fig .
[ fig7 ] shows that our choice ( @xmath113 ) is close to the case of discrete cutoff ( which corresponds to @xmath45 . ) .
( a1)(a2 ) @xmath179 .
( b1)(b2 ) @xmath180 .
( c1)(c2 ) @xmath181 .
( d1)(d2 ) @xmath174 .
upper and lower figures show the results at @xmath24 and @xmath9 , respectively . in each figure
, we offset the results by 0.2 .
, width=566 ]
figure [ fig8 ] shows the angle resolved density of states @xmath117 ( ardos ) at @xmath24 ( upper figures ) and @xmath9 ( lower figures ) . in fig .
[ fig8](a1 ) , a dip structure is shown around @xmath12 . since the superfluid order parameter vanishes at @xmath24 , this is just a pseudogap originating from @xmath0-wave pairing fluctuations .
this many - body phenomenon is non - monotonic in the sense that , while this pseudogap is more remarkable in fig .
[ fig8](b1 ) , it gradually becomes obscure with further increasing the interaction strength , to eventually vanish , as shown in figs .
[ fig8](c1 ) and ( d1 ) .
figures [ fig8](a1)-(c1 ) also show that the pseudogap structure at @xmath24 is anisotropic in momentum space , and is the most remarkable in the @xmath7 direction ( @xmath182 ) . to simply explain this anisotropic pseudogap phenomenon , it is convenient to employ the static approximation for pairing fluctuations @xcite . noting that the particle - particle scattering matrix @xmath183 in the @xmath7-wave channel diverges at @xmath24 @xcite , we may approximate the ( 1,1 ) component of the tma self - energy in eq .
( [ eq.12 ] ) in the normal state near @xmath24 to @xmath184 where @xmath185 p_x^2f_n^2({\bm p } ) \equiv { b_{\rm pg}^x}^2p_x^2f_n^2({\bm p } ) \label{eq.29}\end{aligned}\ ] ] is the so - called pseudogap parameter @xcite . in obtaining eq .
( [ eq.28 ] ) , we have only retained effects of the strongest @xmath7-wave pairing fluctuations near @xmath24 , and have ignored fluctuation contributions from the @xmath14-wave and @xmath15-wave cooper channels .
substituting eq .
( [ eq.28 ] ) into the ( 1,1 ) component of the tma green s function in eq .
( [ eq.10 ] ) , one obtains @xmath186 the first line in eq .
( [ eq.30 ] ) means that the pseudogap parameter @xmath187 works as a coupling between the particle branch @xmath188 and the hole branch @xmath189 . on the viewpoint of this particle - hole coupling
, the pseudogap may be interpreted as a result of the level repulsion between the particle and hole branches around @xmath12 @xcite . the last expression in eq .
( [ eq.30 ] ) is just the same form as the diagonal component of the green s function in the ordinary mean - field bcs theory .
this coincidence immediately gives the bcs - type single - particle excitation spectra @xmath190 , having the excitation gap , @xmath191 where the cutoff function @xmath175 has been approximated to unity .
( note that @xmath192 . )
this gap is actually a pseudogap , when one correctly includes a finite lifetime of preformed cooper pairs , which is ignored in the static approximation @xcite
. equation ( [ eq.32a ] ) indicates that , as expected , the anisotropic pseudogap phenomenon shown in fig.[fig8](a1)-(c1 ) originates from the anisotropic @xmath7-wave pairing fluctuations , described by the pseudogap parameter @xmath193 .
when @xmath111 , fluctuations in all the three @xmath4-wave cooper channels ( @xmath38 ) are equally enhanced near the superfluid instability , so that the pseudogap parameter in eq .
( [ eq.29 ] ) is replaced by the isotropic one , @xmath185 p^2f_n^2({\bm p } ) \label{eq.31}\end{aligned}\ ] ] where we have used the symmetry property , @xmath194 .
the resulting pseudogap is isotropic in momentum space @xcite .
equation ( [ eq.32a ] ) also shows that the ( pseudo)gap size @xmath195 becomes isotropic in the strong coupling regime where the fermi chemical potential is negative @xcite .
this is because most fermi atoms form tightly bound molecules in the strong - coupling regime , so that the threshold energy of single - particle excitations is simply dominated by the dissociation of these molecules with the binding energy @xmath196 , as in the strong - coupling bec regime of the @xmath16-wave case @xcite .
the pseudogap parameter @xmath187 in eq .
( [ eq.29 ] ) also explains the non - monotonic behavior of the pseudogap structure in terms of the interaction strength shown in figs .
[ fig8](a1)-(d1 ) @xcite . since pairing fluctuations are stronger for a stronger pairing interaction , the factor @xmath197 appearing in eq .
( [ eq.29 ] ) also becomes larger , which enhances the pseudogap parameter @xmath187 . at the same , since
strong pairing fluctuations are known to decrease the fermi chemical potential @xmath34 @xcite as shown in fig .
[ fig3 ] , the effective fermi momentum defined by @xmath198 becomes small .
this decreases the pseudogap parameter at the effective fermi momentum because @xmath199 . as a result , while the pseudogap first becomes remarkable with increasing the interaction strength in the weak - coupling region because of the enhanced pairing fluctuations , it gradually shrinks when the decrease of the fermi chemical potential dominantly contributes to @xmath187 .
in the case of fig .
[ fig8](d1 ) , one has @xmath200 .
thus , the low momentum region @xmath201 dominantly contributes to the density of states around @xmath12 , leading to the vanishing pseudogap in this figure @xcite .
we briefly note that the vanishing pseudogap in the intermediate coupling regime ( @xmath202 ) is quite different from the @xmath16-wave case , where the pseudogap monotonically develops , as one passes through the bcs - bec crossover region .
this is simply because the contact - type @xmath16-wave pairing interaction is independent on the momentum @xmath203 , so that the factor @xmath7 is absent in the @xmath16-wave pseudogap parameter . at @xmath9 , since the @xmath7-wave superfluid order parameter @xmath204 is already present , ardos @xmath205 in the low energy region is dominated by the superfluid energy gap , as shown in figs . [
fig8](a2)-(c2 ) . on the other hand ,
such a gap structure is not shown in @xmath206 in the weak - coupling regime ( fig .
[ fig8](a2 ) ) , because of the vanishing @xmath7-wave superfluid order parameter there .
however , ardos in the nodal direction ( @xmath207 ) gradually exhibits a dip structure around @xmath12 , with increasing the pairing interaction .
( see figs . [ fig8](b2 ) and ( c2 ) . ) since the @xmath11-wave superfluid order parameter still vanishes at @xmath9 , this is a pseudogap induced by fluctuations in the @xmath14- and @xmath15-wave cooper channels .
indeed , when we apply the static approximation to the region near @xmath9 , the ( 1,1 ) component of the single - particle green s function with @xmath208 is reduced to eq .
( [ eq.30 ] ) where the pseudogap parameter is replaced by @xmath209 p_i^2f_n^2({\bm p } ) \equiv \sum_{i = y , z}{b_{\rm pg}^i}^2p_i^2f_{n=3}^2({\bm p } ) .
\label{eq.32}\ ] ] although eq .
( [ eq.32 ] ) is similar to eq .
( [ eq.29 ] ) , the former involves effects of @xmath7-wave superfluid order parameter @xmath87 . since gapless
fermi excitations only remain along the line node of the @xmath7-wave superfluid order parameter , pairing fluctuations described by @xmath210 in eq .
( [ eq.32 ] ) are weaker than pairing fluctuations described by @xmath211 in eq .
( [ eq.29 ] ) .
this explains why the pseudogap appearing in ardos @xmath206 at @xmath9 ( figs .
[ fig8](b2 ) and ( c2 ) ) is less remarkable , compared to the dip structure in @xmath212 at @xmath24 ( figs . [ fig8](a1)-(c1 ) ) .
the reason for the vanishing superfluid gap and pseudogap gap in fig .
[ fig8](d2 ) is the same as that in the case of fig .
[ fig8](d1 ) .
that is , at this interaction strength , the chemical potential is very small ( @xmath213 ) , so that the low - momentum region ( @xmath201 ) dominantly contributes to ardos @xmath117 around @xmath12 .
thus , the @xmath7-wave superfluid order parameter @xmath26 , as well as effects of the pseudogap parameter @xmath214 in eq .
( [ eq.32 ] ) , do not almost affect ardos around @xmath12 in this case . at various temperatures , when @xmath215 .
( a1)(a2 ) @xmath182 .
( b1)(b2 ) @xmath216 .
( c1)(c2 ) @xmath207 .
the upper and lower figures show the results in the normal state , and in the @xmath7-wave superfluid state , respectively . evaluating the pseudogap temperature @xmath217
as the temperature below which a dip structure appears in @xmath117 around @xmath12 , one obtains @xmath218 , @xmath219 , and @xmath220 . in each figure
, we offset the results by 0.3.,width=566 ] below which a dip structure appears in ardos @xmath117 .
we take @xmath128 .
the dashed - dotted line shows the temperature at which the fermi chemical potential @xmath34 vanishes .
the chemical potential @xmath34 is negative in the right side of this line , so that this strong - coupling regime may be regarded as a gas of two - body bound molecules with the binding energy @xmath221 @xcite , rather than a gas of fermi atoms.,width=377 ] figures [ fig9](a1 ) and ( a2 ) show that the pseudogap in @xmath212 in the normal state continuously changes to the superfluid gap , as one passes through the superfluid instability at @xmath24 .
the same phenomenon is also shown when @xmath216 , as shown in figs . [
fig9](b1 ) and ( b2 ) . on the other hand , since the @xmath7-wave superfluid order parameter vanishes when @xmath207 , figs .
[ fig9 ] ( c1 ) and ( c2 ) show how the pseudogap in the nodal direction continues developing in the @xmath7-wave superfluid phase , to be the most remarkable at @xmath9 .
when we introduce the characteristic temperature @xmath217 as the temperature below which a dip structure appears in @xmath222 , we obtain fig .
[ fig10 ] . because @xmath91 in the nodal direction ( @xmath207 ) , we may regard @xmath223 as the pseudogap temperature @xcite in this momentum direction , below which strong pairing fluctuations induce a pseudogap in ardos @xmath206 . in this case , one may call the region surrounded by @xmath223 and @xmath9 the pseudogap regime . in the @xmath7-wave superfluid phase .
we take @xmath224 .
each line is offset by 0.03 .
( b ) mean - field result at @xmath225 , which is obtained by ignoring the self - energy correction @xmath226 in eq .
( [ eq.10 ] ) in calculating ardos .
, width=302 ] in the case of @xmath227 , @xmath217 also has the meaning of the pseudogap temperature , when @xmath228 . on the other hand , @xmath229 is lower than @xmath24 around @xmath174 in fig.[fig10 ] , which means that the superfluid gap does not appear in ardos when @xmath230 . as mentioned previously , since @xmath231 in this intermediate coupling regime ,
single - particle excitations around @xmath232 dominantly contribute to ardos around @xmath12 . because of this
, a small superfluid excitation gap by a small @xmath7-wave superfluid order parameter , @xmath233 , around @xmath232 is easily smeared out by strong pairing fluctuations existing in this regime even below @xmath24 . since this strong - coupling effect
is gradually suppressed below @xmath24 , ardos starts to exhibit a superfluid gap structure below @xmath217 ( see fig . [
fig11](a ) . ) , to approach the bcs - type superfluid density of states shown in fig .
[ fig11](b ) .
thus , @xmath217 in this regime may be regarded as the characteristic temperature , below which the @xmath7-wave superfluid order overwhelms pairing fluctuations . in fig .
[ fig10 ] , we also plot the temperature at which the fermi chemical potential @xmath34 changes its sign @xcite .
as mentioned previously , in the strong - coupling regime where @xmath234 , since the system is dominated by tightly bound molecules , the @xmath0-wave character of cooper pairs is less important . in the normal state near @xmath24
, this fact gives the isotropic pseudogap size @xmath195 in eq .
( [ eq.32a ] ) . in the @xmath7-wave superfluid phase , the bogoliubov single particle excitation spectrum in the strong coupling regime
, @xmath235 also has the isotropic energy gap @xmath236 , reflecting that the threshold energy of fermi excitations is simply dominated by the binding energy ( @xmath221 ) of a two - body bound molecule .
thus , the @xmath0-wave anisotropy is not important in the right side of this line in fig .
[ fig10 ] , as far as we consider low - energy single - particle excitations .
to summarize , we have discussed strong - coupling properties of a one - component superfluid fermi gas with a uniaxially anisotropic @xmath0-wave pairing interaction ( @xmath1 ) . including @xmath0-wave pairing fluctuations within a @xmath6-matrix approximation
, we determined the two superfluid phase transition temperatures @xmath24 , which gives the phase boundary between the normal state and the @xmath7-wave superfluid state , and @xmath9 ( @xmath10 ) , which gives the phase boundary between the @xmath7-wave and @xmath11-wave superfluid states .
we examined single - particle excitations near @xmath24 , as well as near @xmath9 . in the normal state near @xmath24
, we showed that strong pairing fluctuations in the @xmath7-wave cooper channel induce an anisotropic pseudogap phenomenon where a pseudogap structure in the angle - resolved density of states ( ardos ) is the most remarkable in the @xmath7 direction .
we also showed that this pseudogap continuously changes to the @xmath7-wave superfluid gap below @xmath24 . on the other hand ,
the pseudogap was found to continue developing below @xmath24 in the nodal direction ( @xmath25 ) of the @xmath7-wave superfluid order parameter , to be the most remarkable at @xmath9 . since pairing fluctuations
are simply suppressed in an isotropic @xmath16-wave superfluid state , this phenomenon is characteristic of a @xmath0-wave fermi superfluid with a nodal superfluid order parameter and with plural superfluid phases . to characterize the anisotropic pseudogap phenomenon , we determined the characteristic temperature @xmath217 , below which a dip structure appears in ardos . in this paper
, we have considered the normal state , as well as the @xmath7-wave superfluid phase . to obtain the complete understanding of a @xmath0-wave superfluid fermi gas with a uniaxially anisotropic @xmath0-wave interaction
, we need to also examine the @xmath11-wave superfluid phase below @xmath9 .
in addition , for simplicity , we employed the bcs hamiltonian with a @xmath0-wave pairing interaction , which implicitly assumes a broad feshbach resonance . in this
regard , all current experiments are using a narrow @xmath0-wave feshbach resonance , so that it is an important problem to clarify how the resonance width affects strong - coupling properties of a @xmath0-wave superfluid fermi gas . since
a @xmath0-wave superfluid state is known to be sensitive to spatial inhomogeneity , inclusion of a realistic harmonic trap also remains as our future problem .
the realization of a @xmath0-wave superfluid fermi gas is an exciting challenge , in order to qualitatively go beyond the current stage of cold fermi gas physics that the @xmath16-wave fermi superfluid has only been realized .
since the anisotropic pairing is a crucial key in a @xmath0-wave fermi superfluid , our results would contribute to understanding how this character affects many - body properties of an ultracold fermi gas , especially in the superfluid phase .
we would like to thank m. sigrist , s. tsuchiya , s. watabe , r. watanabe and t. kashimura for useful discussions .
this work was supported by kipas project in keio university .
yo was also supported by grant - in - aid for scientific research from mext and jsps in japan ( 25105511 , 25400418 , 15h00840 ) .
99 see , for example , m. sigrist , and k. ueda , rev .
63 * , 239 ( 1991 ) .
d. vollhardt , and p. wlfle , _ the superfluid phases of helium 3 _ , ( taylor and francis , new york , 2002 ) .
v. p. mineev , and k. v. samokhin , _ introduction to unconventional superconductivity _ , ( gordon and breach , new york , amsterdam , 1999 ) . c. a. regal , m. greiner , and d. s. jin , phys .
lett . * 92 * , 040403 ( 2004 ) .
m. w. zwierlein , c. a. stan , c. h. schunck , s. m. f. raupach , a. j. kerman , and w. ketterle , phys .
lett . * 92 * , 120403 ( 2004 ) .
j. kinast , s. l. hemmer , m. e. gehm , a. turlapov , and j. e. thomas , phys .
. lett . * 92 * , 150402 ( 2004 ) .
m. bartenstein , a. altmeyer , s. riedl , s. jochim , c. chin , j. h. denschlag , and r. grimm , phys .
92 * , 203201 ( 2004 ) . c. a. regal , c. ticknor , j. l. bohn , and d. s. jin , phys
* 90 * , 053201 ( 2003 ) . c. a. regal , c. ticknor , j. l. bohn , and d. s. jin , nature ( london ) * 424 * , 47 ( 2003 ) . c. ticknor , c. a. regal , d. s. jin , and j. l. bohn , phys .
a * 69 * , 042712 ( 2004 ) .
j. zhang , e. g. m. van kempen , t. bourdel , l. khaykovich , j. cubizolles , f. chevy , m. teichmann , l. tarruell , s. j. j. m. f. kokkelmans , and c. salomon , phys .
a * 70 * , 030702(r ) ( 2004 ) . c. h. schunck , m. w. zwierlein , c. a. stan , s. m. f. raupach , w. ketterle , a. simoni , e. tiesinga , c. j. williams , and p. s. julienne , phys
a * 71 * , 045601 ( 2005 ) .
k. gnter , t. stferle , h. moritz , m. khl , and t. esslinger , phys .
lett . * 95 * , 230401 ( 2005 ) .
j. p. gaebler , j. t. stewart , j. l. bohn , and d. s. jin , phys .
lett . * 98 * , 200403 ( 2007 ) .
y. inada , m. horikoshi , s. nakajima , m. kuwata - gonokami , m. ueda , and t. mukaiyama , phys .
lett . * 101 * , 100401 ( 2008 ) .
j. fuchs , c. ticknor , p. dyke , g. veeravalli , e. kuhnle , w. rowlands , p. hannaford , and c. j. vale , phys .
a * 77 * , 053616 ( 2008 ) .
t. nakasuji , j. yoshida , and t. mukaiyama , phys .
a * 88 * , 012710 ( 2013 ) .
r. a. w. maier , c. marzok , and c. zimmermann , and ph . w. courteille , phys . rev . a * 81 * , 064701 ( 2010 ) .
v. gurarie , l. radzihovsky , and a. v. andreev , phys .
lett . * 94 * , 230403 ( 2005 ) .
y. ohashi , phys .
lett . * 94 * , 050403 ( 2005 ) . t. l. ho and r. b. diener , phys .
rev . lett .
* 94 * , 090402 ( 2005 ) .
s. s. botelho and c. a. r. s demelo , j. low temp . phys . *
140 * , 409 ( 2005 ) .
m. iskin and c. a. r. s de melo , phys .
rev b * 72 * , 224513 ( 2005 ) .
m. iskin and c. a. r. s de melo , phys .
lett . * 96 * , 040402 ( 2006 ) .
cheng and s .- k .
yip , phys . rev .
95 * , 070404 ( 2005 ) , phys .
b * 73 * , 064517 ( 2006 ) .
j. levinsen , n. r. cooper , and v. gurarie , phys .
lett . * 99 * , 210402 ( 2007 ) . v. gurarie , l. radzihovsky , ann .
phys . * 322 * , 2 ( 2007 ) .
e. grosfeld , n. r. cooper , a. stern , and r. ilan phys .
b * 76 * , 104516 ( 2007 )
. m. iskin and c. j. williams , phys .
a * 77 * , 041607(r ) ( 2008 ) .
t. mizushima , m. ichioka , and k. machida , phys .
* 101 * , 150409 ( 2008 ) .
han , y .- h .
chan , w. yi , a. j. daley , s. diehl , p. zoller , and l .- m .
duan phys .
lett . * 103 * , 070404 ( 2009 ) .
t. mizushima and k. machida , phys .
a * 81 * , 053605 ( 2010 ) , _
ibid_. , * 82 * , 023624 ( 2010 ) .
d. inotani , r. watanabe , m. sigrist , y. ohashi , phys .
a. * 85 * , 053628 ( 2012 ) .
m. jona - lasinio , l. pricoupenko , and y. castin , phys . rev . a * 77 * , 043611 ( 2008 ) .
j. levinsen , n. r. cooper , and v. gurarie , phys .
a * 78 * , 063616 ( 2008 ) . f. chevy , e. g. m. van kempen , t. bourdel , j. zhang , l. khaykovich , m. teichmann , l. tarruell , s. j. j. m. f. kokkelmans , and c. salomon , phys .
a * 71 * , 062710 ( 2005 ) .
j. t. stewart , c. a. regal , and d. s. jin , nature * 454 * , 744 ( 2008 ) .
gaebler , j. t. stewart , t. e. drake , d. s. jin , a. perali , p. pieri , and g. c. strinati , nature physics , * 6 * , 569 ( 2010 ) .
a. perali , f. palestini , p. pieri , g. c. strinati , j. t. stewart , j. p. gaebler , t. e. drake , and d. s. jin , phys .
* 106 * , 060402 ( 2011 ) .
a. perali , p. pieri , g. c. strinati , and c. castellani , phys .
b * 66 * , 024510 ( 2002 ) .
s. tsuchiya , r. watanabe , y. ohashi , phys .
a * 80 * , 033613 ( 2009 ) , _ ibid .
_ , * 82 * , 033629 ( 2010 ) , * 84 * , 043647 ( 2011 ) .
q. chen , and k. levin , phys .
lett . * 102 * , 190402 ( 2009 ) .
h. hu , x .- j .
liu , p. d. drummond , and h. dong , phys .
lett . * 104 * , 240407 ( 2010 ) .
p. magierski , g. wlazlowski , and a. bulgac , phys .
* 107 * , 145304 ( 2011 ) .
r. watanabe , s. tsuchiya , y. ohashi , phys .
a. * 82 * , 043630 ( 2010 ) .
y. ohashi , and s. takada , j. phys .
. jpn . * 66 * , 2437 ( 1997 ) .
j. r. schrieffer , _ theory of superconductivity _ addison - wesley , ny 1964 ) . under the nambu representation , the bubble diagrams in fig .
[ fig1](b ) describe pairing fluctuations .
d. j. thouless , ann .
* 10 * , 553 ( 1960 ) .
p. nozires and s. schmitt - rink , j. low .
* 59 * , 195 ( 1985 ) .
c. a. r. s de melo , m. randeria , and j. r. engelbrecht , phys .
lett . * 71 * , 3202 ( 1993 ) . j. r. engelbrecht , m. randeria , and c. a. r. s de melo , phys .
b * 55 * , 15153 ( 1997 )
. y. ohashi , and a. griffin , phys .
a * 67 * 063612 ( 2003 ) .
t. kashimura , r. watanabe , and y. ohashi , phys .
a * 86 * , 043622 ( 2012 ) .
y. ohashi , j. phys .
. jpn . * 74 * , 2659 ( 2005 ) . since we are taking a finite value of the momentum cutoff @xmath114 , the
strong - coupling limit " in this paper means the region where @xmath237 , but ( 1 ) the molecular binding energy , ( 2 ) the absolute value of the chemical potential @xmath238 , as well as ( 3 ) the superfluid order parameter , are still much smaller than the cutoff - energy @xmath239 , to avoid any unphysical effect coming from this high - energy cutoff .
j. leggett , in _ modern trends in the theory of condensed matter _
( springer - verlag , 1980 ) . c. j. pethick and h. smith , _ bose - einstein condensation in dilute gases _
( cambridge university press , cambridge , 2008 ) sec.8.1 .
we note that , in the strong - coupling region ( @xmath240 ) , the pseudogap opens again with the size @xmath241 , as in the strong - coupling region of @xmath16-wave case , although we do not explicitly show the result here .
of course , this line also exists in the @xmath11-wave superfluid phase below @xmath9 .
however , since we have not calculated @xmath34 in this region , fig , [ fig10 ] only shows @xmath242 . | we investigate superfluid properties of a one - component fermi gas with a uniaxially anisotropic @xmath0-wave pairing interaction , @xmath1 ( where @xmath2 ( @xmath3 is a @xmath4-wave pairing interaction ) .
this type of interaction is considered to be realized in a @xmath5k fermi gas . including pairing fluctuations within a strong - coupling @xmath6-matrix theory , we determine the @xmath7-wave superfluid phase transition temperature @xmath8 , as well as the other phase transition temperature @xmath9 ( @xmath10 ) , below which the superfluid order parameter has the @xmath11-wave symmetry . in the normal state near @xmath8
, @xmath7-wave pairing fluctuations are shown to induce an anisotropic pseudogap phenomenon , where a dip structure in the angle - resolved density of states around @xmath12 is the most remarkable in the @xmath7 direction . in the @xmath7-wave superfluid phase ( @xmath13 ) , while the pseudogap in the @xmath7 direction continuously changes to the superfluid gap , the pseudogap in the perpendicular direction to the @xmath7 axis is found to continue developing , because of enhanced @xmath14-wave and @xmath15-wave pairing fluctuations around the node of the @xmath7-wave superfluid order parameter . since pairing fluctuations
are always suppressed in the isotropic @xmath16-wave superfluid state , this phenomenon is peculiar to an unconventional fermi superfluid with a nodal superfluid order parameter .
since the @xmath0-wave fermi superfluid is the most promising non @xmath16-wave pairing state in an ultracold fermi gas , our results would contribute to understanding how the anisotropic pairing fluctuations , as well as the existence of plural superfluid phases , affect many - body properties of this unconventional fermi superfluid . |
up to the present , most ground state charm baryons containing one @xmath0-quark , from @xmath1 to @xmath2 , have been observed @xcite .
there have been much efforts to find excited charm baryons and recently the experimental evidences for @xmath3 @xcite , @xmath4 and @xmath5 @xcite are reported .
although their quantum numbers are not identified yet , the spin - parity of the @xmath6 is interpreted as @xmath7 , and @xmath8 and @xmath5 decaying to @xmath9 are regarded as candidates for @xmath10 and @xmath11 excited states , respectively , in accordance with the quark model predictions @xcite .
the small mass splittings between @xmath6 and @xmath12 and between @xmath13 and @xmath14 are consistent with the heavy quark spin symmetry @xcite , according to which the hadrons come in degenerate doublets with total spin @xmath15 ( unless @xmath16 , the total angular momentum of the light degrees of freedom , is zero ) in the limit of the heavy quark mass going to infinity .
on the other hand , the excited heavy baryons have been extensively studied not only in various quark / bag models @xcite but also in heavy hadron chiral perturbation theory @xcite and in the bound state approach of the skyrme model @xcite . in the bound state model ,
the heavy baryons are described by bound states of heavy mesons and a soliton @xcite
. a natural explanation of low - lying @xmath17 is one of the success of the bound state approach @xcite , where this @xmath18 state is described by a loosely bound @xmath19-wave @xmath20 meson to soliton .
the same picture was straightforwardly applied to the excited @xmath21 in ref .
the lack of the heavy quark symmetry in this first trial is later supplied by treating the heavy vector mesons on the same footing as the heavy pseudoscalar mesons @xcite , and a generic structure of the heavy baryon spectrum of orbitally excited states is established @xcite .
however , these works were done under the approximation that both the soliton and the heavy mesons are infinitely heavy and so they sit on top of each other .
it is evident that this approximation can not describe well the orbitally and/or radially excited states due to the ignorance of any kinetic effects . in ref .
@xcite , the kinetic effects for the excited states are estimated by approximating the static potentials for the heavy mesons to the quadratic form with the curvature determined at the origin .
such a harmonic oscillator approximation is valid only when the heavy mesons are sufficiently massive so that their motions are restricted to a very small range .
the situation is improved in ref .
@xcite by making an approximate schrdinger - like equation and by incorporating the light vector mesons .
in a recent paper @xcite , we have obtained all the energy levels of the heavy meson bound states by solving _
exactly _ the equations of motion from a given model lagrangian without using any approximations .
( see also refs . @xcite . ) in this paper , we quantize those states by following the standard collective coordinate quantization method to investigate the excited heavy baryon spectra . in the next section ,
we briefly describe our model lagrangian and the way of solving the equations of motion to obtain the bound states .
then , in sec .
iii , we quantize the soliton heavy - meson bound system based on the standard collective coordinate quantization method and derive the mass formula . the resulting mass spectra of @xmath22 , @xmath23 , @xmath24 , and @xmath25 baryons are presented in sec .
iv and compared with the recent experimental observations and with the quark model predictions .
some detailed expressions are given in appendix .
we will work with a simple effective chiral lagrangian of goldstone mesons and heavy mesons @xcite : @xmath26 here , @xmath27 is an effective chiral lagrangian of goldstone pions presented by an @xmath28 matrix field @xmath29 , which is simply taken as the skyrme model lagrangian , @xmath30
^ 2,\ ] ] with the pion decay constant @xmath31 and the skyrme parameter @xmath32 . with the help of the skyrme term , it supports a stable baryon - number-1 soliton solution under the hedgehog configuration @xmath33,\ ] ] where @xmath34 satisfies the boundary conditions , @xmath35 and @xmath36 .
the heavy pseudoscalar and vector mesons containing a heavy quark @xmath37 are represented by anti - isodoublet fields @xmath38 and @xmath39 , with their masses @xmath40 and @xmath41 , respectively . as an example , in case of @xmath42 , they are @xmath43 and @xmath44 meson anti - doublets , @xmath45 the chiral transformations of the fields are defined as @xmath46 where @xmath47 , @xmath48 , @xmath49 , and @xmath50 is an @xmath28 matrix depending on @xmath51 , @xmath52 , and @xmath53 .
the field @xmath53 defines vector and axial vector fields as @xmath54 then , the covariant derivatives are expressed in terms of @xmath55 as @xmath56 and a similar equation for @xmath39 , which defines the field strength tensor of the heavy vector meson fields as @xmath57 . in our lagrangian
, we have a few parameters , @xmath31 , @xmath32 , @xmath40 , @xmath41 , @xmath58 , and @xmath59 , which in principle have to be fixed from the meson dynamics .
we will use the experimental values for the heavy meson masses . in order for the quantized soliton to fit the nucleon and @xmath60 masses
@xcite , the pion decay constant has been adjusted down to @xmath61 mev ( with @xmath62 ) . however , recently it is shown that , taking into account the casimir effect of the fluctuating pions around the soliton configuration @xcite , one can get reasonable nucleon and @xmath60 masses with the empirical value of the pion decay constant ( @xmath6393 mev ) . as for the coupling constants @xmath58 and @xmath59 , there is no sufficient experimental data to fix them .
what is known to us is that , in order for the lagrangian to respect the heavy quark symmetry , they should be related to each other as @xmath64 and the nonrelativistic quark model prediction on the universal constant @xmath65 is @xmath66 @xcite , while the experimentally determined upper bound is @xmath67 @xcite . we will take @xmath58 and @xmath59 as free parameters with keeping the relation ( [ hqr ] ) and the nonrelativistic prediction in mind
. the parameter dependence of the results will be discussed in detail in sec .
the lagrangian ( [ lag ] ) is the simplest version of the heavy meson effective lagrangian and one may include the light vector meson degrees of freedom such as @xmath68 and @xmath69 @xcite and the higher derivative terms to improve the model predictions .
the equations of motion for the heavy mesons can be read off from the lagrangian ( [ lag ] ) as @xmath70 with an auxiliary condition for the vector meson fields @xmath71 which reduces to the lorentz condition @xmath72 in case of the free vector meson fields . to avoid any unnecessary complications associated with the anti - doublet structure of @xmath38 and @xmath39 , we work with @xmath73 and @xmath74 .
when the static hedgehog configuration of @xmath75 is substituted , the vector and axial vector fields become @xmath76 \bbox { ) } , \end{array } \label{va}\ ] ] where @xmath77/r $ ] , @xmath78 and @xmath79 . then , the problem becomes to find the classical eigenmodes ( especially the bound states ) of the heavy mesons moving under the static potentials formed by the soliton field .
the equations are invariant under parity operations and the grand spin " rotation generated by the operator @xmath80 where @xmath81 , @xmath82 , and @xmath83 are the orbital angular momentum , spin , and isospin operator of the heavy mesons , respectively .
this allows us to classify the eigenstates by the grand spin quantum numbers @xmath84 and the parity @xmath85 .
we will denote the set of quantum numbers by @xmath86 @xmath87 , @xmath88 is a quantum number to distinguish the radial excitations ) . for a given grand spin @xmath84 with parity @xmath89 ,
the general wave function of an energy eigenmode can be written as @xmath90 , \end{array } \label{wf}\ ] ] where @xmath91 and @xmath92 is the ( iso)spinor spherical harmonic obtained by combining the eigenstates of @xmath81 and @xmath83 .
the wave functions should be normalized so that the eigenmodes carry a unit heavy flavor number ( @xmath93 and @xmath94 ) .
the normalization condition is given in appendix .
note the different sign convention of the energy in the exponent for the time evolution of the eigenmodes and that @xmath95 @xmath96 $ ] is absent in case of @xmath97 .
substituting eq .
( [ wf ] ) into the equations of motion ( [ eom ] ) and auxiliary condition ( [ ac ] ) , one can obtain coupled differential equations for the radial functions , @xmath98 and @xmath99 ( @xmath100 ) .
( see ref .
@xcite for more details . ) given in fig .
1 is a typical energy spectrum of bound heavy meson states obtained by solving the equations numerically with experimental heavy meson masses ( @xmath101 mev , @xmath102 mev and @xmath103 mev , @xmath104 mev ) , @xmath61 mev , @xmath62 , and @xmath105 .
we present the binding energies defined as @xmath106 . comparing it with the energy spectrum obtained in the infinite heavy mass limit @xcite , one can see that the parity doubling " artifact is removed by the centrifugal energy contribution and there appear many radially excited states . as a trace of the heavy quark symmetry ,
the energy levels come in nearly degenerate doublets with grand spin @xmath107 ( unless @xmath108 ) and parity @xmath109 @xcite , where @xmath110 with the heavy quark spin @xmath111 .
the energy levels are obtained in the soliton - fixed frame , which must be a crude approximation .
the soliton - recoil effects should be incorporated in order for the bound state approach to work well with heavy flavors . in this work , however , we will proceed without incorporating the soliton - recoil effects .
we will discuss some possible modifications of the results in sec .
iv , leaving the rigorous and detailed investigations to our future study .
the soliton heavy - meson bound system described so far does not carry any good quantum numbers except the grand spin , parity , and baryon number . in order to describe baryons with definite spin and isospin quantum numbers
, we should quantize the system by going to the next order in @xmath112 @xcite .
this can be done by introducing collective variables to the zero modes associated with the invariance of the soliton heavy - meson bound system under simultaneous isospin rotation of the soliton together with the heavy meson fields : @xmath113 where @xmath114 and @xmath115 is an su(2 ) matrix . the subscript
bf " is to denote that they are the fields in the body - fixed ( isospin co - moving ) frame .
( hereafter , we will drop it to shorten the notation and all the heavy meson fields appearing in equations are those in the body - fixed frame unless specified . ) assuming sufficiently slow collective rotation , we will work in the born - oppenheimer approximation where the bound heavy mesons remain in an unchanged classical eigenmode .
introduction of the collective variables as eq .
( [ cv ] ) leads us to an additional lagrangian of @xmath116 , @xmath117 where the angular velocity @xmath118 of the collective rotation is defined by @xmath119 and @xmath120 is the moment of inertia of the soliton @xcite .
the explicit form of @xmath121 is @xmath122 \\ & & \hskip 3em \mbox { } - { \textstyle\frac12 } f_q ( \phi { \bf t}_a \phi^{*\dagger}_0 + \phi^*_0 { \bf t}_a \phi^\dagger ) - { \textstyle\frac{i}{2 } } g_q \varepsilon^{ijk } \phi^{*i } \ { { \bf t}_v , a^j \}_+ \phi^{*k\dagger } \\ & & \hskip 3em \mbox { } - { \textstyle\frac12 } g_q \varepsilon^{ijk } [ ( d^i \phi^{*j } ) { \bf t}_a \phi^{*k\dagger } + \phi^{*k } { \bf
t}_a d^i \phi^{*j\dagger } ] \bigr\ } , \end{array}\label{pth}\ ] ] where @xmath123 and @xmath124 with @xmath125 , @xmath126 , and @xmath127 . note that it is nothing but the isospin current of the heavy mesons interacting with goldstone bosons ( modulo the sign ) as discussed in ref .
@xcite .
the spin operator @xmath128 and isospin operator @xmath83 of the system can be obtained by applying the nther theorem to the invariance of the lagrangian under the corresponding rotations : @xmath129 where @xmath130 is the rotor spin conjugate to the collective variables , @xmath131 @xmath132 is the grand spin operator of the heavy meson fields ( in the isospin co - moving frame ) and @xmath133 $ ] is the adjoint representation of the collective variables .
note that the grand spin operator plays the role of the spin operator for the heavy mesons , that is , their isospin is transmuted into a part of the spin .
the physical heavy baryon states with spin - parity @xmath134 and isospin @xmath135 can be obtained by combining the rotor spin eigenstates and the heavy meson bound states of grand spin @xmath136 with the help of the clebsch - gordan coefficients : [ state ] @xmath137 with @xmath138 . here , @xmath139 ( @xmath140 )
denotes the eigenstate of the rotor - spin operator @xmath141 : @xmath142 and @xmath143 is the single - particle fock state of the heavy meson fields where one classical eigenmode of the corresponding grand spin quantum number is occupied : @xmath144 as an artifact of large @xmath145 feature of the skyrme model , one may have baryon states with higher isospin @xmath146 .
we will restrict our considerations to the heavy baryons with @xmath147 @xmath148 and @xmath149 @xmath150 . to be precise , in eq .
( [ state ] ) , one may have to sum over the possible @xmath136 as @xmath151 with the expansion coefficients @xmath152 to be determined by diagonalizing the hamiltonian .
however , as far as the heavy baryon states are concerned , the mixing effects are rather small . it is shown in ref .
@xcite that there is no mixing even when two states become degenerate in the @xmath153 limit .
( such a mixing effect plays the most important role in establishing the heavy quark symmetry in the pentaquark baryons @xcite . )
thus , we will involve only one single - particle fock state in the combination ( [ state ] ) .
the physical baryons should be the eigenstates of the hamiltonian and their masses come out as eigenvalues .
the hamiltonian can be obtained by taking the legendre transformation with the collective variables and the heavy meson fields taken as dynamical degrees of freedom . up to the order of @xmath112
, we have [ h ] @xmath154 where @xmath155 is the hamiltonian of @xmath156 .
the hamiltonian of the leading order in @xmath145 is the soliton mass @xcite : @xmath157 and @xmath158 is the hamiltonian of the heavy meson fields which yields the eigenenergy @xmath159 when acts on the single - particle fock state @xmath160 : @xmath161 finally , the hamiltonian of order of @xmath112 arising from the collective rotation is in a form of @xmath162 we will take the @xmath112 order term as a perturbation .
then , the mass of the heavy baryon state ( [ state ] ) is obtained as @xmath163 where @xmath159 is the eigenenergy of the heavy meson bound state involved in the construction of the state @xmath164 . if only one single - particle fock state @xmath165 is involved in eq .
( [ state ] ) , we can write the mass formula in a more convenient form as @xmath166.\ ] ] here , we have used the wigner - eckart theorem to express the expectation value of @xmath121 as @xmath167 which defines the hyperfine splitting " constant @xmath168 .
the explicit expressions for @xmath168 are given in appendix .
in evaluating the expectation value of @xmath169 , we have used the fact that @xmath121 is the isospin operator ( with opposite sign ) , which implies @xmath170
as for the @xmath171 baryons that are constructed with the @xmath147 rotor spin state and one single - particle fock state @xmath172 , the mass formula can be further simplified as @xmath173 where @xmath174 is the nucleon mass ( @xmath175 ) and @xmath176 .
thus , the mass spectrum of @xmath171 baryons is exactly the same as fig .
1 with replacing the @xmath177 line by the @xmath178 threshold . however , the @xmath171 spectrum obtained with the parameters of fig .
1 ( set 1 ) is not at the level of being compared with experiments . as can be seen in fig .
1 , the binding energy ( @xmath179 mev ) and the mass splitting ( @xmath180 mev ) between the first excited @xmath181 and the ground state are too small compared with the experimental values , @xmath182 and @xmath183 mev , respectively . however , we can easily improve the situation by adjusting the parameters within a reasonable range .
table i summarizes the parameter sets that we will examine and the parameter dependence of @xmath22 spectra are shown in fig .
2 . what we want to have is more deeply bound states with wider level splittings , which can be achieved if we have a deeper and narrower interacting potentials in the equations of motion for the heavy mesons .
one way of obtaining such potentials in a given model lagrangian is to take the empirical value for @xmath31 instead of @xmath61 mev . since the soliton wave function @xmath34 is only a function of a dimensionless variable @xmath184 ( in the chiral limit ) , the functions @xmath185 and @xmath186 appearing in potentials scale with the factor @xmath187 .
furthermore , the soliton mass @xmath188 and the moment of inertia @xmath120 come out in the form of @xcite @xmath189 with dimensionless quantities @xmath190 and @xmath191 that are independent of @xmath32 and @xmath31 .
if we are to have a correct @xmath60-@xmath192 mass splitting ( @xmath193 ) , we have to fix the value of @xmath32 so that @xmath194 does not change .
this condition yields @xmath195 when @xmath196 mev , which implies that the soliton mass becomes so heavy as 1.4 gev .
we expect that the casimir energy of fluctuating pions @xcite can reduce it down to 0.87 gev . in this work , for a comparison ,
we fix the nucleon mass to 940 mev for all parameter sets . compared with @xmath31=64.5 mev and @xmath32=5.45
, @xmath187 becomes 1.3 times larger and thus the potential becomes deeper and narrower by the same factor .
this is shown by the dashed line of fig .
3 . the change in @xmath31 alone ( set 2 ) helps the ground @xmath22 mass to come down to 2339 mev with the @xmath197-@xmath22 mass difference being 270 mev .
another way of improving the results is simply to take a larger @xmath198 value ( with putting aside the experimental upper limit on @xmath198 for a while ) , which makes the potential deeper . the dotted line in fig .
3 is what obtained by varying @xmath59 and @xmath199 to @xmath200 while keeping @xmath31=64.5 mev ( set 3 ) .
surprisingly , this nearly 20% change in coupling constants results in about 50% enhancement in the binding energy , while the mass splitting is not so much improved compared with that of set 1 . by the same way
, we take the empirical value for @xmath31 and vary @xmath59 so that the ground @xmath22 mass becomes close to the experimental value , which is achieved with @xmath201 ( set 4 ) .
this parameter set yields comparable @xmath22 mass spectrum to the experiments , which looks quite encouraging .
furthermore , if we break the heavy quark spin symmetric relation , @xmath202 , between the two coupling constants , we can obtain more realistic mass splitting between @xmath203 and @xmath204 .
as an example , we choose @xmath205 and @xmath206 with @xmath196 mev ( set 5 ) .
unfortunately , these coupling constants are not close to the recent estimates of @xmath207 @xcite .
we regard this fact as an indication of the important role of higher order corrections such as light vector mesons . in fig .
4 , we present our results on @xmath22 spectrum ( obtained with parameter set 5 ) together with the experimental values and the other model calculations ; sm ( skyrme model with the heavy pseudoscalar mesons only ) @xcite , qm1 ( quark model ) @xcite , qm2 @xcite , and qm3 @xcite . our result can compete with the quark model calculations quantitatively .
especially , one can notice that it becomes much more improved compared with the first trial , sm @xcite , in the skyrme model .
one may improve the result by adjusting all the parameters for the best fit .
what we have done in this work is just to vary two coupling constants around the values given by the nonrelativistic quark model prediction and the heavy quark symmetric relation , _ i.e. _ , @xmath208 . as for the other parameters
, we used the empirical value for @xmath31 with @xmath32 being fixed by @xmath60-@xmath192 mass splitting , and experimental values for the heavy meson masses .
given in fig .
5 are the spectrum of @xmath23 of our prediction ( obtained with set 5 ) and the other model calculations . as a reference line
, the ground state of @xmath22 is adopted .
our result is again comparable to the others .
however , the splitting between @xmath209 [ @xmath210 and @xmath23 appears too small compared with the experimental data .
note that the same is true for the quark models except qm2 .
it is also interesting to note that one can not improve this situation simply by making the hyperfine constant @xmath0 larger in any way . by eliminating @xmath168 and @xmath120 in the mass formula ( [ massform ] )
, we obtain a model - independent relation @xmath211 ( the same model - independent mass relation holds in the nonrelativistic quark model of de rjula , georgi , and glashow @xcite . ) thus , when our model successfully reproduces all the experimental values for @xmath174 , @xmath212 , @xmath213 , and @xmath214 , we get @xmath215 as our best prediction .
it is only the half of the value evaluated with the recent @xmath216 @xcite .
we present the @xmath24 mass spectrum obtained with this parameter set in fig . 6 with the other model calculations .
the parameter set used for the charm baryons does not work well in the bottom sector ; set 5 yields the ground @xmath24 mass as 5492 mev which is @xmath217 mev below the experimental value .
we may repeat the same process of varying the @xmath59 ( with keeping the empirical value 93 mev for @xmath31 ) to fit the @xmath24 mass of 5641 mev .
we also expect that the heavy quark symmetry relation ( [ hqr ] ) holds well in the bottom sector .
this process leads us to @xmath218 ( set 6 ) . the results with this parameter set
are also given in fig .
the mass splitting ( @xmath219 mev ) between the excited @xmath220 and the ground state @xmath24 appears much smaller than that of the @xmath22 given in fig .
4 , while the quark model calculations show nearly independent mass splittings whether the heavy constituent is a @xmath0-quark ( @xmath221 mev ) or a @xmath222-quark ( @xmath223@xmath224 mev ) . together with the differences in coupling constants fitting the charm baryons and the bottom baryons ,
this apparent difference in the mass splitting is certainly at odds with the _ heavy quark flavor symmetry .
_ such a heavy quark flavor symmetry is expected to be somehow broken because of the mass difference between the @xmath0-quark and @xmath222-quark .
however , since both are much heavier than the typical scale of the strong interaction ( @xmath225 mev ) , the actual amount of the symmetry breaking in nature that occurs at the order of @xmath226 would not be so large .
such a behavior can be seen also in the @xmath25 spectrum given in fig .
since there is no experimental data for the @xmath25 baryons , we can only compare our results with the quark model predictions .
one can find that the mass splitting between the ground @xmath24 and the @xmath227 is 180 @xmath63 190 mev , which is comparable to the quark model predictions .
also the small mass splitting ( @xmath63 10 mev ) between @xmath227 and @xmath228 is still consistent with the quark model predictions .
however , as in the @xmath24 spectrum , the excitation energy ( @xmath63 170 mev ) of @xmath229 appears again smaller than the quark model values ( @xmath63 280 mev ) .
it may be the ignorance of the soliton - recoil effect in our work that causes the larger break down of the heavy quark flavor symmetry than what is actually implied in the model . in order to see this ,
let us go back to fig .
we can see that the kinetic effect reduces the binding energy of the lowest @xmath43 @xmath230 meson bound state by 410 ( 240 ) mev from its infinitely heavy mass limit @xmath231 mev @xcite .
note that the ratio of the kinetic effects ( @xmath232 ) and the ratio of energy splittings between the first excited state and the ground state @xmath233 are very close to the square root of the ( inverse ) mass ratio ( @xmath234 ) .
one can easily understand this feature in the harmonic oscillator approximation .
thus , in our working frame , the fact that @xmath235 mesons are 2.6 times heavier than @xmath43 mesons becomes directly reflected in the results . a simple way of estimating the soliton - recoil effect is to use the reduced mass " of the soliton heavy - meson system , as discussed in refs .
@xcite . with the soliton mass about 1 gev ,
the reduced masses of the @xmath43 mesons and @xmath235 mesons become @xmath236 gev and @xmath237 gev , respectively .
then , the use of these small reduced masses can widen the energy splittings and their small ratio @xmath238 will not break the heavy quark flavor symmetry so seriously .
( see also fig .
@xcite . ) on the other hand , it will require stronger potentials to overcome the larger kinetic energies , which should be supplied by including the light vector mesons and/or higher derivative terms into the lagrangian @xcite . in summary
, we have studied the heavy baryon spectrum in the bound state approach to the skyrme model by using the exactly - solved heavy meson bound states of a given lagrangian .
our results are _ qualitatively _ and/or _ quantitatively _ comparable to the experimental observations and the quark model calculations in the charm / bottom sector .
the nearly degenerate doublets in the spectrum are consistent with the heavy quark _ spin _
symmetry , and our work has a great improvement compared with the first trial @xcite of this model . however , the absence of the soliton - recoil in our framework breaks the heavy quark _ flavor _
symmetry more than the model really implies . to be consistent with both the heavy quark spin and flavor symmetry , such a soliton - recoil effect should be incorporated into the picture .
one of us ( b .- y.p ) thanks the institute for nuclear theory at the university of washington and the department of physics and astronomy of the university of south carolina for their hospitality during the completion of this work .
this work was supported in part by the national science council of roc under grant no .
nsc84 - 2811-m002 - 036 and in part by the korea science and engineering foundation through the src program . in this appendix
, we present the normalization condition of the heavy meson fields and the explicit form of the hyperfine constants . as discussed in sec .
ii , the heavy meson fields are normalized to give a unit heavy flavor number . for a given grand spin @xmath136 with parity @xmath239 , this condition can be written explicitly as @xmath240 \bigr\},\end{aligned}\ ] ] where the constants @xmath241 , @xmath242 , and @xmath243 are written in terms of @xmath136 as @xmath244 the @xmath0-value defined in eq .
( [ theta ] ) can be written as @xmath245 where the functionals @xmath246 , @xmath247 , and @xmath248 are obtained by inserting eqs .
( [ va ] ) , ( [ wf ] ) , and ( [ tva ] ) into eq . ( [ pth ] ) .
their explicit expressions are as follows : @xmath249 \nonumber \\ & & \mbox { } + \varepsilon_n |\varphi^*_3|^2 [ ( t_1+t_2)\lambda_\mp + t_2\mu_\mp ] + \varepsilon_n ( \varphi_2^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_2^ * ) \mu_\pm \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\prime\dagger } \varphi_1^ * + \varphi_1^{*\dagger } \varphi_0^{*\prime } ) ( t_1+t_2 ) \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\dagger } \varphi_2^ * + \varphi_2^{*\dagger } \varphi_0^ * ) \ { \upsilon [ ( t_1+t_2 ) \lambda_\pm + t_2 \mu_\pm ] \gamma_\pm + \left ( \frac{1}{r } + \upsilon\gamma_\mp \right)\mu_\pm \ } \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_0^ * ) \ { \upsilon \gamma_\pm \mu_\pm + \left(\frac{1}{r } + \upsilon \gamma_\mp \right ) [ ( t_1+t_2)\lambda_\mp + t_2\mu_\mp ] \ } \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_2^ * ) ( a_1+a_2)(t_1+t_2)\lambda_\mp - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } k_2 + k_2^\dagger \varphi_2^ * ) t_3 \mu_\mp \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } k_3 + k_3^\dagger \varphi_2^ * ) t_3 \mu_\pm + { \textstyle\frac12 } g_q ( \varphi_3^{*\dagger } k_2 + k_2^\dagger \varphi_3^ * ) t_3 ( 2+\mu_\pm ) \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_3^{*\dagger } k_3 + k_3^\dagger \varphi_3^ * ) t_3 \mu_\pm \bigr\},\end{aligned}\ ] ] @xmath250 + \varepsilon_n ( \varphi_2^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_2^ * ) ( 1-\gamma_\pm\mu_\pm ) \nonumber \\ & & \mbox { } - { \textstyle\frac12 } ( \varphi_0^{*\prime\dagger } \varphi_1^ * + \varphi_1^{*\dagger } \varphi_0^{*\prime } ) ( 2t_1+t_2 ) \gamma_\pm \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\dagger } \varphi_2^ * + \varphi_2^{*\dagger } \varphi_0^ * ) \ { \upsilon t_2 ( 1-\lambda_\pm-\mu_\pm ) \gamma_\pm^2 + \left ( \frac{1}{r } + \upsilon\gamma_\mp \right)(1-\gamma_\pm\mu_\pm ) \ } \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_0^ * ) \ { \upsilon \gamma_\pm ( 1-\gamma_\pm\mu_\pm ) + \left ( \frac{1}{r } + \upsilon\gamma_\mp \right ) [ ( 2t_1+t_2 ) \mu_\mp - t_2 \gamma_\pm \mu_\mp ] \ } \nonumber \\ & & \mbox { } - { \textstyle\frac12 } f_q ( \varphi^\dagger \varphi_0^ * + \varphi_0^{*\dagger } \varphi ) t_3 \gamma_\pm - { \textstyle\frac12 } g_q ( \varphi_1^{*\dagger } \varphi_2^ * + \varphi_2^{*\dagger } \varphi_1^ * ) a_1 t_1 \gamma_\pm \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_1^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_1^ * ) a_1 t_1 ( 1+\gamma_\pm ) \nonumber \\ & & \mbox { } - { \textstyle\frac12 }
g_q [ |\varphi_2^*|^2 + |\varphi_3^*|^2 ( 1+\gamma_\pm ) ] ( a_1+a_2)(t_1+t_2 ) \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_2^ * ) ( a_1+a_2)(t_1+t_2 ) \mu_\mp + { \textstyle\frac12 } g_q ( \varphi_1^{*\dagger } k_1 + k_1^\dagger \varphi_1^ * ) t_3 \gamma_\pm \nonumber \\ & & \mbox { } + { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } k_2 + k_2^\dagger \varphi_2^ * ) t_3 ( 1+\gamma_\pm)(1+\mu_\mp ) \nonumber \\ & & \mbox { } + { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } k_3 + k_3^\dagger \varphi_2^ * ) t_3 ( 1+\mu_\pm ) \gamma_\pm \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_3^{*\dagger } k_2 + k_2^\dagger \varphi_3^ * ) t_3 ( 2+\mu_\pm ) \gamma_\pm \nonumber \\ & & \mbox { } + { \textstyle\frac12 } g_q ( \varphi_3^{*\dagger } k_3 + k_3^\dagger \varphi_3^ * ) t_3 [ 1 + \mu_\pm ( 1+\gamma_\pm ) ] \bigr\},\end{aligned}\ ] ] @xmath251 \nonumber \\ & & \mbox { } - \varepsilon_n |\varphi_3^*|^2 [ ( 2t_1+t_2)\mu_\mp+t_2(\gamma_\mp\mu_\mp-1 ) ] + \varepsilon_n ( \varphi_2^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_2^ * ) \gamma_\pm\mu_\pm \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\prime\dagger } \varphi_1^ * + \varphi_1^{*\dagger } \varphi_0^{*\prime } ) ( 2t_1+t_2 ) \gamma_\pm \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\dagger } \varphi_2^ * + \varphi_2^{*\dagger } \varphi_0^ * ) \ { \upsilon t_2 [ 1+\gamma_\mp(1-\lambda_\pm-\mu_\pm ) ] \gamma_\pm + \left ( \frac{1}{r } + \upsilon\gamma_\mp \right)\gamma_\pm\mu_\pm \ } \nonumber \\ & & \mbox { } + { \textstyle\frac12 } ( \varphi_0^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_0^ * ) \ { \upsilon \gamma_\pm^2 \mu_\pm - \left ( \frac{1}{r } + \upsilon\gamma_\mp \right ) [ ( 2t_1+t_2 ) \mu_\mp + t_2 ( \gamma_\mp \mu_\mp - 1 ) ] \ } \nonumber \\ & & \mbox { } + { \textstyle\frac12 } f_q ( \varphi^\dagger \varphi_0^ * + \varphi_0^{*\dagger } \varphi ) t_3 \gamma_\pm - { \textstyle\frac12 } g_q ( \varphi_1^{*\dagger } \varphi_2^ * + \varphi_2^{*\dagger } \varphi_1^ * ) a_1 t_1 ( 1+\gamma_\mp ) \nonumber \\ & & \mbox { } + { \textstyle\frac12 } g_q ( \varphi_1^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_1^ * ) a_1 t_1 \gamma_\pm + { \textstyle\frac12 } g_q |\varphi_3^*|^2 ( a_1+a_2)(t_1+t_2 ) \gamma_\pm \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } \varphi_3^ * + \varphi_3^{*\dagger } \varphi_2^ * ) ( a_1+a_2)(t_1+t_2 ) ( 1-\mu_\mp ) - { \textstyle\frac12 } g_q ( \varphi_1^{*\dagger } k_1 + k_1^\dagger \varphi_1^ * ) t_3 \gamma_\pm \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } k_2 + k_2^\dagger \varphi_2^ * ) t_3 [ \gamma_\pm+\mu_\mp(1+\gamma_\pm ) ] \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_2^{*\dagger } k_3 + k_3^\dagger \varphi_2^ * ) t_3 [ \gamma_\pm(1+\mu_\pm)-1 ] \nonumber \\ & & \mbox { } + { \textstyle\frac12 } g_q ( \varphi_3^{*\dagger } k_2 + k_2^\dagger \varphi_3^ * ) t_3 [ 1+\gamma_\pm(2+\mu_\pm ) ] \nonumber \\ & & \mbox { } - { \textstyle\frac12 } g_q ( \varphi_3^{*\dagger } k_3 + k_3^\dagger \varphi_3^ * ) t_3 ( 1+\gamma_\pm ) \mu_\pm \bigr\}.\end{aligned}\ ] ] the functionals @xmath252 , @xmath253 , and @xmath254 are defined by @xmath255 which gives @xmath256 to obtain those formulas , we have used the conjugate form of eq .
( [ wf ] ) : @xmath257 , \end{array}\ ] ] where all the operators act on the _ right - hand side_. as a specific example , the @xmath0-values for the @xmath260 states are given below : @xmath261 \nonumber \\ & & \mbox { } + { \textstyle\frac13 } ( \varphi_0^{*\prime\dagger } \varphi^*_1 + \varphi_1^{*\dagger } \varphi_0^{*\prime } ) ( 3t_1+t_2 ) + { \textstyle\frac23 } ( \varphi_0^{*\dagger } \varphi^*_2 + \varphi_2^{*\dagger } \varphi_0^ { * } ) \upsilon ( t_1+t_2 ) \nonumber \\ & & \mbox { } + { \textstyle\frac23 } f_q ( \varphi^\dagger \varphi^*_0 + \varphi_0^{*\dagger } \varphi ) t_3 + { \textstyle\frac23 } g_q ( \varphi_1^{*\dagger } \varphi^*_2 + \varphi_2^{*\dagger } \varphi_1^ { * } ) [ a_1 t_1 + 2 \left ( \frac{1}{r } - \upsilon \right ) t_3 ]
\nonumber \\ & & \mbox { } - { \textstyle\frac23 } g_q |\varphi^*_2|^2 ( a_1+a_2)(t_1+t_2 ) \bigr\},\end{aligned}\ ] ] and @xmath262 \nonumber \\ & & \mbox { } + { \textstyle\frac13 } ( \varphi_0^{*\prime\dagger } \varphi^*_1 + \varphi_1^{*\dagger } \varphi_0^{*\prime } ) ( -t_1+t_2 ) - { \textstyle\frac23 } ( \varphi_0^{*\dagger } \varphi^*_3 + \varphi_3^{*\dagger } \varphi_0^ { * } ) \left ( \frac{1}{r } - \upsilon \right)(t_1+t_2 ) \nonumber \\ & & \mbox { } - { \textstyle\frac23 } f_q ( \varphi^\dagger \varphi^*_0 + \varphi_0^{*\dagger } \varphi ) t_3 - { \textstyle\frac23 } g_q ( \varphi_1^{*\dagger } \varphi^*_3 + \varphi_3^{*\dagger } \varphi_1^ { * } ) ( a_1 t_1 - 2 \upsilon t_3 ) \nonumber \\ & & \mbox { } - { \textstyle\frac23 } g_q |\varphi^*_3|^2 ( a_1+a_2)(t_1+t_2 ) \bigr\}.\end{aligned}\ ] ] cleo collaboration , d. acosta _ et al .
_ , in lepton and photon interactions , proc . of the 16th international symposium , ithaca , new york , 1993 , edited by p. drell and d. rubin ( aip , new york , 1994 ) ; cleo collaboration , m. battle _
et al . _ , in lepton and photon interactions , proc .
of the 16th international symposium , ithaca , new york , 1993 , _ op cit_. m. b. voloshin and m. a. shifman , yad .
* 45 * , 463 ( 1987 ) ; * 47 * , 801 ( 1988 ) [ sov . j. nucl . phys . * 45 * , 292 ( 1987 ) ; * 47 * , 511 ( 1988 ) ] ; n. isgur and m. b. wise , phys . lett .
b * 232 * , 113 ( 1989 ) ; * 237 * , 527 ( 1990 ) .
e. jenkins , a. v. manohar , and m. b. wise , nucl .
b396 * , 27 ( 1993 ) ; e. jenkins and a. v. manohar , phys .
b * 294 * , 273 ( 1992 ) ; z. guralnik , m. luke , and a. v. manohar , nucl .
b390 * , 474 ( 1993 ) ; m. nowak , m. rho , and i. zahed , phys .
b * 303 * , 130 ( 1993 ) ; h. k. lee and m. rho , phys .
d. * 48 * , 2329 ( 1993 ) ; k. s. gupta , m. a. momen , j. schechter , and a. subbaraman , _ ibid .
_ * 47 * , 4835 ( 1993 ) ; a. momen , j. schechter , and a. subbaraman , _ ibid . _ * 49 * , 5970 ( 1994 ) ; d .-
min , y. oh , b .- y . park , and m. rho , int .
e * 4 * , 47 ( 1995 ) .
r. casalbuoni , a. deandrea , n. di bartolomeo , r. gotto , f. feruglio , and g. nardulli , phys .
b * 292 * , 371 ( 1992 ) ; * 299 * , 139 ( 1993 ) ; p. ko , phys .
d * 47 * , 1964 ( 1993 ) ; j. schechter and a. subbaraman , _ ibid . _ * 48 * , 332 ( 1993 ) .
p. colangelo , g. nardulli , a. deandrea , n. di bartolomeo , r. gatto , and f. feruglio , phys .
b * 339 * , 151 ( 1994 ) ; p. jain , a. momen , and j. schechter , syracuse report no .
su-4240 - 581 , 1994 , hep - ph/9406338 . | we obtain the spectra of excited heavy baryons containing one heavy quark by quantizing the exactly - solved heavy meson bound states to skyrme soliton .
the results are comparable to the recent experimental observations and quark model predictions , and are consistent with the heavy quark spin symmetry .
however , somewhat large dependence of the results on the heavy quark mass strongly calls for the incorporation of the soliton - recoil effects . |
frequency spectrum of photonic crystals consists of large number of continuous bands .
this is a direct consequence of spatial periodicity of the permittivity @xmath0 which defines the structure . in the limit of infinitesimally small variations of the permittivity
, all frequency bands could be constructed from the dispersion relation of electromagnetic wave in homogeneous medium by reduction of the momenta @xmath1 to the first brillouin zone @xcite . with increasing permittivity contrast , frequency gaps open at the edges of brillouin zone @xcite .
we will call the resulting frequency bands periodic ( @xmath2 ) bands . for higher permittivity contrast , the frequency spectrum is more complicated .
it contains , besides the @xmath2 bands , also other , usually very narrow ( almost dispersionless ) frequency bands . as an example
, we show in fig .
[ uvod](a ) the frequency spectrum of two dimensional square array of dielectric cylinders embedded in vacuum .
only five of 17 displayed frequency bands are @xmath3 bands .
( @xmath4 is the spatial periodicity in @xmath5 and @xmath6 directions ) and permittivity @xmath7 calculated by the plane wave expansion method @xcite . in the frequency region
@xmath8 we found five @xmath9 bands ( black full lines ) , five even ( dashed red lines ) and seven odd ( dot blue lines ) @xmath10 bands .
inset shows schematically the structure in the @xmath11 plane , red arrow indicates the propagation direction of the electromagnetic wave .
( b ) transmission coefficient of the plane wave incident on the linear array of cylinders ( the geometry of the problem is shown in ( c ) ) .
the incident angle is @xmath12 ( red line ) and @xmath13 ( blue line ) .
note the one - to - one correspondence between resonant frequencies of excited fano resonances and @xmath14 bands observed in the frequency spectrum of the infinite 2d structure .
( c ) transmission of incident electromagnetic wave through linear array of dielectric cylinders .
( d ) geometry of the photonic structure composed from @xmath15 rows of dielectric cylinders .
( @xmath16 throughout this paper . )
both structures displayed in ( c ) and ( d ) are infinite in the @xmath5 direction . , title="fig : " ] ( @xmath4 is the spatial periodicity in @xmath5 and @xmath6 directions ) and permittivity @xmath7 calculated by the plane wave expansion method @xcite . in the frequency region
@xmath8 we found five @xmath9 bands ( black full lines ) , five even ( dashed red lines ) and seven odd ( dot blue lines ) @xmath10 bands .
inset shows schematically the structure in the @xmath11 plane , red arrow indicates the propagation direction of the electromagnetic wave .
( b ) transmission coefficient of the plane wave incident on the linear array of cylinders ( the geometry of the problem is shown in ( c ) ) .
the incident angle is @xmath12 ( red line ) and @xmath13 ( blue line ) .
note the one - to - one correspondence between resonant frequencies of excited fano resonances and @xmath14 bands observed in the frequency spectrum of the infinite 2d structure .
( c ) transmission of incident electromagnetic wave through linear array of dielectric cylinders .
( d ) geometry of the photonic structure composed from @xmath15 rows of dielectric cylinders .
( @xmath16 throughout this paper . )
both structures displayed in ( c ) and ( d ) are infinite in the @xmath5 direction .
, title="fig : " ] the aim of this paper is to explain physical origin these additional frequency bands .
we prove that they originate from fano resonances @xcite which can be excited in linear chain of dielectric cylinders by incident electromagnetic wave .
we find numerically the spectrum of these resonances and demonstrate how each resonance develops into narrow fano ( @xmath10 ) band in corresponding two - dimensional photonic crystal .
the band structure shown in fig . [ uvod](a ) was calculated for the square array of thick dielectric cylinders . for other structures , for instance photonic crystals
composed from thinner cylinders , mutual coupling of @xmath10 and original @xmath3 band occurs which results in irregularities of the band spectrum .
two of them , namely the split of the @xmath3 band , and an overlap of two bands , will be discussed later .
fano resonances in photonic structures @xcite were studied mostly in process of interaction of two dimensional photonic slabs with incident electromagnetic wave @xcite and were used for experimental identification of spectra of leaky modes in photonic structures @xcite . for the case of individual dielectric cylinder , fano resonances were observed as a result of the coupling of mie resonant states @xcite with an incident electromagnetic wave in spherical @xcite and cylindrical dielectric object @xcite .
fano resonances play an important role in design of metamaterials @xcite and influence considerably the transport properties of disordered systems @xcite . here
, we are interested in fano resonances which could be excited by plane electromagnetic wave incident to the linear periodic array of dielectric cylinders ( fig . [ uvod](c ) ) as a result of interference of leaky guided modes of periodic array of cylinders @xcite with incident electromagnetic wave .
these resonances can be characterized by the resonant frequency and lifetime which is inversely proportional to the width of the resonant peak .
surprisingly , they are narrower than resonances excited in individual dielectric cylinders . if the electromagnetic wave propagates across @xmath15 chains of cylinders ( fig . [ uvod](d ) ) then resonances excited in individual chains couple together and create narrow transmission band obtained in the frequency spectra of photonic crystals .
we concentrate on periodic photonic structures composed from dielectric cylinders parallel to the @xmath17 axis .
cylinders possess a frequency independent permittivity @xmath7 and permeability @xmath18 .
the embedding medium is vacuum .
radius of cylinders is @xmath19 and the distance between two nearest - neighbor cylinders is @xmath4 .
the first structure is the linear chain of dielectric cylinders lying in the @xmath20 axis ( fig . [ uvod](c ) ) .
incident electromagnetic wave of wavelength @xmath21 excites in this structure fano resonances which manifests themselves as sharp irregularities in the frequency dependence of the transmission coefficient visible in fig .
[ uvod](b ) .
transmission coefficient is calculated also for finite photonic slab constructed from @xmath16 rows of the same cylinders located in planes @xmath22 ( @xmath23 , see fig . [ uvod](d ) ) . observed transmission spectra
will be compared with the band structure of an infinite square array of cylinders calculated by plane wave expansion method @xcite .
transmission coefficient is calculated by the transfer matrix method @xcite . for more detailed analysis of the spatial distribution of electric field ,
we use another algorithm based on the expansion of electromagnetic field into cylinder functions @xcite .
if an incident electromagnetic wave is polarized with electric field @xmath24 parallel to cylinders ( this polarization is considered throughout this paper ) then the intensity of electric field scattered at a cylinder centered in @xmath25 can be expressed as @xmath26 for @xmath27 and @xmath28 , respectively .
similar expressions for radial and tangential components of magentic field could be derived from maxwell equations @xcite . in eq .
( [ eq : inc ] ) , @xmath29 , @xmath30 , @xcite @xmath31 , @xmath32 are bessel and hankel function , @xmath33 determines the wavelength of electromagnetic field in vacuum and @xmath34 is the index of refraction . for another cylinder , centered at @xmath35 and @xmath36 the field @xmath24
is again expressed by eq .
[ eq : inc ] but with new set of coefficients @xmath37 , @xmath38 and cylindrical coordinates @xmath39 and @xmath40 associated with the center of the cylinder . in numerical simulations ,
we calculate coefficients @xmath41 and @xmath42 from the requirement of continuity of tangential components of the intensity of electric and magnetic field at the boundary of cylinders .
note that spatial periodicity of the structure along the @xmath5 direction considerably reduces the number of unknown coefficients since coefficients @xmath37 and @xmath38 fulfill the bloch theorem @xmath43 where @xmath44 is the transverse component of the wave vector .
this enables us to reduce the number of unknown coefficients @xmath45 to @xmath46 , where @xmath47 is the highest order of bessel function used in numerical calculations ( @xmath48 in most cases ) . to find the transmission coefficient @xmath49 , poynting vector
is calculated on the opposite side of the structure .
details of the method are given elsewhere @xcite .
shown in fig .
[ uvod](c ) .
only even resonances are excited since the electromagnetic wave propagates perpendicularly to the cylinder chain @xcite .
( b ) the frequency dependence of parameters @xmath50 .
( c ) transmission coefficient for @xmath16 rows of cylinders .
fano resonances observed in linear chain develop themselves to narrow frequency bands .
( d ) detail of the the frequency dependence of the transmission coefficient shown in ( a ) and ( c ) in the vicinity of the fano resonance of coefficient @xmath51 ( @xmath52 ) . ] [ cols="<,^,^,^,^,^ " , ]
) and ( b ) transmission coefficient @xmath49 and the coefficient @xmath53 as a function of the frequency @xmath54 . owing to fano resonance
the transmission decreases to zero at @xmath55 .
( c ) the band structure of two - dimensional square array .
fano resonance causes the broad gap between the second and the third bands .
( d ) the transmission coefficient of the electromagnetic wave propagating through an array of 24 rows of cylinders .
( e ) the intensity of electric field @xmath24 inside six cylinders along the @xmath6 direction at band edges x@xmath56 and @xmath57 .
, title="fig : " ] ) and ( b ) transmission coefficient @xmath49 and the coefficient @xmath53 as a function of the frequency @xmath54 . owing to fano resonance
the transmission decreases to zero at @xmath55 .
( c ) the band structure of two - dimensional square array .
fano resonance causes the broad gap between the second and the third bands .
( d ) the transmission coefficient of the electromagnetic wave propagating through an array of 24 rows of cylinders .
( e ) the intensity of electric field @xmath24 inside six cylinders along the @xmath6 direction at band edges x@xmath56 and @xmath57 .
, title="fig : " ] ( e ) ) and ( b ) transmission coefficient @xmath49 and the coefficient @xmath53 as a function of the frequency @xmath54 . owing to fano resonance
the transmission decreases to zero at @xmath55 .
( c ) the band structure of two - dimensional square array .
fano resonance causes the broad gap between the second and the third bands .
( d ) the transmission coefficient of the electromagnetic wave propagating through an array of 24 rows of cylinders .
( e ) the intensity of electric field @xmath24 inside six cylinders along the @xmath6 direction at band edges x@xmath56 and @xmath57 .
, title="fig : " ] ) and ( b ) transmission coefficient @xmath49 and the coefficient @xmath53 as a function of the frequency @xmath54 . owing to fano resonance
the transmission decreases to zero at @xmath55 .
( c ) the band structure of two - dimensional square array .
fano resonance causes the broad gap between the second and the third bands .
( d ) the transmission coefficient of the electromagnetic wave propagating through an array of 24 rows of cylinders .
( e ) the intensity of electric field @xmath24 inside six cylinders along the @xmath6 direction at band edges x@xmath56 and @xmath57 .
, title="fig : " ] ) and ( b ) transmission coefficient @xmath49 and the coefficient @xmath53 as a function of the frequency @xmath54 . owing to fano resonance
the transmission decreases to zero at @xmath55 .
( c ) the band structure of two - dimensional square array .
fano resonance causes the broad gap between the second and the third bands .
( d ) the transmission coefficient of the electromagnetic wave propagating through an array of 24 rows of cylinders .
( e ) the intensity of electric field @xmath24 inside six cylinders along the @xmath6 direction at band edges x@xmath56 and @xmath57 .
, title="fig : " ] ) and ( b ) transmission coefficient @xmath49 and the coefficient @xmath53 as a function of the frequency @xmath54 . owing to fano resonance
the transmission decreases to zero at @xmath55 .
( c ) the band structure of two - dimensional square array .
fano resonance causes the broad gap between the second and the third bands .
( d ) the transmission coefficient of the electromagnetic wave propagating through an array of 24 rows of cylinders .
( e ) the intensity of electric field @xmath24 inside six cylinders along the @xmath6 direction at band edges x@xmath56 and @xmath57 .
, title="fig : " ] @xmath58 .
solid lines represent bands with even spatial symmetry , dot lines correspond , similarly to fig .
[ uvod](a ) , to odd fano bands .
note an overlap of the 3rd and 4th bands , shown in detail in the right panel .
this overlap results from the coupling of two bands displayed by green dot - dashed and red dashed line .
( b ) frequency dependence of the transmission coefficient for the linear chain of cylinders .
three fano resonances could be easily identified .
( c ) top panel shows the transmission coefficient for 24 rows of cylinders as a function of frequency @xmath54 .
two lower panels present detail of the frequency dependence of the transmission coefficient in the vicinity of the fano resonance .
@xcite , title="fig : " ] .
solid lines represent bands with even spatial symmetry , dot lines correspond , similarly to fig .
[ uvod](a ) , to odd fano bands .
note an overlap of the 3rd and 4th bands , shown in detail in the right panel .
this overlap results from the coupling of two bands displayed by green dot - dashed and red dashed line .
( b ) frequency dependence of the transmission coefficient for the linear chain of cylinders .
three fano resonances could be easily identified .
( c ) top panel shows the transmission coefficient for 24 rows of cylinders as a function of frequency @xmath54 .
two lower panels present detail of the frequency dependence of the transmission coefficient in the vicinity of the fano resonance .
@xcite , title="fig : " ] consider first an array of thick dielectric cylinders with radius @xmath59 .
figures [ uvod](b ) and [ r40](a ) shows the transmission coefficient of plane electromagnetic wave propagating through linear chain of cylinders .
a series of very narrow resonances could be identified .
similar maxima and minims has been found numerically in the reflection coefficient @xcite .
following @xcite we interpret these resonances as fano resonances which results from the interference of incident plane wave with leaky guided modes excited in the periodic cylinder row @xcite .
indeed , these fano resonances are accompanied by sharp maxima of coefficients @xmath42 defined in eq .
[ eq : inc ] ( figure [ r40](b ) ) . more detailed analysis of resonances is given in sect .
[ sect : fano ] .
comparison of figures [ uvod](a ) and ( b ) confirms that fano resonances develop to narrow fano bands in the spectra of 2d structures .
this is also shown in figs .
[ r40](c , d ) which present the frequency dependence of the transmission coefficient of plane wave propagating through the slab composed from @xmath16 rows of cylinders .
different character of @xmath3 and @xmath10 bands is clearly visible from the spatial symmetry of the electric field @xmath24 shown in fig .
[ r40-fano ] .
the top panel displays the field @xmath24 for frequencies chosen in the center of five lowest @xmath2 band shown in fig .
[ r40](c ) .
as expected , the field symmetry changes when the frequency increases from one frequency band to the next one @xcite .
the bottom panel shows the field @xmath24 for the five lowest resonant frequencies identified in fig .
[ r40](b ) . as shown in fig .
[ r40](d ) , these frequencies correspond to the center of @xmath10 bands .
the symmetry of the field is unambiguously determined by the order of excited fano resonance . for thinner cylinders , the @xmath3 band and the @xmath10 band can overlap .
then , the resulting frequency spectrum depends on the @xmath44-dependence of the frequency in two bands and on the strength of their mutual coupling .
consider a simple model of two bands @xmath60 and @xmath61 , ( centered at the same frequency for simplicity ) coupled together with coupling constant @xmath62 .
resulting spectrum has a form @xmath63 two bands @xmath64 and @xmath65 are separated by gap if @xmath66 .
this happens either when @xmath67 is large or when one of two bands is narrow .
_ band splitting_. we found the above mentioned band splitting in the frequency spectrum of the square array of thin ( @xmath68 ) dielectric cylinders ( fig .
[ r10 ] ) .
the transmission coefficient through a linear chain of cylinders ( fig . [ r10](b ) ) decreases to zero for the frequency @xmath55 .
this decrease is accompanied by an increase of coefficient @xmath53 ( eq . [ eq : inc ] ) .
we interpret this decrease of the transmission as a result of excitation of broad fano resonance .
figure [ r10](c ) shows the band structure of the square array of thin cylinders .
note that both the second and the third bands have a minimum at the x point .
these two bands result from the coupling of the @xmath9 band with broad @xmath14 band displayed by dot dashed ( x@xmath69 ) and dashed ( @xmath70x@xmath71 ) lines , respectively .
this statement is supported also by the analysis of the spatial symmetry of electric field within two bands shown in fig .
[ r10](e ) .
note that the symmetry of electric field changes along the line x@xmath72 .
also , the field in points x@xmath73 and @xmath74 have the same symmetry ; of course , the same holds for pair x@xmath71 and @xmath70 .
while the field at x@xmath73 and @xmath74 possesses the symmetry of the @xmath9 band , field close to inner band edges x@xmath71 and @xmath70 has a symmetry of excited fano resonance . _ band overlap_. the overlap of @xmath3 and @xmath10 bands
is observed in the band structure of the infinite square array of cylinders with radius @xmath75 displayed in fig .
[ eps12-t](a ) as shown in the right panel , the overlap of the 3rd and 4th bands can be interpreted as a result of coupling of two bands shown by dot dashed and dashed lines , respectively .
this overlap can be identified also from the complicated frequency dependence of the transmission coefficient @xcite shown in fig .
[ eps12-t](b ) for the slab of @xmath16 rows of cylinders .
similar analysis could be done for the overlap of the 5th and 6th bands in fig .
in previous section , we have shown that fano resonances excited in linear array of dielectric cylinders create the @xmath10 bands in spectra of photonic crystals .
now we will discuss physical origin of fano resonances .
fano resonances have been observed recently in the most simple dielectric structure the single dielectric cylinder @xcite . if the frequency of incident electromagnetic wave coincides with the eigenfrequency of any cylinder leaky eigenmode , the last can be excited .
then , the electromagnetic field in the neighbor of the cylinder is given by a superposition of two fields with the same frequency : the incident plane wave and field radiated by excited resonance @xcite .
the excitation of resonance manifests itself as a maximum of coefficient @xmath42 shown in fig .
the width of the resonance is proportional to inverse of its lifetime .
similarly , excitation of resonant guided mode in linear array of cylinders can be identified from sharp maxima in frequency dependence of corresponding coefficient @xmath42 ( figs .
2(b ) , 6(b ) ) .
the interference of two modes is the responsible for narrow maxima and minima in the transmission coefficient of incident electromagnetic wave displayed for instance in figs .
1(b ) and 2(a ) . for ( a ) single dielectric cylinder with radius @xmath59 .
( b ) linear row of cylinders .
( c ) coefficients @xmath76 ( red ) and an inverse determinant @xmath77 ( black ) for the clusters of @xmath78 dielectric cylinders ( @xmath79 ) along the @xmath5 axis .
thick blue line shows the position of the @xmath10 band @xmath80 displayed in fig .
1(a ) by red dashed line .
multiple points for a given frequency corresponds with different intensity of excited resonance at individual cylinders as shown in ( d ) for the cluster with @xmath81 cylinders .
, title="fig : " ] + for ( a ) single dielectric cylinder with radius @xmath59 .
( b ) linear row of cylinders .
( c ) coefficients @xmath76 ( red ) and an inverse determinant @xmath77 ( black ) for the clusters of @xmath78 dielectric cylinders ( @xmath79 ) along the @xmath5 axis .
thick blue line shows the position of the @xmath10 band @xmath80 displayed in fig .
1(a ) by red dashed line .
multiple points for a given frequency corresponds with different intensity of excited resonance at individual cylinders as shown in ( d ) for the cluster with @xmath81 cylinders .
, title="fig : " ] for ( a ) single dielectric cylinder with radius @xmath59 .
( b ) linear row of cylinders .
( c ) coefficients @xmath76 ( red ) and an inverse determinant @xmath77 ( black ) for the clusters of @xmath78 dielectric cylinders ( @xmath79 ) along the @xmath5 axis .
thick blue line shows the position of the @xmath10 band @xmath80 displayed in fig .
1(a ) by red dashed line .
multiple points for a given frequency corresponds with different intensity of excited resonance at individual cylinders as shown in ( d ) for the cluster with @xmath81 cylinders . ]
we start with the comparison of the frequency dependence of coefficients @xmath42 for isolated dielectric cylinder ( fig .
[ fig : mie](a ) ) and for an infinite periodic array of cylinders ( figs .
[ r40](b ) and [ fig : mie](b ) ) . for single cylinder and @xmath24 polarization
, @xmath82 is a solution of system of linear equations @xcite @xmath83 where @xmath84 , @xmath85 is an impedance , and the r.h.s is given by the expansion of incident plane wave into bessel functions @xcite @xmath86 as shown in fig .
[ fig : mie](a ) , fano resonances of cylinder lye very close to those of an array of cylinders ( figs .
2(b ) and fig .
[ fig : mie](b ) ) .
the first two resonances ( @xmath87 and @xmath88 are relatively broad , especially for an infinite number of cylinders . on the other hand , fano resonances with @xmath89 are significantly narrower when excited in an infinite linear chain of cylinders than in individual cylinder . to explore how the shape of the resonance depends on the number of cylinders , we analyze the scattering of incident electromagnetic wave on finite cluster consisting from @xmath78 cylinders along the @xmath5 direction ( fig .
coefficients @xmath90 were calculated as a solution of system of linear equations @xmath91 where @xmath92 is a matrix of the size @xmath93 and vector @xmath94 represents incident electromagnetic wave .
figure [ fig : mie](c ) presents coefficients @xmath76 for finite cluster of @xmath78 cylinders excited by perpendicularly incident plane wave .
the resonance indeed becomes narrower when number of cylinders increases .
since the spatial distribution of electromagnetic field along the finite cluster is not homogeneous , @xmath76 acquires various values for individual cylinders inside the cluster . as an example , we plot in fig .
[ fig : mie](d ) spatial distribution of @xmath95 as well as its real and imaginary part calculated in cluster of 48 cylinders .
figure [ fig : mie](c ) shows also the frequency dependence of an inverse of the determinant of the matrix @xmath92 ( eq . [ eq : lin ] ) which determines the eigenfrequencies and lifetimes of leaky guided modes @xcite exited in the cluster .
comparison of the frequency dependence of an inverse determinant and @xmath96 confirms fano character of observed excitation and enables to estimate the sign of the fano parameter @xmath97 @xcite .
but with small imaginary part of the cylinder permittivity .
absorption is large at band edges , where the group velocity is small , and in the region of the fano bands .
( b ) detailed frequency dependence of the absorption in the resonant frequency region for imag @xmath98 and 0.01 .
, title="fig : " ] + + but with small imaginary part of the cylinder permittivity .
absorption is large at band edges , where the group velocity is small , and in the region of the fano bands .
( b ) detailed frequency dependence of the absorption in the resonant frequency region for imag @xmath98 and 0.01 .
, title="fig : " ] finally , we note an another difference between @xmath3 and @xmath10 bands : we expect that the @xmath10 bands are much more sensitive to the absorption loses than the @xmath3 bands .
one reason is that typical @xmath10 band is narrow , therefore the group velocity of transmitted wave is small .
however , more important is that @xmath10 bands are associated with the resonance which led to higher intensity of propagating electric field .
figure [ abs ] presents the absorption of electromagnetic field in the array of @xmath75 cylinders with small imaginary part of the permittivity .
fano bands could be identified from the position of large maxima of the absorption .
note that very small imaginary part of the permittivity ( imag @xmath0/real @xmath99 ) causes the absorption of 20% of energy when wave propagates through an array of 24 rows of cylinders .
in conclusion , we showed that the band structure of square arrays of cylinders can be completely described in terms of two kinds of frequency bands .
the @xmath3 bands originates from the reduction of the dispersion relation to the first reduced zone .
the @xmath10 bands have origin in fano resonances observed in the linear chain of cylinders . the field distribution in @xmath10 bands is determined by the symmetry of fano resonance .
two bands , @xmath3 and @xmath10 might overlap , which complicates resulting band structure . since resonance is typically accompanied by strong electric field , we expect that absorption is stronger in @xmath10 bands than in the @xmath3 bands . to assure the numerical stability , hankel functions are normalized by the frequency - dependent function @xmath100 .
consequently , this function scales coefficients @xmath82 shown in figures throughout the paper .
note that the overlap frequency interval , found by the the transfer matrix method in fig .
[ eps12-t](c ) differs slightly from that calculated by the plane wave expansion ( fig .
[ eps12-t](a ) .
the origin of this discrepancy lies in the numerical limitation of the transfer matrix method due to the finite discretization of space . | we show that the frequency spectrum of two dimensional photonic crystals is strongly influenced by fano resonances which can be excited already in the linear array of dielectric cylinders .
to support this claim , we calculate the transmission of electromagnetic wave through linear array of dielectric cylinders and show that frequencies of observed fano resonances coincides with position of narrow frequency bands found in the spectra of corresponding two - dimensional photonic crystals .
split of frequency band or overlap of two bands , observed in the band structure of photonic structures are also associated with fano resonances . |
in the last few decades dusty plasma attracts the researchers most , not only because of different types of wave propagation in dusty plasma due to their involvement in the study of astrophysical and space environments but also for its great variety of new phenomena associated with waves and instabilities.@xcite dusty plasmas essentially consists of fairly massive , charged particulates in addition to electrons and ions in the usual two component plasma .
acoustic wave modes in dusty plasma have received a great deal of attention since the last decade .
depending on different time scales , there can exists two or more acoustic waves in a typical dusty plasma .
dust acoustic ( da ) and dust ion - acoustic ( dia ) waves are two such acoustic waves in a plasma containing electrons , ions , and charged dust grains .
shukla and silin @xcite were the first to show that due to the quasi neutrality condition @xmath11 and the strong inequality @xmath12 ( @xmath13 , @xmath14 , and @xmath15 are , respectively , the number density of electrons , ions , and dust particle , where @xmath16 is the number of electrons residing on the dust grain surface ) , a dusty plasma ( with negatively charged static dust grains ) supports low - frequency dia waves with phase velocity much smaller ( larger ) than electron ( ion ) thermal velocity . in case of long wavelength limit
the dispersion relation of dia wave is similar to that of ia wave for a plasma with @xmath17 and @xmath18 , where @xmath19 is the average ion ( electron ) temperature . due to the usual dusty plasma approximations ( @xmath12 and @xmath20 ) , a dusty plasma can not support the usual ion - acoustic ( ia ) waves , but the dia waves of shukla and silin @xcite can .
thus dia waves are basically ia waves , modified by the presence of heavy dust particulates .
the theoretical prediction of shukla and silin @xcite was supported by a number of laboratory experiments.@xcite the linear properties of dia waves in dusty plasma are now well understood .
@xcite due to a balance between the wave dispersion and nonlinearity , there exists the possibility of dia solitary waves ( diasws ) in a dusty plasma .
the diasws studied using the reductive perturbation technique which gives korteweg - de vries ( kdv ) equation or korteweg - de vries burgers ( kdvb ) equation as nonlinear evolution equations.@xcite bharuthram and shukla @xcite studied the nonlinear waves associated with dia waves .
mamun and shukla @xcite have investigated the condition for existence of positive and negative potential diasws , and examined how the ion - fluid temperature modifies the amplitude of diasws . in presence of negatively charged dust grains , one can have diasws with negative potentials , when the electron number density is less than one third of the positive ion number density @xcite .
et al _ @xcite have shown that in the dust - modified ion acoustic regime , negative structures can also be generated , beside positive potential soliton for electrons with a polytropic index @xmath21 , as well as for boltzmann electrons .
the effect of ion - fluid temperature on diasws structures have been investigated by sayed and mamun @xcite in a dusty plasma containing adiabatic ion - fluid , boltzmann electrons , and static dust particles . in the above mentioned works the nonlinear properties of diasws investigated for @xmath22 . in most of the earlier works ,
maxwellian velocity distribution function for lighter species of particles has been used to study diasws and dia double layers ( diadls ) . however the dusty plasma with nonthermally / suprathermally distributed electrons observed in a number of heliospheric environments @xcite .
therefore , it is of considerable importance to study nonlinear wave structures in a dusty plasma in which lighter species ( electrons ) are nonthermally / suprathermally distributed .
recently baluku _ et al _ @xcite investigated diasws in an unmagnetized dusty plasma consisting of cold dust particles and kappa distributed electrons using both small and arbitrary amplitude techniques .
in the present investigation we have considered the problem of existence as well as the polarity of diasws and diadls in a nonthermal dusty plasma consisting of negatively charged dust grains , adiabatic positive ions and nonthermal electrons at the smallest possible value of the mach number @xmath2 .
the sagdeev potential approach@xcite has been considered to investigate the existence of solitary wave and double layer at @xmath23 , where @xmath24 is the lower bound of the mach number @xmath2 , i.e. , solitary wave and/or double layer solutions of the energy integral start to exist for @xmath22 . in
most of the earlier works , solitary wave and/or double layer solutions have been investigated for @xmath22 .
however , some recent investigations @xcite have shown that finite amplitude solitary wave can exist at @xmath23 in the parameter regime where solitons of both polarities exist .
the numerical observations @xcite of the solitary wave solution of the energy integral at @xmath23 , motivate us to investigate the existence of diasw and diadl .
however , we have approach in a different way without considering the parameter regime of the system for @xmath22
. three basic parameters of the present dusty plasma system are @xmath8 , @xmath25 and @xmath26 , which are respectively the ratio of unperturbed number density of nonthermal electrons to that of ions , the ratio of average temperature of ions to that of nonthermal electrons , a parameter associated with the nonthermal distribution of electrons .
nonthermal distribution of electrons becomes isothermal one if @xmath27 and consequently , for isothermal electron species , the present dusty plasma contains only two basic parameters @xmath8 and @xmath25 .
it has been found that solitary wave or double layer do not exist for @xmath28 . for isothermal distributed electrons ,
ppsws exist for @xmath29 , whereas for all @xmath8 lies within @xmath30 , there exist npsws . at @xmath31 , no solitary wave or
double layer solution is possible .
however , there do not exists any double layer solution for isothermal electrons . for nonthermal distributed electrons ,
we have found two critical values of @xmath3 , viz .
, @xmath32 and @xmath33 are the upper bounds of @xmath3 for the existence of ppsws and npsws , respectively . however , the coexistence of both npsws and ppsws are not possible in the present system . again , at @xmath34 , no solitary wave or double layer solution is possible .
we have also found the existence of negative potential double layer ( npdl ) in a portion of the solution space .
depending on the nature of solutions of dia waves , we have @xmath35 , @xmath36 , and @xmath37 such that the entire parametric space can be delimited into three subintervals of @xmath8 , viz . , @xmath38 , @xmath39 , and @xmath40 .
for each @xmath8 in @xmath40 , there exists a value @xmath10 of @xmath3 , such that one can found the existence of npdl .
there is an interesting phenomena occur around the point @xmath41 .
there exists a finite jump in amplitude of npsws at @xmath41 , the jump being the amplitude of the double layer exists at @xmath41 .
the amplitude of npsw decreases with increasing @xmath3 and ultimately , demolished at @xmath34 . on the other hand
, ppsw creates from @xmath34 and the amplitude of ppsw increases with @xmath3 , having maximum amplitude at @xmath42 . throughout the investigation we have considered @xmath43 as the upper bound of @xmath8 , which is strictly less than 1 , since for @xmath44 , the effect of charged dust on
dia wave is negligible .
the present paper is organized as follows : in sec .
[ sec : non eqn ] , the basic equations are given .
the energy integral and the sagdeev potential has been constructed in sec .
[ sec : energy int ]
. physical interpretation of the energy integral for the existence of solitary wave and double layer solutions is also given in this section . in sec .
[ sec : theory ] , an analytical approach has been presented to determine the polarity of the solitary waves .
the analytical approach has been verified numerically in sec .
[ sec : numerical ] .
finally , a brief conclusion has been given in sec .
[ sec : conclusion ] .
the governing equations describing the nonlinear behavior of dia waves , propagating along @xmath45-axis , in collisionless , unmagnetized dusty plasma consisting of negatively charged immobile dust grains , adiabatic positive ions and nonthermal electrons are the following : @xmath46 @xmath47 @xmath48 @xmath49 where the parameter @xmath50 . here
we have used the notation @xmath51 or @xmath52 for @xmath53 and @xmath54 , @xmath55 , @xmath56 , @xmath57 , @xmath58 , @xmath45 and @xmath59 are , respectively , the ion number density , electron number density , ion velocity , ion pressure , electrostatic potential , spatial variable and time , and they have been normalized by @xmath14 ( unperturbed ion number density ) , @xmath13 ( unperturbed electron number density ) , @xmath60(@xmath61 ) ( ion - acoustic speed ) , @xmath62 , @xmath63 , @xmath64(debye length ) , and @xmath65 ( ion plasma period ) . here
@xmath66 is the adiabatic index , @xmath67 is the boltzmann constant , @xmath68 and @xmath69 are , respectively , the average temperatures of ions and electrons , @xmath70 is the mass of an ion , @xmath15 is the dust number density , @xmath16 is the number of negative unit charges residing on dust grain surface , and @xmath71 is the charge of an electron .
the above equations are supplemented by nonthermally distributed electrons as prescribed by cairns _
@xcite for the electron species .
actually , in a number of heliospheric environments , dusty plasma contains nonthermally distributed ions or electrons .
therefore , it is of considerable importance to study diasws and diadls in dusty plasma in which electrons are nonthermally distributed .
nonthermal distribution of any lighter species of particles ( as prescribed by cairns _
@xcite for the electron species ) can be regarded as population of boltzmann distributed particles together with a population of energetic particles .
this can also be regarded as a modified boltzmannian distribution , which has the property that the number of particles in phase space in the neighbourhood of the point @xmath72 is much smaller than the number of particles in phase space in the neighbourhood of the point @xmath72 for the case of boltzmann distribution , where @xmath73 is the velocity of the particle in phase space .
this type of velocity distribution is often termed as cairns distribution and was considered by many authors in various studies of different collective processes in plasmas and dusty plasmas @xcite . under the above mentioned normalization of the dependent and independent variables , the normalized number density of nonthermal cairns
distributed electrons can be written as @xmath74 where @xmath75 with @xmath76 . here
@xmath3 and consequently @xmath77 are the parameters associated with nonthermal distribution of electrons . from ( [ beta1 ] ) and inequality @xmath76
, it can be easily checked that the nonthermal parameter @xmath3 is restricted by the following inequality : @xmath78 .
however we can not take the whole region of @xmath3 ( @xmath79 ) .
plotting the nonthermal velocity distribution of electrons against phase velocity ( @xmath73 ) of electrons in phase space , it can be easily shown that the number of electrons in phase space in the neighbourhood of the point @xmath72 decreases with increasing @xmath3 and the number of electrons in phase space in the neighbourhood of the point @xmath72 is almost zero when @xmath80 .
therefore , for increasing values of @xmath3 distribution function develops wings , which become stronger as @xmath3 increases , and at the same time the center density in phase space drops , the latter as a result of the normalization of the area under the integral .
so , we should not take values of @xmath81 since that stage might stretch the credibility of the cairns model too far @xcite . so here
we consider the effective range of @xmath3 as follows : @xmath82 , where @xmath83 .
when @xmath27 , nonthermal distribution of electrons becomes isothermal one . now introducing a new parameter @xmath84 , the charge neutrality condition
, @xmath85 can be written as @xmath86 so that the poisson equation ( [ poisson eq without mu ] ) can be written as @xmath87 we note from ( [ charge neutrality with mu ] ) that @xmath88 must be greater than zero , i.e. , @xmath89 .
when @xmath44 , the effect of negatively charged dust grains on dia wave is negligible and so , we restrict @xmath8 by the inequality @xmath90 , where @xmath43 is strictly less than 1 .
to study the arbitrary amplitude time independent dia solitary structures we make all the dependent variables depend only on a single variable @xmath91 , where the mach number @xmath2 is normalized by @xmath60 .
thus in the steady state , ( [ continuity of ions])-([pressure eq ] ) and ( [ poisson eq ] ) can be written as @xmath92 @xmath93 @xmath94 @xmath95 using the boundary conditions , @xmath96 and solving ( [ continuity of ions steady state ] ) - ( [ pressure eq steady state ] ) , we get a quadratic equation for @xmath97 and the solution of this quadratic equation can be put in the following form : @xmath98 where @xmath99 from the expression of @xmath97 as given by ( [ n_i with phi_m psi_m ] ) we see that @xmath97 is nonnegative if and only if @xmath100 and @xmath101 both are real . again
@xmath100 and @xmath101 both are real if and only if @xmath102 and @xmath103 both are real and nonnegative . but @xmath102 and @xmath103 both are real and nonnegative if and only if @xmath104 and @xmath105 , respectively , and finally the conditions @xmath104 and @xmath105 hold simultaneously if and only if @xmath106 . now from ( [ psi_m phi_m ] ) we see that @xmath107 ( @xmath108 ) and consequently ( [ n_i with phi_m psi_m ] ) gives theoretically valid expression of @xmath97 if and only if @xmath109 .
therefore , for the present problem we are restricted by the inequality : @xmath110 .
but numerically it is not possible to get the correct value of @xmath54 from ( [ n_i with phi_m psi_m ] ) for small value of @xmath25 because the denominator of the right hand side of ( [ n_i with phi_m psi_m ] ) is @xmath111 . to remove @xmath25 from the denominator of the right hand side of ( [ n_i with phi_m psi_m ] ) we first of all note that @xmath112 for all @xmath109 and consequently we can multiply the numerator and the denominator of the right hand side of ( [ n_i with phi_m psi_m ] ) by @xmath113 and we get the following expression of @xmath54 .
@xmath114 from ( [ n_i with phi_m psi_m 2 ] ) , we see that this equation gives both theoretically and also numerically correct expression of @xmath54 even when @xmath115 if @xmath109 .
now if we put @xmath116 in ( [ n_i with phi_m psi_m 2 ] ) , we get the same expression of @xmath54 for the cold dust fluid with @xmath117 . now integrating ( [ poisson eq steady state ] ) with respect to @xmath58 and using the boundary conditions ( [ boundary cond ] ) , we get the following equation known as energy integral with @xmath118 as sagdeev potential or pseudo - potential .
@xmath119 where @xmath120 @xmath121 @xmath122 although @xmath123 is a function of @xmath58 , @xmath2 , @xmath3 , @xmath8 and @xmath25 , but for simplicity , we can omit any argument from @xmath124 when no particular emphasis is put upon it .
for example , when we write @xmath125 , it is meant that the arguments @xmath3 , @xmath8 and @xmath25 assume fixed values in their physically admissible range and for the fixed values of @xmath3 , @xmath8 and @xmath25 , nature of solitary structures as obtained from the energy integral ( [ energy int ] ) depends only on @xmath2 .
the energy integral ( [ energy int ] ) can be regarded as the one - dimensional motion of a particle of unit mass whose position is @xmath58 at time @xmath126 with velocity @xmath127 in a potential well @xmath123 .
the first term of the energy integral can be regarded as the kinetic energy of a particle of unit mass at position @xmath58 and time @xmath126 whereas @xmath123 is the potential energy of the same particle at that instant . since kinetic energy is always a non - negative quantity , @xmath128 for the entire motion , i.e. , zero is the maximum value for @xmath123 . again from ( [ energy int ] ) , we find @xmath129 , i.e. the force acting on the particle at the position @xmath58 is @xmath130 .
suppose , @xmath131 , therefore , the particle is in equilibrium at @xmath132 because the velocity as well as the force acting on the particle at @xmath132 are simultaneously equal to zero .
now if @xmath133 can be made an unstable position of equilibrium , the energy integral can be interpreted as the motion of an oscillatory particle if @xmath134 for some @xmath135 , i.e. , if the particle is slightly displaced from its unstable position of equilibrium then it moves away from its unstable position of equilibrium and it continues its motion until its velocity is equal to zero , i.e. , until @xmath58 takes the value @xmath136 .
now the force acting on the particle of unit mass at position @xmath137 is @xmath138 .
for @xmath139 , the force acting on the particle at the point @xmath140 is directed towards the point @xmath132 if @xmath141 , i.e. , if @xmath142 . on the other hand , for @xmath143
, the force acting on the particle at the point @xmath140 is directed towards the point @xmath132 if @xmath144 , i.e. , if @xmath145 . therefore ,
if @xmath145 ( for the positive potential side ) or if @xmath142 ( for the negative potential side ) then the particle reflects back again to @xmath133 . again , if @xmath146 then the velocity @xmath147 as well as the force @xmath148 both are simultaneously equal to zero at @xmath137 .
consequently , if the particle is slightly displaced from its unstable position of equilibrium ( @xmath133 ) it moves away from @xmath133 and it continues its motion until the velocity is equal to zero , i.e. , until @xmath58 takes the value @xmath137 .
however it can not be reflected back again at @xmath133 as the velocity and the force acting on the particle at @xmath137 vanish simultaneously . actually , if @xmath145 ( for @xmath143 ) or if @xmath142 ( for @xmath139 ) the particle takes an infinite long time to move away from the unstable position of equilibrium .
after that it continues its motion until @xmath58 takes the value @xmath136 and again it takes an infinite long time to come back its unstable position of equilibrium .
therefore , for the existence of a ppsw ( npsw ) solution of the energy integral ( [ energy int ] ) , we must have the following : ( a ) @xmath132 is the position of unstable equilibrium of the particle , ( b ) @xmath134 , @xmath149 @xmath150 for some @xmath151 @xmath152 , which is nothing but the condition for oscillation of the particle within the interval @xmath153 and ( c ) @xmath154 for all @xmath155 @xmath156 ) , which is the condition to define the energy integral ( [ energy int ] ) within the interval @xmath153 . for the existence of a ppdl ( npdl ) solution of the energy integral ( [ energy int ] ) , the conditions ( a ) and ( c ) remain unchanged but here ( b )
has been modified in such a way that the particle can not be reflected again at @xmath133 , i.e. , the condition ( b ) assumes the following form : @xmath157 , @xmath158 for some @xmath151 @xmath159 ) .
the above discussions for the existence of solitary waves and double layers are valid if @xmath133 is an unstable position of equilibrium , i.e. , if @xmath160 along with @xmath131 . in other words ,
@xmath132 can be made an unstable position of equilibrium if the potential energy of the said particle attains its maximum value at @xmath132 .
now , the condition @xmath161 gives a lower bound @xmath24 of @xmath2 , i.e. , @xmath162 , @xmath163 , and @xmath164 .
this @xmath24 is , in general , a function of the parameters involved in the system , or a constant .
therefore , if @xmath165 , the potential energy of the said particle attains its minimum value at @xmath132 , and consequently , @xmath132 is the position of stable equilibrium of the particle , and in this case , it is impossible to make any oscillation of the particle even when it is slightly displaced from its position of stable equilibrium , and consequently there is no question of existence of solitary waves or double layers for @xmath165 . in other words , for the position of unstable equilibrium of the particle at @xmath132 , i.e. , for @xmath166 , the function @xmath123 must be convex within a neighborhood of @xmath132 and in this case both type of solitary waves ( negative or positive potential ) may exist if other conditions are fulfilled .
now suppose that @xmath167 and also @xmath168 , then if @xmath169 , the potential energy of the said particle attains its maximum value at @xmath132 and consequently , @xmath132 is the position of unstable equilibrium . on the other hand
if @xmath167 , @xmath168 , and @xmath170 , the potential energy of the said particle attains its minimum value at @xmath132 and consequently , @xmath132 is the position of stable equilibrium of the particle and in this case there is no question of existence of solitary wave solution and/or double layer solution of the energy integral ( [ energy int ] ) .
however if @xmath171 and @xmath172 , then with the help of simple mathematics , it can be shown that @xmath173 is a concave function in @xmath174 and @xmath173 is a convex function in @xmath175 , where @xmath176 is strictly positive quantity but can be made sufficiently small . _ the out line of proof of this result is given in appendix .
_ therefore , if a particle , placed at @xmath132 , be slightly displaced towards the positive potential side it falls within the interval @xmath177 and due to the convexity of the function @xmath173 within the interval @xmath177 , it moves away from @xmath133 and it continues its motion until its velocity is equal to zero , i.e. , until @xmath58 takes the value @xmath178 , where @xmath179 and in this case if @xmath180 ( @xmath181 and @xmath182 ) , one may get a ppsw ( ppdl ) as a solution of the energy integral ( [ energy int ] ) .
on the other hand , if the same particle is slightly displaced from @xmath132 towards the negative potential side , it falls within the interval @xmath183 and due to the concavity of the function @xmath173 within the interval @xmath183 , it moves towards the point @xmath133 and consequently , there does not exist any solitary wave or double layer solution in the negative potential side . by similar arguments
, we can conclude that if @xmath184 and @xmath185 then one may obtain a npsw ( npdl ) as a solution of the energy integral ( [ energy int ] ) , whereas there does not exist any ppsw or ppdl . at this point
, it is important to note that we can not expect the coexistence of both negative and positive potential solitary waves at @xmath23 .
therefore , the necessary conditions for the existence of ppsw or ppdl at @xmath23 are @xmath171 and @xmath172 whereas the necessary conditions for the existence of npsw or npdl at @xmath23 are @xmath171 and @xmath185 .
for the present problem , it can be easily checked that @xmath186 for any value of @xmath2 and also for any values of the parameters @xmath3 , @xmath8 and @xmath25 , whereas the condition @xmath187 gives @xmath22 , where @xmath188 from ( [ mc nonthermal ] ) , we see that for @xmath24 to be real and positive , we must have @xmath189 and @xmath190 .
as the effective range of @xmath3 is @xmath82 , where @xmath83 , @xmath24 is well - defined as a real positive quantity for all @xmath191 and @xmath82 . from the discussions of the previous section we have seen that to determine the polarity of solitary wave or double layer at @xmath23 , the function @xmath192 plays an important role and for the present problem that function
is given by @xmath193 where @xmath194 now we can write ( [ v3 at mc non ] ) in the following convenient form : @xmath195 where @xmath196 @xmath197},\ ] ] @xmath198 now as @xmath199 , from ( [ betan ] ) , we get @xmath200 , consequently , from ( [ b ] ) , we see that @xmath201 for all @xmath202 and for all @xmath203 and the sign of equality holds for @xmath116 and for all @xmath202 , i.e. , @xmath204 only when @xmath205 and for any admissible value of @xmath206 .
therefore , from ( [ mu2c ] ) , we see that @xmath207 is well defined as strictly positive real number for all @xmath202 and for all @xmath208 whereas from ( [ mu1c ] ) , we see that @xmath36 is well defined as strictly positive real number for all @xmath202 and for all @xmath203 . again we note that @xmath207 occurs in the expression of @xmath192 when @xmath209 , i.e. , when @xmath208 , therefore , in any case ( @xmath209 or @xmath205 ) sign of @xmath192 depends only on the factor @xmath210 , more specifically , @xmath211 or @xmath212 according to @xmath213 or @xmath214 . as @xmath8
is restricted by the inequality @xmath215 for the case of negatively charged dust particles , then the condition @xmath216 is always true if @xmath217 .
but if @xmath218 then @xmath211 or @xmath212 according to @xmath213 or @xmath214 and when @xmath219 , @xmath220 .
in this case , i.e. , when @xmath220 , for solitary wave solution at @xmath23 , it is necessary that @xmath221 .
so , now our problem , is to find the sign of @xmath222 when @xmath220 , i.e. , when @xmath219 .
when @xmath209 , with the help of simple algebra , we get the following expression of @xmath222 at @xmath219 @xmath223}{9 \beta_{n}^{3}(b+1)^{3}},\ ] ] whereas for @xmath115 , the expression of @xmath224 at @xmath219 is given by @xmath225 as @xmath226 for all @xmath227 , from ( [ v4mc alpha eq zero ] ) , we see that @xmath228 for @xmath205 .
again as @xmath229 for @xmath209 , i.e. , for all @xmath208 and for all @xmath202 , from ( [ v4mc alpha eq zero ] ) , we see that @xmath228 for @xmath209 .
therefore , in any case , @xmath205 or @xmath209 , for the present problem , we see that for any admissible values of parameters of the system @xmath228 when @xmath219 , i.e. , when @xmath230 .
therefore , there does not exist any solitary wave solution of the energy integral ( [ energy int ] ) at @xmath23 when @xmath220 .
again , differentiating ( [ modified v3 at mc non ] ) with respect to @xmath3 , we get @xmath231 now as @xmath226 for all @xmath199 and @xmath232 , from ( [ diff wrt beta1 v3 at mc non ] ) , we see that @xmath233 for any given value of @xmath25 , and consequently , @xmath192 is strictly decreasing function of @xmath3 for all @xmath199 .
again , it can be easily checked that @xmath212 for @xmath27 if @xmath234 whereas @xmath211 for @xmath27 if @xmath235 .
again , it can be easily checked that @xmath236 , where @xmath237 and @xmath3 approaches to 1 , i.e. , @xmath3 approaches to 1 from the left side of 1 .
therefore , if @xmath234 , @xmath192 strictly decreases with increasing @xmath3 for all @xmath199 starting from a positive value and ending with a negative value whereas if @xmath235 , @xmath192 strictly decreases with increasing @xmath3 for all @xmath199 starting from a negative value and ending with a negative value .
so , @xmath192 intersects the axis of @xmath3 at @xmath238 only if @xmath234 and consequently , we have the following conclusions regarding the sign of @xmath192 . 1 .
if @xmath239 and @xmath240 , we get @xmath241 for all @xmath242 , whereas @xmath243 for all @xmath244 and @xmath245 at @xmath4 .
2 . if @xmath239 and @xmath246 , we get @xmath241 for all @xmath247 .
3 . if @xmath248 , we get @xmath243 for all @xmath247 . from the above discussions , it is interesting to note that when @xmath34 then @xmath31 , since for @xmath34 , we have @xmath245 and the equation @xmath245 has only one solution @xmath31 within the admissible range of @xmath8 .
on the other hand when @xmath31 then @xmath34 , since for @xmath31 , we have @xmath245 and the equation @xmath245 has only one solution @xmath34 within the admissible range of @xmath3 .
therefore , we can conclude that when @xmath31 then @xmath34 and conversely , when @xmath34 then @xmath31 and consequently , at @xmath34 , we have @xmath249 , i.e. , there does not exist any solitary wave solution and/or double layer solution of the energy integral at @xmath23 when @xmath34(@xmath250 @xmath31 ) .
so far , we have found that if @xmath213 , then only ppsws are possible at @xmath23 for all @xmath247 . on the other hand , if @xmath214 and @xmath240 , then there exist npsw for all @xmath242 and ppsw for all @xmath244 .
however , there do not exist any solitary wave or double layer solution at @xmath4 .
now , the question is whether @xmath251 is really the upper bound of @xmath3 for the existence of ppsws . for this purpose , consider the existence of ppsw at @xmath23 for some @xmath252 and for some fixed @xmath253 .
therefore , there exists a @xmath143 such that @xmath254 now as @xmath173 is real for @xmath255 we must have @xmath256 , otherwise @xmath257 is not a real quantity .
therefore , @xmath258 defines a large amplitude ppsw , which is in conformity with ( [ ppsw ] ) , if @xmath259 and @xmath260 .
again let @xmath32 be the maximum value of @xmath3 up to which solitary wave solution can exist at @xmath23 .
as @xmath261 increases with @xmath2 and @xmath24 increases with @xmath3 then we have @xmath262 .
therefore , @xmath263 defines the largest amplitude solitary waves if @xmath264 therefore , for the existence of the ppsws , @xmath3 is restricted by the following inequality : @xmath265 , where @xmath32 is the largest positive root of the equation @xmath266 and @xmath267 for all @xmath268 .
thus for @xmath214 , ppsw exists in @xmath269 , where @xmath270 .
with the help of the above discussions , we have investigated the entire system numerically and we have found the following solitary wave solution at @xmath23 : ( i ) there exists a value @xmath35 of @xmath8 such that for @xmath28 there do not exist any solitary wave or double layer at @xmath23 for any value of @xmath3 in @xmath271 .
in fact , for any value of @xmath8 lies within @xmath28 , there do not exist any non - zero @xmath58 such that @xmath272 ; ( ii ) for any value of @xmath8 lies within @xmath273 , there exist only ppsws for all @xmath274 ; ( iii ) in @xmath275 , npsws exist for @xmath276 and ppsws exist in @xmath277 ; ( iv ) for all @xmath8 lies within @xmath278 , npsws exist in both @xmath279 and @xmath280 , whereas ppsws exist for all @xmath277 . for any point on the curve @xmath41 in @xmath278
, one can always find a ppdl solution .
these findings can be summarized in fig .
[ fig : sol space ] .
this figure gives the exact scenario of the nature of solutions of the system .
it looks that the solution space is corresponds to nonthermal distributed electrons .
however , if we move along @xmath27 axis , we can find different intervals of @xmath8 on the basis of the polarity of solitary wave solutions in presence of isothermal electrons . in this case , there does not exist any solitary wave or double layer for @xmath28 .
solitary wave starts to exist for @xmath281 .
only ppsws can be found for all @xmath8 lies within the interval @xmath273 , whereas for @xmath8 lies within @xmath282 , there exist only npsws .
we have taken two values of @xmath8 , one lies in @xmath273 and the other in @xmath282 and draw @xmath283 vs. @xmath58 in fig . [ fig : verify iso ] .
[ fig : verify iso](a ) and [ fig : verify iso](b ) verifies the solution space for isothermal electrons . at @xmath219 , there does not exist any solitary wave solution , the reasons of which being described in our theoretical section .
the amplitude of ppsw decreases with increasing @xmath8 in @xmath38 and ultimately the ppsws demolished at @xmath31 . on the other hand , npsw arrises in the frame after @xmath8 crosses @xmath36 and the amplitude of npsw gradually increases with @xmath8 in @xmath284 .
thus one may assume that the point @xmath31 acts as a source for npsws and as well as a sink for ppsws .
it is to be noted that double layer solution can not be found for maxwellian electrons .
now we are to verify the solution space for nonthermal electrons . for this purpose ,
we have taken a value of @xmath8 from each of the subintervals @xmath273 , @xmath275 , and @xmath278 and draw @xmath283 against @xmath58 in figs .
[ fig : verify mup to muc ] - [ fig : verify mur to mut ] for different values of @xmath3 . in fig .
[ fig : verify mup to muc ] , we have taken two values of @xmath3 lies within @xmath274 and both the curves verify the existence of ppsw .
[ fig : verify muc to mur ] shows that for a value of @xmath8 lies within @xmath275 there exist npsws for all @xmath285 , whereas only ppsws exist for all @xmath3 lies within @xmath286 .
again , ( a ) and ( b ) of fig . [ fig : verify mur to mut ] show the existence of npsws for all @xmath285 and ppsws for all @xmath3 lies within @xmath286 , respectively . fig .
[ fig : verify mur to mut](c ) shows the existence of npdl solution at @xmath41 for each @xmath8 lies within @xmath278 . a closer look at the solution space given by fig .
[ fig : sol space ] reveals an interesting feature near the curve @xmath41 .
we have seen that npsws exist for both @xmath287 and @xmath288 ( at least in a neighborhood of @xmath41 ) and a npdl exists at @xmath41 .
the immediate question is whether the characteristic of npsws in @xmath289 and @xmath290 are same . in search of the answer , we draw fig . [
fig : jump ] . in figs .
[ fig : jump](a ) and [ fig : jump](b ) , @xmath283 is plotted against @xmath58 for three different values of @xmath3 , viz . ,
@xmath10 , @xmath291 and @xmath292 . fig .
[ fig : jump](b ) shows that at @xmath5 , there exists a npdl whereas , for @xmath293 , there is a negative potential soliton .
however the curve corresponding to @xmath294 has no real root of @xmath58 in the neighborhood of @xmath295 , where @xmath296 is the amplitude of the npdl occurs at @xmath5 .
we notice from figure ( a ) that @xmath283 again vanishes far beyond @xmath295 for @xmath5 , @xmath293 , and @xmath294 . however , the curves corresponding to @xmath5 and @xmath293 are unable to produce any solitary wave solution but the curve corresponding to @xmath294 gives a npsw .
the amplitude of this solitary wave is much higher than that of @xmath292 .
thus there is a jump in amplitude of solitary waves and this phenomena has also been observed in some recent works.@xcite therefore , we can say that there exists two types of npsw , where the first type solitary wave is restricted by @xmath289 and the second type solitary waves exist for all @xmath290 .
moreover , there is a jump in amplitude between two types of solitary waves corresponding to the nonthermal parameter @xmath297 and @xmath298 , where @xmath299 is a sufficiently small positive quantity . in fig .
[ fig : amp neg beta1c ] variation in amplitude of npsws have been shown for values of @xmath3 lies within @xmath300 whenever @xmath40 .
this figure shows that the potential drop is maximum at @xmath41 and this is reason behind the occurrence of npdl solution at @xmath41 .
this figure also shows that amplitude of npsw decreases with increasing @xmath3 and ultimately , these npsws demolished at @xmath34 .
[ fig : dls at beta 1c ] verifies the fact that there exist npdl solution at @xmath41 for values of @xmath8 in @xmath40 .
this figure also shows that the amplitude of npdls at @xmath23 decreases with increasing @xmath8 in @xmath278 . again
[ fig : sol space ] , shows that @xmath10 is an increasing function of @xmath8 and hence the amplitude of npdls at @xmath23 decreases with increasing @xmath3 as well . in fig .
[ fig : amp both betac ] the variation in amplitude of npsw and ppsw have been shown for values of @xmath3 lies within @xmath242 and @xmath286 for @xmath275 .
this figure shows that amplitude of npsw decreases with increasing @xmath3 and ultimately , these npsws demolished at @xmath34 , whereas , ppsws start to occur beyond @xmath34 and amplitude of ppsw increases with @xmath3 in @xmath301 having maximum amplitude at @xmath42 .
therefore , we can conclude that @xmath34 acts like a sink for npsws and source for ppsws .
in fact the same phenomena occurs in @xmath40 when @xmath3 lies within @xmath302 and @xmath286 .
the entire numerical investigation depends on the solution space or compositional parametric space given by fig .
[ fig : sol space ] .
so for compactness of our investigation , it is customary to show that the solution space given by fig .
[ fig : sol space ] is unique . to show that this solution space is unique we have to show that the values @xmath35 , @xmath36 , and @xmath37 maintain the same orderings for any value of @xmath25 as they are in fig .
[ fig : sol space ] . to do this
, we have numerically investigated these values of @xmath8 and obtain fig .
[ fig : muprc ] .
this figure shows that for any value of @xmath25 , we always have @xmath303 .
therefore , for any value of @xmath25 , the nature of solutions follow fig .
[ fig : sol space ] .
the existence of dust ion acoustic solitary wave and double layer at the lowest possible mach number @xmath24 have been investigated in an unmagnetized nonthermal plasma consisting of negatively charged dust grains , adiabatic positive ions and nonthermal electrons .
an analytical approach has been constructed to find the polarity of the solitary wave or double layer . in the analytical approach
a function @xmath304 has been constructed , the real root @xmath33 of which is a critical value of the nonthermal parameter @xmath3 .
solitary waves exist for @xmath305 as well as @xmath252 and the polarity of the solitons have been determined by the sign of @xmath304 .
this critical value @xmath33 of @xmath3 acts as sink of negative solitons and the source of positive solitons . however , it is impossible to find any solitary wave at @xmath34 .
we have found the upper bounds of the nonthermal parameter for the existence of negative and positive potential solitary waves .
we have also found the existence of negative potential double layer at some point of the nonthermal parameter and observe a jump in amplitude of negative solitons at that point , the jump being the amplitude of the double layer formed thereat .
depending on these bounds or critical values of the nonthermal parameter , we have constructed a compositional parameter space for the system . in this solution space
, we have delimit the admissible range of @xmath8 into three disjoint subintervals @xmath29 , @xmath306 , and @xmath278 , by means of three critical points @xmath35 , @xmath36 , and @xmath37 of @xmath8 .
however , there do not exists any solution in @xmath307 . from this solution space
we have also found the delimitation of @xmath8 for isothermal electrons . in this case ,
positive solitons exist for @xmath29 , whereas only positive solitons are possible in @xmath282 .
it is impossible to find any solitons at @xmath219 .
we have not found any double layer solution for isothermal distributed electrons . here , we have considered @xmath43 as the upper bound of @xmath8 , which is strictly less than 1 , since for @xmath44 , the effect of charged dust on dia wave is negligible .
the amplitude of negative potential solitary waves decreases with increasing @xmath3 while the amplitude of positive potential solitary waves increases with @xmath3 .
50 c. k. goertz , rev .
geophys * 27 * , 271 ( 1989 ) .
n. n. rao , p. k. shukla , and m. y. yu , planet .
space sci . * 38 * , 543 ( 1990 ) .
p. k. shukla and v. p. silin , phys .
scr . * 45 * , 508 ( 1992 ) .
f. verheest , planet .
space sci . * 40 * , 1 ( 1992 ) .
barkan a , marlino r l and dangelo n. 1995 _ phys .
plasmas _ * 2 * 3563 pieper j b and goree j 1996 _ phys .
_ * 77 * 3137 mamun a a , cairns r a and shukla p k 1996 _ phys
. plasmas _ * 3 * 702 mamun a a , cairns r a and shukla p k 1996 _ phys .
plasmas _ * 3 * 2610 shukla p k 2001 _ phys
. plasmas _ * 8 * 1791 barkan a , dangelo n and marlino r l 1996 _ planet .
space sci . _
* 44 * 239 marlino
r l , barkan a , thompson c and dangelo n 1998 _ phys .
plasmas _ * 5 * 1607 nakamura y and sarma a 2001 _ phys . plasmas _ * 8 * 3921 shukla p k and rosenberg m 1999 _ phys
. plasmas _ * 6 * 1038 verheest f 2000 _ waves in dusty space plasmas _ ( kluwer academic , dordrecht ) shukla p k and mamun a a 2002 _ introduction to dusty plasma physics _ ( iop , bristol ) bharuthram r and shukla p k 1992 _ planet .
space sci .
_ * 40 * 973 mamun a a and shukla p k 2002 _ phys .
_ * 98 * 107 verheest f , cattaert t and hellberg m a 2005 _ phys . plasmas _ * 12 * 082308 sayed f and mamun a a 2008 _ pramana _ * 70 * 527 asbridge j r
, bame s j , and i. b. strong , j. geophys . res . *
73 * , 5777 ( 1968 ) .
feldman w c , anderson r c , bame s j , gary s p , gosling j t , mccomas d j , thomsen m f , paschmann g and hoppe m m 1983 _ j. geophys .
res . _ * 88 * 96 .
lundin r , zakharov a , pellinen r , borg h , hultqvist b , pissarenko n , dubinin e m , barabash s w , liede i and koskinen h 1989 _ nature _ ( london ) * 341 * 609 .
futaana y , machida s , saito y , matsuoka a and hayakawa h 2003 _ j. geophys .
_ * 108 * 1025 .
baluku t k , hellberg m a , kourakis i and saini n s 2010 _ phys
. plasmas _ * 17 * 053702 .
r. z. sagdeev , reviews of plasma physics , vol .
4 , ed m a leontovich ( consultant bureau , new york ) 1966 .
f. verheest and m. a. hellberg , phys .
plasmas * 17 * , 023701 ( 2010 ) .
f. verheest , phys .
plasmas * 17 * , 062302 ( 2010 ) .
t. k. baluku , m. a. hellberg , and f. verheest , epl * 91 * , 15001 ( 2010 ) .
cairns r a , mamun a a , bingham r , b@xmath308strom r , dendy r o , nairn c m c and shukla p k 1995 _ geophys . res .
* 22 * 2709 cairns r a , bingham r , dendy r o , nairn c m c , shukla p k and mamun a a 1995 _ j. phys .
iv _ * 5 * c6 - 43 cairns r a , mamun a a , bingham r and shukla p k 1996 _ phys .
_ * t63 * 80 mamun a a and cairns r a 1996 _ j. plasma phys . _ * 56 * 175 mamun a a , cairns r a and shukla p k 1996 _ phys
. plasmas _ * 3 * 702 bandyopadhyay a and das k p _ j. plasma phys .
_ 1999 * 62 * 255 , 2001 * 65 * 131 , 2002 * 68 * 285 ; _ phys .
_ 2000 * 61 * 92 , 2001 * 63 * 145 , _ phys .
plasmas _ 2000 * 7 * 3227 , 2002 * 9 * 465 , 2002 * 9 * 3333 .
das j , bandyopadhyay a and das k p _ j. plasma phys .
_ 2006 * 72 * 587 , 2007 * 73 * 869 , 2008 * 74 * 163 ; _ phys .
2007 * 14 * 092304 islam sk a , bandyopadhyay a and das k p _ j. plasma phys .
_ 2008 * 74 * 765 ; _ phys . plasmas _ 2009 * 16 * 022307 ; _ phys .
_ 2010 * 82 * 045505 das a , bandyopadhyay a and das k p _ phys . plasmas _ 2009
* 16 * 073703 , 2010 * 17
* 014503 verheest f and pillay s r 2008 _ phys . plasmas _ * 15 * , 013703 verheest f 2009 _ phys
. plasmas _ * 16 * 013704 a graphical presentation of different solitary structures have been given with respect to different subintervals of @xmath8 within the admissible interval of the nonthermal parameter @xmath3 .
in this solution space ` p ' stands for the existence of ppsw and ` n ' stands for the existence of npsw . for any given value of @xmath25 ,
npsws exist in @xmath309 and ppsws exist in @xmath277 . at any point on the curve @xmath41
, one can always find a npdl at @xmath23 .
this solution space has been drawn for @xmath310 .
we have found @xmath311 , @xmath312 , and @xmath313 . ]
@xmath283 is plotted against @xmath58 for isothermal distribution of electrons with values of @xmath8 lies within the intervals @xmath314 ( figure ( a ) ) and @xmath282 ( figure ( b ) ) . for @xmath310 ,
we have found @xmath315 and @xmath316 . ]
@xmath283 is plotted against @xmath58 for two different values of @xmath3 lies within @xmath318 ( figure ( a ) ) and @xmath244 ( figure ( b ) ) with values of @xmath8 lies within the intervals @xmath275 . for @xmath310 ,
we have found @xmath316 , @xmath319 and for @xmath320 , we have found @xmath321 . ] in ( a ) and ( b ) , @xmath283 is plotted against @xmath58 for values of @xmath3 lies within @xmath318 and @xmath244 , respectively with values of @xmath322 lies within the intervals @xmath275 . in ( c ) , the same has been plotted at @xmath41 . for @xmath310 and for @xmath323 , we have found @xmath324 and @xmath325 . ]
@xmath283 is plotted against @xmath58 for three different values of @xmath3 , viz .
, @xmath10[ ] , @xmath291[@xmath326 and @xmath292[- - - ] in ( a ) and ( b ) . in ( b ) we have shown the region of @xmath58 from @xmath327 to @xmath133 , whereas in ( a ) a region of @xmath58 from @xmath328 to @xmath329 has been shown . the npdl profile corresponding to @xmath41 has been shown in ( c ) , whereas the profiles of npsws corresponding to @xmath294 and @xmath293 have been drawn in ( d ) .
profiles in ( d ) shown a jump in amplitude of solitary wave by going from @xmath294 to @xmath293 . in all these four figures we have used @xmath310 , @xmath322 . ] variation in amplitude of npsws have been shown for values of @xmath3 lies within @xmath300 .
this figure shows that amplitude of npsw decreases with increasing @xmath3 and ultimately , these npsws demolished at @xmath34 . for @xmath310 and @xmath322 ,
we have found @xmath330 and @xmath324 . ]
@xmath283 is plotted against @xmath58 and @xmath41 for four different values of @xmath8 , viz .
, @xmath331 ( - @xmath332 - ) , @xmath333 ( @xmath334 ) , @xmath322 ( - - - ) and @xmath335 ( ) and respective @xmath10 s lies in the neighborhood of @xmath336 , @xmath337 , @xmath338 and @xmath339 . ]
variation in amplitude of npsw and ppsw have been shown for values of @xmath3 lies within @xmath242 and @xmath286 .
this figure shows that amplitude of npsw decreases with increasing @xmath3 and ultimately , these npsws demolished at @xmath34 , whereas , ppsws start to occur beyond @xmath34 and amplitude of ppsw increases with @xmath3 in @xmath301 having maximum amplitude at @xmath42 . for @xmath310 and @xmath340 ,
we have found @xmath321 and @xmath341 . ] | the sagdeev potential technique has been considered to find the necessary condition for the existence as well as the polarity of dust ion acoustic solitary structures at @xmath0 , in an unmagnetized nonthermal plasma consisting of negatively charged dust grains , adiabatic positive ions and nonthermal electrons , where @xmath1 is the smallest possible value of the mach number @xmath2 .
depending on the analytical theory , a compositional parameter space showing the nature of existence of solitary structures at @xmath0 has been presented .
the compositional parameter space clearly indicates the upper bounds of the nonthermal parameter ( @xmath3 ) associated with the nonthermal distribution of electrons for the existence of negative and positive potential solitons at @xmath0 .
the compositional parameter space also shows that the present system does not support the coexistence of both negative and positive solitons at @xmath0 .
a critical value of the nonthermal parameter ( @xmath4 ) , which acts as sink of negative solitons and the source of positive solitons at @xmath0 , has been found . however , no solitary wave is possible at @xmath4 .
the compositional parameter space again shows that the present system supports the negative potential double layer along the curve @xmath5 in @xmath6-parametric plane provided that @xmath7 , i.e. , for each @xmath8 lying within the interval @xmath9 , there exists one and only one @xmath10 of @xmath3 such that one can get negative potential double layer at @xmath0 when @xmath5 , where @xmath8 is the ratio of unperturbed number density of nonthermal electrons to that of ions . from the compositional parameter space
, it has been observed that the negative potential solitons at @xmath0 are separated by the curve @xmath5 . a finite jump between amplitudes of negative potential solitons separated by the curve @xmath5 has been observed .
the effect of the parameters on the amplitude of the solitary structures have been discussed . |
let @xmath2 be a @xmath3-variate random variable following a gaussian mixture model ( gmm ) with density @xmath4 where the @xmath5 s are the mixing proportions , with @xmath6 @xmath7 and @xmath8 , and the component @xmath9 represents the density of a @xmath3-variate normal distribution with mean vector @xmath10 and covariance matrix @xmath11 ; furthermore , let us indicate the set of model parameters with @xmath12 and the parameter space with @xmath13 where the symbol @xmath14 refers to lwner ordering on symmetric matrices and , in this case , is equivalent to requiring that @xmath11 be positive definite .
the gmm is frequently used to classify a sample of observations .
the idea is to consider the sample as drawn from a heterogeneous population where each sub - population is described by one component of the mixture . in other terms
, each observation is assumed to come from one of the @xmath15 different groups characterized by the mixture components .
the observations are classified into the groups by computing the posterior probabilities @xmath16 and assigning each observation to the group with the largest posterior probability .
the parameters of the gmm are generally unknown and estimated from the data . given a sample of i.i.d .
observations @xmath17 , the estimation is usually done by maximizing the likelihood @xmath18 the likelihood in equation ( [ like ] ) is known to be unbounded and it is cursed by the presence of several local maxima . as a consequence , the em algorithm may fail to converge , leading to such degenerate solutions . to face degeneracy ,
several methods have been proposed by the literature in which constraints or penalties are added to the log - likelihood .
their main objective is to keep the eigenvalues of the class conditional covariance matrices bounded away from zero .
this paper considers the sufficient condition formulated by ingrassia ( 2004 ) such that hathaway s ( 1985 ) constraints hold : we propose a generalization that enforces the equivariance with respect to linear affine transformations of the data .
the idea is to shrink the class conditional covariance matrices towards a pre - specified matrix @xmath0 we investigate possible data - driven methods for choosing the matrix @xmath1 when _ a priori _ information on the group - specific covariance structure is not available , and we let the data determine the optimal amount of shrinkage .
the equivariance property the method possesses is a key feature for twofold reasons .
first , it means that irrespective of the kind of standardization performed on the data , the final clustering will be the same .
second , whatever the scale of the data is as they come in , there will be no _ best _ pre - processing of the data ensuring a _ better _ result , as the final clustering is not affected by changes in scale .
the plan of the paper is the following .
section [ degen ] gives insights on the notion of degeneracy for multivariate gmm , and section [ remdegen ] reviews some of the workarounds proposed by the existing literature . in section [ invar ] we state the property of equivariance of gmm and we show , in section [ constr1 ] , that the property holds in the constrained approach of hathaway ( 1985 ) , whereas it does not hold in the sufficient condition provided by ingrassia ( 2004 ) . in section [ constr2 ]
we illustrate how these constraints can be generalized to become equivariant under linear affine transformations of the data , and how their configuration can be tuned from the data ( section [ crossvalid ] ) .
section [ alg ] summarizes the algorithm .
the proposal is evaluated through a simulation study ( section [ simulation ] ) and an empirical application ( section [ wineapp ] ) .
section [ concl ] concludes with a final discussion .
in the univariate case , the likelihood function increases without bound if some variances tend to zero and one of the components mean coincides with a sample observation ( kiefer and wolfowitz , 1956 ; day , 1969 ) .
biernacki and chrtien ( 2003 ) showed that if mixture parameters are close to a degenerate solution , then the em is attracted by it and the divergence is extremely fast .
although kiefer ( 1978 ) proved that maximum likelihood does not fail , as there exists a local maximizer strongly consistent and asymptotically efficient , several local maximizers can exist for a given sample .
that is , some local maximizers are spurious , i.e. with a high likelihood but of little practical use because highly biased .
they are characterized by some component variances and mixing proportions very small relative to the others ( day , 1969 ; mclachlan and peel , 2000 ) .
detecting the desired solution , among the many available , can therefore be a complicated task .
the same problems hold in the multivariate case ( as an example , see ingrassia and rocci , 2011 , for an extension of biernacki and chrtien , 2003 ) , with additional complications . to notice how unboundedness is caused ,
first of all let us express the density of the @xmath19-th observation on the @xmath20-th component as follows @xmath21 where @xmath22 is the square @xmath23 matrix whose @xmath24th column is the eigenvector @xmath25 of @xmath26 and @xmath27 is the diagonal matrix whose diagonal elements are the corresponding eigenvalues @xmath28 ordered such that @xmath29 equation can be rewritten as @xmath30[\mathbf{q}_{jg}'(\mathbf{x}_{i } - \boldsymbol{\mu}_{g } ) ] \right\ } \\ & = \frac{1}{\sqrt{2\pi \prod_{j=1}^{j}\lambda_{jg } } } \exp\left\ { -\frac{1}{2 } \sum_{j=1}^{j}\lambda_{jg}^{-1}[(\mathbf{x}_{i } - \boldsymbol{\mu}_{g})'\mathbf{q}_{jg}]^{2 } \right\}. { \addtocounter{equation}{1}\tag{\theequation}}\label{eq : li2parts}\end{aligned}\ ] ] as policiello ( 1981 ) argued , the likelihood can be written as the sum of non negative terms . among them , it is possible to isolate the product of the density of the @xmath19-th observation on the @xmath20-th component - equation - and the densities of the other observations on the other components and the corresponding mixing proportions . if observation @xmath19 is such that @xmath31 then , as @xmath32 there would be no exponential term involving @xmath33 who can attenuate the effect of @xmath34 in words , the sample likelihood diverges when in one component the covariance matrix is close to singularity and the projection of the component s mean on the eigenvector corresponding to the smallest eigenvalue coincides with the projection of one of the observations on the same eigenvector .
the easiest way to handle degeneracy is to initialize the em algorithm from several starting points until a local maximum is found ( biernacki and chrtien , 2003 ) .
mclachlan and peel ( 2000 ) proposed monitoring the local maximizers by inspecting the relative size of the estimated mixing proportions and component variances .
this leads , in practice , to performing maximum likelihood estimation by looking for the correct local maximum and discarding those that seem to be spurious .
further methods exploit constraints on the covariance matrices .
this approach is based on the seminal work of hathaway ( 1985 ) , where he studied how to avoid the divergence of the likelihood in the univariate case by imposing a lower bound , say @xmath35 , to the ratios of the scale parameters . in this way the variances can not be arbitrarily different . hathaway proved the boundedness of the likelihood and the consistency of the ml estimator under such constraints . in the multivariate case , the lower bound
is imposed on the generalized eigenvalues of each pair of covariance matrices and the ml estimator results to be equivariant under linear affine transformations of the data .
this implies that , as in the unconstrained case , if the data are linearly transformed , the estimated posterior probabilities do not change and the clustering remains unaltered ( see sections [ invar ] and [ constr1 ] ) .
an important issue is the choice of the constant @xmath36 which controls the strength of the constraints . in the context of univariate mixtures of gaussians or linear regression models
, some authors have shown that the maximum likelihood constrained estimator is consistent if @xmath35 decreases to zero at a certain rate as the sample size increases to infinity ( e.g. tanaka and takemura ( 2006 ) , tan et al .
( 2007 ) , xu et al . ( 2010 ) ) .
nevertheless , finite - sample sensible choice of @xmath35 is still an open issue .
hathaway s constraints are very difficult to apply within iterative procedures like the em algorithm . to solve this problem , ingrassia ( 2004 ) proposed to simplify the constraints by putting bounds on the eigenvalues of the covariance matrices .
although putting lower bounds on the group conditional covariance matrices was already common practice , ingrassia ( 2004 ) found a way to reconcile hathaway s contribution with the common practice : his bounds on the eigenvalues give a sufficient condition such that hathaway s constraints are satisfied .
the simplification is such that the constraints can be easily implemented within the em algorithm , preserving its monotonicity property ( as shown in ingrassia and rocci , 2007 ) .
several authors extended the constrained setup of ingrassia ( 2004 ) .
greselin and ingrassia ( 2013 ) applied this setup to mixtures of factor analyzers .
they proposed a tuning procedure for selecting the bounds for the eigenvalues of the covariance matrices , based on the final likelihood over a set of runs .
ingrassia and rocci ( 2011 ) modified the constrained algorithm , allowing for stringent constraints which are lifted during the iterations .
browne et al . ( 2013 ) combined the ideas in ingrassia and rocci ( 2007 , 2011 ) , constraining dynamically the smallest eigenvalue , the largest eigenvalue and both the smallest and the largest ones .
all of these proposals share the drawback of not being affine equivariant .
gallegos and ritter ( 2009a ; 2009b ) , and ritter ( 2014 ) applied hathaway s constraints to robust clustering .
they proposed to obtain all local maxima of the trimmed likelihood and , for each solution , investigate the value of @xmath35 such that it fulfills the constraints .
the idea is to choose , _ a posteriori _ , the solution with the highest trade - off between scale balance ( @xmath35 ) and fit ( log - likelihood ) .
this approach can be viewed as a refined version of what was proposed in mclachlan and peel ( 2000 ) .
garcia - escudero et al .
( 2008 ) , from the same strand of literature , introduced the tclust algorithm , based on controlling the relative sizes of the eigenvalues of the cluster scatter matrices .
the tclust algorithm implies solving several complex optimization problems .
fritz et al . ( 2013 ) and garcia - escudero et al . ( 2014 ) proposed further improvements to the algorithm in order to make it more efficient .
the constraints considered therein are not affine equivariant .
seo and kim ( 2012 ) pointed out that singular and spurious solutions overfit random localized patterns composed of few observations in the dataset .
such observations have a strong influence on the formation of the likelihood - based solutions .
their proposal was to take out such , say , @xmath37 observations with the highest likelihood ( likelihood - based @xmath37-deleted method ) , or with the highest value for a score - based statistic ( score - based @xmath37-deleted method ) . in this way
the likelihood of the reduced samples is evaluated at each local maximizer previously found : the root they suggested to select is the one with the highest @xmath37-deleted likelihood .
kim and seo ( 2014 ) show that their score - based method can be fairly well approximated with the computationally more efficient gradient - based version of the @xmath37-deleted method .
the degeneracy problem may also be addressed by adding a penalty to the log - likelihood ( penalized approach ) .
ciuperca et al .
( 2003 ) have shown the consistency of the penalized likelihood estimators proposed in ridolfi and idier ( 1999 , 2000 ) for univariate gmm .
chen and tan ( 2009 ) extended the consistency result for the multivariate case . in this framework
, the penalty term on the component covariance is added to the log - likelihood ( snoussi and djafari , 2001 ; chen et al , 2008 ) .
this penalty can be interpreted as the log of the prior distribution in a maximum - a - posteriori estimation setup . yet , the penalized methods are not affine equivariant , unless the prior s hyperparameters are suitably transformed .
map estimation , with an _ a priori _ distribution for the covariance matrices , is what fraley and raftery ( 2007 ) suggested to use , instead of maximum - likelihood , to circumvent the issues of degeneracy and spurious solutions .
the maximum likelihood estimators ( mle ) of equation ( [ md ] ) are equivariant with respect to linear affine transformations of the data .
that is , if the data are linearly transformed , the mle are transformed accordingly .
this property is particularly important in classification because it implies that linear affine transformations of the data do not change the posterior estimates ( kleinberg , 2002 ; ritter , 2014 ) .
the equivariance property can be shown in the following way .
let us define a linear affine transformation @xmath38 , where @xmath39 is non singular .
it is well known that @xmath40 where @xmath41 and @xmath42 .
this implies that , denoting the likelihood of the original data with @xmath43 and the likelihood of the transformed data with @xmath44 , we have , with obvious notation @xmath45 it follows that there exists a one to one correspondence among the local maxima of @xmath43 and @xmath44 .
in particular , if @xmath46 is a local maximizer for @xmath47 then @xmath48 will be a local maximizer for @xmath44 .
analogously , if @xmath49 is a local maximizer for @xmath44 , then @xmath50 will be a local maximizer for @xmath43 .
it is interesting to note that every pair of local maximizers produces the same estimates of the posterior probabilities , that is @xmath51 the above equality proves that the classification obtained via the gmm model is invariant under the group of linear affine transformations on the data @xmath52 .
this property is crucial when dealing with practical applications as it implies that the clustering does not depend on the choice of a particular method of data standardization - which could instead affect the inference .
hathaway ( 1985 ) proposed to impose the following restrictions on the covariance matrices @xmath53 where @xmath54 is the @xmath55-th eigenvalue of @xmath39 and @xmath56 .
this prevents the likelihood from diverging and reduces the number of spurious maximizers .
however , the method is difficult to implement and a correct choice of @xmath35 is not simple in practice
. a value of @xmath35 close to @xmath57 could exclude the correct solution , whereas a value too close to @xmath58 is likely to increase the chance of converging to a spurious maximizer .
ingrassia ( 2004 ) simplified hathaway s constraints as @xmath59 it is easy to show that ( [ nc ] ) implies hathaway s constraints ( [ ch ] ) while the reverse is not necessarily true ( ingrassia , 2004 ) .
this ensures a bounded likelihood , and a reduction in the number of spurious maximizers .
the constraints are easy to implement , as shown in rocci and ingrassia ( 2007 ) ; however , choosing an optimal _ c _ is still an issue .
it is important to check if the above constrained approaches offer equivariant estimators under linear affine transformations .
the property can be shown to hold for hathaway s approach as follows .
let @xmath17 be a sample of i.i.d .
the estimates are computed as the solution of the optimization problem @xmath60 given the transformation @xmath61 the maximand in equation ( [ eq : maxcon1 ] ) can be rewritten ( see section [ invar ] ) as @xmath62 where @xmath41 and @xmath42 .
noting that @xmath63 we can equivalently write the optimization problem in equation ( [ eq : maxcon1 ] ) as @xmath64 it follows that if @xmath46 is a maximizer for ( [ eq : maxcon1 ] ) , then @xmath65 is a maximizer for ( [ eq : maxcon12 ] ) and vice - versa , and the two maximization problems are equivalent . as in the unconstrained case , every pair of local maximizers produces the same estimates of the posterior probabilities .
this property does not hold for the constraints given in ( [ nc ] ) .
that is , if @xmath66 is a constrained local maximizer for @xmath43 subject to ( [ nc ] ) , @xmath67 does not necessarily satisfy ( [ nc ] ) . as an example , let us suppose that @xmath68 where @xmath69 is the largest singular value of @xmath39 . in this case , for a given @xmath70 @xmath71 we conclude that @xmath72 does not satisfy the constraints in ( [ nc ] ) because @xmath73 , and then it can not be a constrained local maximizer for @xmath44 .
constrains in ( [ nc ] ) are such that there is no one to one correspondence between the set of local maximizers of @xmath43 and @xmath44 .
thus , the method suffers the disadvantage that the clustering depends on the choice of matrix @xmath39 . to fix this ,
data standardization is not the best way to go for two main reasons .
first , the standardization requires a choice for the matrix @xmath39 and , second , there is no single best approach to data standardization ( milligan and cooper , 1988 ; doherty _ et al _ , 2007 ) .
it is now clear that affine equivariance is not just a desirable property .
it is one of the basic requirements of any clustering method , which should not be sensitive to the changes in the units of measurement of the data . in the next section
, our goal will be that of deriving a new set of constraints that are affine equivariant . with an affine equivariant clustering method ,
researchers and practitioners shall not be concerned anymore with choosing what method to adopt to standardize their data .
our proposal is to generalize the constraints ( [ nc ] ) by @xmath74 where @xmath75 is a symmetric positive definite matrix representing our _ prior _ information about the covariance structure .
clearly , is equal to ( [ nc ] ) when @xmath76 .
it can be shown that the above constraints imply hathaway s constraints .
it is known that ( anderson and gupta , 1963 ) @xmath77 where @xmath39 is a positive semi - definite matrix and @xmath78 and @xmath79 are positive definite matrices .
now , if holds , then @xmath80 thus , implies ( [ ch ] ) . furthermore , it can be shown that is invariant under linear and affine transformations provided that @xmath75 is replaced by @xmath81 .
if @xmath66 is a constrained local maximizer for @xmath43 subject to , then @xmath82 is a local maximizer for @xmath44 subject to for @xmath83 .
we have that @xmath84 in words , if a linear affine transformation is performed on the data , @xmath75 must be changed accordingly .
this scheme of transforming @xmath75 ensures the equivariance of the method .
the constraints have the effect of shrinking the covariance matrices to @xmath1 and the level of shrinkage is given by the value of @xmath85 note that for @xmath86 , @xmath87 whereas for @xmath88 @xmath89 equals the unconstrained ml estimate .
furthermore we can show that the stein s discrepancy - known as stein s loss ( james and stein , 1961 ) - between the matrices @xmath89 and @xmath75 goes to zero as @xmath35 approaches one .
the stein s discrepancy between the matrices @xmath89 and @xmath75 is @xmath90 let us rewrite equation as follows .
@xmath91 using the constraints in , we can derive the following majorizing function @xmath92 which is decreasing in @xmath85 this can be shown by noting that the first derivative of the right - hand side of with respect to @xmath35 is equal to @xmath93 and is negative when @xmath56 .
this implies that the function is decreasing when @xmath35 increases within the interval @xmath94 $ ] .
intuitively , the constraints provide with a way to obtain a model in between a too restrictive model , the homoscedastic , and an ill - conditioned model , the heteroscedastic .
issues arise when _ a priori _ information about the structure of the class conditional covariance matrices is not available . in that case ,
@xmath75 and @xmath35 have to be selected from the data . from the previous discussion , for a given @xmath36
every @xmath89 can not be too far from @xmath75 in terms of stein s discrepancy .
thus @xmath75 can be seen as the barycenter of the cloud of the @xmath89 s : the _ average _ conditional covariance matrix .
therefore , the most natural choice is to estimate such _ average _ as the within covariance matrix of the homoscedastic gaussian model .
how close the final clustering will be to the homoscedastic model will depend on the value of the tuning constant : for values of @xmath35 close to @xmath58 , the resulting clustering will be close to that of the heteroscedastic mixture model , whereas @xmath95 implies a clustering close to that of the homoscedastic mixture model . other possible choices of @xmath1 which guarantee the equivariance of the constraints , are available : the sample covariance matrix , which is computationally faster and is frequently used as hyperparameter in bayesian gaussian mixtures ( for instance , see fraley and raftery , 2007 ) , or the within covariance matrix of a homoscedastic mixture of student-_t_. to motivate this , let us recall that a random vector conditionally distributed as a multivariate gaussian , given wishart inverse covariance matrix , has a multivariate student-@xmath96 distribution ( dawid , 1981 ; dickey , 1967 ) .
using similar arguments as in peel and mclachlan ( 2000 ) , if @xmath97 is a gmm , and @xmath98 are i.i.d .
wishart random variables , the marginal distribution of @xmath2 is a homoscedastic mixture of student-@xmath96 s .
the choice of @xmath35 is crucial .
a value of @xmath35 too large could exclude the right solution , whereas a too small value of @xmath35 is likely to increase the chance to converge to spurious local maxima : such solutions overfit random localized pattern composed of few data points being almost co - planar ( ritter , 2014 ; seo and kim , 2012 ) .
hence , selecting @xmath35 jointly with the mixture parameters by maximizing the likelihood on the entire sample would trivially yield a scale balance approaching zero .
a practical alternative would be to split the data into a training set , where model parameters are estimated , and a test set , where the log - likelihood is evaluated for a given value of @xmath85 the optimal tuning parameter @xmath35 would then be selected such that the test set log - likelihood is maximized .
the use of the test set log - likelihood as a model selection tool is advocated by smyth ( 1996 ; 2000 ) , in the context of estimating the number of mixture components .
the motivation behind its use is that it can be showed to be an unbiased estimator ( within a constant ) of the kullback - leibler divergence between the _ truth _ and the model under consideration ( smyth , 2000 ) .
this means that , even under a misspecified model , the procedure renders a @xmath35 such that the kullback - leibler divergence is minimized . in spite of the usual unavailability of large independent test sets ,
a valid alternative is to use the _ cross - validated _ log - likelihood in order to estimate the test set log - likelihood .
this consists in repeatedly partitioning the data into training and test sets and , for a given @xmath36 estimate the mixture parameters on the training sets .
the model fit is then measured summing the log - likelihoods of the test sets evaluated at the parameters computed on the training sets , obtaining the so - called _ cross - validated _ log - likelihood .
the constant @xmath35 is chosen such that the _ cross - validated _ log - likelihood is maximized .
this can be viewed as a function of @xmath35 only ( smyth , 1996 ) , and would solve the issue of overfitting as training and test sets are independent ( arlot and celisse , 2010 ) . in details ,
let us partition @xmath99 times the full data set @xmath100 into two parts , a training set @xmath101 and a test set @xmath102 with @xmath103 and @xmath104 for the @xmath37-th partition , let @xmath105 be the constrained maximum likelihood estimator based on the training set @xmath106 furthermore , let @xmath107 $ ] be the log - likelihood function evaluated at the test set @xmath108 the _ cross - validated _ log - likelihood is defined as the sum of the contributions of each test set to the log - likelihood @xmath109.\ ] ] the best @xmath35 is chosen as the maximizer of @xmath110 .
further details on the choice of the number of random partitions @xmath99 and of the sizes of training and test sets are given in section [ simulation ] .
the objective is to maximize ( [ like ] ) under the constraints .
thanks to the equivariance property of the constraints , we can act any linear affine transformation to the data .
this is useful since it will suffice to transform the data so to have @xmath111 and the existing algorithm of ingrassia and rocci ( 2007 ) can be applied on the transformed data .
the transformation is @xmath112 where @xmath113 is the singular value decomposition of @xmath0 this leads to @xmath114 . for sake of completeness
, we recall briefly the updates of the algorithm proposed by ingrassia and rocci ( 2007 ) .
+ + * update * @xmath115 , @xmath116 @xmath10 + as in the case of a normal mixture , the updates are @xmath117
@xmath118 @xmath119 * update * @xmath11 + compute @xmath120 and set @xmath121 where @xmath122 is the diagonal matrix of the eigenvalues in non decreasing order of @xmath123 and @xmath124 its singular value decomposition .
letting @xmath125 , the update of @xmath11 is given by @xmath126
in this section we perform a simulation experiment in order to compare the performance of the proposed methods with respect to some existing approaches in the literature .
in particular we consider the following seven algorithms : 1 .
unconstrained 1 .
homoscedastic normal ( homn ) , within covariance matrix @xmath127 2 .
heteroscedastic normal ( hetn ) , @xmath128 to prevent degeneracy and numerical instability ; 3 .
homoscedastic student _
_ ) , scale matrix @xmath129 @xmath130 ( mclachlan and peel , 1998 ) ; 4 . gradient - based @xmath37-deleted mle ( kdel ; kim and seo , 2014 ) , @xmath128 to prevent degeneracy and numerical instability , k@xmath131 .
2 . constrained 1 .
sample covariance ( con@xmath132 ) , @xmath133 2 .
normal ( conn ) , @xmath134 3 . student _ t _
( con__t _ _ ) , @xmath135 for each sample , we randomly split the data @xmath136 times into a training set @xmath137 and a test set @xmath138 choosing how many times to partition the full data set is a trade - off between variability of the estimates and computational burden . as smyth ( 2000 )
argues in the context of model selection for probabilistic clustering using cross - validation , the larger the value of @xmath139 the less the variability in the log - likelihood estimates . in practice - the author argues - values of @xmath99 between 20 and 50 appear adequate for most applications . the choice of the size of the test set must be such that the training set has all components represented .
if one component is not represented in the test set , but the parameters are correctly estimated using the training set , the test set log - likelihood will correctly display the fit of the model .
by contrast , if one component is not represented in the training set , although estimation of the other components parameters can be correct , the fit displayed by the test set log - likelihood will be poor .
van der laan , dudoit , and keles ( 2004 ) found , in their simulation study , that the likelihood - based cross - validation procedure is performing equally well with any choice of the relative size of the test set between 0.1 and 0.5 .
as argued in kearns ( 1996 ) , the importance of choosing an optimal size for the training set increases as the target function becomes more complex relative to the sample size .
bearing this in mind , we choose to consider a training set of size @xmath140 and a test set @xmath141 of size @xmath142 then the cross - validation scheme , as described in section [ crossvalid ] , is applied and the optimal @xmath35 is chosen by using a line search with six function evaluations .
the sample data have been generated from @xmath15-class mixtures of heteroscedastic @xmath3-variate normal distributions with : * @xmath143 @xmath144 @xmath145 * @xmath146 @xmath147 * @xmath148 @xmath149 * prior probabilities @xmath150 @xmath151 @xmath152 * component means @xmath153 , independent ; * eigenvalues of the covariance matrices @xmath154 , independent with @xmath155 ; * eigenvectors of the covariance matrices generated by orthonormalizing matrices generated independently from a standard normal distribution .
this yields a total of @xmath156 simulation conditions . for each simulation condition
, we generate 250 data sets , each with different means and covariance matrices .
it is well known that the em for gmm is sensitive to the initial position , especially in the multivariate context ( among others , mclachlan and peel , 2000 ) .
we choose to adopt the standard _
multiple random starts _ strategy .
that is , for each data set , 10 random initial partitions are generated : these are used as starting values for the m - step of all the seven algorithms under analysis . for conn , cons , and cont , a constrained algorithm with arbitrary lower and upper bounds of respectively @xmath157 and @xmath158
is run in order to exclude degenerate ( and some spurious ) solutions , and the estimated clustering is used to initialize the cross - validation scheme .
the alternative option of directly generating 10 different starts for each training set - within the cross - validation scheme - would have added little in terms of accuracy of the final estimates . concerning the root selection criterion , for the unconstrained algorithm , we select the roots yielding the highest likelihood , whereas for the constrained algorithms we select the roots based on the cross - validated likelihood . the performance of the different techniques has been analyzed in terms of : * mad ( mean absolute deviation ) : @xmath159 * arand ( adjusted rand index ; hubert and arabie , 1985 ) ; * computational time needed to analyze a single data set ; * the value of the calibrated constant @xmath35 ( for the constrained approach only ) . the mad is computed evaluating the above expression for all possible permutations of the estimated classes .
the final mad reported refers to the permutation which yields the lowest difference , and measures inaccuracy of estimated _ fuzzy _ classification - whereas arand measures accuracy of estimated _ crisp _ classification .
in addition , we tested the robustness of the results with respect to changes in 1 ) the cross - validation settings , and 2 ) the level of class separation . in order to test robustness with respect to cross - validation settings
, we considered a subset of the above simulation conditions as follows .
250 samples , of 50 , 100 , and 200 observations , were generated from a 3-group 8-variate heteroscedastic gaussian mixture model , with prior class membership probabilities of 0.1 , 0.4 , and 0.5.-deleted method within @xmath3 and @xmath160 , as they suggested , on this reduced set of simulation conditions .
on average , k=@xmath3 yielded the best clustering .
the full set of results is available from the corresponding author upon request . ]
the same setting was used in order to count how many different local maxima each algorithm converged to over the 10 random initializations considered .
this serves the purpose of providing some information on the likelihood surface .
class separation has been manipulated by controlling the dispersion of the group conditional covariance matrices eigenvalues ( through the above sep value ) : higher dispersion levels correspond to overlap between the classes . considering the above full simulation as corresponding to fixed moderate separation ( @xmath155 ) , this final setup compares results for low , moderate , and high separation levels - respectively @xmath161 @xmath162 and @xmath163 .
the subset of simulation conditions considered is as follows .
250 samples , of 50 observations each , were generated from a 3-group and 5-group heteroscedastic gaussian mixture model , with prior class membership probabilities of respectively 0.2 , 0.3 , and 0.5 ; 0.1 , 0.4 , and 0.5 ; 0.1 , 0.1 , 0.2 , 0.3 , and 0.3 .
table [ tabsimcon ] summarizes the conditions explored in all testing setups .
.cross - table of simulation condition and simulation type . [ cols="<,^,^,^,^,^,^,^,^ " , ] the homoscedasticity assumption seems to fit well the data .
the constrained approach conn equals homn in terms of arand , whereas cont yields an arand of 0.93 , compared to 0.87 of homt .
confirming the results obtained in the simulation study , cont seems to be the most accurate approach among the ones considered in this work .
interestingly , however , all of the constrained approaches improve upon the unconstrained heteroscedastic approach and the gradient - based @xmath37-deleted method .
in this paper we have proposed affine equivariant constraints for the class conditional covariance matrices of multivariate gmm in order to circumvent the well - known issue of degenerate and spurious solutions in ml estimation . our approach generalizes the sufficient condition for hathaway ( 1985 ) s constraints to hold as formulated by ingrassia ( 2004 ) .
previous constrained approaches lacked affine equivariance and suffered the choice of an optimal finite - sample scale balance ( @xmath35 ) .
the setup we propose is such that the class specific covariance matrices are shrunk towards a pre - specified matrix @xmath0 we have been able to show that this yields a clustering method which is equivariant with respect to linear affine transformations of the data . a natural choice for the shrinkage target matrix , whenever _ a priori _ information on the covariance structure of the components is not available , seems to be the covariance matrix of a homoscedastic mixture of normals . for a given choice of the target matrix
, we let the data decide , through the constant @xmath35 , how close to the target the final clustering will be .
the tuning constant @xmath35 is chosen by cross - validation .
we have also shown that , given a matrix @xmath1 our constrained ml estimate can be computed by applying the algorithm of ingrassia and rocci ( 2007 ) to the data appropriately linearly transformed .
this allows us to interpret our proposal as a way to decide how to standardize the data before applying ingrassia ( 2004 ) s constraints .
the validity of the proposal has been assessed through a simulation study and an empirical example .
all constrained approaches yield more accurate estimates than the unconstrained ones .
more specifically , cont has been shown to be the best among the constrained approaches this work has been concerned with .
this is not surprising , since a random vector conditionally distributed as a gaussian mixture , given random covariance matrices , has a marginal homoscedastic mixture of student @xmath96 s distribution .
however , different choices of @xmath75 can as well be considered , according to the data specificity .
the equivariant method developed in gallegos and ritter ( 2009a ; 2009b ) and extended in ritter ( 2014 ) requires to obtain all local maxima of the trimmed likelihood .
our method has the virtue of being easily implementable with a minimal extra computational effort , as we have shown in the simulation study and in the empirical example . by comparing our results on the gradient - based @xmath37-deleted method with those obtained by kim and seo ( 2014 )
, we have observed that the @xmath37-deleted method suffered our choice of focusing on multiple random start strategy only .
indeed , the @xmath37-deleted method , being a root selection method , shows its full value with a more vast and diversified set of initialization strategies ( kim and seo , 2014 ) .
our research aim was rather to find a data - driven equivariant approach leading to an optimal compromise between the homoscedastic and the heteroscedastic model , irrespectively of the kind of initialization strategy employed .
there are cases where the clustering model might assume a specific structure on the relationship between the variables , like local independence ( within - cluster diagonal matrices ) .
such a model is not affine equivariant because some ( non diagonal ) affine transformations on the data might destroy the local independence . in cases like these
, the affine equivariance property of the constraints is not required . yet
our approach can be applied using a diagonal matrix as target .
this would prevent the likelihood from degenerating , still improving upon the unconstrained algorithm thanks to the cross - validation strategy we have proposed .
clearly , when all variables in a data set are measured in a common scale , non equivariant constraints are a competitive choice .
an additional issue , pointed out by both the simulation study and the empirical example , is the computational time cross - validation requires to select an optimal @xmath85 whether different cross - validation schemes can speed up the constrained routines can be a topic for future research .
biernacki , c. , celeux , g. , govaert , g. ( 2003 ) .
choosing starting values for the em algorithm for getting the highest likelihood in multivariate gaussian mixture models . _
computational statistics and data analysis , 41(3 ) , 561 - 575 . _ di mari , r. , oberski , d.l . ,
vermunt , j.k .
bias - adjusted three - step latent markov modeling with covariates . _
structural equation modeling : a multidisciplinary journal_. doi:10.1080/10705511.2016.1191015 .
ridolfi a. , idier j. ( 2000 ) .
penalized maximum likelihood estimation for univariate normal mixture distributions .
_ bayesian inference and maximum entropy methods _ , maxent workshops .
gif - sur - yvette , france , july 2000 .
smyth , p. ( 1996 ) . _
clustering using monte - carlo cross validation .
_ in _ proceedings of the second international conference on knowledge discovery and data mining , menlo park , ca , aaai press , pp
. 126133 .
_ snoussi h. , mohammad - djafari a. ( 2001 ) .
_ penalized maximum likelihood for multivariate gaussian mixture . _ in r.l .
fry editor _ maxent workshops : bayesian inference and maximum entropy methods , 36 - 46 , aug .
2001 . _ tan x. , chen j. , zhang r. ( 2007 ) . consistency of the constrained maximum likelihood estimator in finite normal mixture models .
proceedings of the american statistical association , american statistical association , alexandria , va , 2007 , pp .
2113 - 2119 .
[ cd - rom ] .
tanaka k. , takemura a. ( 2006 ) strong consistency of the maximum likelihood estimator for finite mixtures of location scale distributions when the scale parameters are exponentially small .
_ bernoulli , 12 ( 6 ) , 1003 - 1017_. | maximum likelihood estimation of gaussian mixture models with different class - specific covariance matrices is known to be problematic .
this is due to the unboundedness of the likelihood , together with the presence of spurious maximizers .
existing methods to bypass this obstacle are based on the fact that unboundedness is avoided if the eigenvalues of the covariance matrices are bounded away from zero .
this can be done imposing some constraints on the covariance matrices , i.e. by incorporating _ a priori _ information on the covariance structure of the mixture components .
the present work introduces a constrained equivariant approach , where the class conditional covariance matrices are shrunk towards a pre - specified matrix @xmath0 data - driven choices of the matrix @xmath1 when _ a priori _ information is not available , and the optimal amount of shrinkage are investigated .
the effectiveness of the proposal is evaluated on the basis of a simulation study and an empirical example . + * keywords * : model based clustering , gaussian mixture models , equivariant estimators |
among the continuous variable ( cv ) entangled states @xcite , non gaussian states are generally believed to possess much more robust entanglement vis a vis the gaussian states states characterized by a gaussian quasi probability distributions and hence by their first and second moments .
though mathematically not as well understood as the gaussian states @xcite in so far as their entanglement properties are concerned , non gaussian states , by virtue of the robustness of their entanglement have in recent years emerged as strong contenders for potential applications in quantum information technology .
the fact that the non gaussian states are defined by what they are not makes a general discussion of their entanglement properties impossible and one is forced to restrict oneself to sub families of non gaussian states such as states obtained by adding or subtracting a fixed number of photons to gaussian states @xcite and forming suitable superpositions thereof .
one such widely discussed family of non gaussian states parameterized by an integer @xmath0 is the family of noon states @xcite @xmath1.\ ] ] these states , similar in structure to the epr states , have attracted much attention in recent years and can be viewed as a two mode state consisting of a superposition of states containing @xmath0 photons in one mode and none in the other and vice versa .
schemes for reliable production of such states have been proposed @xcite and their usefulness as a practical tool in making super - precision measurements in optical interferometry , atomic spectroscopy and in sensing extremely small magnetic fields than hitherto possible have been highlighted @xcite .
this circumstance makes it imperative to investigate their behavior under attenuation and amplification . while studies on the decoherence effects on noon states under specific models for system - bath interactions already exist in the literature @xcite , in the present work we focus on the question of amplification of noon states and the consequent degradation of entanglement therein and to compare and contrast it with the behavior of entanglement in gaussian states under amplification investigated in @xcite .
it is important to understand the amplification of noon states as quantum communication protocols would use both amplifiers and attenuators @xcite . in the present work we consider only phase insensitive amplification and model this in the standard way as a bath consisting of @xmath0 two level atoms of which @xmath2 are in the excited state and @xmath3 in the ground state with @xmath4 . under the assumptions that atomic transitions have a large width and that the bath is maintained in a steady state
, the time evolution of the density operator @xmath5 for a single mode of radiation field on resonance with the atomic transition is described , in the interaction picture , by the master equation @xmath6 where @xmath7 and @xmath8 are the annihilation and creation operators of the field mode .
a brief outline of this work is as follows . in section
ii , with the noon state as the input to the amplifier , we obtain expressions for the output density operator for the case when both the modes are symmetrically amplified and for the case when only one mode is subjected to amplification and the other is not amplified at all .
we investigate how the entanglement in the output state varies with the amplifier gain using logarithmic negativity as the quantifier for entanglement @xcite .
section iii contains our concluding remarks and further outlook .
our study complements the work done by vitelli et al @xcite on amplification of noon state by a phase sensitive amplifier .
in our earlier work @xcite , we found that for a two mode squeezed vacuum as the input , there are limits on the gain beyond which the output of the amplifier has no entanglement between the two modes and the limiting values of the gain in the two cases considered , symmetric and asymmetric amplification , were found to be @xmath9 where the gain @xmath10 and @xmath11 . in particular , when @xmath12 , the limiting gain in the symmetric case remains finite , that for the asymmetric case the limiting value moves off to infinity @xcite . in this section
we discuss how an input noon state evolves under the action of a phase insensitive amplifier as modeled by the master equation @xmath13 confining ourselves for simplicity to the @xmath12 limit .
the solution of master equation ( 2 ) can be written in terms of the fock state matrix elements .
however such a solution is rather involved @xcite .
it is instructive to work in terms of phase space distributions .
a distribution which is especially useful is the q function introduced by kano ; sudarshan and mehta @xcite .
this function is defined by @xmath14 the master equation ( 2 ) for @xmath15 then leads to @xmath16 note that this differential equation for @xmath17 function involves only the first order derivatives with respect to phase space variables and hence its solution is simple @xcite @xmath18 let us see what the result @xmath19 means .
let us consider the input state to be vacuum then the output would be @xmath20 such an output @xmath17 function is equivalent to a thermal density matrix with mean number of photons equal to @xmath21 .
thus the vacuum state on amplification becomes a thermal state with a mean number of photons that grows with the gain of the amplifier .
we note that a result like ( 6 ) extends to multi - mode case .
we now examine two cases , the symmetric case in which both the modes @xmath7 and @xmath22 in the input noon state are symmetrically amplified and the asymmetric case in which only one mode , say @xmath7 , is amplified . before discussing the amplification of the noon state
we examine quantitatively the entanglement in the state ( 1 ) .
we compute the log negativity parameter which is defined as @xmath23 where @xmath0 is the absolute value of the sum of all the negative eigenvalues of the partial transpose of the density matrix @xmath5 .
it is clear that the partial transpose of the density matrix associated with the state ( 1 ) is @xmath24~,\end{aligned}\ ] ] which can be written in the diagonal form as @xmath25 the partial transpose has a negative eigenvalue @xmath26 and hence the logarithmic negativity parameter @xmath27 .
_ symmetric case _ : the @xmath17- function corresponding to the density operator for the input noon state @xmath28\nonumber\\ & = & \frac{1}{2n!}[a^{\dagger n}\rho_0a^n+a^{\dagger n}\rho_0b^n+b^{\dagger n } \rho_0a^n + b^{\dagger n}\rho_0b^n],\nonumber\\~~ \rho_0 & = & |0,0><0,0|~,\end{aligned}\ ] ] is found to be @xmath29\nonumber\\\end{aligned}\ ] ] following the prescription in @xmath19 , under symmetric phase insensitive amplification , the q - function evolves as follows : @xmath30\nonumber\\ \label{12}\end{aligned}\ ] ] the noon state is highly nonclassical . a quantitative measure for nonclassicality
is obtained by examining zeroes of the @xmath17 function @xcite .
we note that the zeroes of the function @xmath31 are identical to the zeroes of the function @xmath32 and hence we have the remarkable result that the nonclassical character of the input noon state is preserved .
we can now find the density matrix after amplification by using the results ( [ 7 ] ) and ( [ 12 ] ) : @xmath33,\nonumber\\ \rho_{g}&=&\frac{1}{g^4}e^{-\beta(a^\dagger a+b^\dagger b)},\nonumber\\ & & \nonumber\\~\beta & = & { \rm ln}\left(\frac{g^2}{g^2 - 1}\right ) .
\label{density}\end{aligned}\ ] ] we note that the structure of ( [ density ] ) is such that it can not be written in a separable form .
this is seen more clearly if we write ( [ density ] ) as @xmath34 we further note that the output state has the structure of a two mode photon added thermal state in which either mode has added photons . the single mode version of photon added thermal state was introduced by agarwal and tara @xcite .
these states have been experimentally studied recently @xcite .
writing @xmath35 in the number state basis as @xmath36 we have @xmath37\end{aligned}\ ] ] which immediately gives us the expression for the operator @xmath38 obtained by partially transposing @xmath39 ( with respect to the b mode ) : @xmath40\nonumber\\ & = & \frac{1}{2n!g^{2n}}[a^{\dagger
n}\rho_{g}a^n+b^{\dagger n}\rho_{g}b^n \nonumber\\ & & + a^{\dagger n}b^{\dagger n } \rho_{g}+\rho_{g}a^n b^n]\end{aligned}\ ] ] the object of interest now is to calculate the logarithmic negativity @xmath41 , the sum of the logarithmic negativity eigen - values of @xmath38 and to see how it varies as a function of @xmath42 .
we carry out this task numerically and the results are displayed in fig . 1 where we plot @xmath41 as a function of @xmath42 for @xmath43 and 6 .
_ asymmetric case _
: proceeding as above , one finds that @xmath44\end{aligned}\ ] ] and hence @xmath45\nonumber\\ \tilde{\rho } & = & \frac{1}{g^2}e^{-\beta(a^\dagger a)}|0><0|;~\beta={\rm ln}\left(\frac{g^2}{g^2 - 1}\right)\nonumber\\\end{aligned}\ ] ] writing @xmath46 as @xmath47 we can write @xmath39 in terms of number states as @xmath48\end{aligned}\ ] ] and hence @xmath49\label{b}\nonumber\\ & = & \frac{1}{2n!g^{2n}}[a^{\dagger n}\tilde{\rho}a^n+g^{2n}b^{\dagger n}\tilde{\rho}b^n\nonumber\\ & & + g^n a^{\dagger n}b^{\dagger n } \tilde{\rho}+g^n \tilde{\rho}a^n b^n]\end{aligned}\ ] ] the logarithmic negativity of @xmath38 is computed numerically and the results are shown in fig . 2 for @xmath50 .
clearly in the asymmetric case the loss of entanglement is much slower .
finally we compare the amplification of the noon state with that of photon added two mode squeezed vacuum state i.e. the state ( for brevity the normalization factors are ignored ) @xmath51 this is a non gaussian state .
the input and output @xmath17 functions are found to be @xmath52 @xmath53 where @xmath54 is the @xmath17 function for the two mode squeezed vacuum and @xmath55 is the @xmath17 function obtained by amplification of the squeezed vacuum .
thus the density operator after amplification of the non gaussian state can be written as @xmath56 where @xmath57 is the density operator for the squeezed vacuum after amplification .
now @xmath57 becomes separable for @xmath58 greater than that given by ( 3 ) and hence @xmath39 becomes separable if @xmath59 .
thus the non gaussian states obtained from gaussian states by the addition of photons would behave under symmetric amplification in a manner similar to gaussian states .
we have therefore found that the noon states behave quite differently under amplification .
in conclusion we have found that the noon states are more robust under amplification than their gaussian counter parts such as two mode squeezed vacuum produced by a down converter .
we have presented results for the logarithmic negativity as a function of the gain of the amplifier .
we have presented numerical results for cases of states which already have been realized experimentally .
we also found that the noon state does better than say photon added two mode squeezed vacuum state .
s. l. braunstein and p. van loock , rev .
phys . * 77 * , 513 ( 2005 ) ; n. cerf , g. leuchs , and e. s. pollzik ( eds . ) , _ quantum information and continuous variables of atoms and light _ , ( imperial college press .
london , 2207 ) .
r. simon , phys .
lett . * 84 * , 2726 ( 2000);l .- m .
duan , g. giedke , j.i .
cirac , p. zoller , phys .
rev . lett . * 84 * , 2722 ( 2000 ) ; g. giedke , b. kraus , m. lewenstein , j.i .
cirac , phys . rev
. lett . * 87 * , 167904 ( 2001 ) ; r.f .
werner , m.m .
wolf , phys .
86 , 3658 ( 2001 ) ; m.b .
plenio , phys rev .
lett * 95 * , 090503 ( 2005 ) ; g. adesso , a. serafini , f. illuminati , phys . rev .
a * 70 * , 022318 ( 2004 ) ; j. eisert , m.b .
plenio , j. mod .
46 , 145(1999 ) ; s. virmani , m.b .
plenio phys .
a * 268 * , 31 ( 2000 ) .
agarwal , k. tara , phys .
a * 47 * , 3160 ( 1993 ) ; t. opratrny , g. kuriziki , d .-
welsch , phys .
a * 61 * , 032302 ( 2000 ) ; p. t. cochrane , t. c. ralph , g. j. milburn , phys .
rev . a*65 * , 063206 ( 2002 ) ; s. olivares , m. g. a. paris and r. bonifacio , phys .
rev . a*67 * , 032314 ( 2003 ) .
m. dangelo , m. v. chekhova and y. shih , phys .
lett . * 87 * 013002 ( 2001 ) ; m. w. lundeen , a. m. steinberg nature * 429 * , 161 ( 2004 ) ; p. walther , j .- w .
pan , m. aspelmeyer , r. ursinand and a. zeillinger nature * 429 * , 158 ( 2004 ) ; anne e. b. nielsen , klaus molmer , phys .
rev . a * 75 * , 063803 ( 2007 ) ; p.
kok arxiv : quant - ph/0604139 s. boixo , a. datta , m. j. davis , a. shaji , a. b. tacla , and c. m. caves , phys . rev . a * 80 * , 032103 ( 2009 ) ; j a. jones , s d. karlen , j fitzsimons , a ardavan , s c. benjamin , g. a. d. briggs , j j. l. morton , science , * 324 * , 1166 ( 2009 ) ; d. w. berry , b. l. higgins , s. d. bartlett , m. w. mitchell , g. j. pryde , and h. m. wiseman , phys rev . a * 80 * , 052114 ( 2009 ) . s. scheel , t. opatrn and d .-
welsch ; s. scheel , l. knl , t. opatrn and d .-
welsch , phys .
a * 62 * , 043803 ( 2000 ) ; m hillery and m. s. zubairy , phys . rev . a * 74 * , 032333 ( 2006 ) ; p. s. y. poon and c. k. law , phys rev .
a * 76 * 012333 ( 2007 ) . | we examine the behavior of a non gaussian state like noon state under phase insensitive amplification .
we derive analytical result for the density matrix of the noon state for arbitrary gain of the amplifier .
we consider cases of both symmetric and antisymmetric amplification of the two modes of the noon state .
we quantitatively evaluate the loss of entanglement by the amplifier in terms of the logarithmic negativity parameter .
we find that noon states are more robust than their gaussian counterparts . |
the trapping potential used in this study is @xmath87 , \nonumber\end{aligned}\ ] ] where @xmath88 , @xmath89 are defined in the main text .
an illustration of this potential is given in fig .
[ fig : potential ] . the external barrier heights in @xmath88 were set to be @xmath90mev and the length of the system set to @xmath91 m as discussed in the main text .
we choose a value of @xmath16 such that the the degeneracy between the two possible ground states @xmath92 and @xmath93 is lifted sufficiently with respect to the temperature , where @xmath92 is the mirror inversion of @xmath93 about the @xmath94-axis .
the probability of excitation from @xmath92 to @xmath93 at temperature @xmath64 in the presence of the symmetry breaking potential @xmath95 is roughly @xmath96 .
we assume that the system is realised in gaas , at a temperature of @xmath97 , with @xmath98 . inserting these parameters @xmath99 , so @xmath100 nm is used to give a sufficiently low error and since we would expect this level of precision to be possible in experiments .
the internal barrier height @xmath12 was chosen so that the coupling between the dot electrons and the chain was significant , but the change in the positions of the electrons was still small . a change in @xmath12 from 3mev to 50@xmath68ev
was found to change the classical equilibrium electron positions by around @xmath101 of average electron separation , whilst still giving a reasonable coupling .
the variation in dot - chain coupling for differing barrier heights can be seen in fig .
[ fig : barrierheights ] .
the remaining parameters are chosen based on the electron configuration without the dot potentials present ( i.e. when @xmath102 ) .
for each particular situation , the minimum energy configuration @xmath103 was found without the dot potentials , and then the parameters were chosen such that the internal barriers would sit roughly halfway between the first two electrons , and such that the internal barriers have as little effect as possible on the separation between the dot electron and the chain .
thus @xmath104 , and empirically @xmath105 , @xmath106 was found to have little effect on the electron equilibrium positions .
we set @xmath107 , so that the effect of the two quadratic potentials for the dots is confined to within the barriers .
we set @xmath108 so that the confinement of the dots is much stronger than the confinement in the wire , and to further reduce the susceptibility to thermal noise .
the final exchange couplings that we have calculated for this model are well within the experimentally achievable regimes @xcite which justify the choice of above parameters .
before discussing the path - integral instanton method , it will be useful to introduce a characteristic length scale @xmath109 , to ease the notation .
following @xcite , we define this as the length scale @xmath110 at which the coulomb repulsion between electrons is comparable to the potential energy of the parabolic confinement in the @xmath111-direction for a particular strength of the @xmath111-potential @xmath112 , chosen as @xmath113 based on the energy spacing in gated quantum dots @xcite .
it is also useful to define the dimensionless parameter @xmath114 , where @xmath2 is the bohr radius , which gives @xmath115 for the chosen value of @xmath112 . for a system of @xmath3 electrons constrained to move in 2d ,
we define collective spatial and spin coordinates @xmath116 , @xmath117 , and we use @xmath118 to denote the @xmath119 element of @xmath120 , such that @xmath121 , @xmath122 . in the wigner regime
these electrons are localised near to @xmath3 lattice sites @xmath123 .
we assume that the conditions are such that the low - energy hamiltonian has the form in ( [ eqn : hthouless ] ) . then using the path integral instanton approximation , the exchange coupling corresponding to a permutation @xmath124 of the electron labels
is approximately given by @xcite @xmath125 \left ( \frac{\eta[\mathbf{r}_{cl } ] } { 2 \pi } \right)^{1/2 } e^{-\sqrt{r_\omega } \eta[\mathbf{r}_{cl}]}.\end{aligned}\ ] ] where @xmath126 is the path of least action connecting the configurations @xmath127 and @xmath128 , @xmath129 for two body exchange and 1 otherwise , and @xmath130 $ ] is the dimensionless euclidean action defined as @xmath131 : = \int_0^{t_\tau } d\tau \left [ \sum_j \frac{\dot{\mathbf{r}}_j^2}{2 } + \frac{1}{\hbar \sqrt{r_\omega } } v(\mathbf{r})\right].\end{aligned}\ ] ] @xmath132 $ ] is the result of quadratic fluctuations about the classical path , and is given as @xmath133 : = \sqrt { \frac { \det ( -\partial_{\tau}^2 { \mathbbm 1}+ \frac{1}{\hbar \sqrt{r_\omega}}\mathbf{h}^{(0)}(\tau ) ) } { \det ' ( - \partial_{\tau}^2 { \mathbbm 1}+ \frac{1}{\hbar \sqrt{r_\omega } } \mathbf{h}(\tau ) ) } } \end{aligned}\ ] ] where @xmath134 is the time - dependent hessian matrix , defined in terms of the potential @xmath135 as @xmath136 in order to numerically compute the exchange couplings @xmath137 we first find the classical " configuration that minimises the potential energy of the electrons .
the electrons are initialised evenly spaced along the @xmath94-axis with a slight zig - zag perturbation in the @xmath111-direction . the configuration that minimises the potential energy is labelled as @xmath138 .
typical minimum energy configurations in a nano - wire of length @xmath139 for @xmath18 electrons are shown for two different transverse potential @xmath140 and @xmath141 in fig . 1 ( b ) and ( c ) in the main manuscript .
for large values of @xmath6 ( i.e. strong confinements ) , the electrons are constrained to lie along the @xmath94-axis , whereas for smaller @xmath6 ( i.e. weak confinements ) a zig - zag pattern forms .
once the minimum energy classical configuration @xmath127 has been found , the classical path of least action for each permutation can be numerically found by discretising the exchange path into @xmath142 equal time steps @xmath143 , so that the action integral becomes a riemann sum : @xmath144 = \int_0^{t_\tau } d\tau \left [ \sum_j \frac{\dot{\mathbf{r}}_j^2}{2 } + \frac{1}{\hbar \sqrt{r_\omega}}v(\mathbf{r})\right ] \rightarrow \sum_{m=1}^m \sum_j \frac{1}{\delta \tau } \frac { ( \mathbf{r}_j(\tau_m ) - \mathbf{r}_j(\tau_{m-1}))^2}{2 } + \frac{\delta \tau}{\hbar \sqrt{r_{\omega } } } v(\mathbf{r}_m)\end{aligned}\ ] ] where @xmath145 and @xmath146 .
this can be minimised using a steepest descent algorithm , as this is equivalent to the expression for the energy of a chain of beads linked together by springs . for our calculations , we used @xmath147 and @xmath148 , based on the work in @xcite that shows reasonable convergence with these parameters for 27 electrons in a wigner crystal .
the accuracy of discretising time in this way has also been studied in @xcite by comparing numerical results to an analytically solvable system .
here it is found that for @xmath149 , @xmath147 the error is on the order of 1% .
once the classical path @xmath126 has been found , the prefactor @xmath150 $ ] is calculated by diagonalising the numerator and denominator in eqn .
( [ eqn : prefactor ] ) so that the determinant can be found .
this matrix is calculated along the classical path , and in discretised form becomes a @xmath151 matrix .
for example , the discretised form of the numerator in eqn .
( [ eqn : prefactor ] ) ( excluding the square root ) is a matrix @xmath152 with elements given by @xmath153
the multi - spin exchange ( mse ) hamiltonian is given by @xcite @xmath154 where @xmath155 denotes a permutation operator permuting only the spins of the particles according to the permutation @xmath124 of the electron labels .
@xmath156 is the number of 2-body swaps that @xmath124 can be decomposed into .
@xmath137 is the exchange energy , which depends on the energy splitting between the eigenstates of a permutation @xmath157 .
an illustration of the types of permutation processes that we consider in this work .
@xmath158 indicates a pairwise permutation of electrons @xmath159 and @xmath25 .
@xmath160 , @xmath161 , @xmath162 indicates 3- 4- 5- and 6-body ring permutations starting clockwise from electron @xmath25.,scaledwidth=50.0% ] in this appendix we detail the conversion of the mse hamiltonian in ( [ eqn : hthouless ] ) into a hamiltonian with pauli exchange operators . for the range of parameters considered in this study ,
permutations of more than 5 electrons and more than nearest neighbour pairwise permutations were found to be 1000 times smaller than the nearest - neighbour exchange couplings over the regime considered here , and so are left out . with this restriction ,
the hamiltonian used to simulate the evolution is @xmath163 where e.g. @xmath164 means a cycle of length 4 which permutes the spins according to @xmath165 .
we use the notation @xmath166 to mean @xmath159 and @xmath25 are up to 2@xmath167-nearest neighbours , @xmath168 means that @xmath169 are on the vertices of a triangle , @xmath170 means that @xmath171 are on the vertices of a parallelogram and @xmath172 means that @xmath173 are on the vertices of trapezium ( see fig . [
fig : wigexch ] ) . the spin permutation operators can be written in terms of pairwise spin exchange interactions , using the following identities @xcite : @xmath174 by applying the above identities to hamiltonian ( [ eqn : permham ] ) , this gives the overall hamiltonian , ignoring the factors of identity since they do not affect the dynamics : @xmath175 \nonumber\\ & -\frac{1}{8 } \sum _ { \substack { j , k , l , m , n \in \pentagon \\ j < k < n < l < m } } j_{j}^{(r5 ) } \left [ \upsilon_{jklm } + \upsilon_{jkln } + \upsilon_{jnkm } + \nonumber\right .
\upsilon_{jnlm } + \upsilon_{knlm } + \bm{\sigma}_j \cdot \bm{\sigma}_k + \bm{\sigma}_k \cdot \bm{\sigma}_l + \bm{\sigma}_l \cdot \bm{\sigma}_m \right .
\nonumber\\ & \left .
+ \bm{\sigma}_m \cdot \bm{\sigma}_n + \bm{\sigma}_j \cdot \bm{\sigma}_n + \bm{\sigma}_j \cdot \bm{\sigma}_l + \bm{\sigma}_k \cdot \bm{\sigma}_m \right .
+ \bm{\sigma}_l \cdot \bm{\sigma}_n + \bm{\sigma}_k \cdot \bm{\sigma}_n + \bm{\sigma}_j \cdot \bm{\sigma}_m \right ] \ ] ] where @xmath26 .
note the minus sign in front of the @xmath176 and @xmath177 terms since they involve permutations of an odd number of particles .
overall then , this hamiltonian can be written @xmath178 where @xmath179
the two - point steepest descent algorithm proceeds as follows , with the positions of all of the particles at the @xmath180 iteration labelled as @xmath181 @xcite : * make an initial guess @xmath182 . * for each step , @xmath183 , where @xmath184 $ ] , @xmath185 is a small step size ( we use @xmath186 ) and @xmath187 for @xmath188 @xcite . *
stop if @xmath189 ( convergence to a minimum ) , or if @xmath190 ( convergence to a point ) .
here @xmath191 indicates the euclidean norm , and the tolerance used is @xmath192 ( although in this study the algorithm always converged to a minimum rather than a point ) .
a plot showing the convergence of @xmath193 is shown in fig .
[ fig : convergence ] , for all of the exchange processes starting at electron 1 for @xmath194 .
the plot shows the change in action relative to the final value @xmath195 , given as @xmath196 .
clearly the 3 , 4 and 5 body processes take much longer to converge than 2-body processes , which is intuitive given the electrons tend to move less for the 2-body processes .
the small spikes in the plot can be explained by the algorithm occasionally overshooting the local minimum .
as the barrier is dropped to couple the quantum dot to the nanowire , there may be phonon oscillations induced by the slight change in equilibrium positions .
one effect of this would be to cause the exchange couplings to oscillate . to estimate the effects of this on information transfer , we calculated the average change in the nearest - neighbour exchange coupling for different barrier heights .
we found that for a change in internal barrier height from @xmath197mev to @xmath198ev used in the study , the exchange couplings changed by less than @xmath199 .
additionally , we calculated the average classical vibration energy @xmath200 for the @xmath201th electron where @xmath202 these vibration frequencies are typically on the order of 1 - 5@xmath203mev for the regimes covered in this paper , roughly 100 times larger than the exchange coupling .
we can approximately model the effects of these phononic vibrations as adding small random oscillations in the exchange couplings according to @xmath204\ ] ] where @xmath38 is a random phase and @xmath205 is a random amplitude chosen from a normal distribution with mean 0 and standard deviation @xmath206 .
our numerical simulations show that such fast fluctuation cancels out in time evolution and its impact on the fidelity is negligible .
this makes the proposed protocol resistive against phonon excitations created through changing the barrier height @xmath12 . | technological applications of many - body structures that emerge in gated devices under minimal control are largely unexplored . here
we show how emergent wigner crystals in a semiconductor quantum wire can facilitate a pivotal requirement for a scalable quantum computer , namely transmitting quantum information encoded in spins faithfully over a distance of micrometers .
the fidelity of the transmission is remarkably high , faster than the relevant decohering effects , independent of the details of the spatial charge configuration in the wire , and realizable in dilution refrigerator temperatures .
the transfer can evidence near unitary many - body nonequilibrium dynamics hitherto unseen in a solid - state device .
it could also be useful in spintronics as a method for pure spin current over a distance without charge movement .
spin chains can facilitate several important technological applications such as pure spin currents in non - itinerant systems @xcite , quantum state transfer @xcite and quantum gates @xcite .
most of the above applications are facilitated by a nearly unitary dynamics of the spin chain .
such dynamics is not only interesting for potential quantum technology @xcite but also fundamentally important to address questions of equilibration , quantum thermodynamics and information propagation @xcite .
thus , the physical realization of artificial spin chains that have the potential for long time unitary dynamics is an important quest
they have been realized only very recently , and exclusively in atomic physics systems : in cold atom systems @xcite , ion traps @xcite and rydberg systems @xcite . in the realm of solid - state , on the other hand , nonequilibrium dynamics of engineered spin chains is far from unitary and primarily driven by equilibration / relaxation @xcite . although some bulk magnetic materials @xcite , nv centre chains @xcite and josephson junction arrays @xcite hold the potential for long chain unitary dynamics , that is still somewhat distant from experimental realization .
as far as the semiconductor realm is concerned , permanently fabricated spin chain structures with dangling bonds @xcite or phosphorus dopants @xcite may also hold the potential , but is yet to be examined either theoretically or experimentally .
a natural question is thus whether one can realize spin chains exhibiting nearly unitary nonequilibrium dynamics in a feasible manner in a two dimensional electron gas ( 2deg ) and thereby open the door to the aforementioned applications in solid state .
individual electrons trapped in gate defined quantum dot arrays have been proposed for simulating spin chains in the limit of one electron in each dot @xcite .
although the fabrication of large quantum dot arrays is an active current effort @xcite and the complex electronics for gate addressing is also being designed @xcite , it is worth considering the potential of simpler gated structures .
in particular , self assembled charge configurations , such as wigner crystals , are naturally formed without demanding local control . by controlling the density of electrons one can vary the distance between the charges and consequently engineer their exchange interactions .
this , in comparison with quantum dot arrays , allows for stronger exchange couplings which then potentially provide more thermal stability , faster dynamics and less sensitivity to decoherence . here
we show that emergent self assembled " electronic spin chains arising due to wigner crystallization in quasi-1d nano - wires can be probed with two spatially separated accessible interfaces ( two quantum dots ) so that nonequilibrium dynamics and its applications can be probed .
particularly we show how this setting can be used to transfer spin qubits between two quantum dots separated by @xmath0 scales , which is currently being actively considered as an important problem , with very few suggested solutions @xcite . additionally we suggest a feasible way of observing this phenomenon through pa scale currents , which , in turn , opens up a new option in low dissipation spintronics for a spin current without a charge current .
_ wigner crystalization. _ by applying strong confining potentials on a 2deg one may trap a few electrons in a quasi-1d region and effectively make a nano - wire . in such nano - wires , when the electron density is below a critical value , the coulomb interactions between electrons overtakes their kinetic energies resulting in a quasi-1d wigner crystal in which the electrons are extremely localised near to the classical equilibrium configurations . in quasi-1d nano - wires where the electrons are strongly confined in two directions , a wigner crystal is predicted to emerge when the average electron - electron distance is greater than @xmath1 , where @xmath2 is the bohr radius @xcite . _
model. _ we consider trapping @xmath3 electrons ( with @xmath3 even ) in a quasi-1d region by using surface electrodes over a 2deg in gaas .
the trapping potential is modelled as @xmath4 where @xmath5 is the electron effective mass , @xmath6 is the strength of the transverse potential and the two gaussian potentials with height @xmath7 and width @xmath8 define a nano - wire of length @xmath9 extending between @xmath10 .
two small quantum dots , for trapping single electrons , are formed at both ends of the wire by applying proper voltages to the gates .
the quantum dots are modelled by the following potential @xmath11 where the first two gaussian potentials with heights @xmath12 and width @xmath13 , centered at the distance @xmath14 from the boundaries of the wire , create two quantum dots at both sides of the nano - wire and the potentials in the second line break the mirror symmetry by displacing the minima in the two quantum dots in opposite directions .
the symmetry broken system will have a single equilibrium position for electrons and thus numerical convergence is easier to reach . choosing appropriate values of the parameters @xmath15 , @xmath16 and @xmath17
is discussed in the supplemental material .
we consider @xmath18 electrons trapped over a distance @xmath19 m , such that the quantum dots confine one electron each and the remaining 8 electrons are placed in the wire as shown in fig .
[ fig : fig1](a ) .
the whole potential is @xmath20\cr \nonumber\end{aligned}\ ] ] the barrier @xmath12 is varied from 3mev to @xmath21ev to decouple or couple the quantum dots to the wire , respectively .
thus , the two endmost electrons can act as sender and receiver when sending quantum information through the chain . unlike the middle electrons , we assume full control over these two end electrons for initialization and measurement .
_ exchange couplings. _ at low temperatures the dynamics of the wigner crystal is approximately governed by the multi - spin - exchange ( mse ) hamiltonian @xcite . to calculate the exchange couplings in this hamiltonian we exploit the semi - classical path integral instanton method @xcite .
this approach is applicable for the regimes where the quantum effects are small , like the case for wigner crystals where the electrons are well separated @xcite ( see the supplemental material for more details ) . while our methodology assumes a closed system it is potentially amenable to extension to directly incorporate an environment following the methodology of ref .
@xcite . using the instanton method ,
it was found that only processes involving up to @xmath22-nearest neighbour pairwise exchange and up to 4-body exchange were significant . with this restriction ,
the full spin hamiltonian can be written ( see supplemental material for the details ) @xmath23 where @xmath24 is the pauli vector acting on site @xmath25 and @xmath26 .
exchange couplings for @xmath27 and @xmath28 are given in units of @xmath29 in the tables next to each charge configurations of figs .
[ fig : fig1 ] ( b ) and ( c ) .
the general features are low couplings at the boundary ( due to the barrier between the dots and the chain ) and a u - shaped coupling along the chain .
the next - nearest neighbour couplings are always ferromagnetic ( i.e. @xmath30 ) for the linear configuration while for the zig - zag geometry they show a more complex pattern varying from negative to positive values .
the u - shaped feature of the couplings is because the off - center electrons are pushed towards the boundaries due to an unbalanced coulomb repulsion from the majority of electrons on the opposite side .
this makes the effective distance between the electrons shorter in the boundaries and thus results in stronger couplings . _
quantum communication. _
we assume that the quantum dots are initially decoupled from the wire ( i.e. @xmath12 is large ) .
furthermore , we consider zero temperature so that the electrons are confined to their lowest vibrational mode and are prepared in their spin ground state @xmath31 .
the electron in one quantum dot is prepared in an arbitrary quantum spin state @xmath32 , which is supposed to be transferred to the opposite dot , which confines an electron in unpolarized mixed state @xmath33 .
the initial state of the system is thus @xmath34 .
the barriers are then simultaneously lowered ( i.e. @xmath12 is lowered ) in order to start a unitary dynamics under the action of the new hamiltonian .
one can compute the density matrix of the last electron @xmath35 by tracing out the others . to quantify the quality of transfer
one can compute the fidelity @xmath36 which is independent of @xmath37 and @xmath38 due to the su(2 ) symmetry of the hamiltonian . in figs .
[ fig2](a ) and ( b ) we plot @xmath39 as a function of time for the two charge configurations with @xmath40 ( linear chain ) and @xmath28 ( zig - zag chain ) .
it is clear from these figures that the fidelity peaks at a time @xmath41 and the takes its maximum value @xmath42 .
although the linear chain gives a faster dynamics with @xmath43ns ( due to fairly larger couplings ) in comparison with the zig - zag charge configuration with @xmath44ns , the maximum fidelity @xmath45 is remarkably high for both configurations , certainly larger than a uniform chain @xcite .
these results illustrate the key point that the details of the charge configurations are not important for the quality of spin transport . to see the scalability , we plot @xmath45 as a function of @xmath3 in fig .
[ fig2](c ) , keeping the density of electrons fixed and using only nearest - neighbour couplings since these are by far the largest and computing higher order interactions becomes computationally prohibitive for @xmath46 electrons .
as the figures shows the fidelity @xmath45 remains very high even for chains up to @xmath47 electrons . as a function of time , when @xmath12 suddenly changes from 3mev to @xmath21ev , using both real and effective hamiltonians for : ( a ) the linear chain ( @xmath40 ) ; and ( b ) the zig - zag chain ( @xmath28 ) . ( c )
@xmath45 versus @xmath3 .
( d ) the variation in @xmath45 when a random magnetic field of variance @xmath48 is included in the hamiltonian . , scaledwidth=50.0% ] _ explanation. _ to understand the remarkably high fidelities achieved through the time evolution of the wigner crystal , one has to consider the nearest - neighbor couplings , which dominate the hamiltonian @xmath49 . as shown in the tables in fig .
[ fig : fig1 ] ( b ) and ( c ) , the couplings @xmath50 ( which are identical to @xmath51 ) are almost four times stronger than the couplings @xmath52 ( which are identical to @xmath53 ) in both charge configurations .
this implies that the ground state and first excited state of the hamiltonian @xmath49 show delocalized strong correlations between the boundary electrons . by computing the reduced density matrix of the two ending spins @xmath54 from the ground state of @xmath49
one can see that @xmath55 , where @xmath56 is the singlet state , for both charge configurations .
similarly , by computing @xmath54 from the first excited state of @xmath49 one gets @xmath57 , where @xmath58 is the triplet state .
the delocalized eigenvectors of the hamiltonian create an effective rkky - like interaction between the two boundary electrons in the dots @xcite , namely @xmath59 where @xmath60 in which @xmath61 stands for the energy gap of the hamiltonian @xmath49 . in figs .
[ fig2](a)-(b ) the time evolution of average fidelities using both @xmath62 and @xmath49 are plotted which shows that the effective hamiltonian can qualitatively explain the real dynamics of the system .
in fact , the effective model becomes more precise by decreasing the boundary couplings @xmath63 and @xmath53 . _
imperfections. _ so far , we have assumed that the system operates at zero temperature and thus the electrons in the wire are initialized in their ground state .
in order to guarantee that the proposed protocol remains valid at finite temperature @xmath64 , one has to satisfy @xmath65 , where @xmath66 is the boltzmann constant . for the given set of couplings the energy gaps
are @xmath67 @xmath68ev for the linear and @xmath69 @xmath68ev for the zig - zag configurations giving the range of temperature as @xmath70 mk which can be achieved in current dilution refrigerators @xcite . in gaas hetero - structures the electron spins
interact with the nuclear spins of the host material . due to the very slow dynamics of nuclei spins in comparison to the time scales of our protocol
one can describe their average effect on electron spin @xmath25 as an effective random magnetic field @xmath71 . while the direction of this field is fully random
its amplitude has a gaussian distribution @xcite @xmath72 / ( 2\pi b_{nuc}^{2})^{3/2},\ ] ] in which @xmath73 is the variance of the distribution .
the total hamiltonian thus changes as @xmath74 in fig .
[ fig2](d ) we plot the maximum fidelity @xmath45 versus @xmath48 which shows the destructive effect of the hyperfine interaction .
the linear configuration performs better for larger values of @xmath48 since the faster dynamics reduces the time exposed to nuclear spins .
a realistic value for @xmath48 is @xmath75 mt ( i.e. @xmath76 ) @xcite , at which the fidelity of @xmath77 is attainable .
using spin - orbit coupling one may effectively suppress @xmath48 to @xmath78 nev @xcite . at experimental temperatures ( @xmath79 ) thermal phonons are absent in the material @xcite .
moreover , the bulk phonon wavelengths ( @xmath80 ) exceed electron wavelengths ( @xmath81 nm ) so much that the electrons do not couple to them either .
however , sudden quench in @xmath12 may cause the electrons to jiggle around their positions causing fluctuations in exchange interactions .
as shown in the supplemental material , these high frequency vibrations ( @xmath82 ) are integrated out during the transmission time @xmath83 . _
experimental realization. _ the quantum dot - quantum wire - quantum dot system can be realized in a iii - v material such as gaas / algaas heterostructure @xcite , using gates and quantum point contacts technique , as shown in fig .
[ fig3 ] .
we assume that the quantum wire has achieved a state of wigner crystallisation @xcite .
spin initialization in the left dot and read - out in the right dot after the transfer time , can be performed using demonstrated techniques @xcite such as singlet - triplet charge measurements by appending double - dots to both ends of the wire . however , here we show how transport measurements can also probe our mechanism by effectively detecting the transfer of a spin current .
the prototype of this measurement consists of a left electron pump @xcite injecting polarised electrons ( say , from a standard source @xcite ) whereas the right pump injecting unpolarised electrons in the respective quantum dots as shown in fig .
[ fig3 ] . as the electrons are injected into the dots they interact with the electrons in the quantum wire as the barrier height
is reduced and their quantum state is swapped .
thus upon exiting the dots , the initially polarized electrons will be unpolarized and vice - versa .
as the interaction time for spin swap is about @xmath84ns ( @xmath85ns ) for the linear ( zig - zag ) configuration the electrons must stay in the dot for such time .
this implies that the pumps have to inject current of @xmath86 depending on the configuration of electrons in the quantum wire . a spin polarization of the current carried off from the right dot by the right pump
could then be detected by the halving of the conductance through a quantum point contact in series with the right dot .
_ discussion and conclusions. _ we have shown that entirely within the remit of gate defined structures , one can use a wigner crystal in a quantum wire for transferring spin qubits between quantum dots separated by micrometers reaching very high fidelities .
compared to varied proposals for mediating between separated quantum dot spin qubits @xcite , the proposed mechanism can have various advantages .
for example , it could have higher speed making it more resilient to decoherence , simpler fabrication compared to hybrid systems , and a greater spatial extension over few electron quantum dot mediators .
furthermore , thanks to the self - assembled nature of the wigner crystal , our scheme does not require complex electronics to artificially create a regular structure of exchange coupled spins .
our analysis shows the proposed protocol can operate in current dilution refrigerators , and remarkably the details of the charge configuration does not influence the performance of the system , adding additional robustness to the device fabrication .
it also opens an avenue in spintronics by facilitating the spatial transport of a spin current without a charge current , and provides a smoking gun for the near unitary nonequilibrium dynamics in many - body solid - state systems .
alternative realizations of wigner crystals , in liquid helium @xcite , ion traps @xcite and carbon nanotubes @xcite can also be used for implementing our mechanism . _
acknowledgements. _ the authors thank martin uhrin for illuminating discussions on use of the 2-point steepest descent algorithm .
this work was supported by the engineering and physical sciences research council ( epsrc ) , uk .
the research leading to these results has received funding from the european research council under the european union s seventh framework programme ( fp/2007 - 2013 ) / erc grant agreement n. 308253 .
54ifxundefined [ 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 _ _ ( , ) link:\doibase 10.1103/physrevlett.106.040505 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ ( , ) pp .
f. a. zwanenburg , _
et al .
_ , rev .
mod .
phys . * 85 * , 961 ( 2013 ) . @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) j. m. nichol , _ et al . _ , nat
. comm .
* 6 * , 7682 ( 2015 ) .
p. v. e. mcclintock , _
d. j. meredith , j. k. wigmore , low - temperature physics _ ( springer , 1992 ) . @noop * * , ( ) link:\doibase 10.1103/physrevlett.98.126404 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevb.90.201304 [ * * , ( ) ] @noop * * , ( ) p. m. platzman and m. i. dykman , science * 284 * , 1967 ( 1999 ) .
j. baltrusch , a. negretti , j. taylor , and t. calarco , phys .
rev a * 83 * , 042319 ( 2011 ) . v. v. deshpande and m. bockrath , nat .
phys . * 4 * 314 ( 2008 ) .
ifxundefined [ 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 * * , ( ) * supplemental materials * |
as exotic natures of unstable nuclei such as the new magic numbers and the neutron halos are disclosed by experiments , microscopic studies based on the nucleon - nucleon ( @xmath4 ) interaction become even more desired in nuclear structure physics . while the fully microscopic @xmath4 ( and @xmath5 ) interaction is still too complicated to cover large volume of nuclei in the periodic table despite significant progress @xcite , the semi - realistic @xmath4 interactions have been developed @xcite by modifying the michigan 3-range yukawa ( m3y ) interaction @xcite , which was originated from brueckner s @xmath6-matrix at the nuclear surface and expressed by the yukawa functions .
the modification has been made so that the saturation and the spin - orbit ( @xmath7 ) splitting should be reproduced within the mean - field approximation ( mfa ) .
owing to the recently developed numerical methods @xcite , self - consistent calculations in the mfa @xcite and in the random - phase approximation ( rpa ) @xcite have been implemented using the semi - realistic interactions . among the advantages of the m3y - type semi - realistic interactions
is that they contain realistic tensor channels as well as correct longest - range central channels originating from the one - pion exchange , which have been pointed out to play significant roles in @xmath8- or @xmath9-dependence of the shell structure @xcite .
the semi - realistic interactions are suitable to investigate effects of these channels within the self - consistent mfa and rpa @xcite . while the parameter - sets of the m3y - type interactions in refs .
@xcite were adjusted to the data on the nuclear structure , some of them have been applied to the nuclear reactions @xcite and to the neutron stars @xcite as well . in studying structure of the neutron stars ,
density - dependence of the symmetry energy is crucially important @xcite .
it has been pointed out that the symmetry energy at low density is significant in nuclear reactions : _
e.g. _ the charge - exchange reactions @xcite and the multi - fragmentation processes @xcite
. the symmetry energy at low density may also affect the so - called pygmy dipole resonance in neutron - rich nuclei @xcite . however , the symmetry energy , particularly its density - dependence , was not sufficiently reliable in the previous parameter - sets in refs .
@xcite as pointed out in ref .
@xcite , giving rise to instability of the symmetric nuclear matter at the density @xmath10 , which is not consistent with microscopic calculations @xcite . in this article
, we shall propose new parameter - sets of the m3y - type semi - realistic @xmath4 interaction .
as far as the energy of the symmetric nuclear matter is fixed , the symmetry energy at each density is well connected to the energy of the neutron matter .
the new parameters are determined by fitting the neutron - matter energy to microscopic result in ref .
@xcite ( fp ) or @xcite ( apr ) .
moreover , we additionally take into consideration the binding energy of @xmath0sn .
as argued later , the symmetry energy at the saturation point tends to be fixed with good precision by fitting the parameters both to @xmath0sn and @xmath11sn .
the symmetry energy is thus constrained to good degree in the new parameter - sets .
corresponding to the microscopic results on the neutron - matter energy , we obtain two parameter - sets ` m3y - p6 ' ( fitted to fp ) and ` m3y - p7 ' ( to apr ) .
although the parameters are determined from a limited number of data , they will be useful for investigating various aspects of nuclear properties , as will be illustrated by separation energies of the proton- and neutron - magic nuclei and by @xmath8- or @xmath9-dependence of shell structure .
we take a non - relativistic isoscalar nuclear hamiltonian of @xmath12 with @xmath13 and @xmath14 representing the indices of individual nucleons .
we set @xmath15 throughout this paper , where @xmath16 ( @xmath17 ) is the mass of a proton ( a neutron ) @xcite . for the effective @xmath4 interaction @xmath18 , the following form is considered , @xmath19^{\alpha^{(\mathrm{se } ) } } + t_\rho^{(\mathrm{te } ) } p_\mathrm{te}\cdot [ \rho(\mathbf{r}_i)]^{\alpha^{(\mathrm{te})}}\big ) \,\delta(\mathbf{r}_{ij})\ , , \label{eq : effint}\end{aligned}\ ] ] where @xmath20 is the spin operator of the @xmath13-th nucleon , @xmath21 , @xmath22 , @xmath23 , @xmath24 , and @xmath25 denotes the nucleon density .
the tensor operator is defined by @xmath26 $ ] with @xmath27 .
the projection operators on the singlet - even ( se ) , triplet - even ( te ) , singlet - odd ( so ) and triplet - odd ( to ) two - particle states are @xmath28 where @xmath29 ( @xmath30 ) expresses the spin ( isospin ) exchange operator .
the yukawa function @xmath31 is assumed for all channels except @xmath32 .
the density - dependent contact term @xmath32 is added in order to reproduce the saturation properties .
physically , @xmath32 may carry effects of the @xmath5 interaction and of the density - dependence that is dropped in the original m3y interaction .
we start from the m3y - paris interaction @xcite , which will be denoted by m3y - p0 in this article as in ref .
the range parameters @xmath33 of m3y - p0 are maintained in any of @xmath34 , @xmath35 and @xmath36 . as in m3y - p0 ,
the longest - range part in @xmath34 is kept identical to the central channels of the one - pion exchange potential ( opep ) , @xmath37 .
although the @xmath7 splitting plays a significant role in the nuclear shell structure , the @xmath6-matrix is known to underestimate the @xmath7 splitting .
even though effects beyond the mfa may cure this problem @xcite , we introduce an overall enhancement factor to @xmath35 in order to describe the shell structure within the mfa .
effects of the tensor force on the single - particle ( s.p . )
levels could be relevant to the new magic numbers in unstable nuclei @xcite .
we keep @xmath36 without any modification from m3y - p0 . because of this @xmath36 having realistic nature , the present m3y - type interactions are useful to investigate the tensor - force effects within the mfa and rpa , as shown in refs .
@xcite with the previous parameter - sets .
the parameters in m3y - p6 and p7 are tabulated in table [ tab : param_m3y ] , together with m3y - p0 .
.parameters of m3y - type interactions .
[ tab : param_m3y ] [ cols="^,^,>,<,>,<,>,<,>,<",options="header " , ]
we have developed new parameter - sets of the semi - realistic effective interactions to describe low energy phenomena of nuclei .
they are obtained by phenomenologically modifying several parameters in the m3y - paris interaction , while the tensor force and the opep part in the central force are not changed , as before . unlike the previous parameters ,
the new sets m3y - p6 and p7 are adjusted also to the microscopic ( fp and apr ) results of the neutron - matter energies and to the binding energy of @xmath0sn .
we therefore attain improvement on the symmetry energy , up to its density - dependence .
in contrast to instability of the spin - saturated symmetric nuclear matter in the sly5 , d1s and d1 m results , neither of m3y - p6 nor p7 predicts such phase transition in the density range of @xmath3 .
the new parameter - sets m3y - p6 and p7 have been applied to the doubly magic nuclei in the spherical hf calculations , and to the proton- or neutron - magic nuclei in the spherical hfb calculations .
fair agreement with experimental data has been demonstrated for the binding energies of the doubly magic nuclei and for the nucleon separation energies of the proton- or neutron - magic nuclei .
owing to the realistic tensor force and the opep central force , the @xmath8- or @xmath9-dependence of the shell structure is well described with m3y - p6 and p7 , as with the previous set m3y - p5 .
the isotope shifts of the pb nuclei have also been argued .
future study includes application of the semi - realistic interactions to the excitations in the rpa , as well as to deformed nuclei .
moreover , extensive applications to nuclear reactions and to the neutron stars may be within the reach , which give further test of the effective interactions and a step toward unified description of nuclear structure , reactions and neutron stars .
this work is financially supported as grant - in - aid for scientific research ( c ) , no . 22540266 , by japan society for the promotion of science , and as grant - in - aid for scientific research on innovative areas , no .
24105008 , by the ministry of education , culture , sports , science and technology , japan .
numerical calculations are performed on hitac sr16000s at institute of media and information technology in chiba university , yukawa institute for theoretical physics in kyoto university , research institute for information technology in kyushu university , information technology center in university of tokyo , and information initiative center in hokkaido university .
pieper , k. varga and r.b .
wiringa , phys .
c * 66 * , 044310 ( 2002 ) .
p. navratl and b.r .
barrett , phys .
c * 57 * , 3119 ( 1998 ) ; p. navratl , j.p .
vary and b.r .
barrett , phys .
c * 62 * , 054311 ( 2000 ) ; r. roth , j. langhammer , a. calci , s. binder and p. navratl , phys .
lett . * 107 * , 072501 ( 2011 ) .
g. hagen , t. papenbrock , d.j . dean and m. hjorth - jensen , phys .
* 101 * , 092502 ( 2008 ) ; g. hagen , m. hjorth - jensen , g.r .
jansen , r. machleidt and t. papenbrock , phys .
lett . * 109 * , 032502 ( 2012 ) .
h. nakada , phys . rev .
c * 68 * , 014316 ( 2003 ) .
h. nakada , phys .
c * 78 * , 054301 ( 2008 ) ; _ ibid .
_ * 82 * , 029902(e ) ( 2010 ) .
g. bertsch , j. borysowicz , h. mcmanus and w.g .
love , nucl .
a * 284 * , 399 ( 1977 ) .
n. anantaraman , h. toki and g.f .
bertsch , nucl .
a * 398 * , 269 ( 1983 ) .
h. nakada and m. sato , nucl .
a * 699 * , 511 ( 2002 ) ; _ ibid . _
* 714 * , 696 ( 2003 ) .
h. nakada , nucl .
phys . a * 764 * , 117 ( 2006 ) ; _ ibid . _
* 801 * , 169 ( 2008 ) .
h. nakada , nucl .
a * 808 * , 47 ( 2008 ) .
h. nakada , k. mizuyama , m. yamagami and m. matsuo , nucl .
a * 828 * , 283 ( 2009 ) .
h. nakada , _ proceedings of the international symposium a new era of nuclear structure physics _ , edited by y. suzuki , m. matsuo , s. ohya and t. ohtsubo , p. 184
( world scientific , singapore , 2004 ) .
h. nakada , eur .
j. a * 42 * , 565 ( 2009 ) .
h. nakada , phys .
c * 81 * , 051302(r ) ( 2010 ) .
t. shizuma _
et al . _ ,
c * 78 * , 061303(r ) ( 2008 ) . t. otsuka _
lett . * 87 * , 082502 ( 2001 ) .
t. otsuka , t. suzuki , r. fujimoto , h. grawe and y. akaishi , phys .
lett . * 95 * , 232502 ( 2005 ) .
h. nakada , phys .
c * 81 * , 027301 ( 2010 ) ; _ ibid .
_ * 82 * , 029903(e ) ( 2010 ) .
khoa , private communication .
d.t . loan , n.h .
tan , d.t .
khoa and j. margueron , phys .
c * 83 * , 065809 ( 2011 ) .
lattimer and m. prakash , phys . rep . *
442 * , 109 ( 2007 ) .
a. ono , p. danielewicz , w.a .
friedman , w.g .
lynch and m.b .
tsang , phys .
c * 68 * , 051601(r ) ( 2003 ) .
a. klimkiewicz _
et al . _ ,
c * 76 * , 051603(r ) ( 2007 ) ; a. carbone _ et al .
c * 81 * , 041301(r ) ( 2010 ) .
than , d.t .
khoa and n.v .
giai , phys .
c * 80 * , 064312 ( 2009 ) .
a. akmal , v.r .
pandharipande and d.g .
ravenhall , phys .
c * 58 * , 1804 ( 1998 ) . b. friedman and v.r .
pandharipande , nucl .
a * 361 * , 502 ( 1981 ) .
particledatagroup , j. phys .
g * 33 * , 1 ( 2006 ) .
k. suzuki , r. okamoto and h. kumagai , phys .
c * 36 * , 804 ( 1987 ) ; s.c .
pieper and v.r .
pandharipande , phys .
lett . * 70 * , 2541 ( 1993 ) .
e. chabanat , p. bonche , p. haensel , j. meyer and r. schaeffer , nucl .
a * 635 * , 231 ( 1998 ) .
berger , m. girod and d. gogny , comp .
comm . * 63 * , 365 ( 1991 ) .
s. goriely , s. hilaire , m. girod and s. pru , phys .
lett . * 102 * , 242501 ( 2009 ) .
wiringa , v. fiks and a. fabrocini , phys .
c * 38 * , 1010 ( 1988 ) .
s. shlomo , m. kolomietz and g. col , eur .
j. a * 30 * , 23 ( 2006 ) .
p. danielewicz , nucl .
a * 727 * , 233 ( 2003 ) .
m. kortelainen _
et al . _ ,
c * 85 * , 024304 ( 2012 ) . c. mahaux , p.f .
bortignon , r.a .
broglia and c.h .
dasso , phys . rep . *
120 * , 1 ( 1985 ) .
cao , g. col and h. sagawa , phys .
c * 81 * , 044302 ( 2010 ) . c. gaarde
_ et al . _ , nucl .
phys . a * 369 * , 258 ( 1981 ) ; t. suzuki , nucl .
a * 379 * , 110 ( 1982 ) ; g. bertsch , d. cha and h. toki , phys . rev .
c * 24 * , 533 ( 1981 ) ; t. suzuki and h. sakai , phys .
b * 455 * , 25 ( 1999 ) .
m. anguiano , j.l .
egido and l.m .
robledo , nucl . phys .
a * 683 * , 227 ( 2001 ) ; t. lesinski , t. duguet , k. bennaceur and j. meyer , eur .
j. a * 40 * , 121 ( 2009 ) ; h. nakada and m. yamagami , phys .
c * 83 * , 031302(r ) ( 2011 ) .
et al . _ ,
* 99 * , 032502 ( 2007 ) .
g. audi and a.h .
wapstra , nucl .
a * 595 * , 409 ( 1995 ) ; g. audi , a.h .
wapstra and c. thibault , nucl .
phys . a * 729 * , 337 ( 2003 ) .
khoa , h.s . than and
m. grasso , nucl .
a * 722 * , 92c ( 2003 ) .
phys . a * 691 * ( 2001 ) 599 .
alkhazov , s.l .
belostotsky and a.a .
vorobyov , phys . rep . * 42 * , 89 ( 1978 ) .
_ , _ table of isotopes _ , 8th edition ( john wiley & sons , new york , 1996 ) .
m. centelles , x. roca - maza , x. vias and m. warda , phys .
lett . * 102 * , 122502 ( 2009 ) .
et al . _ ,
* 107 * , 062502 ( 2011 ) .
s. perez - martin and l.m .
robledo , phys .
c * 78 * , 014304 ( 2008 ) .
n. schunck _
et al . _ ,
c * 81 * , 024316 ( 2010 ) .
although the set m3y - p5@xmath38 was claimed to be fitted to the even - odd mass difference in the @xmath1 nuclei by the fully self - consistent hfb calculations @xcite , inconsistency had remained in evaluation of contribution of @xmath32 to pairing energy .
sharma , g. lalazissis , j. knig and p. ring , phys .
lett . * 74 * , 3744 ( 1995 ) .
p. aufmuth , k. heilig and a. steudel , at .
data nucl .
data tables * 37 * , 455 ( 1987 ) .
i. angeli , at .
data nucl .
data tables * 87 * , 185 ( 2004 ) .
reinhard and h. flocard , nucl .
a * 584 * , 467 ( 1995 ) .
r. kanungo , i. tanihata and a. ozawa , phys .
b * 528 * ( 2002 ) 58 .
et al . _ ,
b * 510 * , 17 ( 2001 ) .
m. honma , t. otsuka .
brown and t. mizusaki , eur .
j. a * 25 * , s01 , 499 ( 2005 ) .
et al . _ ,
74 * , 868 ( 1995 ) .
t. aumann , prog .
nucl . phys . *
59 * , 3 ( 2007 ) ; s. gales , prog .
* 59 * , 22 ( 2007 ) ; t. motobayashi , prog . part .
* 59 * , 32 ( 2007 ) . | new parameter - sets of the semi - realistic nucleon - nucleon interaction are developed , by modifying the m3y interaction but maintaining the tensor channels and the longest - range central channels .
the modification is made so as to reproduce microscopic results of neutron - matter energies , in addition to the measured binding energies of doubly magic nuclei including @xmath0sn and the even - odd mass differences of the @xmath1 and @xmath2 nuclei in the self - consistent mean - field calculations .
separation energies of the proton- or neutron - magic nuclei are shown to be in fair agreement with the experimental data . with the new parameter - sets m3y - p6 and p7 , the isotropic spin - saturated symmetric nuclear matter remains stable in the density range as wide as @xmath3 , while keeping desirable results of the previous parameter - set on finite nuclei .
isotope shifts of the pb nuclei and tensor - force effects on shell structure are discussed . |
the interplay between disorder and interactions in bose systems has been a vital field of research in condensed matter both in theory and experiment .
the line of investigation was mainly initiated by the seminal work of fisher et al.@xcite , who first provided a detailed study of localization of interacting bosons in a random potential , which led to the notion of the superfluid - insulator transition and the bose glass ( bg ) .
while disorder effects in fermi systems are relevant to a broad range of experimentally accessible scenarios like correlated electron systems , the status was less diverse for bose systems for a considerable time period , as superfluid @xmath0he situated in random pores of vycor had been the predominant setup which could be studies with sufficient precision @xcite .
this changed dramatically when the realization of the superfluid - mott insulator transition of ultracold bosonic atoms in an optical lattice opened up a new field of investigation @xcite .
in particular , optical lattices provide a relatively pure and tunable simulation of effective models used to describe solid state systems@xcite , where effects like disorder can also be realized in a controlled manner .
while several alternative realizations of disorder in optical lattices , such as multichromatic lattices with non - commensurate wavelengths@xcite or multi - species gases with strongly differing tunneling rates @xcite have been proposed recently , speckle laser patterns are probably by now one of the most efficient methods to establish disorder in cold atoms @xcite .
therein , it is possible to combine the speckle beam with the remaining apparatus of the optical lattice to simulate disordered lattice systems with a high tuning accuracy and without other side effects .
a variety of theoretical approaches @xcite has by now been applied to the disordered bose hubbard model ( bhm ) , first introduced for ultracold atoms by jaksch et al .
@xcite , which is described by the hamiltonian @xmath1 where @xmath2 ( @xmath3 ) annihilates ( creates ) a particle in the lowest band wannier state at site @xmath4 , @xmath5 is the local particle number operator , @xmath6 denotes the nearest neighbor hopping energy in the lowest band , and @xmath7 is the chemical potential .
@xmath8 is an on - site energy shift , which in our case is a spatially uncorrelated random variable drawn from a distribution @xmath9 and @xmath10 is the on - site repulsive interaction .
the subscript @xmath11 indicates the sum over all neighboring pairs of sites . unless stated otherwise , we use the unit @xmath12 .
several quantum phases can exist within this model , such as the mott insulator ( mi ) , the bose glass , the condensed phase , commonly referred to as the superfluid ( sf ) , as well as the normal phase at finite temperature .
the transitions between these phases , which constitute some of the first experimentally feasible quantum phase transitions in bosonic systems , have attracted much attention .
numerically , a powerful approach is quantum monte carlo @xcite ( qmc ) .
while mi and sf phase can be characterized efficiently @xcite , the bg phase and the vicinities of the transition lines are significantly more complicated to be adequately described .
the main reason is that for finite size calculations in general , it is problematic to capture the correct description of the phase borders , which are essentially dominated by rare events , which is also a problem for qmc methods . in most cases ,
exact diagonalization studies are simply inadequate due to limited size and number of particles , which often obscures essential physics ( however , there may appear aspects that can indeed be suitably captured by small clusters @xcite ) . on the other hand , with a similar range of treatable system sizes as qmc , density matrix renormalization group ( dmrg ) is an efficient complementary method @xcite .
however , dmrg is currently only applicable in one spatial dimension and thus does not allow for a description of effects in higher dimensional lattices .
analytically , renormalization group analysis @xcite , slave boson theory @xcite , the strong coupling approach @xcite , and various different kinds of mean field theory descendants have been applied to the bhm with and without disorder @xcite .
arithmetically averaged mean - field theories , on the other hand , are incapable of resolving the bg phase at all for @xmath13 @xcite , but also for @xmath14 impose an unphysically strong phase coherence , leading to an overestimation of the sf phase .
other methods like the random phase approximation ( rpa ) which has successfully been applied to the system without disorder @xcite again suffer from finite size effects for the bhm with disorder , as the absence of translational symmetry constrains its applicability to much smaller system sizes . to circumvent this type of problems in describing the different phases of the disordered bhm
, we use the smft which has been previously introduced and applied to the disordered bhm at zero temperature @xcite .
there , it was found that the method efficiently describes localization , is both valid in high dimensions and in the thermodynamic limit , capturing rare events with their respective statistical weight and includes dimensional effects . in particular , it was found that at fixed @xmath7 , there exists a critical hopping strength , above which the system remains superfluid for arbitrarily strong disorder . in this article , we present the smft in detail and investigate how the results found for the disordered bhm at zero temperature are modified at finite temperature .
in addition to prototypical box disorder , we consider exponential speckle disorder to better simulate systems realized in current experiments . here , the results are qualitatively different in the sense that for any finite disorder strength the mi gives way to the bg .
the paper is organized as follows : in sec .
[ sec : method ] , the smft is explained in detail . first the general scope is outlined , followed by the definition of quantities computed within smft , such as compressibility , local greens functions and condensate fraction . in sec .
[ sec : phases ] , the essential results for the disordered bhm at @xmath13 are given , followed by the extension of the smft calculations to @xmath15 . in sec .
[ sec : box ] we extend the results presented for box disorder @xcite , including finite temperature effects .
alternative types of disorder , in particular disorder induced by speckle lasers are discussed for the bhm in sec .
[ sec : speckle ] .
as another possible source of disorder , we discuss the effect of kinetic ( hopping ) disorder in the bhm in sec .
[ sec : hopping ] .
however , we find no sensible dependence of the system on this parameter . the applicability of these results to current experiments essentially relies on the estimate of experimental temperature , which is discussed on in sec .
[ sec : temp ] .
we find that the experimentally realized temperatures are still far above the regime for which we resolve the previously stated interesting phenomena , such as disorder - induced condensation and reentrant behavior .
lda + smft calculations are discussed in sec .
[ sec : lda ] to provide a closer connection to experimentally measurable quantities . in sec .
[ sec : conc ] , we conclude that the smft is an efficient theoretical approach to disordered bose systems and promises an adequate description of ongoing experiments .
as pointed out previously@xcite , performing a self - consistent disorder average over all local on - site energies is not sufficient to generally describe the insulating bose - glass phase . from spatially resolved bosonic gutzwiller calculations
, it becomes apparent that this method overestimates long range correlations , predicting the formation of a global condensate into a single particle orbital which consists of the superposition of a large ( extensive ) number of distinct localized single particle states . in the true bose glass phase ,
off - diagonal long range order does not prevail and a large ( although not extensive ) number of particles may occupy each of these localized modes independently , leading to a condensate fraction @xmath16 in the thermodynamic limit .
however , by imposing phase rigidity and averaging over all mean - field parameters @xmath17 with the same complex phase , the spatially resolved ( as well as the arithmetically averaged ) gutzwiller theory leads to a finite average mean - field parameter ( mfp ) and a finite condensate fraction @xmath18 in the expected bg regime .
in contrast to the approaches mentioned above , the smft is constructed as a single - site theory in the thermodynamic limit , which effectively describes fluctuations in the mfps using a probability density function ( pdf ) @xmath19 . set up in this fashion , the smft method
is not restricted to incorporating disorder fluctuations only , but may also be a powerful approach to treat fluctuations of different origins , such as of thermal or quantum type in a unified framework .
the concept of approximating quantum mechanical operators by random variables has been applied previously in a variety of physical scenarios @xcite . in this paper , we will focus on disorder - induced fluctuations at finite temperature . in this section
, we will discuss the construction of smft in detail , concentrating on the case of including on - site disorder .
extensions of including disorder in the hopping parameter @xmath6 or thermal fluctuations explicitly are discussed in sec .
[ sec : hopping ] and app .
[ sec : explicit_thermal_fluctuations ] respectively . the central quantity , which effectively describes a disordered bosonic lattice gas in the thermodynamic limit is the probability distribution @xmath19 .
it is assumed to be equal for all sites and in particular independent of the nearest neighbors on - site energies .
the validity of this assumption has been checked using spatially resolved mean - field theory , which is known to become exact in the non - interacting limit , while retaining interaction effects beyond gross - pitaevskii theory for finite @xmath20 .
it is found that it yields correlation coefficients below @xmath21 in all regimes considered@xcite , justifying the above assumption .
the aim of smft is to find a self - consistent solution for @xmath19 , which is restricted and uniquely specified by self - consistency equations and minimization of energy .
consider a cluster of lattice sites composed of a central site @xmath4 and @xmath22 ( the coordination number ) nearest neighbors , where the corresponding part of the hamiltonian contains the operators on site @xmath4 within the bosonic gutzwiller approximation @xcite : @xmath23 within this approximation , the site @xmath4 is coupled to ( and the sum extends over ) the nearest neighbors via the scalar mfps @xmath24 , which are random variables within smft .
due to global particle number conservation , the mfps can all be chosen to be real and positive , as any variation in complex phase corresponds to a boost in local kinetic energy .
the expectation value @xmath25 is to be evaluated in the local ground state for @xmath13 or taking the thermal trace for @xmath14 .
thermal fluctuations can , however , also be incorporated on an explicit stochastic level in the distribution @xmath19 within smft , as discussed in app .
[ sec : explicit_thermal_fluctuations ] . inspecting the hamiltonian ( [ eff_ss_hamiltonian ] ) , it does not depend on the @xmath26 s individually , but only on the scaled sum @xmath27 since the @xmath26 s ( and in the hopping disorder sec . [ sec : hopping ] also @xmath6 ) are random variables , the newly defined quantity @xmath28 is also a random variable obeying some distribution @xmath29 . as discussed above , the random variables @xmath26 are assumed independent within smft , allowing us to express the distribution @xmath30 by a @xmath22-fold scaled convolution @xmath31 making use of the convolution theorem , this can be reduced to two one - dimensional fourier transforms by introducing the characteristic function @xmath32 in terms of which the function @xmath33 can be expressed as @xmath34^z \ , e^{-i t \eta / j } \label{qdistribution}.\ ] ]
this can be calculated efficiently for arbitrary coordination numbers @xmath22 using the fft algorithm .
the numerical procedure is discussed in app .
[ multi_grid_fft ] .
the self - consistency condition can now be formulated in the following way : if the on - site energy @xmath35 is randomly drawn from @xmath9 and @xmath26 is randomly drawn from the self - consistently determined distribution @xmath19 for each of the @xmath22 nearest neighbors , this defines the single site hamiltonian ( [ eff_ss_hamiltonian ] ) , which can be diagonalized providing the ground state @xmath36 .
from there , a new mfp @xmath37 can be calculated , with the self - consistency requiring the distribution of this new random variable to be exactly the distribution @xmath19 we initially assumed for the neighboring @xmath26 s . to cast this condition into functional form for a probability distribution in the thermodynamic limit ( i.e. for an infinitely large system at fixed density ) ,
we first define the function @xmath38 where @xmath39 is the ground state of @xmath40 given in ( [ eff_ss_hamiltonian ] ) .
it is useful to introduce the _ conditional _ pdf , which is a function of @xmath41 and @xmath42 , giving the probability density for a specified value @xmath42 if the value of @xmath41 is fixed and @xmath35 is distributed according to @xmath9 .
this can be obtained by using the transformation property of a pdf under a variable transform and takes the form @xmath43 this function does not obey any self - consistency condition and can be evaluated directly , which is the first step in finding @xmath19 .
remembering that the relation between @xmath30 and @xmath19 is given by ( [ def_q ] ) , so that we can express the self - consistency condition as @xmath44 the right hand side can be understood as follows : for every fixed @xmath41 we have a pdf @xmath45 , which yields a contribution with the respective weight @xmath30 , leading to a marginal distribution for the considered site . if this agrees with the initially assumed distribution @xmath19 , a self consistent solution has been found . once this is determined , expectation values of local , self - averaging operators @xmath46 can be directly be determined by @xmath47 where @xmath48 is an effective disorder averaged density operator , incorporating thermal , on - site energy and mfp fluctuations , depending explicitly on @xmath9 and the self - consistently determined @xmath30 . to solve the smft equations numerically , we iterate the self - consistency equations on a discretized grid for @xmath42 , consisting of a superposition of a variable number of equidistantly spaced grids , as explained in app .
[ multigrid ] . for every fixed set of physical parameters
, we first numerically determine the conditional cumulative density distribution function @xmath49 for all values of @xmath41 and @xmath42 which constitute the numerical grids for @xmath30 and @xmath19 respectively ( discussed in app .
[ app : numerical_cdf ] ) . working with the cumulative distribution on a numerical level , as opposed to the pdf itself , is far more controlled and circumvents divergences in the pdf @xmath19 , but also in the conditional pdf @xmath45 . the self - consistency condition ( [ selfconsistency ] )
is not influenced by this approach . as can be seen by inspection ,
the insulating solution @xmath50 is always a self - consistent solution , equivalent to the @xmath51 solution in the single site theory .
however , in the sf regime , there also exists a second , non - trivial self - consistent solution , which corresponds to a lower grand canonical potential and is therefore the physical solution in this case .
furthermore , the physical solution is always found to be the attractive fixed point of the self - consistency mapping in the space of probability distributions , i.e. if the iteration procedure is started at any @xmath52 , the successive distributions continuously converge towards the physical distribution .
we start the iterative procedure with an initial pdf @xmath53 , where all the weight is distributed at small , but non - zero values of @xmath42 , assuring fast convergence in the insulating state and in the vicinity of the phase border .
the distribution in the @xmath4-th iteration step for the scaled sum of mfps from the nearest neighboring sites @xmath54 is calculated from @xmath55 using the convolution theorem for independent random variables and the fft algorithm ( see app .
[ multigrid ] ) . the new distribution @xmath56 is then obtained by integration over @xmath41 .
numerically this is done by using the trapezoidal rule , as we found that using higher order techniques , such as simpson s rule , is not robust and lead to incorrect results if @xmath57-peaks appear in the pdf . in the vicinity of the phase border , the computational effort increases due to two effects .
firstly , the convergence is critically slowed down , increasing the required number of iterations .
secondly , the discretization of the @xmath42-grid plays an ever increasing role and in some parameter regimes directly on the outside of the the mi lobes , where the converged distributions are very close to a @xmath57-peak at @xmath51 , the numerically determined form of the distributions depends on the discretization ( resolution ) , which is clearly unphysical . in these cases , we have to determine @xmath19 by examining a sequence of converged distributions at ever increasing resolution and define the physical distribution as the limit of this sequence .
for the disordered bhm , three different phases exist at zero temperature : a mott insulating state , where number fluctuations are suppressed and the particles are localized due to a repulsive interaction .
this state exhibits a finite energy gap of order @xmath10 , thus the single - particle density of states ( dos ) at @xmath58 vanishes and the state is incompressible . if tunneling - induced delocalization dominates , the system is in a condensate ( sf ) phase , where a macroscopic number of particles can lower their energy by condensing into one single - particle state , thus exhibiting quantum coherence and leading to a finite condensate fraction . within a grand - canonical mean - field description ,
this phase breaks the @xmath59-symmetry of the bh hamiltonian ( [ bh_hamiltonian ] ) and leads to a non - zero order parameter .
the phase border from the sf to any of the insulating phases is thus determined by smft , where finite weight in @xmath19 moves to finite values of @xmath42 .
finally , there is the bose glass phase , where particles are localized by an interplay of disorder and interactions .
however , there exists no single particle state which is occupied macroscopically and thus the bg is not a condensate , i.e. the condensate fraction vanishes ( @xmath16 ) .
however , there does exist an extensive number of localized single particle states , each of which is occupied by an arbitrarily large , but not macroscopic number of particles .
this may be understood as a highly fragmented system of _ incoherent _ localized ` non - macroscopic quasi - condensates ' .
the transition from a bg to a mi is not determined from the self - consistent distribution @xmath19 , but by analyzing the compressibility @xmath60 or the single particle dos .
we consider the latter quantity within two frameworks : 1 . considering the purely sing site particle- and hole excitations as specified in @xcite .
the bg extends over the region where @xmath61 and values of the chemical potential @xmath62 , i.e. the borders are independent of @xmath6 and a direct mi - sf transition is possible .
a more detailed analysis presented in @xcite relies on the analysis of an effective hamiltonian in the subspaces of localized single particle- and hole excitations . for finite
hopping @xmath6 in the pure system , these hybridize , lifting the degeneracy and form superpositions with quantum numbers @xmath63 .
increasing ( decreasing ) @xmath64 , the mott insulating state remains the ground state until the energy difference between the mi state and the @xmath65 particle ( hole ) state vanishes , at which point particles delocalize and condense into these states .
now let us return to the disordered case : here the local particle ( hole ) excitations will not hybridize into fully delocalized states with well - defined quasi - momentum @xmath63 , but into inhomogeneous states , which depend on the individual disorder configuration .
it is however possible to make exact statements about the eigen - energy spectrum for a disordered system in the thermodynamic limit : the lowest kinetic energy is obtained in locally homogeneous regions and approaches the energy of the @xmath65 particle ( hole ) state of the pure system , as the size of this locally homogeneous lifshitz region increases .
furthermore , scaling predicts that the dependence of the kinetic energy on the specific boundaries to the lifshitz region will reduce , as the size of the region increases . on the other hand ,
the potential energy of the particle ( hole ) state is minimized , when the local on - site energy takes on the lowest ( highest ) possible value over the whole region , i.e. lies at the extrema of @xmath9 .
therefore , upper ( lower ) phase boundary of the mott lobe in the disordered system is obtained by shifting the upper ( lower ) boundary down ( up ) by @xmath66 ( @xmath67 ) .
the mi for box disorder in 3 dimension at @xmath13 obtained by strong coupling theory using this criterion , is the area enclosed by the orange phase boundaries in fig .
[ box_phase_diags_t ] .
the insulating region outside of these lobes , bounded from the sf by the dashed white lines , corresponds to the bg . using this criterion
, the transition from mi to sf always occurs through the bg phase for box disorder within smft . at finite temperature @xmath14 ,
the system is always compressible and the incompressible mott insulator is replaced by a normal ( non - superfluid ) phase with a thermally induced compressibility .
a central statement in @xcite is , that the disorder - averaged single particle dos calculated by considering purely local excitations only , allows for a clear distinction at @xmath14 between the bg and the normal phase , as it is zero in the latter .
however , when considering delocalized excitations in the particle and hole sector , this statement holds no longer , as there are generally degenerate states in the @xmath68-particle particle - hole band and the particle ( hole ) band in the @xmath69 sector ( or the @xmath70 sector for the hole band ) , which can be seen from the single particle dos in the lehmann representation @xmath71 .
\end{split}\end{aligned}\ ] ] [ fig_energy_spec ] without disorder .
the levels correspond to the energies of the exact many - particle energy eigenstates of the system with interactions . due to the @xmath59-symmetry
the energy eigenstates can be chosen to be of well - defined particle number .
for commensurate particle number @xmath68 , the ground state ( mi ) is separated from a band of particle - hole excitations by a charge gap , while for @xmath69 particles , all eigenstates are delocalized , forming a band of hybridized particle states . at @xmath13 ,
the single particle dos has finite weight at frequencies corresponding to energy differences @xmath72 of transitions between @xmath68 and @xmath69 states . at @xmath14 transitions between different excited states also
contribute , causing the gap to vanish if the two bands overlap and the respective matrix elements do not vanish . ] here , @xmath73 denotes the energy of the @xmath74 many - particle eigenstate @xmath75 of @xmath76 , which can always be chosen to have well - defined particle number @xmath68 , since @xmath77=0 $ ] , while @xmath78 is a label for the eigenstate in this subspace . @xmath79 and @xmath80 are the amplitudes for the various possible transitions . if the system is in a mott insulating state ( considering no disorder for the clarification of this argument ) as shown in fig .
[ fig_energy_spec ] , the ground state with @xmath68 particles is separated from a band of particle - hole excited states by a gap @xmath81 ( corresponding to the height of the mott lobe at the respective @xmath82 ) . as transitions between various excited states
can also occur at finite temperature , the spdos thus vanishes if the particle - hole band with @xmath68 particles , overlaps with the particle- or hole band containing @xmath69 or @xmath70 particles , but the weight is suppressed exponentially by @xmath83 .
we thus conjecture , that the bg and the normal phase are not fundamentally distinguishable , and only connected by a crossover .
following along the lines of @xcite , we use the disorder averaged local single particle density of states ( dos ) at zero frequency @xmath84 ( see eq . , ) , to determined the normal phase / bg crossover at finite @xmath85 . in the presence of disorder
the spdos containing localized excitations can be calculated explicitly from the local green s function in this regime ( see app .
[ app_loc_dos ] ) and leads to @xmath86 . \end{split}\ ] ] @xmath87 is the local partition function at an effective chemical potential @xmath88 and the two @xmath57-distributions correspond to local particle and hole excitations respectively . using this criterion leads to an overestimation of the mi / normal phase over the bg and thereby an absence of the bg phase around the tips of the mott lobes .
although phase transitions are , strictly speaking , only well defined in the thermodynamic limit , crossovers observed in current experiments may indicate phase borders that do not coincide with the borders obtained in systems of infinite size .
as discussed above and in@xcite , the mi / bg phase borders in an infinitely large disordered system simply correspond to the shifted phase borders of the pure system .
this argument relies on the existence of arbitrarily large lifshitz regions , which is clearly no longer given for finite systems .
the phase diagram for a finite system therefore strongly depends on the specific disorder realization , with the critical values for the phase borders becoming random variables .
therefore , a better question to ask for a finite system for instance is : for a randomly chosen disorder realization in a system consisting of @xmath89 sites , what is the probability @xmath90 that the energy gap will be lower than a given value @xmath91 ? in the limit of @xmath92 this probability is given by @xmath93 with the restriction that @xmath94 and @xmath95 with increasing @xmath82 the mi / bg phase border in the @xmath82-@xmath64-diagram remains a random variable , but with a reduced steepness in the slope for finite systems , as shown in @xcite .
the bg / sf would also be very likely ( in a statistical sense ) to move to larger critical values of @xmath6 in a finite system , as the occurrence probability of ` rare events ' , favoring a sf , is suppressed .
results for box disorder at @xmath13 obtained by smft have been presented in @xcite , here we extend the phase diagram by taking collective excitation in the bg into account and focus on finite temperature .
all numerical results presented in this paper are for a three - dimensional cubic lattice ( z=6 ) .
box disorder is characterized by a constant probability density for the on - site energies @xmath96 over a bounded interval of width @xmath97 , where @xmath98 denotes the heaviside function .
changing the disorder strength may have an influence on the system properties on various levels . coming from the insulating state within a local mean - field picture , where the total state of the system is a direct product of local fock states minimizing the total energy , an increase in disorder reduces the smallest possible particle or hole excitation energy .
the local energy gap for a site with an effective chemical potential @xmath88 and a fock state as ground state @xmath99 is @xmath100 for adding a particle and @xmath101 for creating a hole , if @xmath102 . in the presence of disorder in an infinitely large system , this implies that the energy gap is necessarily reduced by @xmath103 and vanishes in the @xmath104 limit as soon as the interval of realizable effective chemical potentials contains a positive integer multiple of @xmath10 , i.e. @xmath105 \neq \emptyset$ ] .
( color online ) compressibility @xmath60 and local single particle dos in an insulator for box disorder at @xmath106 at various temperatures ( in units of @xmath107 ) . at zero temperature @xmath60
vanishes in the mi , but not in the bg and may thus be taken as a quantity to distinguish between them . at finite temperature
however , @xmath60 becomes non - zero in the normal phase and the sharp features become rounded off by thermal fluctuations . in this case , the disorder averaged dos of local excitations @xmath108 , which retains all sharp features at @xmath14 , can be used to determine the crossover from a bg to the normal phase , where it remains zero . ] within the picture of purely local excitations , this argument is independent of temperature and directly reflected by the disorder averaged local dos @xmath109 .
\end{split}\ ] ] in the case of bounded disorder , this directly implies that the system can not be in a gapped state as soon as the carrier width ( @xmath97 for box disorder ) equals or exceeds @xmath10 and that the system is then always in the bg or sf state , i.e. the mi and normal state can only exist at @xmath110 ( and not at all for unbounded disorder ) .
when the system is either of these two insulating states at zero temperature , the compressibility @xmath111 is determined by @xmath84 within a picture of local excitations , where locally degenerate sites change their occupation number by @xmath112 if @xmath7 is slightly altered .
since for finite temperature thermal particle or hole excitations are also present when the system is gapped , @xmath60 is always driven to a finite value .
the behavior of @xmath60 and @xmath84 is shown in fig .
[ fig_compressibility_of_delta_box ] at the specific chemical potential of @xmath113 .
starting at @xmath114 , the system is in a gapped mi / normal state , but increasing @xmath97 , the system undergoes a phase transition at a critical disorder strength of @xmath115 . at this point the local dos @xmath84 takes on a non - zero value , as two different local fock states can become degenerate at the edge of the box disorder distribution . at low temperatures the compressibility exhibits a significant change from an exponentially small value in @xmath85 in the normal phase to a large value due to the existence of local configurations of on - site energies which give rise to almost degenerate states with different particle number .
there , a small change in the chemical potential leads to a local jump in particle number .
in the zero temperature limit the compressibility locally behaves as @xmath116 , since with increasing disorder strength the statistical weight of these events is reduced . when the effective chemical potentials at the borders of the probability distribution enter a region corresponding to a new local particle number @xmath117 , the compressibility increases with a jump , corresponding to subsidiary phase transitions between different bg phases at @xmath13 , which turn into less pronounced crossovers with increasing @xmath85 . considering the limit of strong disorder @xmath118 at fixed @xmath64 in fig .
[ fig_compressibility_of_delta_box ] , the system approaches a state in which only half the number of sites are occupied and @xmath119 .
however , disorder does not only affect local excitation properties , but leads to an intricate interplay between the hopping and interaction energy scale , influencing the overall coherence properties of the system .
whereas an increase in temperature or interaction energy generally tends to counteract the formation of a condensate , this is true in most , but not all scenarios when increasing the disorder strength . in certain parameter regimes at sufficiently low temperature , an increase in @xmath97 can actually lead to the formation and stabilization of a condensate ( i.e. disorder induced condensation ) , as previously predicted by various methods , including smft @xmath120 .
( color online ) finite temperature phase diagrams for box disorder with @xmath121 for three temperature regimes : zero temperature ( upper row ) , intermediate @xmath122 ( middle row ) and high temperature regime @xmath123 ( lower row ) .
the left column shows the expectation value of the self - consistently determined distribution , which is the order parameter for the sf - insulator transition .
it vanishes in the mi , bg , and the @xmath14 normal phase . in the @xmath13 diagrams ,
the mi / bg phase borders from strong coupling theory are indicated by the orange solid lines .
the bg phase is marked by white stripes , at finite @xmath14 this is determined by a vanishing local dos only ( i.e. not by strong coupling theory ) . at finite temperature
there is only a crossover between the normal and bg phase , and the borders determined by the dos of purely local excitations is marked by the horizontal white lines .
the right column shows the compressibility @xmath60 in units of @xmath107 . whereas this is a suitable quantity to distinguish the mi ( incompressible ) from the bg at @xmath13 , it is non - zero , but exponentially small in @xmath85 in the mi in the intermediate regime and of order @xmath107 in the high temperature regime .
the white borders indicate the transition to the sf , determined from the respective diagram on the left . ] in fig .
[ box_phase_diags_t ] , the effect of increasing temperature is exemplified in three phase diagrams and the compressibility in the three main regimes : the zero , low and high temperature regime at fixed disorder strength .
the tips of the mi /normal lobes remain almost unchanged under an increase in @xmath85 , while the bg region between the initial lobes is strongly enhanced and stabilized , even at a very low temperature of @xmath124 ( central plots of fig .
[ box_phase_diags_t ] ) .
since the sf / insulator phase border still possesses a distinctive lobe structure , but the bg region is strongly enhanced in contrast to the t=0 case ( upper plots of fig .
[ box_phase_diags_t ] ) , we therefore refer to this situation as the intermediate temperature regime .
furthermore , a large value of compressibility is still a strong indicator for the bg at this temperature , suggesting that it may still useful as an indicator of the transition between the bg and the normal phase in experiment .
in the high temperature regime @xmath125 , on the other hand , ( see lower diagrams in fig . [ box_phase_diags_t ] ) the compressibility is large throughout , approaching the value @xmath126 in the high @xmath85 limit .
the typical temperature of @xmath127 at which the mi and normal phase melts is consistent with previous previous studies @xcite . in this temperature regime
the lobe structure of the sf / insulator phase border is totally wiped out and the critical border to the sf follows a @xmath128 decay with increasing filling .
the sf / bg transition is highly sensitive to the system size . when the system approaches the sf phase from the bg phase , the localized single particle orbitals occupied by a large number of bosons increase in size and are occupied by an ever increasing , but never extensive number of particles in the bg phase . at the transition point
the localization length , being a measure for the size of these orbitals , diverges and driven by percolation , phase coherence between neighboring orbitals is established , eventually driving the system into the sf phase . in finite size systems ,
the detection of the transition point thus critically depends on the system size , as the bg phase may be mistaken for the sf phase if the localization length is larger than the system size .
smft has the advantage that it is constructed in the thermodynamic limit for an infinitely large system in the grand canonical ensemble and takes all possible disorder realizations into account within a functional description for the probability distributions .
the numerical error in the discretization performed for the distribution @xmath19 is well controlled and to be distinguished from the finite size deviations made in real space calculations . in the phase diagram for box disorder fig .
[ box_phase_diags_t ] , smft does not give rise to a direct transition from the mi to the sf , if the extended criterion for the mi / bg border , including collective excitations in rare regions is used .
this is furthermore demonstrated in the finite temperature phase diagram at constant filling @xmath129 in fig .
[ box_phase_diags_n1 ] . due to the absence of a clear distinction between the normal and bg phase at @xmath14
, the orange border only indicates a crossover between these regimes , but would go over into a mi / bg phase border for @xmath13 . at any @xmath130
a finite bg region intervenes between the mi and sf phases .
the question whether this transition always occurs via the bg phase has been a highly debated topic since the introduction of the disordered bhm @xcite , and was established for the one and two dimensional case @xcite . in a recent work @xcite
it was shown that this scenario is true in any finite dimension for bounded disorder , due to the statistical certainty that any possible configuration of on - site energies for a cluster of sites will occur in the limit of an infinitely large system . in the previous work @xcite these collective excitations in the bg phase
were not considered and within a simpler framework of purely local particle and hole excitations , a direct transition was predicted .
it is known that arithmetically averaged mft , as well as smft providing an improvement in any finite dimension , both become exact in the limit of infinite dimensions , where no bg exists at @xmath13 . however , as argued recently@xcite the theorem of inclusions guarantees the existence of an intervening bg phase between the sf and the mi in any arbitrarily high , but finite number of dimensions for bounded disorder .
it is instructive to understand the decrease of the bg region , including the collective excitation in the lifshitz regions in terms of percolation physics : for any finite dimension the outer border between the bg and sf phase specifies the critical value @xmath131 specifies the lowest energy at which it becomes energetically favorable for the particles to form a global condensate ( in smft this is the border where @xmath132 takes on a finite value ) .
the border between the mi and bg inside the global insulator , specifies the critical value @xmath133 at which it becomes possible for the system to form large local superfluid patches ( locally resembling pure systems ) without phase coherence between different patches . with increasing dimensionality of the system , the connectivity between different patches increases ( percolation is enhanced ) and the required tunneling energy @xmath131 to form one large percolated patch , i.e. a global condensate , decreases . in the limit of high dimensions , the critical values of @xmath6 at which these two phenomena occur approach the arithmetically averaged mean - field value@xcite @xmath134^{-1 } , \end{split}\end{aligned}\ ] ] and the bg disappears , where @xmath135 is the filling and @xmath7 and @xmath97 are given in units of @xmath10 .
( color online ) phase diagram for box disorder at fixed density @xmath129 in the @xmath20-@xmath136 plane at @xmath122 showing the mean order parameter @xmath137 .
reentrant superfluidity is reflected by the protruding sf lobe ( we also find a second less pronounced lobe at higher @xmath97 ) , where increasing @xmath97 can drive the system through a sequence of sf - insulator transitions .
the orange line specifies the crossover from the normal to the bg phase ( at @xmath13 this line becomes the mi - bg phase border ) as determined by by shifted mean - field phase borders at this temperature , giving a better approximation than the strong - coupling approach for three dimensions .
the simpler criterion of looking at purely local particle and hole excitations eq .
[ do_av_dos_box ] would lead to the mi / bg phase border at @xmath138 and predict a direct mi - sf transition at small @xmath97 . including collective excitations in the bg , as done here ,
always leads to an intermediate bg phase between the mi and sf in the @xmath13 limit .
] in fig .
[ box_phase_diags_n1 ] we present a phase diagram calculated at fixed density @xmath129 in the low temperature regime @xmath122 . at every point in the diagram , the self - consistent distribution is calculated for a fixed @xmath7 , enabling the calculation of the density @xmath139 .
thereafter @xmath7 is iteratively determined using ridder s algorithm @xcite until the density obtained from smft does not deviate more than @xmath140 from the specified density . in fig .
[ box_phase_diags_n1 ] , the disorder averaged mfp @xmath141 ( within smft , this is exactly @xmath142 ) clearly shows the usual sf / insulator phase transition ( at fixed low temperature ) along the line @xmath114 , where the disorder localizes the particles with increasing @xmath20 .
moving outwards into the @xmath136 at fixed interaction @xmath20 , the condensed phase is surprisingly robust , surviving local on - site fluctuations @xmath97 several hundred times larger than the hopping energy @xmath6 , as pointed out in a recent work@xcite .
this can be understood from the bosons filling up the low - lying sites and forming a background sea via the repulsive interactions , creating an effective smoother potential in which it is energetically favorable for the remaining bosons to delocalize@xcite .
a remarkable effect at sufficiently low temperature is the appearance of a sf lobe , protruding into the insulating domain at finite @xmath97 . in this regime
the interplay between disorder and interactions is non - monotonic in these two effects , and , for the regime @xmath143 , an increase in @xmath97 drives the system into the sf phase , delocalizing the particles .
this effect can be understood from the pure bhm @xmath64-@xmath82 phase diagram and relies on the existence of a lobe structure , i.e. requires a sufficiently low @xmath85 . to keep the particle number constant with increasing @xmath97
, @xmath7 is required to increase .
in certain regimes the majority of sites in the system may enter from an insulating regime between the lobes into a regime inside the lobes , thereby favoring condensation .
qualitatively , the smft phase diagram agrees well and shows the disorder induced sf lobe , as found in recent qmc calculation @xcite at @xmath13 on relatively small lattices ( @xmath144 ) . at large @xmath97 the order parameter
is non - monotonic in @xmath20 , vanishing at sufficiently small @xmath20 which indicates a transition into an anderson localized state , where the localization almost exclusively disorder - induced .
however , the region of extremely small @xmath20 is problematic as @xmath145 and @xmath64 diverge , since very few sites have to contain an ever increasing number of particles to keep the disorder - averaged density fixed , when asymptotically half the number sites ( due to the symmetry of the box distribution @xmath9 ) have such a high effective on - site energy , that they contain no particle . due to the diverging local occupation number
, this limit transcends the constraints imposed in the derivation of the bhm in an optical lattice and is , in this sense , unphysical .
although a homogeneous box distribution is most commonly used for disorder calculations in theory , it is currently not an experimentally feasible choice . in this section
we discuss and compare the results for a realistic disorder distribution created by a speckle laser to those of a box disorder distribution .
a laser passing through an inhomogeneous disordered plate leads to a disordered optical potential , which is the fourier transform of the disordered pattern on the plate . in recent experiments ,
it has become possible to reduce the autocorrelation length of this disordered potential to the order of the lattice spacing ( @xmath146 ) @xcite . with this experimental achievement , the priorly
most criticized artifact of a speckle laser for creating uncorrelated disorder has been overcome , thereby making speckle potentials the most promising method for future disorder experiments in optical lattices .
the resulting distribution for uncorrelated on - site energies is well approximated by @xmath147 although it may be argued that an optical speckle potential in experiment is fundamentally bounded by its finite size , it is only essential that the width of the on - site energy distribution exceeds @xmath10 ( which is fulfilled in essentially all experimentally relevant regimes ) .
this justifies the use of @xmath148 .
( color online ) typical self - consistent distributions @xmath19 for speckle disorder for @xmath149 , @xmath150 , @xmath151 , and @xmath152 at different temperatures . for low temperatures the mean - field parameters in the condensed phase are robust against finite temperature fluctuations , but with increasing temperature the system is eventually driven into an insulating bg phase , as indicated by the distribution at @xmath61 in the at @xmath153 .
] to treat this disorder distribution using smft , it is useful to perform a transformation of variables @xmath154 , which on a formal level transforms the smft conditional probability functions into a form analogous to homogeneous disorder .
this step enters only on the level of calculating the conditional cumulative distribution function ( cdf ) , @xmath155 apart from this , the smft method remains identical to the homogeneous disorder case .
similarly , arbitrary disorder distributions may also be incorporated into smft , although an analytical transformation of the random variable will not exist in general .
in contrast to box disorder , which has been the distribution primarily focused on so far when considering the disordered bhm , speckle disorder is unbounded and arbitrarily high values of @xmath35 have a finite probability to occur .
this leads to the effect ( strictly only possible for an infinitely large system ) that _ turning on _ the disorder by an arbitrarily small amount immediately changes a large part of the phase diagram from the mi / normal phase into the bg phase . in a physical picture
, the effective chemical potential @xmath88 can then take on integer multiple values of @xmath10 ( the on - site potential can become arbitrarily high ) with a non - zero probability density , where local fock states with different particle number become degenerate , leading to a finite compressibility and local dos @xmath84 at @xmath13 .
( color online ) disorder averaged mfp @xmath156 characterizing the sf - insulator transition ( left column ) , and compressibility ( right column , in units of @xmath107 ) in the @xmath157-@xmath64-plane for _ speckle disorder_. diagrams are shown for increasing disorder strength ( @xmath158 , @xmath159 , @xmath160 ) in the zero or low temperature regime ( @xmath13 , @xmath161 , @xmath161 ) for the ( upper , middle , lower ) row respectively .
in contrast to the behavior for box disorder , the structure of each lobe does not change symmetrically , but rapidly extends in the direction of decreasing @xmath64 with increasing disorder strength @xmath97 .
since the mi / normal phases do not exist for speckle disorder , the insulating region ( black in the left figures ) is always a bg .
the white lines denote the sf / insulator phase boundaries , indicating where @xmath162 takes on a finite value . ] in the absence of disorder and at sufficiently high tunneling coupling @xmath6 , a macroscopic number of particles occupy the @xmath163 bloch state . within an effective , symmetry breaking gutzwiller description the local order parameters @xmath164 then take on a finite and constant value , reflecting the translational symmetry of the system . within smft this state
is characterized by a @xmath57-distribution @xmath165 , where @xmath166 is the order parameter of conventional bosonic gutzwiller theory . turning on the disorder in such a system in the sf state breaks
the translational symmetry of the system , i.e. the condensate state deviates from the @xmath65 bloch state , which is reflected by the distribution of mfps @xmath19 taking on a finite width .
initially for weak disorder , an increase in disorder always leads to a broadening of @xmath19 , but for stronger values of @xmath97 the system may eventually be driven toward an insulating state , driving @xmath167 and thereby decreasing the fluctuations in the mfps . on the other hand , increasing the temperature suppresses the sf and leads to a decrease of the mfps above a certain temperature , up to which the sf remains stable , as shown in fig .
[ fig : dists_speckle_various_t ] .
( color online ) the compressibility , local single particle dos and chemical potential as a function of disorder strength @xmath97 for speckle disorder at constant filling . comparison of the upper two plots shows a diverging compressibility and local dos for @xmath168 for non - integer filling ( the system remains superfluid ) , while for integer filling it drops to an exponentially small value in @xmath97 at a value of @xmath169 and vanishes in this limit .
the lowest figure shows the same quantities at finite temperature @xmath170 , where thermal fluctuations have totally smeared out the sharp features in the compressibility . however , these persist in the local dos , although their position is changed due to a temperature induced shift in @xmath7 . ]
the influence which speckle disorder has on the @xmath64-@xmath82-phase diagrams is shown in fig .
[ fig : dists_speckle_various_delta ] . in contrast to box disorder , where the distribution of on - site energies is symmetric around @xmath7 and the insulating lobes give way to the sf in the same way on the upper and the lower side of the lobe , the insulator forms on the lower side of the lobes with increasing @xmath97 for speckle disorder .
this can be understood from the fact that only lower values of the effective local chemical potential can occur .
for strong disorder @xmath171 , the lobe structure of the insulator / sf phase boundary is washed out , which is similar to the effect of finite temperature . for speckle disorder
, @xmath60 can not be used to identify a phase transition , since it is non - zero in both the bg and the sf .
a question of interest , regarding the disappearance of the mi /normal phase for an arbitrarily small amount of speckle disorder , is how the compressibility behaves as a function of @xmath97 , as some interpretation is needed , that an infinitesimal amount of disorder can instantaneously convert the whole mi / normal area of the phase diagram into a bg . in fig .
[ fig : kappa_of_delta_speckle ] , @xmath172 is shown for different parameters to clarify this dependence . in the insulating state ,
the compressibility can be calculated explicitly ( [ insulator_kappa ] ) and one obtains @xmath173 \end{split}\end{aligned}\ ] ] essentially , two different scenarios have to be considered .
first , if @xmath64 is positive and integer , @xmath60 diverges in the limit of vanishing disorder , as is well known from the pure bhm phase diagram , where the density is a step function in @xmath64 for @xmath174 .
this is equivalent to the case for non - integer , fixed particle density @xmath135 , where the system remains sf for any non - zero @xmath6 .
second , if @xmath64 is non - integer , the compressibility vanishes with decreasing @xmath97 , as the system approaches a point in a mott lobe away from the border .
this corresponds to the case of fixed , integer - valued density @xmath135 .
we will now discuss the behavior of @xmath60 and @xmath175 at fixed particle density @xmath135 , shown in fig .
[ fig : kappa_of_delta_speckle ] . keeping the density constant with rising disorder ,
requires the chemical potential to be increased , as an ever increasing number of sites shifts to weights with lower occupation numbers . at every point when @xmath64 passes a positive integer number ,
a new fock state becomes potentially occupied , but with an ever decreasing statistical weight as @xmath97 increases . as a result
the compressibility experiences a jump at each of these points , as highlighted by the gray regions in fig .
[ fig : kappa_of_delta_speckle ] , where @xmath64 ( dotted gray lines in fig .
[ fig : kappa_of_delta_speckle ] ) passes an integer value .
this leads to the characteristic series of ever smaller kinks in @xmath172 . at finite temperature ,
these features in @xmath60 are smeared out over a typical scale of @xmath176 , whereas the sharp features in the local single particle dos survive at @xmath14 within smft ( solid blue lines in fig .
[ fig : kappa_of_delta_speckle ] ) .
the disorder averaged single particle density of states @xmath177 at zero frequency for @xmath178 for different speckle disorder intensities . for @xmath114
this consists of a sequence of @xmath57-peaks at positive integer values of @xmath64 , however for any @xmath130 this quantity is non - zero for @xmath179 and the system is in the bg phase . ] to clarify the effect speckle disorder has on @xmath84 and the immediate disappearance of the mi / normal phase at any @xmath130 , the local dos is plotted for weak ( @xmath180 ) , intermediate ( @xmath181 ) and strong ( @xmath182 ) in fig .
[ fig : speckle_dos_along_mu ] as a function of @xmath7 . in the pure system
@xmath183 consists of a sum of @xmath57-peaks at integer values of @xmath64 , i.e. at these values of the chemical potential there are two degenerate fock states @xmath184 and @xmath185 at all sites in the insulator and the local single particle dos diverges .
as soon as speckle disorder is turned on , these @xmath57-peaks are broadened according to the on - site energy distribution ( [ speckle_distribution ] ) and @xmath84 takes on the form of a sequence of superimposed exponential functions , each decaying with the constant @xmath97 . from this
it is clear that @xmath84 takes on a non - zero value as soon as @xmath130 at any @xmath186 , although it is exponentially suppressed for most values of @xmath7 at weak disorder @xmath187 . at zero temperature
the different amplitudes of the various peaks in @xmath183 at integer @xmath7 can be exclusively attributed to the @xmath188 factor from the action of the bosonic operators , whereas at @xmath14 the amplitudes ( but not the positions of the sharp features ) may also be modified by the boltzmann factors in ( [ dos_fct_speckle ] ) .
( color online ) main -@xmath20-plane for fixed @xmath189 at fixed filling @xmath129 in three dimensions .
for this specific disorder strength the sf region is enlarged by the disorder in comparison to the pure case ( i.e. disorder induced condensation , see sec .
[ sec : reentrant_sf]).in the lower left corner , the sf exists at sufficiently low @xmath85 and @xmath97 and the value of disorder - averaged @xmath132 is color - coded . outside the sf region , the system is always in a bg phase , but undergoes two crossovers from a regime with exponentially low compressibility @xmath190 for intermediately low @xmath191 ( for @xmath192 ) into a strongly compressible bg regime , in the limits of both high and very low @xmath193 . to clarify the quantitative behavior ,
the compressibility @xmath60 , the chemical potential @xmath7 and the local single particle dos @xmath194 along the dashed line @xmath195 is shown in the inset . ]
[ fig : speckle_pd_t_u ] shows the phase diagram at constant disorder strength @xmath196 and constant density @xmath129 in units of @xmath6 .
the sf region prevails in the lower left region at low temperature and weak interaction @xmath20 . in this parameter regime
the disorder stabilizes the sf phase , actually extending the sf region of the phase diagram in contrast to the pure ( @xmath114 ) case ( disorder - induced condensation , see sec .
[ sec : reentrant_sf ] ) .
all of the non - sf region in fig .
[ fig : speckle_pd_t_u ] is a bg , since we are dealing with unbounded disorder with @xmath130 , but we can identify a weakly and two strongly compressible regimes in the phase diagram . since @xmath61 in the bg , the energy scale @xmath6 can not influence thermodynamic quantities beyond a scaling relation , implying that the compressibility @xmath190 may be a function of @xmath193 only .
therefore the compressibility has a radial structure and it suffices to consider its behavior along a single line ( such as @xmath195 , depicted in the inset of fig . [ fig : speckle_pd_t_u ] ) .
this reveals that there are three regimes in this phase diagram : at high temperature @xmath197 the system is strongly compressible ( @xmath190 is of order unity ) , as thermal fluctuations have wiped out the sharp peaks over a wide range in @xmath7 . at intermediately low temperature ,
the compressibility is exponentially low , as the typical thermal excitation energy scale does not suffice to excite the majority of sites to higher states .
somewhat surprisingly , within a certain parameter regime for @xmath97 and the density @xmath135 , the system undergoes a further crossover at very low temperature @xmath198 into a second highly compressible regime . at @xmath13 and at integer filling , the chemical potential @xmath64 approaches an integer value from below ( green dotted line in the inset of fig .
[ fig : speckle_pd_t_u ] ) . at these points @xmath60 diverges in the pure limit of @xmath168 , which reveals that the compressibility is grows with the inverse of @xmath97 and only persists in the limit @xmath199 if the density @xmath135 is integer . at low temperature and fixed density , we find that an increase in @xmath97 can actually drive the system from an insulating into a condensed state within a certain window of @xmath20 , as shown in fig . [
fig : speckle_pd_delta_u_low_t ] and fig .
[ fig : speckle_kappa_delta_u_low_t ] . within a very small window of @xmath20
, the system may be driven through an additional sequence of bg and sf phases .
( color online ) low temperature @xmath200 phase diagram showing the disorder - averaged mfp @xmath132 for speckle disorder at constant filling @xmath129 in three dimensions .
multiple reentrant behavior can be seen within a small window of @xmath20 .
the red line at @xmath114 indicates the presence of mi / normal state , for any @xmath130 the insulator is a bg . ] on the @xmath114 line , the usual sf - insulator transition occurs , whereas along @xmath97 for small values of @xmath10 , disorder only suppresses the sf slightly up to reasonably large values of @xmath136 . for even larger disorder values
, the system will eventually undergo the transition into an ( color online ) compressibility @xmath60 in units of @xmath201 at @xmath200 for speckle disorder at constant filling @xmath129 in three dimensions ( same parameters as in fig .
[ fig : speckle_pd_delta_u_low_t ] ) .
the incompressible normal phase only exists on the line @xmath202 , but the region where @xmath60 is exponentially small in the insulator extends linearly with @xmath20 .
the white line is the sf - insulator phase border , where @xmath162 becomes zero .
the compressibility along the blue dashed line can be understood as the low temperature case and compared to @xmath60 in the lower two plots of fig .
[ fig : kappa_of_delta_speckle ] , where the sharp features have been washed out by temperature , but are still pronounced peaks . ] anderson localized state . in comparison to the corresponding phase diagram for box disorder fig
. [ box_phase_diags_n1 ] , the extruding superfluid lobe appears at a considerably smaller values of @xmath203 for speckle disorder ( @xmath204 for box disorder , not the different @xmath97-axis scale ) .
however , it should be noted that the measure @xmath97 is not the same ( in a statistical sense ) for the two disorder types : whereas for speckle disorder @xmath97 is also the standard deviation ( std ) , in the case of box disorder the std is only @xmath205 , corresponding to weaker disorder for the same @xmath97 . examining the position of the superfluid lobe in units of the standard deviation , actually reveals that it appears at slightly smaller disorder strengths for box disorder .
( color online ) main background phase diagram as in fig .
[ fig : speckle_pd_delta_u_low_t ] , but in the experimentally relevant high temperature regime @xmath170 at @xmath129 , where disorder can not induce condensation .
insets : disorder averaged mfp ( a ) , compressibility ( b ) and local dos @xmath84 in the @xmath64-@xmath157-plane at the same temperature @xmath206 for @xmath160 . in this regime
most interesting structure has been washed out by thermal fluctuations and disorder and an increase in either @xmath10 , @xmath97 or @xmath85 _ always _ suppresses condensation . ] in the vicinity of the phase border , @xmath60 is generally larger in the sf state than in the neighboring insulator , as a shift in @xmath7 leads to a locally continuous shift in density . on the lower side of the lobe at disorder strengths @xmath207 , the compressibility jumps across the phase border ( i.e. second order transition ) . at higher values of @xmath97 ,
@xmath60 is almost unaffected by the disappearance of the condensate , i.e. the compressibility is almost exclusively induced by disorder . in the insulating state
the ensemble of mean - field states determined by smft can not depend on @xmath6 ( since different sites are only coupled via @xmath42 ) .
therefore any physical quantity , such as @xmath60 , can only be a function of @xmath193 in this regime ( reflected in the radial structure of @xmath60 in the insulator ) . in the high temperature regime
relevant for current experiments ( see app .
[ sec : temp ] for an approximation of @xmath85 ) , the lobe in the @xmath208plane at fixed density has vanished completely and an increase in disorder always counteracts the condensate formation .
this might explain the recent finding , that no reentrant behavior or disorder - induced condensation was observed in experiments so far @xcite .
furthermore , we calculated the @xmath209phase diagram for this temperature and disorder regime , where the lobe structure is also fully washed out and the system is dominated by thermal fluctuations , as seen in the insets of fig . [
fig : speckle_pd_delta_u_high_t ] . we therefore conclude that an upper critical temperature , which may depend on the filling , exists for the occurrence of reentrant superfluidity .
our temperature estimation suggests that @xmath85 is too high in current experiments to observe this effect , which agrees with recent experimental evidence @xcite .
in addition to diagonal on - site energy disorder , it is possible to incorporate off - diagonal hopping disorder @xcite into the smft formalism . in this case
, the local hopping energy @xmath210 between site @xmath4 and @xmath211 becomes a random variable , each described by a distribution @xmath212 , which we assume to be independent of the on - site energies @xmath8 .
this leads to the bh hamiltonian with on - site , as well as hopping disorder @xmath213 the pdf @xmath212 has been calculated for a speckle disorder potential @xcite using imaginary time evolution .
in contrast to diagonal disorder ( on - site energy or interaction ) , where the fluctuations are incorporated into the conditional pdf ( [ conditional_pdf ] ) before the iteration procedure of the smft self - consistency equations , the hopping disorder acts as an additional source of fluctuations during this iteration procedure , methodically entering at a different point in the method .
the corresponding mean - field hamiltonian @xmath214 , ( analogous to ( [ eff_ss_hamiltonian ] ) ) now depends on the rest of the system only through the new random variable @xmath215 where both @xmath216 and @xmath26 are random variables , the latter again being assumed to be distributed according to @xmath19 .
it is therefore convenient to introduce the an intermediate random variable @xmath217 , distributed according to the pdf @xmath218 .
as explained in app .
[ pdf_of_products ] , this new pdf can be expressed explicitly in terms of the pdf s @xmath19 and @xmath212 as @xmath219 numerically allowing the use of the fft algorithm on a suitable grid .
once @xmath218 is known , the pdf for the random variable @xmath220 can subsequently be calculated by @xmath221^z \ , e^{-i t \eta } \label{qdistribution_hop},\ ] ] where @xmath222 is the characteristic function of @xmath218 . for the numerical computation of the previous two
, the fft algorithm can be used . the conditional pdf ( [ conditional_pdf ] ) @xmath223 , incorporating the effect of any diagonal ( here : on - site ) disorder remains unchanged under the inclusion of hopping disorder and is calculated before the smft iteration procedure in full analogy to the previous case for the relevant on - site disorder type @xmath9 . the final self - consistency equation , closing the iteration procedure is also left unchanged to the previous case ( [ selfconsistency ] ) @xmath224 except that the @xmath30 entering is calculated from ( [ qdistribution_hop ] ) .
of course the case of pure on - site disorder with a constant hopping energy @xmath225 can be obtained as a limit of this extension by setting @xmath226 , causing ( [ p_phi_conv ] ) and ( [ qdistribution_hop ] ) to reduce to ( [ def_q ] ) .
the numerical iteration procedure is carried out analogously to sec .
[ sec : numerical_sol ] , except that only a single equidistant grid , as restricted by ( [ p_phi_conv ] ) , can be used .
this limits the numerically obtained precision of @xmath19 at small values of @xmath42 . in our calculation
we use the appropriately scaled distribution @xmath212 , matching the on - site disorder strength , as obtained by zhou and ceperley@xcite for a lattice depth of @xmath227 and at a speckle disorder strength of @xmath228 , where we assume that the standard deviation in @xmath212 is proportional to the disorder intensity @xmath229 ( as motivated by their analysis ) .
since a change in the disorder strength does not change the most probable value of the distribution@xcite @xmath212 , we model the hopping disorder distribution to have its most likely value at the given value @xmath230 , with the width being independent thereof .
if any weight lies at values of negative @xmath6 , this is set to zero and the distribution is subsequently renormalized , however this is only relevant for points deep in the insulator and does therefore not influence the results in any way .
however , this only occurs at very small values of @xmath230 deep in the insulator , making this formal alteration of @xmath212 irrelevant to the result .
( color online ) diagrams showing the effect of _ hopping disorder _ at @xmath161 and an on - site speckle disorder strength @xmath231 .
main 3d figure : disorder - averaged mfp @xmath42 for speckle on - site and experimentally corresponding hopping disorder . for orientation purposes , the mf phase boundary for the pure system
is plotted as a thin dashed white line .
the value @xmath82 refers to the most likely value of the hopping disorder distribution @xmath212 , whereas the width of the distribution @xmath212 is constant .
upper left inset : comparison of the order parameter at fixed @xmath232 with and without hopping disorder , clearly demonstrating that the addition of hopping disorder stabilizes the sf phase and leads to a lower critical value of @xmath6 in this regime .
upper right inset : subsequent comparison of phase borders in the @xmath64-@xmath82-plane . ] as shown in the low temperature phase diagram in the @xmath82-@xmath64-plane ( fig.[fig : speckle_hopping_pd_jz_mu ] ) , the additional inclusion of hopping disorder leads to a stabilization of the sf phase and a small shift in the phase boundary . for disorder distributions in the typical experimentally relevant parameter regime ,
the standard deviation of the hopping parameter distribution @xmath212 is three orders of magnitude smaller than the standard deviation of the on - site energy distribution @xmath9 ( as found in @xcite , specifically for the distributions used here @xmath233 ) .
this explains the minor , but clearly resolved modification of the phase boundaries in contrast to the pure on - site disordered case , shown in the upper right inset of fig .
[ fig : speckle_hopping_pd_jz_mu ] .
the initial temperature in the trap , prior to the optical lattice ramp - up , can be determined from the expansion profile of the cloud .
if the ramp - up of the optical lattice is performed adiabatically , the entropy of the system is conserved and can actually lead to a cooling of the atoms .
the initial entropy of a weakly interacting cloud in the trap using bogoliubov theory leads to the expression @xcite @xmath234 \right).\ ] ] after ramping the lattice to a sufficiently high intensity , the entropy can be calculate up to first order ( neglecting terms containing @xmath6 ) @xcite @xmath235,\ ] ] where @xmath236 is the number of sites and @xmath237 is the grand canonical partition function . equating these two expressions for a sufficiently high initial temperature
@xmath238 leads to the relation @xcite @xmath239 for @xmath240rb in an optical with a wave length of @xmath241 and intensity @xmath242 , the disorder - averaged interaction constant is @xmath243 . as a typical , conservative estimate of the initial temperature before the lattice ramp - up in current experiments
@xcite we use the value @xmath244 , for which the relation ( [ final_temp ] ) predicts a final temperature of @xmath245 after the ramp - up . to compare the results obtained via smft to experimental data on a quantitative level , we performed a lda+smft calculation to incorporate the effect of the trapping potential .
two effects of the lattice laser beams are taken into account : the red - shifted lattice laser beam with a gaussian profile and beam width @xmath246 leads to an attractive potential via the ac stark effect .
furthermore , the local energies of states within a localized wannier basis are also increased in regions of high intensity , leading to the renormalized effective lower trapping frequency within an harmonic approximation @xcite . @xmath247 with the addition of an external magnetic trap with trapping frequency @xmath248 , the total trapping frequency , which the atoms are exposed to is given by @xmath249 for every fixed value of the lattice height @xmath250 , @xmath6 and @xmath10 are extracted from the single particle wannier function ( beyond the approximation for deep lattices @xcite ) , the total trapping frequency ( [ total_trap_freq ] ) is calculated and the local effective trap energies are assigned on a sufficiently large @xmath251 lattice .
the chemical potential @xmath7 is adjusted using ridder s method @xcite to obtain a specified total particle number .
the condensate fraction , depicted in fig .
[ fig : lda_fc ] , is subsequently averaged over the local values obtained by smft , weighted by the respective density .
( color online ) the condensate fraction as a function of the lattice strength @xmath250 for various total particle numbers in a harmonic trap , calculated within lda and smft ( connected dotted lines ) .
the calculations were performed at fixed temperature @xmath170 and disorder strength @xmath149 . ]
we used the following values for the experimental parameters : @xmath252 , the lattice laser wavelength was set to @xmath253 , and the magnetic trapping frequency @xmath254 .
in contrast to the behavior of the order parameter at the transition point , the condensate fraction does not follow a power law decay in the finite trap , but is smeared out a the transition point , as superfluid regions in the trap decay in size with increasing @xmath250 .
we have described the stochastic mean - field theory in detail on a methodological level , and extend it to incorporate finite temperature effects .
subsequently , we have applied it to ultracold atoms in an optical lattice with uncorrelated on - site box disorder distribution and discussed the intricate interplay between interaction , tunneling energy , disorder , filling and finite temperature effects .
furthermore , we have presented , to the best of our knowledge , the first quantitative theoretical calculations for speckle disorder , which leads to a qualitatively different phase diagram than for box disorder and are of immediate experimental relevance @xcite . for this case
, we have discussed the characteristic features of the various phases and presented phase diagrams , both at fixed chemical potential and at fixed density . below a critical temperature , we find disorder - induced condensation and multiple reentrant behavior , both for box and speckle disorder .
the temperature in recent experiments @xcite is estimated and found to be too high yet to observe disorder - induced condensation .
we also find that including hopping disorder in addition to local on - site disorder for a realistic distribution of speckle parameters enhances the insulator and jumps in the order parameter within the sf phase indicate a series of transitions .
an lda+smft calculation has been performed to incorporate the effects of an external trap for on - site speckle disorder .
we thank i. bloch , l. carr , b. demarco , e. demler , a. pelster , d. semmler , m. snoek , h. stoof and w. zwerger for useful discussions .
this work was supported by the german science foundation ( dfg ) via forschergruppe for 801 .
ub and rt acknowledge support by the studienstiftung des deutschen volkes .
calculations were performed at the center for scientific computing at the university of frankfurt / main and the tkm computation cluster at the university of karlsruhe .
[ multi_grid_fft ] to employ the fft algorithm an equidistant grid is required , which is not compatible with having a very high resolution at small values of @xmath42 ( @xmath255 ) to capture the behavior in the vicinity of the sf - insulator transition , as well as simultaneously correctly describing the distribution at large values ( @xmath256 ) . to circumvent this problem
, we use a superposition of equidistant grids , which enables us to use the fft algorithm on each of these grids individually ( see fig . [
fig : grids ] ) .
illustration of the multi - grid procedure . ]
this procedure relies on the following property of the convolution ( [ def_q ] ) : the @xmath22-fold convolution of the truncated function @xmath257 with @xmath258 for @xmath259 and @xmath260 is identical to the @xmath22-fold convolution of @xmath19 up to an easily determinable normalization constant on the interval @xmath261 $ ] . here
@xmath262 is the largest value of the grid , if this is finite ( as in the discretized numerical case ) . in our calculations
the number of grids used ( typically @xmath263 ) with @xmath264 points per grid is adjusted dynamically within the iteration procedure , depending on the position of the most likely value of @xmath19 and the convergence properties .
we found the numerically most efficient method for calculating and tabulating the two - dimensional cdf @xmath265 to be the following : for every fixed value of @xmath41 , consider the function ( [ def_g ] ) @xmath266 on the interval @xmath267 $ ] .
we define the local minima and maxima of this function ( including the end points ) in increasing order as @xmath268 and @xmath269 .
furthermore the @xmath135-th monotonically increasing and decreasing function on the restricted interval , provided that this interval exists , is denoted by @xmath270 where @xmath271 by construction , these functions are invertible in @xmath88 within the defined range , allowing us to introduce the functions @xmath272 in terms of which the cdf can be written as a superposition @xmath273
including hopping disorder into the smft leads to the task of having to calculate the probability distribution for the newly defined random variable @xmath217 , if the pdfs for @xmath6 and @xmath42 are known functions @xmath212 and @xmath19 .
this can be done by taking the logarithm @xmath274 and using the convolution theorem for the pdfs @xmath275 ( analogously for @xmath276 ) of the random variables @xmath277 and @xmath278 .
the resulting distribution for @xmath279 can be denoted in compact form as @xmath280 using the fft - algorithm for the numerical computation of this operation leads to a vast increase in performance , however it requires the functions @xmath281 and @xmath276 to be interpolated on an equidistant grid .
in the special case with @xmath61 , the disorder integration can be performed explicitly and expressions for the density and the compressibility can be found . using partial integration and the property @xmath282 for the on - site probability distribution , the disorder averaged density at finite temperature
can be expressed as @xmath283 with @xmath284 . in the specific case of box disorder with @xmath285\ ] ] the disorder averaged density takes on the form @xmath286\ ] ] and the compressibility @xmath111 can be directly evaluated @xmath287 for speckle disorder , on the other hand , one has @xmath288 and the compressibility takes on the form @xmath173 . \end{split}\end{aligned}\ ] ]
to obtain the local dos in an insulating phase , we calculate the single particle green s functions @xmath289 \end{split}\end{aligned}\ ] ] for local fock states at finite temperature , where @xmath290 is the on - site particle annihilation operator in the heisenberg representation .
the fourier transformed green s function can be calculated and takes on the form @xmath291 , \end{split}\end{aligned}\ ] ] where @xmath292 is the local partition function .
this is related to the single particle dos by @xmath293 . \end{split}\end{aligned}\ ] ] averaging over the on - site energy distribution for speckle disorder leads to the final expression @xmath294 , \end{split}\end{aligned}\ ] ] whereas for a box disorder distribution one obtains @xcite @xmath295 .
\end{split}\end{aligned}\ ] ]
instead of performing the thermal average before constructing the conditional probability distribution for @xmath42 , the smft furthermore also allows to explicitly facilitate the thermal fluctuations of @xmath42 in the probability distribution .
we have not yet performed the numerical calculation , but will outline the procedure .
this can be formulated in the following way : for fixed external parameter @xmath41 ( hence _ conditional _ probability distribution ) the on - site energy is randomly drawn from @xmath9 and the resulting single site hamiltonian is diagonalized .
for the different eigenstates @xmath297 with eigenenergies @xmath298 the respective expectation value @xmath299 is calculated . within the grand canonical ensemble
, this expectation value has the probability @xmath300^{-1 } \ , e^{-\beta e_i}$ ] of occurring .
this probability is uncorrelated to the probability of a certain on - site energy @xmath35 occurring . expressing this mathematically
, we obtain the explicit formula for the conditional probability distribution @xmath301^{-1 } \\ & \times \sum_{i=1}^\infty e^{-\beta e_i(\eta,\epsilon ) } \ , \theta\left[\psi- { \left\langle i(\epsilon,\eta)\right| } b { \left|i(\epsilon,\eta)\right\rangle } \right ] \end{split}\ ] ] equivalently on a formal level , one may work with the cumulative conditional probability function @xmath302^{-1 } \\ & \times \sum_{i=1}^\infty e^{-\beta e_i(\eta,\epsilon ) } \ , \theta\left[\psi- g_i(\mu-\epsilon,\eta ) \right ] \end{split}\ ] ] with @xmath303 to evaluate the expression @xmath304 we perform a variable substitution @xmath305 and subsequently @xmath306 using the central theorem of calculus one can explicitly construct @xmath307 since @xmath308 is a non - negative function the monotonously increasing function @xmath309 is invertible in the relevant range .
the conditional cumulative density function can then be explicitly calculated from @xmath310 as in the smft for @xmath13 , the self - consistency condition reads @xmath311 where @xmath30 is the @xmath312fold convolved and rescaled function of @xmath19 with the random variable @xmath313 .
the conditional cumulative density @xmath265 now also contains the thermal fluctuations explicitly in the distribution ( i.e. they are not averaged over within the self - consistency loop ) , so the subsequent determination of @xmath19 is identical to the previous cases . | we discuss the stochastic mean - field theory ( smft ) method which is a new approach for describing disordered bose systems in the thermodynamic limit including localization and dimensional effects .
we explicate the method in detail and apply it to the disordered bose - hubbard model at finite temperature , with on - site box disorder , as well as experimentally relevant unbounded speckle disorder .
we find that disorder - induced condensation and reentrant behavior at constant filling are only possible at low temperatures , beyond the reach of current experiments @xcite .
including off - diagonal hopping disorder as well , we investigate its effect on the phase diagram in addition to pure on - site disorder . to make contact to present experiments on a quantitative level
, we also combine smft with an lda approach and obtain the condensate fraction in the presence of an external trapping potential . |
in school we all dealt , in one way or another , with solids such as prisms and pyramids , but maybe also with bipyramids .
the aim of this paper is to generalize these solids as different products of abstract polytopes , and study their symmetry and combinatorial properties .
prisms , pyramids and bipyramids over polygons ( see figure [ pentagons ] ) can be seen as a product of a polygon by either a segment or a point . however , these are three different kinds of products . while prisms are the cartesian product of a segment with a polygon , pyramids are the join product of a point with a polygon and bipyramids are the direct product of a segment with a polygon . in the theory of convex polytopes the generalization of these three notions
are the cartesian product , the join product and the direct sum , respectively ( @xcite ) . given two convex polytopes @xmath0 and @xmath1 ,
their products are defined as follows .
the join of @xmath2 and @xmath3 is obtained by embedding @xmath2 and @xmath3 in disjoint affine subspaces of @xmath4 and taking the convex hull of their vertices .
for example , for each @xmath5 , a @xmath6-simplex can be seen as the join of a point and a @xmath7-simplex ( figure [ joinpowersofapoint ] ) .
-simplex is the join product of a point and a @xmath7-simplex.,width=377 ] the cartesian product of @xmath2 and @xmath3 is obtained by taking the convex hull of @xmath8 in @xmath9 .
the classical example in this case , is to see a @xmath6-cube as the cartesian product of an edge -or line segment- with a @xmath7-cube ( as in figure [ cartesianpowersofaedge ] ) .
-cube is the cartesian product of an edge and a @xmath7-cube.,width=377 ] the direct sum of @xmath2 and @xmath3 is slightly more complicated to state .
we first require that @xmath2 and @xmath3 contain in their relative interiors the origins of @xmath10 and @xmath11 , respectively .
then the direct sum is the convex hull of all the points of the form @xmath12 and @xmath13 , where @xmath14 and @xmath15 .
for example , cross polytopes can be generated in this way .
note that as where in the join product and cartesian product of convex polytopes @xmath2 and @xmath3 , every face of @xmath2 and of @xmath3 is again a face of the product , for the direct sum this is no longer the case . on the other hand , for both the join product and the direct sum , the vertices of the product is the union of the vertices of both polytopes , while the cartesian product of two polytopes , in general , has more vertices .
it is straightforward to see that the only convex polyhedra ( or convex 3-polytopes ) that arise as one of these products are precisely the prisms , the pyramids and the bipyramids over polygons .
it is also well - know that , in @xmath16 , the product of two orthogonal circles @xmath17 is precisely the _ flat torus _ ( also known as the clifford torus , @xcite ) . if we place @xmath18 points on each of the circles , evenly spaced , we obtain @xmath18 congruent line segments on each circle .
then , take the cartesian product of each point of each @xmath19 with line segment of the other @xmath19 .
what you obtain is a tessellation of the flat torus by squares ( see figure [ torus ] ) .
hence , some maps on surfaces can also be seen as products of polygons .
abstract polytopes generalize the ( face lattice ) of convex polytopes .
moreover , they also generalize non - degenerated maps .
hence , it is natural to generalise the four products described above and define them for abstract polytopes , and we do so in section [ sec : allproducts ] . as we show , the four products are closed for abstract polytopes , meaning that the product of two abstract polytopes is again an abstract polytope ( under any of the four products ) .
we shall also study , for each product , which polytopes are trivial , in the sense that the product of them with any polytope @xmath2 is simply @xmath2 . with that in mind
, it is natural to say that a polytope is prime with respect to a given product , if it can not be decomposed as the product of non - trivial polytopes .
we show a unique prime factorization theorem and use it to investigate the structure of the automorphism group of a product .
theorem a summarizes the main results of sections [ sec : allproducts ] , [ sec : fact ] and [ sec : auto ] .
let @xmath2 and @xmath3 be two abstract polytopes and @xmath20 be a product of polytopes ( either the direct sum , the join , cartesian or topological product ) .
then , 1 . the product @xmath21 is an abstract polytope . in particular
, @xmath22 is also an abstract polytope .
the polytope @xmath2 can be uniquely factorized as a product of prime polytopes .
3 . if @xmath23 , where the @xmath24 are distinct prime polytope with respect to @xmath20 , then @xmath25 4 .
if @xmath2 is decomposed as above , where @xmath24 is a @xmath26-orbit polytope , then @xmath2 is a @xmath27-orbit polytope , with @xmath28 where if @xmath20 is the join product , then @xmath29 is the rank of @xmath24 , if @xmath20 is either the cartesian product or the direct sum , then @xmath30 is the rank of @xmath24 , and if @xmath20 is the topological product , then @xmath31 is the rank of @xmath24 as a corollary of part @xmath32 of theorem a , for each product , we also obtained the families of regular and two - orbit polytopes that are not prime .
the monodromy group of a polytope ( either convex or abstract ) encodes all the combinatorial information of the polytope .
it was first studied by hartley in @xcite and he used it to construct regular covers of ( non - regular ) polytopes .
it is well - know that the monodromy group of a regular polytope is isomorphic to its automorphism group .
however , little is known about monodromy groups of non - regular polytopes . in the last decade
, there has been an effort to understand these groups ( see for example @xcite , @xcite , @xcite ) . in particular in @xcite
hartely et al .
study the monodromy group of the prism oven an @xmath18-gon and compute it , in terms of generators and relations .
moreover , in @xcite , berman et al . study that of the pyramid over an @xmath18-gon and show that it is an extension of the symmetric group @xmath33 by a cyclic group which sometimes splits ( and determine when ) .
we show that products of polytopes are useful to understand monodromy groups of some non - regular polytopes .
the results of monodromy groups of polytopes are summarized in the following theorem let @xmath34 be polytopes of ranks @xmath35 , respectively and let @xmath20 be a product of polytopes .
let @xmath36 .
then , 1 .
the monodromy group @xmath37 is an extension of @xmath38 , where @xmath39 , and @xmath40 , if @xmath41 , @xmath42 if @xmath43 and @xmath44 if @xmath45 .
the extension of @xmath46 splits ( at least ) in the following cases : * if @xmath2 is the prism ( or the bipyramid ) over an @xmath18-gon . in this case
@xmath47 , where @xmath48 is an extension @xmath49 by @xmath50 ( @xmath51 ) ; moreover , this extension splits whenever @xmath18 in not congruent to @xmath52 modulo @xmath53 , in which case @xmath54 . * if @xmath2 is the prism ( or the bipyramid ) over a @xmath55-polytope having the property that all its vertex figures ( faces ) are isomorphic to an @xmath18-gon , with @xmath18 not congruent to @xmath52 modulo @xmath56 ; * if @xmath45 and each @xmath24 has rank 2 . by @xcite , we already know that the monodromy group of a pyramid over an @xmath18-gon is an extension of @xmath33 by @xmath57 , where @xmath58 , and that such extension splits whenever @xmath18 is not congruent to @xmath52 modulo @xmath56 . our techniques to show theorem b can also be used to show this result .
abstract polytopes generalise the ( face - lattice ) of classical polytopes as combinatorial structures . in this chapter
we review the basic theory of abstract polytopes and refer the reader to @xcite for a detail exposition of the subject .
an _ ( abstract ) @xmath18-polytope _ ( or _ ( abstract ) polytope of rank @xmath18 _ ) @xmath2 is a partially ordered set whose elements are called _ faces _ that satisfies the following properties .
it contains a minimum face @xmath59 and a maximum face @xmath60 .
these two faces are the _ improper _ faces of @xmath2 ; all other faces are said to be _
proper_. there is a rank function from @xmath2 to the set @xmath61 such that @xmath62 and @xmath63 .
the faces of rank @xmath64 are called _
@xmath64-faces _ , the @xmath52-faces are called _ vertices _ , the @xmath65-faces are called _ edges _ and the @xmath7-faces are called _ facets_.
every maximal totally ordered subset ( called _ flag _ ) contains precisely @xmath66 elements including @xmath59 and @xmath60 . if @xmath67 is a flag of @xmath2 we shall often denote by @xmath68 the @xmath64-face of @xmath67 . for incident
faces @xmath69 , we define the _ section _
@xmath70 , and when convenient , we identify the section @xmath71 with the face @xmath72 itself in @xmath73 .
the section @xmath74 , when @xmath75 is a vertex , is called the _ vertex - figure _ of @xmath73 at @xmath75 , and if @xmath76 is a face of rank @xmath77 , then @xmath78 is a _ co - face _ of @xmath2 .
all sections @xmath79 of @xmath2 are by themselves posets with a rank function , minimum and maximum faces and satisfy that all their maximal chains have the same number of elements . a section @xmath79 is said to be _ connected _ , if @xmath80 or if whenever @xmath81 , with @xmath82 , there exists a sequence of faces @xmath83 such that @xmath84 and either @xmath85 or @xmath86 , for every @xmath87 .
we further ask that @xmath2 is _ strong connectivity _ , meaning that all the sections of @xmath2 , including itself , are connected .
the last condition that @xmath2 should satisfy to be a polytope , known as the _ diamond condition _ , is the following .
if @xmath72 and @xmath88 are incident faces such that @xmath89 , then there exist precisely two faces @xmath90 and @xmath91 such that @xmath92 .
this property implies that for any flag @xmath67 and any @xmath93 there exists a unique flag @xmath94 differing from @xmath67 only in the @xmath64-face .
the flag @xmath94 is called the _
@xmath64-adjacent flag _ of @xmath67 . up to isomorphism
, there is a unique @xmath18-polytope for @xmath95 .
polygons ( including infinite ones ) are @xmath96-polytopes and no - degenerated maps are 3-polytopes . in general
, a convex @xmath6-polytope can be regarded as an abstract @xmath6-polytope .
it is not difficult to see that the diamond condition implies that the incidence structure consisting of all vertices and edges of a polytope @xmath2 , together with the incidence given in @xmath2 is a graph .
we shall refer to this graph as the _
@xmath65-skeleton _ of @xmath2 .
an _ automorphism _ of a polytope is an order - preserving permutation of its faces .
we denote the group of automorphisms of @xmath2 by @xmath97 .
it is straightforward to see that @xmath98 acts on the set of flags , denoted by @xmath99 , in the natural way .
moreover , the strong connectivity of @xmath2 implies that such action is free ( or semi - regular ) .
an @xmath18-polytope @xmath73 is said to be _ regular _ whenever @xmath97 acts transitively on the flags .
we say that @xmath2 is a _
@xmath27-orbit polytope _ if @xmath97 has precisely @xmath27 orbits on @xmath99 .
( hence , regular polytopes and 1-orbit polytopes are the same . ) given a polytope @xmath2 , one can define the dual of @xmath2 , denoted by @xmath100 , as the poset whose elements coincide with the elements of @xmath2 , but the order is reversed . in other words
, @xmath100 is the dual of @xmath2 if there exists a bijection @xmath101 that reverses the order .
note that @xmath102 and that @xmath103 .
the _ monodromy group _ @xmath104 of an @xmath18-polytope @xmath2 is the subgroup of the permutations of the set of flags @xmath99 that is generated by the permutations @xmath105 ( see @xcite ) .
the elements of the monodromy group are , in general , far from being automorphisms of the polytope . by the connectivity of @xmath2 , @xmath37 is transitive on @xmath99 .
one can think of the generators of the monodromy group as the instructions to assemble the flags of the polytope .
in fact the monodromy group possesses all the combinatorial information of the polytope .
given @xmath106 , @xmath107 and @xmath108 it is straightforward to see that @xmath109 . the generators @xmath110 of @xmath37 are involutions and satisfy , at least , the relations @xmath111 whenever @xmath112
whenever @xmath2 is a regular polytope , its monodromy group and its automorphism group are isomorphic .
however , in other cases little is known about the structure of the monodromy group of a polytope ( see @xcite for further discussion on the subject ) . given a poset @xmath2 and @xmath113 , we shall say that _
@xmath72 is covered by @xmath88 _ if @xmath114 and there exists no @xmath115 such that @xmath116 .
in particular , if @xmath2 is a polytope , then a face @xmath72 is covered by a face @xmath88 whenever @xmath114 and @xmath117 .
hasse diagram _ of the poset @xmath2 , denoted by @xmath118 is the directed graph whose vertices are the elements of @xmath2 and there is an arc from a face @xmath88 to a face @xmath72 whenever @xmath72 is covered by @xmath88 .
note that if @xmath2 is a polytope , then the digraph @xmath118 has one sink , one source , is acyclic and directed paths of maximal length have @xmath66 vertices .
moreover , @xmath119 in @xmath2 if and only if there is a directed path from @xmath88 to @xmath72 in @xmath118 .
therefore if @xmath2 and @xmath3 are two polytopes such that there exists and isomorphism between @xmath118 and @xmath120 , then it induces an isomorphism between @xmath2 and @xmath3 ( and viceversa : isomorphisms between @xmath2 and @xmath3 induce isomorphisms between their hasse diagrams ) .
note further that @xmath121 .
a poset @xmath2 is said to be _ discrete _ if the transitive closure of the hasse diagram @xmath118 is @xmath2 itself .
for example @xmath122 is a discrete poset , while @xmath123 is not .
all abstract polytopes are discrete posets .
hence , in this paper , unless otherwise indicated , all posets are discrete .
as we have seen before , one can identify an abstract polytope with its hasse diagram . for the purpose of this paper it shall prove helpful to often think of abstract polytopes as directed graphs ( with the induced properties ) .
thus , we study here some properties about products of posets and digraphs .
given posets @xmath24 with @xmath124 , the _ ( cardinal ) product _ , @xmath125 is the ordered set on their cartesian product , with component - wise order . denoting by @xmath126 the product of two posets @xmath2 and @xmath3 ,
it is then straightforward that @xmath127 and that , if @xmath128 is yet another poset , then @xmath129 . we denote by @xmath130 to the product of @xmath27 copies of @xmath2 . given @xmath131 , with @xmath132 , the set @xmath133 is the _ closed interval _ between @xmath72 and @xmath88 .
similarly , @xmath134 is said to be an _
open interval_. ( hence , sections of a polytope are in fact closed intervals of the poset . ) if the poset @xmath2 does not have a minimum or a maximum , the sets @xmath135 , @xmath136 , @xmath137 and @xmath138 are also said to be ( closed / open , resp . ) intervals of @xmath2 .
we say that a poset @xmath2 is _ factorable _ if there exist non - trivial posets @xmath139 and @xmath140 such that @xmath141 and that @xmath2 is _ prime _ if no such factorization exists
. in @xcite hashimoto shows that if we have two proper factorizations of a poset @xmath2 , then there exists another proper factorization of @xmath2 that is a refinement of the two original ones .
hashimoto then uses this result to show the following theorem .
[ uniquefactposet ] every connected poset has a unique prime factorisation ( up to isomorphism ) . in this context
, a poset @xmath2 is said to be _ connected _ if for any two elements @xmath113 , there exists a sequence @xmath142 such that @xmath143 or @xmath144 for every @xmath145 .
recall now that a graph @xmath88 is said to be connected if for any two vertices @xmath146 and @xmath147 , there is a @xmath146-@xmath147 path in @xmath88 .
a digraph is said to be _ weakly connected _ if its underlying graph ( where the arcs are replaced by edges ) is connected .
a discrete poset is connected if and only if is its hasse diagram is weakly connected .
given two digraphs @xmath148 and @xmath149 , the _ cartesian product _ of @xmath150 and @xmath151 is a digraph @xmath152 whose vertex set is @xmath153 , and such that there is an arc @xmath154 if @xmath155 and @xmath156 or if @xmath157 and @xmath158 .
it was shown in @xcite that the hasse diagram of a product of orders is a product of hasse diagrams .
in fact we have the following proposition .
let @xmath24 , with @xmath124 , be a family of posets . if @xmath159 , then @xmath160 .
moreover , in @xcite , walker showed that if @xmath2 is a poset such that @xmath161 , for some digraphs @xmath162 , then there exist posets @xmath24 , such that @xmath163 for each @xmath124 and @xmath164 . given a digraph @xmath165 , if there exists digraphs @xmath150 and
@xmath151 such that @xmath166 , where @xmath167 , then we say that @xmath168 is _ cartesian - factorable _ ( or , simply , _
factorable _ ) and that @xmath166 is a _ proper factorization _ of @xmath168 .
if no such factorization exists , we shall say that @xmath168 is _
prime_. hence , a poset @xmath2 is prime if and only if its hasse diagram @xmath118 is a prime digraph .
as pointed out in the introduction , geometrically , there are several kinds of products of polytopes . in this section
we define each of them as products of abstract polytopes .
we shall see that although the different products that we define have different geometric interpretations , they can all be expressed in terms of cardinal products of posets .
geometrically the most natural product might be the cartesian one , however when considering abstract polytopes the natural product arrises from the product of posets .
we therefore start by studying such product of polytopes . given two polytopes @xmath2 and @xmath3 , the _
join product _ of @xmath2 and @xmath3 , denoted @xmath169 , is defined as the set @xmath170 where the order is given by @xmath171 in other words , @xmath169 is simply the cardinal product @xmath126 of the posets @xmath2 and @xmath3 .
( we have changed the notation as we shall only use the join product @xmath169 when both @xmath2 and @xmath3 are polytopes , while we shall keep referring to the product @xmath172 as the product of any two posets . )
it is therefore straightforward to see that @xmath169 is indeed a poset .
moreover , @xmath173 and if @xmath128 is another abstract polytope , then @xmath174 .
hence , for every natural number @xmath27 , @xmath130 denoted the join product of @xmath2 , @xmath27-times .
observe further that a section of @xmath169 is the join of a section of @xmath2 and a section of @xmath3 .
that is , let @xmath2 and @xmath3 be two polytopes and consider the join @xmath169 .
let @xmath175 , @xmath176 such that @xmath177 and @xmath178 .
then @xmath179 if @xmath2 and @xmath3 are two polytopes of ranks @xmath18 and @xmath180 respectively , then the rank functions of @xmath2 and @xmath3 naturally induce a rank function on @xmath169 , namely , @xmath181 hence , the rank function of @xmath169 has range from @xmath182 to @xmath183 , and therefore @xmath184 shall have rank @xmath183 .
it is not difficult to see that , if @xmath185 and @xmath186 denote the minimal faces of @xmath2 and @xmath3 respectively , then the vertices of @xmath169 are of the form @xmath187 or @xmath188 , where @xmath147 is a vertex of @xmath3 and @xmath146 is a vertex of @xmath2 . hence , the vertices of @xmath169 are in bijection with the union of the vertices of @xmath2 and @xmath3 . in general , for @xmath189 , if @xmath72 is an @xmath64-face of either @xmath2 ( or @xmath3 ) , then @xmath190 ( or @xmath191 ) is an @xmath64-face of @xmath169 , but these are not all the @xmath64-faces of @xmath169 .
let @xmath2 and @xmath3 be two polytopes of ranks @xmath18 and @xmath180 , respectively .
then @xmath184 is a polytope of rank @xmath183 .
first note that , if @xmath59 and @xmath192 , and @xmath193 and @xmath194 are the minimal and maximal faces of @xmath2 and @xmath3 respectively , then @xmath195 is the minimal face of @xmath169 , while @xmath196 is its maximal face . given two elements @xmath197 such that @xmath198 , there are flags @xmath67 and @xmath199 of @xmath2 and @xmath3 , respectively , such that @xmath200 , and @xmath201 .
it is straightforward to see that the set @xmath202 ( which is a subset of @xmath184 ) , is a chain of the order @xmath184 that has one element of each rank from @xmath203 to @xmath204 .
this implies that every flag of @xmath184 has exactly @xmath205 elements , including @xmath206 and @xmath196 .
we now turn our attention to show that @xmath184 is strongly connected .
consider a section @xmath207 of @xmath184 and let @xmath208 be two proper elements of @xmath209 . then @xmath210 and @xmath211 . by the strong connectivity of @xmath2 and @xmath3 , there exist sequences @xmath212 and @xmath213 of elements of @xmath214 and @xmath215 , respectively , such that consecutive elements of each sequence are incident and with @xmath216 , @xmath217 , for all @xmath218 and @xmath219 . without loss of generality
we may assume that @xmath220 .
hence , the sequence @xmath221 is such that any two consecutive elements are incident and are all proper faces of the section @xmath207 of @xmath184 .
hence @xmath207 is connected and therefore @xmath184 is strongly connected .
finally , we show that the join product @xmath184 satisfies the diamond condition .
let @xmath222 be such that @xmath223 and @xmath224 then @xmath225 since we have that @xmath177 and @xmath226 , the following possibilities arise : * @xmath227 and @xmath228 ; * @xmath229 and @xmath230 ; * @xmath231 and @xmath232 .
note that the first and the last case are symmetric , so it suffices to consider one of them . in the first case , @xmath233 and , by the diamond condition of @xmath3
, there are exactly two elements @xmath234 such that @xmath235 .
therefore the only elements between @xmath236 and @xmath237 are @xmath238 and @xmath239 . in the second case
the only two faces between @xmath236 and @xmath237 are @xmath240 and @xmath241 .
therefore @xmath184 satisfies the diamond condition , and @xmath169 is an abstract polytope .
the join product on abstract polytopes coincides with the join of two convex polytopes . to see
this one just needs to note that the @xmath65-skeleton of @xmath169 consists of the union of the @xmath65-skeleton of @xmath2 and the @xmath65-skeleton of @xmath3 , together with all the edges from vertices of @xmath2 to vertices of @xmath3 .
the most common example of this product is a pyramid over a polygon : if @xmath147 is a vertex ( or a @xmath52-polytope ) and @xmath2 is an @xmath18-gon ( a @xmath96-polytope with @xmath18 vertices ) , then @xmath242 is simply the pyramid over the @xmath18-gon .
another common example is to consider two edges @xmath243 and @xmath244 ( or line segments ) , and the join of them : @xmath245 is a tetrahedron .
suppose for a moment that we were to regard the _ empty polytope _ @xmath246 as an abstract polytope of rank @xmath182 .
then , the join of @xmath246 with an @xmath18-polytope @xmath2 would simply be the set @xmath247 , and the order will be inherited by that of @xmath2
. it should be then clear that @xmath248 .
conversely , if @xmath3 is a @xmath180-polytope such that @xmath249 for every polytope @xmath2 , then @xmath250 .
therefore we shall say that the only _ trivial polytope with respect to the join product _ is the empty polytope .
if now we consider @xmath147 to be a @xmath52-polytope ( that is , a vertex ) , then @xmath251 that is , the join product of @xmath147 with a polytope @xmath2 gives us two copies of @xmath2 .
however , the rank of an element of the type @xmath252 is @xmath253 , while one of the type @xmath254 is @xmath255 , so the two copies of @xmath2 are at `` different levels '' .
we further note that @xmath256 if and only if @xmath257 ( see figure [ pyramid ] ) .
.,width=181 ] the @xmath258-polytope @xmath259 is called the _ pyramid _ of @xmath2 and we shall denote it by @xmath260 .
it is then straightforward to see that if @xmath2 is a 2-polytope ( or a polygon ) , then @xmath260 is simply the pyramid over @xmath2 . furthermore , @xmath261 is the @xmath262-simplex , which is a regular polytope ( see figure [ joinpowersofapoint ] ) .
note that the join product interacts nicely with the dual operation .
if @xmath263 and @xmath264 are dualities from @xmath2 to @xmath265 and @xmath266 to @xmath267 , respectively , then @xmath268 sending @xmath269 to @xmath270 is a bijection between @xmath169 and @xmath271 such that @xmath272 if and only if @xmath273 .
that is , @xmath274 is a duality from @xmath184 to its dual , implying that @xmath275 in particular if both @xmath2 and @xmath3 are self - dual polytopes , then so is @xmath184 .
the cartesian product of two abstract polytopes is the natural product when thinking on the geometry : it generalises the cartesian product of two convex polytopes .
the direct sum can ( and will ) be defined in terms of the cartesian product and dual polytopes . given two posets @xmath2 and @xmath3 , with minimum faces @xmath59 and @xmath192 , respectively , the _ cartesian product _ of @xmath2 and @xmath3 , denoted @xmath276 , is defined as the set @xmath277 where the order is given by @xmath278 note that the cartesian product of two polytopes @xmath2 and @xmath3 , as a set , is a subset of the @xmath169 , the join of @xmath2 and @xmath3 .
hence , it follows at once that @xmath276 is a poset .
the rank function on @xmath276 is defined in a different way as for the join product : given a face @xmath279 , with @xmath280 , we define the rank of @xmath237 as , @xmath281 and we define the rank of @xmath195 to be @xmath182 . hence , if @xmath2 is an @xmath18-polytope and @xmath3 is an @xmath180-polytope , then @xmath282 is a function from @xmath276 to the set @xmath283 .
in contrast with the join product , we no longer consider the empty set to be a rank @xmath182 polytope , as if we did , we would only have that the cartesian product of any polytope with the empty set is the empty set again .
so the product is of no interest .
[ cartesianpolytope ] let @xmath2 and @xmath3 be two polytopes of ranks @xmath18 and @xmath180 , respectively .
then @xmath276 is a polytope of rank @xmath284 . as pointed out above ,
@xmath276 is a poset .
clearly , @xmath195 is its minimal face and , if @xmath60 and @xmath194 denote the maximal faces of @xmath2 and @xmath3 , respectively , then @xmath196 is the maximal face of @xmath276 . to see that all the flags of @xmath276 have the same number of elements , and that @xmath276 is strongly connected
, one can simply adapt the proofs given in the previous section for @xmath169 .
alternatively , one can think of @xmath276 as a subset of @xmath169 and use this contention to obtain the two properties .
hence , one only needs to see that @xmath276 satisfies the diamond condition .
let @xmath285 such that @xmath286 and @xmath287 note that if @xmath288 , the result holds as it did for @xmath169 .
hence , without loss of generality we may assume that @xmath289 .
this immediately implies that both @xmath72 and @xmath88 are proper faces of @xmath2 and @xmath3 , respectively , and that @xmath290 . if @xmath291 is such that @xmath292 , then @xmath293 .
both @xmath294 and @xmath48 are proper faces of @xmath2 and @xmath3 , respectively , and therefore @xmath295 . since @xmath290 , then @xmath296 , which in turns implies that either @xmath297 and @xmath298 or @xmath299 and @xmath300 .
in the first case , by the diamond condition of @xmath2 we have that there exist two @xmath52-faces @xmath234 such that @xmath301 .
this implies that @xmath302 or @xmath303 .
the second case is similar and hence the diamond condition is satisfied .
the cartesian product of an edge with a polygon is precisely the prism over the polygon and the cartesian product of an edge with any polytope @xmath2 is the prism over @xmath2 .
the only _ trivial polytope with respect to the cartesian product _ is the @xmath52-polytope @xmath147 .
it is straightforward to see that for any polytope @xmath2 , @xmath304 , as the only @xmath52-face of @xmath147 is @xmath147 itself . and
conversely , if @xmath3 is a polytope such that @xmath305 for any polytope @xmath2 , then by considering the rank of @xmath306 one deduces that the rank of @xmath3 is zero and hence @xmath307 .
one interesting example for the cartesian product is to consider a @xmath65-polytope @xmath308 ( that is , an edge ) .
let @xmath309 and @xmath310 be the two @xmath52-faces of @xmath308 .
then , given an @xmath18-polytope @xmath2 , @xmath311 in this case , @xmath312 has two isomrphic copies of @xmath2 ( at the same `` level '' ) , and a third copy of @xmath2 with the minimum removed , at one level higher ( see figure [ prismp ] ) .
we note further that while @xmath313 whenever @xmath314 , for @xmath315 , two faces of the type @xmath316 and @xmath317 can never be incident .
.,width=321 ] the @xmath258 polytope @xmath312 is called the _ prism _ over @xmath2 and shall be denoted by @xmath318 . hence , @xmath319 is the @xmath6-cube , which is a regular polytope .
the direct sum of a segment and a polygon is the bipyramid of the polygon . in the introduction we gave a definition of the direct sum of two convex polytopes .
the direct sum of two convex polytopes can be described , using duality , in terms of a cartesian product .
in fact , we have that for convex polytopes , @xmath320 ( see for example ( * ? ? ? * lemma 2.4 ) ) where @xmath100 denotes the polar dual of @xmath2 .
hence , given two abstract polytopes @xmath2 and
@xmath3 , we define the _ direct sum _ of @xmath2 and @xmath3 , denoted by @xmath321 , simply as @xmath322 it is straightforward to see that if @xmath60 and @xmath194 are the maximal elements of @xmath2 and @xmath3 , respectively , then we have that @xmath323 where the order is given by @xmath324 an immediate corollary of proposition [ cartesianpolytope ] is the following result .
let @xmath2 and @xmath3 be two polytopes of ranks @xmath18 and @xmath180 , respectively . then @xmath321 ( as defined in ( [ setdir ] ) and ( [ orderdir ] ) ) is a polytope of rank @xmath284 .
similarly as above , the only _ trivial polytope with respect to the cartesian product _ is the @xmath52-polytope @xmath147 .
in fact , for any polytope @xmath2 , @xmath325 , as @xmath326 is @xmath147 itself . and
conversely , if @xmath3 is a polytope such that @xmath327 for any polytope @xmath2 , then by considering the rank of @xmath328 one deduces that the rank of @xmath3 is zero and hence @xmath307 .
given an @xmath18-polytope @xmath2 , the @xmath258-polytope @xmath329 is called the _ bipyramid _ over @xmath2 , and shall be deonted by @xmath330 . in this case , @xmath331 is the @xmath6-cross - polytope , which is a regular ( convex ) polytope , that is dual to the @xmath6-cube . the last product that we consider in this paper ,
does not have a convex analogue .
the name is given with the following example in mind : the topological product of two polygons ( homeomorphic to circles @xmath19 ) gives us a map on the torus ( the product of @xmath17 ) .
given an @xmath18-polytope @xmath2 with minimum element @xmath59 and maximum element @xmath60 , and an @xmath180-polytope @xmath3 with minimal element @xmath192 and maximal element @xmath194 , the _ topological product _ of @xmath2 and @xmath3 , denoted by @xmath332 , is defined as @xmath333 where the order is given by @xmath334 here , we say that the rank of the faces @xmath335 are @xmath182 and @xmath336 , respectively , and given @xmath337 with @xmath338 , then @xmath339 note that if @xmath2 has rank @xmath52 , then @xmath340 for every polytope @xmath3 .
moreover , if @xmath2 has rank @xmath65 , and @xmath3 has rank at least @xmath65 , then @xmath332 is not connected , implying that it is not a polytope .
however , using a similar proof as that of proposition [ cartesianpolytope ] , we have the following proposition .
let @xmath2 and @xmath3 be two polytopes of ranks @xmath18 and @xmath180 , respectively , with @xmath341 . then @xmath332 ( as defined in ( [ settopo ] ) and ( [ ordertopo ] ) ) is a polytope of rank @xmath336 .
there are no trivial polytopes for the topological product .
if @xmath342 is a collection of @xmath96-polytopes , then @xmath343 is a @xmath6-torus tessellated by @xmath6-cubes . in particular
if every @xmath344 is isomorphic to a @xmath345-gon , then @xmath343 is the regular @xmath346-polytope @xmath347 ( see @xcite ) .
the purpose of this section is to show that , for any of the four products described in the previous section , any polytope can be factored in a unique way ( up to isomorphism ) as the product of prime polytopes . as the proofs of this result for each of the four products
are very similar , we shall first view all four products as cardinal products of posets and show some results for such products .
we start by noticing that given a @xmath18-polytope @xmath2 with minimum element @xmath59 and maximum element @xmath60 , and an @xmath180-polytope @xmath3 with minimal element @xmath192 and maximal element @xmath194 , we have that @xmath348 in other words , @xmath349 which says that maybe with exception of the minimum and maximum faces , the four products of polytopes can be seen as cardinal products of posets .
theorem [ uniquefactposet ] then implies that each of the four products of polytopes has a unique prime factorisation in terms of posets , however , we want to show that the prime factors are also abstract polytopes .
the following lemma is straightforward .
[ factminmax ] let @xmath2 be poset with minimum element ( resp .
maximumm ) and suppose there exist posets @xmath3 and @xmath350 such that @xmath351
. then @xmath3 has a minimum element ( resp .
maximum ) . by the commutativity of the product , the above lemma
implies that also @xmath352 has a minimum and/or maximum , whenever @xmath2 has it too . in what follows , for an @xmath18-polytope @xmath2 , @xmath353
will be denoting a polytope @xmath2 without its minimum and/or maximum elements .
hence , @xmath353 satisfies the following properties . 1 .
@xmath353 is a poset with a rank function , in which all the maximal chains have the same number of elements . 2 .
@xmath353 satisfies the diamond condition for @xmath354 , and for every face of rank @xmath355 ( 1 , resp . ) , there are exactly two @xmath356-faces ( @xmath52-faces , resp . )
incident to it .
3 . @xmath357 is a weakly connected digraph , and every open interval of @xmath353 either has two elements or it is also connected . note that if a poset @xmath3 satisfies the three above properties , we can extend @xmath3 by defining a minimum and a maximum elements of the order , and then the resulting new poset is indeed a polytope . in the next lemmas
we shall establish that the factors of a factorable poset @xmath353 have the properties p1 , p2 and p3 . [ factrankflag ]
let @xmath2 be an @xmath18-polytope and suppose there exist posets @xmath3 and @xmath350 such that @xmath358
. then @xmath3 ( and therefore @xmath352 ) has a rank function .
furthermore , all the flags of @xmath3 ( and therefore of @xmath352 ) have the same number of elements .
we assume that @xmath359 does not have a maximum and minimum elements .
the arguments are similar in the case it has one of them .
hence , @xmath359 has a rank function with range @xmath360 .
fix @xmath361 to be a maximal face of @xmath359 , thus , @xmath361 has rank @xmath362 . consider @xmath363 then @xmath364 and @xmath365 .
note that the maximality of @xmath361 implies that for every @xmath366 , we have that @xmath367 .
hence , @xmath368 is a maximal element of @xmath3 ( though most likely it is not maximum ) .
moreover , as @xmath364 , then there exists @xmath369 such that @xmath370 and @xmath371 for every @xmath372 .
thus , @xmath373 is a minimal ( but not minimum ) element of @xmath3 .
let @xmath374 .
we now show that if @xmath375 is another minimal element of @xmath3 , then @xmath376 .
since @xmath377 , then we can complete @xmath378 to a maximal chain @xmath67 of @xmath359 .
hence , the minimal element @xmath379 of @xmath67 has rank @xmath52 in @xmath359 and @xmath27 is a minimal element of @xmath352 .
consider the set @xmath380 of all that elements of @xmath67 that have rank less or equal to @xmath381 . since @xmath373 is minimal in @xmath3
, then the first coordinate of all such elements is in matter of fact @xmath373 .
thus , the set @xmath382 consisting of the second coordinates of @xmath380 is a maximal chain of @xmath352 .
now , let @xmath383 be a minimal element ( of @xmath3 )
. we can complete @xmath384 to a maximal chain @xmath199 of @xmath359 in such a way that all the elements of @xmath199 with rank less than @xmath385 are of the form @xmath386 with @xmath387 . since @xmath2 is a polytope ,
then all the maximal chains of @xmath359 have @xmath18 elements .
hence , both @xmath67 and @xmath199 have @xmath18 elements .
as the number of elements in @xmath67 of rank less than @xmath381 equals the number of elements in @xmath199 of rank less than @xmath388 , then the number of of elements in @xmath67 of rank greater than @xmath381 equals the number of elements in @xmath199 of rank greater than @xmath388 , implying that @xmath389 . since @xmath2 is an @xmath18-polytope , then there exists a rank function @xmath390 .
hence , @xmath391 thus , by defining for each @xmath372 , @xmath392 we obtain a rank function from @xmath3 to the set @xmath393 , and the first part of the lemma has been stablished .
note now that any maximal chain of @xmath3 must have at most @xmath394 faces .
let @xmath67 be a maximal chain of @xmath3 , and suppose @xmath67 has less than @xmath395 elements .
let @xmath396 be the minimal and maximal elements of @xmath67 , respectively .
we have shown that all minimal elements of @xmath3 have the same rank and one can similarly show that all maximal elements also have the same rank .
hence , @xmath397 have ranks zero and @xmath395 , respectively .
so let @xmath398 be such that there is no element in @xmath67 of rank @xmath399 and that @xmath399 is minimal in that sense .
then , there exists @xmath400 such that @xmath401 .
let @xmath402 .
then @xmath403 is a maximal chain of @xmath404 .
extend @xmath403 to a flag @xmath199 of @xmath359 , and consider its faces @xmath405 and @xmath406 ( of ranks @xmath407 and @xmath408 , respectively ) . since @xmath401 , then @xmath409 . as there exists no element of @xmath3 of rank @xmath399
, there exists no element of @xmath404 of rank @xmath408 , and therefore @xmath406 is not an element of @xmath404 .
this implies that there exists @xmath410 and @xmath411 with @xmath412 such that @xmath413 .
but since @xmath199 is a flag and @xmath414 , then @xmath415 .
this immediately implies that @xmath416 , which is a contradiction .
therefore for every @xmath417 there is an element of @xmath67 of rank @xmath399 and thus all flags of @xmath3 have the same number of elements , namely @xmath395 . [ factconn ]
let @xmath2 be an @xmath18-polytope and suppose there exist posets @xmath3 and @xmath350 such that @xmath358
. then @xmath3 ( and therefore @xmath352 ) is connected and so is every interval of it .
we start by showing that @xmath3 is connected .
suppose otherwise .
then @xmath120 is a disconnected digraph and hence @xmath418 is disconnected .
this in turns implies that @xmath359 is disconnected , which is a contradiction .
we shall now see that every interval of @xmath3 is in fact isomorphic to an interval of @xmath359 .
the proposition will follow then at once .
in fact , given @xmath419 with @xmath114 , the intervals @xmath420 , @xmath421 and @xmath422 are respectively isomorphic to the intervals of @xmath423 , @xmath424 and @xmath425 , where @xmath426 is a fixed element of @xmath352 .
thus every interval of @xmath3 is connected .
[ factdiam ] let @xmath2 be an @xmath18-polytope and suppose there exist posets @xmath3 and @xmath350 such that @xmath358 .
let @xmath427 be such that @xmath428 is the rank function defined in ( [ rankq ] ) .
then , 1 .
if @xmath419 are such that @xmath132 with @xmath429 , then there are exactly two faces @xmath430 such that @xmath431 .
2 . if @xmath432 is such that @xmath433 , then there are exactly two faces @xmath430 such that @xmath434 .
3 . if @xmath410 is such that @xmath435 , then there are exactly two faces @xmath430 such that @xmath436 .
we start by showing part @xmath46 .
let @xmath426 be an element of @xmath352 .
then , @xmath437 and @xmath438 . by the diamond condition of @xmath2 there exist exactly two elements @xmath439 such that @xmath440 . by the definition of the cardinal product @xmath441 ,
the second coordinate of @xmath442 must be @xmath48 .
hence , part @xmath46 of the proposition follows .
parts @xmath443 and @xmath444 follow in a similar fashion watching for the details .
in fact for part @xmath443 , we must set @xmath445 a maximal element of @xmath352 , while for part @xmath444 , @xmath446 a minimal element of @xmath352 .
using lemmas [ factminmax ] , [ factrankflag ] , [ factconn ] and [ factdiam ] we can now establish the following theorem .
[ factpoly ] let @xmath2 be an @xmath18-polytope with minimum element @xmath59 and maximum element @xmath60 and @xmath3 and @xmath352 be posets .
then we have the following . 1 . if @xmath351 , then both @xmath3 and @xmath352 are polytopes
2 . if @xmath447 , then both @xmath3 and @xmath352 have a maximum element and satisfy properties p1 , p2 and p3 .
3 . if @xmath448 , then both @xmath3 and @xmath352 have a minimum element and satisfy properties p1 , p2 and p3 .
4 . if @xmath449 , then both @xmath3 and @xmath352 satisfy properties p1 , p2 and p3 .
let @xmath2 be an abstract polytope and let @xmath20 denote a product of polytopes ( either the join , cartesian or topological product , or the direct sum )
. then @xmath2 can be uniquely factorised as a @xmath20-product polytopes that are prime with respect to the product @xmath20 .
the corollary follows from theorems [ uniquefactposet ] and [ factpoly ] .
in sections [ sec : auto ] and [ sec : mono ] we shall deal with the groups and orbits of products . to study these groups
it shall prove very helpful to have a better understanding of the structure of the flags of a product .
that is the purpose of this section .
we start by analyzing the join product .
let @xmath2 be an @xmath356-polytope and suppose that @xmath450 , for some polytopes @xmath451 such that @xmath24 has rank @xmath452 .
this implies that @xmath453 without loss of generality we may assume that @xmath454 . let @xmath67 be a flag of @xmath2 .
then @xmath455 , where @xmath68 has rank @xmath64 .
since @xmath2 is a product , then for each @xmath456 there exist @xmath457 such that @xmath458 by definition of the join product , we have that for each @xmath459 , @xmath460 where @xmath461 and @xmath462 are the minimum and maximum elements , respectively , of @xmath463 .
note that many of the @xmath464 are repeated in the above sequence , as otherwise @xmath2 would be just a trivial product .
that means that the set @xmath465 has cardinality @xmath30 and , after erasing the repeated faces , it is a flag of @xmath463 . call @xmath466 such flag .
now , since @xmath455 is a flag of @xmath2 , then @xmath467 for each @xmath64 . again , by the definition of the join product , for each @xmath64 , @xmath468 and @xmath469 differ in exactly one entry .
denote by @xmath470 such entry .
in other words , @xmath471 if and only @xmath472 and @xmath68 differ in their @xmath473 entry .
thus , we can naturally identify each flag @xmath67 of @xmath2 with the ordered pair @xmath474 , where each @xmath475 is the flag of @xmath24 described above and @xmath476 .
clearly , two different flags of @xmath2 define different ordered pairs .
note further that for each flag @xmath108 , the sequence @xmath477 has exactly @xmath478 times the integer @xmath473 , for each @xmath479 .
let @xmath480 be the set of all ordered @xmath18-tuples @xmath481 with @xmath482 and such that each @xmath483 appears exactly @xmath478 times in @xmath381 .
given an ordered pair @xmath484 , where @xmath466 is a flag of @xmath463 and @xmath485 , we can define the flag @xmath486 of @xmath2 as follows .
the minimum face of @xmath67 , @xmath487 is the @xmath488-tuple @xmath489 , where each @xmath490 is the minimum face of the polytope @xmath463 .
suppose that we have defined the @xmath491-face @xmath492 of @xmath67 , in such a way that @xmath493 is a face of the flag @xmath466 of @xmath463 .
hence , for each @xmath494 , the face @xmath493 of @xmath463 has some rank , say @xmath495 , with @xmath496 .
the @xmath64-face @xmath497 is the @xmath488-tuple that coincides with @xmath472 in all its entries , except in the entry @xmath498 .
the entry @xmath498 of @xmath68 is the face of the flag @xmath499 of rank @xmath500 .
hence , in particular , @xmath501 has all its entries equal the minimum face of the corresponding polytope ( all of rank @xmath182 ) , except for its @xmath502 entry , which is the @xmath52-face of the flag @xmath503 .
if we now take one of the other three products , the analysis of the flags is very similar .
the main differences are in the way the set @xmath480 should be defined for each product and , thus , in the way to construct a flag of the product , given one flag of each factor and an element of @xmath480 .
alternatively , one can keep @xmath480 fixed and adjust the ranks of the polytopes @xmath24 as well as the definition of the vertices or the facets of the product , depending on the product we are dealing with . using similar methods as the ones explained above one can show the following lemma .
[ flagbijection ] let @xmath504 be polytopes and @xmath20 denote one of the four products discussed in section [ sec : allproducts ] ( that is , @xmath505 ) .
let @xmath506 and let @xmath507 denote the set @xmath508 .
then there exists a bijection @xmath509 between @xmath510 and @xmath511 , where , @xmath480 is the set of ordered @xmath18-tuples with entries in the set @xmath512 and such that each @xmath483 appears exactly @xmath478 times in @xmath381 , @xmath513 , and @xmath478 is related to the rank of the polytope @xmath463 in the following way : @xmath514
in this section we turn our attention to the automorphism group of a product of polytopes . throughout the section , @xmath20 will denote one of the four products discussed in section [ sec : allproducts ] ( i.e. , @xmath515 , and we shall refer to the @xmath20-product simply as the product . likewise , a prime polytope will be a prime polytope with respect to @xmath20 .
although @xmath20 can not always be seen as a cardinal product of posets , we note that the automorphism group of a polytope @xmath2 coincides with the automorphism group of @xmath2 taking away the minimum or maximum elements or both .
hence , for proposes of computing the automorphism group of a product @xmath20 of polytopes , without loss of generality we may assume that @xmath20 is in fact the cardinal product of posets .
we shall say that two polytopes @xmath2 and @xmath3 are _ relatively prime _ if their ( unique ) prime factorization does not have any prime polytopes in common .
in particular , if both @xmath2 and @xmath3 are different prime polytopes , then they are relatively prime . in @xcite ( corollary 2 )
, duffus shows that every automorphism @xmath516 of a product @xmath172 of relatively prime posets is the product of an automorphism @xmath517 of @xmath2 times an automorphism @xmath518 of @xmath3 . from this fact
, we obtain the following proposition .
[ relprime ] let @xmath2 and @xmath3 be two relatively prime polytopes .
then , @xmath519 .
corollary 2 of @xcite also states that if @xmath3 is a prime poset and @xmath520 , then for any automorphism @xmath516 of @xmath521 there exist a permutation @xmath522 of the set @xmath523 and automorphisms @xmath524 of @xmath3 such that for every @xmath525 , and every @xmath526 , the @xmath64-th coordinate of the element @xmath527 is precisely @xmath528 ( where @xmath529 is precisely the image of @xmath64 under the permutation @xmath522 ) .
that is , @xmath530 .
moreover , it is clear that an element @xmath531 acts naturally on the elements of @xmath2 .
namely , @xmath532 . from these two facts
, one can show the following proposition .
[ power ] let @xmath3 be a prime polytope and let @xmath533 .
then @xmath534 . by denoting @xmath535 simply by @xmath536 and @xmath537 as @xmath538 , from propositions [ relprime ] and [ power ] , we obtain the following corollary , which settles part @xmath444 of theorem a. if @xmath23 , where the @xmath24 are distinct prime polytope , then @xmath25 the main interest in the study of abstract polytopes has been the highly symmetric ones , being the regular ones the most studied ones in the last 30 years ( see for example @xcite ) .
as one naturally expects , the product of two regular polytopes in general is not a regular polytope anymore .
in fact , we shall see that with exception of one family per product , regular polytopes are prime .
although different products are described in a slightly different way , we can study them all under the same scope , so we let @xmath539 . by lemma [ flagbijection ] , there is a bijection between the flags of @xmath540 and @xmath541 , where @xmath542 and @xmath480 is the set of all sequences @xmath543 such that @xmath544 and that for each @xmath483 the integer @xmath473 appears exactly @xmath478 times in the sequence , @xmath478 stands for the rank of @xmath463 plus or minus @xmath65 , depending which product of polytopes we would want to consider and @xmath545 is the rank of @xmath2 .
note that the cardinality of @xmath546 is @xmath547 , and let us denote @xmath548 by @xmath549 .
suppose now that @xmath550 ( that is , @xmath551 ) and that the posets @xmath552 and @xmath553 are relatively primes .
let @xmath554 .
then , @xmath555 , where @xmath556 , and the action of @xmath516 on an element @xmath557 , of @xmath549 is given by @xmath558 note that given @xmath559 with @xmath560 , there is no element of @xmath98 that can send an element of @xmath549 with second coordinate @xmath381 to one with second coordinate @xmath561 .
furthermore , in order to have @xmath557 and @xmath562 in the same orbit , we need @xmath563 and @xmath564 in the same orbit under @xmath565 . if for each integer @xmath29 , we let @xmath566 then we have established the following lemma .
[ orbitstworelatprime ] let @xmath2 be an @xmath356-polytope , and suppose @xmath567 with @xmath552 and @xmath553 relatively prime with respect to @xmath20 .
let @xmath568 ( as in ( [ not : ranks ] ) ) denote the rank of @xmath24 , and let @xmath26 denote the number of orbits of @xmath569 on @xmath570 .
then the number of orbits of @xmath99 under the action of @xmath98 is @xmath571 . by induction
we can now obtain the following corollary [ coro : orbitsrelatprimes ] let @xmath572 with @xmath24 and @xmath463 relatively prime for every @xmath573 .
let @xmath568 ( as in ( [ not : ranks ] ) ) be the rank of @xmath24 and @xmath26 denote the number of orbits of @xmath569 on @xmath570 .
then the number of orbits of @xmath99 under the action of @xmath98 is @xmath574 we now turn our attention to the case when @xmath575 , where @xmath3 is a prime polytope with respect to @xmath20 and @xmath180 is a natural number . in this case , the action of @xmath576 on the elements of @xmath549 is given by : @xmath577 we have seen that @xmath578 . in this case
this means that , if @xmath579 is the rank of @xmath3 , then @xmath580 and @xmath581 note that each @xmath485 can be sent by an element of @xmath98 to @xmath582 elements .
indeed , each element of @xmath583 acts on the second coordinate of the elements of @xmath584 , and the only element of @xmath38 that fixes a given @xmath585 is the identity .
moreover , only the elements of @xmath583 can permute the second coordinates of the elements of @xmath586 .
this implies that the action of @xmath583 on @xmath480 has @xmath587 orbits .
in particular we note that this number is always an integer . by now taking into consideration the number of orbits of @xmath588 under the action of @xmath589
, we have the following lemma .
[ lemma : orbitspowers ] let @xmath590 for some prime polytope @xmath3 with respect to @xmath20 .
let @xmath579 denote the rank of @xmath3 , and @xmath27 denote the number of orbits of @xmath589 on @xmath588 .
then the number of orbits of @xmath99 under the action of @xmath98 is @xmath591 we are now ready to compute the number of flag orbits of a product .
[ orbitsposet ] let @xmath23 , where the @xmath24 are distinct prime posets with respect to @xmath20 .
let @xmath568 be the rank of @xmath24 and let @xmath26 denote the number of orbits of @xmath569 on @xmath570
. then the number of orbits of @xmath99 under the action of @xmath98 is @xmath592 note that in lemmas [ orbitstworelatprime ] and [ lemma : orbitspowers ] , corollary [ coro : orbitsrelatprimes ] and proposition [ orbitsposet ] , we never use the fact that the posets are strongly connected . hence similar propositions for pre - polytopes ( in the sense of @xcite ) , and their products as posets hold .
we would like to find out when a product of polytopes is a regular or a 2-orbit polytope .
we start by analyzing the case when a product is regular .
then , the number of orbits of proposition [ orbitsposet ] has to equal one .
start by noticing that @xmath593 since @xmath594 is an integer for every @xmath64 , then @xmath595 is also an integer .
furthermore , whenever @xmath596 , then @xmath597 .
hence , if we want @xmath598 , then @xmath599 ( that is , @xmath2 is a power of a prime poset @xmath3 ) .
thus , @xmath27 becomes @xmath600 again , since @xmath601 is an integer , this immediately implies that @xmath602 ( that is , @xmath3 is regular ) and that @xmath603 .
the last equality holds if and only if either @xmath604 or @xmath605 . in the first case
, this implies that @xmath2 is a prime poset .
the second case implies that @xmath606 , where @xmath3 has maximal chains of size @xmath96 .
let @xmath2 be a regular polytope .
then , @xmath2 is prime with respect to all four products except in the following cases : 1 . if @xmath2 is an @xmath18-simplex , then @xmath2 is not prime with respect to the join product .
in fact , @xmath607 , where @xmath147 is a @xmath52-polytope .
2 . if @xmath2 is an @xmath18-cube , then @xmath2 is not prime with respect to the cartesian product .
in fact , @xmath608 , where @xmath308 is a @xmath65-polytope .
3 . if @xmath2 is an @xmath18-crosspolytope , then @xmath2 is not prime with respect to the direct sum .
in fact , @xmath609 , where @xmath308 is a @xmath65-polytope .
4 . if @xmath610 , where @xmath3 is a @xmath96-polytope , then @xmath2 is a regular polytope that is not prime with respect to the topological product .
each of the cases follows from the above discussion and by the following facts . in every case @xmath2
must be the product of identical copies of prime polytopes ( with respect to the given product ) .
the join product is a product of posets and hence , with the above notation , the size of a maximal chain of the poset @xmath3 coincides with the number of flags of @xmath3 as polytope , implying that the rank of @xmath2 is zero . the cartesian product and the direct sum taking away one element are products of posets , hence , the size of a maximal chain of the poset @xmath3 is in fact one more than the rank of @xmath3 , that is , the rank of @xmath2 must be one .
finally , the topological product , when taking away the minimum and maximum elements , is a product of posets .
hence the rank of @xmath3 must in fact coincide with the number of elements in a maximal chain , when seen as a factor in the product .
that is , @xmath3 has to be a @xmath96-polytope .
following a similar analysis we can obtain an analogous theorem for two - orbit polytopes .
let @xmath2 be a two - orbit polytope .
then @xmath2 is prime with respect to the four products , except in the case where @xmath2 is a torus @xmath611 . in this case
@xmath612 , where @xmath3 and @xmath352 are non - isomorphic @xmath96-polytopes , and @xmath2 is prime with respect to the other four products .
if @xmath2 is a two - orbit polytope , then @xmath27 in ( [ k ] ) must equal 2 .
we divide the analysis into two cases , when @xmath599 , and when @xmath613 .
suppose first that @xmath614 , with @xmath3 prime . then , ( setting @xmath615 ) @xmath616 .
since @xmath617 , then either @xmath602 and @xmath618 or @xmath619 and @xmath620 . the first case can never happen ( as @xmath621 implies @xmath622 ) , and in the second case @xmath2 is simply a prime two - orbit polytope .
suppose now that @xmath613 .
since @xmath623 and @xmath624 , we have that @xmath625 and hence @xmath626 . as for every @xmath64
, @xmath627 is an integer , this in turn implies that every @xmath628 and that for every @xmath64 , @xmath629 .
let @xmath630 , for every @xmath64 .
then we have that @xmath631 .
the last equality holds if and only if @xmath550 and @xmath632 .
hence , @xmath633 . as pointed out before , for the join product
, the cartesian product and the direct product , @xmath634 implies that @xmath635 and they are either a @xmath52-polytope or a @xmath65-polytope .
however , we are under the assumption that @xmath552 and @xmath553 are relatively primes . hence , we only have left the case when @xmath552 and @xmath553 , as polytopes , have rank @xmath96 and they are relatively prime with respect to @xmath636 . since all rank 2 polytopes are prime with respect to the topological product , then @xmath552 and @xmath553 are only required to be non isomorphic 2-polytopes .
this establishes the theorem .
the monodromy groups of an abstract polytope encapsulates all the combinatorial information of the polytope ( see @xcite,@xcite ) . however , monodromy groups of non - regular abstract polytopes have been proven hard to understand ( see @xcite ) . here
, we study some basic properties of the monodromy group of a product . to this end
, we use the description of the flags of a product given in section [ sec : flags ] . as we have seen throughout ,
the four products of polytopes studied in this paper behave very much alike . here , we give the details of our proofs only for the join product .
the details of the other three products can be recovered from this one by making small modifications .
let @xmath637 be an @xmath356-polytope , where @xmath638 and @xmath452 is the rank of the polytope @xmath24 . for convenience
, throughout this section we shall make use of lemma [ flagbijection ] , and write each @xmath639 as @xmath640 , where each @xmath641 and @xmath642 .
( recall that @xmath480 is the set of ordered @xmath18-tuples @xmath643 with @xmath482 and such that each @xmath483 appears exactly @xmath478 times in @xmath381 . )
hence , we regard the set of flag @xmath99 and the set @xmath644 as the same object and use one or the other indistinctly .
note that we can regard @xmath38 as the permutation group on the symbols @xmath645 , and hence @xmath38 acts on @xmath480 in a natural way .
that is , given @xmath646 and @xmath647 , then @xmath648 .
let @xmath649 be the generators of the monodromy group @xmath650 of @xmath24 .
hence , each @xmath651 permutes every flag of @xmath24 with its @xmath473-adjecent one .
let @xmath652 .
we shall start by showing that the wreath product @xmath653 of @xmath426 by @xmath38 acting on @xmath480 as described above , acts on the set @xmath99 in a faithful way . recall that if @xmath654 , @xmath655 , then @xmath656 let @xmath657 and @xmath658 .
since for every @xmath659 , we have that @xmath660 , then @xmath661 is an @xmath488-tuple @xmath662 , with @xmath663 .
hence , the action on @xmath664 on @xmath67 is given by @xmath665 it is not difficult to see that ( [ monoaction ] ) in fact defines an action of @xmath666 on @xmath99 and that if an element @xmath667 fixes every flag of @xmath2 , then @xmath668 and for each @xmath669 we have that @xmath670 , implying that the action is faithful .
let @xmath671 be the generators of the monodromy group of @xmath2 , @xmath37 . since each @xmath672 permutes every flag of @xmath2 with its @xmath27-adjacent one , in order to understand @xmath37 we need to understand the flag adjacencies in @xmath2 .
let @xmath673 .
consider the @xmath27-adjacent flag to @xmath67 , @xmath674 .
then we can write @xmath675 and @xmath676 , where each @xmath68 , @xmath677 as well as @xmath382 are faces of the product .
for each @xmath64 , we write @xmath678 and @xmath679 . using the definition of the order of the product @xmath2
, we observe that , for each @xmath64 , @xmath68 and @xmath680 differ in exactly one element ( in fact , they differ in their @xmath681 element ) .
hence , @xmath682 and @xmath683 differ on at most two elements and on at least one .
our study then naturally splits into two cases : when @xmath682 and @xmath683 differ on one or two elements .
we start by assuming that @xmath682 and @xmath683 differ in exactly one element , say @xmath684 .
that is , @xmath685 and @xmath686 .
this immediately implies that @xmath687 for all @xmath688 , and that , in @xmath463 , @xmath689 are three different faces that are incident and whose ranks are consecutive .
the same holds true for @xmath690 . in other words ,
when @xmath691 , then @xmath692 where @xmath693 is the @xmath694-adjacent flag to @xmath695 , and @xmath694 is the number of times that @xmath498 appears in the sequence @xmath696 .
suppose now that @xmath682 and @xmath683 differ in exactly two elements , say on those corresponding to @xmath697 and @xmath698 .
then @xmath699 for all @xmath700 .
furthermore , either @xmath701 or @xmath702 in other words , when @xmath703 , @xmath704 we are now ready to relate the monodromy group of @xmath2 with the wreath product @xmath666 . given @xmath705 and @xmath706 a product polytope , if @xmath29 is the rank of the polytope @xmath24 , then we define @xmath18 as , @xmath707 [ prop : monext ] let @xmath705 .
given polytopes @xmath708 and @xmath706 , the monodromy group of @xmath2 , @xmath37 , can be embedded as a subgroup of the wreath product @xmath709 , where @xmath426 is the direct product of the monodromy groups of the @xmath24 and @xmath18 is as in ( [ n ] ) .
moreover the projection on the second factor @xmath710 is surjective .
we give a proof for when @xmath41 , the other three cases are similar . since we know that @xmath666 acts faithfully on @xmath711 ) , to settle the first part of the proposition , it is enough to show that we can embed each of the generators of @xmath37 in @xmath666 .
let @xmath712 be fixed , and let @xmath713 and for each @xmath714 let @xmath715 .
consider @xmath716 , where @xmath717 and @xmath718 where @xmath694 is the number of times that @xmath473 appears in the sequence @xmath719 .
using ( [ eq : ad1 ] ) and ( [ eq : ad2 ] ) is straightforward to see that for every @xmath720 , @xmath721 .
hence each generator of @xmath37 can be embedded into @xmath666 , implying that @xmath37 can be embedded as a subgroup of the wreath product @xmath722 .
since @xmath723 , the second part of the proposition follows .
let @xmath705 and @xmath706 be a polytope .
then the monodromy group of @xmath2 is an extension of a symmetric group @xmath38 , where @xmath18 is as in ( [ n ] ) .
this corollary tells us that the monodromy group of a product is always an extension of a symmetric group .
however , this extension does not always splits , and figuring out when it does is not always easy or straightforward . in what follows
we show how can this be computed in some simple examples .
for the remainder of this section , we let @xmath724 be the natural projection and @xmath48 be the kernel of the restriction of @xmath725 to @xmath726 .
let @xmath2 be an @xmath18-polytope and consider its pyramid @xmath727 .
let @xmath728 denote the symmetric group on the symbols @xmath729 , and for @xmath730 , let @xmath731 .
the set @xmath480 is the set @xmath732 , where @xmath733 is the vector with @xmath66 entires such that the entire @xmath734 is @xmath96 and the rest of them are @xmath65 . since @xmath147 is a @xmath52-polytope , its monodormy group is trivial .
hence @xmath735 is embedded as a subgroup of the wreath product @xmath736 .
let @xmath737 be the generators of @xmath37 .
following the proof of proposition [ prop : monext ] , we can see that the generators @xmath738 of @xmath735 can be regarded as : @xmath739 where the identity element @xmath65 of @xmath37 is in the @xmath734 and @xmath740 entries .
observe that for each @xmath741 , @xmath742 where the identity element @xmath65 of @xmath37 is in the @xmath734 , @xmath740 and @xmath743 entries and @xmath744 denotes the identity of @xmath745 .
hence , the order of @xmath746 is @xmath747 $ ] , the largest common multiple of @xmath55 , @xmath748 and @xmath749 , where @xmath750 is the order of @xmath751 in @xmath37 .
computing the kernel @xmath48 of the restriction of the projection @xmath752 to @xmath37 is rather difficult in general .
one can use similar techniques to the ones we shall use in section [ sec : monoprism ] to show that when @xmath2 is a @xmath345-gon ( that is , the simplest case of the pyramid ) then @xmath753 , where @xmath58 .
hence , the monodromy group is an extension of @xmath33 by @xmath57 .
moreover , the extension splits if and only if @xmath345 is not congruent to @xmath52 modulo @xmath56
. we do not give the details of this here , as this group has been previously computed in @xcite .
let @xmath2 be an @xmath18-polytope and consider its prism @xmath754 .
we start by making some general remarks about @xmath755 to exemplify how the above discussion would apply to one example in the cartesian product , and then proceed to compute the monodromy group of the prism over a polygon as an extension of @xmath756 .
let @xmath745 denote the symmetric group on the symbols @xmath757 , let @xmath744 denote the identity of @xmath745 and for @xmath758 , let @xmath759 .
the set @xmath480 is the set @xmath760 , where @xmath733 is the vector with @xmath761 entries such that the entry @xmath64 is @xmath96 and the rest of them are @xmath65 . since @xmath308 is a @xmath65-polytope , its monodromy group is a cyclic group of order 2 .
we let @xmath762 denote its generator . and let @xmath737 be the generators of @xmath37 . by proposition [ prop : monext ] , @xmath763 can be embedded in @xmath764 . note that in this case , the @xmath64-adjacency of the flags of @xmath755 is not described anymore by ( [ eq : ad1 ] ) and ( [ eq : ad2 ] ) .
let @xmath765 be a flag of @xmath755 , where @xmath766 , @xmath767 and @xmath642 .
then , the @xmath52-adjacency of a flag @xmath67 is determined only by the value of the first entry of @xmath381 .
@xmath768 for @xmath77 , the @xmath64-adjacency is similar as that in ( [ eq : ad1 ] ) and ( [ eq : ad2 ] ) .
there is a small modification that has to be done , obtaining that , @xmath769 hence , using this to modify the ideas of the proof of proposition [ prop : monext ] , if @xmath770 denote the generators of @xmath763 , then @xmath771 where the identity element @xmath65 of @xmath37 is in the @xmath64 and @xmath734 entries . computing the kernel @xmath48 is not always easy and depends on the monodromy group of @xmath2 . in
what follows , we compute @xmath48 , whenever @xmath2 is a @xmath345-gon , that is , @xmath755 is simply the prism over a polygon . in @xcite
the monodromy groups of prisms over polygons were computed in terms of generators and relations . here , we also have the generators and can infer the relations of the group , but we shall focus on computing such group as a split extension of @xmath756 .
let @xmath2 be a @xmath345-gon , and let @xmath3 be the prism over @xmath2 , that is @xmath772 . by proposition [ prop : monext ]
, @xmath773 can be embedded into the wreath product @xmath774 , where @xmath775 and @xmath776 .
hence , @xmath773 is in fact an extension of @xmath756 by the group @xmath48 , the kernel of the restriction of @xmath777 to @xmath773 .
furthermore , the generators in ( [ genmonoprism ] ) become : @xmath778 we start by noticing that @xmath779 has order 3 .
moreover , observe that @xmath780 and hence @xmath781 from here is straightforward to see that the order of @xmath782 is @xmath783 , where @xmath784 moreover , @xmath785 and so @xmath786 .
in fact , it is not too difficult to see that @xmath787 and @xmath788 generate @xmath48 .
we now study the structure of the group @xmath48 . in order to slightly simplify our notation
, we let @xmath789 , @xmath790 , @xmath791 and @xmath792 . hence , @xmath793 and @xmath794 .
observe that @xmath795 hence , the group @xmath796 is isomorphic to @xmath50 ( this group @xmath294 actually coincides with the group @xmath294 in ( * ? ? ?
* section 6 ) ) .
it is straightforward to see ( using the description of the generators as elements of @xmath666 ) that @xmath294 is normal in @xmath48 .
moreover , the elements of the quotient @xmath797 are simply @xmath798 , implying that @xmath799 . in other words
, @xmath48 is an extension of @xmath49 by @xmath50 .
whenever @xmath180 is odd ( that is , if @xmath345 is different than @xmath52 mod @xmath53 ) , then @xmath800 generate the @xmath49 , implying that the extension splits . otherwise , there are no elements of @xmath801 that generate the @xmath49 and the extension does not split .
hence , we have the following proposition . let @xmath3 be the prism over a @xmath345-gon .
then , @xmath802 , where @xmath48 is an extension of @xmath49 by @xmath50 , with @xmath784 .
this extension splits whenever @xmath345 is not congruent to @xmath52 modulo @xmath53 . in this case , @xmath803 once we know the structure of the monodromy group of prisms over polygons , we consider prisms over some @xmath55-polytopes .
let @xmath345 be an integer , and @xmath2 be a @xmath55-polytope such that all its vertex figures are isomorphic to @xmath345-gons ( these polytopes are sometimes called _ uniform _ maps , however , the notation is not standard as _ uniform _ in other contexts means that the polytope is vertex - transitive ) . consider the prism over @xmath2 .
then , the generators of @xmath763 are @xmath804 note that @xmath805 is isomorphic to the monodromy group of the pyramid over a @xmath345-gon .
although this suggests that computing the kernel @xmath48 and hence knowing the structure of the monodromy group of the prism is easy ( as we have already done the work for the pyramid ) , this is far from true .
the reason for this is that now @xmath806 , so finding the generators of @xmath48 is not easy .
it is true , however , that @xmath48 contains a normal subgroup isomorphic to @xmath807 and that the extension of @xmath33 by @xmath48 splits whenever @xmath345 is not @xmath52 modulo @xmath56 ( since the elements we need to use to recover @xmath33 in @xmath763 are the same as the ones needed in the pyramid ) . in other words
, we have the following proposition .
let @xmath2 be a @xmath55-polytope such that all its vertex - figures are isomorphic to a @xmath345-gon .
then the monodromy group of @xmath808 ) , the prism over @xmath2 , is a split extension of @xmath33 by some normal group whenever @xmath345 is not congruent to @xmath52 modulo @xmath56 .
let @xmath2 be an @xmath18-polytope , and consider @xmath809 , where @xmath3 is a @xmath345-gon .
note that @xmath810 has rank @xmath761 .
the analysis of @xmath811 is very similar to that of @xmath755 .
the two main differences are that @xmath811 is now an extension of @xmath38 ( as opposed to @xmath745 ) , and that the generators @xmath812 and @xmath813 of @xmath811 are now @xmath814 where @xmath815 are the generators of @xmath773 and @xmath816 are the generators of @xmath37 .
again , computing in general @xmath48 is not easy in general .
however , whenever @xmath2 is also a @xmath96-polytope , say a @xmath373-gon , is rather simple . in this case ,
@xmath817 , so @xmath818 is a split extension of @xmath819 by @xmath48 .
moreover , @xmath820 and @xmath821 hence , the kernel @xmath48 is generated by @xmath822 and @xmath823 . a simple computation shows
then that @xmath824 , where @xmath825 $ ] is the least common multiple of @xmath345 and @xmath373 . moreover , it is straightforward to see that these two groups commute implying that @xmath826 and @xmath827 in fact , it is not difficult to extend these techniques to show that , if @xmath24 is a @xmath748-gon , then @xmath828 where @xmath345 is the least common multiple of @xmath829 .
we note here that this result is not surprising at all , since the monodromy group of a polytope @xmath2 is isomorphic to the minimal regular cover of @xmath2 , whenever such cover is unique ( see for example @xcite ) .
it is easy to see that the minimal regular cover of @xmath830 is the regular polytope @xmath831 , where @xmath3 is a @xmath345-gon ( @xmath832 $ ] ) and the power is taken over the @xmath636-product .
as we have pointed out before , computing the monodromy group of non - regular polytopes is a difficult task . in this paper
we showed that by regarding some polytopes as products this task can be simplified .
in particular we think that the computations needed to calculate the monodromy groups of prisms and pyramids over polygons are fairly easy , specially if one compares them to those of @xcite and @xcite . for this reason
we strongly believe that the techniques used here can be extended in order to compute monodromy groups of other interesting products and think it is an interesting project to pursue .
the authors would like to thank toma pisanski and ricardo strausz for suggesting studying certain products on polytopes and ricardo strausz and deborah oliveros for some discussions on the subject .
the completion of this work was done while the second author was on sabbatical at the laboratoire dinformatique de lcole polytechnique .
she thanks lix and vicent pilaud for their hospitality and the program paspa - dgapa and to the unam , the support for this sabbatical stay .
we also gratefully acknowledge financial support of the papiit - dgapa , under grant in107015 and of conacyt , under grant .
m. henk , j. richter - gebert , and g. m. ziegler , basic properties of convex polytopes . in j. e. goodman and j. orourke , editors , _ handbook of discrete and computational geometry , _ chapter 13 , pages 243270 .
crc press , boca raton , ( 1997 ) . | given two convex polytopes , the join , the cartesian product and the direct sum of them are well understood . in this paper
we extend these three kinds of products to abstract polytopes and introduce a new product , called the topological product , which also arises in a natural way .
we show that these products have unique prime factorization theorems .
we use this to compute the automorphism group of a product in terms of the automorphism groups of the factors and show that ( non trivial ) products are almost never regular or two - orbit polytopes .
we finish the paper by studying the monodromy group of a product , show that such a group is always an extension of a symmetric group , and give some examples in which this extension splits . |
x - rays may have played an important role in heating the early universe @xcite .
however , direct study of the sources that produced x - rays in the early universe is currently impractical due to their high redshifts . @xcite
have suggested that blue compact dwarf ( bcd ) galaxies may be local analogues of the x - ray producing , metal - deficient galaxies found in the early universe .
the x - ray emission of bcds is dominated by point - like sources that are likely high - mass x - ray binaries ( hmxb ) @xcite .
recently @xcite , @xcite , and @xcite have reported enhanced production of x - ray binaries in low - metallicity , or extremely metal poor galaxies ( xmpg ) , relative to the star formation rate ( sfr ) . to further study this enhanced production , we analyse data from a sample of 25 metal - poor bcds .
we compare our results with those of near - solar metallicity galaxies from a study done by @xcite .
initially , we assume that the x - ray luminosity function ( xlf ) power law slope is the same for low and near - solar metallicities . using a bayesian approach
, we show that the distributions of the xlf normalisations for the low - metallicity bcds are significantly different from that found for near - solar metallicity galaxies .
we also find the most probable values for both the power law slope and normalisation of the metal poor xlf and find that the slope is consistent with that derived for near - solar metallicity galaxies .
in section [ sect : sample ] , we discuss the selection criteria for our sample of dwarf galaxies .
section [ sect : analysis ] contains the full description of the procedures we used to prepare and analyse our dataset .
section [ sect : bayes ] introduces the bayesian approach we used in determining the xlf normalisation . in section [
sect : results ] , we present our results along with a short discussion regarding uncertainty in sfrs .
the analysis was done using a sample of bcds that is given in table [ proptable ] .
we selected galaxies with metallicity such that @xmath6 and a distance @xmath7 mpc .
the metallicity cut - off was chosen to select galaxies similar to those found at high redshifts @xcite .
the metallicities in table [ proptable ] were taken from @xcite .
we selected nearby galaxies so that the individual , compact x - ray sources are bright enough to be detected .
we searched the literature for bcds fitting our criteria , with sources including @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , and references therein .
we found 27 bcds , of which 25 had been observed by the _
_ chandra x - ray observatory__. none of the galaxies in our sample were originally discovered in x - rays .
we use the 25 bcds that have been observed by _
chandra _ as our sample .
the luminosity limits for which the _ chandra _ data are complete varies and are listed in table [ proptable ] .
we discuss the method for determining completeness limits in section [ subsect : completeness ] .
l r c c c c c name & obsid & 12+log(o / h ) & ra & dec & exposure & distance + & & & ( j2000 ) & ( j2000 ) & ( ks ) & ( mpc ) + ddo68 & 11271 & 7.15 & 09 56 45.7 & + 28 49 35 & 10.00 & 5.90 + j081239.5 + 483645 & 11298 & 7.16 & 08 12 39.5 & + 48 36 45 & 4.78 & 9.04 + i zw 18 & 805 & 7.18 & 09 34 02.4 & + 55 14 23 & 40.77 & 18.20 + ugc 772 & 11281 & 7.24 & 01 13 39.6 & + 00 52 31 & 5.08 & 11.50 + j210455.3 - 003522 & 11282 & 7.26 & 21 04 55.3 & @xmath800 35 22 & 5.01 & 13.70 + ugca 292 & 11295 & 7.27 & 12 38 40.0 & + 32 46 01 & 5.01 & 3.50 + j141454.1 - 020823 & 11293 & 7.32 & 14 14 54.1 & @xmath802 08 23 & 16.68 & 24.60 + 6df j0405204 - 364859 & 11292 & 7.34 & 04 05 20.3 & @xmath836
49 01 & 5.01 & 11.00 + hs 0822 + 3542 & 11284 & 7.35 & 08 25 55.5 & + 35 32 32 & 5.12 & 12.70 + sbs 1129 + 576 & 11283 & 7.41 & 11 32 02.5 & + 57 22 46 & 14.75 & 26.30 + kug 0937 + 298 & 11301 & 7.45 & 09 40 12.8 & + 29 35 30 & 5.01 & 11.20 + j085946.9 + 392306 & 11299 & 7.45 & 08 59 46.9 & + 39 23 06 & 4.78 & 10.90 + sbs 0940 + 544 & 11288 & 7.48 & 09 44 16.6 & + 54 11 34 & 16.83 & 22.10 + rc2 a1116 + 51 & 11287 & 7.51 & 11 19 34.3 & + 51 30 12 & 11.65 & 20.80 + ugc 4483 & 10559 & 7.54 & 08 37 03.0 & + 69 46 31 & 3.09 & 3.44 + j120122.3 + 021108 & 11286 & 7.55 & 12 01 22.3 & + 02 11 08 & 8.10 & 18.40 + kug 0201 - 103 & 11297 & 7.56 & 02 04 25.6 & @xmath810 09 35 & 13.59 & 22.70 + kug 1013 + 381 & 11289 & 7.58 & 10 16 24.5 & + 37 54 46 & 9.40 & 19.60 + sbs 1415 + 437 & 11291 & 7.59 & 14 17 01.4 &
+ 43 30 05 & 5.11 & 13.70 + hs 1442 + 4250 & 11296 & 7.64 & 14 44 12.8 & + 42 37 44 & 5.19 & 8.67 + rc2 a1228 + 12 & 11290 & 7.64 & 12 30 48.5 & + 12 02 42 & 12.04 & 21.20 + sbs 1102 + 606 & 11285 & 7.64 & 11 05 53.7 & + 60 22 29 & 10.20 & 19.90 + kug 0942 + 551 & 11302 & 7.66 & 09 46 22.8 & + 54 52 08 & 16.02 & 24.40 + kug 0743 + 513 & 11300 & 7.68 & 07 47 32.1 & + 51 11 28 & 5.07 & 8.60 + vii zw 403 & 871 & 7.69 & 11 28 00.4 & + 78 59 39 & 10.25 & 3.87 +
all observations were obtained with the back - illuminated acis - s3 chip aboard _
chandra _ and the galaxies were near the aimpoint on that chip .
we reprocessed the level 1 event files using the latest versions of ciao ( 4.5.2 ) and caldb ( 4.5.6 ) .
x - ray sources were located in the 0.5 - 8 kev band using the ciao tool ` wavdetect ` . using the same settings as @xcite
, we created psf maps using the ciao tool ` mkpsfmap ` , with an enclosed counts fraction ( ecf ) of 90 percent .
the ` wavdetect ` tool was run on the @xmath9 series of pixel scales from 1 to 8 .
the significance threshold was set to @xmath10 , resulting in less than one false detection per image .
we used ` maxiter ` = 10 , ` iterstop ` @xmath11 , and ` bkgsigthresh ` = 0.0001 . in order to determine if the sources we found using ` wavdetect ` were associated with a target galaxy , @xmath12 ellipses were constructed using the given dimensions in the hyperleda database . in comparing the @xmath12 ellipse for hs 1442 + 4250 with its optical counterpart
, we noticed that the ellipse was not centered and needed to be realigned .
the new coordinates were found to be ; @xmath13 , @xmath14 . using this method ,
eight sources were found in the sample of 25 galaxies .
some observations of x - ray binaries have shown them to be removed from the region in which they formed due to kicks during the death of the progenitor to the compact object @xcite .
these kicks can send the x - ray binaries ( that have luminosities @xmath15erg s@xmath16 ( [email protected] kev ) ) to a distance nearly 200 pc from the star - forming region over the lifetime of the binary .
each of the @xmath12 ellipses also have an associated error in their position of 1 arcsec . taking this into account
, we enlarge the dimensions of our ellipses to allow for an extra 200 pc of extent as well as a 1 arcsec positional error . allowing our ellipses to increase in size
, we gain an extra two sources bringing the total to ten . additionally , we examined the number of x - ray sources in our sample as a function of distance from their respective bcd . as the size of each ellipse
is increased the number of sources it contains grows larger .
this is no surprise .
however , the background - subtracted number of sources is insensitive to ellipse size ( remaining at @xmath17 10 sources ) once the dimensions of the ellipses are increased by 20 per cent of their original size .
increasing the dimensions of the ellipses as previously described corresponds to a 25 per cent increase on average .
all ten sources are less than 1.5 arcmin off - axis . the number of background sources , @xmath18 , was estimated using the @xmath19 curves of @xcite , which combines both shallow and deep x - ray observations .
the flux limits used for these estimates were calculated assuming a minimum of 9 counts for source detection ( as discussed in the following section ) .
we find that upon increasing the size of each ellipse , the total number of expected background sources is @xmath20 , an increase of @xmath21 sources .
thus the additional two sources are likely associated with their respective target galaxies and we continue the analysis , keeping these additional sources while also taking into account the expanded @xmath12 when calculating sfr estimates . from @xcite , the probability of detecting a source given the background counts per pixel and the source counts is given by , @xmath22 here @xmath23 is the total number of counts from the source and @xmath24 are parameters that depend on the background counts per pixel .
the probability relation was found using sources less than 1.5 arcmin off - axis .
all 10 of our sources meet this criterion . for our sample ,
the background is less than 0.025 counts per pixel for all observations .
this is the lowest background @xcite used in their data sets .
thus , the completeness estimate is a conservative since we are assuming a higher background for this calculation .
their fit for this background level gives us parameters such that , @xmath25 in order to detect sources with at least a 95 percent probability , there must be a minimum of 9 source counts in the @xmath26 kev band . from this threshold
, we calculated a flux using the ` webpimms ` tool .
the input parameters for this calculation are the count rate ( 9 counts/(exposure time ) ) , the neutral hydrogen column density ( found using ` colden ` ) , the photon index ( @xmath27 ) and the energy range ( @xmath26 kev ) .
this minimum flux was then converted to a luminosity using the distance to the target given in table [ proptable ] .
this luminosity is the assumed completeness luminosity .
all sources associated with our sample of bcds meet the required 9 count minimum .
l c c l r c c c c name & @xmath28 & & angle@xmath29 & @xmath30 & @xmath31 & @xmath32 & @xmath33 + & @xmath34 & & ( deg ) & & & @xmath35 & @xmath35 + i zw 18 & 1.99 & 22.7 & 17.9 & 145.0 & 1 & 0.1002 & 0.737 & 38.10 + vii zw 403 & 3.91 & 71.0 & 48.3 & 6.0 & 2 & 0.0774 & 0.137 & 3.982 , 0.151 + ugc 4483 & 3.23 & 93.3 & 63.0 & 169.3 & 0 & 0.0379 & 0.355 & @xmath8 + ddo 68 & 1.97 & 135.7 & 56.5 & 13.4 & 2 & 0.1597 & 0.316 & 0.489 , 0.620 + ugc 772 & 3.10 & 81.3 & 63.9 & 0.0 & 0 & 0.0417 & 2.404 & @xmath8 + j210455.31 - 003522.2 & 5.66 & 24.2 & 18.5 & 177.9 & 0 & 0.0036 & 3.609 & @xmath8 + sbs 1129 + 576 & 0.87 & 47.6 & 19.2 & 162.0 & 1 & 0.0372 & 4.173 & 10.14 + hs 0822 + 3542 & 4.82 & 25.8 & 16.8 & 141.4 & 0 & 0.0036 & 2.995 & @xmath8 + sbs 1102 + 606 & 0.59 & 41.5 & 24.7 & 104.8 & 0 & 0.0215 & 3.438 & @xmath8 + j120122.3 + 021108.5 & 1.88 & 33.3 & 20.6 & 129.2 & 0 & 0.0100 & 3.784 & @xmath8 + rc2 a1116 + 51 & 1.19 & 24.1 & 18.8 & 90.9 & 1 & 0.0122 & 3.322 & 32.07 + sbs 0940 + 544 & 1.34 & 39.5 & 22.2 & 157.2 & 1 & 0.0458 & 2.820 & 9.80 + kug 1013 + 381 & 1.41 & 26.5 & 24.3 & 40.7 & 0 & 0.0119 & 3.670 & @xmath8 + rc2 a1228 + 12 & 2.47 & 24.9 & 20.3 & 78.8 & 1 & 0.0147 & 3.414 & 3.414 + sbs 1415 + 437 & 1.21 & 28.8 & 18.2 & 58.0 & 0 & 0.0042 & 3.288 & @xmath8 + 6df j0405204 - 364859 & 0.88 & 37.6 & 32.3 & 16.5 & 0 & 0.0094 & 2.150 & @xmath8 + j141454.13 - 020822.9 & 4.17 & 24.8 & 18.5 & 28.4 & 0 & 0.0249 & 3.413 & @xmath8 + ugca 292 & 1.34 & 84.2 & 66.1 & 0.0 & 0 & 0.0433 & 0.219 & @xmath8 + hs 1442 + 4250 & 1.53 & 63.8 & 23.2 & 63.2 & 1 & 0.0119 & 1.303 & 2.164 + kug 0201 - 103 & 2.08 & 31.9 & 16.1 & 101.3 & 0 & 0.0185 & 3.445 & @xmath8 + j081239.52 + 483645.3 & 4.58 & 36.7 & 25.2 & 73.7 & 0 & 0.0071 & 1.619 & @xmath8 + j085946.92 + 392305.6 & 2.44 & 35.2 & 27.7 & 85.0 & 0 & 0.0074 & 2.272 &
@xmath8 + kug 0743 + 513 & 5.17 & 62.7 & 36.6 & 124.2 & 0 & 0.0188 & 1.395 & @xmath8 + kug 0937 + 298 & 1.81 & 46.4 & 24.1 & 92.2 & 0 & 0.0087 & 2.265 & @xmath8 + kug 0942 + 551 & 1.23 & 27.7 & 20.1 & 68.2 & 0 & 0.0266 & 3.327 & @xmath8 + + * notes . * + @xmath36 column densities were found using the ` colden`tool .
+ @xmath29 properties of the @xmath12 ellipses were taken from the hyperledadatabase and modified as described in section [ subsect : observations ] .
+ @xmath37 number of observed sources within the @xmath12 ellipse with luminosity greater than @xmath38 .
+ @xmath39 number of expected background sources within the @xmath12 ellipse , as determined by @xmath19 curves of @xcite .
+ @xmath40 luminosities limits calculated for the @xmath41 kev energy band .
+ @xmath42 luminosity of detected source(s ) in the @xmath41 kev energy band .
star formation rates were determined using two methods .
the first method relates the fuv luminosity of the dwarf galaxy to sfr , as found by @xcite and allows us to calculate a sfr for each galaxy in the sample .
the second method was employed in order to be consistent with @xcite and accounts for uv light escaping the galaxy as well as ir emission due to heating of dust by young stars . however
, this method can not be applied across the whole sample due to lack of observations in the ir .
this problem is discussed further in section [ subsect : sfr2 ] .
observations made by _
galaxy evolution explorer _
( _ galex _ ) were used for the uv data and observations by _ spitzer _
were used in determining the ir contribution .
images corresponding to each galaxy in the sample were found in the _ galex _ archive . in order to calculate the sfr for each galaxy , we follow the procedure of @xcite .
we first needed to extract the count rate for each galaxy .
the pixel values for the image files ( * int.fits ) report photons / pixel / second corrected for relative response . by using the previously constructed @xmath12 ellipses as source apertures , we extracted background - subtracted fuv count rates for the galaxies . for background apertures , we used annular ellipses centered on the source galaxy with dimensions scaled to give @xmath43 the area of the source aperture . using the ciao ` dmstat ` tool
, we obtained the background - subtracted count rates for each galaxy in the fuv band . for _ _ galex _ _ ] ,
the specific flux is proportional to the count rate @xmath44 , @xmath45}.\ ] ] following @xcite , we convert units from to hz in order to have the correct units for luminosity in eq .
( [ sfr_eq ] ) .
@xmath46}\ ] ] from these fluxes , we calculate a luminosity @xmath47 that is corrected for extinction due to galactic dust .
this is done using the method given in @xcite .
the reddening magnitudes found in table [ analysistablesfr ] , were computed using the infrared science archive ( irsa ) tool dust .
then , using the relation found in @xcite , we relate the specific luminosity to the sfr , @xmath48 observations in the nuv band were found in the _ galex _ archive for each galaxy in the sample .
however , only 13 out of 25 galaxies in our sample had been observed by _
this smaller sample of galaxies contains 9 of the 10 observed x - ray sources and nearly 70 per cent of the total star - formation rate as determined by the previous method . in order to compare our results with @xcite
, we determine sfrs for this smaller sample of galaxies , which dominates in its overall contribution to sfr , and find a relation that allows us to estimate the sfr for the other 12 galaxies . for the _ galex _ nuv images
, we use the same approach that we used for the fuv band to calculate a flux , @xmath49},\ ] ] where @xmath44 is still the background - subtracted count rate . using this flux
we calculate a monochromatic luminosity @xmath50erg s@xmath16@xmath51 at 2312 for nuv which is not corrected for extinction .
we used multiband imaging photometer for _ spitzer _ ( mips )
70-@xmath52 m large field images to find the ir component of the sfr .
following @xcite , we used the post basic calibrated data ( pbcd ) products found in the the _ spitzer _ archive[foot : spitzer ] .
we used the same process to extract background - subtracted counts as we did for both uv data .
however , for some images the background aperature s size needed to be reduced or rotated due to the sources proximity to the image s edge and/or other bright sources .
the images are calibrated in units of mjy sr@xmath16 and were converted to jy using the following conversion factor : @xmath53 the flux for each galaxy was then converted to a spectral luminosity @xmath50erg s@xmath16 hz@xmath51 which was then used to estimate the total ir luminosity@xmath54m@xmath55 using , @xmath56 where @xmath57 erg s@xmath16 . from our luminosity measurements we calculated the total sfr , @xmath58 where sfr@xmath59 and sfr@xmath60 are obtained from , @xmath61 in order to account for ir emission due to an old stellar population in normal star - forming galaxies , a correction factor @xmath62 is used . for starburst galaxies , @xmath63 .
@xmath64 is the observed nuv ( 2312 ) luminosity , which is uncorrected for dust attenuation , and @xmath65 is the 8@xmath81000 @xmath52 m luminosity .
results for both methods can be found in table [ analysistablesfr ] .
+ in figure [ fig : sfr_comparison ] , we compare the two methods used to calculate sfr for the sample of 13 galaxies that were observed in all bands .
we find a linear correlation between the two methods that shows the method used by @xcite produces a sfr that is a factor of @xmath66 larger than the fuv method of @xcite .
this is consistent with the 26 per cent increase that @xcite find when comparing their method to that of @xcite that was used by @xcite .
@xcite designed their method for normal , solar metallicities galaxies , whereas @xcite constructed their relation specifically for dwarf galaxies with sub - solar metallicities .
thus , the @xcite relation should be more appropriate for our sample .
however , we used the relation of @xcite to allow a direct comparison with their paper .
we did this by using the correlation to estimate the sfrs for the remaining 12 galaxies that were not observed in the ir .
larger than using the method employed by @xcite.,scaledwidth=50.0% ] l c c c c c c c name & @xmath67 & count rate@xmath68 & @xmath69 & @xmath70 & @xmath71 & count rate@xmath72 & flux@xmath73 + & ( mag ) & ( cps ) & @xmath74 & @xmath75 & @xmath75 & ( cps ) & ( mjy ) + i zw 18 & 0.032 & 10.766 & 5.528 & 70.20 & 68.27 & 28.64 & 14.01 + vii zw 403 & 0.037 & 24.061 & 0.590 & 7.49 & 13.98 & 76.31 & 586.96 + ugc 4483 & 0.033 & 12.575 & 0.247 & 3.14 & 4.16 & 40.24 &
95.74 + ddo 68 & 0.018 & 25.189 & 1.011 & 12.83 & 20.14 & 61.97 & 218.56 + ugc 772 & 0.028 & 4.951 & 0.751 & 9.54 & @xmath8 & 12.32 & @xmath8 + j210455.31 - 003522.2 & 0.066 & 1.241 & 0.468 & 5.94 & @xmath8 & 3.58 & @xmath8 + sbs 1129 + 576 & 0.013 & 3.576 & 3.112 & 39.59 & 62.88 & 11.25 & 20.51 + hs
0822 + 3542 & 0.047 & 1.076 & 0.306 & 3.88 & 4.45 & 3.63 & 3.31 + sbs 1102 + 606 & 0.006 & 4.344 & 2.053 & 26.08 & 34.57 & 11.28 & 14.05 + j120122.3 + 021108.5 & 0.024 & 1.209 & 0.519 & 6.59 & @xmath8 & 3.33 & @xmath8 + rc2 a1116 + 51 & 0.015 & 4.579 & 2.699 & 34.28 & 40.18 & 13.31 & 2.32 + sbs 0940 + 544 & 0.013 & 1.631 & 1.140 & 14.47 &
21.59 & 5.61 & 7.94 + kug 1013 + 381 & 0.015 & 7.945 & 4.155 & 52.77 & 73.19 & 24.25 & 36.13 + rc2 a1228 + 12 & 0.027 & 2.920 & 1.955 & 24.83 & @xmath8 & 9.09 & @xmath8 + sbs 1415 + 437 & 0.009 & 9.962 & 2.217 & 28.16 & 40.38 & 25.95 & 54.10 + 6df j0405204 - 364859 & 0.006 & 5.592 & 0.857 & 10.88 & @xmath8 & 17.01 & @xmath8 + j141454.13 - 020822.9 & 0.058 & 0.670 & 0.733 & 9.31 & @xmath8 & 2.20 & @xmath8 + ugca 292 & 0.016 & 8.797 & 0.139 & 1.77 & 2.40 & 23.28 & 44.50 + hs 1442 + 4250 & 0.013 & 9.156 & 0.849 & 10.78 & 14.74 & 26.52 & 18.67 + kug 0201 - 103 & 0.021 & 1.383 & 0.952 & 12.08 & @xmath8 & 4.47 & @xmath8 + j081239.52 + 483645.3 & 0.051 & 0.897 & 0.130 & 1.65 & @xmath8 & 2.99 & @xmath8 + j085946.92
+ 392305.6 & 0.026 & 0.743 & 0.131 & 1.67 & @xmath8 & 3.32 & @xmath8 + kug 0743 + 513 & 0.070 & 13.290 & 2.041 & 25.92 & @xmath8 & 38.53 & @xmath8 + kug
0937 + 298 & 0.018 & 5.649 & 0.976 & 12.40 & @xmath8 & 17.00 & @xmath8 + kug 0942 + 551 & 0.012 & 1.555 & 1.223 & 15.53 & @xmath8 & 5.14 & @xmath8 + + * notes . *
+ @xmath36 extinctions were computed using the infrared science archive ( irsa ) tool dust .
+ @xmath29 count rate in the far ultraviolet band ( fuv ) of galex observations .
+ @xmath37 luminosity in the fuv band , @xmath76 .
+ @xmath39 star - formation rate as determined by the @xcite method .
+ @xmath40 star - formation rate as determined by the @xcite .
+ @xmath42 count rate in the near ultraviolet band ( nuv ) of galex , @xmath77 .
+ @xmath78 infrared flux in mjy from spitzer in the @xmath79 m range .
from @xcite , the luminosity function for normal metallicity galaxies is of the form , @xmath80 where @xmath81 , @xmath82 is a normalisation constant , @xmath83 is the power law index , @xmath84 is the star formation rate @xmath85 , and @xmath86 is the cut - off luminosity for the xlf . for data that are complete down to some luminosity , @xmath38 ,
the expected number of sources in the range @xmath87 $ ] @xmath88 erg s@xmath51 is found by integrating eq .
( [ lumfunct ] ) , @xmath89 given that eq .
( [ expnum ] ) describes an expected number of sources over a certain luminosity range , the number of sources that are actually observed should follow a poisson distribution .
that is , @xmath90 it should be mentioned that a poisson distribution is being assumed as an approximation to what is actually a binomial distribution , as has been pointed out by @xcite .
since the total number of x - ray sources at * all * luminosities in a galaxy is finite , and the number seen in a certain luminosity range is some fraction of this total , the appropriate distribution to use would be a binomial distribution . however , if the number of observed sources is much smaller than the total number of sources , the binomial distribution may be approximated by the poisson distribution . in order to calculate a probability distribution for our normalisation parameter @xmath82 from the sfr and luminosity limits , we apply bayes theorem .
bayes theorem states that , @xmath91 where @xmath92 represents some observed data , @xmath82 is the desired parameter , the function @xmath93 is the posterior probability distribution , @xmath94 is the likelihood function , and @xmath95 is the prior probability distribution . in our case , we calculate a posterior probability distribution for the xlf normalisation @xmath82 .
we do this by using the poisson distribution in eq .
( [ poisson ] ) as our likelihood function @xmath96 , where the expected number of sources @xmath97 comes from eq .
( [ expnum ] ) .
the data @xmath92 will be composed of our observed number of x - ray sources , sfr , and luminosity limits .
the prior , @xmath98 , provides an initial guess for the probability distribution of @xmath82 .
the bayesian method requires calculating @xmath99 , using eq .
( [ bayesthm ] ) , iteratively for each galaxy .
that is , @xmath99 for one galaxy becomes @xmath98 for the next . in doing the calculation associated with eq .
( [ bayesthm ] ) , one then finds that the posterior has the form of the gamma distribution , @xmath100 where @xmath101 , @xmath102 is the number of observed sources in the @xmath103 galaxy , and @xmath104 . as an initial step , we assume a prior of @xmath105 .
we call this prior the _ _ know - nothing _ _ gamma distribution . the assumption of this prior yields the same results as a frequentist approach .
the gamma distribution is also known as the _ conjugate prior _ to the poisson distribution .
that is , if the likelihood is modelled by a poisson distribution , the posterior will be a gamma distribution .
this allows us to skip the task of iteratively calculating a posterior and simply jump straight to eq .
( [ gammadist ] ) .
some of the important properties of the gamma distribution are given in table [ gammadistprops ] .
our calculation should weight the luminosity function by the completeness function , given by eq .
[ eqn : zezas_fit ] , in order to account for the probability that a source exists but is not detected . in eq .
[ eqn : zezas_fit ] , the number of counts , @xmath23 , depends linearly on the luminosity , so we replace it with @xmath106 , where @xmath107 contains the distance and conversion factors . thus the weighted luminosity function should be written as , @xmath108 integrating this equation from @xmath38 , which is equivalent to @xmath109 by definition , to @xmath110 erg s@xmath51 , one finds the total number of expected x - ray sources weighted by completeness .
if one replaces eq .
[ expnum ] with this result and carries out the calculation as before , one finds that the final result changes the normalization @xmath82 by 0.4 per cent .
the error on the final value of @xmath82 is about 33 per cent ( see below ) , thus we are justified in neglecting the source detection probability .
l l form & @xmath111 + mean & @xmath112 + mode & @xmath113 + median & @xmath114 , where @xmath115 + variance & @xmath116 +
in this paper , the galaxies observed are metal poor or extremely metal poor ( xmpg ) . the luminosity function found by @xcite is for normal metallicity galaxies .
we applied the same functional form for the luminosity function given by @xcite to these metal poor galaxies . following @xcite
, we assumed @xmath117 , @xmath118 , and we used the sfrs determined by their method for galaxies that had uv and ir observations .
for the 12 galaxies that were not observed in the ir , we scaled the sfrs determined from the fuv by a factor of 1.23 , as discussed at the end of section [ subsect : headsfr ] . from figure [ probdist_brorby_grimm_prior1 ] , one can see the probability distribution for our sample of bcds . in the @xcite paper , a value of @xmath119
is given as the normalisation parameter in eq .
( [ lumfunct ] ) . from our analysis
, we calculate an xlf normalisation of @xmath120 .
thus , by calculating the overlap between the two distributions we find that we can exclude the hypothesis that normal and bcd galaxies have the same normalisation parameter , @xmath82 , with probability of @xmath1 . as an additional check to these statistical calculations , we use a frequentist approach to find the likelihood of detecting 10 or more x - ray sources given the model of @xcite . from our sfrs , luminosity limits , expected background sources , and using the xlf normalisation of 1.49
, we expect to detect 1.75 sources .
assuming poisson statistics for the number of sources detected with a mean of 1.75 , detecting 10 or more sources has a probability of @xmath121 . using the sfrs we calculated from the fuv measurements
, we find a normalisation of @xmath122 .
our results do not take measurement uncertainty of sfr into account .
all quoted uncertainties are statistical in nature . to this end , we now devise a method in which we may account for the uncertainty in sfr measurements . when calculating star formation rates , a typical observational uncertainty of @xmath123 percent is quoted .
this arises from complexity in the physical processes of the system which produce scatter in the data around the simplified models used to describe sfr as functions of luminosity in bands such as uv or ir @xcite .
assuming the errors in sfr are gaussian about our calculated values , we generated @xmath124 samples of galaxies and calculated the mean and standard deviation .
the mean matched the original ( 14.5 , as expected ) and the standard deviation was 1.5 . using this information , we calculated the probability distribution function ( pdf ) for an xlf normalisation with @xmath125 .
this new _ lower limit _ pdf excludes the @xcite value with a confidence of @xmath126 ( see figure [ pdf_lower ] ) .
this gives us further confidence in the significance of our result .
, low metallicity bcds ( dashed line ; brorby ) @xmath120 .
the difference in appearance ( and lack of overlap ) in the probability distributions of the normalisation parameter is evident for the two different populations of galaxies . by calculating their overlap
, we may exclude the hypothesis that the xlf normalisation parameters agree with @xmath1 confidence .
the probability distributions are normalized with respect to their peak values ( as opposed to area ) in order to enhance visual comparison.,scaledwidth=50.0% ] , bcds ( dashed line ; brorby ) @xmath120 , lower limit for bcds ( dotted line ; brorby ) @xmath127 .
same as figure [ probdist_brorby_grimm_prior1 ] , but with a lower limit pdf that was calculated based on sfr uncertainty .
the _ lower limit _ pdf excludes the @xcite value for the xlf normalisation @xmath82 with a confidence of @xmath126,scaledwidth=50.0% ] and @xmath5 , respectively .
the cross represents the location of our reported value of @xmath120 and our assumed values of @xmath117.,scaledwidth=40.0% ] the bayesian calculation may be extended to two or more parameters in order to determine multivariate probability distribution functions . using the methods described in section [ sect : bayes ] , we allowed both the xlf normalisation @xmath82 and the power law index @xmath83 to be simultaneously described by a two - variable probability distribution function .
the 68 per cent and 95 per cent confidence regions are shown as contours in figure [ fig : contour ] .
finding the point of maximum probability for the pdf produces simultaneous estimates for both @xmath82 and @xmath83 . in our case ,
@xmath4 and @xmath5 are the most probable values for the xlf normalisation and power law index , respectively .
these values are consistent with our assumed , fixed value of @xmath117 and the corresponding xlf normalisation of @xmath120 , which are represented as a cross in figure [ fig : contour ] .
being nearby analogues to low - metallicity galaxies in the early universe , evidence for enhanced hmxb populations in bcds could suggest hmxbs played a major role in the epoch of reionization .
the analysis in this paper shows that the population of these x - ray binaries in bcds is indeed significantly larger than the expected value calculated from the xlf for near - solar metallicity galaxies , thus implying a heightened hmxb population in the early universe .
our analysis shows that for our sample of known bcds the xlf normalisation is larger than that of near - solar metallicity galaxies by a factor of @xmath3 .
specifically , we find a normalisation of @xmath120 as compared to the value of @xcite , @xmath119 , which is consistent with previous studies of near - solar metallicity galaxies .
simultaneous determination of the xlf normalisation and power law index yields @xmath4 and @xmath5 .
thus , we find that the power law index for our sample of low - metallicity bcds is consistent with that of near - solar metallicity galaxies .
this suggests that the enhanced production of hmxbs relative to sfr may be modelled by the same xlf as normal galaxies , except with an enhanced normalisation parameter .
previous studies have also found enhanced production of x - ray binaries relative to sfr in low metallicity galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . the low metallicity environments in which these objects were formed is thought to be similar to the first small galaxies of the early universe @xcite .
these findings suggest x - ray production was enhanced relative to sfr in the early universe , thus making x - rays an important and previously overlooked contributor to heating during the epoch of reionisation @xcite .
using observational data from @xmath128 surveys , @xcite showed that the @xmath129/sfr relation evolved with redshift , which they ascribed to changes in the hmxb metallicities . building upon this work
, @xcite used previous large scale simulations to synthesize populations of x - ray binaries from the time galaxies first were forming up to the present time .
the same observational data used by @xcite were used to constrain their models from which they found that at @xmath130 hmxbs began to dominate the x - ray emission over lmxbs and agn , increasing by about a factor of 5 in their x - ray luminosity relative to sfr ( @xmath129/sfr ) .
most simulations to date have assumed a power law spectrum in the x - rays for these hmxbs ( e.g. , * ? ? ? * ; * ? ? ?
however , @xcite has recently shown that the total spectrum of these early universe populations are likely dominated by hmxbs with curved spectra in high luminosity states , affecting and weakening constraints on emission from hmxbs in the soft x - ray background .
curved spectra would also modify the k - correction in such a way that is consistent with a large enhancement of @xmath129/sfr at redshifts of @xmath131 when applied to deep x - ray surveys .
all of these results may be used to provide better estimates of the x - ray contribution to the heating of the early universe .
the conditions associated with this heating affect the ionisation morphology during the epoch of reionisation via thermal feedback , and also affect the time at which reionisation was completed . | x - rays from binaries in small , metal - deficient galaxies may have contributed significantly to the heating and reionization of the early universe .
we investigate this claim by studying blue compact dwarfs ( bcds ) as local analogues to these early galaxies .
we constrain the relation of the x - ray luminosity function ( xlf ) to the star - formation rate ( sfr ) using a bayesian approach applied to a sample of 25 bcds .
the functional form of the xlf is fixed to that found for near - solar metallicity galaxies and is used to find the probability distribution of the normalisation that relates x - ray luminosity to sfr .
our results suggest that the xlf normalisation for low metallicity bcds ( 12+log(o / h ) @xmath0 7.7 ) is not consistent with the xlf normalisation for galaxies with near solar metallicities , at a confidence level @xmath1 .
the xlf normalisation for the bcds is found to be @xmath2 , a factor of @xmath3 higher than for near solar metallicity galaxies .
simultaneous determination of the xlf normalisation and power law index result in estimates of @xmath4 and @xmath5 , respectively .
our results suggest a significant enhancement in the population of high - mass x - ray binaries in bcds compared to the near - solar metallicity galaxies .
this suggests that x - ray binaries could have been a significant source of heating in the early universe .
[ firstpage ] galaxies : blue compact dwarf stars : formation x - rays : galaxies |
one of the most important questions in quantum information theory is the classification of states containing inequivalent types of quantum correlations .
many results exist in the case when the parts of the composite system are effectively distinguishable , but much less is known in the case of identical particles .
when discussing entanglement at short scales , one can expect that indistinguishability of identical particles changes the nature of quantum correlations in a fundamental way .
understanding such correlations and inequivalent types of entanglement of identical particles is also of great importance outside quantum information theory @xcite .
following recent success in describing the algebra of polynomial invariants under the local unitary group of multipartite quantum systems in the limiting case of large hilbert space dimensions @xcite , in this paper we turn to the question of finding a similar description when identical particles are present .
in the case of distinguishable subsystems one has a natural notion of local operations : these can be realized by acting on the parts separately .
lack of distinguishability of particles , however , prevents one to apply this definition directly to identical particles , and indeed , the notion of entanglement in this case is not entirely settled yet and several different definitions exist @xcite . in this paper
we regard unitary evolutions generated by hamiltonians built from single - particle observables as the analogue of local transformations as e.g. in ref .
the state space is then of the form @xmath0 ( bosons ) or @xmath1 ( fermions ) , and a local transformation is described by the appropriate restriction of @xmath2 ( to be denoted by @xmath3 and @xmath4 respectively ) for some unitary operator @xmath5 acting on @xmath6 . in some cases
the physical setting provides a splitting of the single particle hilbert space into the direct sum of orthogonal subspaces .
the most natural example is when the position of a particle can be localized in one of a few possible regions of space having vanishing or negligible overlap . in this situtation
we may assume that the single - particle hamiltonian leaves the orthogonal subspaces fixed , and this further restricts the possible transformations leading to conservation of local particle numbers .
finally , one may have to consider a quantum mechanical system where both bosons and fermions are present and the most general hilbert space obtained this way has the form @xmath7 on which the local unitary operations @xmath8 act where @xmath9 .
note that the same hilbert space can also be interpreted as the state space of a system of @xmath10 types of particles where different types can be distinguished but particles of the same type are identical .
the outline of this paper is as follows . in section [ sec : luinv ]
we introduce the object which we study , namely , the inverse limit of the algebras of invariant polynomials under products of unitary groups starting over any irreducible representation where the limit is taken as the dimensions tend to infinity . here
we state that the inverse limit does not contain `` too many '' elements , that is every element is already represented for some finite values of the dimensions .
section [ sec : symfunc ] summarizes some concepts related to the representation theory of the unitary groups and its connection to symmetric groups and their representations . in section [ sec : hilbser ]
we derive a formula for the dimensions of the homogeneous subspaces of the inverse limit in terms of induced characters .
section [ sec : mat ] prepares the detailed study of bosonic and fermionic systems by collecting some properties of regular bipartite graphs and integer stochastic matrices . in section [ sec : alggen ] we utilize these tools in the simplest nontrivial case : we describe invariants of pure states of a quantum system containing any number of identical bosons .
the main result is stated in section [ sec : various ] which deals with the most general quantum systems containing different types of bosonic , fermionic or distinguishable particles in any combination .
we find a convenient way of encoding elements of an algebraically independent generating set by certain graphs . in section [ sec : mixed ] we show that the problem of finding invariants of mixed states can be completely reduced to the case of pure states .
some concluding remarks follow in section [ sec : conclusion ] . for the readers convenience
some more technical proofs are presented outside the main text in section [ sec : proofs ] .
let @xmath11 , @xmath12 be nonempty partitions and @xmath13 , and let us consider the complex hilbert space @xmath14 .
this space carries an irreducible representation of @xmath15 , the group of local unitary transformations . for @xmath16 ( with respect to the componentwise partial order )
we have the usual inclusions @xmath17 sending an @xmath18-tuple of complex numbers to the first @xmath19 components which gives rise to the inclusion @xmath20 ( obtained by tensoring the images of the former inclusions under the appropriate schur functor ) . regarding @xmath21 as the subgroup of @xmath22 acting on the first @xmath23 entries in the @xmath24th factor and stabilizing the image of @xmath25
, one can see that @xmath25 is an @xmath21-equivariant linear map .
let @xmath26 denote the algebra of @xmath21-invariant polynomial functions over @xmath27 , regarded as a real vector space .
polynomial functions can be identified with elements in @xmath28 , the symmetric algebra on @xmath29 on which an action of @xmath21 is induced and we have @xmath30 as in ref .
@xcite , we use a grading which is different from the usual one in a factor of two , and call homogeneous of degree @xmath31 the polynomials which are of degree @xmath31 both in the coefficients and their conjugates .
this convention is convenient as an invariant must have the same degree in the coefficients and their conjugates , as seen from the fact that @xmath32 acts on @xmath33 by multiplication with @xmath34 .
the inclusions @xmath20 induce morphisms of graded algebras @xmath35 defined by @xmath36 .
thus we obtain the inverse system @xmath37 of graded algebras , the inverse limit of which will be denoted by @xmath38 and called the algebra of lu - invariants : @xmath39 note that @xmath26 is a quotient of @xmath38 .
the next lemma implies that every element of @xmath38 is already represented in some @xmath26 .
[ lem : lowdimiso ] let @xmath11 and @xmath40 .
then the restriction of @xmath41 to the subspace of elements of degree at most @xmath42 is an isomorphism . for the proof see sec .
[ sec : proofs ] .
in this section we collect some well - known facts related to the representations of unitary and symmetric groups and their characters . for more details see e.g. refs .
@xcite .
recall that the isomorphism class of a polynomial representation of the unitary group @xmath43 is uniquely determined by its character which is a symmetric polynomial of the eigenvalues with integer coefficients .
irreducible representations correspond to schur polynomials indexed by partitions of integers into at most @xmath18 parts .
we are interested in the large @xmath18 limit , therefore it is convenient to work with the algebra @xmath44 of symmetric functions in infinitely many variables .
a basis of @xmath44 is the set @xmath45 where @xmath46 is the schur polynomial labelled by @xmath47 and @xmath48 denotes the fact that @xmath47 is a partition of @xmath31 .
the usual inner product on the space of class functions on the compact group @xmath43 translates to an inner product on @xmath44 defined by @xmath49 direct sums and tensor products of the representations correspond to sums and products of the corresponding symmetric functions , respectively . given two representations @xmath50 and @xmath51 , we can form their composition @xmath52 and its character can be calculated as the plethysm of that of @xmath53 with @xmath54 .
this operation may be defined on @xmath44 by @xmath55=p_{nn'}$ ] and @xmath56=p_n[f]$ ] for @xmath57 and extending to be an algebra endomorphism in the outer variable .
the ring of symmetric functions is also connected to the representation theory of the symmetric groups via the characteristic map . denoting the space of class functions on @xmath58 by @xmath59 and the @xmath18th homogeneous part of @xmath44 by @xmath60 ,
we define @xmath61 by @xmath62 where @xmath63 and @xmath64 denotes its value on the conjugacy class labelled by @xmath65 , that is , on elements of cycle type @xmath65 , and @xmath66 if @xmath65 has @xmath67 ones , @xmath68 twos etc .
we may also regard @xmath69 as a map from @xmath70 to @xmath44 .
note also that @xmath59 is equipped with the usual inner product , and hence also their direct sum .
characters of irreducible representations of @xmath58 are indexed by partitions of @xmath18 , and will be denoted by @xmath71 where @xmath72 .
these form an orthonormal basis of @xmath59 , and an important property of @xmath69 is that @xmath73 , in particular , @xmath69 is an isometry .
the sum , product and plethysm operations also have a description in terms of representations of the symmetric group . clearly , the direct sum of representations corresponts to the sum in @xmath44 .
the product translates to the induction product defined as follows .
let @xmath74 be the character of the representation @xmath75 of @xmath76 for @xmath77 .
then @xmath78 can be regarded as the subgroup of @xmath79 containing bijections which permute the first @xmath80 numbers among themselves and similarly the last @xmath81 numbers .
the induction product of the two representations is @xmath82 it can be shown that the image of its character under @xmath69 is @xmath83 finally , plethysm can be described in terms of the wreath product of representations .
similarly as before , @xmath84 can be regarded as a subgroup of @xmath85 .
its normalizer is isomorphic to the wreath product @xmath86 .
we can think of the wreath product @xmath86 as the set @xmath87 with multiplication defined by @xmath88 now let @xmath89 be a representation of @xmath90 with character @xmath91 and @xmath92 a representation of @xmath93 with character @xmath94 .
we can define a representation of @xmath86 on @xmath95 by @xmath96 we denote the character of this representation by @xmath97 .
regarding @xmath86 as the subgroup of @xmath85 above , we have the following equality : @xmath98\ ] ] finally , applying to irreducible representations and expanding in the basis of schur functions ( equivalently : irreducible characters of the symmetric groups ) we write : @xmath99 and @xmath100=\sum_{\nu\vdash km}m_{\lambda\mu\nu}s_\nu\ ] ] or as the composition of the corresponding schur functors @xmath101
let @xmath102 , @xmath12 be nonempty partitions and @xmath103 such that @xmath104 .
our aim is to calculate the dimension @xmath105 of the @xmath31th graded subspace of @xmath26 .
this subspace can be written as @xmath106 hence if we write @xmath107 as the direct sum of irreducible representations of @xmath21 , then @xmath105 is the sum of the squares of multiplicities . for @xmath48
we have the isomorphism @xmath108 where @xmath109 is the multiplicity of the irreducible representation of @xmath93 indexed by the @xmath47 in the tensor product of irreducible representations corresponding to @xmath110 .
now we can write @xmath111 and therefore @xmath112 where we have used that the appearing characters are all real , irreducible characters @xmath113 form an orthonormal basis and introduced the notation @xmath114 and @xmath115 to simplify the last expression we need the following .
[ lem : wrindmany ] let @xmath116 be finite groups , @xmath117 , @xmath118 and let @xmath119 be a class function on @xmath120 for every @xmath24
. then @xmath121 the proof can be found in sec .
[ sec : proofs ] using this result the dimension of the degree @xmath31 homogeneous subspace of @xmath38 can be rewritten as @xmath122 where @xmath114 and @xmath123 .
note that in the @xmath124 case @xmath125 is the trivial group , and therefore @xmath126 corresponding to the one dimensional space of constant polynomials . in the special case @xmath127
the groups reduce to @xmath128 and @xmath129 , and we have that @xmath105 equals the number of orbits of @xmath130 under @xmath131 acting via left and right multiplication , or equivalently , the number of orbits of @xmath132 under @xmath93 acting by simultaneous conjugation @xcite . in the remaining sections
we will restrict ourselves to the study of the cases where the ferrers diagrams of the appearing partitions have either a single row or a single column , which corresponds to a multipartite quantum system with various types of bosonic and fermionic particles .
the next two sections deal with the simplest case of a system of identical bosons .
following the strategy of refs .
@xcite , in this section we introduce certain combinatorial objects which can be used to conveniently label a basis of @xmath133 .
an integer stochastic matrix is a square matrix with nonnegative integer entries and such that the sum of entries in each row and in each column is the same . clearly , if @xmath134 and @xmath135 are two such matrices with line sums @xmath136 , then the block - diagonal matrix @xmath137\ ] ] built from them is also an integer stochastic matrix with line sums @xmath136 .
we will call integer stochastic matrices differing only in a permutation of rows and columns equivalent .
clearly , this is an equivalence relation and @xmath138 gives rise to a well defined operation on the equivalence classes , to be denoted by @xmath139 .
a matrix with nonnegative integer entries may also be thought of as biadjacency matrix of a bipartite graph @xmath140 on a labelled vertex set , say @xmath141 and @xmath142 ( and possibly with multiple edges ) together with a fixed order of colour classes , and this correspondance is a bijection .
our convention will be that rows correspond to vertices in @xmath143 while columns correspond to vertices in @xmath144 .
clearly , the biadjacency matrix is an @xmath145 integer stochastic matrix iff the corresponding graph is @xmath136-regular and @xmath146 . under this map
the binary operation @xmath139 corresponds to the disjoint union of bipartite graphs , which we define so that we keep track of the ordering of colour classes , i.e. for @xmath140 and @xmath147 we have @xmath148 . note that this is important as the bipartition of a disconnected bipartite graph is not unique .
passing to equivalence classes of the integer stochastic matrices corresponds to forgetting the labels of vertices .
we will make use of another description in terms of permutations .
let @xmath149 be a permutation of the numbers @xmath150 .
to @xmath151 we can associate an @xmath136-regular @xmath31 by @xmath31 bipartite graph as follows .
let @xmath141 and @xmath152 be the two vertex sets , and for each @xmath153 add an edge joining @xmath154 with @xmath155 where @xmath156 denotes the smallest integer not less than @xmath157 . in other words , there are @xmath158 edges joining @xmath159 with @xmath160 iff @xmath158 of numbers in the range @xmath161 are mapped by @xmath151 into the range @xmath162 .
it is not hard to see that @xmath163 are mapped to the same graph with labelled vertices iff @xmath164 with @xmath165 regarded as the subgroup whose elements permute the numbers @xmath166 among themselves for each @xmath167 .
reordering the labels of the vertices amounts to left and right multiplication with an element of @xmath168 regarded as the subgroup permuting the @xmath31 blocks of numbers @xmath166 with @xmath169 without reordering the numbers inside the blocks .
the two subgroups generate the normalizer of the former in @xmath170 , which is isomorphic to @xmath171 . in the following
we will always assume @xmath171 to be this subgroup in @xmath170 .
to sum up , equivalence classes of @xmath145 integer stochastic matrices with line sum @xmath136 are in bijection with @xmath136-regular bipartite graphs with a fixed bipartition into two @xmath31-element vertex sets , which in turn are in bijection with elements of the double coset space @xmath172 .
[ fig : l3m3 ] shows the five possible graphs in the @xmath173 case .
-regular bipartite graphs on six vertices .
representatives of the corresponding equivalence classes of permutations ( from left to right ) are : @xmath174 , @xmath175 , @xmath176 , @xmath177 , @xmath178 .
[ fig : l3m3 ] ]
our next aim is to prove that @xmath133 is free by presenting an algebraically independent generating set .
let @xmath6 be a hilbert space , and @xmath179 , and consider the @xmath136-boson state space @xmath180 .
invariant polynomial functions on this space are in bijection with elements of the space @xmath181 note that @xmath182 can be regarded as a subspace of @xmath183 , and as this space comes equipped with an inner product induced from that of @xmath6 , we also have the orthogonal projection @xmath184 .
this implies that @xmath185 can be viewed both as a subspace and as a quotient of @xmath185 .
recall that @xmath186 acts on @xmath183 by the defining representation on each factor and on this space there is an action of @xmath170 permuting the factors and these two actions commute .
schur - weyl duality states that the algebra homomorphism @xmath187 is surjective , and it is injective iff @xmath188 .
in these formulae we can also write @xmath189 instead of @xmath186 . therefore we also have a surjection from the group algebra @xmath190 to the space of degree @xmath31 polynomial invariants .
let @xmath191 be an orthonormal basis of @xmath6 , and @xmath192 its dual basis .
vectors of the form @xmath193 with @xmath194 form a basis of @xmath195 .
then our surjection is defined on the group elements as @xmath196}:=\sum_{j_1,\ldots , j_{lm}=1}^{n}(e_{j_1}\vee\cdots\vee e_{j_k})(e_{j_{l+1}}\vee\cdots\vee e_{j_{2l}})\cdots(e_{j_{(m-1)l+1}}\vee\cdots\vee e_{j_{ml}})\cdot \\ \cdot(e^*_{j_{\sigma(1)}}\vee\cdots\vee e^*_{j_{\sigma(l)}})\cdots(e^*_{j_{\sigma((m-1)l+1)}}\vee\cdots\vee e^*_{j_{\sigma(ml)}})\end{gathered}\ ] ] and extended linearly . as the symmetrized product @xmath197 as well as
the product in @xmath198 is commutative , and we sum over every possible @xmath199-tuple of integers ( in particular , over permutations of any fixed @xmath199-tuple ) , we have that the image of @xmath200 under this map is the same as the image of any element in the double coset @xmath201:=(s_l\wr s_m)\sigma(s_l\wr s_m)$ ] . in other words , we have a well - defined surjection from a vector space freely generated by the double coset space @xmath172 to the space of degree @xmath31 invariant polynomials .
but the dimension of the two spaces can be seen to be equal , so that this surjection is in fact an isomorphism if @xmath202 . indeed , in this special case eq .
( [ eq : stabdim ] ) reduces to @xmath203 using the well - known formula @xmath204 for the number of double cosets .
what we have seen so far is that a basis of @xmath133 can be labelled by @xmath136-regular bipartite graphs or equivalently by integer stochastic matrices with line sums @xmath136 up to permutation of rows and columns .
it is not hard to see ( and will be proved later in a more general setting ) that under these bijections multiplication of the basis elements corresponds to the disjoint union and to the operation @xmath139 defined above , respectively , implying that any element of the basis can be uniquely written as the product of basis elements corresponding to _ connected _ graphs .
our results are summarized in the following @xmath133 is a free algebra , and an algebraically independent generating set is @xmath205 where @xmath206 is the set of elements of @xmath207 corresponding to connected graphs by the above bijection .
we turn to the local unitary invariants of the most general quantum systems containing various types of bosonic , fermionic and distinguishable particles .
the starting point is the stable dimension formula of eq .
[ eq : stabdim ] . in the present case
@xmath10 is arbitrary while the partitions @xmath12 are either of the form @xmath208 or @xmath209 corresponding to bosons or fermions ( so that @xmath210 ) .
let @xmath211 and @xmath212 as before .
the degree @xmath31 subspace of @xmath38 is @xmath213 by mackey s theorem@xcite where @xmath214 and @xmath215 and in the sum @xmath216 runs over a system of representatives of the set of double cosets .
note that the inner product indeed depends on @xmath216 only through the double coset @xmath217 .
observe that each term in the sum is either @xmath218 or @xmath219 , being the inner product of one dimensional and hence irreducible characters , therefore @xmath220 holds . in the following
we will give a set of @xmath221 invariants spanning the degree @xmath31 homogeneous subspace , and we shall see that when strict inequality holds in eq .
( [ eq : dmbound ] ) , some invariants become zero while the remaining ones form a basis .
let @xmath222 be a fixed @xmath10-tuple of integers such that @xmath223 .
first observe that @xmath224 can be thought of both as a subrepresentation and as a quotient of @xmath225 via symmetrization / antisymmetrization .
hence we have a surjection @xmath226 which is @xmath21-equivariant , therefore we also have a surjection @xmath227 here the right hand side can be identified with the degree @xmath31 homogeneous subspace of @xmath38 .
let @xmath228 denote the standard basis of @xmath229 and let us introduce the following notation : @xmath230 so that @xmath231 ( @xmath232 ) corresponds to symmetrized ( antisymmetrized ) tensor products , respectively .
elements of the dual bases will be denoted by @xmath233 etc . by schur - weyl duality
we have the surjection @xmath234 defined by @xmath235 on @xmath10-tuples of permutations and extended linearly . composing this with the surjection
above we have the following map : @xmath236}:=\sum(e_{\lambda_1,i_1 ^ 1i_1 ^ 2\ldots i_1^{l_1}}\otimes e_{\lambda_2,i_2 ^ 1i_2 ^ 2\ldots i_2^{l_2}}\otimes\cdots\otimes e_{\lambda_k , i_k^1i_k^2\ldots i_k^{l_k}})\cdots \\ ( e_{\lambda_1,i_1^{(m-1)l_1 + 1}i_1^{(m-1)l_1 + 2}\ldots i_1^{ml_1}}\otimes \cdots\otimes e_{\lambda_k , i_k^{(m-1)l_k+1}i_k^{(m-1)l_k+2}\ldots i_k^{ml_k}})\cdot \\ ( e^*_{\lambda_1,i_1^{\sigma_1(1)}i_1^{\sigma_1(2)}\ldots i_1^{\sigma_1(l_1)}}\otimes e^*_{\lambda_2,i_2^{\sigma_2(1)}i_2^{\sigma_2(2)}\ldots i_2^{\sigma_2(l_2)}}\otimes\cdots\otimes e^*_{\lambda_k , i_k^{\sigma_k(1)}i_k^{\sigma_k(2)}\ldots i_k^{\sigma_k(l_k)}})\cdots \\ ( e^*_{\lambda_1,i_1^{\sigma_1((m-1)l_1 + 1)}i_1^{\sigma_1((m-1)l_1 + 2)}\ldots i_1^{\sigma_1(ml_1)}}\otimes \cdots\otimes e^*_{\lambda_k , i_k^{\sigma_k((m-1)l_k+1)}i_k^{\sigma_k((m-1)l_k+2)}\ldots i_k^{\sigma_k(ml_k)}})\end{gathered}\ ] ] where the sum is over all possible values of the indices :
@xmath237 ( @xmath238 ) and @xmath239 $ ] denotes the double coset @xmath240 .
note that @xmath241 does not change ( up to sign ) upon changing the order of the appearing indices , and also the factors in each term can be reordered in an arbirary way .
taking into account that we sum over every possible value of the indices , the first transformation is realized when any of the @xmath242 is multiplied from either side by a permutation in the young subgroup @xmath243 while the latter amounts to simultaneous multiplication of the permutations @xmath244 from the left or from the right by the subgroup of @xmath245 isomorphic to @xmath93 which permutes simultaneously the blocks fixed by these young subgroups . but
the subgroup generated by these is precisely @xmath246 so that indeed we have an up - to - sign well defined invariant for each element of @xmath247 .
the sign can be fixed by requiring the invariant to be positive for separable states .
this already implies that when in eq .
( [ eq : dmbound ] ) equality holds , a basis of the degree @xmath31 homogeneous subpace is obtained this way . similarly to the special case considered in sec .
[ sec : alggen ] we can describe @xmath247 in a purely combinatorial way in terms of certain graphs as follows .
let @xmath248 .
now we can draw a bipartite graph with vertices @xmath249 and with edges of @xmath10 different colours , adding an edge of the @xmath250th colour connecting @xmath251 with @xmath252 for every @xmath253 .
in other words , there are @xmath158 edges of colour @xmath250 joining @xmath159 with @xmath254 iff @xmath158 of numbers in the range @xmath255 are mapped by @xmath242 into the range @xmath256 .
finally we forget the labels of the vertices but keep the order of the colour classes , and this way obtain a bijection between @xmath247 and the set of isomorphism classes of bipartite graphs with a fixed bipartition into two @xmath31-element vertex sets and edges of @xmath10 different colours with the subgraph given by edges of colour @xmath250 being @xmath257-regular .
the set of these isomorphism classes will be denoted by @xmath258 , and connected ones by @xmath259 as an illustration fig .
[ fig : bipartite ] shows the graph obtained in the @xmath260 , @xmath261 , @xmath262 , @xmath263 case from the pair of permutations @xmath264 .
( left ) , @xmath265 ( right ) and the combined graph encoding the equivalence class of the pair in the @xmath260 , @xmath261 , @xmath262 , @xmath263 case .
( colour online)[fig : bipartite ] ] note also that the @xmath266 special case corresponds to distinguishable particles , and the resulting graphs can be identified ( after merging pairs of vertices along one of the colours ) with the graph coverings of ref .
@xcite .
alternatively , we can identify the double cosets with orbits of @xmath10-tuples of @xmath145 integer stochastic matrices such that the @xmath250th matrix has line sums @xmath257 , under simultaneous multiplication from the left or from the right by permutation matrices .
the next step is to understand multiplication in @xmath267 in terms of graphs , i.e. to find the structure constants with respect to the basis given above .
let @xmath268 denote the disjoint union @xmath269 .
this set comes equipped with an operation @xmath270 induced by the disjoint union of graphs .
similarly , let @xmath271 denote the disjoint union @xmath272 .
let @xmath273 denote the operation defined on @xmath274 as follows .
for @xmath275 and @xmath276 the usual inclusions @xmath277 of young subgroups determine an element @xmath278 in @xmath279 .
we define @xmath239\star[\sigma'_1,\ldots,\sigma'_k]$ ] to be the double coset @xmath280 $ ] .
it is easy to see that this operation is well defined , and turns @xmath281 into a commutative monoid ( the identity being the only element of @xmath282 ) .
in addition , the map @xmath283 described above becomes this way an isomorphism of monoids .
as @xmath268 is freely generated by the subset @xmath271 , the former is also freely generated by preimages of connected graphs .
[ lem : product ] let @xmath275 and @xmath276 be arbitrary elements
. then @xmath284}\cdot f_{[\sigma'_1,\ldots,\sigma'_k]}=f_{[\sigma_1,\ldots,\sigma_k]\star[\sigma'_1,\ldots,\sigma'_k]}\ ] ] the proof can be found in sec .
[ sec : proofs ] .
it follows that we have a surjective homomorphism from the semigroup algebra @xmath285 defined by @xmath286 and extended linearly .
when equality holds in eq .
( [ eq : dmbound ] ) , this is an isomorphism .
we turn now to the case when we have a strict inequality in eq .
( [ eq : dmbound ] ) .
we will see that this means only a minor modification in the structure of the inverse limit , namely , to each _ nonzero _ term in eq .
( [ eq : dmcosets ] ) corresponds an invariant via eq .
( [ eq : fdef ] ) forming a basis of the degree @xmath31 homogeneous subspace , and for each vanishing term the corresponding invariant is identically zero .
[ lem : vanishing ] let @xmath287 and @xmath246 as above and @xmath288 such that @xmath289 holds . then @xmath290}=0 $ ] . for the proof see sec .
[ sec : proofs ] now we have everything at hand to state the following : let @xmath291 and let @xmath206 denote the subset @xmath292 of @xmath293 .
then @xmath294 is a basis of @xmath38 and @xmath38 is freely generated as an algebra by the set @xmath295 we have seen that @xmath296 generates @xmath38 as a vector space . removing the zero vector does not change this property , hence @xmath297 is also a generating set .
the number of degree @xmath31 elements in @xmath297 is at most the number of double cosets @xmath298\in h_m\backslash g_m / h_m$ ] such that @xmath299 and this number is equal to @xmath105 by eq .
( [ eq : dmcosets ] ) , therefore must form a basis .
an element @xmath300 of @xmath293 can be uniquely written as the disjoint union of connected graphs i.e. elements of @xmath206 , because an invariant corresponding to a connected graph outside @xmath206 is zero . by lemma
[ lem : product ] this means that @xmath301 is in a unique way the product of elements of @xmath302 .
it remains to settle the question whether we have equality or not in eq .
( [ eq : dmbound ] ) for a given @xmath10-tuple @xmath303 .
equivalently , we wish to determine if there exist an element @xmath216 in @xmath304 for some @xmath31 such that @xmath289 clearly this can not happen if @xmath305 is the restriction of some character of @xmath304 .
we may assume that @xmath306 since for @xmath124 and @xmath307 we have @xmath308 .
in the @xmath306 case @xmath309 has exactly @xmath310 distinct one dimensional characters : we can chose the trivial or the alternating character for each factor .
let @xmath311 be such a choice .
we wish to find out its value on an element @xmath312 of @xmath313 . in the image of this element in @xmath304 the @xmath250th factor
is a permutation of @xmath31 blocks of size @xmath257 determined by @xmath314 and inside the blocks the permutations @xmath315 act .
we denote this permutation by @xmath316 . with this notation
we have @xmath317 where @xmath318 is the sign of the permutation and therefore the value of the character of @xmath304 we are looking for is @xmath319 let @xmath320 if @xmath321 and @xmath322 if @xmath323 . in the special case
@xmath324 the two representations are the same , and we can safely choose any of @xmath325 for @xmath326 . this freedom will be used shortly .
the value of @xmath305 on @xmath327 is @xmath328 the two values coincide for every element iff @xmath329 is even and for all @xmath330 either @xmath324 or @xmath331 .
note that when for at least one @xmath250 we have @xmath324 then we can always choose @xmath332 so that @xmath333 is even .
if @xmath334 for all @xmath330 then the condition means that the total number of fermions is even . on the other hand , when @xmath335 and the total number of fermions is odd and @xmath336 we can always find an element @xmath337 such that @xmath289 as an important special case we list some values of @xmath105 for @xmath133 and @xmath338 , that is , for a system of @xmath136 bosons and fermions , respectively in table [ tab : bosonfermion ] .
the numbers of homogeneous invariants in an algebraically independent generating set are listed in [ tab : bosonfermionfree ] .
note that in the case of two particles ( either bosons or fermions ) @xmath105 is the number of partitions of @xmath31 .
accordingly , @xmath339 and @xmath340 is freely generated by traces of nonnegative integer powers of the one - particle reduced density matrix .
note that bipartite entanglement measures introduced previously for bosons@xcite and fermions@xcite can be expressed with invariants of the reduced density matrix .
.stable dimensions of homogeneous subspaces of the algebra of local unitary invariants of bosons ( left ) and fermions ( right ) .
the number of particles grows from left to right , while degree grows downwards.[tab : bosonfermion ] [ cols=">,>,>,>,>,>",options="header " , ] the fact that we have a distinguished subsystem of a single particle makes it possible to give an alternative description of the graphs labelling invariants in the basis or in the algebraically independent generating set given above .
let @xmath341 as before . applying the results of the previous section we have that @xmath342 encodes elements of a basis of @xmath343 while the subset @xmath342 corresponds to a set of free generators .
recall that by definition the edges having the @xmath344th colour give a subgraph which is bipartite and @xmath219-regular , therefore we have a distinguished bijection between the two colour classes of vertices .
we can use this bijection to identify the endpoints of such edges , and at the same time , in order to be able to recover the original graph we direct first the remaining edges from the first colour class to the second one . in this way
@xmath342 can be identified with the set of equivalence classes of finite directed graphs edges of @xmath10 different colours such that the subgraph determined by the @xmath250th colour is @xmath257-regular ( i.e. at each vertex the indegree and the outdegree are both @xmath257 ) .
this alternative description is illustrated in fig .
( [ fig : mixed ] ) for a degree @xmath345 invariant of a mixed state of two bosons or two fermions . ]
in this paper we have studied the algebras of real polynomial invariants under the local unitary groups @xmath346 over the representation @xmath347 .
this can be interpreted as the state space of a composite quantum mechanical system containing various types of identical particles , possibly obeying non - abelian statistics .
the algebras have the property that as @xmath348 for all @xmath250 , the dimension of the homogeneous subspaces stabilize , making it convenient to work with the inverse limit @xmath38 of the algebras ( in the category of graded algebras ) .
the stable dimension is then the dimension of the corresponding homogeneous subspace of the inverse limit , for which a formula in terms of induced characters was derived . in the most important case when only bosonic and fermionic particles are present , the bound ( [ eq : dmbound ] ) on the stable dimension was estabilished , which has a combinatorial interpretation in terms of the number of graphs with a certain property .
curiously the bound is saturated iff the total number of fermions is even .
we would like to remark that in general eq .
( [ eq : dmbound ] ) does not hold .
for example in the simplest case not covered in sec .
[ sec : various ] we have @xmath349 and @xmath350 . in this case the sequence @xmath105 for @xmath351 starts as @xmath352 which is to be compared with the @xmath353 column of table [ tab : bosonfermion ] . the simplest such example in the mixed case is @xmath354 , here the stable dimensions are @xmath355 while the bounds would be the values in column @xmath353 in table [ tab : mixed ] . by the definition of the inverse limit , @xmath38 comes equipped with surjections onto the algebra of invariant polynomials over @xmath347 for every choice of the dimensions @xmath356 .
elements of the inverse limit can therefore be directly interpreted as polynomial invariants , although some of these will coincide when projected onto the smaller algebras . in the case of systems of bosons and fermions
we have also given a combinatorial description of an algebraically independent generating set of the inverse limit in terms of regular graphs , showing in particular that the inverse limit is free in these cases .
this generating set can be interpreted as a set of polynomials distinguishing between physically different types of nonlocal behaviour which is minimal in the sense that any polynomial in its elements is nonzero provided the single - particle hilbert spaces are large enough .
we belive that these invariants will become a useful tool in the exploration of genuine multiparticle quantum correlations in quantum mechanical systems containing identical particles .
for every @xmath103 there is a surjection @xmath357 defined by eq . [ eq : tensorinv ] .
composing with the surjective map @xmath358 we obtain a map @xmath359 that is onto the degree @xmath31 homogeneous subspace , @xmath360 .
for @xmath361 we have the following commutative diagram : @xmath362\ar@{->>}[d ] & i_{k,(\lambda_1,\ldots,\lambda_k),n}^{(m ) } \\
i_{k,(\lambda_1,\ldots,\lambda_k),n'}^{(m ) } \ar[ur]_{\varrho_{n , n'}}}\ ] ] this implies that the restriction of @xmath363 to the degree @xmath31 homogeneous subspace must also be surjective .
but if @xmath364 , then @xmath365 and therefore @xmath363 restricted to the degree @xmath31 homogeneous subspace is an isomorphism . before proving lemma [
lem : wrindmany ] we introduce some notation and summarize some properties of wreath products following ref .
for a set @xmath332 let us denote the set of partition valued functions on @xmath332 by @xmath366 , and let @xmath367 for a finite group @xmath300 the set of conjugacy classes of @xmath300 will be denoted by @xmath368 .
recall that the set of conjugacy classes of @xmath369 for any finite group @xmath300 is in bijection with @xmath370 as follows : if @xmath371 is an element of @xmath369 then @xmath372 can be written as a product of disjoint cycles , and for any cycle @xmath373 we can form the product @xmath374 which is determined up to conjugation . the class of @xmath157 is then labelled by the function @xmath375 for which the number of @xmath10-s in @xmath376 is the number of @xmath10-cycles of @xmath372 such that the corresponding cycle - product is in @xmath377 .
the order of the centralizer of an element @xmath157 with type @xmath378 is @xmath379 where @xmath136 denotes the length of the partition as usual and @xmath380 is the order of the centralizer in @xmath368 of an element in the conjugacy class @xmath377 .
a pair @xmath390 can therefore be identified with a conjugacy class in @xmath391 .
the partitions @xmath392 and @xmath393 depend only on the conjugacy class of the two permutations in an element of type @xmath394 . denoting the corresponding characteristic function by @xmath395 and similarly for conjugacy classes of @xmath396,@xmath397 and @xmath93
we can write @xmath398 where we have used that irreducible characters of @xmath93 as well as characteristic functions are real . the last inner product is @xmath218 if @xmath399 and @xmath400 otherwise . now let us look at the left hand side .
the value of the induced character at an element @xmath401 is @xmath218 whenever @xmath157 is not conjugate to any element of @xmath402 which happens precisely when the two permutations of @xmath157 are not conjugate to each other , that is when @xmath399 .
comparing eq .
( [ eq : lhs ] ) with eq .
( [ eq : rhs ] ) one can see that we need to prove that @xmath408 multiplying both sides by @xmath409 we have on the left hand side the size of the conjugacy class @xmath47 in @xmath93 times the number of @xmath410 whose image in @xmath411 has type @xmath394 and on the right hand side size of the conjugacy class @xmath394 in @xmath412 .
an element @xmath413 of @xmath411 can be uniquely written as the product of an element in @xmath414 and one in @xmath415 as follows : @xmath416 , uniqueness follows from the fact that the intersection of the two subgroups consists of only the identity .
therefore every element of @xmath414 can be reached as the conjugate of some element in @xmath411 with an element of the form @xmath417 where @xmath418 finally , @xmath421 iff @xmath422 iff @xmath423 for a fixed @xmath372 such that @xmath424 , which implies that the inverse image of any element in @xmath411 has precisely @xmath425 elements , finishing the proof .
we prove by induction using the @xmath260 case in the induction step .
the @xmath349 case is easy to check . assuming the statement to be true for @xmath426 we can write @xmath427 using that irreducible characters of @xmath93 are real and form an orthonormal basis . as the map in eq .
( [ eq : fdef ] ) defining @xmath290}$ ] factors through @xmath428 , we can also work in the symmetric algebra @xmath429 for some large @xmath18 . in this algebra
the image of @xmath430 is @xmath431 where the sum is over the possible values of the indices @xmath432 and similarly for @xmath433 .
it is convenient to denote the indices in the sum corresponding to @xmath433 by @xmath434 and to regard the permutation @xmath435 as a bijection from @xmath436 to itself .
this convention clearly does not affect the definition of @xmath437}$ ] and it is consistent with the definition of @xmath273 .
then we have @xmath438 where @xmath439 is the image of @xmath440 under the inclusion @xmath441 .
but by the definition of @xmath273 we have @xmath442=[\sigma_1,\ldots,\sigma_k]\star[\sigma'_1,\ldots,\sigma'_k]$ ] .
the sum in eq .
( [ eq : fdef ] ) can be rewritten as a nested sum , grouping together the possible tuples of indices such that the sets of multisets @xmath443 @xmath444 @xmath445 and also the sets @xmath446
@xmath447 @xmath448 are kept fixed .
the inner sum is therefore over permutations stabilizing the above structure which is precisely @xmath449 , since @xmath246 and @xmath450 is the stabilizer of the first and last @xmath10 sets , respectively .
taking into account the sign changes introduced when flipping a pair in a wedge product we have that the inner sums are proportional to ( the projection of ) @xmath451= \\ = & [ \ldots]\langle((\chi_{\lambda_1}\times\cdots\times\chi_{\lambda_k})\wr 1)^s,\operatorname{res}_{h_m}^{{h_m}_s}(\chi_{\lambda_1}\times\cdots\times\chi_{\lambda_k})\wr 1\rangle_{{h_m}_s } \end{split}\ ] ] and therefore vanish by assumption .
-1 mm l. amico , r. fazio , a. osterloh , v. vedral , rev .
80 , 517 - 576 ( 2008 ) p. vrana , j. phys .
a : math . theor . * 45 * , 225304 ( 2011 ) y. shi , phys .
a * 67 * , 024301 ( 2003 ) h. barnum , e. knill , g. ortiz , r. somma , l. viola , phys .
* 92 * , 107902 ( 2004 ) t. ichikawa , t. sasaki , i. tsutsui , j. math
. phys . * 51 * , 062202 ( 2010 ) t. sasaki , t. ichikawa , i. tsutsui , phys .
a * 83 * , 012113 ( 2011 ) k. eckert , j. schliemann , d. bru , m. lewenstein , annals of physics * 299 * , 88 ( 2002 ) w. fulton , j. harris , _ representation theory : a first course _ , springer - verlag ( 1991 ) i. g. macdonald , _ symmetric functions and hall polynomials _ , clarendon press , oxford , ( 1979 ) m. w. hero , j. f. willenbring , _ stable hilbert series as related to the measurement of quantum entanglement _ , discrete math
. 309 ( 2009 ) m. w. hero , j. f. willenbring , l. k. williams , _ the measurement of quantum entanglement and enumeration of graph coverings _ , arxiv:0911.0222 b. simon , _ representations of finite and compact groups _ , graduate studies in mathematics vol . 10 , american mathematical society , ( 1996 ) r. pakauskas , l. you , phys .
a * 64 * , 042310 ( 2001 ) j. schliemann , j. i. cirac , m. ku , m. lewenstein , d. loss , phys .
a * 64 * , 022303 ( 2001 ) s. albeverio , l. cattaneo , s .-
fei , x .- h .
wang , _ multipartite states under local unitary transformations _
phys . * 56 * , 341 ( 2005 ) | we investigate the properties of the inverse limit of the algebras of local unitary invariant polynomials of quantum systems containing various types of fermionic and/or bosonic particles as the dimensions of the single particle state spaces tend to infinity .
we show that the resulting algebras are free and present a combinatorial description of an algebraically independent generating set in terms of graphs .
these generating sets can be interpreted as minimal sets of polynomial entanglement measures distinguishing between states showing different nonclassical behaviour .
pter vrana department of theoretical physics , institute of physics + budapest university of technology and economics + h-1111 budapest , hungary ( ) |
low energy supersymmetry ( susy)@xcite is one of the most active fields of research for physics beyond the standard model ( sm)@xcite . a minimal extension of the standard model when supersymmetry is incorporated is the minimal supersymmetric standard model ( mssm)@xcite that includes two higgs doublets .
the model however has a large number of susy breaking parameters and this motivates one into studying models with specific mechanisms for breaking susy .
the later involves high scale physics input and renormalization group analyses .
this in general leads to a large reduction of the number of unknown parameters .
the minimal supergravity ( msugra)@xcite model is one of the well studied susy models .
it requires a very few input parameters at the gauge coupling unification scale or grand unified theory ( gut ) scale , @xmath3 gev .
the parameters indeed quantify our ignorance of the exact nature of susy breaking .
the model incorporates radiative breaking of electroweak symmetry ( rewsb ) .
the unification scale universal input parameters are : i ) the gaugino mass parameter @xmath4 , ii ) the scalar mass parameter @xmath5 and iii ) the tri - linear susy breaking parameter @xmath6 .
additionally , one has to provide with @xmath7 , the ratio of higgs vacuum expectation values and the sign of the higgsino mixing parameter @xmath8 .
renormalization group evolutions are used to obtain the electroweak scale parameters of mssm .
there is however no _ a priori _ necessity of considering such universalities of parameters .
indeed , it is worthwhile to explore scenarios with non - universalities in the scalar or in the gaugino masses at the unification scale@xcite .
non - universal scalar masses may appear because of non - flat khler potential@xcite .
however , one must be careful to accommodate the stringent constraints from phenomena involving flavor changing neutral currents ( fcnc ) .
satisfying fcnc constraints demands near - degeneracy of the first two generations of scalar masses but the requirements on the third generation of scalars as well as the higgs scalars are not so stringent @xcite .
thus it is seen that fcnc constraints as well as constraints from cp - violating phases ( for example , those arising from the electric dipole moments of the electron and neutron ) may be managed by introducing multi - tev scalar masses for the first two generations of scalars@xcite .
the third generations of scalar masses and the higgs scalar masses however should be adequately light in order to satisfy naturalness@xcite .
there have been several efforts for obtaining the desired features as outlined above .
analyses with radiatively generated inverted mass hierarchy models ( rimh ) were made in ref.@xcite . in ref.@xcite
the authors achieved the above - mentioned requirements at the electroweak scale by using @xmath9 yukawa unification with special non - universal relationship among the scalar masses at @xmath10 . here
, the higgs and the third generation of scalar masses are rapidly diminished at the electroweak scale via rg evolutions . however , the yukawa unification and the consideration of rewsb constrain such models heavily .
a second realization of the above idea is partially possible via the hyperbolic branch ( hb)/ focus point ( fp)@xcite scenarios where the scalars may become considerably massive ( multi - tev ) in a subset of the typical msugra parameter space satisfying universal boundary conditions while fine - tuning@xcite still remains small .
a third possibility was considered in ref.@xcite , where plain decoupling arguments motivated the authors in using explicit splitting at @xmath10 between the scalars belonging to the first two generations and the same of the third generation ( along with the higgs scalars ) . here the first two generations of scalars were chosen in the multi - tev domain whereas the third generation of scalars as well as the higgs scalars were considered to be in the sub - tev zone .
additionally , a large value of the tri - linear coupling parameter was chosen in ref.@xcite for the first two - generations .
universality of scalar masses in the first two - generations along with a choice of a different scalar mass parameter for the third generation as well as the higgs scalars , or even splitting of squarks and sleptons within the third generation itself have also been considered in refs.@xcite . in ref.@xcite
the authors additionally considered non - universality in the gaugino masses and analyzed the fine - tuning aspect of computing the relic density of dark matter in addition to obtaining a parameter zone of the mssm corresponding to a well tempered neutralino .
similarly , analyses with only non - universalities in the higgs scalar may be seen in refs.@xcite .
a comprehensive set of characteristic possibilities for varieties of non - universal sugra scenarios may be seen in ref.@xcite . in this analysis
we explore a sugra scenario with non - universal scalar masses that i ) allows to have very large first two - generation of scalar masses so as to obey the fcnc and the cp - violation limits easily ( _ ie . _ without requiring any ultra - small phases ) and that would not impose any additional price on fine - tuning , ii ) spans a large amount of mssm parameter space satisfying the neutralino relic density constraint from wmap data by not requiring any delicate mixing of bino and higgsinos , so that we would find a bino - dominated lightest neutralino for most of the parameter space and , iii ) satisfies the higgs mass lower bound from lep2 data as well as other low energy constraints .
certainly , a model with rewsb and universal scalar mass like msugra is not friendly to achieve these objectives if we consider a common gaugino mass parameter below a tev or so .
the above phenomenologically inspired requirements in combination motivate us to introduce non - universality between the third and the first two - generation of scalars as well as the higgs scalars .
this has to be such that the rewsb conditions would not become prohibitive to have a multi - tev first two - generation of scalars .
here we would like to point out that we would not be able to satisfy the mentioned objectives by considering non - universalities only in the higgs scalar masses .
a simpler and purely phenomenological attempt in this direction in a universal gaugino mass framework could be to generate all the third generation of scalar masses and the higgs scalar masses radiatively , starting from zero values at the unification scale , while keeping the first two - generations of scalar masses universal ( and this may be in the multi - tev zone ) .
we point out that starting with a similar range of values for the higgs scalars as with the third generation of scalars at @xmath10 would be desirable since this would lead to similar range of mass values for all the soft - susy breaking terms contributing to rewsb after rg evolutions .
however , if we consider all the third generation of scalars at @xmath10 to be light , we would find stau ( @xmath11 ) becoming the lightest stable particle ( lsp ) or even tachyonic at the electroweak scale . on a similar note ,
we remind that such an appearance of tachyonic sleptons also arise in the anomaly mediated supersymmetry breaking ( amsb ) analyses , and it is avoided purely by a phenomenologically inspired way of adding an appropriate non - zero mass value to all the scalars at the unification scale in the minimal amsb scenario@xcite . thus with a simple motivation of managing with fcnc and cp - violation , naturalness as well as the dark matter constraints concurrently we consider a non - zero mass value for the third generation of sleptons , while having the masses of squarks of the third generation and the higgs scalars vanishing at @xmath10 .
additionally , for convenience we set the third generation of slepton mass parameters the same as that of the first two - generation of scalars at @xmath10 .
thus , the parameters of our non - universal scalar mass model which will henceforth be called the nusm model is given by , , a_0 , sign ( ) , and m_0 where the scalar mass input assignments at @xmath10 are as follows .
+ i ) the unification scale mass parameter for the first two - generations of squarks , sleptons and the third generation of sleptons is @xmath5 , where @xmath5 is allowed to span up to a very large value .
+ ii ) the mass parameters for the third generation of squarks and higgs scalars are set to zero .
+ we could have also chosen a non - vanishing value for ii ) as long as it is sufficiently small .
we may note here that different values of mass parameters for squarks and sleptons at @xmath10 may appear in orbifold models with large threshold corrections .
this is related to having different modular weights associated with the squarks and the sleptons in a given generation@xcite .
as we will see below , the nusm model provides with a highly bino - dominated neutralino dark matter over almost its full range of parameter space .
it has an interesting feature of having a large funnel region , a region of parameter space where the associated annihilation channels are characterized by the direct channel pole @xmath12 .
we note that unlike msugra where one finds the funnel region only for a large value of @xmath1 , here in the nusm one finds it for almost all possible values of @xmath13 .
we will also see that the nusm is further characterized by a lighter @xmath14 or @xmath15 particularly when @xmath5 is large and this is found even for small values of @xmath16 .
a small part of the parameter space may be far from the decoupling@xcite region of higgs boson mixing that causes the lighter cp - even higgs boson ( @xmath17-boson ) to be non - standard model like@xcite .
this in turn reduces the lower bound of @xmath18 much below the lep2 higgs boson mass limit . in this work
we include a semi - analytic calculation that first points out a large mass effect in the solution of the renormalization group equations ( rge ) of @xmath0 .
the effect causes a hyperbolic branch / focus point like behavior is associated with the first minimization condition of radiative electroweak symmetry breaking . in this analysis
we do not have such an effect . on the contrary we have a similar hb
/ fp effect for small values of @xmath1 in connection with the second minimization condition that is associated with @xmath2 . ] in @xmath19 that may cause @xmath14 to become smaller even for a small @xmath1 as mentioned above .
here we remind that @xmath14 typically becomes smaller in msugra only for large @xmath1 .
the paper is organized as follows . in sec.[nusmhbfp
] we will primarily discuss the large mass rge effects in the nusm on @xmath2 and its consequent reduction for large @xmath5 via an hb / fp - like effect .
we will obtain the semi - analytic results which will also be verified by numerical computation . in sec.[cdmandfunnel ] we will study the cold dark matter constraint from wmap data including also the constraints from the lep2 higgs bound , @xmath20 , and @xmath21 .
we will also discuss a few sample points in the context of lhc reach . in sec.[detectionrates
] we will discuss the direct and indirect detection ( via continuous gamma ray ) rates of the lsp .
finally we will conclude in sec.[concl ] .
in this section we focus on the key elements that are important for the nusm .
the primary quantities of phenomenological interest are @xmath8 , and @xmath14 .
these in turn have important significance on the following : i ) the issue of fine - tuning , ii ) the presence of non - standard model like lighter higgs boson mass bound for a limited region of parameter space , iii ) dark matter and iv ) collider discovery possibilities .
additionally , we remind ourselves about obtaining a lighter third generation of squarks at the electroweak scale and this is of course related to their vanishing values at the unification scale . to start with
we write down the rewsb results , ^2 = - m^2_z + + , [ rewsbmusqeqn ] and , 2= 2b/(m_h_d^2+m_h_u^2 + 2 ^ 2+_1+_2 ) [ rewsbbeqn ] where @xmath22 represent the one - loop corrections @xcite , that become small in the scale where the higgs potential @xmath23 is minimized .
we may approximately consider @xmath24 ( for @xmath25 ) and @xmath26 at tree level .
the associated one - loop rges are given by ( neglecting the very small first two - generations of yukawa contributions ) , @xmath27 @xmath28 here we have @xmath29 with @xmath30 being the renormalization scale .
@xmath31 for @xmath32 are the scaled gauge coupling constants ( with @xmath33 ) and @xmath34 are the running gaugino masses , where @xmath32 refers to u(1 ) , su(2 ) and su(3 ) gauge groups respectively .
@xmath35 represent the scaled and squared yukawa couplings , e.g , @xmath36 where @xmath37 is a yukawa coupling ( @xmath38 stands for @xmath39 respectively ) .
the quantity @xmath40 which is shown below may become relevant for a general set of non - universal boundary condition for scalars .
however in this analysis it is zero at @xmath10 and it causes only a negligible effect .
s_0=m_h_u^2-m_h_d^2 + _ k ( m_q_k^2 + m_d_k^2 + m_e_k^2 - m_l_k^2 - 2 m_u_k^2)+(m_q^2+m_d^2+m_e^2-m_l^2- 2 m_u^2 ) where the subscript @xmath41 for @xmath42 or @xmath43 indicates the first two generations .
the solutions for @xmath44 and @xmath2 valid up to a moderately large value of @xmath7 ( 10 or so ) follow from ref.@xcite after appropriately considering the nusm parameters .
this involves ignoring @xmath45 and @xmath46 in the limit of a small @xmath1 .
the solutions read ^2=m_0 ^ 2c_1 + a_0 ^ 2 c_2 + ^2c_3+a_0 c_4-m_z^2 + s_0 p , [ mueqn ] m_a^2=m_0 ^ 2d_1 + a_0 ^ 2 d_2 + ^2d_3+a_0 d_4-m_z^2 + s_0 p. [ maeqn ] here @xmath47 is given by : @xmath48 $ ] .
the unification scale conditions for scalar mass squares for the nusm may be given as @xmath49 . here ,
the subscript @xmath50 for @xmath51 stands for : @xmath52 . for nusm one has @xmath53 and @xmath54 .
these unification scale conditions allow us to define a quantity @xmath55 that appears in the expressions of the coefficients @xmath56 and @xmath57 as shown below .
it follows that @xmath58 for the nusm , and @xmath59 for msugra .
we note that only @xmath56 among @xmath60 s and @xmath57 among @xmath61 s depend on @xmath55 and one obtains and @xmath62 .
however , nusm additionally requires non - universality in @xmath63 . ] , + c_1&=&(1 + ) ( 1-^2 ) , + d_1&=&(1+)(1-d_0 ) .
[ cdeqn ] here , @xmath64 .
the definitions of the quantities @xmath60 and @xmath61 for @xmath65 may be seen in the appendix . as mentioned above , for the nusm one has @xmath58 .
hence , eqs .
[ mueqn]-[cdeqn ] show that at the level of approximation where @xmath45 and @xmath46 are ignored , @xmath44 and @xmath2 are independent of @xmath5 .
we will however see that both @xmath44 and @xmath2 would actually depend on @xmath5 .
while the dependence of @xmath44 on @xmath5 would be caused by two - loop rge effects in @xmath66 , the same for @xmath2 is very prominent even at the one - loop level of rge and this is found when we include the effects due to @xmath45 and @xmath46 which are specially important for the nusm .
the following subsection describes our improved results .
numerical computation shows that unlike what is seen from eqs .
[ maeqn ] and [ cdeqn ] , @xmath2 may indeed decrease very rapidly with an increase of @xmath5 for large values of the later .
there are two reasons behind the above behavior : i ) the choice of vanishing higgs scalars in the nusm at the unification scale @xmath10 and ii ) a large slepton mass ( lsm ) effect in the rge of eq.[hdeqn ] via the tau - yukawa term .
henceforth , we will refer the combined effect of i ) and ii ) as the lsm effect .
in this analysis we first compute @xmath0 analytically without ignoring the terms involving @xmath45 and @xmath46 .
the calculation essentially keeps the terms involving the bottom and the tau - yukawa couplings similar to what was used for the top - yukawa term of eq.[hdeqn ] in refs.@xcite .
this results into , m_h_d^2=c_h_dm_0
^ 2+m_1/2 ^ 2 g(t)+s_0p .
[ mhdsoln ] here @xmath67 is a function of @xmath1 @xcite .
the term involving @xmath40 vanishes in the nusm . at the level of approximation where the bottom and tau yukawa couplings are straightway ignored in eq.[hdeqn ]
, @xmath68 would become zero in the nusm .
the result of our computation of @xmath68 that leads to a non - vanishing value is given below .
c_h_d = c_h_d(u)+c_h_d(nu ) where , @xmath69 the quantities @xmath70 and @xmath71 are functions of @xmath72 for @xmath73 .
small @xmath1 solutions of @xmath74s , computed at the electroweak scale@xcite are shown in the appendix . @xmath70 and
@xmath71 defined below are computed numerically .
i_2=_0^t y_2(t^)dt^ , i_3=_0^t y_3 ( t^)dt^ here @xmath75 .
one finds @xmath76 . for nusm ,
the above reduces to c_h_d=-2i_3 .
[ chdneweqn ] since @xmath71 is proportional to the square of @xmath77-yukawa coupling @xmath46 where @xmath78 , we find @xmath79 to be an increasing function of @xmath1 .
for @xmath80 , we find @xmath81 .
the above clearly shows the large slepton mass effect because the dependence of @xmath68 or @xmath0 on @xmath71 arises from the term of eq.[hdeqn ] that is associated with @xmath82 .
thus with the nusm parameters , the scalar mass term that appears as the first term in eq.[mhdsoln ] is essentially contributed by the third generation of slepton masses .
eq.[mhdsoln ] and eq.[chdneweqn ] clearly show that even with a small @xmath1 , larger values of @xmath5 may reduce @xmath0 appreciably and may turn the later negative .
this demonstrates the lsm effect as mentioned before at the unification scale , the resulting coefficient of @xmath83 in @xmath0 would be positive and the later would no longer be a decreasing function of @xmath5 .
( ii ) we further note that the fact that the third generation of squark masses are vanishing in nusm also makes the lsm effect prominent .
for example , had we considered non - vanishing values for @xmath84 and @xmath85 ( @xmath86 , ie .
@xmath87 ) we would find the terms in eq.[chdeqn ] that are associated with @xmath70 to become more prominent .
this would have caused @xmath0 to be further negative .
however , it could also cause @xmath88 to turn tachyonic at the electroweak scale .
we have also checked this fact numerically . ] .
we are not aware of any past reference that pointed out this large mass effect or provided with semi - analytic expressions . for larger @xmath1
, @xmath68 may be appreciably large and negative that causes the lsm effect to become prominent even for a smaller value of @xmath5 .
we note that there is no lsm effect that may modify @xmath44 of eq.[mueqn ] at the one - loop level .
this is simply because @xmath24 and there is no possibility of having a lsm effect in the corresponding rge of eq.[hueqn ] .
on the other hand , an improved @xmath0 as obtained above in eqs.[mhdsoln ] and [ chdneweqn ] changes the result of @xmath2 ( eq.[maeqn ] and [ cdeqn ] ) but the later does not receive any additional @xmath83 dependence other than what is already contributed from eq.[chdneweqn ] .
thus it follows that the coefficient of @xmath83 when eq.[maeqn ] is modified would be @xmath68 . as a result , in the above analysis that is valid for small @xmath1 below 10 or 15
, we find that the lsm effect causes @xmath2 to have a hb / fp like behavior for its dependence on @xmath5 and @xmath4 .
hence , @xmath14 may become significantly small for a large @xmath5 because of a cancellation between the terms .
this of course may happen even for a small value of @xmath1 .
a very large @xmath5 would cause @xmath2 to become tachyonic , or would result into an absence of rewsb ( eq.[rewsbbeqn ] ) . for a large @xmath1 on the other hand ,
the lsm effect is drastically enhanced .
we comment here that in spite of showing the one - loop results we performed a complete numerical solution of the rges up to two loops in this analysis using suspect@xcite .
we now point out that a simple non - universal higgs scalar mass scenario as described in ref.@xcite may also provide a small @xmath14 along with a funnel type of region that satisfies the wmap data , for a smaller value of @xmath1 .
we emphasize in particular the case where non - universality was analyzed with a unified higgs scalar mass in ref.@xcite .
however , unlike the nusm these scenarios are very much constrained via rewsb so that considering larger values of masses for the first two - generation of scalars , an easier way to control fcnc and cp - violation effects is not possible . in these non - universal higgs scalar models @xmath14 becomes small for small values of @xmath1 , only when tachyonic values of @xmath0 and @xmath66 ( assumed equal ) are considered at the unification scale and this severely reduces the available parameter space .
additionally , one may obtain an @xmath89-pole annihilation region or a funnel region for a small value of @xmath1 in the so called sub - gut cmssm scenario@xcite .
we will now describe our results as obtained by using suspect@xcite .
fig.[m0onmuandma ] shows the variation of @xmath8 and @xmath14 when @xmath5 is varied in the nusm and msugra .
fig.[m0onmuandmaa ] shows the effect of varying @xmath5 in msugra and in the nusm for @xmath90 gev and 1 tev .
a small reduction of @xmath91 ( @xmath92 ) is seen in the nusm when @xmath5 is increased up to 5 or 10 tev .
the approximate one - loop result of eq.[mueqn ] and eq.[cdeqn ] with @xmath58 as in nusm however would indicate a flat @xmath8 over a variation of @xmath5 .
the figure of course shows a moderately varying @xmath8 and we have checked that this variation has its origin in two - loop rge effects .
the two different @xmath93 contours for msugra however show a decreasing behavior of @xmath8 when @xmath5 is enhanced .
this happens simply because of the hb / fp effect existing in msugra .
fig.[m0onmuandmab ] shows the results for @xmath14 which have two sets of contours for @xmath90 gev and 1 tev corresponding to msugra and the nusm .
in contrast to msugra where @xmath14 rapidly increases with @xmath5 , the behavior in nusm is opposite so that we may find a very light pseudoscalar higgs boson via the lsm effect .
the decrease of @xmath14 is even more pronounced for a large value of @xmath1 and this may be easily seen in fig.[m0onmuandmac ] where a variation of @xmath14 _ vs _ @xmath5 is shown for a given @xmath94 for three different values of @xmath1 .
the curves ends in the larger @xmath5 sides for a few different reasons . for @xmath95 ,
the largest @xmath5 limit is caused by stop mass becoming very light or unphysical . on the other hand for @xmath96 the largest @xmath5 limit is given by absence of rewsb ( ie @xmath2 turning negative ) .
the lsm effect may be explicitly seen in fig.[masssqvsq ] where we plot squared scalar masses of higgs scalars as well as their difference @xmath97 with respect to a variation over the renormalization scale @xmath30 for msugra and the nusm for @xmath80 , @xmath98 gev and @xmath99 .
fig.[masssqvsqa ] corresponds to msugra for @xmath100 tev whereas fig.[masssqvsqb ] shows the case of the nusm for @xmath101 tev .
we have chosen different values of @xmath5 in the two figures because of the fact that a large @xmath5 is prohibited in msugra via the rewsb constraint whereas a choice of a small @xmath5 in the nusm would have a negligible lsm effect .
clearly , @xmath0 stays positive at the electroweak scale in msugra ( fig.[masssqvsqa ] ) . on the other hand the same for the nusm turns toward a negative value while running from @xmath10 to the electroweak scale ( fig.[masssqvsqb ] ) and this essentially shows the lsm effect in @xmath2 .
thus a typical large @xmath1 phenomenon that occurs in msugra is obtained in the nusm for a small @xmath1 .
in supergravity type of models @xmath102 becomes the lsp for most of the parameter space@xcite and we assume that the cold dark matter relic density is entirely due to @xmath102 .
considering wmap data@xcite one finds a 3@xmath103 limit as shown below .
@xmath104 where @xmath105 is the dm relic density in units of the critical density . here
, @xmath106 is the hubble constant in units of @xmath107 . in the thermal description ,
the lsp was in thermal equilibrium with the annihilation products at a very high temperature of the early universe ( @xmath108 ) .
the annihilation products include fermion pairs , gauge boson pairs , higgs boson pairs or gauge boson - higgs boson combinations and they are produced via @xmath109 and @xmath110-channel processes .
as the temperature decreased the annihilation rate fell below the expansion rate of the universe and the lsp went away from the thermal equilibrium and freeze - out occured .
the current value of @xmath111 is computed by solving the boltzmann equation for @xmath112 , the number density of the lsp in a friedmann - robertson - walker universe .
the above computation essentially involves finding the thermally averaged quantity @xmath113 , where @xmath114 is the relative velocity between two annihilating neutralinos and @xmath115 is the neutralino annihilation cross section that includes all the final states .
in addition to annihilations one considers coannihilations@xcite which are annihilations of the lsp with sparticles close in mass values with that of the lsp .
the cross section sensitively depends on the nature of the composition of the lsp . in the mssm ,
the lsp is a mixed state of bino ( @xmath116 ) , wino ( @xmath117 ) and higgsinos ( @xmath118 @xmath119 ) : @xmath120 here @xmath121 are the elements of the matrix that diagonalizes the neutralino mass matrix .
the gaugino fraction @xmath122 of the lightest neutralino is defined by @xmath123 .
a gaugino - like lsp may be defined to have @xmath124 very close to 1(@xmath125 ) . on the other side , a higgsino - like
lsp would have @xmath126 . for values in between
, the lsp could be identified as a gaugino - higgsino mixed state .
the mssm with gaugino masses universal at @xmath10 has a few distinct regions in general that satisfy the wmap constraint . in this section
we point out the existence of such regions in the context of the nusm .
i ) the _ bulk annihilation _ region in msugra is typically characterized by smaller scalar masses and smaller values of @xmath93 where the lsp is bino - dominated .
a bino dominated lsp couples favorably with right handed sleptons .
thus , the lsp pair annihilation in the bulk annihilation region occurs primarily via a t - channel sfermion in msugra .
there are two important constraints that disfavors the bulk region in msugra .
these are the constraints a ) from the slepton mass lower limit from lep2@xcite and b ) from the lower limit of higgs boson mass @xmath18 of @xmath127 gev@xcite@xcite . ] .
ii ) the _ focus point_@xcite or the _
hyperbolic branch_@xcite region of msugra is typically characterized by a small @xmath91 region that is close to the boundary of the lighter chargino mass lower bound .
because of a small @xmath91 here the lsp has a significant amount of the higgsino component or it can even be almost a pure higgsino .
additionally the lighter chargino @xmath128 becomes lighter and coannihilations@xcite with lsp reduce the relic density to an acceptable level .
the hb / fp region is however absent in the nusm .
+ iii ) coannihilations of lsp may also occur with sleptons , typically staus ( @xmath129)@xcite in msugra .
these regions are associated with small @xmath93 and small @xmath5 zones near the boundary of the discarded zone where staus become the lsp .
stau coannihilation is also an effective way to bring the neutralino relic density to an acceptable level in the nusm .
coannihilations of lsp may also occur with stop ( @xmath130 ) in a general mssm scenario@xcite or even in msugra@xcite .
however , in spite of having relatively lighter @xmath130 , the nusm does not have such a region unless one reduces the mass of @xmath130 further via appropriately considering non - zero values for @xmath6 .
+ iv ) the most important region satisfying wmap data for our study is the higgs - pole annihilation or _ funnel _ region@xcite . the funnel region that satisfies the wmap data
is characterized by the direct - channel pole @xmath131 .
this occurs in msugra typically for large @xmath1 extending to larger @xmath5 and larger @xmath93 regions . in the nusm
however , the funnel region occurs in all possible @xmath13 .
indeed apart from the lsp - stau coannihilation appearing in a small region , higgs - pole annihilation is the primary mechanism in the nusm to satisfy the wmap constraint throughout the parameter space .
we now show the results of the computation of the neutralino relic density using the code micromegas@xcite . in fig.[t10nuspectrum ] that corresponds to @xmath80 the region with red dots in the @xmath132 plane satisfy the wmap constraint ( eq.[relicdensity ] ) for the neutralino relic density . for a small @xmath5
we find the stau coannihilation region marked with red dots .
the gray region below the stau coannihilation region is discarded because of the appearance of charged lsps .
the upper gray region is discarded broadly via @xmath2 turning negative except near the boundary where there can be interplay with other constraints as described below .
thus if we concentrate on the boundary of the discarded region , the smallest @xmath93 zone ( below 160 gev or so ) is ruled out by the lep2 lower limit of sparticle masses@xcite .
this is followed by obtaining tachyonic sfermion scalars ( particularly stop scalars ) when @xmath4 is increased further up to @xmath133 600 gev .
the same boundary zone for the larger @xmath93 region is eliminated because of the appearance of the charge and color breaking ( ccb ) minima@xcite . in the nusm
this happens via the ccb conditions that involve @xmath0 , the later becoming negative makes the ccb constraint stronger . in the region between the stau coannihilation area and the upper gray shaded discarded area one finds a long red region that satisfies the wmap constraint for @xmath111 . as mentioned before , the reason for satisfying the wmap data is the direct channel annihilation of two lsps via neutral higgs bosons .
fig.[t10nufunnel ] shows a scanned output for @xmath80 when @xmath5 is varied as in fig.[t10nuspectrum ] .
the lsp mass is plotted against @xmath134 so as to show the extent of the @xmath89-annihilation or funnel effect .
clearly the wmap satisfied points shown in red fall around the zero of the y - axis confirming that the funnel region appears in the nusm for small @xmath1 as well .
we note that the @xmath89-width can be quite large ( @xmath135-@xmath136 gev ) and @xmath137 can be appreciably away from the exact resonance zone still giving a @xmath138-channel annihilation consistent with the wmap data .
the heavy scalar higgs boson @xmath139 also significantly contributes to the total annihilation cross - section .
we now discuss the result of using a few constraints .
first , we use the lep2 limit@xcite of @xmath127 gev for the sm higgs boson . for the cp - even lighter susy higgs boson mass @xmath18
we use the same constraint as long as we are in the decoupling region@xcite where the susy @xmath17-boson is sm - like . in the nusm
this is true for almost all the parameter space except a very small region with small @xmath4 and for @xmath96 which we will discuss later .
additionally , we note that there is an uncertainty of about 3 gev in computing the mass of the light higgs boson@xcite .
this theoretical uncertainty primarily originates from momentum - independent as well as momentum - dependent two - loop corrections , higher loop corrections from the top - stop sector etc .
hence , we have drawn a contour for @xmath140 gev in order to consider an effective lower limit .
gev and @xmath141 gev . ]
we have further drawn the @xmath142 contour by considering a 3@xmath103 limit@xcite , @xmath143 clearly in fig.[t10nuspectrum ] this constraint would keep the @xmath144 gev region alive and this is indeed the region of super - large @xmath145 tev that satisfies the wmap data . in this context
we should however note that the @xmath146 constraint may be of less importance if some additional theoretical assumptions are taken into consideration .
the computation of the constraint in models like msugra assumes a perfect alignment of the squark and quark mass - matrices .
this essentially considers an unaltered set of mixing angle factors than the cabibbo - kobayashi - maskawa ( ckm ) factors at the corresponding sm vertices .
even a small set of off - diagonal terms in squark mass matrices at the unification scale may cause a drastic change in the mixing pattern of the squark sector at the electroweak scale .
this however does not cause any effective change in the sparticle mass spectra or in the flavor conserving process of neutralino annihilation or in generating events in a hadron collider in any significant way .
a brief review of the model - dependent assumptions for the @xmath147 analyses may be seen in refs.(@xcite , @xcite ) and references therein .
we now explain an interesting aspect of the nusm higgs boson as a consequence of the lsm effect . as we have seen before
, @xmath14 decreases with increasing @xmath5 in the nusm .
we find that with larger @xmath5 , @xmath14 may become very light for the small @xmath4 zone so that we may find a spectrum where @xmath148 and this indeed is a consequence of moving into the intense - coupling@xcite region of higgs bosons .
this results in a change in the value of the coupling @xmath149 , where @xmath150 is the mixing angle between two neutral higgs bosons @xmath17 and @xmath139 .
typically in the decoupling region @xmath151 stays very close to @xmath152 , but in the intense coupling region this can be considerably small like 0.5 or lesser .
this effectively reduces the lower limit of @xmath18 to 93 gev , a value close to @xmath153 . in the nusm
we obtain this effect for a small region of parameter space ( small @xmath4 ) if @xmath96 .
fig.[t15nuspectrum ] shows such a region ( this may also satisfy the wmap limits ) for @xmath154 gev and this appears very close to the top gray shaded discarded zone .
fig.[t15nuhiggsmixing ] demonstrates the existence of a small @xmath151 in the nusm as discussed above .
however with @xmath99 , we will see that such a very light @xmath14 or @xmath18 region is almost discarded via the present limit of the @xmath155 .
the current experimental limit for the @xmath156 coming from cdf@xcite puts a strong constraint on the mssm parameter space .
the experimental bound is given by ( at @xmath157 ) @xmath158 the estimate of @xmath21@xcite in the mssm sensitively depends on the mass of a - boson ( @xmath159 ) and on the value of @xmath160 ( @xmath161 ) .
@xmath21 constraint eliminates the thin region along the boundary of the rewsb in nusm where @xmath14 is very light . as mentioned before , for @xmath96 a part of the above region for small @xmath4 zone has non - standard model like @xmath17-boson .
apart from the above constraints we would like to remind that there is no non - universality in the first two - generations in nusm .
this saves from the stringent fcnc violating limits such as those coming from the @xmath162-@xmath163 mass difference or from the @xmath164 bound .
splitting of the first generation and the third generation of scalars may also cause violations of fcnc bounds , although to a lesser extent .
following ref.@xcite we see that for no violation of fcnc one would need ( for equal gluino and average squark masses ) @xmath165 . here
@xmath166 refers to the average squark mass . considering the analysis performed in ref.@xcite for different gluino masses
we conclude that the results of our analysis stay in the safe zone regarding the fcnc bounds in spite of the inter - generational splitting between the squarks .
fig.[t40nuspectrum ] shows the results for @xmath167 . here
the funnel region is extended up to @xmath168 tev .
a large value of @xmath1 increases @xmath45 and @xmath46 and this would enhance the width @xmath169@xcite .
we point out that unlike fig.[t10nuspectrum ] , here the reach of @xmath5 decreases for a given @xmath93 .
the mass of a - boson decreases with increase of @xmath5 much more rapidly because of larger @xmath45 and @xmath46 arising out of larger values of @xmath1 . here , the very light @xmath14 region that satisfies the wmap data and that evades the lep2 higgs boson limit exists near the top gray boundary for @xmath170 gev which is again ruled out via the @xmath155 limit .
the upper gray region is typically discarded here via the rewsb constraint of @xmath2 that turns negative at the tree level .
finally , we have not imposed any limit from the muon @xmath171 data that may or may not show a discrepancy from the standard model result .
it is known that using the recent @xmath172-@xmath173 data leads to a @xmath174 level of discrepancy . on the other hand , using hadronic @xmath77-decay data in computing the leading order hadronic contribution to muon @xmath171 washes away@xcite any deviation from the sm result .
we would focus on a few characteristic parameter points to discuss the nature of the nusm spectra that satisfy the wmap limits .
clearly one is able to reach a considerably large @xmath5 satisfying dark matter relic density constraints and all other necessary constraints for @xmath175 tev . as an example with @xmath176 tev ,
this limit is around @xmath177 tev for @xmath80 , @xmath178 tev for @xmath179 , and @xmath180 tev for @xmath167 .
thus , for larger @xmath1 the reach of @xmath5 decreases considerably . on the other hand ,
it is not unusual to obtain an @xmath89-pole annihilation region for a large @xmath1 in popular models like msugra .
hence , if we are interested in focusing on a region of the mssm that is not available in msugra type of models we would rather explore the smaller @xmath1 domain of the nusm .
the other important zone of parameter space could be the region with smaller @xmath4 because this would be easily accessible in lhc in its early run .
we pick up a point on fig.[t10nuspectrum ] ( _ scenario - a _ ) that just satisfies the minimum value of @xmath140 gev besides being consistent with the wmap data for cold dark matter .
the input parameters are : @xmath181 .
the scenario - a of table [ tabnusm ] thus has light stop and light sbottom quarks , light charginos and neutralinos and at the same time it would have a light higgs spectra , all of which are promising for an early lhc detection .
we remind ourselves that typically the nusm is associated with a heavy first two - generation of scalars and heavy sleptons for all the three generations .
we note that we have relaxed the @xmath182 constraint keeping in mind of the argument given after eq.[bsgammalimits ] .
we could of course respect the constraint , only at a price of having a little heavier spectra , still that would be very much accessible in lhc for the third generation of squarks , higgs bosons , charginos etc .
this however would not cause any essential change in the general pattern . on the other hand , with
not so light @xmath14 and with a small @xmath1 the scenario - a satisfies the @xmath155 limit .
the _ scenario - b _ of fig.[t15nuspectrum ] refers to a special point ( @xmath183 , @xmath184 ) for which the higgs sector is not in the decoupling region thus reducing the limit of @xmath18 to a value near the z - boson mass .
this parameter point obeys the @xmath155 limit but as with the _ scenario - a _ it has an inadequate @xmath185 . with smaller @xmath4 and smaller @xmath5
_ scenario - b _ has a lighter spectrum in general than _ scenario - a_. particularly , we find a much smaller higgs boson spectra .
the _ scenario - c _ ( with @xmath186 , @xmath187 ) , that satisfy all the constraints including the @xmath182 bound provides a relatively heavier spectra .
this is however still within the lhc reach for the relevant part of the nusm as mentioned above .
we note that the large @xmath5 domain of the nusm is typically associated with lighter higgs spectrum and this region may in general be probed via the production of all mssm higgs bosons and subsequently their decays .
the squarks corresponding to the first two generations and sleptons of all the generations are large when @xmath5 is large . as a result
the production cross - section of these particles will be too low at the lhc .
thus , to see any susy signal for the nusm one should primarily analyze productions and decays of gluino , stop and sbottom in addition to charginos and neutralinos . in a part of the parameter space
where the higgs spectra is light , all the higgs states may be produced via @xmath188 either via loop - induced processes like ( @xmath189 ) or through cascade decays via heavier charginos and neutralinos @xmath190 .
the latter decays are only allowed if enough phase space is available .
.data for point a , b , and c. masses are in gev [ cols="<,^,^,^",options="header " , ] fig.([photonfluxvsm0 ] ) shows the result of continuum photon - flux _ vs _
@xmath5 for @xmath80 and @xmath191 corresponding to two different values of @xmath93 ( @xmath192 and @xmath193 gev ) and three different halo profiles as mentioned above .
the photon flux ( in @xmath194 ) in the nusm is computed using darksusy@xcite with @xmath195 gev , for a solid angle aperture of @xmath196 sr .
as @xmath5 increases the mass of psuedoscalar higgs boson decreases and the lsp pair annihilation through s - channel resonance causes a peak corresponding to a halo profile .
broadly , such a peak covers the region of @xmath5 for a given @xmath93 where wmap data for the neutralino relic density is satisfied . with an increase in @xmath1 the width @xmath169 of @xmath14 increases and the peak associated with a given halo profile broadens .
we see that in spite of having a broad range of halo profile characteristics , the photon flux in the region of resonance annihilation in the nusm where the wmap data is satisfied may be probed in the upcoming glast@xcite experiment at least for the cuspier profiles .
glast would be able to probe photon - flux as low as @xmath197 @xmath198@xcite .
we note that this conclusion remains valid in spite of the fact that glast will use an aperture of @xmath199 sr so that an appropriate scaling of the photon - flux in fig.([photonfluxvsm0 ] ) needs to be done from table ( [ profiletable ] ) and eq.([totalphotonflux ] ) .
we further note that as mentioned before , the use of the adiabatic compression mechanism would modify a given halo profile significantly and this may increase the photon flux by a few orders of magnitude .
fig.([photonfluxvsmchi ] ) shows the plots of photon - flux ( in @xmath194 with @xmath195 gev ) vs the mass of the lsp for @xmath80 and @xmath191 for nfw halo profile in the nusm . here
, @xmath93 and @xmath5 are varied such that @xmath200 tev and @xmath201 tev for @xmath80 and @xmath200 tev and @xmath202 tev for @xmath167 .
only wmap allowed parameter points are shown .
glast would be able to probe up to @xmath203 to @xmath204 gev .
in this analysis we have worked with a non - universal scalar mass scenario in a supergravity framework .
we started with purely phenomenological motivations namely , i ) to manage the fcnc and cp - violation type of constraints by decoupling , ii ) to obtain wmap satisfied values for neutralino relic density for a broad region of parameter space without depending on any delicate mixing of gauginos and higgsinos , iii ) to have radiative electroweak symmetry breaking and iv ) to keep naturalness within control . keeping the above in mind and considering a unified gaugino mass scenario , we used a common scalar mass parameter @xmath5 at the gauge coupling unification scale for the first two - generation of scalars as well as the third generation of sleptons .
the item ( i ) mentioned above would require @xmath5 to be large , and the item ( iv ) would prefer light third generation of squarks and light higgs scalars . for simplicity
we used vanishing third generation of squarks and higgs scalar masses at the unification scale .
in such a scenario with a possibly multi - tev @xmath5 , we first found a large mass effect or in particular large slepton mass effect in the rge of @xmath0 ( _ ie .
_ for the down type of higgs scalar ) that turns the later negative at the electroweak scale almost irrespective of a value of @xmath1 .
extending the semi - analytic solution of @xmath2 by considering the @xmath45 and the @xmath46 terms ( this is required here even for a small @xmath1 ) relevant for the large slepton mass effect we found a hyperbolic branch / focus point like effect in @xmath205 for small @xmath1 .
this causes @xmath14 to be almost independent of @xmath5 for a large domain of the later .
but , with a very large @xmath5 this causes @xmath14 to become very light or this may even turn @xmath2 negative giving rise to no radiative electroweak symmetry breaking .
we further found that because of such large slepton mass effect in the rge , the higgs sector may reach an intense coupling region with all the higgs bosons becoming very light and this may evade the lep2 limit of @xmath18 for a limited region of parameter space . however , constraint from @xmath206 becomes stringent in this region because of lighter @xmath14 . in general , we find relatively lighter @xmath14 or @xmath15 and this fact leads to a large @xmath89-pole annihilation region or funnel region of dark matter even for a small @xmath1 .
this is in contrast to minimal supergravity type of scenarios where funnel region may occur only for large values of @xmath1 .
the nature of the lsp is bino - dominated , thus there is no need of any delicate mixing of binos and higgsinos in order to satisfy the neutralino relic density constraint .
we have also computed the direct detection rates of lsp - nucleon scattering .
the upcoming detectors like xenon-1 t would be able to probe almost the entire region of parameter space .
we have further estimated the indirect detection prospect via computing continuous photon fluxes
. the ongoing glast experiment will be succesfully able to probe the parameter space even for a less cuspy halo profile .
we have also briefly discussed the detection prospect of sparticles in the lhc .
the higgs bosons and the third generation of squarks are light in this scenario .
in addition to charginos and neutralinos the above may be easily probed in the early runs of lhc .
the coefficients appearing in eqs .
[ mueqn ] and [ maeqn ] are given by , here the functions @xmath209 and @xmath210 may be seen in ref .
electroweak scale result of @xmath211 with @xmath212 are shown below .
y_1(t ) = , y_2(t ) = , y_3(t ) = y_3(0)e_3(t ) [ yukawa ] the quantities @xmath213 are defined as follows .
@xmath214 here , @xmath215 , @xmath216 and @xmath217 . * acknowledgments * + dd would like to thank the council of scientific and industrial research , govt .
of india for the support received as a junior research fellow .
uc is thankful to the organizers of the workshop `` tev scale and dark matter ( 2008 ) '' at nordita where the last part of the work was completed .
uc acknowledges useful discussions with m. guchait , a. kundu , b. mukhopadhyaya , d. choudhury , d.p .
roy , jan kalinowski and s. roy . 1 for reviews on supersymmetry , see , _ eg _ , h. p. nilles , phys . rep . * 110 * , 1 ( 1984 ) ; h. e. haber and g. kane , phys . rep . *
117 * , 75 ( 1985 ) ; j. wess and j. bagger , _ supersymmetry and supergravity _ ,
2nd ed . , ( princeton , 1991 ) ; m. drees , p. roy and r. m. godbole , _ theory and phenomenology of sparticles _ , ( world scientific , singapore , 2005 ) .
a. h. chamseddine , r. arnowitt and p. nath , phys .
* 49 * , 970 ( 1982 ) ; r. barbieri , s. ferrara and c. a. savoy , phys .
b * 119 * , 343 ( 1982 ) ; l. j. hall , j. lykken and s. weinberg , phys .
d * 27 * , 2359 ( 1983 ) ; p.
nath , r. arnowitt and a. h. chamseddine , nucl .
b * 227 * , 121 ( 1983 ) ; n. ohta , prog .
phys . * 70 * , 542 ( 1983 ) ; for reviews see @xcite and p. nath , r. arnowitt and a.h .
chamseddine , _ applied n = 1 supergravity _
( world scientific , singapore , 1984 ) .
p. nath and r. arnowitt , phys .
d * 56 * , 2820 ( 1997 ) d. g. cerdeno and c. munoz , jhep * 0410 * , 015 ( 2004 ) [ arxiv : hep - ph/0405057 ] .
j. ellis , k. a. olive and p. sandick , arxiv:0805.2343 [ hep - ph ] ; j. r. ellis , k. a. olive , y. santoso and v. c. spanos , phys . lett .
b * 603 * , 51 ( 2004 ) ; j. r. ellis , t. falk , k. a. olive and y. santoso , nucl .
b * 652 * , 259 ( 2003 ) [ arxiv : hep - ph/0210205 ] ; a. de roeck , j. r. ellis , f. gianotti , f. moortgat , k. a. olive and l. pape , eur .
j. c * 49 * , 1041 ( 2007 ) [ arxiv : hep - ph/0508198 ] .
h. baer , a. mustafayev , e. k. park , s. profumo and x. tata , jhep * 0604 * , 041 ( 2006 ) ; s. bhattacharya , a. datta and b. mukhopadhyaya , arxiv:0804.4051 [ hep - ph ] ; s. bhattacharya , a. datta and b. mukhopadhyaya , arxiv:0809.2012 [ hep - ph ] ; h. baer , a. mustafayev , s. profumo , a. belyaev and x. tata , jhep * 0507 * , 065 ( 2005 ) ; h. baer , a. mustafayev , s. profumo , a. belyaev and x. tata , phys . rev .
d * 71 * , 095008 ( 2005 ) .
u. chattopadhyay , d. choudhury and d. das , phys . rev .
d * 72 * , 095015 ( 2005 ) ; u. chattopadhyay and d. p. roy , phys .
d * 68 * , 033010 ( 2003 ) ; a. corsetti and p. nath , phys .
d * 64 * , 125010 ( 2001 ) ; u. chattopadhyay , a. corsetti and p. nath , phys .
d * 66 * , 035003 ( 2002 ) ; u. chattopadhyay and p. nath , phys .
d * 65 * , 075009 ( 2002 ) ; s. bhattacharya , a. datta and b. mukhopadhyaya , jhep * 0710 * , 080 ( 2007 ) ; s. bhattacharya , a. datta and b. mukhopadhyaya , arxiv:0804.4051 [ hep - ph ] ; k. huitu , j. laamanen , p. n. pandita and s. roy , phys .
d * 72 * , 055013 ( 2005 ) ; k. huitu , r. kinnunen , j. laamanen , s. lehti , s. roy and t. salminen , arxiv:0808.3094 [ hep - ph ] ; s. f. king , j. p. roberts and d. p. roy , jhep * 0710 * , 106 ( 2007 ) [ arxiv:0705.4219 [ hep - ph ] ] ; g. belanger , f. boudjema , a. cottrant , a. pukhov and a. semenov , nucl .
b * 706 * , 411 ( 2005 ) [ arxiv : hep - ph/0407218 ] .
m. dine , r. leigh , and a. kagan , _ phys .
rev . _ * d48 * , 4269 ( 1993 ) ; p. pouliot and n. seiberg , _ phys .
lett . _ * b318 * , 169 ( 1993 ) ; y. nir and n. seiberg , _ phys .
lett . _ * b309 * , 337 ( 1993 ) . s. dimopoulos and g.f .
giudice , _ phys .
lett . _ * b357 * , 573 ( 1995 ) ; a. cohen , d.b . kaplan , and a.e .
nelson , _ phys .
_ * b388 * , 588 ( 1996 ) ; m. misiak , s. pokorski and j. rosiek , adv .
high energy phys . * 15 * , 795 ( 1998 ) [ arxiv : hep - ph/9703442 ] ; p. h. chankowski , k. kowalska , s. lavignac and s. pokorski , arxiv : hep - ph/0507133 .
j. bagger , j. l. feng and n. polonsky , nucl .
b * 563 * , 3 ( 1999 ) ; .
k. agashe and m. graesser , phys .
d * 59 * , 015007 ( 1998 ) ; j. a. bagger , j. l. feng , n. polonsky and r. j. zhang , phys . lett .
b * 473 * , 264 ( 2000 ) .
h. baer , c. balazs , p. mercadante , x. tata and y. wang , phys .
d * 63 * , 015011 ( 2000 ) [ arxiv : hep - ph/0008061 ] .
j. l. feng , k. t. matchev and t. moroi , phys .
d * 61 * , 075005 ( 2000 ) ; phys .
lett . *
84 * , 2322 ( 2000 ) ; j. l. feng , k. t. matchev and f. wilczek , phys .
b * 482 * , 388 ( 2000 ) ; j. l. feng and f. wilczek , phys .
b * 631 * , 170 ( 2005 ) ; u. chattopadhyay , t. ibrahim and d. p. roy , phys .
d * 64 * , 013004 ( 2001 ) ; u. chattopadhyay , a. datta , a. datta , a. datta and d. p. roy , phys .
b * 493 * , 127 ( 2000 ) ; s. p. das , a. datta , m. guchait , m. maity and s. mukherjee , eur .
j. c * 54 * , 645 ( 2008 ) , [ arxiv:0708.2048 [ hep - ph ] ] .
k. l. chan , u. chattopadhyay and p. nath , phys .
d * 58 * , 096004 ( 1998 ) ; [ arxiv : hep - ph/9710473 ] ; u. chattopadhyay , a. corsetti and p. nath , phys .
d * 68 * , 035005 ( 2003 ) [ arxiv : hep - ph/0303201 ] ; + see also : 1st article of ref.@xcite . c. d. mcmullen and s. nandi , phys .
d * 75 * , 095001 ( 2007 ) d. feldman , z. liu and p. nath , jhep * 0804 * , 054 ( 2008 ) ; d. feldman , z. liu and p. nath , phys .
lett . * 99 * , 251802 ( 2007 ) [ erratum - ibid .
* 100 * , 069902 ( 2008 ) ] .
l. randall and r. sundrum , * b557*(1999 ) 79 .
giudice , m.a .
luty , h. murayama , and r. rattazzi , j. high energy phys . * 9812 * , 027 ( 1998 ) ; t. gherghetta , g.f .
giudice , and j.d .
wells , * b559 * ( 1999 ) 27 .
a. brignole , l. e. ibanez and c. munoz , nucl .
b * 422 * , 125 ( 1994 ) [ erratum - ibid .
b * 436 * , 747 ( 1995 ) ] l. e. ibanez , d. lust and g. g. ross , phys .
b * 272 * , 251 ( 1991 ) .
d. g. cerdeno , t. kobayashi and c. munoz , arxiv:0709.0858 [ hep - ph ] .
h.e . haber and y. nir , phys
b306 * 327 ( 1993 ) ; h.e .
haber , hep - ph/9505240 ; a. dobado , m.j .
herrero and s. penaranda , eur .
j. * c17 * 487 ( 2000 ) ; j.f .
gunion and h.e .
haber , phys .
d67 * 075019 ( 2003 ) .
a. djouadi and y. mambrini , jhep * 0612 * , 001 ( 2006 ) ; e. boos , v.bunichev , a. djouadi and h.j .
schreiber , phys .
* b622 * 311 ( 2005 ) ; e. boos , a. djouadi and a. nikitenko , phys . lett . *
b578 * , 384 ( 2004 ) ; e. boos , a. djouadi , m. muhlleitner and a. vologdin , phys . rev .
* d66 * 055004 ( 2002 ) .
g. gamberini , g. ridolfi and f. zwirner , nucl .
b * 331 * , 331 ( 1990 ) ; v. d. barger , m. s. berger and p. ohmann , phys .
d * 49 * , 4908 ( 1994 ) ; for two - loop effective potential see : s. p.
martin , phys .
d * 66 * , 096001 ( 2002 ) .
a. b. lahanas , n. e. mavromatos and d. v. nanopoulos , int .
j. mod .
d * 12 * , 1529 ( 2003 ) [ arxiv : hep - ph/0308251 ] ; c. munoz , int .
j. mod .
a * 19 * , 3093 ( 2004 ) [ arxiv : hep - ph/0309346 ] ; manuel drees , plenary talk at 11th international symposium on particles , strings and cosmology ( pascos 2005 ) , gyeongju , korea , 30 may - 4 jun 2005 ( published in aip conf.proc . , * 805 * , 48 - 54 , ( 2006 ) .
e. komatsu _ et al .
_ [ wmap collaboration ] , arxiv:0803.0547 [ astro - ph ] . j. r. ellis , t. falk and k. a. olive , phys .
b * 444 * , 367 ( 1998 ) ; j. r. ellis , t. falk , k. a. olive and m. srednicki , astropart .
* 13 * , 181 ( 2000 ) [ erratum - ibid .
* 15 * , 413 ( 2001 ) ] ; a. lahanas , d. v. nanopoulos and v. spanos , phys .
d * 62 * , 023515 ( 2000 ) ; r. arnowitt , b. dutta and y. santoso , nucl . phys .
b * 606 * , 59(2001 ) ; t. nihei , l. roszkowski and r. ruiz de austri , jhep * 0207 * , 024 ( 2002 ) ; v. a. bednyakov , h. v. klapdor - kleingrothaus and v. gronewold , phys .
d * 66 * , 115005 ( 2002 ) . c. boehm , a. djouadi and manuel drees , phys .
d * 62 * , 035012(2000 ) ; j. r. ellis , k. a. olive and y. santoso , astropart .
* 18 * , 395 ( 2003 ) ; + see also ref.@xcite .
r. arnowitt and p. nath , phys .
* 70 * , 3696 ( 1993 ) ; h. baer and m. brhlik , phys . rev . * d53 * , 597 ( 1996 ) , phys . rev .
d * 57 * , 567 ( 1998 ) ; h. baer , m. brhlik , m. diaz , j. ferrandis , p. mercadante , p. quintana and x. tata , phys . rev . * d63 * , 015007 ( 2000 ) ; j. r. ellis , t. falk , g. ganis , k. a. olive and m. srednicki , phys .
b * 510 * , 236 ( 2001 ) ; a. b. lahanas and v. c. spanos , eur .
j. c * 23 * , 185 ( 2002 ) ; a. djouadi , m. drees and j. kneur , phys .
b * 624 * 60 ( 2005 ) .
s. heinemeyer , w. hollik and g. weiglein , ( 425,265,2006 ) ; s. heinemeyer , hep - ph/0408340 ; s. heinemeyer , int .
j. mod .
a * 21 * , 2659 ( 2006 ) ; g. degrassi , s. heinemeyer , w. hollik , p. slavich and g. weiglein , eur .
j. c * 28 * , 133 ( 2003 ) ; b. c. allanach , a. djouadi , j. l. kneur , w. porod and p. slavich , jhep * 0409 * , 044 ( 2004 ) .
s. chen , et al .
, cleo collaboration , phys .
lett . * 87 * , 251807 ( 2001 ) , hep - ex/0108032 ; p. koppenburg _ et al . _ [ belle collaboration ] , phys .
lett . * 93 * , 061803 ( 2004 ) b. aubert , et al . , babar collaboration , hep - ex/0207076 .
t. aaltonen _ et al . _
[ cdf collaboration ] , phys .
lett . * 100 * , 101802 ( 2008 ) .
ellis , k.a .
olive and v.c .
spanos , phys .
b * 624 * , 47 ( 2005 ) ; t. ibrahim and p. nath , phys .
d * 67 * , 016005 ( 2003 ) ; s. baek , p. ko , and w. y. song , jhep * 0303 * , 054 ( 2003 ) ; j. k. mizukoshi , x. tata and y. wang , phys . rev .
d * 66 * , 115003 ( 2002 ) ; r. arnowitt , b. dutta , t. kamon and m. tanaka , phys .
b * 538 * ( 2002 ) 121 ; a. dedes , h. k. dreiner , u. nierste , and p. richardson , phys .
lett . * 87 * , 251804 ( 2001 ) ; k. s. babu and c. kolda , phys .
lett . * 84 * , 228 ( 2000 ) ; s. r. choudhury and n. gaur , phys .
b * 451 * , 86 ( 1999 ) .
h. baer , a. belyaev , t. krupovnickas and j. ofarrill , jcap * 0408 * , 005 ( 2004 ) ; h. baer and j. ofarrill , jcap * 0404 * , 005 ( 2004 )
. m. passera , w. j. marciano and a. sirlin , phys .
d * 78 * , 013009 ( 2008 ) .
g. bertone , d. hooper and j. silk , phys .
* 405 * , 279 ( 2005 ) ; g. bertone and d. merritt , mod .
a * 20 * , 1021 ( 2005 ) ; j. carr , g. lamanna and j. lavalle , rept .
phys . * 69 * , 2475 ( 2006 ) .
m. w. goodman and e. witten , phys .
d * 31 * , 3059 ( 1985 ) ; k. greist , phys . rev .
* d38 * , 2357(1988 ) ; j. ellis and r. flores , nucl . phys . * b307 * , 883(1988 ) ; r. barbieri , m. frigeni and g. giudice , nucl . phys . * b313 * , 725(1989 ) ; a. bottino et.al .
, phys . lett . *
b295 * , 330(1992 ) ; m. drees and m.m .
nojiri , phys .
rev.*d48*,3483(1993 ) ; j. r. ellis , a. ferstl and k. a. olive , phys . lett .
b * 481 * , 304 ( 2000 ) ; j. d. vergados , lect .
notes phys .
* 720 * , 69 ( 2007 ) ; v. k. oikonomou , j. d. vergados and c. c. moustakidis , nucl .
b * 773 * , 19 ( 2007 ) ; j. ellis , k. a. olive and c. savage , phys .
d * 77 * , 065026 ( 2008 ) ; d. feldman , z. liu and p. nath , phys .
b * 662 * , 190 ( 2008 ) ; d. feldman , z. liu and p. nath , arxiv:0808.1595 [ hep - ph ] .
m. drees and m. nojiri , phys . rev .
* d47 * , 4226 ( 1993 ) and phys . rev .
* d48 * , 3483 ( 1993 ) ; m. a. shifman , a. i. vainshtein and v. i. zakharov , phys . lett . * 78b * , 443 ( 1978 ) ; a. i. vainshtein , v. i. zakharov and m. a. shifman , usp
nauk * 131 * , 537 ( 1980 ) . ....
http://www.lngs.infn.it/lngs_infn/contents/lngs_en/research/ experiments_scientific_info / experiments / current / xenon / collaboration.htm ; for a limit plot see : http://dendera.berkeley.edu/plotter/entryform.html .... a. bouquet , p. salati and j. silk , phys .
d * 40 * , 3168 ( 1989 ) ; f. w. stecker , phys .
b * 201 * , 529 ( 1988 ) ; m. urban , a. bouquet , b. degrange , p. fleury , j. kaplan , a. l. melchior and e .
pare , phys .
b * 293 * , 149 ( 1992 ) ; v. s. berezinsky , a. v. gurevich and k. p. zybin , phys .
b * 294 * , 221 ( 1992 ) ; v. berezinsky , a. bottino and g. mignola , phys .
b * 325 * , 136 ( 1994 ) ; g. bertone , g. sigl and j. silk , mon . not .
* 337 * , 98 ( 2002 ) [ arxiv : astro - ph/0203488 ] ; t. bringmann , l. bergstrom and j. edsjo , jhep * 0801 * , 049 ( 2008 ) .
y. mambrini , c. munoz , e. nezri and f. prada , jcap * 0601 * , 010 ( 2006 ) ; y. mambrini , c. munoz and e. nezri , jcap * 0612 * , 003 ( 2006 ) ; l. roszkowski , r. r. de austri , j. silk and r. trotta , arxiv:0707.0622 [ astro - ph ] ; s. profumo and a. provenza , jcap * 0612 * , 019 ( 2006 ) ; h. baer , a. mustafayev , e. k. park and s. profumo , jhep * 0507 * , 046 ( 2005 ) ; s. dodelson , d. hooper and p. d. serpico , phys . rev .
d * 77 * , 063512 ( 2008 ) ; d. hooper , g. zaharijas , d. p. finkbeiner and g. dobler , phys .
d * 77 * , 043511 ( 2008 ) ; d. hooper and e. a. baltz , arxiv:0802.0702 [ hep - ph ] ; n. bernal , a. goudelis , y. mambrini and c. munoz , arxiv:0804.1976 [ hep - ph ] ; a. birkedal , k. t. matchev , m. perelstein and a. spray , arxiv : hep - ph/0507194 ; l. bergstrom and p. ullio , ` full one - loop calculation of neutralino annihilation into two photons ' , _ nucl .
phys . _ * b504 * , 27 ( 1997 ) ; z. bern p. gondolo and m. perelstein , ` neutralino annihilation into two photons ' , _ phys .
lett . _ * b411 * , 86 ( 1997 ) ; u. chattopadhyay , d. das , p. konar and d. p. roy , phys . rev .
d * 75 * , 073014 ( 2007 ) ; u. chattopadhyay , d. choudhury , m. drees , p. konar and d. p. roy , phys .
b * 632 * , 114 ( 2006 ) .
p. ullio and l. bergstrom , phys .
d * 57 * , 1962 ( 1998 ) [ arxiv : hep - ph/9707333 ] .
b. moore , t. quinn , f. governato , j. stadel and g. lake , mon . not .
soc . * 310 * , 1147 ( 1999 ) [ arxiv : astro - ph/9903164 ] .
g. r. blumenthal , s. m. faber , r. flores and j. r. primack , astrophys .
j. * 301 * , 27 ( 1986 ) ; r. jesseit , t. naab and a. burkert , astrophys .
j. lett * 571 * , l89 ( 2002 ) , arxiv : astro - ph/0204164 ; o. y. gnedin , a. v. kravtsov , a. a. klypin and d. nagai , astrophys .
j. * 616 * , 16 ( 2004 ) [ arxiv : astro - ph/0406247 ] ; f. prada , a. klypin , j. flix , m. martinez and e. simonneau , arxiv : astro - ph/0401512 ; | a non - universal scalar mass supergravity type of model is explored where the first two generation of scalars and the third generation of sleptons may be very massive .
the lighter or vanishing third generation of squarks as well as higgs scalars at the unification scale cause the radiative electroweak symmetry breaking constraint to be less prohibitive .
thus , both fcnc / cp - violation problems as well as the naturalness problem are within control .
we identify a large slepton mass effect in the rge of @xmath0 ( for the down type of higgs ) that may turn the later negative at the electroweak scale even for a small @xmath1 .
a hyperbolic branch / focus point like effect is found for @xmath2 that may result in very light higgs spectra .
the lightest stable particle is dominantly a bino that pair annihilates via higgs exchange , giving rise to a wmap satisfied relic density region for all @xmath1 .
detection prospects of such lsps in the upcoming dark matter experiments both of direct and indirect types ( photon flux ) are interesting .
the higgs bosons and the third generation of squarks are light in this scenario and these may be easily probed besides charginos and neutralinos in the early runs of lhc .
+ pacs no : 04.65.+e , 13.40em , 14.60ef , 13.85.-t , 14.80.ly -1 cm 16.5 cm 21.5 cm -0.5 cm # 1 # 1 # 1 # 1 ( # 1,#2,#3)*#1 * , # 2 ( # 3)#1#2#3#4#1 * # 2 * , # 3 ( # 4 ) # 1#2#3jhep * # 1 * , # 3 ( # 2 ) # 1#2#3jhep * # 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
d * # 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
d * # 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
lett . *
# 1 * , # 3 ( # 2 ) # 1#2#3phys .
rev .
lett . * # 1 * , # 3 ( # 2 ) # 1#2#3nucl .
phys .
b * # 1 * , # 3 ( # 2 ) # 1#2#3nucl .
phys .
b * # 1 * , # 3 ( # 2 ) # 1arxiv : hep - ph/#1 ( # 1,#2,#3)phys . rep .
( # 1,#2,#3 ) ( # 1,#2,#3)eur .
phys .
j. c ( # 1,#2,#3 ) 0_b_0 22 2 ^ 2_w 2m_z^2 2b2 v#1v_#1 # 1 ^ 0_#1 # 1m_#1 # 1m_#1 # 1m_^0_#1 # 1m_^+_#1 2wsec^2_w 2(g-2 ) _ = 11 # 1#2 # 1#1| # 1| # 1 # 1#1 # 1#1 versim#1#2 2_2^+ 2_2 ^ - # 1#2 * higgs funnel region of susy dark matter for small @xmath1 and renormalization group effects on pseudoscalar higgs boson with scalar mass non - universality + * 0.5 cm utpal chattopadhyay and debottam das 0.2 cm _ department of theoretical physics + indian association for the cultivation of science + 2a and 2b raja s.c .
mullick road + jadavpur , kolkata 700 032 , india _ |
the relative number and astrophysical properties of dwarf galaxies represent one of the major challenges to the widely accepted @xmath12+cold dark matter ( @xmath12cdm ) cosmological model of structure formation .
for instance , the predicted number of dm haloes around a milky way ( mw)-sized halo exceeds the observed number by at least an order of magnitude ( e.g. * ? ? ? * ; * ? ? ? * ) , and their central densities at a given mass are lower than inferred from simulations ( e.g. * ? ? ? * ; * ? ? ? * ) .
understanding these challenges has largely been left to studies of the local group ( lg ; with exceptions , e.g. * ? ? ?
* ) where the faintest satellite galaxies can be discovered and studied in detail . however , a diversity of halo to halo substructure is expected due to varying accretion histories @xcite and inhomogeneous reionization @xcite ; it is also possible that the lg is an outlier in its satellite properties .
a systematic study of the faint satellite populations of a large number of galaxies in a variety of environments is critical before the @xmath12cdm picture of galaxy formation can be confirmed .
pioneering studies of faint satellite systems beyond the lg have begun .
the m81 group of galaxies ( at @[email protected] mpc ) has been the subject of a resolved stellar population search for satellites utilizing deep ground based imaging and _ hubble space telescope _ ( hst ) follow - up @xcite , finding satellites down to @xmath13 and largely confirming the faint satellite deficit with respect to @xmath12cdm seen in the lg .
studies of unresolved low surface brightness satellites have additionally been undertaken for more distant groups ( e.g. , m101 , * ? ? ?
* ) , where dwarf galaxy distance measurements are still necessary to produce conclusive results .
we have begun the panoramic imaging survey of centaurus and sculptor ( pisces ) to significantly increase the sample of massive galaxy halos with a complete census of satellites down to @xmath14@xmath0@xmath48 mag .
pisces is focused on two nearby massive galaxies the spiral ngc253 and the elliptical ngc5128/centaurusa ( cena ) , that reside in a loose group of galaxies ( sculptor ) and in a rich group ( cena ) , respectively ( see * ? ? ?
* for more details ) .
pisces will identify faint satellites and streams from resolved stellar light out to @xmath15@xmath0150 kpc in each system , allowing for direct comparison with the mw , the m31 pan - andromeda archaeological survey ( pandas ; * ? ? ? * ) and m81 @xcite .
out of the planned survey area of @xmath16 deg@xmath10 , we have already collected data over @xmath17 deg@xmath10 around cena . in this _ letter _ , we present the first results of our survey around cena : the discovery of a close pair of faint dwarf galaxies , likely associated with cena .
the dwarfs were immediately visible in our pisces survey data and are located near cena s major axis in the north - east direction , at a projected distance of @xmath18 kpc . for the more luminous dwarf , a modest surface brightness enhancement can be recognized in archival digital sky survey images .
in [ sec : obs ] we describe our data and reduction procedure , while in [ analysis ] we present the physical properties derived for the newly discovered dwarfs . in [ concl ]
we discuss the possibility of these dwarfs constituting a pair of galaxies and draw our conclusions .
the data presented here were acquired on 2013 june 4 ( ut ) , as part of the pisces survey , with megacam @xcite at the magellan clay 6.5 m telescope , which has a @xmath024@xmath1924 field of view and a binned pixel scale of 016 .
the seeing was excellent throughout the night , with a median of @xmath20 , and the conditions were photometric .
the final stacked images had a total exposure time of @xmath21 sec for the @xmath22-band and @xmath23 sec for the @xmath24-band .
initial data reduction ( image detrending , astrometry and stacking ) was performed by the smithsonian astrophysical observatory telescope data center , using a code developed by m. conroy , j. roll and b. mcleod .
we perform point spread function ( psf)-fitting photometry on the final stacked images , adopting the suite of programs daophot and allframe @xcite .
we construct a psf for each band selecting @xmath25 bright stars across the image .
we then fit the psf to all objects @xmath26 above the background for each image , and compute the coordinate transformations between filters with daomatch / daomaster @xcite . finally , we perform simultaneous photometry of all the objects detected in both filters in order to obtain deeper catalogues .
we only keep those objects satisfying the criteria @xmath27 and @xmath28 . the same night ,
several equatorial sloan digital sky survey fields were observed at different airmasses to obtain zeropoints , color terms and extinction coefficients , and thus calibrate the instrumental magnitudes to the sdss system .
stars have been individually corrected for galactic extinction @xcite , which has an average value of @[email protected] .
the photometric errors and incompleteness of our data have been computed via artificial star experiments .
we have injected @xmath0@xmath30 fake stars with a uniform distribution on each of the stacked images , divided into 20 experiments in order not to increase the stellar crowding artificially .
the fake stars cover the whole relevant range in color - magnitude space , and additionally reach @xmath31 mag ( @xmath02 mag fainter than the faintest real recovered stars ) . the same psf - fitting photometry procedure used for the real data has been performed on the fake stars .
we find the overall ( color - averaged ) @xmath32 completeness limit to be @xmath33 and @xmath34 .
note that in the central @xmath01 arcmin of cena - mm - dw1 the crowding is higher than in the rest of the pointing , and the @xmath32 completeness limits correspond to @xmath35 and @xmath36 .
figure [ cmds ] shows the dereddened color magnitude diagrams ( cmds ) for our newly found dwarfs , cena - mm - dw1 and cena - mm - dw2 .
the error bars ( from the artificial star experiments ) show the typical magnitude and color uncertainty as a function of @xmath22-band magnitude , and the @xmath32 completeness limits ( for uncrowded regions ) are drawn .
we begin this section by broadly discussing the stellar populations of our newly found dwarfs , and visualizing these new systems via red giant branch ( rgb ) stellar maps .
we then discuss the distance , structure , metallicity and luminosity of this faint pair of cena dwarfs .
figure [ cmds ] shows our dereddened cmds where stars within the half light radius ( calculated in [ sec : struct ] ) are plotted .
the field cmds ( rescaled to the area of each dwarf ) in the right panels are from 3 rectangular regions with an area of @xmath37 deg@xmath10 each , two of which are located @xmath38 arcmin closer / further away than our targets along cena s major axis ( see figure [ rgb_maps ] ) , and the third one located perpendicular to it , to the east of the two dwarfs .
this choice is driven by the necessity of taking into account a possible density gradient and small - scale substructures in cena s halo .
the field cmds show the main contaminants , i.e. galactic foreground ( almost vertical ) sequences at @[email protected] and @xmath01.3 , unresolved background galaxies centered at @xmath39@xmath40 , and cena rgb halo stars ( sparsely populated at these large projected galactocentric distances , @xmath090 - 95 kpc ) .
a prominent rgb is evident for both dwarfs .
we plot the overall distribution of rgb stars in the megacam field in figure [ rgb_maps ] , utilizing all stars from the red selection box in the cmds ( figure [ cmds ] ) .
these predominantly old stars are clearly clumped at the position of the dwarfs ( figure [ rgb_maps ] , right panel ) .
we can not rule out the presence of younger , main sequence stars : for cena - mm - dw2 , a @xmath01@xmath41 excess of blue sources can be seen at the center of the dwarf , but we do not consider this conclusive due to the strong background galaxy contamination .
cena - mm - dw1 s cmd shows an overdensity of sources above the rgb with respect to foreground contaminants , which may be luminous asymptotic giant branch stars .
if real , they belong to an episode of star formation that occurred @xmath42 gyr ago .
a more definitive assessment of the dwarfs star formation history will be forthcoming after planned hst observations of the pair .
we derive the distance to the new dwarfs using the tip of the rgb ( trgb ) method , which utilizes the brightest , metal poor rgb stars as a standard candle ( e.g. * ? ? ?
* ; * ? ? ?
we employ a sobel edge detection filter to identify a sharp transition in the @xmath22-band luminosity function , with a color cut of @xmath43 in order to minimize contamination from foreground galactic stars ( at red colors ) and unresolved background galaxies ( bluer than the rgb ) .
we find @xmath44 for cena - mm - dw1 and @xmath45 for cena - mm - dw2 , where the errors mainly depend on photometric uncertainties rather than the number of stars .
these values correspond to distance moduli of @xmath46 and @xmath47 , respectively , once adopting a theoretically calibrated trgb absolute value of @xmath48 ( as computed in @xcite for sdss bands ) . as a test
, we derive trgb distances for the adopted field regions , i.e. for stars belonging to cena s outer halo .
we obtain @xmath49 , in excellent agreement with the average of literature values using several methodologies ( @xmath50 ; * ? ? ?
the derived distances put the two dwarfs at roughly the same distance as cena , thus suggesting they are its satellites .
follow - up hst data will be used to refine these distances and investigate the possibility that the two dwarfs form a physical pair ( see discussion in [ concl ] ) . to quantify the structure and luminosities of our new dwarf discoveries , we work with rgb stars from the selection boxes shown in figure [ cmds ] .
first , the dwarf centers are determined via an iterative process , computing the average of the stellar positions within circles of decreasing radius .
from there , the position angle and ellipticity are measured using the method of moments for the rgb spatial distributions ( e.g. * ? ? ?
the final values are reported in table [ tab1 ] .
for cena - mm - dw2 we only obtain upper limits on the ellipticity ( and thus have a poorly determined position angle ) because of the low number of available stars .
we derive radial density profiles by counting the number of rgb stars in elliptical ( circular ) radii for cena - mm - dw1 ( cena - mm - dw2 ) .
the field level is estimated from the average number of rgb stars present in our adopted field regions , and is entirely consistent with the number density for cena - mm - dw1 beyond @xmath51 arcmin .
we subtract the field level from the derived radial profiles and correct them for incompleteness ( including a correction for the excess crowding incompleteness within @xmath52 arcmin of cena - mm - dw1 ) . given the faint nature of these dwarfs , resolved rgb stars are necessary to trace their radial profiles as opposed to integrated light .
however , we adopt the latter to convert the observed number density into proper surface brightness values . this can be done by performing integrated aperture photometry in the central regions of the dwarfs , and scaling the number density profile to the central surface brightness values ( e.g. , @xcite ) .
the zeropoint between the two profiles is computed from their overlapping region , i.e. the innermost @xmath53 arcmin for cena - mm - dw1 and cena - mm - dw2 , respectively .
the integrated light photometry is performed by summing up the flux within these apertures , subtracting the median sky level and applying zeropoint , color terms and airmass corrections ( sect .
[ sec : obs ] ) .
the composite surface brightness profiles are shown in figure [ sbprofs ] .
note that we reach a remarkable surface brightness of @xmath54 mag / arcsec@xmath10 ( before field subtraction ) .
srsic models are fit via least squares minimization to the composite surface brightness profiles ( figure [ sbprofs ] ) , for datapoints within @xmath11 arcmin for cena - mm - dw1 and @xmath55 arcmin for cena - mm - dw2 . beyond these radii ,
the profile of each dwarf is contaminated by the light of the other dwarf ( cena - mm - dw2 can be recognized as an enhancement in cena - mm - dw1 s profile at @xmath56 arcmin ) .
the resulting parameters of the sersic profiles are reported in table [ tab1 ] , and are consistent with being exponential profiles .
the uncontaminated profiles trace out to @xmath57 times the derived half - light radii for the dwarfs .
we derive their luminosities by integrating the best - fit srsic profiles , and obtain @xmath58 for cena - mm - dw1 and @xmath59 for cena - mm - dw2 ( uncertainties from error propagation ) . using the transformation between sdss and johnson - cousins filters reported by @xcite , the @xmath60-band total magnitudes of the dwarfs are @xmath5 for cena - mm - dw1 and @xmath6 for cena - mm - dw2 .
the two dwarfs host relatively metal - poor stellar populations , as can be seen from the isochrones overplotted on the cmds in figure [ cmds ] . to quantify this ,
we use the standard method for computing photometric metallicities ( e.g. * ? ? ?
* ) : we interpolate between solar - scaled isochrones with a fixed age of 12 gyr and [ fe / h ] varying from @xmath61 to @xmath62 , adopting the dartmouth stellar evolutionary database @xcite . a metallicity value is obtained for each individual rgb star with magnitude @xmath63 .
the mean values we deduce are [ fe / h]@xmath64 for cena - mm - dw1 and [ fe / h]@xmath65 for cena - mm - dw2 ( as plotted in figure [ cmds ] ) .
the spread of the rgb is consistent with photometric errors , however we can not exclude an intrinsic range of metallicities for cena - mm - dw1 in particular .
follow - up data will be crucial to assess the metallicity content of these dwarfs more precisely .
we have presented the discovery of a faint pair of ( likely ) satellites of the elliptical galaxy cena , discovered within our pisces survey .
cena - mm - dw1 and cena - mm - dw2 are located at a projected distance of 92 kpc from the center of their parent galaxy and lie at its approximate distance ( @xmath66 mpc ) , although the uncertainties call for deeper photometric follow - up .
both dwarfs contain predominantly old ( @xmath012 gyr ) and metal - poor stellar populations ( [ fe / h]@xmath67 ) , while cena - mm - dw1 likely also contains intermediate age asymptotic giant branch stars .
neither cena - mm - dw1 or cena - mm - dw2 are detected in the hi parkes all sky survey ( hipass ; * ? ? ?
* ) , with 3-@xmath41 upper mass limits of @xmath68 . the derived structural and luminosity parameters of cena - mm - dw1 and cena - mm - dw2 ( table [ tab1 ] )
place them within the main locus defined by mw and m31 dwarfs with similar properties ( figure [ rhmv ] ) .
of some note is the low central surface brightness of cena - mm - dw1 ( @xmath69 mag / arcsec@xmath10 ) , which lies at the edge of the m31 satellites distribution in a surface brightness versus luminosity plot .
its properties are comparable to a few m31 satellites ( andxxiii , andxix , laci and casiii ; * ? ? ?
* ; * ? ? ?
* ) with large half light radii ( @xmath70 kpc ) and unusually low central surface brightness ( @xmath71 mag / arcsec@xmath10 ) .
no mw satellite shows such a low central surface brightness and large half light radius at the same time .
it has been suggested that m31 companions with @xmath72 have lower surface brightness and are more extended with respect to mw companions with similar luminosity @xcite , although the inclusion of the latest results for the m31 subgroup have substantially decreased this difference @xcite .
the new dwarfs distances are consistent with each other , and their very small angular separation ( 3 arcmin or 3.1 kpc at the distance of the dwarfs ) suggests we are looking at the first , faint satellite of a satellite in an external group of galaxies . according to simulations , groups of dwarfs infalling into the potential well of a giant host are common @xcite , although the predicted pairs of satellites ( with comparable luminosities ) are fewer than those observed in the lg ( e.g. * ? ? ?
examples of dwarf associations are present within the lg : the magellanic clouds , ngc147/ngc185 ( @xcite , but see also @xcite ) , leo iv / leo v / crater ( * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* but see @xcite ) , and for all these pairs a common infall to the lg has been suggested @xcite .
note that all the galaxies within these pairs display similar luminosities , while cena - mm - dw1 is @xmath73 mag brighter than cena - mm - dw2 . in some cases , stellar streams / overdensities with small angular separation are interpreted as dynamically associated structures , or even remnants of infalling groups onto the mw s halo , where the most massive of the dwarfs has already been disrupted ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
@xcite suggest that @xmath74 of faint dwarfs are likely to have fallen onto the mw as satellites of more massive dwarfs . for some of the cited pairs , tidal perturbances / streams
have been observed , even though their origin is still under discussion .
intriguingly , cena - mm - dw2 lies only @xmath11 arcmin in projection away from its more massive companion and does not display clear signs of distortion , however the upper limit for its ellipticity ( @xmath75 ) is high , and possible substructures may be hiding at fainter magnitudes .
we compute cena - mm - dw1 s tidal radius from a king profile , obtaining @xmath76 kpc , while its jacobi radius is @xmath77 kpc ( assuming @xmath78 , a projected radius from cena of 92 kpc and a cena mass of @xmath79 , @xcite ) , thus cena - mm - dw2 may be within its gravitational influence .
the cmds derived from deeper hst imaging will allow for an improved rejection of background galaxies , and thus a more accurate investigation of cena - mm - dw2 s structural parameters and possible tidal distortions .
if this speculation is confirmed , we might be looking at the last moments of cena - mm - dw2 before it is accreted by its more massive companion cena - mm - dw1 .
lcc + ra ( h : m : s ) & 13:30:14.26@xmath80 & 13:29:57.34@xmath80 + dec ( d : m : s ) & @xmath81:53:35.8@xmath82 & @xmath81:52:22.6@xmath82 + @xmath83 ( mag ) & @xmath84 & @xmath85 + d ( mpc ) & @xmath86 & @xmath87 + @xmath88 & @xmath89 & @xmath900.67 + @xmath91 ( n to e ; @xmath92 ) & @xmath93 & unconstrained + @xmath94 ( mag / arcsec@xmath10 ) & @xmath95 & @xmath96 + @xmath97 ( arcmin ) & @xmath98 & @xmath99 + @xmath97 ( kpc ) & @xmath100 & @xmath8 + @xmath101 ( sersic index ) & @xmath102 & @xmath103 + @xmath104 ( mag / arcsec@xmath10 ) & @xmath105 & @xmath106 + @xmath107 ( mag ) & @xmath108 & @xmath6 + @xmath109 ( @xmath110 ) & @xmath111 & @xmath112 + [ tab1 ]
we warmly thank maureen conroy , john roll and sean moran for their prolonged efforts and help related to megacam . dc wishes to kindly thank the hospitality of the mullard space science laboratory , university college of london , where part of this work has been carried out .
djs , js and pg acknowledge support from nsf grant ast-1412504 ; pg acknowledges additional support from nsf grant ast-1010039 .
this paper uses data products produced by the oir telescope data center , supported by the smithsonian astrophysical observatory . , s. , martin , n. , de jong , j. , et al .
2012 , in astronomical society of the pacific conference series , vol .
458 , galactic archaeology : near - field cosmology and the formation of the milky way , ed .
w. aoki , m. ishigaki , t. suda , t. tsujimoto , & n. arimoto , 153 | as part of the panoramic imaging survey of centaurus and sculptor ( pisces ) we report the discovery of a pair of faint dwarf galaxies ( cena - mm - dw1 and cena - mm - dw2 ) at a projected distance of @xmath090 kpc from the nearby elliptical galaxy ngc 5128 ( cena ) .
we measure a tip of the red giant branch distance to each dwarf , finding @[email protected] mpc for cena - mm - dw1 and @[email protected] mpc for cena - mm - dw2 , both of which are consistent with the distance to ngc 5128 .
a qualitative analysis of the color magnitude diagrams indicates stellar populations consisting of an old , metal - poor red giant branch ( @xmath312 gyr , [ fe / h]@[email protected] to @xmath41.9 ) .
in addition , cena - mm - dw1 seems to host an intermediate - age population as indicated by its candidate asymptotic giant branch stars .
the derived luminosities ( @xmath5 for cena - mm - dw1 and @xmath6 for cena - mm - dw2 ) and half - light radii ( @xmath7 kpc for cena - mm - dw1 and @xmath8 kpc for cena - mm - dw2 ) are consistent with those of local group dwarfs .
cena - mm - dw1 s low central surface brightness ( @xmath9 mag / arcsec@xmath10 ) places it among the faintest and most extended m31 satellites .
most intriguingly , cena - mm - dw1 and cena - mm - dw2 have a projected separation of only 3 arcmin ( @xmath11 kpc ) : we are possibly observing the first , faint satellite of a satellite in an external group of galaxies . |
for many years now , it has been appreciated that there are distinct advantages in describing quantum resonances and the quantum phenomena associated to the production , evolution and decay of resonances in terms of resonant or gamow states , since many physical effects are then readily expressed and evaluated@xcite . in this work
, we will give closed analytical expressions for the complex berry phase of an open quantum system in a resonant state of a hermitian hamiltonian with non - self - adjoint boundary conditions , and we will discuss some of its properties . during the last fourteen years
the geometric phase factors arising in the adiabatic evolution of quantum systems@xcite have been the subject of many investigations@xcite .
the early literature was mostly concerned with the geometric phase factors of closed systems driven by hermitian hamiltonians@xcite .
more recently there has been a substantial interest in the complex geometric phase acquired by the eigenstates of open quantum systems .
this problem arises naturally in connection with various experiments which , by their very essence , require the observation of the geometric phase in metastable states .
the berry phase in the optical supermode propagation in a free laser , which is a classical system described by a schrdinger - like equation with a non - hermitian hamiltonian , was studied by dattoli et al.@xcite .
the measurement of the geometric phase in atomic systems with two energy levels , one of which at least is metastable , was also described in terms of a non - hermitian hamiltonian by miniatura et al.@xcite .
the validity of the adiabatic approximation for dissipative , two level systems driven by non - hermitian hamiltonians was examined by nenciu and rasche@xcite , and by kvitsinsky and putterman@xcite , who also established that the berry phase is complex in this case .
sun@xcite proposed a higher - order adiabatic approximation for two - level non - hermitian hamiltonians , and showed that the holonomy structure associated to the berry phase factor in this non - hermitian case is the non - unitary generalization of the holonomy structure of the hermitian case .
in two previous papers@xcite , we gave explicit expressions for the geometric phase of true resonant states , defined as complex energy eigenstates of a hermitian hamiltonian which satisfy purely outgoing wave boundary conditions at infinity@xcite , and pointed out some of the mathematically interesting and physically relevant properties resulting from the extended nature of the singularities in parameter space associated with the occurrence of accidental degeneracies of two resonances .
in this conection , in another paper we showed that the codimension of the accidental degeneracy of @xmath1 resonances differ significantly from those of bound states , we also showed that , close to a crossing of two resonant states , the topological structure of the energy hypercomplex surfaces differ significantly from the double conical point singularity typical of bound states@xcite . later , by means of a numerical analysis of the experimental data on the 2@xmath2 doublet of resonances with @xmath3 in the energy spectrum of @xmath4be , we showed , in a realistic example , that a true crossing of resonances mixed by a hermitian interaction may be brought about by the variation of only two real linearly independent parameters@xcite . in this paper
we show that the geometric interpretation of the berry phase factor , first given by b. simon@xcite for the adiabatic evolution of closed quantum systems , may be generalized , in a very natural way , to the case of resonant states of open quantum systems .
that is , the adiabatic evolution of resonant states may be interpreted as parallel transport in a complex line bundle defined over the space of parameters with structure group @xmath5 ( the multiplicative group of the non - zero complex numbers ) .
then , the berry phase factors of resonant states arise as the holonomy group elements due to a connection in the bundle such that during the adiabatic evolution the resonant state is parallel transported along the fibre .
the horizontal spaces are perpendicular to the fibre with respect to a generalized inner product of resonant states defined in a rigged hilbert space .
let us consider the time evolution of a quantum system in a state which is a superposition of unstable eigenstates moving in some strong external field of force which changes slowly with time .
in order to have some concrete example in mind , although a very hypothetical one , we may think of an @xmath6 nucleus which has only unstable energy eigenstates moving in the field of forces of a double magic nucleus , like @xmath7 in a peripheral collision in which the distance between the two nuclei is never smaller than the sum of the nuclear radii . in a semiclassical treatment of the collision , when the centers of the nuclei move along classical trajectories , the parameters in the nucleus - nucleus interaction change with time@xcite .
the evolution of the system under the influence of the external perturbation is governed by the time dependent schrdinger equation @xmath8 the hamiltonian @xmath9 is the sum of the time - independent hamiltonian @xmath10 describing the evolution of the unperturbed system plus a perturbation term @xmath11 which is a function of a number @xmath12 of external " parameters @xmath13 which may change with time , @xmath14 the energy eigenfunctions of the unperturbed hamiltonian are the solutions of the equation , @xmath15 the wave functions @xmath16 satisfy the boundary conditions appropriate to a decaying state .
assuming that , in the absence of perturbation , the unstable system decays spontaneously in two stable fragments , the unperturbed energy eigenfunctions may be written as cluster model wave functions@xcite @xmath17 where @xmath18 and @xmath19 are the wave functions of the clusters a and b , @xmath20 is the radial part of the wave function of the relative motion of the two clusters , @xmath21 is a spherical harmonic and @xmath22 is the antisymmetrizer . in our example , @xmath23 would be the eigenfunction of a state of @xmath4be which decays spontaneously in two @xmath24he clusters . in this case
@xmath20 is a gamow function .
gamow functions are the eigenfunctions of the time independent schrdinger equation which vanish at the origen , and behave as purely outgoing waves for large values of the relative distance @xmath25 , @xmath26 and @xmath27 = 0,\ ] ] where the function @xmath28 is an outgoing spherical wave of complex wave number @xmath29 and angular momentum @xmath30@xcite .
the boundary condition ( [ seis ] ) is not self - adjoint , in consequence , the energy eigenvalues are complex , with @xmath31 and @xmath32 hence , gamow functions are a generalization of bound state eigenfunctions in that they belong to complex wave numbers , @xmath33 , with @xmath34 , instead of purely imaginary ones .
this generalization leads out of the hilbert space based quantum mechanics .
therefore , the quantum mechanical rules for normalization , orthogonality and completeness in their usual form , do not apply . nevertheless , bound and resonant states form a bi - orthonormal set with their adjoints , which may be extended by a continuum of suitably chosen scattering states of complex wave number to form a complete set in terms of which any square integrable function may be expanded @xcite .
the symmetry properties of the schrdinger equation and the boundary conditions under the operations of complex conjugation and reversal of time suggest a definition for the adjoint @xmath35 of the gamow function @xmath36@xcite , @xmath37 now , if the adjoint @xmath35 is identified with the bra - eigenfunction , the quantum mechanical inner product , or bra - c - ket rule , may be generalized to include gamow eigenfunctions . with this prescription ,
matrix elements of operators and inner products are computed in configuration space representation as integrals over the radial variable @xmath25 . since the gamow function @xmath36 and its adjoint @xmath35
oscillate between envelopes that grow exponentially with @xmath25 , the integrals over @xmath25 must be properly defined .
this may be done , either by analytic continuation in the complex @xmath38-plane from above @xcite or by means of an abel regulator@xcite with a suitable convergence factor and a limiting procedure@xcite .
both procedures give the same result@xcite .
in this paper , we will adopt the second definition .
then , it may be shown that @xmath39 i.e , the gamow eigenfunctions are orthonormalized in a generalized sense .
it may also be shown@xcite that , for any two square integrable functions , @xmath40 and @xmath41 , the following relation holds @xmath42\times \left[\int_{0}^{\infty } u^*_{nl}(-k^*_n , r')g(r')dr ' \right]\right\ } + \\ \int_{c } dk\left[\int^{\infty}_{0 } f^*(r)\phi_l(k , r)dr\right]\left[\int_{0}^{\infty } \phi^*_l(-k^*,r')g(r')dr ' \right ] .
\end{array}\ ] ] in this expression , the functions @xmath43 are bound states or gamow state eigenfuctions belonging to real negative or complex eigenvalues @xmath44 , the functions @xmath45 are scattering partial wave functions of angular momentum @xmath30 and complex wave number @xmath38 .
the integration contour @xmath46 , in the wave number plane @xmath38 , starts from the origin as a straight line with slope -1 , it goes down to @xmath47 and then , it continues as a straight line parallel to the real axis@xcite .
finally , the square brackets around the integrals mean that , when necessary , the integrals are defined by means of a gaussian regulator and a limit as in ( [ ocho ] ) or by analytical continuation from above as in mondragn and hernndez@xcite . since @xmath40 and @xmath41 are arbitrary functions , we are justified in writing the expansion @xmath48 the index @xmath49 runs over bound and resonant states .
the expansion coefficients are given by @xmath50 @xmath51 once the validity of the quantum mechanical inner product has been generalized to apply to bound and resonant eigenfunctions of the relative motion of the two clusters , the generalization of ( [ nueve ] ) and ( [ diez ] ) to expansions of many body wave functions in terms of cluster model bound and resonant eigenfunctions is fairly straightforward@xcite .
we may , now , go back to our problem , namely the time evolution of the resonant states of a many body system moving in a slowly time - dependent external field of force .
since we are interested in the time evolution of a state @xmath52 which is a superposition of unstable states , we make an expansion of the wave function @xmath53 in terms of bound and resonant states of @xmath54 @xmath55 in general , the index @xmath49 runs over bound and resonant states . in our example
, @xmath56 would be the complex energy eigenfunctions of the @xmath4be nucleus which decays spontaneously in two @xmath24he clusters .
the scattering states @xmath57 of complex wave number @xmath38 and the integration contour @xmath46 are defined in the previous subsection .
substitution of ( [ trece ] ) in ( [ uno ] ) gives the set of coupled equations @xmath58 and a similar expression for @xmath59 we will use the notation @xmath60 for the gamow function , and @xmath61 for its adjoint .
hence , the matrix element of the perturbation term @xmath62 taken between bound or resonant states of the unperturbed system is given by @xmath63 the integral over the radial variable @xmath25 is defined as in ( [ ocho ] ) .
when the interactions are time reversal invariant , the dual of the complex gamow function @xmath20 is the same function@xcite .
but , when the interactions are not time reversal invariant , the gamow function and its dual are not the same function .
therefore , when the forces acting on the system are time reversal invariant , the complex matrix @xmath64 , with matrix elements @xmath65 is symmetric but non - hermitian .
when the forces acting on the system are not time reversal invariant , @xmath64 is , in general , complex , non - symmetric and non - hermitian . the contribution of the non - resonant background integral over the continuum of scattering wave functions to ( [ trece ] ) and ( [ catorce ] ) will not be relevant to the following discussion .
therefore , to ease the notation , we will disregard the contribution of the background of scattering functions . with this truncation , @xmath66.\ ] ] then , the set of coupled equations([catorce ] ) reduces to @xmath67 where @xmath68 is the time derivative of @xmath69
the time evolution of the quantum system is governed by the hamiltonian @xmath9 , or the matrix @xmath64 , which are functions of @xmath12 real , linearly independent parameters .
therefore , we may consider the matrix @xmath64 embedded in a population of @xmath64 matrices smoothly parametrized by @xmath12 external parameters which take values in some domain @xmath70 of a manifold or parameter space .
each point in @xmath70 represents an @xmath64 matrix .
when the external parameters change with time , the system traces a path @xmath46 in parameter space . in the following , we will study the behaviour of the system in the time interval [ 0,t ] , assuming that at the initial time , @xmath71 , it is in an eigenstate of @xmath64 , at @xmath72 the parameters @xmath73 have returned to their initial values , and the adiabatic theorem holds .
then , the unstable system traces a closed path @xmath46 in parameter space while it remains in an eigenstate of @xmath74 at all times . in the absence of symmetry
, @xmath64 has no repeated eigenvalues at almost all points in the domain @xmath70 .
the set of points in @xmath70 where @xmath64 has one twofold repeated eigenvalue is a subdomain @xmath75 . if @xmath70 has @xmath12 dimensions , and @xmath76 has @xmath77 dimensions , then @xmath78 , @xmath79 is the codimension of the twofold degeneracy .
similar relations hold in the case of an m - fold degeneracy .
it will be assumed that the complex non - hermitian matrix @xmath64 has no repeated eigenvalues at all points on the path @xmath46 . in consequence , at all points on @xmath46 , it may be brought to diagonal form by means of a similarity transformation @xmath80 where @xmath81 is the diagonal matrix of the complex energy eigenvalues .
the columns in the matrix @xmath82 are the instantaneous right eigenvectors of @xmath83 . in an obvious notation @xmath84 which satisfy the eigenvalue equation @xmath85
the rows in @xmath86 are the corresponding left eigenvectors of @xmath83 , properly normalized , @xmath87 the adiabatic basis , @xmath88 of instantaneous bound and resonant energy eigenstates of the complete hamiltonian @xmath9 , is obtained from the set of unperturbed bound and resonant states with the help of the matrix @xmath89 @xmath90 and their adjoints ( duals ) are given by @xmath91 it follows that the instantaneous bound and resonant energy eigenstates of @xmath92 satisfy an orthogonality relation similar to ( [ ocho ] ) , which they inherit from the unperturbed complex energy eigenfunctions @xmath93 if we multiply both sides of eq.([veintitres ] ) by @xmath86 , we obtain @xmath94 substitution of this expresion in ( [ trece ] ) , gives the expansion of @xmath52 in instantaneous energy eigenfunctions of @xmath64 , @xmath95 where @xmath96 in ( [ veintisiete ] ) we have kept only the summation over bound and resonant states and , as in the previous section , we have disregarded the contribution from the integral over the continuum of scattering functions of complex wave number .
similarly , the expansion of @xmath97 becomes @xmath98 substitution of ( [ veintisiete ] ) and ( [ veintinueve ] ) in the time dependent schrdinger equation gives the set of coupled equations @xmath99 where , @xmath100 is the gradient of @xmath101 in parameter space , @xmath102 it will be assumed that the non - adiabatic transition amplitudes are very small @xmath103 then , we can make the approximation @xmath104 integrating both sides , we get @xmath105 exp [ { i\gamma _ s}\hat ] a_s(0),\ ] ] the first factor is the complex dynamical phase , whereas the second one is the complex berry phase given by @xmath106 direct evaluation of @xmath107 requires a locally single valued basis for @xmath108 and can be awkward .
such difficulties are avoided by transforming the path integral into a surface integral with the help of stokes theorem @xmath109 where @xmath110 is any surface in parameter space whose boundary is the curve @xmath111 , and @xmath112 means wedge product . this expression may be written in a more convenient form by means of the identity @xmath113 then , @xmath114 provided the surface @xmath115 does not cross any point in @xmath70 where the denominator vanishes .
since the dependence on @xmath116 has been eliminated in ( [ treintaiocho ] ) , phase relations between eigenstates with different parameters are no longer important , any complete set of instantaneous eigenfunctions of the time dependent hamiltonian @xmath9 may be used to evaluate the integral in ( [ treintaiocho ] ) .
explicit expressions for the berry phase in terms of our parametrization of the interaction hamiltonian may easily be obtained . since the unperturbed hamiltonian and its bound and resonant states are independent of time , the time dependence of the bound and resonant instantaneous energy eigenstates of the perturbed system is entirely contained in the matrix @xmath89 , which is a function of @xmath117 through the time dependence of the external parameters .
therefore , from eqs.([veintitres ] ) , ( [ veinticuatro ] ) and ( [ veinticinco ] ) we get @xmath118 _ { ss}\cdot d\vec \re.\ ] ] the columns in @xmath119 are the gradients of the instantaneous eigenvectors of @xmath74 .
therefore , @xmath120 may also be written as @xmath121 ) we may derive a sum rule for the geometric phases of the interfering resonant states . writing the integrand in the right hand side of ( [ treintaiseis ] ) in terms of @xmath89 and @xmath122 , and taking the sum over @xmath123 , we get @xmath124\cdot d\sigma\ ] ] but @xmath125 = c_1(n),\ ] ] is the first chern class of a complex vector bundle defined by the matrix @xmath64 over the parameter space m. its integral is the first chern number@xcite @xmath126\cdot d\vec \sigma = c_1.\ ] ] hence , @xmath127 \cdot d \vec \re
. = -2\pi c_1,\ ] ] is a topological invariant .
since 1983 , b. simon@xcite pointed out that for hermitian closed quantum systems , the adiabatic evolution can be interpreted as a parallel translation in a hermitian line bundle and the berry phase factor is the holonomy in such a bundle . in this section , it will be shown that this geometric interpretation may be generalized in a very natural way to the berry phase factor of resonant states of open quantum systems .
the resonant or gamow functions @xmath128 of the unperturbed system are eigenfunctions of the self - adjoint unperturbed hamiltonian @xmath10 , which satisfy purely outgoing wave boundary conditions for large values of the separation distance @xmath25 of the decay fragments . since the boundary condition is not self - adjoint , the corresponding energy eigenvalues @xmath129 , are complex , with @xmath1300 and @xmath1310 . in configuration space representation , the gamow functions , as functions of @xmath25 , behave as outgoing waves which oscillate between envelopes that increase exponentially . in consequence ,
gamow functions are not square integrable and can not be characterized as elements of a hilbert space .
gamow functions are usually associated with the resonance poles of the scattering matrix and the resolvent operator of the time independent schrdinger equation which lie in the lower half - plane of the unphysical sheet of the riemann energy surface@xcite . in order to give a proper mathematical characterization of gamow functions as elements of a space
, one has to specify a rigged hilbert space in which the gamow states @xmath132 are defined as generalized eigenvectors of the hamiltonian @xmath10 with generalized complex eigenvalues @xmath129 . following bohm and gadella@xcite , we will associate to the resonant poles of the resolvent operator of the schrdinger equation of the unperturbed system , a rigged hilbert space @xmath133 in this expression , @xmath134 is the space of well - behaved functions of the position coordinates ( @xmath135 ) which are hardy class 2 functions of the complex energy e from above , @xmath136 is the hilbert space of square integrable functions and @xmath137 is the space of continuous antilinear functionals on @xmath138 . in this way , the gamow states of the unperturbed system are defined as continuous antilinear functionals on @xmath134 , that is , @xmath139 the adiabatically evolving gamow functions @xmath140 of the complete hamiltonian @xmath9 , introduced in section 3 , eq.([veintitres ] ) , are linear combinations of the gamow functions @xmath141 of @xmath10 , @xmath142 ) > = \sum _ m|\varphi _ m(\xi_i,\eta_j , r ) >
{ \bf k}_{ms}[\vec \re(t)],\ ] ] where @xmath143 $ ] is the matrix which diagonalizes the complex non - hermitian matrix @xmath144 $ ] . therefore , the adiabatically evolving gamow functions are also elements of @xmath145 , @xmath146 ) >
\epsilon \phi ^*_{+}.\ ] ] the dual of the gamow eigenvector @xmath147>$ ] is the eigenvector @xmath148|= \sum_m({\bf k}^{-1})_{sm}<\varphi_m(\xi_i,\eta_j)|$ ] defined in eq.([veinticuatro ] ) , corresponding to the same complex eigenvalue @xmath149 $ ] .
since , by assumption @xmath150 $ ] has no repeated eigenvalues for @xmath151 , the adiabatically evolving gamow eigenvectors and their duals satisfy the orthogonality relation @xmath152|\hat \varphi_{s ' } [ \vec \re ] > = 0,\hspace{1.5cm}\hat{\cal e}_s\neq \hat { \cal e}_s'\ ] ] inherited from the gamow eigenfuctions @xmath153 of @xmath10 , and they may be normalized to one @xmath154|\hat \varphi_s [ \vec \re ] > = 1\ ] ] having characterized the adiabatically evolving gamow functions as elements of a rigged hilbert space we may turn to the question of the geometric interpretation of the berry phase factors of resonant states . let us suppose that , as time varies , the self - adjoint hamiltonian @xmath155 $ ] and its instantaneous energy eigenstates @xmath156 make a cyclic excursion in a closed circuit c in parameter space .
we assumed that the time dependence of the hamiltonian justifies the adiabatic approximation .
we will further assume that , if @xmath157 is any point on c , the complex matrix @xmath144 $ ] has no repeated eigenvalues in an open neighbourhood of @xmath158 c , that is , the hamiltonian @xmath92 has no repeated instantaneous complex eigenvalues when @xmath158 c , and that both the generalized eigenvectors @xmath159 and the eigenvalues @xmath160 $ ] are smooth functions of @xmath158c .
moreover , we will also assume that the evolution of each state is such that there are no level crossings along c. to avoid a clumsy notation in the rest of this section we will not write the position coordinates ( @xmath135 ) .
let us call @xmath161 the complex line bundle defined by the gamow eigenfunction @xmath162>$ ] over the parameter space m of the adiabatically evolving system @xmath163 > ) | { h } [ \vec \re]|\hat \varphi _
s[\vec \re ] > = { \cal e}_s [ \vec \re ] |\hat \varphi _ s[\vec \re ] > , \vec
\re \epsilon m \right\}\ ] ] its fibre is a complex , one - dimensional linear space @xmath164>{\bf | } | \hat \psi _
s[\vec \re ] > = e^{i\alpha } \epsilon \phi^*_{+ } , \alpha
[ \vec \re ] \epsilon { \bf c^ * } , \vec \re \epsilon m \right \}\ ] ] where @xmath165 is the space of antilinear functionals defined over the space @xmath134 of well - behaved functions of hardy class 2 from above and @xmath166 $ ] is a complex function of @xmath167 . under the assumptions made above , the quantum number @xmath123 , labelling the eigenvalue @xmath168 $ ] and the eigenstate @xmath169>$ ] of the instantaneous hamiltonian , is an adiabatic invariant . hence , a system prepared in a state @xmath170>$ ] such that @xmath171 > = c_s [ \vec \re(t_o)]|\hat \varphi
_ s[\vec \re(t_o)]>\ ] ] will evolve with @xmath172 $ ] and be in a state @xmath173 > = c_s[\vec \re(t)]|\hat \varphi _ s[\vec \re(t)]>\ ] ] at @xmath117 .
now , consider the decomposition of a tangent vector in vertical and horizontal parts @xmath174 the horizontal part , orthogonal to the fibre is @xmath175 and the vertical part , along the fibre is @xmath176|\frac{d}{dt}| \hat \psi _ s(t)>|\hat \varphi _ s[\vec \re]>\ ] ] if @xmath177 is an evolution state in an adiabatic change , @xmath178}{dt } + < \ hat \varphi _
s[\vec \re]|\frac{d}{dt}\hat \varphi _ s [ \vec
\re(t)]>c_s[\vec \re]\right)|\hat \varphi _ s[\vec \re]>\ ] ] the condition for parallel translation along the curve c is @xmath179|\frac{d } { dt}\hat \psi _ s[\vec \re ] > = 0,\hspace{1.5 cm } \vec \re \epsilon c\ ] ] which gives a one - form equation @xmath180 + < \hat \varphi _ s[\vec \re ] its solution gives the berry phase factor of the resonant state @xmath181 = exp [ { i\gamma _ s[\vec \re]}].\ ] ] in the case of a cyclic evolution , such that @xmath182 and c is a closed curve , the complex phase @xmath120 may be expressed as @xmath183 ) , may be written as @xmath184 now , making use of the normalization condition , eq.([veinticinco ] ) , this expressions may be written as @xmath185| \left\{|\hat \varphi _ s[\vec \re(t)]>\frac{dc_s(t)}{dt } + \frac{|d\varphi _
s[\vec \re(t)]>}{dt}c_s(t)\right\ } = 0,\ ] ] where @xmath186 is given by @xmath187 multiplying both sides of ( [ sesentaicuatro ] ) by the dynamic phase factor , the condition for adiabatic evolution of a resonant state takes the form @xmath188 which is the condition for parallel translation of the resonant state @xmath189 >
= \hat c_s[\vec \re(t)]|\hat \varphi _ s[\vec \re(t)]>$ ] along the curve c. it follows that , except for the effect of the dynamical phase factor , the condition of adiabatic evolution of a resonant state is equivalent to the condition of parallel translation on the complex line bundle @xmath161 .
the geometric phase factors @xmath190 $ ] , occurring in the adiabatic evolution , are holonomy group elements of the complex line bundle @xmath161 @xcite .
as is apparent from ( [ treintaiocho ] ) , non - trivial phase factors of the energy eigenvectors or eigenfunctions are related to the occurrence of accidental degeneracies of the corresponding eigenvalues . in the absence of symmetry
, degeneracies are called accidental for lack of an obvious reason to explain why two energy eigenvalues @xmath191 and @xmath129 of @xmath64 should coincide . however ,
if the matrix @xmath64 is embedded in a population of complex non - hermitian matrices @xmath192 $ ] smoothly parametrized by @xmath12 external parameters @xmath73 , degeneracy in the absence of symmetry is a geometric property of the hypersurfaces representing the real or complex eigenvalues of @xmath64 in a @xmath193-dimensional euclidean space with cartesian coordinates @xmath194 .
the energy denominators in ( [ treintaiocho ] ) show that when the circuit @xmath46 lies close to a subdomain @xmath195 , in parameter space , where the matrix @xmath64 has an m - fold repeated eigenvalue , and the state @xmath196 is involved in this degeneracy , the berry phase @xmath197 is determined by the geometry of the energy hypersurfaces close to the crossing of eigenvalues and the other ( m-1 ) states involved in the degeneracy .
it is in this connection of the berry phase with the accidental degeneracy of energy eigenvalues that the non - hermiticity of the matrix @xmath64 plays an important role .
in contrast with the case of hermitian matrices , square , complex , non - hermitian matrices with repeated eigenvalues can not always be brought to diagonal form by a similarity transformation .
however , any n - dimensional , square complex matrix @xmath64 may always be brought to a jordan canonical form @xmath198 by means of a similarity transformation .
if @xmath64 has @xmath199 @xmath200 different eigenvalues , @xmath201 , with multiplicities @xmath202 , the jordan canonical form @xmath203 is the direct sum of @xmath199 square jordan blocks @xmath204 .
each jordan block @xmath204 is the sum of a diagonal matrix @xmath205 , and a nilpotent matrix @xmath206 .
corresponding to each jordan block , there is a cycle of lenght @xmath207 of generalized eigenvectors . when the lenght @xmath208 of the cycle is @xmath2092 , the codimension of the accidental degeneracy , the geometry of the energy hypersurfaces at the crossing and the properties of the generalized eigenvectors involved in the degeneracy of resonant states differ substantially from those of bound states(@xcite ) . rather than trying to develop a theory of the most general case , in the following we will examine the simplest possible case , namely the accidental degeneracy of two resonances and the topology of the energy surfaces close to a crossing of two resonances in parameter space .
let us consider a system with two resonant states strongly mixed by a hermitian interaction , all other bound or resonant eigenstates being non - degenerate .
we may suppose that we already know the correct eigenvectors of @xmath64 for all the real and complex eigenenergies @xmath191 , except for the two the crossing of which we want to investigate .
using for this two states two vectors which are not eigenvectors but which are orthogonal to each other and to all other eigenvectors , we obtain a complete basis to represent @xmath64 . in this basis
, @xmath64 will be diagonal except for the elements @xmath210 and @xmath211 .
the diagonal elements @xmath212 and @xmath213 will , in general , be non - vanishing and different from each other . there is no loss of generality in this supposition , since any complex matrix @xmath64 may be brought to a jordan canonical form by means of a similarity transformation .
when the eigenvalues are equal , @xmath214 is either diagonal or equivalent to a jordan block of rank two .
hence , we need consider only the conditions for degeneracy of the submatrix @xmath214 . the matrix @xmath214 may be written in terms of the pauli matrix valued vector @xmath215 and the @xmath216 unit matrix as @xmath217 where @xmath218 and @xmath219 are real vectors with cartesian components @xmath220and @xmath221 .
when the forces acting on the system are time reversal invariant , @xmath222 and @xmath223 vanish . in the absence of more specific information about the external parameters @xmath224
, we will parametrize @xmath214 in terms of @xmath218 and @xmath225 . then ,
@xmath226 from ( [ sesentaisiete ] ) , the eigenvalues of @xmath214 are given by @xmath227 where @xmath228 then , @xmath229 and @xmath230 coincide when @xmath231 vanishes .
since , real and imaginary parts of @xmath232 should vanish , the condition for accidental degeneracy of the two interfering resonances may be written as @xmath233 and @xmath234 these equations admit two kinds of solutions corresponding to @xmath235 being or not being diagonal at the degeneracy : \i ) when both @xmath236 and @xmath237 vanish , eqs . ( [ setentaiuno ] ) and ( [ setentaidos ] )
define a point in parameter space , @xmath83 is diagonal at the degeneracy and the submatrix @xmath214 has two cycles of eigenvectors of lenght one each .
the vanishing of @xmath238 implies that @xmath239 also vanish@xcite , therefore the two complex eigenvalues of @xmath64 which become degenerate fuse into one real positive energy eigenvalue embedded in the continuum . since in this case
all the cartesian components of @xmath236 and @xmath240 should vanish , the minimum number of linearly independent , real , external parameters that should be varied to produce a degeneracy ( codimension ) of two resonances to form a bound state embedded in the continuum is four or six depending on the quantum system being or not being time reversal invariant .
\ii ) in the second case , when the degeneracy conditions ( [ setentaiuno ] ) and ( [ setentaidos ] ) are satisfied for non - vanishing @xmath236 and @xmath240 , these equations define a circle in parameter space . in this case
@xmath241 is not diagonal at the degeneracy , it is equivalent to a jordan block of rank two and has one cycle of generalized eigenvectors of lenght two . in this case , the two complex eigenvalues become one two - fold repeated eigenvalue @xmath242 . since , the two linearly independent conditions ( [ setentaiuno ] ) and ( [ setentaidos ] ) should be satisfied for non - vanishing values of @xmath243 and @xmath244 , at least two real , linearly independent parameters should be varied to produce a rank two degeneracy of resonances .
hence , the codimension of a degeneracy of resonances of second rank is two , independently of the time reversal invariance character of the interactions .
let us consider now the computation of the berry phase in the simplest situation when the degeneracy involves only two resonant states .
the matrix @xmath136 that mixes the two interfering resonances off degeneracy is @xmath245 this matrix has two right and two left eigenvectors and may be diagonalized by a similarity transformation @xmath246 where @xmath247 @xmath248 and @xmath249 are short hand for @xmath250 and @xmath251 respectively.the matrices @xmath86 and @xmath252 are readily obtained from ( [ setenta ] ) and ( [ setentaicinco ] ) . if we assume that the interfering resonances are mixed by a hermitian interaction , @xmath225 will be kept fixed
. then a straightforward calculation gives @xmath253 and @xmath254 these expressions are very similar to the well known results obtained for the geometric phase of bound states@xcite .
an obvious difference is that the geometric phase of resonant states is complex since @xmath231 and @xmath255 are complex functions of the parameters @xmath218 and @xmath219 .
there is another important but less apparent difference : in the case of an accidental degeneracy of resonances @xmath256 the denominator in the right hand side of ( [ setentaiseis ] ) and ( [ setentaisiete ] ) vanishes on the continuous line of singularities defined by eqs.([setentaiuno ] ) and ( [ setentaidos ] ) , which will be called the diabolical circle , and not at one isolated point as is the case for bound states .
it follows that two kinds of non - trivial , topologically inequivalent closed paths are possible .
first , those paths which surround the diabolical circle but are not linked to it .
second , the closed paths which are linked to the diabolical circle .
paths of the first kind are clearly analogous to the non - trivial paths that go around the diabolical point of bound state degeneracies while paths of the second kind have no analogue in accidental degeneracies of bound states . for paths of the first kind
it is always possible to find a surface @xmath110 which spans the closed path @xmath111 and does not cross the diabolical circle .
then , stokes theorem applied to ( [ setentaiseis ] ) and ( [ setentaisiete ] ) gives @xmath257^{\frac 32}},\ ] ] where @xmath258 since @xmath259 changes into @xmath260 when @xmath261 and @xmath262 are exchanged and the sign of @xmath263 is changed , the normals for @xmath262 and @xmath264 should be oppositely oriented .
if we say that @xmath219 points upwards , then @xmath264 is above and @xmath262 is below @xmath265 this is the same result as would have been obtained from the general expression ( [ treintaiocho ] ) and our parametrizacion of the perturbation term in the hamiltonian ; no summation over intermediate states occurs in ( [ setentaiocho ] ) since in the simple case of only two interfering resonant states , the summation in ( [ treintaiocho ] ) has only one term . adding @xmath260 and @xmath259 , we get @xmath266^{\frac 32}},\ ] ] where @xmath110 is a sphere with the diabolical circle completely contained in its interior .
the integral is easily computed when @xmath267 , the result is @xmath268 in agreement with our identification of the integral in ( [ cuarentaicuatro ] ) as the first chern number of the complex line bundle defined by the eigenvector of the non - hermitian matrix @xmath235 .
it is now easy to show that the resonance degeneracy produces a continuous distribution of singularities on the diabolical circle .
the surface integral in ( [ setentainueve ] ) may be written as a volume integral by means of gauss theorem .
then , @xmath269 ^{\frac 32}}\right ) dv,\ ] ] where @xmath270 is the volume inside @xmath115 and bounded by it .
the term in round brackets under the integration sign vanishes when @xmath271 therefore , the non - vanishing value of @xmath272 implies the occurrence of @xmath273-function singularities of the integrand on those points where @xmath274 vanishes .
hence , @xmath275 ^{\frac 32}}\right ] = - \frac{\delta \left ( r-\frac 12\gamma \right ) } { r^2 } \delta \left(\cos \theta \right),\ ] ] the factor @xmath276 multiplying the delta function is needed to reproduce the value @xmath277 of @xmath272 .
let us turn to the case of closed paths of the second kind , that is , those paths which are linked to the diabolical circle .
in this case there is no surface @xmath278 which spans the closed path @xmath279 without crossing the diabolical circle .
therefore , it is not possible to use stokes theorem to convert the path integral into a surface integral .
however , we may still compute the geometric phase from the path integral . to this end , we change from cartesian coordinates @xmath280 with @xmath281 parallel to @xmath225 , to spherical coordinates in parameter space .
then , eqs .
( [ setentaiseis ] ) and ( [ setentaisiete ] ) take the form @xmath282 the path @xmath283 is specified when @xmath284 and @xmath285 are given as functions of @xmath286 . from ( [ ochentaitres ] )
, it follows that , for closed paths which are linked to the diabolical circle , @xmath287 since , in this case , the angle @xmath288 starts out at some value @xmath289 and , as the system traces the path @xmath279 , it oscillates between a minimum and a maximum values and finally ends at the same initial value @xmath289 .
there is no analogue to this case in the geometric phase of bound states .
the purpose of the foregoing has been to discuss the geometric phase acquired by the resonant states when they are adiabatically transported in parameter space around a degeneracy of resonances . as in the case of bound states
, the condition of adiabatic evolution of resonant states may be given a geometric interpretation as parallel translation in a complex line bundle @xmath161 .
the geometric phase factors acquired by the resonant states when adiabatically transported in a closed circuit in parameters space are holonomy group elements of the complex line bundle @xmath161 . in the case of two resonant states mixed by a hermitian interaction we find two kinds of accidental degeneracies which may be characterized by the number and length of the cycles of instantaneous energy eigenfunctions at the degeneracy . in the first case
there are two linearly independent eigenfunctions belonging to the same real positive repeated energy eigenvalue , that is , two cycles of lenght one . in the second case
there is only one resonant eigenstate and one generalized resonant eigenstate belonging to the same degenerate ( repeated ) complex energy eigenvalue , @xmath290 .
one cycle of lenght two . at degeneracy
, the hamiltonian matrix has one jordan block of second rank . in the generic case of a non - time reversal invariant system ,
when the degeneracy is of the second rank , the topology of the energy surfaces is different from that at a crossing of bound states .
the energy surfaces of the two resonant states that become degenerate touch each other at all points in a circle .
close to the crossing , the energy hypersurface has two pieces lying in orthogonal subspaces in parameter space .
the surface representing the real part of the energy has the shape of a hyperbolic cone of circular cross section , or an open sandglass , with its waist at the diabolical circle .
the surface of the imaginary part of the energy is a sphere with the equator at the diabolical circle .
the two surfaces touch each other at all points on the diabolical circle@xcite . in the case of two interfering resonant states , the geometric phase acquired by the resonant states when transported around the diabolical circle in a closed path which is not linked to it ,
may be written as the sum of two terms .
the first term , @xmath292 , is the real geometric phase which a negative energy eigenstate would have acquired when transported around a diabolical point in a closed path in the same parameter space .
the second term is complex , it gives rise to a change of the phase and a dilation of the resonant state eigenfunction .
its imaginary part may be positive or negative , in consequence , it may produce an amplification or a damping of the wave function which may compensate or reinforce the attenuation due to the imaginary part of the dynamical phase factor .
since it is not possible to find a continuous surface @xmath115 which spans the closed path @xmath295 without crossing the diabolical circle , we can not make use of the theorem of stokes to convert the path integral into a surface integral .
however , it may readily be computed as a path integral from the expression which is obtained from ( [ ochentaitres ] ) and ( [ ochentaicuatro ] ) . as in the previous case , @xmath297 is complex and produces changes of phase and dilations in the resonant state wave function .
this case has no analogue in bound states .
the sum of the geometric phases acquired by two interfering resonant states which are transported around a degeneracy in a closed path of the first kind in parameter space is a topological invariant , namely the first chern number@xcite .
its value is the `` magnetic charge '' on the diabolical circle . for paths of the second kind
the sum of the geometric phases vanishes . for reviews on berry s phase see : jackiw , r. _ commun . at .
phys_. * 27 * , 71 , ( 1988 ) ; vinitskii s.j . , derbov v.l . , dubovik v.n
, markovski b.l . and stepanovskii yu .
sov.phys . usp_. * 33 * , 403 , ( 1990 ) : | we study the complex geometric phase acquired by the resonant states of an open quantum system which evolves irreversibly in a slowly time dependent environment . in analogy with the case of bound states ,
the berry phase factors of resonant states are holonomy group elements of a complex line bundle with structure group @xmath0 . in sharp contrast with bound states , accidental degeneracies of resonances produce a continuous closed line of singularities formally equivalent to a continuous distribution of `` magnetic '' charge on a `` diabolical '' circle , in consequence , we find different classes of topologically inequivalent non - trivial closed paths in parameter space . |
_ c - loops _ are loops satisfying the identity @xmath2 as we shall see , they are in a sense dual to moufang loops the most intensively studied variety of loops and they are closely related to steiner triple systems .
they are thus important both algebraically and combinatorially , and they are amenable to analysis by techniques from both fields . but in spite of this , little is known about them .
it is the intention of this paper to remedy this situation by laying a foundation for the systematic study of c - loops .
we assume that the reader is familiar with the reasoning and notational conventions of loop theory , however , we do not hesitate to include loop - theoretical folklore and to point out some of the pitfalls of nonassociativity mostly because we fell into many of them ourselves .
c - loops were named by ferenc fenyves @xcite , who investigated the inclusions between varieties of loops of _ bol - moufang type_. these are varieties of loops defined by a single identity that : ( i ) involves three distinct variables on both sides , ( ii ) contains variables in the same order on both sides , ( iii ) exactly one of the variables appears twice on both sides .
fenyves s program was completed by the authors in @xcite
. there are 60 identities of bol - moufang type , and they happen to define 14 distinct varieties of loops .
figure [ fg : l ] gives the hasse diagram of these varieties , with the largest varieties ( with respect to inclusion ) at the bottom .
a superficial glance at the diagram suggests that c - loops could behave analogously to moufang loops .
this impression is further strengthened by the fact that c - loops are exactly those loops that are both lc - loops and rc - loops ( * ? ? ?
* theorem 4 ) , just as moufang loops are exactly those loops that are both left bol and right bol @xcite .
there are additional analogies , especially between commutative moufang loops and commutative c - loops , as we shall see .
in combinatorics , moufang loops have connections to projective geometry ( moufang planes , moufang polygons , etc . , cf .
@xcite ) , while c - loops have connections to steiner triple systems : consider the complete graph @xmath3 on @xmath4 vertices .
steiner triple system _ is a decomposition of the edges of @xmath3 into disjoint triangles .
it is well known ( cf .
@xcite ) that such a decomposition exists if and only if @xmath5 or @xmath6 ; the case @xmath7 being degenerate .
there is a canonical way of constructing a quasigroup from a steiner triple system .
namely , if @xmath8 is a steiner triple system , we define multiplication on @xmath9 by @xmath10 , and ( for @xmath11 ) by @xmath12 if and only if @xmath13 is a triangle of @xmath14 .
the resulting quasigroup clearly satisfies @xmath15 conversely , any quasigroup satisfying @xmath16 gives rise to a steiner triple system in a canonical way ( cf .
@xcite , @xcite ) .
quasigroups satisfying @xmath16 are therefore called _
steiner quasigroups_. any steiner quasigroup can be made into a loop by introducing a new element @xmath17 and by letting @xmath18 , @xmath19 .
such loops satisfy @xmath20 and are called _
steiner loops_. it is now clear that the steiner quasigroup that gave rise to a steiner loop @xmath21 can be reconstructed from @xmath21 .
steiner loops are therefore in one - to - one correspondence with steiner triple systems , too . intuitively , the reason why c - loops are related to steiner loops is the presence of the term @xmath22 in the defining equation @xmath23 . more formally : [ lm : cs ] every steiner loop is a c - loop .
note that @xmath24 is a part of the definition @xmath25 , and that @xmath26 follows from @xmath25 immediately by commutativity .
thus @xmath27 .
not every c - loop is a steiner loop , as is witnessed by any nonabelian group .
another connection between c - loops and steiner loops becomes apparent upon investigating the nucleus of c - loops .
recall that for a loop @xmath21 , the set @xmath28 for every @xmath29 , @xmath30 is called the _ left nucleus_. similarly , the _ middle nucleus _ @xmath31
consists of all elements @xmath32 satisfying @xmath33 for every @xmath29 , @xmath34 ; and the _ right nucleus _ @xmath35 consists of all elements @xmath32 satisfying @xmath36 for every @xmath29 , @xmath37 .
the _ nucleus _ @xmath38 of @xmath21 is a subgroup of @xmath21 .
there are several equivalent ways in which normality can be defined for loops .
the following definition works best with elementary calculations .
a subloop @xmath39 of a loop @xmath21 is said to be _ normal _ in @xmath21 if @xmath40 , @xmath41 , and @xmath42 for every @xmath43 , @xmath44 . the _ factor loop _
@xmath45 is then defined in the usual way .
a loop @xmath21 with neutral element @xmath17 is a _ left inverse property loop _ if @xmath46 for every @xmath43 , @xmath44 , where @xmath47 is the unique element satisfying @xmath48 .
dually , @xmath21 is a _
right inverse property loop _ if @xmath49 for every @xmath43 , @xmath44 , where @xmath50 is the unique element satisfying @xmath51 .
a loop that has both the left and right inverse property is an _
inverse property loop_. if @xmath32 is such that @xmath52 for every @xmath29 , we have @xmath53 .
therefore , inverse property loops possess _ two - sided inverses _
( i.e. , @xmath54 ) , and it is easy to check that they satisfy the _ antiautomorphic inverse property _ ( i.e. , @xmath55 .
pflugfelder shows @xcite that steiner loops are exactly commutative inverse property loops of exponent 2 .
in fact , steiner loops are exactly inverse property loops of exponent 2 .
this fact belongs to loop - theoretical folklore and is sometimes used as a definition of steiner loops ( cf .
since we did not manage to find a reference for the proof , here it is : [ lm : folklore ] steiner loops are exactly inverse property loops of exponent two .
let @xmath21 be a steiner loop .
since @xmath18 , every element is its own two - sided inverse .
from @xmath56 we see that @xmath21 has the left inverse property . by commutativity
, it has the right inverse property , too .
conversely , let @xmath21 be an inverse property loop of exponent @xmath1 .
let @xmath57 .
then @xmath58 , and similarly , @xmath59 , @xmath60 .
thus @xmath21 is commutative . as @xmath56 by the right inverse property
, @xmath21 is a steiner loop .
also notice that @xmath18 is not necessary in the definition @xmath25 of steiner loop , since @xmath61 .
hence steiner loops are exactly loops satisfying @xmath62 quasigroups satisfying @xmath63 are called _ totally symmetric _ , and thus steiner loops can also be found under the name _ totally symmetric loops _ in the literature .
let us now mention some basic properties of lc - loops and c - loops that we will use without reference throughout the paper .
the first three properties are due to fenyves ( * ? ? ?
* theorem 2 ) .
the fourth property first appeared in @xcite .
[ pr : sum ] let @xmath21 be an lc - loop . then : 1 . @xmath21 is left alternative , 2 .
@xmath21 has the left inverse property , 3 .
@xmath21 is a left nuclear square loop , 4 .
@xmath21 is a middle nuclear square loop .
we will often derive theorems from their one - sided versions .
let @xmath21 be a c - loop .
then : 1 .
@xmath21 is both left alternative and right alternative , 2 .
@xmath21 has the inverse property , 3 .
@xmath21 is a _ nuclear square loop _ , i.e. , @xmath64 belongs to the nucleus of @xmath21 for every @xmath32 .
[ cr : nuc ] the three nuclei of a c - loop coincide .
the three nuclei coincide for any inverse property loop , by ( * ? ? ?
* theorem vii.2.1 ) .
the nucleus @xmath65 of a loop @xmath21 is always a subgroup of @xmath21 , but it is not necessarily a normal subgroup of @xmath21 . even when @xmath21 is an inverse property loop , its nucleus does not have to be normal in @xmath21 .
( see example [ ex : ipnuc ] ) . throughout the paper , if we claim without explanation that a loop with given properties is as small as possible , or that there are @xmath66 such nonisomorphic loops of given order , we rely on the finite model builder mace4 @xcite . [ ex : ipnuc ] the smallest inverse property loop with nucleus that is not normal . @xmath67 check that 1 is in the nucleus and @xmath68 .
but @xmath1 is not in the nucleus , since @xmath69 .
fortunately , all is well for c - loops .
we will use the following notation in the proof of proposition [ pr : n ] .
any element @xmath32 determines two permutations of @xmath21 : the _ left translation _
@xmath70 defined by @xmath71 , and the _ right translation _
@xmath72 defined by @xmath73 .
[ pr : n ] the nucleus of a c - loop is a normal subgroup .
let @xmath65 be the nucleus of a c - loop @xmath21 .
our task is to show that @xmath74 for every @xmath32 , or , equivalently , that @xmath75 for every @xmath32 , @xmath76 .
since the nuclei of a c - loop coincide , it suffices to show @xmath77 , which in the language of translations becomes @xmath78 for every @xmath44 . because squares of elements in a c - loop are in the nucleus and because @xmath79 , the last identity is equivalent to @xmath80 , which is what we prove below .
the following permutations coincide : @xmath81 , @xmath82 ( by the left inverse property and the right alternative property ) , @xmath83 ( since @xmath84 ) , @xmath85 ( since @xmath76 ) , @xmath86 ( by the left inverse property ) .
using similar arguments , we see that @xmath87 .
therefore @xmath88 .
the last translation @xmath89 is equal to @xmath90 , because @xmath91 , and we are done .
[ pr : ln ] let @xmath21 be a c - loop with nucleus @xmath65 . then @xmath92 is a steiner loop .
we have @xmath93 for every @xmath32 .
thus @xmath92 is an inverse property loop of exponent @xmath1 . by lemma [ lm : folklore ]
, @xmath92 is a steiner loop .
the following lemma will be useful in the next section .
[ lm : n2 ] there is no c - loop with nucleus of index @xmath1 .
assume , for a contradiction , that @xmath21 is a c - loop with nucleus @xmath65 of index @xmath1 .
let @xmath65 , @xmath94 be the two cosets of @xmath92 . we show that @xmath95 .
since the three nuclei of @xmath21 coincide , it suffices to show that @xmath96 for every @xmath97 , @xmath98 .
in fact , it suffices to prove this for all elements @xmath97 , @xmath99 , since all other elements are nuclear .
let us write @xmath100 , @xmath101 , for some @xmath102 , @xmath103 .
since @xmath102 , @xmath104 , @xmath64 and @xmath105 are all nuclear , and since @xmath106 , we have @xmath107 .
we now construct the @xmath108 smallest nonassociative c - loops ( 1 of order 10 , 1 of order 12 , and 2 of order 14 ) .
three of these loops are well - known steiner loops .
the remaining c - loop of order 12 belongs to an infinite family of nonassociative noncommutative c - loops constructed here for the first time . although the four loops are constructed by hand , we do not have sufficiently strong theoretical tools to show that no other nonassociative c - loops of order less than @xmath109 exist .
this is easily verified by mace4 , though .
recall that steiner loops of order @xmath1 , @xmath108 and @xmath110 are elementary abelian @xmath1-groups @xcite .
[ pr : orders ] let @xmath21 be a nonassociative c - loop of order @xmath4 with nucleus @xmath65 of order @xmath66
. then 1 .
@xmath111 or @xmath112 , 2 .
@xmath4 is even , 3 . if @xmath113 for some prime @xmath114 and positive integer @xmath115 , then @xmath116 and @xmath117 .
moreover , there is a nonassociative non - steiner c - loop of order @xmath118 for every @xmath117 .
part ( i ) follows from proposition [ pr : ln ] and from the already mentioned fact that steiner quasigroups of order @xmath119 exist if and only if @xmath120 or @xmath121 .
part ( ii ) follows immediately from part ( i ) .
assume that @xmath113 , @xmath114 a prime . by ( ii ) ,
when @xmath122 , @xmath21 must be a group , since there is no nonassociative loop of order less than @xmath123 .
assume that @xmath124 .
if @xmath125 , proposition [ pr : ln ] implies that @xmath21 is a steiner loop of order @xmath110 , thus the elementary abelian @xmath1-group of order @xmath110 .
if @xmath126 , we reach a contradiction by lemma [ lm : n2 ] .
we were not able to find a one - line argument that shows that there is no nonassociative c - loop of order @xmath110 with nucleus of size @xmath1 .
it can be checked tediously by hand .
example [ ex:16 ] gives a nonassociative non - steiner c - loop of order @xmath127 .
direct products of this loop with @xmath1-groups provide all needed examples .
let @xmath21 , @xmath4 , @xmath66 be as assumed in proposition [ pr : orders ] .
the only admissible values of @xmath128 with @xmath129 are then @xmath130 , @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 and @xmath136 .
lemma [ lm : n2 ] further reduces the possibilities to @xmath131 , @xmath133 and @xmath135 .
as we shall see , there is at least one nonassociative c - loop for each of these parameters .
the smallest nonassociative commutative inverse property loop is of order @xmath137 , and it is unique .
its multiplication table is in example [ ex:10 ] .
we can see immediately that this loop has exponent @xmath1 .
it is therefore a steiner loop . by lemma [ lm : cs ]
, it is a nonassociative c - loop , hence the smallest nonassociative c - loop .
[ ex:10 ] the smallest nonassociative c - loop .
we now construct an infinite family of nonassociative noncommutative c - loops whose smallest member is the smallest nonassociative noncommutative c - loop . the construction is best approached via extensions of loops .
our notation is based on @xcite .
let @xmath139 be a multiplicative group with neutral element @xmath140 , and @xmath141 an abelian group written additively with neutral element @xmath142 .
any map @xmath143 satisfying @xmath144 for every @xmath145 is called a _
factor set_. when @xmath143 is a factor set , we can define multiplication on @xmath146 by @xmath147 the resulting groupoid is clearly a loop with neutral element @xmath148
. it will be denoted by @xmath149 .
additional properties of @xmath149 can be enforced by additional requirements on @xmath150 .
let @xmath143 be a factor set .
then @xmath149 is a c - loop if and only if @xmath151 for every @xmath152 , @xmath153 , @xmath154 .
the loop @xmath149 is a c - loop if and only if @xmath155 holds for every @xmath152 , @xmath153 , @xmath154 and every @xmath97 , @xmath156 , @xmath157 .
straightforward calculation with @xmath158 shows that this happens if and only if @xmath159 is satisfied .
we call a factor set @xmath150 satisfying @xmath159 a _ c - factor set_. when @xmath139 is an elementary abelian @xmath1-group , the equation @xmath159 reduces to @xmath160 we now use a particular c - factor set to construct the above - mentioned family of c - loops .
[ pr : constr ] let @xmath161 be an integer .
let @xmath141 be an abelian group of order @xmath4 , and @xmath162 an element of order bigger than @xmath1 .
let @xmath163 be the klein group with neutral element @xmath140 .
define @xmath143 by @xmath164 then @xmath149 is a non - flexible @xmath165hence nonassociative@xmath166 noncommutative c - loop with nucleus @xmath167 .
the map @xmath150 is clearly a factor set .
it can be depicted as follows : @xmath168 to show that @xmath169 is a c - loop , we verify @xmath170 .
since @xmath150 is a factor set , there is nothing to prove when @xmath171 .
assume that @xmath172 .
then @xmath170 becomes @xmath173 , and both sides of this equation are equal to @xmath142 , no matter what @xmath115 , @xmath145 are .
assume that @xmath174 .
then @xmath170 becomes @xmath175 , and both sides of this equation are again equal to @xmath142 .
assume that @xmath176 .
then @xmath170 becomes @xmath177 , and both sides of this equation are equal to @xmath178 .
since @xmath179 , the c - loop @xmath180 is not commutative . by corollary [ cr : nuc ]
, the three nuclei of @xmath180 coincide and will be denoted by @xmath65 .
let @xmath181 .
since @xmath182 , we have @xmath183 .
this shows that : ( i ) @xmath180 is not flexible , ( ii ) @xmath184 , @xmath185 .
similarly , @xmath186 shows that @xmath187 .
finally , for @xmath152 , @xmath188 and @xmath156 , @xmath157 we have @xmath189 , and @xmath190 follows .
[ cr : nnnuc ] for any integer @xmath191 there is a nonassociative noncommutative @xmath180-loop with nucleus of size @xmath4 .
it remains to show that there is a nonassociative noncommutative @xmath180-loop with nucleus of size @xmath1 .
consider the octonion loop @xmath192 of order @xmath127 .
this loop is moufang .
recall that extra loops are precisely moufang loops with squares in the nucleus . since the squares in @xmath192 are equal to @xmath140 or @xmath193 , @xmath192 is an extra loop , hence a @xmath180-loop .
it is well - known that @xmath194 .
the bound @xmath191 of corollary @xmath195 can not be improved , since a c - loop with nucleus of size @xmath140 is steiner , hence commutative , by proposition @xmath196 .
[ ex:12 ] the smallest group @xmath141 satisfying the assumptions of proposition [ pr : constr ] is the @xmath197-element cyclic group @xmath198 .
the construction of proposition [ pr : constr ] with @xmath199 then gives rise to the smallest noncommutative nonassociative c - loop : @xmath200 the following properties of this loop will be revoked later : the associator @xmath201 $ ] has order @xmath197 .
note that @xmath202 , @xmath203 , @xmath204 , but @xmath205 .
there are two nonisomorphic nonassociative c - loops of order 14 , both of them steiner loops .
these loops are well known .
we include their multiplication tables for the sake of completeness , and also because we will refer to some of their properties later .
their multiplication tables are given in examples [ ex:14a ] and [ ex:14b ] .
[ ex:14a ] one of the two nonassociative c - loops of order @xmath206 . @xmath207
as we know from proposition [ pr : orders ] , this loop has a trivial nucleus .
thus the associator @xmath208 $ ] is not in the nucleus .
[ ex:14b ] one of the two nonassociative c - loops of order 14 .
@xmath209 we must go beyond @xmath210 to find a nonassociative non - steiner commutative c - loop .
there is one of order @xmath127 , and its multiplication table is in example [ ex:16 ] . we were unable to determine the number of nonassociative c - loops of order @xmath127 .
[ ex:16 ] a nonassociative non - steiner commutative c - loop of the smallest possible order . @xmath211
many properties that we take for granted in groups do not hold in c - loops .
this section is concerned with subloops generated by one or two elements ; with the relations between the order of a loop , the order of a subloop , and the order of an element ; and with alike properties .
a loop @xmath21 is _ power associative _ if for every @xmath32 and every @xmath212 the power @xmath213 is well - defined .
clearly , the powers @xmath214 , @xmath43 , and @xmath64 are always well - defined .
note that , up to this point , we have carefully avoided all higher powers in our calculations .
but we did not have to : lc - loops are power associative .
the power @xmath213 is clearly well - defined for @xmath215 , @xmath140 , @xmath1 , and since , by the left alternative law , @xmath216 , it is also well - defined for @xmath217 .
assume that @xmath218 and that @xmath219 is well - defined for every @xmath220 .
let @xmath119 , @xmath221 be such that @xmath222 .
we now show that @xmath223 can be rewritten canonically as @xmath224 .
since @xmath225 , we can assume that @xmath226 .
then @xmath227 , which is by the lc - identity and by the induction hypothesis equal to @xmath228 .
c - loops are power associative . _
a _ loop @xmath21 is _ diassociative _ if any subloop of @xmath21 generated by two elements is a subgroup .
c - loops are not necessarily diassociative . to see this , consider the c - loop @xmath21 of order 12 with multiplication table given in example [ ex:12 ] .
note that @xmath229 .
it is the nobvious from the multiplication table that @xmath230 .
thus @xmath21 is generated by two elements , yet it is not associative . in @xcite , _ arif loops _ are defined to be flexible loops satisfying @xmath231 .
[ lm : arif ] flexible c - loops are arif loops . since c - loops are alternative and have all squares in the nucleus , we have @xmath232 .
we use this identity twice to obtain @xmath233 .
[ lm : fd ] flexible c - loops are diassociative . in particular , commutative c - loops are diassociative . by lemma [ lm : arif ] , flexible c - loops are arif loops .
by @xcite , arif loops are diassociative .
c - loops are alternative . in the presence of commutativity ,
the two alternative laws are not only equivalent to each other , but also to the flexible law .
power alternativity is best expressed in terms of translations .
a power associative loop @xmath21 is _ left power alternative _ if @xmath234 for every @xmath235 and @xmath32 .
similarly , @xmath21 is _ right power alternative _ if @xmath236 for every @xmath235 and @xmath32 .
power associative loops that are both left and right power alternative are called _ power alternative_.
[ lm : lcpa ] lc - loops are left power alternative .
we have @xmath237 by left alternativity .
let @xmath161 , and assume that @xmath238 for every @xmath239 .
we have @xmath240 by power associativity .
since lc - loops are middle nuclear square , we have @xmath241 , which , by induction , is equal to @xmath242 .
c - loops are power alternative . a finite loop @xmath21 is said to have the _ weak lagrange property _ if the order of any subloop of @xmath21 divides the order of @xmath21 . a finite loop @xmath21 has the _ weak monogenic lagrange property _ if the order of any monogenic subloop of @xmath21 divides the order of @xmath21 . to any weak property
, there is its strong version : a loop has the _
strong property p _ if every subloop of @xmath21 has the weak property @xmath243 .
the weak property does not always imply its strong version .
for instance , there are loops with the weak but not the strong lagrange property .
steiner loops ( and hence c - loops ) do not have the weak lagrange property .
this is illustrated by the loop in example [ ex:14a ] that is of order 14 and possesses a subloop @xmath244 of order @xmath108 . however , c - loops have the strong monogenic lagrange property
. we can establish this by imitating the proof of the lagrange theorem for groups .
[ lm : smlp ] let @xmath21 be a finite loop that is left power alternative and has the left inverse property
. then @xmath21 has the strong monogenic lagrange property .
it suffices to prove that @xmath21 has the weak monogenic lagrange property .
let @xmath32 , @xmath245 .
we claim that any two right cosets of @xmath246 are either disjoint or coincide .
let @xmath247 , @xmath248 be two such cosets .
assume that @xmath249 .
then @xmath250 for some @xmath4 , @xmath251 . by the left inverse property , @xmath252 . by the left power alternative law , @xmath253
. then @xmath254 , by the left power alternative law again .
[ cr : smlp ] let @xmath43 be an element of a finite lc - loop @xmath21 .
then the order of @xmath43 divides the order of @xmath21 .
lc - loops are left power alternative , by lemma [ lm : lcpa ] , and have the left inverse property , by proposition [ pr : sum](ii ) .
we are done by lemma [ lm : smlp ] .
a finite power associative loop is said to have the _ weak cauchy property _ if for any prime @xmath114 dividing the order of the loop there is an element of order @xmath114 .
since there are steiner loops of order different from @xmath118 , yet all steiner loops have exponent @xmath1 , it is clear that steiner loops ( and thus c - loops ) do not have the weak cauchy property .
example [ ex:10 ] illustrates this nicely ( and minimally ) for @xmath255 .
since our main structural result for commutative c - loops ( corollary [ cr : decomposition ] ) requires the notion of a @xmath1-c - loop , let us talk about @xmath1-loops .
a group @xmath139 of order @xmath4 is said to be a _ @xmath1-group _ if @xmath256 for some @xmath119 , or , equivalently , if @xmath139 is of exponent @xmath257 for some @xmath258 .
the trouble with power associative loops is that the above two properties are not necessarily equivalent .
the smallest possible counterexample is a nonassociative power associative loop of order @xmath123 and exponent @xmath1 , cf .
* example i.4.5 ) . throughout this paper ,
we therefore postulate : a finite power associative loop @xmath21 is said to be a _
@xmath1-loop _ if it is of exponent @xmath257 , for some @xmath258 .
note that if @xmath21 is a c - loop of order @xmath118 then @xmath21 is a @xmath1-loop , by corollary [ cr : smlp ] .
the converse is not true , even for the smaller class of steiner loops , as is demonstrated by the nonassociative steiner loop of order @xmath137 .
example [ ex:12 ] demonstrates that the subset @xmath259 of a c - loop @xmath21 is not necessarily a subloop of @xmath21 .
we are going to see that in the commutative case , @xmath39 is not only a subloop , but a normal subloop .
[ lm : xyxy ] let @xmath21 be a commutative , alternative , inverse property loop .
then @xmath260 for every @xmath43 , @xmath44 .
consider @xmath261 . by alternativity , @xmath262 . by the inverse property , @xmath263 . by commutativity , @xmath264 . by alternativity ,
@xmath265 . finally , by the inverse property , @xmath266 .
thus @xmath267 .
[ pr : lk ] let @xmath21 be a commutative @xmath180-loop , and let @xmath39 consist of all elements of exponent @xmath1 in @xmath21 . then @xmath39 is a normal subloop of @xmath21 and @xmath45 is a group . by lemma [ lm : xyxy ] ,
the map @xmath268 is an endomorphism of @xmath21 .
its kernel @xmath39 is therefore a normal subloop of @xmath21 .
it remains to show that @xmath45 is associative .
this is true if and only if @xmath269 for every @xmath43 , @xmath29 , @xmath34 , @xmath270 , or , equivalently , if @xmath271 .
since squaring is a homomorphism and since all squares are in the nucleus , we can rewrite the last equation as @xmath272 , which certainly holds , as @xmath270 .
let @xmath21 be a commutative c - loop and let @xmath141 be the subloop of @xmath21 generated by all associators @xmath273 $ ] , where @xmath43 , @xmath29 , @xmath34 .
then @xmath141 is of exponent at most @xmath1 .
let @xmath39 be as in proposition [ pr : lk ] .
since @xmath45 is associative , all associators must be in @xmath39 .
thus @xmath274 . even in the noncommutative case we can say something about the associators .
let @xmath21 be a loop with normal nucleus
. then all associators of @xmath21 commute with all nuclear elements of @xmath21 .
this is ( * ? ? ?
* lemma 4.2(vii ) ) .
let @xmath21 be a c - loop
. then all associators of @xmath21 commute with all nuclear elements .
in particular , associators commute with all squares
. it would be nice if products of associators were again associators .
unfortunately , this fails already for extra loops ( hence for c - loops ) , by @xcite .
the smallest variety of nonassociative loops of bol - moufang type is that of extra loops ( cf .
we would like to describe an analogy between extra loops and commutative c - loops .
a loop @xmath21 is _ conjugacy closed _ , if for every @xmath43 , @xmath44 , @xmath275 is a left translation , and @xmath276 is a right translation .
it is well known that extra loops are exactly conjugacy closed moufang loops ( see , for instance , @xcite ) .
basarab @xcite showed that @xmath92 is an abelian group for any conjugacy closed loop @xmath21 and its nucleus @xmath65 .
( also see @xcite , @xcite . )
since extra loops are also precisely moufang loops with squares in the nucleus , @xmath92 is an elementary abelian @xmath1-group whenever @xmath21 is extra .
proposition [ pr : ln ] shows that @xmath92 is a steiner loop when @xmath21 is a c - loop . since steiner loops are commutative of exponent @xmath1 , they differ from elementary abelian @xmath1-groups `` only '' in their lack of associativity . the analogy can be extended little further for commutative c - loops .
we have seen that all associators in a commutative c - loop are of order at most @xmath1 .
the same is true for extra loops .
however , all associators of an extra loop are in the nucleus .
this is not the case for commutative c - loops , as example [ ex:14a ] illustrates .
we finish this paper with a decomposition theorem for finite commutative c - loops .
[ lm : uv ] let @xmath21 be a finite commutative c - loop .
let @xmath277 is a power of @xmath278 , @xmath279 is relatively prime to @xmath278 . then : 1 .
@xmath280 , @xmath281 , 2 .
@xmath282 is contained in the nucleus of @xmath21 , hence @xmath282 is a commutative group , 3 .
@xmath283 , 4 .
@xmath284 , 5 .
@xmath285 , 6 .
@xmath286 .
first of all , by commutativity and diassociativity ( lemma [ lm : fd ] ) , we have @xmath287 for every @xmath43 , @xmath44 and every integer @xmath4 .
let @xmath43 , @xmath288 .
let @xmath4 be the least common multiple of @xmath289 , @xmath290 .
since @xmath43 , @xmath29 are powers of @xmath1 , @xmath4 is a power of @xmath1 ( the maximum of @xmath289 and @xmath290 ) . as @xmath291
, we see that @xmath292 divides @xmath4 , and is therefore a power of @xmath1 .
let @xmath43 , @xmath293 .
let @xmath4 be the least common multiple of @xmath289 , @xmath290 .
since both @xmath43 , @xmath29 are relatively prime to @xmath1 , so is @xmath4 . as @xmath291
, we see that @xmath292 divides @xmath4 , and is therefore relatively prime to @xmath1 . we have proved ( i ) .
let @xmath294 .
we want to show that @xmath43 belongs to the nucleus of @xmath21 .
let @xmath295 .
then @xmath296 is even , and @xmath297 is a square . since c - loops are nuclear square loops , @xmath43 is in the nucleus .
any subloop contained in the center of @xmath21 is normal in @xmath21 . by ( ii ) , @xmath282 is contained in the center of @xmath21 , and so @xmath283 .
we now show that @xmath298 is normal in @xmath21 .
thanks to commutativity , all we have to show is that @xmath299 for every @xmath43 , @xmath44 .
this is equivalent to showing that @xmath300 for every @xmath301 .
let @xmath302 be the order of @xmath303 .
then @xmath304 , by lemma [ lm : xyxy ] .
thus the order of @xmath305 divides @xmath302 , and @xmath306 follows . since ( vi ) follows immediately from the definition of @xmath298 and @xmath282 , it remains to prove that @xmath307 .
let @xmath32 , and let @xmath308 , where @xmath309 and @xmath221 is an odd integer .
( there is nothing to prove when @xmath310 or @xmath311 . ) since @xmath118 , @xmath258 are relatively prime , there are integers @xmath66 , @xmath4 such that @xmath312 ( * ? ? ? * theorem 2 - 4 ) .
then @xmath313 , where @xmath314 , @xmath315 .
since @xmath316 , we see that @xmath317 divides @xmath118 , hence @xmath317 is a power of @xmath1 , and @xmath301 follows .
similarly , since @xmath318 , we see that @xmath319 divides @xmath258 , hence @xmath319 is odd , and @xmath320 follows .
universal algebraists define loops equivalently as sets with three binary operations @xmath321 , @xmath322 , @xmath323 , and one nullary operation @xmath17 such that @xmath324 , @xmath325 , @xmath326 , @xmath327 , @xmath328 .
thus @xmath329 is the solution @xmath305 to the equation @xmath330 , and similarly for @xmath331 .
we will use this notation in the proof of the following theorem , that could be called _ the internal direct product for loops_. the theorem appears in a more general form in bruck s book ( * ? ? ? * lemma iv.5.1 ) .
since he does not give a proof , we provide it .
[ th : dpl ] let @xmath21 be a loop with normal subloops @xmath39 , @xmath246 such that @xmath332 , @xmath333 , @xmath334 . then @xmath21 is the direct product of @xmath39 , @xmath246 .
we first show that any element @xmath32 decomposes uniquely as a product @xmath335 , @xmath336 , @xmath337 .
at least one decomposition exists since @xmath338 .
let @xmath339 be two such decompositions . then @xmath340 .
now , @xmath341 belongs to @xmath342 ( since @xmath39 is a normal subloop ) , and there is therefore @xmath336 such that @xmath343 .
then @xmath344 belongs to @xmath332 , and thus @xmath345 , which in turn implies @xmath346 .
define @xmath347 by @xmath348 . by the preceding paragraph ,
@xmath349 is one - to - one and onto .
it remains to show that @xmath349 is a homomorphism , i.e. , that @xmath350 for every @xmath351 , @xmath352 , @xmath353 , @xmath354 .
since @xmath39 is normal , we have @xmath355 , and there is @xmath336 such that @xmath356 . since @xmath246 is normal , we have @xmath357 , and there is @xmath337 such that @xmath358 . since @xmath43 has a unique decomposition , we must have @xmath359 , and so @xmath360 .
[ th : decomposition ] let @xmath21 be a finite commutative c - loop
. then @xmath361 , where @xmath362 is a power of @xmath278 , @xmath363 is odd@xmath364 .
combine lemma [ lm : uv ] and theorem [ th : dpl ] .
[ cr : decomposition ] every finite commutative c - loop is a direct product of a finite commutative group and a finite commutative @xmath1-c - loop , and _
vice versa_. it is worth noting that we did not assume finiteness of the loop in this section , only the fact that every element has finite order .
power associative loops with all elements of finite order are called _
torsion loops_. thus , lemma [ lm : uv ] , theorem [ th : decomposition ] and corollary [ cr : decomposition ] remain valid if all occurrences of `` finite '' in their statements are replaced by `` torsion '' .
we conclude this paper with yet another analogy between c - loops and moufang loops .
theorem [ th : decomposition ] shows that finite commutative c - loops are of the form @xmath365 , where @xmath298 consists of elements whose order is a power of @xmath1 , and @xmath282 consist of elements whose order is relatively prime to @xmath1 . by
* corollary of theorem 3 ) , finite commutative moufang loops are of the from @xmath365 , where @xmath298 consists of elements whose order is a power of @xmath197 , and @xmath282 consist of elements whose order is relatively prime to @xmath197 .
we thank michael k. kinyon , indiana university south bend , for many useful comments , especially for the proof of theorem [ th : dpl ] ; scott feller , department of chemistry , wabash college , for generously allowing us to use his beowulf cluster for our mace4 computations ; the anonymous referee who suggested a deeper look at c - loops of small order we came up with the construction of proposition [ pr : constr ] in response to his / her inquiry . | c - loops are loops satisfying @xmath0 .
they often behave analogously to moufang loops and they are closely related to steiner triple systems and combinatorics .
we initiate the study of c - loops by proving : ( i ) steiner loops are c - loops , ( ii ) c - loops are alternative , inverse property loops with squares in the nucleus , ( iii ) the nucleus of a c - loop is a normal subgroup , ( iv ) c - loops modulo their nucleus are steiner loops , ( v ) c - loops are power associative , power alternative but not necessarily diassociative , ( vi ) torsion commutative c - loops are products of torsion abelian groups and torsion commutative @xmath1-c - loops ; and several other results .
we also give examples of the smallest nonassociative c - loops , and explore the analogy between commutative c - loops and commutative moufang loops . |
the frequency band structure of artificial periodic dielectrics formally known as photonic crystals is the electromagnetic ( em ) counterpart of the electronic band structure in ordinary atomic solids .
recently , a new analogy between electron and photon states in periodic structures has been proposed by raghu and haldane , @xcite namely the one - way chiral edge states in two - dimensional ( 2d ) photonic - crystal slabs which are similar to the corresponding edge states in the quantum hall effect .
@xcite the photonic chiral edge states are a result of time - reversal ( tr ) symmetry breaking which comes about with the inclusion of gyroelectric/ gyromagnetic material components ; these states are robust to disorder and structural imperfections as long as the corresponding topological invariant ( chern number in this case ) remains constant . in certain atomic solids , tr
symmetry breaking is not prerequisite for the appearance of topological electron states as it is the case in the quantum hall effect .
namely , when spin - orbit interactions are included in a tr symmetric graphene sheet , a bulk excitation gap and spin - filtered edge states emerge @xcite without the presence of an external magnetic field , a phenomenon which is known in literature as quantum spin hall effect .
its generalization to three - dimensional ( 3d ) atomic solids lead to a new class of solids , namely , topological insulators .
@xcite the latter possess a spin - orbit - induced energy gap and gapless surface states exhibiting insulating behavior in bulk and metallic behavior at their surfaces .
apart from topological insulators where the spin - orbit band structure with tr symmetry defines the topological class of the corresponding electron states , other topological phases have been proposed such as topological superconductors ( band structure with particle - hole symmetry ) , @xcite magnetic insulators ( band structure with magnetic translation symmetry ) , @xcite and , very recently , topological crystalline insulators .
@xcite in the latter case the band structure respects tr symmetry as well as a certain point - group symmetry leading to bulk energy gap and gapless surface states . in this work ,
we propose a photonic analog of a topological crystalline insulator .
our model photonic system is a 3d crystal of weakly interacting resonators respecting tr symmetry and the point - symmetry group associated with a given crystal surface . as a result
, the system possesses an omnidirectional band gap within which gapless surface states of the em field are supported .
it is shown that the corresponding photonic band structure is equivalent to the energy band structure of an atomic topological crystalline insulator and , as such , the corresponding states are topological states of the em field classified by a @xmath0 topological invariant .
the frequency band structure of photonic crystals whose ( periodically repeated ) constituent scattering elements interact weakly with each other can be calculated by a means which is similar to the tight - binding method employed for atomic insulators and semiconductors .
photonic bands amenable to a tight - binding - like description are e.g. , the bands stemming from the whispering - gallery modes of a lattice of high - index scatterers @xcite the defect bands of a sublattice of point defects , within a photonic crystal with an absolute band gap , @xcite the plasmonic bands of a lattice of metallic spheres @xcite or of a lattice of dielectric cavities within a metallic host .
@xcite in the latter case , the frequency band structure stems from the weak interaction of the surface plasmons of each individual cavity @xcite wherein light propagates within the crystal volume by a hopping mechanism .
such type of lattice constitutes the photonic analog of a topological crystalline insulator presented in this work whose frequency band structure will be revealed based on a photonic tight - binding treatment within the framework of the coupled - dipole method .
@xcite the latter is an exact means of solving maxwell s equations in the presence of nonmagnetic scatterers .
we consider a lattice of dielectric cavities within a lossless metallic host .
the @xmath1-th cavity is represented by a dipole of moment @xmath2 which stems from an incident electric field @xmath3 and the field which is scattered by all the other cavities of the lattice . this way the dipole moments of all the cavities
are coupled to each other and to the external field leading to the coupled - dipole equation @xmath4 .
\label{eq : cde}\ ] ] @xmath5 is the electric part of the free - space green s tensor and @xmath6 is the @xmath7 polarizability tensor of the @xmath1-th cavity .
( [ eq : cde ] ) is a @xmath8 linear system of equations where @xmath9 is the number of cavities of the system .
we assume that the cavities exhibit a uniaxial em response , i.e. , the corresponding polarizability tensor is diagonal with @xmath10 and @xmath11 . for strong anisotropy ,
the cavity resonances within the @xmath12-plane and along the @xmath13-axis can be spectrally distinct ; thus , around the region of e.g. , the cavity resonance @xmath14 within the @xmath12-plane , @xmath15 ( see appendix ) . in this case , one can separate the em response within the @xmath12-plane from that along the @xmath13-axis and eq .
( [ eq : cde ] ) becomes a @xmath16 system of equations , @xmath17 .
\label{eq : cde_no_field}\ ] ] where we have set @xmath18 since we are seeking the eigenmodes of the system of cavities .
also , now , @xmath19 . for a particle / cavity of electric permittivity @xmath20 embedded within a material host of permittivity @xmath21 , the polarizability @xmath22 is given by the clausius - mossotti formula @xmath23 where @xmath24 is the volume of the particle/ cavity . for a lossless plasmonic ( metallic ) host in which case
the electric permittivity can be taken as drude - type , i.e. , @xmath25 ( where @xmath26 is the bulk plasma frequency ) , the polarizability @xmath22 exhibits a pole at @xmath27 ( surface plasmon resonance ) . by making a laurent expansion of @xmath22 around @xmath14 and keeping the leading term ,
we may write @xmath28 where @xmath29 . for sufficiently high value of the permittivity of the dielectric cavity , i.e. , @xmath30 , the electric field of the surface plasmon is much localized at the surface of the cavity . as a result , in a periodic lattice of cavities ,
the interaction of neighboring surface plasmons is very weak leading to much narrow frequency bands . by treating such a lattice in a tight binding - like framework
, we may assume that the green s tensor @xmath5 does not vary much with frequency and therefore , @xmath31 .
in this case , eq . ( [ eq : cde_no_field ] ) becomes an eigenvalue problem @xmath32 where @xmath33 .
\label{eq : g_tensor}\end{aligned}\ ] ] with @xmath34 , @xmath35 and @xmath36 is the @xmath37 unit matrix .
the form of functions @xmath38 , @xmath39 generally depends on the type of medium hosting the cavities ( isotropic , gyrotropic , bi - anisotropic , etc ) . @xcite the green s tensor of eq .
( [ eq : g_tensor ] ) describes the electric interactions between two point dipoles @xmath40 and @xmath41 each of which corresponds to a single cavity .
the first term of @xmath42 describes an interaction which does not depend on the orientation of the two dipoles whilst the second one is orientation dependent . for an infinitely periodic system , i.e. , a crystal of cavities
, we assume the bloch ansatz for the polarization field , i.e. , @xmath43 the cavity index @xmath1 becomes composite , @xmath44 , where @xmath45 enumerates the unit cell and @xmath46 the positions of inequivalent cavities in the unit cell .
also , @xmath47 denotes the lattice vectors and @xmath48 is the bloch wavevector . by substituting eq .
( [ eq : bloch ] ) into eq .
( [ eq : cde_eigen ] ) we finally obtain @xmath49 where @xmath50 { \bf g}_{n \beta ; n ' \beta'}(\omega_{\parallel } ) .
\label{eq : green_fourier}\ ] ] solution of eq .
( [ eq : cde_eigen_periodic ] ) provides the frequency band structure of a periodic system of cavities . )
corresponding to the green s tensor of eq .
( [ eq : g_elem ] ) with @xmath51.,width=302 ]
since eq . ( [ eq : cde_eigen_periodic ] )
is equivalent to a hamiltonian eigenvalue problem , we adopt the crystal structure of ref . .
namely , a tetragonal lattice with a unit cell consisting of two same cavities at inequivalent positions @xmath52 and @xmath53 [ see fig . [ fig1](a ) ] along the @xmath54-axis . in this case , the index @xmath46 in eq . ( [ eq : cde_eigen_periodic ] ) assumes the values @xmath55 for each sublattice ( layer ) of the crystal .
the above lattice is characterized by the @xmath56 point - symmetry group . in order to preserve the @xmath56 symmetry @xcite in the green s tensor matrix of eq .
( [ eq : cde_eigen_periodic ] ) we assume that the interaction between two cavities within the same layer ( either @xmath52 or @xmath53 ) depends on the relative orientation of the point dipole in each cavity whilst the interaction between cavities belonging to adjacent layers is orientation independent .
also , we take into account interactions up to second neighbors in both inter- and intra - layer interactions . taking the above into account
, the lattice green s tensor assumes the form @xmath57 where @xmath58 { \bf i}_{2}. \nonumber & & \\ \label{eq : g_elem}\end{aligned}\ ] ] of the crystal of fig .
[ fig1 ] made from 80 bilayers . ,
width=302 ] the lattice green s tensor of eq .
( [ eq : g_elem ] ) is completely equivalent to the lattice hamiltonian of ref . .
@xmath59 in eq .
( [ eq : g_elem ] ) generally depend on @xmath60 , the lattice constant @xmath61 and the interlayer distance @xmath54 but hereafter will be used as independent parameters .
namely we choose @xmath51 . in fig .
[ fig2 ] we show the ( normalized ) frequency band structure corresponding to eq .
( [ eq : g_elem ] ) along the symmetry lines of the brillouin zone shown in fig .
[ fig1](b ) .
it is evident that an omnidirectional frequency band gap exists around @xmath62 which is prerequisite for the emergence of surface states . in order to inquire the occurrence of surface states we find the eigenvalues of the green s tensor of eq .
( [ eq : g_elem ] ) in a form appropriate for a slab geometry .
the emergence of surface states depends critically on the surface termination of the finite slab , i.e. , for different slab terminations different surface - state dispersions occur ( if occur at all ) . namely , we assume a finite slab parallel to the ( 001 ) surface ( characterized by the @xmath56 symmetry group ) consisting of 80 alternating @xmath63 layers except the last bilayer which is @xmath64 , i.e. , the layer sequence is @xmath65 .
the corresponding frequency band structure along the symmetry lines of the surface brillouin zone of the ( 001 ) surface [ see fig . [ fig1](c ) ] is shown in fig .
it is evident that there exist gapless surface states within the band gap exhibiting a quadratic degeneracy at the @xmath66-point . in this case
, the corresponding doublet of surface states can be described by an effective theory @xcite similarly to the doublet states at a point of linear degeneracy ( dirac point ) .
@xcite we note that the equivalence of the green s tensor @xmath67 with the atomic hamiltonian of ref . as well as the form of the time - reversal @xmath68 and ( geometric ) @xmath56-rotation @xmath69 operators for the em problem @xcite which are the same as for spinless electrons , allows to describe the photonic band structure with the @xmath0 topological invariant @xmath70 @xmath71 where for real eigenvectors of @xmath72 we have @xcite @xmath73/ { \rm pf}[w({\bf k}_{1 } ) ] \label{eq : pf_frac}\ ] ] and @xmath74 @xmath75 stands for the pfaffian of a skew - symmetric matrix , i.e. , @xmath76^{2}=\det(w)$ ] .
due to the double degeneracy of the band structure at the four special momenta points @xmath77 the frequency bands come in doublets .
since frequency eigenvectors with different eigenfrequencies are orthogonal , all the inter - pair elements of the @xmath78-matrix are zero and the latter is written as : @xmath79 where @xmath80 are anti - symmetric @xmath81 matrices , @xcite i.e. , @xmath82 or @xmath83 , where @xmath84 in this case , @xmath85=w^{1}_{12 } w^{2}_{12 } \cdots w^{n}_{12}=\pm 1 $ ] .
therefore , @xmath86 and @xmath87 which ensures the presence of gapless surface states .
we note that the above analysis relies on the assumption of real frequency bands .
the presence of losses in the constituent materials renders the frequency bands complex , i.e. , the bloch wavevector possesses both a real and an imaginary part .
however , even in this case , one can still speak of real frequency bands if the imaginary part of the bloch wavevector is at least _
hundred _ times smaller than the corresponding real part .
this a common criterion used in calculations of the complex frequency band structure by on - shell electromagnetic solvers such as the layer - multiple scattering method @xcite or the transfer - matrix method
a possible realization of the photonic analogue of topological insulator in the laboratory is depicted in fig .
since our model system requires dielectric cavities within a homogeneous plasma , a lattice of nano - cavities formed within a homogeneous drude - type metal , e.g. , a noble metal ( au , ag , cu ) , would be the obvious answer .
@xcite however , the plasmon bands are extremely lossy due to the intrinsic absorption of noble metals in the visible regime .
a solution to this would be the use of an _ artificial _ plasmonic medium operating in the microwave regime where metals are perfect conductors and losses are minimal .
artificial plasma can be created by a 3d network of thin metallic wires of a few tens of @xmath88 m in diameter and spaced by a few mm .
@xcite a lattice of dielectric particles within an artificial plasma can be modelled with the presented tight - binding green s tensor .
since the interaction among first and second neighbors within the same bilayer ( @xmath52 or @xmath53 ) should depend on the dipole orientation ( in order to preserve the @xmath56 symmetry ) , the dielectric particles in each layer are connected with cylindrical waveguiding elements ( different in each layer @xmath52 or @xmath53 - see fig .
[ fig4 ] ) .
in contrast , between two successive bilayers there are no such elements since interactions between dipoles belonging to different layers should be independent of the dipole orientations ( s orbital - like ) .
another advantage of realizing the photonic analog in the microwave regime is the absence of nonlinearities in the em response of the constituent materials since photon - photon interactions may destroy the quadratic degeneracy of the surface bands in analogy with fermionic systems .
@xcite finally , we must stress that a photonic topological crystalline insulator can be also realized with purely dielectric materials if the host medium surrounding the cavities is not a plasmonic medium but a photonic crystal with an absolute band gap : the cavities would be point defects within the otherwise periodic photonic crystal and the tight - binding description would be still appropriate .
@xcite in this case , maxwell s equations lack of any kind of characteristic length and the proposed analog would be realized in any length scale .
in conclusion , a 3d lattice of weakly interacting cavities respecting tr symmetry and a certain point - group symmetry constitutes a photonic analog of a topological crystalline insulator by demonstrating a spectrum of gapless surface states .
a possible experimental realization would be a 3d lattice of dielectric particles within a continuous network of thin metallic wires with a plasma frequency in the ghz regime .
this work has been supported by the european community s seventh framework programme ( fp7/2007 - 2013 ) under grant agreement no .
228455-nanogold ( self - organized nanomaterials for tailored optical and electrical properties ) .
the @xmath13-component of the polarizability need not be zero but can assume finite values as far as the frequency band stemming from the surface - plasmon resonance corresponding to the z - component is spectrally distinct ( no overlap ) with the bands stemming from the @xmath12-components of the polarizability .
in this case , one can treat separately the two frequency bands ( doublet ) stemming from the xy - resonance from the band coming from the z - resonance ( singlet ) .
the above requirements can be quantified as follows , @xmath89 where @xmath90 is the width of the @xmath12-frequency bands ( in dimensionless frequency units ) .
@xmath90 is obtained from the first term of the right - hand side of eq .
( [ eq : g_tensor ] ) of the paper . to a first approximation
, it is given by @xcite @xmath91 where @xmath92 .
therefore , the condition ( [ eq : app_1 ] ) is written as @xmath93 where @xmath94 is the lattice constant in the @xmath12-plane .
given that @xmath95 , eq .
( [ eq : app_3 ] ) becomes @xmath96 from the above equation it is evident that for a given value of the dielectric anisotropy @xmath97 , one can always find a suitably large lattice constant @xmath22 such that eq .
( [ eq : app_5 ] ) is fulfilled .
the latter allows the easy engineering of the photonic analog of a crystalline topological insulator since there is practically no restriction on the choice of the ( uniaxial ) material the cavities are made from .
it can also be easily understood that if eq .
( [ eq : g_tensor ] ) holds , the same equation is true for the width @xmath98 of the singlet frequency band ( resulting from the @xmath13-resonance ) . _ numerical example_. suppose that the cavities are made from a nematic liquid crystal which is a uniaxial material .
typical values of the permittivity tensor @xmath99 are e.g. , @xmath100 , @xmath101 . in this case ,
@xmath102 and @xmath103 . by choosing a large lattice constant , i.e. , @xmath104 , eq .
( [ eq : app_5 ] ) is clearly fulfilled @xmath105 f. d. m. haldane and s. raghu , , 013904 ( 2008 ) ; _ ibid _ , , 033834 ( 2008 ) .
z. wang , y. d. chong , j. d. joannopoulos , and m. soljai , , 013905 ( 2008 ) ; _ ibid _ , nature ( london ) * 461 * , 772 ( 2009 ) ; z. yu _
et al . _ , , 023902 ( 2008 ) ; h. takeda and s. john , , 023804 ( 2008 ) ; d. han _ et al .
_ , , 123904 ( 2009 ) ; x. ao , z. lin , and c. t. chan , , 033105 ( 2009 ) ; m. onoda and t. ochiai , , 033903 ( 2009 ) ; t. ochiai and m. onoda , , 155103 ( 2009 ) ; r. shen , l. b. shao , b. wang , and d. y. xing , , 041410(r ) ( 2010 ) ; y. poo _ et al .
_ , , 093903 ( 2011 ) . c. l. kane and e. j. mele , , 226801 ( 2005 ) ; _ ibid _ , , 146802 ( 2005 )
l. fu , c. l. kane , and e. j. mele , , 106803 ( 2007 ) ; l. fu and c. l. kane , , 045302 ( 2007 ) ; m. z. hasan and c. l. kane , , 3045 ( 2010 ) .
a. p. schnyder , s. ryu , a. furusaki , and a. w. w. ludwig , , 195125 ( 2008 ) ; a. kitaev , aip conf .
proc . * 1134 * , 22 ( 2009 ) ; y. ran , arxiv : 1006.5454 ; x.- l. qi , t. l. hughes , s. raghu , and s.- c. zhang , , 187001 ( 2009 ) ; r. roy , arxiv : 0803.2868 .
r. s. k. mong , a. m. essin , and j. e. moore , , 245209 ( 2010 ) ; ; r. li , j. wang , x.- l. qi , and s.- c. zhang , nat . phys .
* 6 * , 284 ( 2010 ) .
l. fu , , 106802 ( 2011 ) .
e. lidorikis , m. m. sigalas , e. n. economou , and c. m. soukoulis , , 1405 ( 1998 ) .
m. bayindir , b. temelkuran , and e. ozbay , , 2140 ( 2000 ) .
m. quinten , a. leitner , j. r. krenn , and f. r. aussenegg , opt .
* 23 * , 1331 ( 1998 ) .
n. stefanou , a. modinos , and v. yannopapas , sol .
. commun . * 118 * , 69 ( 2001 ) .
e. m. purcell and c. r. pennypacker , astrophys .
j. * 186 * , 705 ( 1973 ) . for
a detailed study see supplemental material at ...
a. eroglu , _ wave propagation and radiation in gyrotropic and anisotropic media _ ( springer , new york , 2010 ) . v. dmitriev , prog . electromag
. res . * 48 * , 145 ( 2004 ) .
y. d. chong , x.- g. wen , and m. soljacic , , 235125 ( 2008 ) .
r. a. sepkhanov , ya .
b. bazaliy , and c. w. j. beenakker , , 063813 ( 2007 ) .
z. wang , x- .
l.- qi , and s.- c.- zhang , new j. phys .
* 12 * , 065007 ( 2010 ) .
n. stefanou , v. yannopapas , and a. modinos , comput .
. commun . * 113 * , 49 ( 1998 ) .
p. m. bell _ et al .
_ , comput .
. commun . * 85 * , 306 ( 1995 ) .
j. b. pendry , a. j. holden , w. j. stewart , and i. youngs , , 4773 ( 1996 ) .
k. sun , h. yao , e. fradkin , and s. a. kivelson , , 046811 ( 2009 ) .
k. fang , z. yu , and s. fan , , 075477 ( 2011 ) . | we show that a tetragonal lattice of weakly interacting cavities with uniaxial electromagnetic response is the photonic counterpart of topological crystalline insulators , a new topological phase of atomic band insulators .
namely , the frequency band structure stemming from the interaction of resonant modes of the individual cavities exhibits an omnidirectional band gap within which gapless surface states emerge for finite slabs of the lattice . due to the equivalence of a topological crystalline insulator with its photonic - crystal analog ,
the frequency band structure of the latter can be characterized by a @xmath0 topological invariant .
such a topological photonic crystal can be realized in the microwave regime as a three - dimensional lattice of dielectric particles embedded within a continuous network of thin metallic wires . |
consider an edge weighted directed graph as a model of a neural network .
such network can be build up using real life measurements ( see for example , @xcite ) , and neuroscientists are interested in comparing this network with random networks . if the edges were not weighted , the typical approach would be to generate random graphs with prescribed in and out degrees .
this topic has a tremendous literature , see for example , @xcite , just to mention a few . in case of weights
are introduced , one might want to generate a random graph that keeps not only the sum of the weights but also individual weights .
the weights can be transformed into colors , and then we are looking for an edge colored graph with prescribed in and out degrees for each color .
this problem is known as finding edge packing @xcite , edge disjoint realizations @xcite , or degree constrained edge - partitioning @xcite .
the problem has also applications in discrete tomography @xcite . above finding one edge colored graph with given constraints ,
it is also an important problem how to generate a `` typical '' solution , as we can see in the before mentioned neuroscientific problem .
unfortunately , the general edge packing problem is np - complete even if the number of colors is 2 @xcite . although the general problem is np - complete , special cases are tractable .
such special cases include the case when the number of colors is 2 and the subgraph with one of the colors is almost regular @xcite .
another tractable case is when the graph is bipartite , the number of colors is 2 , and there exist constants @xmath2 and @xmath3 such that for each vertex , the total number of edges in one of the vertex class is @xmath4 , @xmath2 or @xmath5 and in the other vertex class is @xmath6 , @xmath3 or @xmath7 @xcite .
when the number of colors is unlimited , tractable solutions exist if the graphs are forests for each color @xcite or the graphs are forests for each color except one of the colors might contain one cycle @xcite .
a special case is when the bipartite graph is the union of @xmath8 1-factors .
such edge packings are simply the latin squares . for any @xmath8 ,
latin squares exist , and even a markov chain is known that explores the space of latin squares for a fixed @xmath8 @xcite .
this markov chain is conjectured to be rapidly mixing , that is , computationally efficient to sample random latin squares .
the conjecture has neither proved nor disproved in the last twenty years . in this paper
, we consider another tractable case with unlimited number of colors .
we require that the subgraphs for each color be half - regular , that is , the degrees are constant on one of the vertex class .
we further require that this regular vertex class be the same for all colors .
we give sufficient and necessary conditions when this edge packing problem has a solution .
we also give necessary and sufficient perturbations to transform solutions into each other .
a markov chain monte carlo method has been given using these perturbations , and we give a proof that the inverse of the acceptance ratio is polynomial bounded .
in this paper , we are going to work with realizations of half - regular degree matrices , defined below . a _ bipartite degree sequence _ @xmath9 is a pair of sequences of non - negative integers .
a bipartite degree sequence is _ graphical _ if there exists a simple bipartite graph @xmath10 whose degrees correspond exactly to @xmath11 .
we say that @xmath10 is a _ realization _ of @xmath11 . a _ bipartite degree matrix _
@xmath12 is a pair of @xmath13 and @xmath14 matrices of non - negative integers .
a bipartite degree matrix is graphical if there exists an edge colored simple bipartite graph @xmath0 such that for all color @xmath15 and for all @xmath16 , the number of edges of @xmath17 with color @xmath15 is @xmath18 and for all @xmath19 , the number of edges of @xmath20 with color @xmath15 is @xmath21 .
such graph is called a _
realization _ of @xmath22 .
a bipartite degree matrix is _ half - regular _ if for all @xmath23 , @xmath24 .
the rows of @xmath22 are bipartite degree sequences that are also called _ _ factor__s and the edge colored realization of @xmath22
is also called an _ @xmath22-factorization_. we consider two problems .
one is the existence problem which asks the question if there is a realization of a half - regular degree matrix .
the other is the sampling problem , which considers the set of all realizations of a half - regular degree matrix and asks the question how to sample uniformly a realization from this set .
markov chain monte carlo ( mcmc ) methods are generally applicable for such problems , and we are also going to introduce an mcmc for sampling realizations of half - regular degree sequences .
below we introduce the main definitions .
a _ discrete time , finite markov chain _ is a random process that undergoes transitions from one state to another in a finite state space .
the process has the markov property which means that the distribution of the next state depends only on the current state .
the transition probabilities can be described with a transition matrix @xmath25 , where @xmath26 , namely , the conditional probability that the next state is @xmath27 given that the current state is @xmath28 . when the state space is a large set of combinatorial objects ,
the transition probabilities are not given explicitly , rather , a random algorithm is given that generates a random @xmath27 by perturbing the current state @xmath28 .
such algorithm is called _
transition kernel_. when the process starts in a state @xmath27 , after @xmath29 number of steps , it will be in a random state with distribution @xmath30 , where @xmath31 is the column vector containing all @xmath32 s except for coordinate @xmath33 , which is @xmath34 .
the markov chain monte carlo method is the way to tailor the transition probabilities such that the limit distribution @xmath35 be a prescribed distribution . to be able to do this , necessary conditions are that the markov chain be irreducible , aperiodic and the transitions be reversible defined below . given a discrete time , finite markov chain on the state space @xmath36 , the _ markov graph _ of the markov chain is a directed graph @xmath37 , with vertex set @xmath36 and there is an edge from @xmath38 to @xmath39 iff @xmath40 .
a markov chain is _ irreducible _ iff its markov graph is strongly connected .
when the transition kernel of an irreducible markov chain generates a class of perturbations , we also say that this class of perturbations is irreducible on the state space .
a markov chain is _ aperiodic _ if the largest common divisor of cycle lengths of its markov graph is 1 .
this automatically holds , if there is a loop in the markov chain , that is , a state @xmath41 exists for which @xmath42 .
the transition kernel of a markov chain is reversible if for all @xmath43 , @xmath44 .
the strength of the theory of markov chain monte carlo is that any irreducible , aperiodic markov chain with reversible transition kernel can be tailored into a markov chain converging to a prescribed distribution as stated below .
@xcite let an irreducible , aperiodic markov chain be given with reversible transition kernel @xmath45 over the finite state space @xmath36 .
let @xmath46 be a distribution over @xmath36 , for which @xmath47 , @xmath48 .
then the following algorithm , called the metropolis - hastings algorithm , also defines a markov chain that converges to @xmath46 , namely , its transition kernel @xmath49 satisfies @xmath50 for all indices @xmath33 . 1 .
draw a random @xmath51 following the distribution @xmath52 where @xmath53 is the current state of the markov chain after @xmath29 steps .
2 . draw a random @xmath54 following the uniform distribution over @xmath55 $ ] . the next state , @xmath56 , will be @xmath51 if @xmath57 and @xmath53 otherwise .
the ratio in the second step of the algorithm is called the metropolis - hastings ratio .
it is simplified to @xmath58 when the target distribution @xmath46 is the uniform one .
when this ratio is small , the markov chain defined by the algorithm tends to remain in the same state for a long time , thus increasing the mixing time , namely , the number of steps necessary to get close to the target distribution . in this paper , we introduce perturbations that are irreducible on the realizations of a half - regular degree matrix . using these perturbations
, we design a reversible transition kernel , for which @xmath59 where @xmath60 is the number of vertices in vertex class @xmath61 .
let @xmath37 be a multigraph , and for each @xmath62 , let @xmath63 be @xmath64 where @xmath65 denotes the multiplicity of edge @xmath66 . the _ exceed number _ of graph @xmath10
is defined as @xmath67 clearly , the exceed number is @xmath32 iff @xmath10 is a simple graph .
[ the : regular - factorization ] let @xmath68 be a half - regular bipartite degree matrix .
the bipartite complete graph @xmath69 has an @xmath22-factorization iff 1 .
@xmath70 , @xmath71 2 .
@xmath72 3 .
@xmath73 , @xmath74 .
@xmath75 if @xmath69 has an @xmath22-factorization , then clearly , all factors are graphical and the degrees sum to the degrees of the complete bipartite graph . conditions @xmath76 and @xmath77 explicitly state that the sum of the degrees in the factors sum up to the degree of the complete bipartite graph .
condition @xmath34 states that in each factor , the sums of the degrees in the two vertex classes are the same , which is a necessary condition for a graphical degree sequence .
@xmath78 conditions @xmath76 and @xmath77 also say implicitly that for each @xmath33 , @xmath79 and for each @xmath80 , @xmath81 . together with condition @xmath34 , they are sufficient conditions that a half - regular bipartite degree sequence be graphical .
namely , the theorem says if all factors are graphical and the sums of the degrees are the degrees of the complete bipartite graph , then @xmath69 has such a factorization . for each @xmath33 ,
let @xmath82 be a realization of the degree sequence @xmath83 .
let @xmath84 .
color the edges of @xmath10 such that an edge is colored by color @xmath15 if it comes from the realization @xmath82 .
@xmath10 might be a multigraph , however , for each color @xmath15 and each pair of vertices @xmath85 , there can be at most one edge between @xmath54 and @xmath86 with color @xmath15 . if @xmath87 , then @xmath10 is a simple graph , and due to the conditions , it is @xmath69 , thus we found an @xmath22-factorization of @xmath69 .
assume that @xmath88 .
then there is a pair of vertices @xmath85 such that there are more than one edge between @xmath54 and @xmath86 .
fix one such pair @xmath66 , and let @xmath89 denote the set @xmath90 .
since the degree of @xmath86 in @xmath10 is @xmath8 , there must be a @xmath91 such that there is no edge between @xmath91 and @xmath86 .
let @xmath92 denote the set of colors that appear as edge colors between @xmath54 and @xmath86 .
since for each color @xmath15 , @xmath54 and @xmath91 has the same number of edges with color @xmath15 , there must be at least one vertex @xmath93 such that there are more edges with color from @xmath92 between @xmath91 and @xmath93 than the edges with color from @xmath92 between @xmath54 and @xmath93 .
let @xmath94 denote the set of vertices for which this condition holds .
there are three possibilities ( see also figure [ fig : firstiteration ] ) : 1 .
there is a vertex @xmath95 such that there are more than one edge between @xmath91 and @xmath93 .
2 . there is a vertex @xmath95 such that there is only a single edge between @xmath91 and @xmath93 , however , there is no edge between @xmath54 and @xmath93 .
3 . for each vertex @xmath95
, there is only a single edge between @xmath91 and @xmath93 and there is at least one edge between @xmath54 and @xmath93 . and @xmath86 .
see text for details.,width=278 ] if case @xmath34 holds , then let @xmath96 be a color in @xmath92 such that there is a @xmath96 colored edge between @xmath91 and @xmath93 and there is no @xmath96 colored edge between @xmath54 and @xmath93 . remove that edge between @xmath54 and @xmath86 and also between @xmath91 and @xmath93 and add a @xmath96 colored edge between @xmath91 and @xmath86 and also between @xmath54 and @xmath93 .
note that @xmath97 , the graph with color @xmath96 remains a simple graph with the prescribed degree sequence .
@xmath63 decreases by @xmath34 .
@xmath98 remain @xmath32 , since the edge with color @xmath96 is the only edge between them .
@xmath99 decreases by @xmath34 .
@xmath100 might increase by @xmath34 if there was already at least one edge between them before the change , otherwise it remains the same .
altogether , @xmath101 is decreased at least by @xmath34 .
if case @xmath76 holds , then do the same edge changing than in case @xmath34 .
this causes a decrease in @xmath63 by one , and no change in @xmath98 , @xmath100 and @xmath99 .
altogether , @xmath101 decreases by @xmath34 . in case @xmath77 , select a subset @xmath102 from @xmath94 such that the colors of edges between @xmath91 and the vertices in @xmath102 are exactly the set @xmath92 .
this is doable , since for each vertex @xmath95 , there is only one edge between @xmath91 and @xmath93 , furthermore , the union of colors of edges between @xmath91 and @xmath94 must contain a subset @xmath92 .
let @xmath103 denote the ( possibly multi)set of colors that appear as edge colors between vertex @xmath54 and the set of vertices @xmath102 .
the multiset of colors between @xmath54 and @xmath104 is @xmath105 , where @xmath106 denote the multiset union , while the set of colors between @xmath91 and @xmath104 is @xmath92 .
therefore , there is at least one vertex @xmath107 such that the number of edges with colors from @xmath103 between @xmath91 and @xmath93 is more than the number of edges with color from @xmath103 between @xmath54 and @xmath93 .
let @xmath94 denote the set of vertices with this property .
if case @xmath77 holds , then we can select a subest @xmath108 from @xmath94 such that the multiset of colors between vertex @xmath91 and @xmath108 is exactly @xmath103 .
then the multiset of colors between @xmath54 and @xmath109 is @xmath110 , while the multiset of colors between @xmath91 and @xmath109 is @xmath105 .
therefore , while case @xmath77 holds , we can select the subset of vertices @xmath111 such that the number of edges with colors from @xmath112 between @xmath91 and @xmath93 is more than the number of edges with colors from @xmath112 between @xmath54 and @xmath93 , and from this set , we can select an appropriate subset @xmath113 .
since there are finite number of vertices in @xmath114 , for some @xmath115 , case @xmath34 or @xmath76 must hold ( see also figure [ fig : iteration ] ) .
let @xmath93 be the vertex for which case @xmath34 or @xmath76 holds , and let @xmath116 be the color such that there is an edge with color @xmath116 between @xmath93 and @xmath91 and there is no such edge between @xmath54 and @xmath93 .
remove that edge between @xmath91 and @xmath93 and add between @xmath54 and @xmath93 .
let @xmath117 be a vertex such that there is an edge between @xmath54 and @xmath118 with color @xmath116 .
remove that edge and add between @xmath91 and @xmath118 .
let the color between @xmath91 and @xmath118 be @xmath119 .
remove it and add between @xmath54 and @xmath118 .
( note that due to the definition of @xmath94 and due to case @xmath77 held in the @xmath120st iteration , before the change , there was no @xmath119 colored edge between @xmath54 and @xmath118 , thus after the change , all edges between @xmath54 and @xmath118 have different colors . ) iterate this process ; in the @xmath121th iteration , let @xmath122 be a vertex in @xmath123 such that there is an edge with color @xmath124 between @xmath54 and @xmath122 . remove that edge and add between @xmath91 and @xmath122 , remove the edge with color @xmath125 between @xmath91 and @xmath122 and add between @xmath54 and @xmath122 .
finally , remove the edge with color @xmath96 between @xmath54 and @xmath86 and add it between @xmath91 and @xmath86 .
let this new graph be @xmath126 . then in @xmath126 , @xmath63 decreases by 1
, @xmath98 remains the same .
for each @xmath127 , neither @xmath128 nor @xmath129 is changed since the total number of edges between them is not changed . finally , @xmath130 is not increased .
altogether , @xmath131 is at least @xmath34 less than @xmath101 . on figure
[ fig : iteration ] , it is easy to verify that all vertices participating in the modification of @xmath10 , the same colored edges were removed and added .
therefore , @xmath126 is still the union of realizations of the prescribed degree sequences .
since the exceed number is a non - negative finite integer , after a finite number of steps , the exceed number will be @xmath32 , thus we get @xmath69 as an @xmath22-factorization .
it is clear that one of the colors might encode `` non - edge '' , namely , we can delete those edges and get a realization of a half - regular degree matrix obtained by deleting one row from @xmath22 .
therefore , the following theorem also holds .
[ theo : extended halfregular ] let @xmath68 be a half - regular bipartite degree matrix
. then @xmath22 has a realization iff 1 .
@xmath70 , @xmath71 2 .
@xmath132 3 .
@xmath73 , @xmath133 .
it is clear that equality in condition 2 holds iff equality in condition 3 holds for all @xmath115 . in case of equality , we get back theorem [ the : regular - factorization ] . in case of inequality , let @xmath134 and let @xmath135 extend @xmath22 with this @xmath136st row , and this matrix will satisfy the conditions of theorem [ the : regular - factorization ] . find a realization of this extended matrix , and delete the edges with the @xmath136st color , thus obtain a realization of @xmath22 . in the following section , we consider the solution space of realizations of half - regular degree matrices .
since there is no difference of realizations of half - regular degree matrices and @xmath22-factorizations of the complete bipartite graph @xmath69 , we will consider this later , namely , we consider non - edges as @xmath136st colors when non - edges exist .
in this section , we give necessary and sufficient perturbations to transform any realization of a half - regular bipartite degree matrix , @xmath22 , into another realization of @xmath22 .
first , we extend the space of graphs on which perturbations are applied .
we show how to transform a realization of @xmath22 into another realization in this extended space , then we show how to transform realizations into each other remaining in the space of realizations .
the concept is very similar to the concept applied in paper @xcite , but different perturbations are necessary to temporarily extend the space of graphs .
first , we introduce the necessary definition , the @xmath137 deficiency .
we will extend the space of graphs that have at most @xmath77 vertices with deficiency .
we show how to transform @xmath138 , a realization of @xmath22 into another realization @xmath139 , via graphs having at most @xmath77 vertices with deficiency . in doing so
, we first arrange the colored edges of a vertex @xmath140 that they agree with realization @xmath139 .
then we fix these edges , and reduce the problem to a similar , smaller problem .
note that @xmath1 is the regular class , and if we fix ( technically : remove ) the vertex @xmath141 together with its edges , the remaining graph is still half - regular .
finally , we prove how to transform realizations of @xmath22 into each other remaining in the space of realizations of @xmath22 .
first , we define deficiency .
let @xmath22 be a bipartite degree matrix , and let @xmath10 be an edge colored bipartite graph .
we say that a vertex @xmath17 in @xmath10 has a @xmath142-deficiency w.r.t .
@xmath22 if the number of its @xmath143 colored edges is @xmath144 , the number of its @xmath145 colored edges is @xmath146 and for all other @xmath147 , the number of @xmath15 colored edges of @xmath17 is @xmath18 .
then we define an auxiliary graph that we use several times .
[ def : auxiliaryk ] let @xmath0 be an edge colored bipartite graph in which the edges are colored with colors @xmath148 , and let @xmath54 and @xmath91 be two vertices in @xmath1 . then the directed , edge labeled multigraph @xmath149 is defined in the following way .
the vertices of @xmath150 are the colors @xmath148 , and for each @xmath151 , there is an edge going from @xmath15 to @xmath152 , where @xmath15 is the color of the edge between @xmath54 and @xmath86 and @xmath152 is the color of the edge between @xmath91 and @xmath86 . such edge is labeled with @xmath86 .
the next two lemmas show how to handle graphs with a small amount of deficiency .
[ lem : broken - permutation ] let @xmath10 be such an edge colored simple graph of @xmath69 which is almost a factorization of a half - regular bipartite degree matrix @xmath12 in the following sense : 1 . for each color @xmath15 and vertex @xmath39 ,
the number of edges with color @xmath15 is @xmath153 .
2 . for each color @xmath15 and vertex
@xmath17 the number of edges with color @xmath15 is @xmath18 except for two colors @xmath143 and @xmath145 and two vertices @xmath154 and @xmath155 , where @xmath154 has a @xmath142-deficiency and @xmath155 has a @xmath156-deficiency .
then there exists a perturbation of @xmath10 that affects only edges of @xmath154 and @xmath155 and transforms @xmath10 into a realization of @xmath22 .
consider @xmath157 .
for each vertex @xmath15 , @xmath158 of the graph @xmath157 , the number of incoming and outgoing edges are the same , while @xmath143 has two more outgoing edges than incoming and @xmath145 has two more incoming edges than outgoing .
therefore , there is a trail from @xmath143 to @xmath145 due to the pigeonhole principle . for each edge @xmath159 labeled by @xmath160 along the trail , swap the corresponding edges in @xmath10 , namely , the edge between @xmath154 and @xmath160 and the edge between @xmath155 and @xmath160 .
this transforms @xmath10 into a realization of @xmath22 .
indeed , @xmath154 will have one less edge with color @xmath143 and one more edge with color @xmath145 while the effect on @xmath155 is the opposite .
the number of edges with other colors are not affected , since in the trail , the number of incoming and outgoing edges are the same for all colors not @xmath143 and not @xmath145 , and swapping the edges in @xmath10 is equivalent with inverting the direction of the corresponding edges in @xmath157 .
[ lem : broken - permutation - ii ] let @xmath10 be such an edge colored simple graph of @xmath69 which is almost a factorization of a half - regular bipartite degree matrix @xmath12 in the following sense 1 . for each color @xmath15 and vertex @xmath39 ,
the number of edges with color @xmath15 is @xmath153 .
2 . for each color @xmath15 and vertex @xmath17 , the number of edges with color @xmath15 is @xmath18 except for three colors @xmath161 and @xmath162 and three vertices @xmath154 , @xmath155 and @xmath163 , for which @xmath154 has @xmath142-deficiency
, @xmath155 has @xmath164-deficiency and @xmath163 has @xmath165-deficiency . then there exists a perturbation of @xmath10 that affects only edges of @xmath166 and @xmath167 and transforms @xmath10 into an edge colored simple graph of @xmath69 satisfying conditions 1 and 2 in lemma [ lem : broken - permutation ] .
consider the graph @xmath168 .
for each vertex @xmath15 , @xmath169 of the graph @xmath168 , the number of incoming and outgoing edges are the same , while @xmath162 has two more outgoing edges than incoming and @xmath143 and @xmath145 has one - one more incoming edges than outgoing .
therefore there is a trail from vertex @xmath162 to @xmath145 ( and also to @xmath143 ) due to the pigeonhole principle .
take a trail from @xmath162 to @xmath143 , and for each edge along the trail , swap the corresponding edges in @xmath10 , namely , the edge between @xmath155 and @xmath160 and the edge between @xmath163 and @xmath160 .
this transforms @xmath10 into a graph satisfying conditions 1 and 2 in lemma [ lem : broken - permutation ] .
indeed , @xmath170 will have one less edge with color @xmath162 and one more edge with color @xmath143 , namely , the transformation cancels its deficiency .
vertex @xmath171 will have one more edge with color @xmath162 and one less edge with color @xmath143 , therefore its @xmath172 deficiency becomes a @xmath173 deficiency .
the number of edges with other colors are not affected , since in the trail , the number of incoming and outgoing edges are the same for all colors not @xmath143 and not @xmath162 , and swapping the edges in @xmath10 is equivalent with inverting the direction of the corresponding edges in @xmath168 .
the following lemma is the key lemma in transforming a realization into another realization . as we mentioned above , the strategy is to transform a realization @xmath138 into an intermediate realization @xmath174 , such that the colors of edges of a vertex @xmath141 in the regular class agrees with the colors of the edges of @xmath141 in the target realization @xmath139 .
we have to permute the edges , and the basic ingredient of a permutation is a cyclic permutation . the following lemma shows how to perturb a realization along a cyclic permutation .
[ lem : single row perturbations ] let @xmath138 and @xmath139 be two realizations of the same half - regular bipartite degree matrix @xmath22 .
let @xmath94 be a subset of vertices such that for some @xmath175 , each possible colors appears at most once on the edges between @xmath54 and @xmath94 in @xmath138 , furthermore , there exists a cyclic permutation @xmath46 on @xmath94 such that for all @xmath176 , the color between @xmath54 and @xmath86 in @xmath138 is the color between @xmath54 and @xmath177 in @xmath139 . then there exists a sequence of colored graphs @xmath178 with the following properties 1 . for all @xmath179
, @xmath180 is a colored graph satisfying either the properties 1 and 2 in lemma [ lem : broken - permutation ] or the properties 1 and 2 in lemma [ lem : broken - permutation - ii ] .
2 . for all @xmath181
, a perturbation exists that transforms @xmath180 into @xmath182 and perturbs only the edges of two vertices in @xmath1 .
@xmath183 is a realization of @xmath22 such that for all @xmath95 , the color of the edge between @xmath54 and @xmath93 is the color between @xmath54 and @xmath93 in @xmath139 , and for all @xmath184 , the color between @xmath54 and @xmath86 is the color between @xmath54 and @xmath86 in @xmath138 .
let @xmath185 denote the cyclic permutation , and for all @xmath186 , let @xmath187 be the color of the edge between @xmath54 and @xmath188 in @xmath138 .
since @xmath139 is a realization of @xmath22 , there is a @xmath91 such that the color between @xmath91 and @xmath189 is @xmath190 .
indeed , the color between @xmath54 and @xmath189 in @xmath139 is @xmath190 ( the permutation @xmath46 moves @xmath191 to @xmath189 ) , thus , @xmath189 has an edge with color @xmath190 .
swap the edges between @xmath54 and @xmath189 and between @xmath91 and @xmath189 .
this will be @xmath192 , which satisfies the conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation ] .
indeed , @xmath54 has a @xmath193 deficiency , while @xmath91 has a @xmath194 deficiency .
clearly , @xmath195 and @xmath192 differ only on edges of @xmath54 and @xmath91 , furthermore , all edges of @xmath54 has the same color than in @xmath138 except the edge between @xmath54 and @xmath189 .
assume that some @xmath196 is achieved for which conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation ] are satisfied with @xmath54 having @xmath197 deficiency and with some @xmath91 having @xmath198 deficiency and for all vertices @xmath199 , the edges between @xmath54 and these vertices have a color as in @xmath139 .
furthermore , all other edges between @xmath54 and @xmath200 did not change color .
the color between @xmath54 and @xmath201 in @xmath139 is @xmath202 , therefore , there is a @xmath203 such that the color between @xmath203 and @xmath201 is @xmath202 in @xmath196 .
swap the edges between @xmath54 and @xmath201 and between @xmath203 and @xmath201 .
this will be the graph @xmath204 .
clearly , @xmath196 and @xmath204 differ only on edges of @xmath54 and @xmath203 . if @xmath205 , then @xmath54 has @xmath206 deficiency , and @xmath91 has @xmath207 deficiency , hence @xmath204 satisfies conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation ]
. we can rename @xmath204 to @xmath196 and iterate the chain of transformations . if @xmath208 , then @xmath54 has @xmath206 deficiency , @xmath203 has @xmath209 deficiency and @xmath91 has @xmath210 deficiency .
thus , the conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation - ii ] hold .
due to lemma [ lem : broken - permutation - ii ] , there exists a perturbation that affects only edges on @xmath91 and @xmath203 and transforms @xmath204 to an @xmath211 for which conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation ] hold with @xmath54 having @xmath206 deficiency and @xmath203 having @xmath207 deficiency
. we can rename @xmath211 to @xmath196 and iterate the chain of transformations . with this series of transformations
, we can reach @xmath196 wich satisfies conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation ] with @xmath54 having @xmath212 deficiency and with some @xmath91 having @xmath213 deficiency , furthermore , all vertices @xmath214 , the edges between @xmath54 and these vertices have a color as in @xmath139 , while all other edges of @xmath54 did not change color . there is a @xmath203 such that the color of the edge between @xmath203 and @xmath191 is @xmath215 .
@xmath196 is transformed into @xmath204 by swapping the edges between @xmath54 and @xmath191 and between @xmath203 and @xmath216 . if @xmath205 , then this transformation leads to a realization of @xmath22 , and we are ready , namely , @xmath217 .
otherwise , @xmath204 satisfies the conditions @xmath34 and @xmath76 in lemma [ lem : broken - permutation ] with @xmath91 having @xmath213 deficiency and @xmath203 having @xmath212 deficiency . according to lemma [ lem : broken - permutation ] , @xmath204 can be transformed into @xmath183 with a perturbation affecting only edges on @xmath91 and @xmath203 .
since @xmath183 is also a realization of @xmath22 , it is desirable to have a series of transformation from @xmath138 to @xmath183 such that all the intermediate graphs are realizations of @xmath22 .
to do this , we need a slightly larger perturbation modifying the edges of three vertices in the regular vertex set , as stated in the following lemma . [
lem : one cycle proper perturbations ] let @xmath138 , @xmath139 , @xmath94 and @xmath46 the same as in lemma [ lem : single row perturbations ] .
then there exists a sequence of colored graphs @xmath218 with the following properties 1 . for all @xmath219
, @xmath220 is a realization of @xmath22 .
2 . for all @xmath221
, a perturbation exists that transforms @xmath220 into @xmath222 and perturbs only the edges of at most three vertices in @xmath1 .
3 . @xmath223 is a realization of @xmath22 such that for all @xmath95 , the color of the edge between @xmath54 and @xmath93 is the color between @xmath54 and @xmath93 in @xmath139 , and for all @xmath184 , the color between @xmath54 and @xmath86 is the color between @xmath54 and @xmath86 in @xmath138 .
consider the series of colored graphs @xmath224 obtained in lemma [ lem : single row perturbations ] .
find the largest @xmath225 satisfying @xmath226 , @xmath227 has two vertices with a deficiency . due to the construction , for all @xmath228 , these two vertices are the same , and @xmath138 and @xmath227 differ only in edges of these two vertices , @xmath54 and @xmath91 .
if @xmath229 , then @xmath138 and @xmath183 differ only on edges of two vertices of @xmath1 , thus @xmath230 will suffice .
otherwise , @xmath231 has deficiency on @xmath77 vertices , @xmath54 , @xmath91 and @xmath203 .
@xmath232 has deficiency on @xmath76 vertices , @xmath54 and @xmath203 , and @xmath138 differ from @xmath232 on edges of three vertices , @xmath54 , @xmath91 and @xmath203 .
apply lemma [ lem : broken - permutation ] on @xmath232 , the so - obtained graph will be @xmath233 .
@xmath233 is a realization of @xmath22 and differ from @xmath138 only on edges of 3 vertices in @xmath1 .
iterate this construction , find the largest @xmath234 such that @xmath235 , @xmath227 has two vertices with deficiency . if @xmath236 , then @xmath237 , and @xmath238 .
otherwise , @xmath239 has 3 vertices with deficiency , and differ from @xmath240 only on the edges of these @xmath77 vertices .
@xmath241 has two vertices with deficiency , on which lemma [ lem : broken - permutation ] is applied to get @xmath242 .
@xmath242 differs from @xmath240 only on edges of @xmath77 vertices , since @xmath240 differs from @xmath243 on the edges of the same 2 vertices that have deficiency for all @xmath227 , @xmath244 .
after a finite number of iterations , @xmath236 , and then @xmath245 , @xmath238 , and this finishes the construction .
[ theo : full perturbations ] let @xmath138 and @xmath139 be realizations of the same half - regular bipartite degree matrix @xmath22 .
then there exists a sequence of colored graphs @xmath246 with the following properties . 1 . for all @xmath247
, @xmath180 is a realization of @xmath22 .
2 . for all @xmath181
, a perturbation exists that transforms @xmath180 into @xmath182 and perturbs only the edges of at most three vertices in @xmath1 .
consider a vertex @xmath54 in the regular vertex class @xmath1 , and define the following directed multigraph @xmath150 .
the vertices of @xmath150 are the colors @xmath148 and the edges are defined by the following way . for each vertex @xmath151
, there is an edge going from @xmath15 to @xmath152 , where @xmath15 is the color of the edge between @xmath54 and @xmath86 in @xmath139 and @xmath152 is the color of the edge between @xmath54 and @xmath86 in @xmath138 ( if @xmath248 , then this edge will be a loop ) .
label this edge with vertex @xmath86 .
since @xmath138 and @xmath139 are realizations of the same degree matrix @xmath22 , @xmath150 is eulerian and thus , can be decomposed into directed cycles .
if all cycles are trivial ( loops ) , then for all vertex @xmath151 , the color of the edge between @xmath54 and @xmath86 in @xmath138 and @xmath139 are the same .
if there is a nontrivial cycle @xmath249 , then each color appears in this cycle at most once , and the edges of a cycle define a cyclic permutation on a subset of vertices @xmath61 .
let @xmath185 denote this cyclic permutation @xmath46 . by the definition of @xmath150 ,
the color between @xmath54 and @xmath86 in @xmath138 is the color between @xmath54 and @xmath177 in @xmath139 , for all @xmath250 .
therefore , we can apply lemma [ lem : one cycle proper perturbations ] to transform @xmath138 into @xmath223 . now
if we construct the same directed multigraph @xmath150 considering @xmath223 , @xmath139 and the same vertex @xmath54 , then cycle @xmath249 becomes separated loops for all of its vertices , while other edges are not affected ( recall the second half of condition 3 in lemma [ lem : single row perturbations ] , `` _
_ for all @xmath184 , the color between @xmath54 and @xmath86 is the color between @xmath54 and @xmath86 in @xmath138 . _ _ '' ) .
while there are non - trivial cycles in @xmath150 , we can apply lemma [ lem : one cycle proper perturbations ] , and in a finite number of steps , @xmath138 is transformed into a realization @xmath251 such that for all vertices @xmath151 , the color of the edge between @xmath54 and @xmath86 in @xmath251 and @xmath139 are the same . furthermore ,
in all steps along the transformation , the intermediate graphs satisfy the two conditions of the lemma .
fix the edges of @xmath54 both in @xmath251 and @xmath139 .
technically , this can be done by deleting vertex @xmath54 and its corresponding edges both from @xmath251 and @xmath139 .
then these reduced graphs will be realizations of the same half - regular degree matrix that can be obtained from @xmath12 by deleting a column from the row - regular @xmath11 ( which thus remains row - regular ) and modifying some values of @xmath252 according to the edge colors of @xmath54 . on these reduced graphs
, we can consider a vertex @xmath54 from the regular vertex class @xmath1 , and do the same .
once there is one remaining vertex in @xmath1 , the two realizations will be the same , and thus , we transformed @xmath138 into @xmath139 .
theorem [ theo : full perturbations ] says that perturbing the edges of at most three vertices in the regular vertex class is sufficient to connect the space of realizations of a half - regular degree matrix @xmath22 , namely , a finite series of such perturbations is sufficient to transform any realization into another such that all intermediate perturbed graphs are also realizations of @xmath22 .
a natural question is if such perturbations are also necessary .
the answer to this question is yes .
latin squares can be considered as 1-factorizations of the complete bipartite graph @xmath253 .
the corresponding degree matrix @xmath12 satisfies the definition of half - regular degree matrices .
indeed , both @xmath11 and @xmath252 are @xmath254 , all-1 matrices .
it is well - known that already @xmath255 latin squares can not be transformed into each other via latin squares if only two rows are perturbed in a step , and the same claim holds for any @xmath256 latin squares , where @xmath257 is prime and @xmath258 @xcite .
in this section , we give a markov chain monte carlo ( mcmc ) method for sampling realizations of a half - regular degree matrix @xmath22 . theorem [ theo : full perturbations ]
says that the perturbations that change the edges of at most 3 vertices of the regular vertex class are irreducible on the realizations of half - regular degree matrices .
a nave approach would make a random perturbation affecting the edges of @xmath77 vertices , and would accept this random perturbation if it is a realization of @xmath22 .
unfortunately , the probability that such random perturbation would generate a realization of @xmath22 would tend to 0 exponentially quickly with the size of @xmath22 , making the mcmc approach very inefficient .
indeed , there were an exponential waiting time for an acceptance event , that is , when the random perturbation generates a realization of @xmath22 .
even if the random perturbation generates a realization of @xmath22 with probability 1 , the metropolis - hastings algorithm applying such random perturbation might generate a torpidly mixing markov chain if the acceptance ratios are small .
an example when this is the case can be found in @xcite . in this section
, we design a transition kernel that generates such perturbations and its transition probabilities satisfy equation [ eq : highratio ] , namely , the inverse of the acceptance ratio bounded by a polynomial function of the size of @xmath22 . to do this , we first have to generate random circuits and trails in auxiliary graphs that we define below .
[ def : trail ] let @xmath149 be the directed , edge labeled multigraph of the edge colored bipartite graph @xmath10 as defined in definition [ def : auxiliaryk ] .
let @xmath259 and @xmath260 be two vertices of @xmath150 such that for any vertex @xmath261 of @xmath150 , the number of outgoing edges of @xmath262 is greater or equal than the number of its incoming edges , and @xmath259 has more outgoing edges than incoming edges if @xmath263 .
we define the following function @xmath264 that generates a random trail from @xmath259 to @xmath260 when @xmath263 and a random circuit when @xmath265 . 1 .
select uniformly a random outgoing edge @xmath266 of @xmath259 which is not a loop .
let the sequence of labels @xmath267 , where @xmath86 is the label of @xmath266 , namely , @xmath268 be a sequence containing one label .
let @xmath262 be the endpoint of @xmath266 .
2 . while @xmath269 , select uniformly a random outgoing edge @xmath266 of @xmath262 which is not a loop and does not appear in @xmath268 .
extend @xmath268 with the label of @xmath266 , and set @xmath262 to the endpoint of @xmath266 .
note that due to the pigeonhole rule , this random procedure will eventually arrive to @xmath260 , since for any vertex @xmath269 reached by the algorithm , there exist at least one outgoing edge of @xmath262 which is not a loop and is not in the sequence @xmath268 .
the probability of a trail @xmath270 or circuit @xmath271 generated by @xmath272 consists of the product of the inverses of the available edges going out from @xmath273 .
these probabilities will be denoted by @xmath274 and @xmath275 .
we are going to perturb the graphs along the circuits and trails as defined below .
let @xmath276 ( resp . ,
@xmath277 ) be a random circuit ( resp .
, random trail ) generated by @xmath278 .
then the transformation @xmath279 ( @xmath280 ) swaps the colors of the edges between @xmath54 and @xmath86 and between @xmath91 and @xmath86 for all edge labels @xmath86 in @xmath276 ( @xmath277 ) . with these types of perturbations
, we are going to define the random algorithm that generates a random perturbation . [
alg : perturbation ] let @xmath0 be a realization of the half - regular degree matrix @xmath22 .
do the following random perturbation on @xmath10 .
a. with probability @xmath281 , do nothing .
this is technically necessary for the markov chain surely be aperiodic .
any constant probability would suffice , @xmath281 is the standard choice for further technical reasons not detailed here , see for example , @xcite .
b. with probability @xmath282 , do the following 1 .
draw uniformly a random ordered pair of vertices @xmath54 and @xmath91 from the vertex set @xmath1 . construct the auxiliary graph @xmath283 ,
draw uniformly a random color @xmath284 and draw a random circuit @xmath276 using @xmath285 .
2 . chose a random integer @xmath121 uniformly from @xmath286 $ ] .
if @xmath287 , then let @xmath126 be @xmath279 .
@xmath126 is the perturbed graph , exit the algorithm .
if @xmath288 , do the following ( see also figure [ fig : perturbation1 ] ) .
let @xmath289 denote the edges of circuit @xmath276 starting with the edge going out from @xmath284 .
let @xmath277 be the trail with edges @xmath290 .
perturb @xmath10 to @xmath280 .
draw uniformly a @xmath203 from @xmath291 and also uniformly a vertex @xmath86 from the subset of @xmath61 satisfying the condition that the color of the edge between @xmath54 and @xmath86 in @xmath280 is the last color in the trail @xmath277 .
let @xmath292 denote the color of the edge between @xmath203 and @xmath86 .
swap the colors of the edges between @xmath54 and @xmath86 and between @xmath203 and @xmath86 .
note that @xmath292 might equal to @xmath262 , and in that case , swapping the colors has no effect .
denote the so far perturbed graph @xmath293 .
4 . construct @xmath294 , and draw a random trail @xmath295 using function @xmath296 .
let @xmath297 be @xmath298 .
5 . construct @xmath299 , and draw a random trail @xmath300 using function @xmath301 .
let @xmath126 , the perturbed graph be @xmath302 .
c. with probability @xmath282 , do the following .
draw uniformly a random ordered pair of vertices @xmath54 and @xmath203 from the vertex set @xmath1 .
construct @xmath303 , draw uniformly a random color @xmath284 , and draw a random circuit @xmath304 using @xmath305 .
2 . draw uniformly a random integer from @xmath286 $ ] .
if @xmath306 , then let @xmath126 be @xmath307 .
@xmath126 is the perturbed graph , exit the algorithm .
3 . if @xmath308 , do the following ( see also figure [ fig : perturbation2 ] ) .
let @xmath309 be the trail from @xmath284 to @xmath310 , and let @xmath311 , and let @xmath292 denote @xmath310 . draw uniformly a random vertex @xmath91 from @xmath312 .
build @xmath313 , and draw a random trail @xmath314 using @xmath315 .
let @xmath316 $ ] denote the prefix of the trail @xmath314 containing only the first @xmath33 edges of the trail .
let @xmath317 denote the @xmath33th edge of trail @xmath314 .
construct the set of vertices @xmath94 which contains all vertex @xmath38 such that the label of @xmath317 is @xmath38 and in the graph @xmath318 , u " , u')$ ] , there is a vertex @xmath86 such that the color of the edge between @xmath54 and @xmath86 is @xmath292 , the color of the edge between @xmath203 and @xmath86 is the color of the edge between @xmath91 and @xmath38 , furthermore , @xmath262 is the first occurrence in the trail , that is , @xmath319 $ ] does not contain @xmath262 .
4 . if @xmath94 is empty , then exit the algorithm , let the perturbed graph be the original graph . otherwise , draw a random vertex @xmath86 from @xmath94 .
shorten @xmath314 such that the last edge have label @xmath86 .
let @xmath320 denote this trail .
let @xmath293 be @xmath321 . if @xmath203 has a deficiency @xmath322 for some @xmath262 in @xmath293 , then draw uniformly a vertex among the vertices @xmath323 for which the color of the edge between @xmath54 and @xmath323 is @xmath292 and the color of the edge between @xmath203 and @xmath323 is @xmath262 .
modify @xmath293 by swapping the colors of these two edges .
+ if @xmath203 has no deficiency in @xmath293 then recall @xmath292 to @xmath262 .
5 . build @xmath324 , and draw a random trail @xmath325 using @xmath326 .
let @xmath126 , the perturbed graph be @xmath327 .
[ theo : perturbation ] the random perturbation in algorithm [ alg : perturbation ] has the following properties . 1 .
the generated @xmath126 is a realization of @xmath22 , thus this random perturbation is the transition kernel of a markov chain on the realizations of @xmath22 .
the perturbations are irreducible on the realizations of @xmath22 3 .
the perturbations form a reversible kernel and for any @xmath10 and any @xmath126 generated from @xmath10 , @xmath328 where @xmath60 is the number of vertices in @xmath61 . furthermore
, if this reversible kernel is used in a metropolis - hastings algorithm , the expected waiting time ( that is , how many markov chain steps is necessary to leave the current state ) at any state is bounded by @xmath329 . 1 . when @xmath287 or @xmath330 ( cases ii.(b ) and iii.(b ) in algorithm [ alg : perturbation ] ) , the constructed @xmath126 differs on edges of two vertices of @xmath1 , @xmath54 and @xmath91 ( or @xmath54 and @xmath203 ) .
since the swapped colors are along a circuit in the auxiliary graph @xmath150 , the effects of color swaps cancel each other , and we get a realization of @xmath22 .
+ when @xmath288 ( case ii.(c)-(e ) in algorithm [ alg : perturbation ] ) , @xmath10 is transformed into an intermediate graph @xmath293 , which has deficiency on three vertices @xmath54 , @xmath91 and @xmath203 .
vertex @xmath54 has @xmath331-deficiency , @xmath91 has @xmath332-deficiency , and @xmath203 has @xmath322-deficiency ( see also the boxed colors on figure [ fig : perturbation1 ] ) .
the generated random trail @xmath295 on @xmath333 goes from @xmath284 to @xmath262 . in lemma
[ lem : broken - permutation - ii ] , we proved that any such trail transforms @xmath293 into a graph satisfying the conditions of lemma [ lem : broken - permutation ] .
the random trail @xmath300 generated in @xmath334 is from @xmath292 to @xmath284 , and lemma [ lem : broken - permutation ] proves that swapping the edge colors along such trail transforms the graph into a realization of @xmath22 .
+ when @xmath308 ( case iii.(c)-(e ) in algorithm [ alg : perturbation ] ) , @xmath10 is transformed into a @xmath297 , in which @xmath54 has a @xmath331-deficiency and @xmath203 has a @xmath335 deficiency ( see also the boxed colors on figure [ fig : perturbation2 ] ) .
then @xmath297 transformed into a @xmath293 , in which two cases is possible .
either @xmath203 has a @xmath322-deficiency and then @xmath54 has a @xmath331-deficiency and @xmath91 has a @xmath332-deficiency or @xmath203 has no deficiency and then @xmath54 has a @xmath331-deficiency and @xmath91 has a @xmath335-deficiency . in the former case , a couple of colors @xmath262 and @xmath292 are swapped between @xmath54 and @xmath203 causing @xmath203 has no deficiency and @xmath54 has @xmath336 deficiency , which are cancelled by the perturbation along the trail @xmath325 , thus arriving to a realization of @xmath22 . in the later case
, the perturbation along the trail @xmath325 directly leads to a realization of @xmath22 .
it is sufficient to show that the algorithm generates the perturbations appear in theorem [ theo : full perturbations ] , which are actually the perturbations appear in lemma [ lem : one cycle proper perturbations ] .
there are two types of perturbations in lemma [ lem : one cycle proper perturbations ] .
the simpler one swaps the colors of edges along a cycle .
this happens when @xmath229 in the proof of lemma [ lem : one cycle proper perturbations ] and also when @xmath236 . since a cycle is a circuit , and any circuit in the auxiliary graph @xmath150 constructed in the algorithm can be generated with non - zero probability , this type of perturbation is in the transition kernel .
+ the larger perturbation in lemma [ lem : one cycle proper perturbations ] first swap the colors of edges of @xmath54 and @xmath91 along a path , then swap the colors of two edges between @xmath54 and @xmath86 and @xmath203 and @xmath86 for some @xmath86 , then swap the edges of @xmath91 and @xmath203 along a path to eliminate the deficiency of @xmath203 , finally , swap the color of the edges @xmath54 and @xmath203 along a path to eliminate their deficiencies , thus generate a realization .
since any path is a trail , the given algorithm can generate such perturbations when @xmath288 ( case ii.(c)-(e ) in algorithm [ alg : perturbation ] ) .
we show that for any perturbation of @xmath10 to @xmath126 , there exists a perturbation from @xmath126 to @xmath10 , and we also compare the probabilities of the perturbations .
recall that the number of vertices in @xmath1 is @xmath8 , the number of vertices in @xmath61 is @xmath60 , and the number of colors is @xmath337 ; @xmath8 , @xmath60 and @xmath337 will appear in the probabilities below several times .
+ when the perturbation affects the edges of only two vertices , @xmath54 and @xmath91 or @xmath54 and @xmath203 ( cases ii.(b ) or iii.(b ) in algorithm [ alg : perturbation ] ) , the algorithm first draws these vertices with probability @xmath338 , where @xmath339 .
also a random color @xmath284 is drawn with probability @xmath340 .
then a random circuit @xmath341 is generated with probability @xmath274 . finally ,
a random @xmath342 is generated , this happens with @xmath343 probability , where @xmath344 and the color of the edges indicated by the circuit are swapped .
the so - obtained @xmath126 can be generated not only in this way , but an arbitrary such @xmath15 can be generated as the starting point of the circuit which is visited only once in the circuit .
+ to perturb back @xmath126 to @xmath10 , we first have to select the same pair of vertices @xmath54 and @xmath91 , with the same , @xmath338 probability .
then the random circuit @xmath345 must be generated in the constructed auxiliary directed multigraph @xmath346 . since for all color , the auxiliary graph @xmath150 constructed from @xmath10 has the same outgoing edges than the auxiliary graph @xmath346 constructed from @xmath126 , @xmath347 .
furthermore , this claim holds for all possible circuits having the possible starting colors @xmath15 . finally ,
generating the same @xmath121 has the same probability , thus we can conclude that for this type of perturbation , @xmath348 , thus equation [ eq : claim ] holds .
+ when the perturbation affects the edges of three vertices , @xmath54 , @xmath91 and @xmath203 , then @xmath10 is first perturbed to @xmath293 , @xmath293 is perturbed to @xmath297 and finally @xmath297 is perturbed to @xmath126 , or the third type of perturbation is applied in the algorithm where @xmath10 first perturbed into @xmath297 then @xmath293 and finally @xmath126 ( case ii.(c)-(e ) and iii.(c)-(e ) in algorithm [ alg : perturbation ] ) .
we show that these two types of perturbations are inverses of each other in the following sense . for any perturbation generated in case ii.(c)-(e ) in algorithm [ alg : perturbation ]
, its inverse can be generated in case iii.(c)-(e ) in algorithm [ alg : perturbation ] , and vice versa , for any perturbation generated in case iii.(c)-(e ) , its inverse can be generated in case ii.(c)-(e ) .
first we consider the perturbation generated by algorithm [ alg : perturbation ] in case ii.(c)-(e ) . in that case , first the ordered pair of vertices ( @xmath54 , @xmath91 ) is generated with probability @xmath338 and a random color @xmath284 with probability @xmath340 .
a circuit @xmath349 is generated together with a random number @xmath121 with probability @xmath350 .
the generated number @xmath121 must be smaller than @xmath351 , and thus , only the trail @xmath352 is important .
the probability of the trail consists of the product of the inverses of the number of available outgoing edges .
let @xmath275 denote this probability .
the colors along the trail are swapped .
then random @xmath203 is generated from the appropriate set of vertices in @xmath353 with probability @xmath354 and also a vertex @xmath86 is generated with probability @xmath355 where @xmath115 is the index of color @xmath310 .
the colors @xmath356 and @xmath292 are swapped .
therefore , @xmath357 .
then a trail @xmath295 is generated from @xmath292 to @xmath284 in the auxiliary graph @xmath333 , and @xmath293 is transformed to @xmath358 thus , @xmath359 .
finally , a random trail @xmath300 from @xmath292 to @xmath284 is generated in @xmath334 with probability @xmath360 , which is @xmath361 .
since @xmath362 we get that @xmath363 + to transform back @xmath126 to @xmath10 , first @xmath126 should be transformed back to @xmath297 , then @xmath297 back to @xmath293 and finally @xmath293 back to @xmath10 . this can be done in case iii.(c)-(e ) by first drawing the same @xmath54 and @xmath203 , then drawing a @xmath364 which is exactly the inverse of @xmath300 , then the same @xmath91 must be selected and the trail @xmath320 must be the inverse of @xmath295 , the same edges with color @xmath262 and @xmath292 must be swapped , and finally the trail @xmath325 must be the inverse of @xmath277 .
first we show that these trails can be inverses of each other .
+ @xmath277 is a shortening of a circuit with start and end vertex @xmath284 . as such
, it contains @xmath284 only once , but might contain @xmath262 several times .
@xmath365 is a trail from @xmath262 to @xmath284 .
therefore , it can contain @xmath262 several times , but contains @xmath284 only once .
thus , the inverse of any @xmath277 might be a @xmath365 and vice versa .
+ @xmath295 is a trail from @xmath284 to @xmath262 .
therefore , it might contain @xmath284 several times , but contains @xmath262 only once .
@xmath366 is a shortening of a trail from @xmath284 to @xmath292 .
it might contain @xmath284 several times , but can contain @xmath262 only once , due to its definition ( see case iii.(c)-(d ) of algorithm [ alg : perturbation ] ) .
hence , the inverse of any @xmath295 can be a @xmath366 , and vice versa .
+ @xmath300 is a trail from color @xmath292 to @xmath284 .
it might hit @xmath292 several times , but only once @xmath284 .
@xmath364 is a shortening of a circuit with start and end vertex @xmath284 . as such
, it contains @xmath284 only once ( as start and end vertex ) , but might contain @xmath292 several times .
thus the inverse of any @xmath300 might be a @xmath364 and vice versa .
+ we are going to calculate the probability of a random perturbation in case iii.(c)-(e ) .
first , the ordered pair of vertices @xmath54 and @xmath203 should be selected with probability @xmath338 .
the auxiliary directed multigraph @xmath367 is constructed , a random @xmath284 is selected , and a random circuit @xmath368 in @xmath367 is generated .
the probability that the randomly generated @xmath121 is exactly the index of the appropriate occurrence of @xmath292 in the trail @xmath364 is @xmath350 .
then only the trail @xmath369 is interesting .
thus , transforming back @xmath126 to @xmath297 has probability @xmath370 .
@xmath364 in @xmath367 is the inverse of the trail @xmath300 in @xmath334 and @xmath367 and @xmath334 differs in inverting the trail @xmath300 .
therefore @xmath360 and @xmath371 differ in the number of outgoing edges from @xmath284 at the begining of the trail @xmath300 and the number of outgoing edges from @xmath292 at the beginning of the trail @xmath364 .
since the number of outgoing edges might vary between 1 and @xmath60 , the ratio of the two probabilities bounded by @xmath372 + after swapping the colors along the trail @xmath373 , @xmath91 should be randomly generated , it has probability @xmath354 .
a random trail @xmath314 is generated using @xmath374 .
a random @xmath86 from the subset @xmath94 is generated in iii.(d ) of the algorithm .
this has probability @xmath375 .
then the trail @xmath314 is shortened to @xmath320 , and @xmath293 is obtained by transforming @xmath297 along this trail .
+ it is not easy to calculate exactly the probability @xmath376 since the set @xmath94 depends on the generated trail @xmath314 .
however , the size of @xmath94 can not be greater than @xmath60 and lower than @xmath34 , therefore the following inequality holds : + @xmath377 the trail @xmath320 is the inverse trail @xmath295 , therefore , the ratio of their probabilities is also between @xmath350 and @xmath60 .
+ finally , the same edge colors @xmath262 and @xmath292 must be swapped back , and @xmath293 must be transformed back to @xmath10 along the trail @xmath325 .
the probability that the selected @xmath323 in iii.(d ) of the algorithm [ alg : perturbation ] is a particular vertex depends on the number of vertices from which @xmath323 is selected .
therefore this probability is again between @xmath350 and @xmath34 .
altogether , the probability of the backproposal probability is bounded between @xmath378 + comparing equations [ eq : g2gprime ] and [ eq : gprime2 g ] , and also considering that the ratio of the probabilities of the trails and corresponding inverse trails are between @xmath350 and @xmath60 , we get for the ratio of proposal and backproposal probabilities that + @xmath379 + case ii . is chosen with probability @xmath282 . given that case ii . is selected , the probability that algorithm [ alg : perturbation ] generates a realization being different from the current realization is 1 .
if edges of two vertices in @xmath1 are perturbed , then the perturbation is accepted with probability 1 .
if edges of three vertices in @xmath1 are perturbed , then the perturbation is accepted with probability @xmath380 according to equation [ eq : ratio ] .
thus , the probability that the markov chain defined by the metropolis - hastings algorithm does not remain in the same state is greater or equal than @xmath381 .
the expected waiting time is upper bounded by the inverse of this probability , that is , @xmath329 .
we considered the half - regular factorizations of the complete bipartite graph , and we proved that it is a tractable version of the edge packing problem . above the existence theorem , we also give sufficient and necessary perturbations to transform solutions into each other .
when these perturbations are the transition kernel of a markov chain monte carlo method , the inverse of the acceptance ratios are polynomial bounded .
this result might be the first step to prove rapid mixing of the markov chain .
however , proving rapid mixing might be particularly hard .
the speed of convergence of a similar markov chain on latin squares is a twenty years old open problem .
approximate sampling and counting are equally hard computational problems for a large set of counting problems , the so - called self - reducible counting problems @xcite .
latin squares are not self - reducible counting problems , since deleting a few rows from a latin square yield a partial latin square .
half - regular factorizations of the complete bipartite graph are closer to be self - reducible , since deleting a few lines from a half - regular factorization yields another half - regular factorization of a smaller complete bipartite graph .
unfortunately , further technical conditions are necessary ; we have to require that a subset of the colors in the first line be fixed .
it is discussed in @xcite , why this restriction is necessary .
a possible further work could be to give a markov chain which is irreducible on such restricted space .
a proof of rapid mixing of that markov chain would yield to an efficient random approximation ( fpras , see also @xcite ) on the number of realizations of half - regular factorizations of the complete bipartite graph .
since latin squares are also half - regular factorizations , it would also yield an efficient random approximation of latin squares , too . at the moment ,
the known lower and upper bounds on the number latin squares are far away each other @xcite .
finally , a further work might be to relax the half - regularity to a condition where we require that the factors be almost half - regular , that is , for each factor , the degrees in the first vertex class are either @xmath382 or @xmath383 .
chrobak , m. , drr , c. : reconstructing polyatomic structures from discrete x - rays : np - completeness proof for three atoms , _ in : l. brim , j. gruska , j. zlatu@xmath384ka ( eds . ) , mathematical foundations of computer science 1998 , in : lecture notes in computer science , springer , berlin heidelberg _ , * 1450 * ( 1998 ) 185193 .
drr , c. , gu@xmath385ez , f. , matamala , f. m. : reconstructing 3-colored grids from horizontal and vertical projections is np - hard : a solution to the 2-atom problem in discrete tomography , _ siam j. discrete math . _ * 26(1 ) * ( 2012 ) 330352 .
markov , n. t. , ercsey - ravasz , m. , ribiero gomes , a.r . , lamy , c. , vezoli , j. , magrou , l. , misery , p. , falchier , a. , quilodran , r. , sallet , j. , gariel , m.a .
, gamanut , r. , huissoud , c. , clavagnier , s. , giroud , p. , sappey - marinier , d. , barone , p. , dehay , c. , toroczkai , z. , knoblauch , k. , van essen , d.c . ,
kennedy , h. : a weighted and directed interareal connectivity matrix for macaque cerebral cortex , _ cereb .
* 24(1 ) * ( 2014 ) 1736 .
mikls , i. , mlykti , b. , swenson , k. : the metropolized partial importance sampling mcmc mixes slowly on minimum reversal rearrangement paths _ acm / ieee transactions on computational biology and bioinformatics _ , * 4(7 ) * ( 2010 ) 763767 . | we consider a bipartite version of the color degree matrix problem . a bipartite graph @xmath0 is half - regular if all vertices in @xmath1 have the same degree .
we give necessary and sufficient conditions for a bipartite degree matrix ( also known as demand matrix ) to be the color degree matrix of an edge - disjoint union of half - regular graphs .
we also give necessary and sufficient perturbations to transform realizations of a half - regular degree matrix into each other .
based on these perturbations , a markov chain monte carlo method is designed in which the inverse of the acceptance ratios are polynomial bounded .
realizations of a half - regular degree matrix are generalizations of latin squares , and they also appear in applied neuroscience . + degree sequences ; degree matrix ; graph factorization ; edge packing ; latin squares ; markov chain monte carlo |
the solution of the @xmath7 worldsheet string theory is one of the cornerstones of the ads / cft duality program . despite continuous effort and progress on classical aspects in particular @xcite , and
the generally accepted presence of both integrability and conformal invariance symmetries , most aspects of the quantum theory remain elusive .
it is natural to try to understand some aspects of this quantum theory by first tackling simpler models with similar properties .
the so called @xmath0 coset model - specifically , a sigma model on the supersphere @xmath8 - is a very attractive candidate for such an exercise : like the @xmath7 worldsheet theory it is conformal invariant and its target space is a supergroup coset .
of course , it lacks other aspects such as the brst structure of the string theory .
apart from the string theory motivation , models such as the @xmath0 coset model are extremely interesting from the pure conformal field theory point of view .
indeed , they are sigma models which are massless without any kind of topological term , and for a large range of values of the coupling constant @xmath1 . to make things more precise let us briefly remind the reader of some generalities .
supersphere sigma models have target super space the supersphere @xmath9 and can be viewed as a `` supersymmetric '' extension of the nonlinear @xmath10 sigma models ( which differs of course from the usual @xmath10 `` supersymmetric '' models ) .
use as coordinates a real scalar field @xmath11 where the first @xmath12 components are bosons , the last @xmath13 ones fermions , and the invariant bilinear form @xmath14 where @xmath15 is the orthosymplectic metric @xmath16 @xmath17 denoting the identity .
the unit supersphere is defined by the constraint @xmath18 the action of the sigma model ( conventions are that the boltzmann weight is @xmath19 ) reads @xmath20 the perturbative @xmath21 function depends only on @xmath22 to all orders ( see , e.g. , ref @xcite ) , and is the same as the one of the @xmath10 model with @xmath23 .
physics can be reliably understood from the first order beta function @xmath24 the model for @xmath25 positive flows to strong coupling for @xmath26 . like in the ordinary sigma models case ,
the symmetry is restored at large length scales , and the field theory is massive . for @xmath27 meanwhile , the model flows to weak coupling , and the symmetry is spontaneously broken .
one expects this scenario to work for @xmath1 small enough , and the corresponding goldstone phase to be separated from a non perturbative strong coupling phase by a critical point .
the case we are interested in here is @xmath28 , where the @xmath21 function vanishes to all orders in perturbation theory , and the model is expected to be conformal invariant , at least for @xmath1 small enough , the goldstone phase being replaced by a phase with continuously varying exponents not unlike the low temperature kosterlitz thouless phase .
how the group symmetry combines with the ( logarithmic ) conformal symmetry in such models is largely unknown .
it is an essential question to be solved before any serious attempts to understanding universality classes in non interacting disordered 2d electronic systems can be contemplated @xcite .
the @xmath0 coset model was considered in particular in two papers by mann and polchinski using the massless scattering and bethe ansatz approaches .
this is indeed a natural idea , since supersphere sigma models are in general integrable , and , when massive ( ie @xmath26 ) can be described by a scattering theory involving particles in the fundamental representation of the group .
the @xmath29 matrix is well known @xmath30 here , @xmath17 is the identity , @xmath31 is the graded permutation operator , and @xmath32 is proportional to the projector on the identity representation . for @xmath33 arbitrary , factorizability requires that @xmath34 where @xmath35 , while @xmath36 itself is determined , up to cdd factors , by crossing symmetry and unitarity .
one immediately observes that when @xmath37 , the amplitude @xmath36 cancels out , leaving a scattering matrix with a simpler tensorial structure , since the @xmath31 operator disappears .
this corresponds to a particular point @xcite on the sigma model critical line ( where , among other things , the symmetry is enhanced to @xmath38 ) , the rest of which is not directly accessible by this construction
. is special and allows for an extension of the s matrix to the whole @xmath39 critical line .
] the idea used in @xcite is to consider an analytical continuation to @xmath33 real , and an approach to @xmath28 with proper scaling of the mass . though interesting results were obtained , the emphasis in these papers was not on conformal properties .
another line of attack , more suited to the conformal aspects , was launched by read and saleur in 2001 @xcite , who proposed to use a lattice regularization to control the integrable features of the model .
they obtained in this way the spectrum of critical exponents for several related sigma models on super target spaces , including the @xmath0 coset one at a particular ( critical ) value of the coupling @xmath1 .
the results exhibited several mysterious features , including a pattern of large degeneracies , and a set of values of the exponents covering ( modulo integers ) all the rationals .
in two subsequent papers @xcite , it was argued further that many algebraic properties of the conformal field theory could be obtained at the lattice level already .
these include fusion , and the structure of conformal `` towers '' ( see below for further details ) . the work we present in this paper and its companion is an attempt at understanding the conformal field theoretic description of the @xmath0 model for all values of the coupling by using a lattice regularization .
foremost in the lattice approach is the understanding of the algebraic structure of the lattice model - the algebra defined by the local transfer matrices and its commutant . while in the cases discussed in @xcite most necessary results
were already available in the mathematical literature , the situation here is much more complicated : in a few words , we have to deal , instead of the temperley lieb algebra , with the _ brauer algebra _ whose representation theory , in the non semi - simple case , is far from fully understood .
an important part of our work has consisted in filling up the necessary gaps of the literature , sometimes rigorously , but sometimes at the price of some conjectures .
this algebraic work is the subject of the first paper , which we realize might be a bit hard to read for a physics reader .
we capitalize on the algebraic effort in the companion paper , where the boundary conformal field theory for the coset sigma model is analyzed thoroughly . in the second section of this paper
we discuss generalities about lattice regularizations of @xmath10 sigma models in 2 dimensions and define the model we shall be interested in . in section 3 , the transfer matrix , the loop reformulation and the associated brauer algebra are introduced and discussed .
section 4 is the main section , where the full decomposition of the hilbert space of the lattice model under the action of @xmath0 and @xmath3 is obtained .
our main result can be found in eqs . and .
section 5 discusses aspects of the hamiltonian limit and section 6 contains conclusions .
technical aspects of representation theory are discussed further in the appendices . for the reader
s convenience , we provide here a list of notations used throughout the paper : * @xmath40 is the lie superalgebra of the supergroup @xmath41 * @xmath42 is the brauer algebra on @xmath43 strings with fugacity for loops @xmath35 * @xmath44 is the mod 2 graded vector space @xmath45 with even part @xmath46 and odd part @xmath47 .
we shall often drop the indices @xmath48 in @xmath49 . *
@xmath2 is considered as a left @xmath40 and right @xmath42 bimodule * @xmath50 stands for `` @xmath51 is a partition of @xmath43 '' and @xmath52 is the partition @xmath51 transposed * @xmath53 and @xmath54 are the symmetric group on @xmath43 objects and its group algebra * @xmath55 is the temperley lieb algebra with fugacity for loops @xmath56 * @xmath57 and @xmath58 are generic elements of @xmath42 and @xmath41 * @xmath59\,\}$ ] is the set of partitions labeling the weights of @xmath42 .
@xmath60 selects those of them which do realize on @xmath2 * associate weights @xmath61 are labels of @xmath41 irreps which are nonequivalent(identical ) and become(split into two ) isomorphic(nonequivalent ) irreps under the restriction to the proper subgroup @xmath62 of supermatrices with @xmath63 * @xmath64 is the set of hook shape partitions labeling the weights of @xmath40 irreps appearing in @xmath2 and @xmath65 * @xmath66 are standard or generically irreducible representations of @xmath42 * @xmath67 , @xmath68 , @xmath69 , @xmath70 are irreducible representations of @xmath71 , @xmath41 , @xmath42 and @xmath55 respectively . * @xmath72 are direct summands of @xmath2 as a @xmath40 , @xmath41 and @xmath42 module * @xmath73 are the supersymmetric generalization of @xmath10 symmetric functions * @xmath74 are characters of @xmath66 , @xmath69 and @xmath75 .
lattice discretizations of @xmath10 sigma models have had a long history .
the simplest way to go is obviously to introduce spins taking values in the target manifold - the sphere @xmath76 - on the sites of a discrete lattice , with an interaction energy of the heisenberg type @xmath77 ( where @xmath78 stands for the bilinear @xmath10-invariant quadratic form ) .
this is however difficult to study technically , as the number of degrees of freedom on each site is infinite .
a possible way to go is to discretize the target space , leading to various types of `` cubic models '' @xcite .
another way which has proved especially fruitful in two dimensions has been to reformulate the problem of calculating the partition or correlation functions geometrically by using the techniques of high or low temperature expansions , thus obtaining graphs with complicated interaction rules and weights determined by properties of the underlying groups .
the simplest of these formulations appeared in @xcite where the authors studied the @xmath10 model on the honeycomb lattice in two dimensions , and replaced moreover the term @xmath79 by its considerably simpler high temperature approximation @xmath80 , @xmath81 @xcite . expanding
the brackets , in say the calculation of the partition function , one can draw graphs by putting a bond between neighboring sites @xmath82 and @xmath83 whenever the term @xmath84 is picked up .
the integral over spin variables leaves only loops , with a fugacity equal to @xmath85 as there are @xmath85 colors one can contract .
note that because of the very low coordination number of the honeycomb lattice , only self - avoiding loops are obtained .
this leads to the well known self - avoiding loop gas partition function : @xmath86 where the sum is taken over all configurations @xmath87 of self avoiding , mutually avoiding closed loops in number @xmath43 , covering a total of @xmath32 edges .
note that once an expression such as ( [ nien_triv ] ) is written down , it is possible to analytically continue the definition of the model for @xmath85 an arbitrary real number . barring the use of superalgebras , only @xmath85 integer greater or equal to one has a well defined meaning as a spin model ( the case @xmath88 coincides with the ising model is different from @xmath89 because of the additional @xmath90 freedom in choosing the sign of the determinant . ] ) . in two dimensions ,
the mermin wagner theorem prevents spontaneous symmetry breaking , so for @xmath85 integer , critical behavior can only occur for @xmath91 . analysis of the same beta function suggests however that lattice models defined by suitable analytic continuation should have a goldstone low temperature phase for all @xmath92 , though it says nothing about whether this phase might end by a second or first order phase transition . model ( [ nien_triv ] ) lacks interaction terms which would appear with less drastic choices of the lattice and the interactions : these are the terms where the loops intersect , either by going over the same edge , or over the same vertex , maybe many times . it has often been argued that such terms are irrelevant for the study of the critical points of the @xmath10 models in two dimensions .
most of the interest has focused on such critical points for @xmath93 $ ] , which have geometrical applications - in particular the case @xmath94 is related with the physics of self - avoiding walks .
it turns out however that intersection terms are crucial for the understanding of low temperature phases .
indeed , the model ( [ nien_triv ] ) _ does have _ a sort of goldstone phase for @xmath93 $ ] called the dense phase , but its properties are not generic , and destroyed by the introduction of a small amount of intersections . a simple way to see that the dense phase is not generic
is that the exponents at @xmath37 are always those of the kosterlitz thouless transition point : model ( [ nien_triv ] ) does not allow one to enter the low temperature phase of the xy model . also , model ( [ nien_triv ] ) has a first order transition for @xmath95 , which is not the behavior expected from the sigma model analysis .
it was suggested in @xcite that model ( [ nien_triv ] ) can be repaired by allowing for some intersections .
the minimal scenario one can imagine is to define a similar model on the square lattice , and allow for self intersections at vertices only , so either none , one or two loops go through the same vertex .
the resulting objects are often called trails .
this gives the new partition function @xmath96 where the sum is taken over all configurations @xmath97 of closed loops , which visit edges of the lattice at most once , in number @xmath43 , covering a total of @xmath32 edges , with @xmath17 intersections .
the phase diagram of this model has not been entirely investigated .
it is expected that at least for @xmath98 small enough , the critical behavior obtained with @xmath99 is not changed ( though @xmath100 is ) , while the low temperature behavior @xmath101 will be . in @xcite
a yet slightly different version was considered corresponding , roughly , to the limit of very large @xmath102 , where all the edges of the lattice are covered .
the partition function of this fully packed trails model then depends on only two parameters : @xmath103 and an example of allowed configuration is given in figure [ f : loop_cov ] .
numerical and analytical arguments suggest strongly that this model has all the generic properties of the @xmath10 sigma model in the spontaneously broken symmetry phase ( such that one might derive from analytic continuation in @xmath85 of the rg equations ) , for all @xmath104 .
[ f : loop_cov ] ) , contractible ( @xmath105 ) , even ( @xmath32 ) and odd ( @xmath106 ) loops .
periodic imaginary time runs vertically ] note that expression ( [ ours ] ) can be obtained very naturally if instead of putting the degrees of freedom on the vertices , one puts them on the edges of the lattice . in this case ,
the minimal form of interaction involves two edges crossing at one vertex .
invariance under the @xmath10 group allows for three invariant tensors as illustrated on the figure [ f : vert ] , while isotropy and invariance under an overall scale change of the boltzmann weights leaves one with a single free parameter , the crossing weight @xmath98 .
graphical representation of the contractions on the invariant tensors reproduces eq .
( [ ours ] ) , as will be discussed below . for @xmath92 ,
model ( [ ours ] ) flows to weak coupling in the ir , and therefore it is expected that the critical properties of the corresponding low temperature ( goldstone ) phase do not depend on @xmath98 , a fact checked numerically in @xcite .
the case @xmath37 is expected to be different : as mentioned already in the introduction , the beta function of the corresponding sigma model is exactly zero so the coupling constant does not renormalize .
it is indeed easy to see that the loop model ( [ ours ] ) with @xmath37 is equivalent to the 6 vertex model with @xmath107 .
consider the vertices of the 6 vertex model as represented on figure [ f:6v ] .
we chose isotropic weights @xmath108 .
we can decide to split the vertices of a configuration into pieces of oriented loops as represented on figure [ f:6vo2map ] . for each vertex , there are two possible splittings , and we assume that they are chosen with equal probability .
the loops obtained by connecting all the pieces together provide a dense covering of the lattice , and come with two possible orientations , hence a fugacity of two once the orientations are summed over .
the loops can intersect , with a weight @xmath98 given from the obvious correspondence : @xmath109 and thus @xmath110 symmetric six vertex model ] we note that there are indeed _ three _ invariant tensors for the case of @xmath39 .
the corresponding projectors are @xmath111 .
they project respectively on two one dimensional representations , and on a single two dimensional one .
the parameter @xmath112 covers the interval @xmath113 $ ] as @xmath114 $ ] .
changing @xmath112 is well known to change the exponents of 6 vertex model , and therefore eq . ( [ ours ] ) for @xmath37 exhibits a critical line , which is in fact in the universality class of the continuous xy ( @xmath39 ) model in the low temperature ( kosterlitz thouless ) phase .
all what was said so far can be easily generalized to the case of spins taking values on a supersphere @xmath115 .
the fugacity of loops is now equal to @xmath22 : this combination is the number of bosonic minus the number of fermionic coordinates , and follows from the usual fact that when contracting fermions along a loop , a minus sign is generated models to the case of orthosymplectic groups dates back to the work of parisi and sourlas @xcite . ] , see sec .
[ sec:1749 ] .
the loop model formulation therefore provides a convenient graphical representation of the discrete supersphere sigma models for all @xmath22 , in particular @xmath116 where interesting physics is expected to occur .
this physics was explored in @xcite , and the expected results were obtained for @xmath27 . the purpose of this paper is to explore the more challenging @xmath28 case .
of course , at the naive level of partition functions and without worrying about boundary conditions , it looks as if there is no difference between the @xmath10 spin model and its supersphere cousins provided @xmath117 .
the point is that the _ observables _ of the models are different or , at the very least , come with different multiplicities .
indeed , consider for instance correlation functions of spin variables . in the @xmath39 case ,
the spin has only two components @xmath118 , so one can not build a totally antisymmetric tensor on three indices .
this means that the corresponding operator ( which has a nice geometrical interpretation to be given in the next paper ) will not be present in this case , though it will be in the @xmath0 model when @xmath119 .
note that in general , correlators involving spins within the first @xmath12 bosonic and the first @xmath13 fermionic labels will be the same for _ any _ choice of group @xmath120 with @xmath121 and @xmath122 .
bosonic and @xmath123 fermionic degrees of freedom are the same in both cases .
] this is immediately proved by performing a graphical expansion of the correlator : variables outside of the set of the first @xmath12 bosonic and the first @xmath13 fermionic labels are not getting contracted with the spins in the correlators , and cancel against each other in the loop contractions . a standard trick to extract the full operator content of a model is to study the partition function with different boundary conditions .
consider for instance the spin model on an annulus with some symmetry preserving boundary conditions in the space direction . with what we will call periodic boundary conditions ( corresponding to taking the supertrace of the evolution operator ) in the time direction , representations of @xmath0
will always be counted with their superdimension , and the partition function will be identical with the one of the @xmath39 case .
but if we take antiperiodic boundary conditions , we will get a modified partition function ( in the sense of @xcite ) which is a _ trace _ over the hilbert space instead of a supertrace , counts all observables with the multiplicities ( not supermultiplicities ) , and will turn out to be a very complex object
. a good algebraic understanding of the lattice model will be essential to make further progress , and , since the area is largely unexplored , this will occupy us for most of the rest of this first paper .
as discussed briefly in the introduction , the @xmath41 spin model we consider is most easily defined on a square lattice with degrees of freedom ( states ) on the edges and interactions taking place at vertices .
the set of states on every edge is a copy of the base space @xmath124 of the fundamental @xmath41 representation .
interactions at a vertex can be encoded in a local transfer matrix @xmath125 acting on @xmath126 and commuting with the @xmath41 supergroup action .
we call @xmath125 an intertwiner and write @xmath127 .
the boltzmann weights of the model are components of the transfer matrix along a basis of intertwiners .
a natural choice of basis are the projectors onto @xmath41 irreducible representations appearing in the decomposition of the tensor product of two fundamental @xmath41 representations .
to find them one can apply the same ( anti)symmetrization and trace substraction techniques used for reducing @xmath10 tensor representations . if @xmath128 is a mod 2 graded set of basis vectors in @xmath124 with grading @xmath129 , the decomposition of @xmath126 will read @xmath130 here @xmath131 is the @xmath41 invariant tensor , @xmath132 , and @xmath133 ( resp .
@xmath134 ) if @xmath82 is fermionic ( resp .
bosonic ) .
each of the three terms on the l.h.s . of transforms according to an irreps of @xmath41 , or , in other words , belongs to a simple @xmath41 module .
introduce the identity @xmath17 , the graded permutation operator @xmath31 ( also known as braid operator ) , and @xmath32 the temperley lieb operator ( proportional to the projector on the trivial representation ) , @xmath135 in terms of projectors onto irreducible @xmath41 modules , eq .
may be written in a more elegant way as @xmath136 let @xmath137 denote as usual the inversion of space , @xmath138 the inversion of time and @xmath139 the charge conjugation with the matrix @xmath15 .
one can check directly from definition that @xmath31 is @xmath140 , @xmath137 and @xmath141 invariant , while @xmath32 is @xmath142 and @xmath143 invariant .
moreover , @xmath32 and @xmath17 transform into each other under the @xmath144 rotation of the lattice @xmath145 , while @xmath32 and @xmath31 are related by the crossing symmetry @xmath146 @xmath147 take @xmath17 , @xmath32 and @xmath31 as basis of intertwiners in @xmath148 .
the local transfer matrix generally depends on three independent weights @xmath149 , @xmath150 and @xmath151 .
however , on a homogeneous and isotropic lattice one can normalize @xmath152 and leave only the weight @xmath153 .
finally , the local transfer matrix takes the form @xmath154 and @xmath155 represented on a diagonal lattice of width 6 . ] on a diagonal lattice with open boundaries represented in fig . [
f : dl ] choose the time in vertical direction .
the notation of sites at a fixed time is such that the left edge @xmath82 and right edge @xmath156 meet at vertex @xmath82 .
let @xmath157 denote a transfer matrix acting nontrivially only at vertex @xmath82 according to eq . . from the figure it is clear that odd and even times are inequivalent
the transfer matrix @xmath158 , propagating one step forward at equivalent times , may be written as a product @xmath159 of one layer transfer matrices @xmath160}t_{2i},\quad y = \prod_{i=1}^{[l/2]}t_{2i-1},\ ] ] schematically shown in fig .
[ f : dl ] .
the simplest way to define a partition function that depends on the whole spectrum of the transfer matrix @xmath158 is by taking the _ trace _ of @xmath158 at a certain power @xmath21 .
selecting other boundary conditions with some nontrivial symmetry generally amounts to restricting the whole space of states of the model to a subspace compatible with the symmetry of chosen boundary conditions .
what exactly we mean by `` symmetry of boundary conditions '' will be explained later in sec .
[ sec:1749 ] . for the moment
let us just say that it is convenient to consider a more general class of boundary conditions , called quasiperiodic , in which @xmath161 is `` twisted '' by the action of an element @xmath58 of the supergroup . define the quasiperiodic partition function to be @xmath162 we must take the _
supertrace _ in eq .
if we want the quasiperiodic partition function to be well defined .
for instance , when @xmath163 we get the usual trace partition function and when @xmath58 equals to the identity matrix we get the supertrace partition function . + note
that because @xmath58 is a supermatrix , the tensor product in @xmath164 has to be graded , that is @xmath165 after inserting the local transfer matrix from eq . in eq . and
expanding the transfer matrix @xmath158 , the quasiperiodic partition function reads as a sum of weighted products of @xmath166 s and @xmath167 s .
such linearly independent products must be considered as words of a _ transfer matrix algebra _ , while intertwiners @xmath166 and @xmath167 are _ generators _ of this algebra . in the next section
we identify this algebra as a representation of the brauer algebra . for an abstract introduction to the brauer algebra see ref .
@xcite while in the context of @xmath40 centralizer algebra see ref .
we collect in this section some well known facts about the brauer algebra we shall use in the next sections .
let @xmath166 and @xmath167 , @xmath168 act
nontrivially as @xmath32 and @xmath31 in eq . only at the sites @xmath169 of @xmath2 .
one can check that for @xmath167 and @xmath166 so defined the following relations hold : @xmath170 in the second line of these relations @xmath82 and @xmath83 are supposed to be nonadjacent sites . relations ( [ e : def_rel ] ) ( is one of the many ways to ) define the @xmath42 brauer algebra ( also denoted sometimes by the names of braid - monoid algebra or degenerate birman - wenzel - murakami algebra @xcite ) . note that this algebra depends on a single , generally complex , parameter @xmath85 , and contains the maybe more familiar temperley lieb algebra , generated by @xmath166 s alone , and the symmetric group algebra , generated by @xmath167 alone . for @xmath85
fixed and @xmath43 big enough , the @xmath41 spin models provide highly unfaithful representations of the brauer algebra @xmath42 .
this is because , in @xmath2 , the generators @xmath167 and @xmath166 satisfy additional higher order relations @xmath171 on top of .
symmetry in the fundamental representation , and corresponding quotients of the hecke algebra . ] for a simple example , consider the @xmath39 spin model on a lattice of width 3 .
the projector onto the antisymmetric tensor of rank 3 is zero , thus , @xmath171 contains the additional relation @xmath172 .
our spin models in general provide representations of the quotient algebras @xmath173 .
the set of relations @xmath171 can be explicitly described for @xmath174 , see @xcite and references therein , and we have little to say about the case @xmath119 .
the first step in understanding the spectrum of the transfer matrix @xmath158 brings up the question of @xmath42 irreducible representations , and of their multiplicities in @xmath158 for a particular choice of @xmath12 and @xmath29 .
this leads us to discussing some results about the representation theory of the brauer algebra . in @xmath175 . ]
the most natural representation to begin with is the _
adjoint representation_. it admits a diagrammatic representation in terms of graphs on @xmath176 points in which every vertex has degree 1 .
usually one orders the @xmath176 points on two horizontal parallel lines as shown in fig .
[ f:1factor ] .
let @xmath177 denote the vector space spanned on the @xmath178 such diagrams .
the product @xmath179 of two diagrams @xmath180 and @xmath181 is performed by putting @xmath180 on top of @xmath181 and replacing each of the loops in the resulting diagram with @xmath85 .
define diagrammatically the identity @xmath17 and the generators @xmath166 , @xmath167 as represented in fig .
[ f : gen_diag ] .
one can check that the graphical representation of generators with the multiplication @xmath182 of diagrams satisfy all of the eqs . .
the left action of generators on @xmath177 via the multiplication @xmath182 of diagrams provide the adjoint representation of @xmath42 . , @xmath166 and @xmath167 . ] from the graphical representation we see that @xmath177 has a series of invariant subspaces @xmath183 , where @xmath184 is spanned on diagrams with fewer than @xmath185 vertical lines and @xmath186 .
the vector space spanned on diagrams with exactly @xmath185 vertical lines may be defined as a @xmath42 module by the coset @xmath187 .
the left action of @xmath42 on this modules may be seen as a modified multiplication @xmath188 of diagrams @xmath189 under the left action of the algebra the position of horizontal lines in the bottom of a diagram does not change . for a given configuration of the horizontal lines in the top of a diagram and a given pattern of intersections of vertical lines
there are @xmath190 possibilities of choosing the configuration of horizontal lines in the bottom of the diagram .
this simply means that @xmath191 decomposes into a direct sum of @xmath192 equivalent modules .
the coset representative @xmath193 of these equivalent left modules is spanned on @xmath194 graphs on @xmath43 points with every vertex having degree 0 or 1 and a labeling with numbers @xmath195 of free vertices .
an example of such a labeled graph is shown in fig .
[ f : lab_1fact ] .
if the labellings are omitted the resulting graph is called a partial diagram .
the labeling of the @xmath185 free points of a labeled graph is a permutation @xmath196 in the symmetric group @xmath197 .
the labeled graphs will provide a representation of the brauer algebra , which is irreducible for generic values of @xmath85 , if we take the labellings @xmath196 in an irreducible representation of @xmath197 .
we call such representations generically irreducible .
let @xmath198 be a partition of @xmath185 , which we write as @xmath199
. in a more algebraic language the definition of generically irreducible left modules translates to @xmath200 where @xmath201 is an irreducible @xmath197 module . in view of later numerical analysis
we give below a basis in @xmath66 and describe the action of @xmath42 on this basis . .
the labeling permutation is @xmath202 . ]
let @xmath203 denote the labeling of a partial diagram @xmath204 with the permutation @xmath196 , @xmath205 be a set of basis vectors in @xmath201 and @xmath206 be the matrix of the permutation @xmath207 in the representation @xmath208 .
a natural basis in @xmath66 is given by all pairs @xmath209 .
the action of a diagram @xmath210 on a basis vector is @xmath211 where @xmath207 is the labeling of @xmath212 and @xmath213 is the partial diagram @xmath204 labeled with the identity permutation .
the dimensions @xmath214 of @xmath66 is @xmath215 . in simple words ,
a generically irreducible module is a span on graphs on @xmath43 points , obtained by choosing @xmath185 points among @xmath43 , pairing all the others ( this gives the multiplicity @xmath216 since intersections are allowed ) , choosing for the @xmath185 unpaired ones a representation of the permutation group and setting to zero the action of any brauer diagram that reduces the number @xmath185 of unpaired points .
the generically irreducible representations labeled by @xmath217 , @xmath218 $ ] appear in the decomposition of the adjoint representation with multiplicity given by their dimension @xmath214 when @xmath42 is semisimple .
let us conclude with a few words about the reducibility of generically irreducible modules @xmath66 .
for integer @xmath85 and a number of strings @xmath219 the brauer algebra is not semisimple and , as a consequence , certain of the modules @xmath66 become reducible , though they remain _ indecomposable_. the irreducible components appearing in such reducible modules @xmath66 are far from being understood ( the situation is much worse than in the case of the nonsemisimple temperley lieb algebra @xcite ) .
numerical computations based on the diagonalization of the transfer matrix in the diagrammatic representation of @xmath42 restricted to @xmath66 decreases in efficiency very fast with increasing @xmath43 , compared to the ideal case where the transfer matrix is restricted to an irreps of @xmath42 .
this is because for big @xmath43 and @xmath198 fixed the number of irreducible components in @xmath66 `` goes wild '' and there are a lot of `` accidental degeneracies '' in the spectrum of the transfer matrix restricted to @xmath66 .
however , a significant progress in this direction has been recently made in @xcite .
let us note that , as described in @xcite , the content of ( at least some ) @xmath66 s can be computed by repeated applications of frobenius reciprocity applied to the short exact sequence of @xcite describing the structure of the induced modules @xmath220 . in the end we recall the basic results for the temperley lieb algebra , to allow a quick comparison with brauer .
temperley lieb algebra diagrams are a subset of brauer algebra diagrams subject to the constraint that no intersections between edges are allowed .
the dimension of the algebra is given by the catalan numbers @xmath221 .
the main line of reasoning for finding generically irreducible modules follows the same way , except there is no available action of the symmetric group on vertical lines .
therefore , the analogue of the labeled graphs will be the partial diagrams , in which no free points may be trapped inside an edge .
the number of such graphs is @xmath222 , where @xmath223 is the number of edges .
the generically irreducible modules @xmath224 are parametrized by the number @xmath225 of free points in the graphs .
the presented facts about the brauer algebra should be enough to understand the loop gas reformulation of @xmath41 spin model , which we give in the next section .
the emergence of dense intersecting loops becomes transparent if we take the local transfer matrices in the adjoint representation of the brauer algebra .
this simply amounts to replacing in eq .
the generators @xmath17 , @xmath166 and @xmath167 defined by eq . with the diagrams in fig .
[ f : gen_diag ] .
the adjoint local transfer matrix is represented in fig .
[ f : vert ] .
we now define a loop model on a diagonal lattice represented in fig .
[ f : dl ] , with reflecting boundaries on the left and right ( ie , free boundary conditions in the space direction ) and identified boundaries on the top and bottom ( quasiperiodic boundary conditions in the time direction . )
the states of this model are coverings of the lattice with dense _ intersecting _ loops .
dense means that every edge on the lattice necessarily belongs to a loop .
avoiding loop vertices have weight 1 and intersections come with weight @xmath98 .
there are two possible ways for a line to close in a loop .
the first one comes from the graphical representation of the relation @xmath226 in fig .
[ f : e2 ] .
we call such loops _ bulk _ loops
. clearly the fugacity of bulk loops is fixed to @xmath85 by the brauer algebra .
the second possibility is that the ends of the line close in the identified points of the top and bottom boundaries of the lattice .
we call such loops _ cycles_. the boundary condition we consider have an annulus geometry and , thus , a cycle can be either _
contractible _ or _ uncontractible_. the fugacity of cycles
is not fixed by the algebra .
in fact , as we explain below , this is exactly the degree of freedom allowing for multiple mappings from the @xmath41 spin models with @xmath117 fixed and the dense intersecting loop model with fugacity @xmath85 for loops .
we start by evaluating the trace @xmath227 of a diagram @xmath57 in the spin representation and then we generalize the result for quasiperiodic boundary conditions given by @xmath228 .
we follow the same line of reasoning as in @xcite . a cycle in a diagram @xmath57 is the subgraph on the set of points belonging to a loop if we identify its top and bottom vertices .
by an abuse of language we call the corresponding loop also cycle .
if we put a diagram @xmath180 to the left of a diagram @xmath181 we get a new diagram which we denote @xmath229 .
let @xmath230 be the cycles in @xmath57 .
we can separate them by permuting the top and bottom vertices of @xmath57 with the same permutation @xmath196 @xmath231 thus the trace of a diagram depends only on the weights of cycles @xmath232 more than that , the weight of a cycle depends only on how many times it winds the annulus .
indeed , if a cycle on @xmath233 points has no horizontal lines , then , by applying the same permutation to the top and bottom vertices we can bring it to the cycle @xmath234 .
this is because permutations with one cycle are conjugate in @xmath197 .
if a cycle @xmath235 has a horizontal edge between the first and the second vertex in the top then it has the same weight as a certain cycle @xmath236 on four points less then @xmath235 @xmath237 if we compare the @xmath235 on the left with @xmath236 on the right it is clear that , in the end of the iterative application of eq . , the final cycle can be interpreted as being the initial cycle @xmath235 maximally contracted on the annulus . in the end , the only weights we need to compute explicitly are that of the cycles @xmath32 and @xmath234 @xmath238 for boundary conditions twisted by the matrix @xmath239 the generalized weights are computed to be @xmath240 in the fundamental representation , every supermatrix @xmath58 is diagonalizable .
the diagonal form of @xmath241 in the fundamental representation is determined by exponentiating elementary weights @xmath242 and @xmath243 introduced in sec .
[ sec : superalg ] .
thus , @xmath58 restricted to @xmath244 has eigenvalues @xmath245 and restricted to @xmath246 has eigenvalues @xmath247 .
the braces in @xmath248 mean that @xmath249 appears for odd @xmath12 only .
eq . can now be rewritten @xmath250 for @xmath251 only the eigenvalues in @xmath244 change with respect to the previous case .
there are of the form @xmath252 for @xmath12 odd , while for @xmath12 even @xmath253 and @xmath254 . instead of eq .
one has now @xmath255 to summarize the basic results in this section , let @xmath87 be a dense loop covering of the lattice , @xmath17 be the number of intersections , @xmath256 be the number of bulk loops , @xmath105 be the number of contractible loops ( cycles ) and @xmath32(@xmath106 ) be the number of loops winding the annulus an even(odd ) number of times . on the annulus
the trace partition function ( which would correspond to antiperiodic boundary conditions in the ( imaginary ) time direction ) of the @xmath41 spin model may be reformulated as a dense intersecting loop model in the following way @xmath257 we see that it does depend on @xmath33 separately and not only on @xmath85
. meanwhile the supertrace partition function ( which would correspond to periodic couplings ) reads @xmath258 where @xmath259 is the total number of `` loops '' and , since it depends on @xmath85 only , is the same as for the @xmath10 model .
one can say that taking the supertrace in the partition function is equivalent to restricting the @xmath41 supersymmetry of the spin model to a , smaller , @xmath10 symmetry .
denote the spectrum of the transfer matrix of the @xmath41 spin model by @xmath260 .
we have the following important inclusion property @xmath261 the only difference between the @xmath41 and @xmath262 quasiperiodic partition functions is the weight of uncontractible cycles . for @xmath239 a matrix with eigenvalues @xmath263 and @xmath264 , except @xmath265 , the weight of uncontractible cycles is , according to eq . either @xmath85 or @xmath266 .
notice that these are exactly the weights of uncontractible cycles in the trace partition function for the @xmath262 spin model , which proves eq . .
we will use the inclusion property in the next section to derive some information about the indecomposable representations of @xmath41 appearing in the decomposition of @xmath2 .
assume the transfer matrix be a generic element of @xmath42 .
the action of @xmath42 on the tensor space @xmath2 was defined in the beginning of sec . [
sec : br_alg ] . the complete picture of the reducibility of the transfer matrix can be conveniently encoded in the decomposition of @xmath267 into a direct sum of @xmath42 indecomposable modules @xmath268 where @xmath269 denotes the multiplicity of isomorphic indecomposable @xmath42 modules @xmath270 ( it does not depend on @xmath43 ) , and the set @xmath271 is defined implicitly by the formula , and will be defined explicitly below . we remind the reader that @xmath2 is not necessarily a semisimple @xmath42 module if @xmath219 , so the modules @xmath270 appearing on the rhs of eq
. can be reducible .
the question of computing degeneracies of eigenvalues of the spin transfer matrix is easier to treat by looking at the centralizer @xmath272 , which acts on @xmath2 from the left if one consider @xmath42 acting from the right .
the dimension of indecomposable modules @xmath273 in the decomposition of @xmath2 as a @xmath274-module @xmath275 will give the desired degeneracies .
this is due to the fact that @xmath276 are dimensions of simple @xmath42 modules @xmath69 appearing as constituents of @xmath270 in eq . , while @xmath269 are dimensions of simple @xmath277 modules @xmath75 appearing as constituents of @xmath273 in eq . .
taking the character of both eqs .
( [ eq : bd],[eq : cd ] ) one can see that the number @xmath278 of irreducible components @xmath69 in @xmath270 is equal to the number @xmath279 of irreducible components @xmath75 in @xmath273 . because the action of @xmath40 commutes with @xmath42 we have that @xmath280 . however , when @xmath2 is semisimple it follows from the wedderburn decomposition theorem that @xmath281 . in the following
we suppose that there is still a schur duality between @xmath40 and the quotient of @xmath42 faithfully represented on @xmath2 .
this allows us to give an algorithm to compute the lhs of eqs .
( [ eq : bd],[eq : cd ] ) for small tensor powers @xmath43 and get some intuition about the general structure of @xmath270 and @xmath273 .
the set of partitions @xmath59\,\}$ ] labels @xmath42 irreps , while @xmath60 selects those of them which do realize on the tensor space @xmath2 .
denote by @xmath282 the double sided ideal defined by @xmath283 .
the annihilator @xmath284 is diagrammatically described in @xcite . under the homomorphism @xmath285 , the indecomposable modules @xmath66 give rise to induced modules @xmath286 .
clearly , @xmath287 is a tensor representation and can be generated by trace substraction and symmetrization as @xmath288 where @xmath289 , @xmath290 extracts all the traces from the tensor space @xmath291 and @xmath292 acts nontrivially only on @xmath291 as a young symmetrizer .
the double sided ideal @xmath293 is completely characterized by the set of weights @xmath294 .
note that @xmath295 .
the surviving indecomposable tensor modules @xmath287 are given by @xmath66 with irreducible components @xmath296 removed . the quotient @xmath297 can be carried out by imposing the vanishing of all words @xmath298 with @xmath299 .
it is useful to notice that not all of these conditions are independent and as one can see from eq .
@xmath300 if @xmath301 . as discussed in sec .
[ sec : superalg ] and [ sec : supergr ] , the @xmath40 irreducible components of @xmath273 are indexed ( up to an equivalence under the action of the outer automorphism @xmath302 induced by the symmetry of the dynkin diagram of @xmath40 when @xmath12 even ) by the set @xmath64 of hook shape partitions . representing the supergroup as a semidirect product @xmath303 , the elements of @xmath271 naturally acquire the structure of a couple of the form @xmath304 or @xmath305 if @xmath306 and @xmath307 if @xmath308 , where @xmath309 are the trivial , alternating ( superdeterminant ) and two dimensional representations of @xmath310 .
thus , every @xmath311 gives rise to two @xmath41 inequivalent irreps with highest weights @xmath312 and the associate @xmath313 if @xmath306 and a single self - associate irreps of highest weight @xmath314 if @xmath308 .
for typical @xmath311 , one can realize the @xmath41 irreps @xmath61 on tensors @xmath315 and describe their symmetry by some young tableaux . as discussed in details in sec .
[ sec : tmb ] , the young tableau corresponding to @xmath316 can be constructed by adding a border strip to the young tableau of shape @xmath51 corresponding to @xmath317 .
ultimately , this is justified by the fact that eq .
( [ eq : elchar],[eq : onchar ] ) gives the right characters for @xmath318 and that they coincide up to @xmath319 .
although atypical representations can not be realized as tensor representations we represent the associate weight @xmath320 of an atypical weight @xmath51 by a young tableau such that @xmath321 is a skew partition described in sec .
[ sec : tmb ] and sec .
[ sec : mod_rules ] .
the idea is to exploit the fact that the characters of indecomposable modules @xmath66 , given in @xcite , do not depend on the semisimplicity of @xmath42 .
this and some properties of generalized schur functions , which are summarized in @xcite , can be used to prove that @xmath322 is true even for all @xmath43 . here
@xmath323 is the character of @xmath210 in the representation provided by @xmath66 .
the functions @xmath324 are polynomials in the eigenvalues of @xmath239 , which where introduced for the first time by bars in @xcite in an early attempt to describe the supercharacters of @xmath41 .
they can be defined recursively as @xmath325 and @xmath326 for @xmath327 the brauer algebra @xmath42 is semisimple and @xmath328 .
consequently , @xmath66 are irreducible and @xmath329 .
because of the commuting actions of @xmath41 and @xmath42 one can naturally consider @xmath2 as a @xmath41-@xmath330-bimodule , with @xmath41 acting from the left and @xmath42 from the right .
then , eq . can be understood as a consequence of the decomposition @xmath331 with @xmath332 being actual characters of tensor irreducible modules @xmath333 as shown in @xcite . for @xmath33
such that @xmath219 , the polynomials @xmath332 can not generally be interpreted as the character of some @xmath41 representation . as we have seen in sec .
[ sec:1749 ] , @xmath334 can be brought to the form @xmath335 and eq .
is not more then a simple equality between two polynomials in eigenvalues of @xmath58 .
moreover , the two eqs .
( [ eq : cd],[eq : bf ] ) are still compatible , even if there are much more elements in @xmath336 then in @xmath337 .
this is possible because @xmath332 are not functionally independent when @xmath219 .
then , for @xmath338 the polynomials @xmath332 can be written in terms of functionally independent @xmath73 with @xmath339 by means of modification rules for characters @xmath340 given in @xcite and discussed in details in sec .
[ sec : mod_rules ] .
the fundamental eq .
is useful for small widths @xmath43 , when it is possible to compute the number @xmath341 of irreducible components @xmath342 in @xmath66 either by repeated applications frobenius reciprocity , as explained in @xcite , or by numerically diagonalizing the transfer matrix of sec . in the adjoint representation of @xmath42 and detecting the `` accidental degeneracies '' in its spectrum .
indeed , from the explicit definition it is clear how to restrict the adjoint transfer matrix to indecomposable modules @xmath66 . after we described in details the action of generators on the basis of @xmath66 in sec .
[ sec : br_alg ] , the algorithm of a numerical diagonalization is straightforward . the information about the structure of @xmath66 and the modification rules in eq . can now be used to bring eq . to the form
@xmath343 we see that @xmath344 iff , @xmath345 are not irreducible @xmath40 characters , one can still use them as a basis for representing the character of any representation .
] there is at least one @xmath339 such that @xmath346 . to determine @xmath279 one has to decompose the factor of @xmath347 in eq . as a sum of @xmath41 irreducible characters , which are explicitly known , as far as we know , only for @xmath348 and @xmath349 .
given the huge order of the set of weights @xmath336 , it may seem that calculations according to eq .
are extremely cumbersome already for small @xmath43 .
the simplifying point is that @xmath350 ( or @xmath351 ) is non zero only if both weights are in the same equivalence class of @xmath271 ( or @xmath337 ) .
the splitting of @xmath271 ( or @xmath337 ) into equivalence classes , called _ blocks _ and described in details in sec .
[ sec : tmb ] , is with respect to an equivalence relation between irreducible components of indecomposable @xmath41 ( or @xmath42 ) modules .
an important consequence of the fact that @xmath41 supertrace partition functions depends only on the @xmath10 part of the spectrum is the vanishing of the superdimension @xmath352 for all indecomposable modules with @xmath353 . a more restrictive criterion for @xmath273 supercharacters deriving from the full inclusion sequence in eq .
can be derived by taking a matrix @xmath58 with eigenvalues @xmath354 and @xmath355 for @xmath356 , @xmath357 .
then , it can be seen from eqs .
( [ eq : trp+],[eq : loop_pf ] ) or eqs .
( [ eq : elchar],[eq : onchar],[eq : bf ] ) that any @xmath41 quasiperiodic partition function will also be an @xmath262 quasiperiodic partition function . as a consequence @xmath358 and
the supercharacters of @xmath273 vanish when @xmath359 and @xmath58 can be embedded in @xmath262 . for @xmath174 the modules @xmath2 is semisimple .
therefore @xmath360 . on the other hand ,
if @xmath29 is big enough @xmath361 . it could be interesting to understand the relation between the sequence @xmath362 and the cohomology of the radical @xmath363 .
observe that a filtration similar to that of @xmath337 is available on @xmath271 by the degree of atypicality of its elements .
in fact , we explain in sec . [
sec : tmb ] of the appendix how the weights in a block of @xmath337 or @xmath271 can be organized by the number of _ removable balanced continuous border strips _ in the corresponding young tableau .
this number can be interpreted as the degree of atypicality when the corresponding partition represents an @xmath41 weight .
let @xmath124 be a two dimensional vector space endowed with an action of @xmath364 .
the action of @xmath39 on the tensor space @xmath2 is @xmath365 @xmath3 acts according to the following definition of generators @xmath366 @xmath367 where @xmath368 have the following representation on @xmath126 @xmath369 ( note that @xmath32 differs in some essential way from the projection operator onto the singlet representation in the usual @xmath370 basis ) .
the decomposition of @xmath2 as a @xmath39-@xmath3-bimodule is simply @xmath371 here @xmath372 the tensor representations @xmath373 are irreducible with dimensions @xmath374 and @xmath375 .
the representation @xmath376 is the trivial one and @xmath377 is the associate one dimensional @xmath378 representation . at the restriction to @xmath89 the representations @xmath376 and @xmath377 become equivalent , while @xmath379 splits into two nonequivalent one dimensional representations @xmath380 where @xmath381 is the @xmath382 angle .
the @xmath3 representations @xmath69 are irreducible as well and are constructed by acting with @xmath3 on the tensor module in eq . .
the dimensions of simple modules @xmath383 are easily computed by looking at the first row of @xmath384 , where @xmath105 is the fusion matrix @xmath385 described by a @xmath386 type dynkin diagram with labeling of the nodes shown in fig .
[ fig : fusion ] .
it is not hard to solve the recurrence relations satisfied by @xmath387 @xmath388 and get that @xmath389+k}$ ] except the case when @xmath43 is even and @xmath390 for which @xmath391 .
these are , as expected , the number of eigenvalues of the transfer matrix for the 6 vertex model in the sector of spin @xmath392 . .
] we are interested in giving an algebraic description of the 6 vertex model transfer matrix algebra . in other words
we want to identify the brauer algebra annihilator @xmath393 of @xmath2 and carry out the quotient @xmath394 .
all the @xmath3 weights @xmath395 such that @xmath396 satisfy either @xmath397 or @xmath398 .
thus , it is enough to consider @xmath399 and impose the vanishing of the double sided ideal of the word @xmath400 projecting onto @xmath401 .
as explained in the beginning of the previous section @xmath402 , where @xmath403 extracts all the traces from @xmath404 and @xmath405 are the young symmetrizers corresponding to @xmath406 .
the projector @xmath403 can be found by looking at the form of an arbitrary tensor @xmath407 after extracting all of its traces @xmath408 which gives @xmath409 and clearly @xmath410 .
the young symmetrizer @xmath411 is @xmath412 and @xmath413 , where @xmath414 are the projectors onto the _ standard _ young tableau @xmath415 $ ] and @xmath416 $ ] .
the two orthogonal projectors @xmath417 and @xmath418 are independent only if we restrict to the right @xmath3 action .
in fact , the left ideal of the word @xmath419 is the same as the double sided ideal of the word @xmath420 . the condition @xmath421 gives the following restriction @xmath422 on generators @xmath423 .
putting @xmath424 one can see that eq
. implies that @xmath425 are temperley lieb operators with @xmath426 .
there are no more restrictions that can be drawn from the conditions @xmath427 , because @xmath428 is a one dimensional projector . before exploring the next vanishing condition
let us revise the the defining relations of @xmath3 given in eq .
@xmath429 which imply @xmath430 observe that although the algebra has now two temperley lieb operators their role is not symmetric yet at this stage .
next , the condition @xmath420 implies @xmath431 which after inserting @xmath432 with the help of eqs .
( [ e : first_step],[e : first_step1 ] ) becomes @xmath433 multiplying eq . by @xmath434 on the right
we get : @xmath435 which can be used to rewrite eq . as @xmath436 multiplying by @xmath437 on the left and on the right of eq .
and using only the relations between @xmath425 , the relations between @xmath166 and eq
. one can get all the eqs .
( [ e : first_step][e : first_step10 ] ) and also @xmath438 which establish a complete symmetry between @xmath166 and @xmath425
. the double sided ideal of @xmath420 is composed of four linearly independent words two generated by the left action and other two generated by the right action of @xmath3 . it is useful to note that after taking the quotient of @xmath439 we are left with 10 independent words instead of 15 , which is exactly what we need for the 6 vertex local transfer matrix . we give the following abstract definition to the 6 vertex model transfer matrix algebra @xmath440 in term of generators @xmath441 latexmath:[\ ] ] where @xmath592 and the content of @xmath587 is represented below @xmath589 . apart
the last irreducible module @xmath593 , we recognize in the terms of eq .
the projective representations of the quiver @xmath594 in fig .
[ fig : einfty ] , which describes the homomorphisms between the @xmath3 tensor modules @xmath531 realized on @xmath2 .
it will turn out in our forthcoming analysis of conformal properties to be easier to study numerically the hamiltonian @xmath595 the expectation which we will confirm in great details is that this hamiltonian will be in the same universality class as the spin model we had started with .
the hamiltonian @xmath596 is obviously local and has only nearest neighbour interactions if the @xmath32 s and @xmath31 s are taken in the spin representation provided by eq . .
however , this is no longer true if we think of @xmath596 as an element of the adjoint representation of @xmath3 .
the lowest eigenvalue of @xmath596 belongs to the @xmath3 irreducible representation labeled by @xmath597 . for generic @xmath112
it is nondegenerate if @xmath43 is even and has degeneracy @xmath598 if @xmath43 is odd . on the other hand
, the highest eigenvalue belongs to the completely antisymmetric representation labeled by @xmath599 . in this representation the @xmath31 s
act as -1 and the @xmath32 s as 0 .
the hamiltonian @xmath596 is determined up to an arbitrary additive constant and multiplicative factor . for numerical diagonalization
it is convenient to fix the additive constant such that the maximal eigenvalue of @xmath596 be zero .
the multiplicative factor is fixed by requiring @xmath600 with @xmath601 as in eq . , @xmath15 as in app .
[ sec : superalg ] and @xmath602 being the _ xxz _ spin chain hamiltonian in its usual form @xmath603 the fact that the eigenvalues of the 6 vertex model appear as a subset of the eigenvalues of the transfer matrix for the @xmath0 model and @xmath604 carries over to a similar result for the hamiltonians .
the velocity of sound for the massless excitations can thus be derived from its value for the _ xxz _ subset , which is well known from @xcite to be @xmath605 the hamiltonian @xmath606 is diagonalized numerically in the adjoint representation of the brauer algebra by studying its action on the diagrams just like for the transfer matrices .
next , once the structure of indecomposable modules @xmath66 is known , eq . can be used as explained in sec .
[ sec : gf ] to select the part of the spectrum which does indeed appear for a fixed @xmath29 spin model .
however , in the two special cases @xmath607 the hamiltonian @xmath596 greatly simplifies . in the following two subsections we discuss the behaviour of the spectrum of @xmath596 in the two limits @xmath608 .
when @xmath609 the temperley lieb operators @xmath166 do not contribute to @xmath596 and , thus , the hamiltonian is no longer a generic element of the brauer algebra @xmath3 , but belongs instead to the subalgebra @xmath610 .
this will translate to additional degeneracies in the spectrum of @xmath596 at the point @xmath611 compared to other points in the range @xmath612 .
hamiltonians of type @xmath613 , with @xmath31 s in the representation provided by eq .
, are integrable and have been studied in @xcite and @xcite .
although the continuum limit of such spin chains is a gapless field theory , it fails to be conformal , because excitations have a @xmath614 scaling law in the thermodynamic limit .
this can readily be seen from the vanishing of the sound velocity in eq . .
we will not enter into the details here , but just mention that the different systems of bethe ansatz equations are indexed by @xmath615-hook shape partitions @xmath50 .
this is exactly the label of irreducible representations of the group algebra @xmath54 realizing in the centralizer of the spin chain @xmath2 , with @xmath124 being the fundamental representation of @xmath38 .
we see that the symmetry of our spin model @xmath0 jumps to @xmath38 at the point @xmath611 .
the additional degeneracies in the spectrum of the @xmath0 spin model at the point @xmath611 can be understood by looking at the decomposition of @xmath3 modules @xmath66 , into a direct sum of @xmath54 irreducible modules @xmath67 .
let @xmath289 and @xmath50 , then it was shown in @xcite that the multiplicity of @xmath67 in the decomposition of @xmath66 is @xmath616 where @xmath617 are littlewood - richardson coefficients . alternatively , @xmath618 is the number of tensors of rank @xmath619 , with index symmetry of some fixed standard young supertableau of shape @xmath198 , that can be obtained from a tensor of rank @xmath43 , with index symmetry of some standard young supertableau of shape @xmath51 , by contracting @xmath620 indices in all the possible ways
. one can apply eq . to understand the degeneracy of the lowest level of @xmath596 at @xmath611 .
first , observe that @xmath613 is minimized in the sector @xmath621 ( where @xmath31 s acts as 1 ) .
the _ only _ @xmath198 such that @xmath622 are one row partitions .
thus , the lowest eigenvalues of @xmath596 restricted to @xmath623 for @xmath218 $ ] become all degenerate at @xmath611 . arguments of this kind can be used to derive information about the critical exponents of the spin model in the limit @xmath624 the same reasoning can be applied to the point @xmath626 . at this point , the hamiltonian @xmath596 belongs to the temperley lieb subalgebra @xmath627 and the model can be considered as a spin chain @xmath628 where @xmath629 are the fundamental representation of @xmath38 and its conjugate .
additional degeneracies can be understood by looking at the decomposition of @xmath3 modules @xmath66 as a direct sum of standard irreducible @xmath630 modules @xmath70 .
let us compute the multiplicity @xmath631 of irreducible modules @xmath70 in the decomposition of @xmath66 with @xmath632 . as explained in sec .
[ sec : br_alg ] , @xmath66 has a natural basis composed of all possible pairings @xmath633 of partial diagrams @xmath204 with @xmath634 free points and basis vectors @xmath635 of @xmath201 .
we say that a horizontal line of a partial diagram @xmath204 is _ intersected _ either if it intersects another horizontal line or if there is a free point in @xmath204 between the two ends of the horizontal line .
let us associate to each partial diagram @xmath204 the number of intersected horizontal lines @xmath502 in @xmath204 .
it is not hard to see that the span on the basis vectors @xmath209 , with @xmath204 s having at most @xmath502 horizontal intersected lines , is a @xmath630 submodule in @xmath66 . if we denote this submodule by @xmath636 there is an obvious filtration @xmath637 of @xmath66 .
consider the natural action of @xmath630 on the quotient modules @xmath638 .
observe that the action of @xmath630 changes the labeling @xmath639 of free points in a labeled graph @xmath203 if and only if it also reduces the number of horizontal intersected lines .
therefore , @xmath640 is isomorphic to a direct sum of @xmath641 modules @xmath642 .
obviously @xmath643 and , therefore , we get @xmath644 for @xmath645 , @xmath646 for @xmath647 and finally @xmath648 . thus , our problem effectively reduces to understanding the action of @xmath630 on the module @xmath649 , which is composed of partial diagrams @xmath204 on @xmath43 points with @xmath185 unlabeled free points . at a closer look
, one can see that the action of @xmath630 on partial diagrams keeps the reciprocal configuration of intersected lines and free points intact .
in other words , if @xmath650 is a map that eliminates all the nonintersected horizontal lines from a partial diagram and acts as identity otherwise , then @xmath650 defines an invariant of @xmath630 , that is @xmath651 to understand the meaning of this invariant let us define a local map @xmath381 between partial diagrams which sends intersected horizontal lines to free points as depicted in fig .
[ fig : themapphi ] and acts as identity otherwise .
the local map @xmath381 is applied repeatedly until there are no more horizontal intersected lines left . .
] it is not hard to see that @xmath381 extends to a homomorphism of @xmath630 modules @xmath652 in fact , the role of the map @xmath381 is to show that @xmath653 is composed of a direct sum of isomorphic @xmath654 modules , while that of the map @xmath650 is to distinguish between these modules .
the set of partial diagrams @xmath204 in @xmath653 splits into subsets of constant @xmath655 and each of these subsets is isomorphic to @xmath654 as a @xmath630 module . according to what was said before , we get that @xmath656 equals to the number of graphs on @xmath657 vertices and @xmath502 intersected edges .
it follows that @xmath656 does not actually depend on @xmath43 and we drop the index @xmath43 in the following .
this fact allows , in principle , for an iterative computation of @xmath658 by simply computing the dimensions of the left and right hand sides of the decomposition formula @xmath659 that is @xmath660 successively for @xmath661 or @xmath662 .
one can give an explicit expression for @xmath658 with a little more combinatorial work .
we call a horizontal line an _ empty cup _ if its ends are adjacent and simply a _ cup _ if there are separated by free points
. observe that all the lines in a partial diagram are intersected if and only if there is at least one free point in each cup .
thus , if the partial diagram has @xmath204 cups with only one free point inside and a total of @xmath502 edges then the remaining @xmath663 free points can be added to the diagram in @xmath664 different ways in such a way that the resulting diagram has only intersected edges . moreover , the number of diagrams on @xmath665 points with @xmath204 empty cups and a total of @xmath502 edges is again @xmath666 .
this is because the condition of no cups in the connection of the remaining @xmath667 edges is similar to the condition of composing a graph with @xmath204 free points and @xmath667 intersected edges . putting everything together we get a new recurrence formula @xmath668 reducing the problem to the computation of @xmath669 .
next , we want to find a recurrence relation for @xmath670 by looking at the connectivity of the first point in the partial diagrams on @xmath671 points with @xmath83 intersected edges .
the leftmost vertex in the partial diagram has to be connected to some other vertex at position @xmath501 .
the connectivity of the @xmath672 points to the left of the point at position 1 is equivalent to that in a partial diagram with @xmath673 intersected edges and a free point except for the case where @xmath513 .
therefore we have that @xmath674 now , eq . yields @xmath675 . using again eq . for @xmath673 we finally get that @xmath676 the solution of the recurrence eq . with the initial conditions @xmath677 and @xmath678 is @xmath679 and coincides with the absolute value of bessel polynomials @xmath680 @xmath681 evaluated at @xmath682 .
besides the careful definition of the spin model and its sectors , the main point of this first paper is the algebraic set up necessary to analyze its symmetries . this is a non trivial task since we are dealing with non semi - simple algebras , and that the action of @xmath0 and @xmath3 are meshed through a complex structure of indecomposable representations .
the main results are the decomposition formulas ( [ eq : rez1],[eq : rez2 ] ) for @xmath481 viewed as a @xmath349 and a @xmath3 module . the decomposition in eq .
has been computed in two essentially different ways : first , by decomposing tensor products between @xmath349 representations and @xmath124 without knowing anything about the brauer algebra and , second , starting from eq . with the assumption that the representations of @xmath349 and @xmath3 on @xmath2 generate the full centralizers of each other ( schur duality ) .
the fact that we arrive at the same result using both methods highly suggests that our assumption about the schur duality between @xmath349 and @xmath3 on @xmath2 is correct . when the question of decomposing @xmath2 is addressed in sec .
[ heart ] , the notion of block appears to be a particularly useful concept for organizing indecomposable representations .
these results will be applied to educated conjectures about the conformal field theory in the next paper .
although there are many things left unclear about the representation theory of @xmath683 , it is very tempting to speculate the form of the decomposition of @xmath684 . before making the guess , observe that as a @xmath349 module @xmath685 , where @xmath31 is a direct sum of projectives organized in blocks , while @xmath158 is a direct sum of simples indexed by the same young tableau ( in the partition notation for dominant weights ) as the irreps of @xmath39 .
more than that , they appear with the same multiplicities as their partners in @xmath686 . that the same phenomenon occurs for @xmath687 . in this case
@xmath158 is the trivial representation . ] therefore , @xmath158 and @xmath688 are similar in all but the internal structure of their simple summands .
the similarity between the two modules has to be understood in terms of their centralizers , because these are precisely the objects that do not `` see '' the internal structure of simples .
is a simple module for the algebra @xmath689 then @xmath690 . ] in conclusion , one should have @xmath691 , which is quite natural once there is a schur duality between @xmath349 and @xmath3 on @xmath684 .
it is tantalizing to speculate that as a @xmath0 module @xmath692 , with @xmath31 projective and @xmath693 .
thus , the problem of the decomposition of @xmath684 as a @xmath0 module is reduced to understanding the projective representations of the supergroup , i.e. to finding the quiver diagram for each block .
it has been suggested in @xcite that the quiver diagram of blocks does not depend on @xmath29 provided the degree of atypicality @xmath501 and the action of the outer automorphism @xmath302 are fixed .
can act in two ways : either leave invariant all the weights in the block or pairwise transform some of them . ]
the discussion of sec .
[ sec : osp42 ] suggests that the two types of quivers for a block of @xmath694 and a fixed @xmath501 will give rise to the same quiver for the induced blocks in @xmath0 .
we also succeeded in computing the multiplicity of temperley lieb representations in a standard @xmath3-module @xmath66 .
finally , we gave a combinatorial description of @xmath42 blocks as the set of minimal partitions dressed by balanced removable border strips and have shown that there is a similar description for @xmath40 blocks .
* acknowledgments : * this work was supported by the agence national pour la recherche under a programme blanc 2006 int - ads / cft .
we would like to thank j. germoni for fruitful correspondence , and j. jacobsen and p. martin for useful discussions .
cc i. bena , j. polchinski and r. roiban , phys .
d 69 ( 2004 ) 046002 , hep - th/0305116 .
f. wegner , nucl .
phys . b316 ( 1989 ) 663 .
m. zirnbauer,``conformal field theory of the integer quantum hall plateau transition '' , hep - th/9905054 .
n. read and h. saleur , nucl .
b613 ( 2001 ) 409 .
n. mann and j. polchinski , phys .
d72 ( 2005 ) 086002 and n. mann and j. polchinski ; hep - th/0508232 .
n. read and h. saleur , nucl .
b 777 , 263 ( 2007 ) .
n. read and h. saleur , nucl .
b 777 , 316 ( 2007 ) .
b. nienhuis , e. k. riedel and m. schick , phys .
b27 , 5625 .
e. domany , d. mukamel , b. nienhuis and a. schwimmer , nucl .
b190 ( 1981 ) 279
. b. nienhuis , j. stat . phys . 34 ( 1984 ) 731
. j. l. jacobsen , n. read and h. saleur , phys .
( 2003 ) 090601 ; cond - mat/0205033 .
g. parisi and n. sourlas , j. physique lett .
( paris ) 41 ( 1980 ) l403 .
a. ram , pac .
j. math . 169
( 1995 ) 173 .
w. f. doran , d. b. wales and p. j. hanlon , j. algebra 211 ( 1999 ) 647 .
g. benkart , c. l. shader and a. ram , j. pure appl .
algebra 130 ( 1998 ) 1 .
j. birman and h. wenzl , trans . am .
( 1989 ) 249 .
j. murakami , osaka j. math .
24 ( 1987 ) 745 .
m. wadati , t. deguchi and y. akutsu , phys .
180 ( 1989 ) 249 .
h. weyl , `` the classical groups '' , princeton university press ( 1946 )
. f. gavarini , `` on the radical of brauer algebras '' , math.rt/0606501 ( 2006 ) .
f. goodman , p. de la harpe and v. jones , `` coxeter graphs and towers of algebras '' , msri publications 14 , springer , berlin 1989 . p.p .
martin , `` potts models and related problems in statistical mechanics '' , world scientific , singapore ( 1991 ) .
a. cox , m. de visscher and p. martin , `` the blocks of the brauer algebra in characteristic zero '' , math.rt/0601387 .
a. cox , m. de visscher and p. martin , `` a geometric characterization of the blocks of the brauer algebra '' , math.rt/0612584 .
a. balantekin and i. bars , j. math .
( 1981 ) 1149 . c. j. cummins and r. c. king , j. phys
. a 20 ( 1987 ) 3103 .
r. green , jktr 7 ( 1998 ) 155 .
d. levy , int .
phys . a6 ( 1991 ) 5127 .
j. germoni , boletin de la academia nacional de ciencias 65 ( 2000 ) 147 ` http://hal.archives-ouvertes.fr/hal-00002750/fr/ ` . c. hamer , j. phys . a 19 ( 1986 ) 3335 .
b. sutherland , phys .
b 12 ( 1975 ) 3795 .
h. saleur , nucl.phys .
b578 ( 2000 ) 552 .
p. hanlon and d. wales , math . comp .
54 ( 1990 ) 771 . v. serganova proceedings of the international congress of mathematicians , 1998 .
v. kac , adv .
( 1977 ) 8 .
b. morel , a. sciarrino , p. sorba , j. phys .
a : math . gen . 18
( 1985 ) 1597 .
j. p. hurni , j. phys . a 20 ( 1987 ) 5755 .
r. farmer , p. jarvis , j. phys . a : math .
17 ( 1984 ) 2365 .
v. kac , lecture notes in math . 676
( 1978 ) 597 .
v. kac , commun .
algebra 5 ( 1977 ) 889 .
j. van der jeugt , j. math . phys . 26 ( 1985 ) 913 .
i. macdonald , `` symmetric functions and hall polynomials '' , clarendon press , oxford ( 1979 ) , 10
on the left and @xmath695 on the right . ] in this section we recall standard facts about the @xmath40 lie superalgebra mainly following the pioneering work of kac @xcite . for more details on @xmath40 young supertableaux
see @xcite .
let @xmath124 be a vector space with an additive @xmath696 grading @xmath129 , that is @xmath697 and @xmath698 .
let @xmath699 , @xmath700 and @xmath701 $ ] . choose in @xmath124 a basis @xmath702 with @xmath703 and @xmath704 .
we take the vector @xmath705 in brackets because it appears for odd @xmath12 only .
the grading of @xmath124 induces a grading on @xmath706 , that is @xmath707 preserves the degree of @xmath708 and @xmath709 changes it .
define the supertranspose of a matrix @xmath710 by @xmath711 let @xmath15 denote the matrix with the only nonzero components @xmath712 the lie superalgebra @xmath40 is realized as a subset of @xmath706 with elements @xmath158 satisfying @xmath713 in terms of elementary matrices @xmath714 the generators of @xmath40 read @xmath715 the generators @xmath716 span the cartan subalgebra @xmath717 .
denote by @xmath718 the basis in @xmath719 dual to @xmath720 .
it can be easily checked that generators in eq .
correspond to roots of the type @xmath721 , generators in eq .
correspond to roots of the type @xmath722 and generators in eq .
correspond to roots of the type @xmath723 .
the bilinear invariant form @xmath724 induces a scalar product on @xmath719 .
the standard basis is recovered by putting @xmath725 for @xmath726 and @xmath727 for @xmath728 .
elementary weights @xmath729 are orthogonal in @xmath719 and @xmath730 .
the first @xmath731 simple roots are chosen to be @xmath732 , @xmath733 , @xmath734 for @xmath728 and @xmath735 .
the last simple root is @xmath736 for odd @xmath12 and @xmath737 .
the roots @xmath738 are called odd and the rest even . the component of a weight @xmath739 along the hidden simple @xmath740 root @xmath741 is @xmath742 according to @xcite , an @xmath40 highest weight is dominant iff it has integer dinkyn labels @xmath743 and integer @xmath744
satisfying the following consistency conditions @xmath745 all irreducible finite dimensional representations are indexed by dominant weights @xmath739 . given a dominant weight @xmath746 in the standard basis , the first @xmath731 dynkin labels are @xmath747 for @xmath748 , @xmath749 for @xmath750 .
the last dynkin label is @xmath751 for @xmath12 odd and @xmath752 for @xmath12 even . from eq .
we also get @xmath753 .
the set of numbers @xmath754 define a partition , shown in fig .
[ f : hook_shape ] , provided that consistency conditions plus some additional constraints depending on @xmath12 are satisfied .
these additional constraints require @xmath755 and @xmath756 to be even if @xmath12 is even , and @xmath757 to be to be even if @xmath12 is odd . representation to which belongs the highest weight state of a @xmath40 representation . ]
the last two conditions define tensorial weights . partitions @xmath51 such that @xmath758 are called hook shape .
let @xmath302 denote the outer automorphism induced by the symmetry of the @xmath695 dynkin diagram under the exchange of the last two roots in fig .
[ f : dynkin_diag ] .
this automorphism is extremely important in understanding the difference between the representation theory of the supergroup @xmath41 and its lie superalgebra .
note that @xmath302 can be explicitly realized through the discrete transformation @xmath759 exchanging the last two basis vectors in @xmath760 .
indeed , @xmath761 for @xmath735 and @xmath762 because @xmath763 are the duals of @xmath764 .
therefore @xmath765 . in the case of
@xmath12 odd , there is a bijective correspondence between hook shape partitions @xmath51 and dominant weights @xmath739 .
the same holds for @xmath12 even , except for @xmath51 with @xmath766 when @xmath51 represents both @xmath739 and @xmath767 . if there is a pair @xmath768 such that at least one of the conditions below are satisfied @xmath769 the weight @xmath51 is called _
atypical_. the highest weight @xmath739 is always atypical . ]
see @xcite for the origin of these conditions and note ref .
@xcite , where these have been presented in the form ( [ e:1_atyp],[e:2_atyp ] ) .
if none of these conditions is satisfied , the weight is called _ typical _ and , according to @xcite , the associated kac module @xmath770 ( which is a finite dimensional quotient of the corresponding highest weight module ) is irreducible , its ( super)character is given by the weyl - kac formula @xcite and , in particular , its superdimension is zero .
let @xmath771 be a grassman algebra .
the supergroup @xmath41 may be realized as a subset of even supermatrices @xmath772 with entries in @xmath689 and @xmath58 belonging to @xmath773 , and entries in @xmath256 and @xmath105 belonging to @xmath774 , which satisfies @xmath775 equivalently , @xmath41 can be seen as the set of linear transformations leaving invariant the graded symmetric form @xmath776 where @xmath777 are arbitrary points in a superspace parametrized by coordinates @xmath778 and @xmath779 and @xmath780 .
representing @xmath781 with infinitesimal @xmath782 and expanding eq .
one gets the definition of the superalgebra @xmath40 .
thus , the subgroup of @xmath41 connected to identity is an exponential of @xmath40 .
the representation theory of both is the same as long as we restrict to tensor representations which are the only ones appearing in the tensor space @xmath2 . from the definition any matrix @xmath783 has superdeterminant @xmath784 .
the supergroup has two disconnected parts @xmath785 , which correspond to the value of the superdeterminant of its elements , that is @xmath786 . to see this
, one can repeat the same reasoning typical of @xmath10 groups .
elementary transformations susceptible to change the sign of the superdeterminant belong to the discrete symmetry group @xmath787 of the @xmath41 invariant form .
the generators of @xmath787 are read out from eq . to be `` reflections '' @xmath788 and @xmath789 , and permutations @xmath790 and @xmath791 . for odd @xmath12
there is also the reflection @xmath792 .
the subgroup @xmath787 is in fact the weyl group of the root system of @xmath793 .
denote by @xmath794 the set of elements of @xmath787 embedded in @xmath785 .
it is easy to see that all elements of @xmath795 are conjugate in @xmath796 to a single reflection @xmath759 , which one can take @xmath797 if @xmath12 is even and @xmath798 if @xmath12 is odd .
therefore , we see that indeed @xmath799 .
let @xmath800 be a vector of weight @xmath801 .
then , as seen in sec .
[ sec : superalg ] , there is an action of @xmath759 on @xmath802 provided by @xmath803 . in the case of @xmath475
the outer automorphism @xmath302 exchanges @xmath804 with @xmath805 .
the representations induced from @xmath40 to @xmath41 are of the form @xmath806 there are two possible cases now : i ) either @xmath807 and then obviously @xmath808 with @xmath809 and @xmath810 or ii ) @xmath811 and the induced module in eq .
is irreducible .
two representations @xmath812 of @xmath696 are called _
associate_. the modules @xmath813 and @xmath814 are also called associate .
in contrast , @xmath815 is isomorphic to its associate because there is an equivalence transformation between @xmath816 and @xmath817 through the change of sing of basis vectors in the subspace @xmath818
. therefore @xmath819 is called selfassociate .
a direct implication following from the definitions of ( self ) associate modules is @xmath820 , where @xmath821 are ( self)associate weights of @xmath41 . for a selfassociate weight @xmath198
this equality implies @xmath822 if @xmath823 .
note that the centralizer of @xmath42 on @xmath2 is the direct product algebra @xmath824 rather then @xmath40 .
this algebra has the same tensor irreducible representations as the supergroup @xmath41 .
this is a compact resum of the results presented in @xcite plus some additional remarks on the representation theory of @xmath349 .
the superalgebra @xmath475 has minor differences with respect to the general context of @xmath40 superalgebras , because of the isomorphism @xmath825 .
the even part of the superalgebra is @xmath826 .
the odd part is a representation of the even part of dimension @xmath827 .
the standard basis vectors @xmath828 of @xmath719 are normalized as @xmath829 .
the even and the odd positive root systems are @xmath830 and @xmath831 .
the simple roots are traditionally chosen as @xmath832 , @xmath833 , @xmath834 .
the hidden root will then be @xmath835 .
consistency conditions for a dominant weight @xmath836 , require @xmath837 and @xmath838 .
we associate to @xmath739 a hook shape partition @xmath51 with symplectic part @xmath839 and orthogonal part @xmath840 . to make the correspondence @xmath841 bijective we mark the partition @xmath51 by @xmath842 when @xmath843 .
irreps can be written as a couple @xmath844 of @xmath845 irreps .
the sign attached to @xmath51 distinguishes between @xmath846 and @xmath847 when @xmath848 . ]
atypicality conditions ( [ e:1_atyp],[e:2_atyp ] ) take the form @xmath849 the solutions can be parametrized by two integers @xmath501 and @xmath502 . for @xmath390
these are @xmath850 , while for @xmath851 , @xmath852 and @xmath853 . and @xmath854 for the blocks of @xmath475 .
] denote by @xmath855 the @xmath475 simple modules . for @xmath51 typical
@xmath856 and @xmath857 can be computed by decomposing @xmath858 into ( at most 16 ) representations of @xmath859 @xmath860 the dimensions of @xmath861 can be computed with the help of character formulas given in @xcite . for @xmath390
we get @xmath862 , @xmath863 and @xmath864 , where @xmath865 . for @xmath851
we get @xmath866 and @xmath867 , and @xmath868 .
the set of weights @xmath534 with @xmath501 fixed belong to the same block of @xmath475 . to ( at least partially ) see this one has to check that the second order casimir invariant takes the same value @xmath869 on the whole block @xmath501 .
define the _ central character _ of @xmath855 .
if two weights @xmath870 are in the same block then @xmath871 have the same central characters .
this is a consequence of the extension of the schur lemma ( in the form known to physicists ) to indecomposable representations . ]
the actual construction of the set of indecomposable modules providing the equivalence relation of sec .
[ sec : tmb ] between the weights of a block is done in @xcite .
the quiver diagram representing the structure of @xmath475 projective modules in a block is represented in fig .
[ f : diag_ind ] .
the projective covers @xmath872 of the modules @xmath873 in the block @xmath390 have the submodule structure @xmath874 while in the block @xmath851 their submodule structure is @xmath875 the dimensions of projective modules in the block @xmath390 are @xmath876 , while in the block @xmath851 there are @xmath877 .
the superdimension of projective ( including typical ) modules vanishes .
let us apply the general discussion of sec .
[ sec : supergr ] to the supergroup @xmath349 .
the outer automorphism @xmath302 acts on @xmath719 by exchanging @xmath804 with @xmath805 .
consequently , @xmath878 for @xmath390 or @xmath879 and @xmath880 otherwise .
we claim that the quiver diagram of type @xmath881 for the block @xmath390 of @xmath475 will give rise to two quiver diagrams of type @xmath881 , as shown in fig .
[ fig : cool ] , and , consequently , to two associate blocks for the algebra @xmath570 , which we call @xmath882 .
as we shall see bellow , the weights in the block @xmath390 are @xmath883 , while the weights in the block @xmath505 are @xmath884 .
we also claim that the quiver diagram of type @xmath854 for the block @xmath885 of @xmath475 will give rise to a single quiver diagram of type @xmath881 , as shown in fig .
[ fig : cool ] , and a selfassociate block for the algebra @xmath570 , which we label also by @xmath501 .
as we shall sea bellow , the weights in the block @xmath886 are @xmath887 . and @xmath854 ( grey dots and dotted lines ) for @xmath475 blocks , represented in fig .
[ f : diag_ind ] , into quiver diagrams of type @xmath881 for @xmath570 blocks .
white and black dots represent weights of the type @xmath888 and @xmath305 respectively .
double circles represent selfassociate weights @xmath889 . ]
our claim follows from the analysis of shift operators @xmath890 introduced in @xcite .
these operators are very practical for decomposing kac modules @xmath891 into @xmath859 irreps .
of arrows in the quiver diagram of a block is irrelevant for the following .
one can take instead of @xmath891 the standard modules @xmath892 as well , which are defined by cohomological induction in @xcite . ] to be more specific , let @xmath893 be a vector of weight @xmath198 maximal for the algebra @xmath859 .
then , @xmath894 is again a maximal vector of weight @xmath895 for @xmath859 . to identify @xmath475 irreducible components in @xmath891 one has to search for @xmath859 maximal vectors with dominant weights @xmath198 in the same block as @xmath51 .
consider first the block @xmath390 of @xmath475 .
then @xmath896 has a @xmath859 maximal vector @xmath897 and one can check that all positive odd generators annihilate it .
therefore @xmath897 is a maximal vector for @xmath475 and @xmath898 contains at least @xmath899 and @xmath900 .
in fact , these are the only two irreducible factors of @xmath901 because , using the weyl - kac formula for characters and the results of sec .
[ sec : osp42 ] , one can check that @xmath902 . in the case @xmath903
one has that @xmath904 and , therefore , @xmath905 contains also the trivial representation . in order to see how the four weights @xmath906 split into two different blocks of @xmath907 one has to check out how @xmath908 transforms under the action of @xmath759 . from the explicit expression of shift operators in @xcite it follows that @xmath909 and , consequently ,
@xmath910 are in the same block and @xmath911 are in an other same block of @xmath570 .
the module @xmath912 has a maximal vector @xmath913 , of weight zero , corresponding to the trivial representation @xmath914 .
with the help of relations in appendix @xcite for the shift operator products of type @xmath915 , one can show that @xmath916 and , therefore , @xmath914 belongs to the block @xmath390 of @xmath570 as claimed .
consider now the block @xmath851 of @xmath475 .
then @xmath917 will have a single @xmath475 maximal vector ( besides @xmath918 ) corresponding to the irrep @xmath919 given by @xmath920 if @xmath921 , @xmath922 if @xmath923 and @xmath924 if @xmath925
. the induced module @xmath926 will be the sum of @xmath917 glued together by the action of @xmath759 .
finally , the induced module @xmath927 has two irreducible components @xmath928 and @xmath929 with @xmath570 maximal vectors @xmath930 .
in this section we explain carefully the notion of block appearing in the representation theory of nonsemisimple algebras .
we also look in details at the similarity between the blocks of @xmath40 and @xmath42 . in the representation theory of non semisimple algebras
the block is an essential notion .
the _ blocks _ are conjugacy classes of irreps with respect to the equivalence relation @xmath931 defined as follows .
let @xmath932 be the category of indecomposable modules of the algebra .
write @xmath933 if there is an indecomposable module in @xmath932 with simple summands @xmath934 .
extend the relation @xmath931 by transitivity in order to get an equivalence . in a semisimple
algebra the notion of block is irrelevant because indecomposable representation are irreducible and the congruence @xmath931 becomes an equality .
a relevant example is the temperley lieb algebra , with fugacity for loops @xmath85 in its adjoint / diagrammatic representation . for generic values of @xmath85
the algebra is semisimple and , thus , has only completely reducible representation . restricting to subsets of planar diagrams , with the number of vertical lines fixed to @xmath185 , and treating all the other diagrams as zero , we get all irreps , which are parametrized by @xmath185 . however , at special points @xmath935 with coprime integers @xmath936 , the algebra becomes nonsemisimple , irreps labeled by @xmath185 become reducible and @xmath185 becomes a label for a whole block of the algebra , see @xcite .
the irreducible components @xmath577 of indecomposable modules @xmath937 when the brauer algebra @xmath42 is nonsemisimple , where first studied by mathematicians hanlon _ et al _ in @xcite .
recently martin _ et al _ gave a complete description for the blocks of the brauer algebra in @xcite .
we introduce the same notation as in @xcite to formulate their block result for @xmath42 .
if the box @xmath938 is in the row @xmath82 and column @xmath83 of the young tableau of a partition @xmath198 , then its content is @xmath939 .
two boxes @xmath940 are called balanced if @xmath941 . for two partitions @xmath301 , the skew partition @xmath942 is called balanced if it is composed of balanced pairs of boxes .
the necessary condition for @xmath66 to contain @xmath577 is : i ) @xmath301 and @xmath942 is balanced ; ii ) if @xmath85 is even and the boxes of content @xmath943 in @xmath942 are configured as shown in case @xmath944 fig .
[ fig:2conf ] , then the number of columns in this configuration is even .
the given necessary criterion has the structure of a partial ordering .
if @xmath945 satisfy i ) and ii ) we write @xmath946 .
the splitting of the set of weights @xmath336 into posets with respect to @xmath947 gives the blocks of @xmath42 .
as shown in @xcite , there is a unique minimal partition in a block , which can serve as a label .
a sufficient criterion for the module @xmath66 to contain @xmath577 was derived in @xcite and requires @xmath51 to be the least weight @xmath948 . .
] we want to give a combinatorial description of the weights in a block .
consider the young tableau of a partition @xmath51 in the block of the minimal partition @xmath198 .
let @xmath949 ( @xmath950 ) be the box with the highest ( lowest ) content in the skew partition @xmath942 .
let @xmath804 ( @xmath951 ) denote the box bellow ( on the left of ) @xmath949 ( @xmath950 ) , if there is one , and the box on the left of ( above ) @xmath949 ( @xmath950 ) otherwise .
define by recurrence the balanced pairs @xmath952 until @xmath953 if @xmath85 is even or @xmath954 if @xmath85 is odd . by construction
, the set of boxes @xmath955 belongs to a balanced removable border strip of width one or , simply , a _
balanced strip_. one can repeat the same reasoning with the young tableau of @xmath956 ( which is not necessarily in the same block as @xmath198 because of ii ) ) .
thus , we clearly see that partitions @xmath51 in the same block can be constructed by dressing up with balanced strips a certain partition @xmath198 with no removable balanced strips .
denote by @xmath957 the balanced strip of smallest length addable to @xmath198 .
if @xmath85 is odd denote by @xmath958 the minimal partition @xmath959 . if @xmath85 is even denote by @xmath958 the minimal partition @xmath959 only if the two boxes with content @xmath960 in @xmath957 are disposed horizontally and @xmath961 otherwise . partitions which are of the form @xmath198 dressed up with an even ( odd ) number of balanced strips are in the same block as @xmath962 .
note that it is irrelevant in what order the strips are dressed on @xmath198 .
also , there can not be two balanced strips of the same length .
thus , a partition @xmath51 in the block @xmath963 is unambiguously specified by the length of balanced strips in the skew partition @xmath942 .
we claim now and show bellow that a block of @xmath40 is composed , in the partition notation of sec .
[ sec : superalg ] , of hook shaped partitions built up by dressing with balanced strips an atypical partition with no removable balanced strips . weights .
in particular the minimal partitions @xmath964 discussed above are in the same block of @xmath40 if there are hook shaped and atypical .
condition ii ) is clearly related to the @xmath696 we neglect by looking at the representation theory of @xmath40 instead of @xmath41 . ] for that we need to reformulate the original block result @xcite for @xmath40 .
let the _ degree of atypicality _
@xmath501 of a dominant weight @xmath739 , be the dimension of the subspace @xmath965 of the root lattice orthogonal to @xmath966 , where @xmath759 is the weyl vector of @xmath40 .
each atypicality condition in eq .
( [ e:1_atyp],[e:2_atyp ] ) is , in fact , an orthogonality condition between an odd root @xmath967 and @xmath968 .
therefore , @xmath501 is the number of odd roots orthogonal to each other and to @xmath968 or , equivalently , the number of atypicality conditions labeled by couples @xmath768 with distinct @xmath82 and @xmath83 . from the definition of the highest weight module
@xmath969 it is clear that irreducible finite dimensional components of @xmath969 must have dominant weights of the form @xmath970 , where the sum is over all odd positive roots @xmath971 spanning @xmath965 .
consider a @xmath40 weight @xmath51 , which , in the notation of app .
[ sec : superalg ] , has symplectic part @xmath759 and orthogonal part @xmath207 .
suppose @xmath972 are the first columns of @xmath51 satisfying @xmath973 and @xmath974 .
then , one can find rows @xmath975 , such that @xmath976 and the pairs @xmath977 satisfy the atypicality condition for @xmath978 if @xmath12 is odd and @xmath979 if @xmath12 is even and the atypicality condition for @xmath980 . indeed , from eq .
with @xmath981 the condition @xmath982 implies @xmath983 and thus @xmath984 , while @xmath985 implies @xmath986 if @xmath12 is even and @xmath987 is @xmath12 is odd . from eq . with @xmath981
the condition @xmath980 implies @xmath985 while @xmath988 follows directly from @xmath989 .
conversely , if @xmath990 satisfies an atypicality condition with @xmath981 , then @xmath991 . as shown in fig .
[ fig : imagine ] , @xmath992 is the width of the foot of the narrowest hook with arm width @xmath993 in which the young tableau of @xmath51 can be drawn in .
two atypicality conditions @xmath768 and @xmath994 are called independent if @xmath995 and @xmath996 .
clearly , conditions @xmath977 are pairwise independent for @xmath997 and for @xmath980 .
let us show that an atypicality condition @xmath998 is independent of conditions @xmath999 iff there is a row shorter then @xmath29 such that the box @xmath938 at the end this row and the box @xmath1000 at the end of column @xmath83 are balanced . in order to do that it is useful to imagine the partition @xmath51 drawn on an infinite square lattice , as in fig .
[ fig : imagine ] , with each square having its content written inside .
the following cases are possible * suppose first that there is no box with at the end of the row @xmath83
. from @xmath987 follows @xmath1001 .
observe that the column @xmath1002 is also empty .
therefore @xmath1003 and the atypicality conditions @xmath977 and @xmath1004 are not independent .
* let @xmath1005 and let @xmath763 be the box at the end of column @xmath83 .
no suppose that @xmath51 has a rightmost box @xmath1006 with content @xmath1007 and let @xmath1008 be the column of that box . condition @xmath1009 gives @xmath1010 .
the equality sign can not hold because otherwise @xmath1011 which contradicts @xmath987 .
thus @xmath1012 and therefore @xmath1013 .
if @xmath1014 then there is a box @xmath938 bellow @xmath1006 , which is balanced with @xmath1000 and has no box to the right , thus it is the end of a row shorter then @xmath29 . if @xmath1015 then comparing the @xmath82 s from the two atypicality conditions we get @xmath1016 and the two atypicality conditions are not independent
* finally , if there is no box with content @xmath235 then the column @xmath1017 gives an atypicality condition @xmath1004 with @xmath1016 .
next , by the definition of @xmath185 , a column @xmath1018 can satisfy an atypicality condition only with a row @xmath1019 . after inserting @xmath1020 in eqs .
( [ e:1_atyp],[e:2_atyp ] ) we get @xmath1021 the lhs in eq . is the hook length of the box in the row @xmath82 and column @xmath83 of @xmath51 , thus , always positive . on the other hand , eq .
requires the box @xmath1000 at the foot of column @xmath83 be balanced with the box @xmath242 at the end of row @xmath82 . in the end
, we see that there are two sources for _ independent _ atypicality conditions satisfied by a weight @xmath51 .
first , if @xmath992 is the width of the foot of the narrowest hook with arm width @xmath993 , in which the young tableau of @xmath51 can be drawn in , then there are @xmath1022 atypicality conditions satisfied by the weight and we call them of type 1 .
second , to each balanced pair of boxes @xmath1023 , such that @xmath938 is a box at the end a row and @xmath938 is a box at the end of a column , corresponds an atypicality condition of type 2 .
if @xmath51 satisfies @xmath1024 atypicality condition of type 2 then the degree of atypicality of the weight is @xmath1025 .
let @xmath1026 be the set of balanced pairs satisfying atypicality conditions of type 2 . applying the iterative construction explained above to the boxes
@xmath1027 one can see that there is a removable balanced strip @xmath1028 in @xmath1029 , with its ends in @xmath1030 .
clearly , by the same reasoning , one can identify a new balanced strip @xmath1031 removable in @xmath1032 .
the end @xmath1033 of this iterative procedure has no more removable balanced strips .
note that @xmath1033 satisfies @xmath501 atypicality conditions all of type @xmath1034 and the sequence of weights @xmath1035 has the same degree of atypicality @xmath501 . in order to complete the proof of the claim it remains to notice two things
first , if @xmath1036 is an odd root generating an atypicality condition of type 1 , then @xmath1037 is not dominant .
second , if @xmath1038 is an odd root generating an atypicality condition of type 2 , then @xmath739 has a removable strip with its ends in the last box @xmath242 of row @xmath82 and @xmath1039 of column @xmath83 and @xmath1040 is dominant and can be represented by a partition of the form @xmath1041 , where @xmath51 is the young tableau of @xmath739 . drawn on a hook shaped square lattice and fitting exactly inside a hook with foot width @xmath992 and arm width @xmath993 .
the black boxes represent the diagonal of squares with content @xmath1042 . ]
the explicit form of the characters of classical groups is easier derived in the limit of infinite rank of the corresponding lie algebra .
the inverse limit exists and is given by the _ modification rules _ for characters .
the concepts of infinite rank and inverse limit are rigorously defined for the case of schur symmetric functions , connected to the irreducible characters of @xmath1043 , in @xcite .
let us clarify this point .
the characters of classical groups , evaluated on a group element , are polynomials in the eigenvalues of that element in the defining representation for the group .
the infinite rank limit corresponds to considering polynomials depending on an infinite number of such variables .
irreducible characters are polynomials with a very specific symmetry , which is not obscured by the restriction of finite number of variables in the infinite rank limit . these objects are known as _
symmetric functions_. when the number of variables is set finite most symmetric functions become functionally dependent .
once an algebraically independent subset of symmetric functions is chosen , which is the actual set of characters in the case of classical groups , the modification rules `` for characters '' represent arbitrary symmetric functions along this basis .
one can introduce generalized symmetric functions @xmath332 for the supergroup @xmath41 according to eqs .
( [ eq : elchar ] , [ eq : onchar ] ) , see @xcite , @xcite .
the major difference with respect to classical groups is that functionally independent generalized symmetric functions are no longer irreducible characters of the supergroup .
however , modification rules for @xmath332 exist and have been derived in @xcite .
we bring them bellow in the form of eq . with the notations of our paper .
suppose that @xmath51 is a typical @xmath40 weight .
then , according to @xcite , only @xmath332 with @xmath198 of the form @xmath51 dressed by balanced strips @xmath1044 modify to @xmath73 @xmath1045 here @xmath1046 is the superdeterminant representation .
we have also introduced the weight function @xmath1047 defined on skew partitions composed of balanced strips and @xmath1048 is the number of columns in @xmath1049 .
consider now , the @xmath40 block labeled by the weight @xmath1050 with degree of atypicality @xmath501 and no removable balanced strips .
then , any weight @xmath51 in the block @xmath1051 of @xmath198 is of the form @xmath1050 dressed up by @xmath1052 balanced strips .
then , according to @xcite , @xmath332 with @xmath198 of the form @xmath1050 dressed up by @xmath1053 balanced strips modifies to @xmath1054 for a typical weight @xmath51 we put @xmath320 equal to @xmath51 dressed up by the balanced strip of minimal length @xmath1055 if @xmath1056 . in order to prove that @xmath1057 one has to show that @xmath1055 runs over an odd number of columns .
indeed , each box in @xmath1055 belongs either to a horizontal or a vertical part of the strip , except for the boxes at the corners of @xmath1055 , which belong to both .
we say the balanced pair @xmath1060 has an allowable configuration if both boxes belong either to horizontal or vertical parts of @xmath1055 , otherwise @xmath1061 has a non allowable configuration . as discussed in app .
[ sec : tmb ] , to every non allowable configurations of @xmath1061 with content @xmath1062 corresponds a removable balanced strip in @xmath51 with its ends in the border boxes with content @xmath1063 or @xmath1064 depending on weather @xmath938 is on a horizontal or a vertical part of @xmath957 .
because @xmath51 is a @xmath40 typical weight and , thus , has no removable balanced strips , there are only allowable configuration of balanced pairs in @xmath1055 .
thus , a balanced pair in @xmath1055 , which is not in the same column , indexes either two different columns or none .
there is at most one column containing the whole balanced pair and it appears always if @xmath85 is odd and only for @xmath1056 if @xmath85 is even , as shown in fig .
[ fig : corner ] .
thus , @xmath1058 is always odd for @xmath12 odd and even only if @xmath1059 for @xmath12 even .
let @xmath1049 denote the @xmath82th lowest length strip addable to @xmath51 and @xmath1048 the number of columns in it .
one can prove by the same method that @xmath1065 has an even number of columns , one of which is already in @xmath1049 and , consequently , @xmath1066 is odd . again by the same method it is possible to prove that @xmath1067 has an even number of columns .
this is because @xmath1068 contains a substrip @xmath1069 which can be obtained by moving down along the diagonal the strip @xmath1055 .
applying what was said above about @xmath1065 to @xmath1070 we see that @xmath1071 .
if @xmath51 is typical and @xmath1072 then @xmath1073 is the next partition in the block of @xmath51 , while if @xmath1074 then @xmath1075 is the next partition in the bloc of @xmath51 .
therefore we have just shown , as claimed in sec .
[ heart ] that @xmath1076 . | we define and study a lattice model which we argue is in the universality class of the @xmath0 supercoset sigma model for a large range of values of the coupling constant @xmath1 . in this first paper
, we analyze in details the symmetries of this lattice model , in particular the decomposition of the space of the quantum spin chain @xmath2 as a bimodule over @xmath0 and its commutant , the brauer algebra @xmath3 .
it turns out that @xmath2 is a nonsemisimple module for both @xmath0 and @xmath3 .
the results are used in the companion paper to elucidate the structure of the ( boundary ) conformal field theory . #
1#2#3#4 # 1#2#3#4 constantin candu@xmath4 and hubert saleur@xmath5 _ service de physique thorique , cea saclay , + gif sur yvette , 91191 , france@xmath4 + _ _ department of physics and astronomy , university of southern california + los angeles , ca 90089 , usa@xmath6 + _ |
since 1998 , a series of independent observations , including type ia supernova @xcite , weak lens @xcite , cosmic microwave background anisotropy @xcite , large scale structure @xcite , baryon oscillation @xcite , etc . , have confirmed that our universe is undergoing a period of accelerated expansion .
this result conflicts fiercely with the prediction of general relativity , and therefore triggers a `` golden age '' of cosmology since researchers in cosmology , astrophysics , general relativity , particle physics are all involved in seeking for viable models to explain this phenomena . within the frame of general relativity , the cosmic speed - up can be viewed as an indication that the present universe is dominated by certain mysterious fluid with large negative pressure , called `` dark energy ''
. however , such simple explanations could hardly be satisfactory .
the simplest @xmath3cdm model , which fits the observation data best so far , assumes a cosmological constant to be responsible for the dark content .
but the observed value of the constant @xmath3 is unnaturally smaller than any estimation by tens of orders @xcite .
other dynamical dark energy models then waive evolving scalar fields to replace and resemble the cosmological constant ( for review see @xcite ) .
these early attempts provide us with initial understanding of the complexity of the problem , though they chiefly serve for empirical fittings with comparatively poor theoretical motivation , and none of such models turn out to be problem - free . in view of the challenge and the fact that gravity is the only dominant long - range interaction insofar as we know , it is reasonable to consider the possibility that we have not fully understood the character of gravity on a cosmological scale .
therefore , the exploration of alternative theories of gravity are proposed .
one of such examples is the kaluza - klein ( kk ) theory , which was initiated by the motivation of unifying the gravitation field and the electromagnetic field in a five - dimensional ( 5-d ) metric .
since the fifth dimension is supposed to be a compact @xmath4 circle with an extremely tiny radius , it would actually yield no observable effect .
graceful though , the original version of kk theory fails in passing the solar - system experiment @xcite .
but recently a new model of modified ( non - compact ) kaluza - klein cosmology was investigated in @xcite , where the universe in turns inflates , decelerates , and then accelerates in respectively early times , radiation dominated era and matter dominated era .
this result meets the observational facts roughly .
but there still exist problems such as that it leaves no decelerated period in the matter - dominant era for the formation of large scale structure .
besides , in @xcite brans - dicke ( bd ) theory of gravity was introduced to match the mach s principle .
it is found in @xcite that 5-d bd theory can naturally predict the cosmic acceleration without the requirement of a time - varying bd parameter @xmath5 or a fabricated potential in its 4-d counterpart @xcite .
thus , it lends confidence and motivation to the effort of extending gravitational theories in 4-d to 5-d space - time .
also , note that in @xcite the allowed range for parameters @xmath6 is considerably wide , where @xmath5 is the bd parameter and @xmath7 represents the interaction between the extra dimension and other spatial ones .
we may suspect that further information concerning the different choices of @xmath6 are masked in the formalism of 5-d bd theory .
given the argument for the equivalence ( by evolutional effect ) of the @xmath0 theory of gravity to certain special cases of bd theory up to a potential term @xcite , we are led to consider the 5-d @xmath0 theories of gravity for such information .
it is worth noting that the scale of the compact extra dimension of 5-d gravity theories , which we are considering , is constrained to be less than 50 @xmath8 by current experiments @xcite .
the @xmath0 theories of gravity extends the hilbert - einstein action in general relativity to @xmath9 where @xmath0 is an arbitrary function of the curvature scalar @xmath10 .
the reason for such exploration in 4-d spacetime is twofold . on the one hand ,
recent research has shown that a small correction to the hilbert - einstein action by adding an inverse term of @xmath10 would lead to acceleration of the universe @xcite , and a large variety of models are proposed since then ( for review paper , see @xcite ) .
however , few of them turn out to be without problem and most of them admit fine - tunings so as to be consistent with experiments . on the other hand , it is highly possible that future quantum theory of gravity , such as string theory , loop quantum gravity , etc . , would bring about modifications to the action for classical general relativity , just like that general relativity has introduced corrections to the newton s gravitation theory .
therefore , this effort would potentially match the low energy effective theory of quantum gravity and therefore accelerate its birth . with above motivations
, we thought that modified gravity without fine - tuning should better be constructed in higher dimensions .
it is shown in this paper that values of the parameters in the 5-d theory are all with reasonable deductions and interpretations . in 5-d spacetimes with a hypersurface - orthogonal killing vector field ,
the 5-d @xmath0 theories of gravity are reduced to the sensible 4-d world in section [ killing ] .
here we have adopted kk s idea that the fifth dimension is a small unobservable compact ring @xmath4 , so that a killing vector field would arise naturally in the low energy environment .
further , in view of that in kk s theory one can find the sensible 4-d universe to be expanding ( without acceleration ) as a result of the contraction of the extra dimension , we assume that there is certain interaction between the fifth dimension and other ones at present , and take their relationship to be the general power - law form .
then the reduced theories are carefully studied in the @xmath11 case of the frw metric in section [ frw ] .
we consider a special class of models with @xmath1 and present the numerical simulation of the evolution history of the matter dominated universe in section [ numeri ] , demonstrating its consistency with the observation .
finally , we end with discussions in section [ jielun ] on the similarity and difference between the 5-d @xmath0 theories and the 5-d bd theory , as well as further indications from our 5-d models .
we start with the action in 5-d space - time : @xmath12 here @xmath13 , where @xmath14 denotes the gravitational constant in the 5-d spacetime , and @xmath15 represents the matter term in the total action @xmath16 .
variation of ( [ action ] ) with respect to the metric @xmath17 gives @xmath18 where @xmath19 , and @xmath20 .
note that we have employed the abstract index @xcite and @xmath21 represent quantities in the 5-d spacetime .
contracting equation ( [ 5dfield ] ) with @xmath17 , we obtain the dynamical equation for the scalar field @xmath22 : @xmath23.\ ] ] since here @xmath24=5 instead of 4 , the dynamical equation of @xmath22 differs from its usual 4-d counterpart .
next we consider the structure of the 5-d spacetime and its relation to our sensible 4-d world .
first , as mentioned in the introduction , we assume that the 5-d spacetime possesses a killing vector field @xmath25 which represents the fifth dimension and is everywhere space - like .
recall that the killing reduction of 4-d spacetime is studied by geroch @xcite and is further extended to the 5-d spacetime by yang _ et al .
therefore , following these works we introduce the 5-d metric as @xmath26 where @xmath27 is the metric in the usual 4-d universe and @xmath28 .
if @xmath25 is not hypersurface - orthogonal , it would be related to electromagnetic 4-potential in the reduced 4-d theory , as demonstrated by kk theory .
since we are chiefly concerned with the cosmological effect of the reduced model , our discussion is restricted to the case where @xmath25 is hypersurface - orthogonal for convenience . without losing generality ,
a coordinate system can be chosen as @xmath29 , with @xmath30 .
then the line element of @xmath31 reads @xmath32 , while ricci tensors and the dynamical equation of @xmath33 in 5-d and 4-d spacetime bear the following relation @xcite : @xmath34 @xmath35 here @xmath36 denotes the covariant derivative on 4-d spacetime and is defined as @xcite @xmath37 satisfying all the conditions for a derivative operator . on the other hand ,
the stress - energy tensor in ( [ 5dfield ] ) is regarded as a perfect fluid in 5-d spacetime with the expression @xcite @xmath38 = l^{-1}{\lambda^{-\frac{1}{2}}}\left[t_{ab}^{(4 ) } + p { \lambda}\xi_a\xi_b\right],\ ] ] where @xmath39 and @xmath40 are the 4-d energy density and the hydrostatic pressure respectively , and l is a constant representing the coordinate scale of the fifth dimension .
this extension of the stress - energy tensor of the 4-d perfect fluid to a 5-d one can be understood in the following way : because the fifth dimension is compact and attached to every point of the 4-d space - time , @xmath39 and @xmath40 experienced by a 4-d observer should be an integrated effect throughout the compact ring .
if we further expect that the fluid distributes homogeneously and does not travel along the fifth dimension , it is clear that @xmath41 , @xmath42 , and hence equation ( [ tab ] ) is obtained . combining ( [ 5dfield ] ) , ( [ 5ddynamics ] ) with ( [ riccireduction ] ) , ( [ lambdareduction ] ) , and ( [ tab ] ) , straightforward calculations lead to the 4-d field equation : @xmath43\\ & & + \frac{1}{f'(r ) } ( d_a d_b -h_{ab}d^c d_c)f'(r ) - \frac{1}{2 f'(r)}\lambda^{-1 } h_{ab } ( d^c{\lambda})d_cf'(r ) , \end{aligned}\ ] ] and the dynamical equations of @xmath33 and @xmath22 : @xmath44 , \end{aligned}\ ] ] @xmath45,\ ] ] where @xmath46 represents the usual 4-d gravitational constant @xcite .
in this section , we will study the cosmological predictions of the reduced @xmath0 gravity .
the metric of a spatially isotropic and homogeneous 4-d spacetime is the friedman - robertson - walker ( frw ) metric with three possible structures of the space . as suggested by the observation @xcite
, we only handle the spatially flat case where the 4-d line element reads @xmath47 then the two components of the field equation ( [ 4dfield ] ) are : @xmath48 and @xmath49 the dynamical equations of @xmath50 and @xmath22 are respectively : @xmath51 and @xmath52 here an over - dot denotes the derivative with respect to the proper time @xmath53 of the isotropic observer , and we have assumed that the scalar field @xmath50 depends only on @xmath53 . note that the conservation equation of the 5-d stress - energy tensor , @xmath54 , gives : @xmath55 which can also be obtained directly from equations ( [ gab1 ] ) ( [ phi1 ] ) . for the common baryonic matter content of the present universe ,
the pressure is negligible compared to the energy density .
thus , we set @xmath56 by approximation in the above equation and get @xmath57 , where @xmath58 is the current observed energy density of the luminary matter . by
now , we have been dealing with the general form of @xmath0 gravity .
next we restrict our discussion to a specific class of @xmath0 models , @xmath1 , to further study the evolutional characteristics of the theory .
it has been shown that this choice in 4-d spacetime , with @xmath59 and the common baryonic matter , could fit the hubble diagram of type ia supernovae without need of dark energy @xcite .
we denote @xmath60 and obtain @xmath61 where @xmath62 is a dimensional constant . with the above analysis
, we find the three cosmological evolution equations for numerical simulations from the combination of equations ( [ gab1 ] ) ( [ phi1 ] ) : @xmath63 @xmath64 @xmath65
let us now find the natural initial values of @xmath66 for the numerical simulation .
firstly , @xmath67 themselves have no direct physical meaning and therefore can be simply fixed as 1 ( with no dimension )
. then we can directly determine the value of @xmath68 by the present value of hubble parameter @xmath69 .
secondly , we adopt the dynamical compactification idea of kk cosmology that the contraction of the extra dimension would result in the expansion of the remaining dimensions @xcite , and assume @xmath70 where for a particular @xmath0 theory with given @xmath2 and @xmath62 in eq.([rphi ] ) , the free parameter @xmath7 can be constrained by the observation .
therefore we have @xmath71 .
finally , comparing the matter term in equation ( [ 4dfield ] ) to the field equation of general relativity , we expect @xmath72 , at least for the present period ; thus @xmath73 ( with no dimension either ) and further @xmath74 . in summary ,
the initial conditions are : @xmath75 then we estimate the value of @xmath62 in equations ( [ ai ] ) ( [ phii ] ) and determine the allowed range for the pair of parameters @xmath2 and @xmath7 .
a calculation of the 5-d curvature scalar shows @xmath76 inserting ( [ rphi ] ) , ( [ lambdai ] ) into the expression of @xmath10 and noting that the deceleration parameter @xmath77 is an observable with its present value denoted as @xmath78 , we have @xmath79 h_0 ^ 2.\ ] ] substituting ( [ alpham ] ) into ( [ ai ] ) , we obtain the relationship between parameters @xmath2 and @xmath7 as follows : @xmath80 where @xmath81@xcite and therefore @xmath82 .
one necessary and sufficient condition for the present accelerated expansion of the universe is that the deceleration parameter @xmath83 .
specifically , the allowed range for @xmath78 according to the present observation is @xmath84 @xcite . here
we apply this criteria to determine the allowed range for @xmath2 and @xmath7 , as depicted in fig .
[ range1 ] . allowed range for the values of m and n with @xmath85 @xcite . the horizontal axis stands for different value of @xmath7 , while the vertical axis represents @xmath2 . ,
width=491,height=264 ] from the figure , it is obvious that there do exist suitable parameters @xmath2 and @xmath7 , that is , suitable formalisms of @xmath1 and corresponding relationships between the extra dimension and the other ones , such that the present cosmic acceleration can be explained without the need of dark energy . besides , fig .
1 also reveals that both positive and negative values of @xmath7 is allowed , indicating that the idea of dynamical compactification can be extended to that both the expansion and the contraction of the extra dimension can be responsible for the accelerated expansion of the other dimensions . besides , we know from equation ( [ alpham ] ) that @xmath86^{-(m-1)}.\ ] ] and the possible range for @xmath62 . the range is spanned by the uncertainty of the deceleration parameter @xmath78.,width=491,height=264 ] using eqs .
( [ mvalue ] ) and ( [ alpha ] ) , we depict the @xmath87 relation in fig .
[ alphan ] with the range spanned by @xmath78 , from which we find @xmath62 increases rapidly as @xmath88 increases .
further , one could find from fig .
[ range1 ] or equation ( [ mvalue ] ) that the lower bound of @xmath2 is 1 , corresponding to @xmath89 , and therefore @xmath90 , which is exactly the kk case .
one subtlety concerning this case is that fig .
[ range1 ] as well as equation ( [ mvalue ] ) shows that @xmath91 for @xmath92 , and therefore equation([kk ] ) apparently indicates @xmath93 , contradicting with the present observations .
thus one might think that the kk theory is inconsistent with the observation .
however , this conclusion is misleading , because from the inverse - solving procedure in equation ( [ rphi ] ) , we have already confined our discussion to @xmath94 cases . on the other hand ,
the asymptotic value of @xmath2 , as @xmath7 approaches infinity , is @xmath95 from which we find the upper bound for @xmath2 to be @xmath96 .
hence , the possible range for @xmath2 is narrowed down as @xmath97 . with equations ( [ ai ] ) ( [ phii ] ) , initial values ( [ ini ] ) and ( [ alpham ] ) , the numerical simulation of the evolution of the scale factor @xmath98 , the scalar field @xmath99 and the deceleration parameter @xmath100 can be conducted . as an example , we choose @xmath101 . then from equation ( [ mvalue ] ) we get @xmath102 .
the corresponding value of @xmath103 is approximately @xmath104 .
it is useful to define the effective gravitational `` constant '' from eqs .
( [ 4dfield ] ) ( [ 4dfr ] ) as @xmath105 note that @xmath106 is actually an evolving scalar .
then the evolution characters of @xmath107 and @xmath108 are illustrated in figs .
[ m15n54q06aq ] [ m15n54q06 g ] respectively . from these figures , it is clear that the present deceleration parameter @xmath102 is in consistent with observation and the above analysis .
moreover , we note the following : 1 . @xmath100 rolls from a positive value to a negative one smoothly in the recent past .
specifically , if the age of the universe is @xmath109gyr @xcite , the universe turns from deceleration to acceleration at @xmath110gyr .
thus , the `` coincidence '' problem in dark energy can be addressed in this case .
after @xmath100 reaches the minimal value , it rolls back again and becomes positive at @xmath111gyr .
thus , the universe will become decelerating in the future rather than be endlessly accelerating .
numerical simulation for a long period of cosmic evolution shows that the value of @xmath100 becomes divergent at @xmath112 , as depicted in fig .
[ diverge ] .
since @xmath113 and its evolution is generally slow , the only reason for the divergence of @xmath100 is @xmath114 . in other words ,
the universe would come to a static point in the finite future .
actually in this period , numerical simulation shows that all quantities , such as @xmath115 and @xmath116 , still remain well - behaved .
after this static point , @xmath98 starts to decrease for a certain period before numerical simulation fails .
thus , around the static point of @xmath98 , numerical simulation is still effective .
@xmath115 decreases from a large value in the past and increases slowly at present , while @xmath116 increases to a maximum value for the present period and decreases slowly in the future .
after the static point of @xmath98 , nonetheless , the decrease of @xmath116 becomes faster and therefore it would approach @xmath117 in the finite future when the numerical simulation fails . actually , since @xmath118 appears in the denominators of eqs .
( [ ai ] ) ( [ phii ] ) , the analysis from this set of equations is no longer applicable for the evolution character beyond the @xmath119 point .
instead , one should take advantage of original equation ( [ 5dfield ] ) ( or eqs .
( [ 4dfield ] ) ( [ 4dfr ] ) by multiplying @xmath22 on both sides ) .
we leave the details about the static point and the evolution beyond to future investigations .
the evolution of @xmath106 is generally slow in the future .
the present value of @xmath106 is maximal with an almost zero evolving speed .
thus , it is also in consistent with the observation data that the present gravitational constant has a negligible evolving speed . on the other hand , as indicated by fig .
[ range1 ] , the expansion of the extra dimension would also bring about the acceleration of other dimensions .
if we take @xmath120 instead of @xmath121 , the numerical simulation will give similar evolution features of @xmath122 and @xmath123 with the present value of deceleration parameter @xmath102 , which are consistent with observations .
in this case , the universe turned from deceleration to acceleration at @xmath124gyr , which was @xmath125gyr earlier than that in the @xmath126 case .
correspondingly , the future transition point of the universe from acceleration to deceleration is also moved ahead to @xmath127gyr , rendering that the total acceleration period of the universe changes only moderately from @xmath128gyr to @xmath129gyr .
thus , the @xmath130 case is very similar to the @xmath131 case , except that the cosmic transition from deceleration to acceleration and the inverse occur both ahead of time .
the reason for such temporal translation can be viewed in the following way . as shown before , when @xmath97 , the future asymptotic evolution of @xmath33 should be increasing .
therefore , if the initial ( present ) evolution rate for @xmath33 is set to be increasing at present rather than be decreasing , there should be a temporal translation with respect to the evolution of @xmath33 and therefore to other quantities that can be affected .
to clarify the argument for the equivalence of the @xmath0 theory of gravity to certain special cases of bd theory up to a potential term @xcite , we now compare the cosmic evolution of 5-d @xmath0 models developed in previous sections to the 5-d bd theory @xcite . from equation ( [ 4dfield ] ) , it is noticeable that the potential term is @xmath132 , and similar terms also exist in equations ( [ 4dlambda ] ) and ( [ 4dfr ] ) .
these terms correspond to the @xmath133 terms in ( [ ai ] ) ( [ phii ] ) , and would be shown to play significant roles in predicting the future evolution of the universe .
there are three independent functions in the 5-d bd theory @xcite , namely @xmath98 , @xmath115 and @xmath116 , where @xmath116 represents the core idea of bd theory on an evolving gravitation `` constant '' due to the interaction between the local gravitational fields with the faraway matter in the universe .
we note that @xmath100 in this case ( see fig
. 2 in @xcite ) evolves to a constant in the future .
this means that the universe would accelerate permanently in the future until its energy density decreases to be extremely low so that certain quantum effects may become significant , such as the situation in @xcite .
it can also be read from figs .
3 5 in @xcite that the asymptotic values of @xmath134 and @xmath135 are both zero , as further illustrated in fig .
[ ratebd ] .
in contrast , in our 5-d @xmath0 theories , fig . [ m15n54q06phi ] has demonstrated that , while @xmath118 evolves to a large value presently and remains positive in the future , the @xmath133 terms render the evolution of @xmath99 to be more complicated .
specifically , the corresponding evolution characteristics of @xmath136 are illustrated in fig .
[ ratefr ] with specific parameters @xmath101 . as
@xmath137 is an observable , it is interesting to show how @xmath133 terms affect its evolution .
since @xmath138 , @xmath83 , from equation ( [ alpham ] ) we have @xmath139 .
moreover , we do not consider any solution with a negative @xmath118 because of its anti - gravity character .
thus , the coefficient of the @xmath133 term would determine the evolutional inclination of @xmath33 and @xmath118 .
specifically , because @xmath140 and @xmath141 correspond to @xmath142 and @xmath143 respectively , the qualitative features of the future evolution of @xmath33 and @xmath118 are as follows : * if @xmath144 , @xmath145 * if @xmath146 , @xmath147 * if @xmath148 , @xmath149 in view of the constraint illustrated by fig .
[ range1 ] , only the last situation is allowed , where the evolution of @xmath33 and @xmath118 counterbalances each other and therefore the evolution of @xmath106 is generally slow . in summary , we present the killing reduction of 5-d @xmath0 theories of gravity to the 4-d sensible world
. then we study its cosmological implication by assuming the spatial homogeneity and isotropy , namely , the friedman - robertson - walker metric . with the illustration of a specific example it is found that the theory is consistent with the present observations in a variety of aspects , including the recent cross from cosmic deceleration to acceleration , its prescription of present cosmic speed - up , the negligible evolving speed of the effective gravitational constant , and etc .. in contrast to the kk cosmology , it is worth mentioning that in 5-d @xmath0 theories of gravity , both expansion and contraction of the extra dimension could result in the present accelerated expansion of other spatial dimensions . even for a simple and generic class of @xmath1 models , the 5-d @xmath0 theories of gravity do not need unreasonable or ill - initiated fine - tuning of certain parameters .
hence it is reasonable to infer that the present accelerated expansion of spatial dimensions could be certain basic character of the 5-d spacetimes . finally , it is pointed out in ref .
@xcite that in the large scale limit , quantum gravity effect would come to be vital and bring about higher - order quantum corrections to the friedman equation to prescribe a static point in the finite future , followed by the recollapse of the universe .
note that the static point in the finite future is also a significant character of the 5-d @xmath0 gravity .
thus , it is of interest to explore whether there are some relation between the 5-d theories of @xmath1 models and the quantum gravity theory applied in @xcite .
in addition , the future recollapse also differs from the prediction of endless acceleration scenario in the 5-d bd cosmology @xcite .
hence , 5-d @xmath0 gravity is not simply equivalent to 5-d bd theory , since the effect of potential terms are shown to be vital enough to prescribe different cosmological scenarios in the distant future .
the authors would like to thank li - e qiang for valuable discussions .
this work is a part of projects 10675019 and 10975017 supported by nsfc .
biao huang would also like to acknowledge support from the beijing undergraduate research and entrepreneurship foundation .
arabic a. g. riess _
et al . _ ,
j. * 116 * , 1009 ( 1998 ) s. perlmutter _
et al . _ ,
j. * 517 * , 565 ( 1999 ) j. l. tonry _
et al . _ ,
j. * 594 * , 1 ( 2003 ) b. jain and a. taylor , phys .
rev . lett . *
91 * , 141302 ( 2003 ) d. n. spergel _
et al . _ ,
. ser . * 148 * , 175 ( 2003 ) m. tegmark _
et al . _ , phys .
d * 69 * , 103501 ( 2004 ) u. seljak _
d * 71 * , 103515 ( 2005 ) d. j. eisenstein _
_ , astrophys
. j. * 633 * , 560 ( 2005 ) c. blake , d. parkinson , b. bassett , k. glazebrook , m. kunz , r. c. nichol , mon . not .
. soc . * 365 * , 255 ( 2006 ) s. m. carroll , living rev
. relativ .
* 4 * , 1 ( 2001 ) p. j. peebles and b. ratra , rev .
mod . phys . * 75 * , 599 ( 2003 ) e. j. copeland , m. sami , and s. tsujikaa , hep - th/0603057v3 ; p. j. peebles and b. ratra , rev .
phys . * 75 * , 599 ( 2003 ) p. xu and y. ma , phys . lett .
b * 656 * , 165 ( 2007 ) f. darabi , gr - qc/0901.0835v1 c. brans and r. h. dicke , phys .
d * 124 * , 925 ( 1961 ) l. qiang , y. ma , m. han and d. yu , phys .
d * 71 * , 061501 ( 2005 ) n. banerjee and d. pavon , phys .
d * 63 * , 043504 ( 2001 ) s. sen and a. a. sen , phys .
d * 63 * , 124006 ( 2001 ) t. chiba , phys . lett .
b * 575 * , 1 ( 2003 ) s. m. carroll , v. duvvuri , m. trodden , and m. s. turner , phys . rev .
d * 70 * , 043528 ( 2004 ) s. nojiri , int .
phys . * 4 * , 115 ( 2007)hep - th/0601213v3 r. m. wald , _ general relativity _
( university of chicago press , chicago , 1984 ) r. geroch , j. math .
( n. y. ) * 12 * , 918 ( 1971 ) x. yang , y. ma , j. shao , and w. zhou , phys .
d * 68 * , 024006 ( 2003 ) y. ma and j. wu , int
. a * 19 * , 29 ( 2004 ) d. n. spergel _
_ , astrophys
. j. suppl .
* 170 * , 377 ( 2007 ) s. capozziello , v. f. cardone , s. carloni and a. troisi , int . j. mod .
d * 12 * , 1969 ( 2003 ) t. appelquist , a. chodos , and p. g. o. freund ( ed . ) , modern kaluza - klein theories , frontiers in physics series vol.*65*(addison - wesley , reading , ma , 1986 ) . j. frieman , m. turner , d. huterer , ann .
. astrophys . * 46 * , 385 ( 2008 ) y. ding , y. ma , j. yang , phys .
* 102 * , 051301 ( 2009 ) | the metric @xmath0 theories of gravity are generalized to five - dimensional spacetimes . by assuming a hypersurface - orthogonal killing vector field representing the compact fifth dimension , the five - dimensional theories are reduced to their four - dimensional formalism .
then we study the cosmology of a special class of @xmath1 models in a spatially flat frw spacetime .
it is shown that the parameter @xmath2 can be constrained to a certain range by the current observed deceleration parameter , and its lower bound corresponds to the kaluza - klein theory .
it turns out that both expansion and contraction of the extra dimension may prescribe the smooth transition from the deceleration era to the acceleration era in the recent past as well as an accelerated scenario for the present universe .
hence five - dimensional @xmath0 gravity can naturally account for the present accelerated expansion of the universe .
moreover , the models predict a transition from acceleration to deceleration in the future , followed by a cosmic recollapse within finite time .
this differs from the prediction of the five - dimensional brans - dicke theory but is in consistent with a recent prediction based on loop quantum cosmology .
pacs numbers : 04.50.-h , 98.80.es |
astrophysical uhe neutrinos is being searched , since four decades , at tev - pev energy via neutrino muons in underground detectors .
their signal is greatly polluted from above ( and below ) by cr secondary muons ( and atmospheric muon neutrinos into muons ) and partially suppressed at pevs , by earth size opacity .
horizontal muons at hundreds tev , originated by astrophysical uhe neutrinos are also polluted by leading prompt atmospheric signals . however , since four decades , the rise of highest energy neutrino astronomy is generally expected also by the cosmogenic eev neutrinos , secondaries of gzk @xcite,@xcite cut - off on ultra high cosmic rays ( uhecr ) .
these ones are difficult to observe in @xmath5 detector .
but since a decade @xcite the uhe tau neutrino astronomy have been foreseen and suggested in present and future detectors .
indeed tau air - shower astronomy is an amplified neutrino astronomy observable .
indeed an air - shower is not a single track , but a wide ( even many @xmath6 ) area raining of billions secondaries at once . one does nt need to collect and count all of them to realize such an event beyond a mountain or earth screen .
one need just a spread sample at once . as for uhecr .
moreover hadronic air - shower below zenith angle @xmath7 are originated far and far and they are filtered : their peak electromagnetic component is exhausted at horizons leaving just late diluted ( and sharp ) inclined muon bundles . on the contrary neutrino , either interacting in air or , better , escaping the earth by @xmath8 may lead to electromagnetic - rich horizontal ( upward ) air - showers . tau decay in flight are better than neutrino interacting in air because earth rock density is three thousand times larger than air one . even if the tau enjoy of a bounded escaping solid angle ( a zenith angle width of @xmath9 degree ) while all down - ward neutrinos interacting on air may reach from wider cone ( a zenith angle width @xmath10 degree ) and are in three flavors .
such skimming events in auger experiment may rise via upward tau air - shower @xcite,@xcite,@xcite . in last years
the upward - horizontal eev @xmath11,@xmath12 appearance , via uhecr @xmath13 has been predicted by many authors ; the most extreme ones were at rate of @xmath14 a year@xcite , or @xmath15 a year in auger @xcite and finally up to @xmath16 a year @xcite .
this rate , has only recently being adjusted and confirmed by last auger group estimates ( now 2008 , cris 2008):@xmath15 a year , in full agreement with our previous ones @xcite .
indeed following the auger evidences ( and hires ones @xcite ) of an uhecr gzk@xcite,@xcite cut - off and the latest auger ( possible ) super - galactic anisotropy due to an eventual proton uhecr guarantee a secondary flux of uhe - gzk neutrino at eevs energy within auger detection via @xmath11,@xmath12 showering @xcite . in fact this occurs because the muon neutrino flavor mixing must feed also a tau neutrino component .
such uhe astrophysical tau neutrino ( noise - free from any atmospheric background ) may interact in and it may rises out the earth as uhe @xmath11 . the uhe @xmath11,@xmath12 decay in flight in atmosphere must lead to loud tau air - showers .
such a detectable flashes may rise in short times within auger sd ( by large electromagnetic curvature signals ) or in fd arrays by horizontal fluorescence signals , namely once in a few years ( @xmath17 ) from now @xcite .
nevertheless a recent alternative uhecr understanding @xcite , based on observed auger uhecr ( nuclei ) mass composition and with cen - a rich clustering map , is in disagreement with uhecr proton understanding @xcite .
this model is leading to different uhe neutrino predictions .
it suggests that uhecr are made by lightest nuclei ( @xmath0 , @xmath1 , maybe a few also @xmath18,@xmath19 ) mostly originated from cen - a : their trajectories are bent and spread by galactic magnetic fields and they are incidentally clustered ( by galactic fields ) around cen - a , nearest ( @xmath20 mpc ) and possibly unique source able to survive the short lightest nuclei gzk cut - off .
these events spread mostly along the same super - galactic arm , just apparently from far @xmath21 mpc centaurs cluster .
the mean random angle bending @xmath0 ( see fig.[uhecr - map ] ) by spiral galactic magnetic fields along the plane is @xmath22 : @xmath23 , li , be ) will be bent more .
the lighter and less energetic @xmath24 less . , width=226 ] this _ lightest nuclei for highest cosmic rays _
model implies and foresees among the other , additional clustering of uhecr events around the nearest agn cen - a ( the lightest uhecr the more correlated to the source , the heavier and with larger charges , the most bent and spread ones ) .
the model explains the absence or a poor signal from virgo ( too far for the fragile nuclei to fly by ) .
lightest nuclei for highest cosmic rays _ are forced in a very confined cosmic volume ( ten mpc ) due to a fragile light nuclear ( few mevs ) binding energy .
usually heavy nuclei fragmentation pour energy only in uhe neutrino ( at @xmath25 eev energy ) spectra , less energetic than common uhe eev @xmath26 neutrino flux .
auger sd or fd are not able to reveal tens - hundreds pev energy easily .
however a very recent and a less spaced auger sub - system , a more dense array amiga and the additional telescopes heats might lower the threshold accordingly .
the @xmath28 mostly hadronic , inclined-_upward _ , tau air - showers @xmath29 occur at much lower altitudes than hadronic inclined _ down_-going air - showers , at much nearer distances from telescopes than hadronic eev air - showers , reducing the area and the rate .
they may offer a tens - hundred pev neutrino windows , secondaries of uhe he nuclear fragmentation .
some estimates are offered assuming , for sake of simplicity a fermi - like uhe gzk spectra , comparable to the well known waxmann - bachall flux derived for cosmic grbs . nevertheless if rarest uhecr , possibly from agn , above @xmath30 , ( a few in auger events ) , are also he nuclei , they may still suffer gzk photo - pions opacity decaying into eev neutrinos too .
such he gzk interaction may be comparable with corresponding proton gzk ones at @xmath31 : therefore eev tau neutrino astronomy , tails of the most energetic lightest nuclei gzk secondaries , may still rise at auger .
eev inclined tau decays , born nearly at sea level ( versus inclined hadronic @xmath32 ones developing at altitudes well above ten km ) occur at air - density at least three times higher ( than hadron horizontal ones ) ; therefore tau air - showers ( mostly behaving as hadron ones ) , their electromagnetic and fluorescence tail sizes ( and times ) are three times shorter ( @xmath33 km ) and their characteristic azimuth speed @xmath34 is slower than common inclined uhecr hadron ones at higher altitudes ( @xmath35 km ) , ( @xmath36 km ) , whose detection threshold ( minimal impact distance ) is very near to the observer ( a dozen of km from the edges and few kms from air - shower baricenter ) .
the consequent azimuth angular velocity is for instance , assuming an eev event , at maximal distance of @xmath37 km , at minimal distance @xmath38 km , total duration 100 @xmath39 s , @xmath40 .
such a brief @xmath41s and distinct up - ward tau lightening versus slow @xmath42s diluted down - ward horizontal showering may be well disentangled within auger ( and amiga sd ) threshold , also by obvious angular resolution and directionality ( up or downward ) ; their strong curvature and their inclinations would rise within heat fd and auger fd , we estimate , once every a couple of years [ array ] .
possibly ( factor two ) originated from the ande side ( west to east ) @xcite,@xcite,@xcite . and the 23.5 @xmath6 ) offer a lower energy threshold for neutrino inclined air - showers respectively at @xmath43 and @xmath44 ev . in the figure a schematic area due to air - showering lobes of an escaping horizontal air - shower , width=226 ] this new neutrino astronomy , the pevs - eevs ratio , may disentangle in future records the real uhecr nucleon or lightest nuclei uhecr nature . indeed the additional rise of the resonant glashow contribute , @xmath45 , in the upgoing tau , while just marginally doubling the signal at 0.01 eev , see fig.[nulenght],[differential],[integral ] it may leave a clear imprint on horizontal differential angular spectra . leading to a flavor uhe neutrino spectroscopy . the role of uhe neutrino interaction with matter
is well understood : they also shape the neutrino survival across the earth .
indeed the highest energy @xmath4 are opaque to earth , but not to smallest cord .
therefore the harder events are the more opaque and survive at shortest cords , the tangent ones .
the resonant @xmath46 are very peculiar signals .
their opacity on earth is extreme .
they are as opaque as eev energetic neutrinos.see [ nulenght ] .
but their lower energy corresponds to higher flux ( for constant energy fluency as waxmann bachall one ) .
but their propagation in earth is much smaller too .
the compensated flux in 5 - 6 times higher than eev one .
unfortunately this rate is ( in auger ) suppressed by detection threshold ( nearly one hundred times smaller than eev showers ) .
nevertheless the mini array amiga ( of 6 @xmath6 ) and the heat telescopes may contribute to make detectable in a few year the same resonant @xmath46 : the distance will be ten times smaller and the showering azimuth angular velocity will appear @xmath47 times faster than already fast eev tau airshower . their curvature in amiga may leave a clear imprint in dense array cherenkov ( and in mostly absent underground ( 3 m ) muon one ) .
, for each neutrino flavor , derived from the observed uhecr gzk@xcite,@xcite at @xmath48 ev .
the inner earth density profile has been taken into account .
note the very peculiar bump due to the resonant antineutrino at @xmath49 pev ; its preferential inclination is due to the severe earth opacity .
the overall integral signal almost double the tau skimming rate in the same energy band , width=226 ] , for each neutrino flavor , derived from the observed uhecr gzk@xcite,@xcite at @xmath48 ev .
the inner earth density profile has been taken into account .
note the very peculiar bump due to the resonant antineutrino at @xmath49 pev ; its preferential inclination is due to the severe earth opacity .
the overall integral signal almost double the tau skimming rate in the same energy band , width=272 ] efficiency for fd .
the fd threshold is growing linearly with energy and it is respectively @xmath50 , @xmath51 , @xmath52 at the energy @xmath53 , @xmath28 , @xmath54 eev . because of it each rate at lower energies than eev must be suppressed in fd efficiency by corresponding factor ( 0.1 , 0.01 ) making the expected event rate at ten pev ( @xmath55 ) and hundred pev ( 0.5 ) , just below unity in 3 years .
the additional mini - array amiga at @xmath56 , @xmath57 , whose array spacing is respectively , @xmath58 , @xmath59 m. is a sd active day and night and it might double the signal , offering a detection in a few years from now ( 0.35 - 0.4 event / year).,width=302 ]
the very exciting neutrino astronomy is beyond the corner .
either by eev gzk @xcite,@xcite secondaries or by tens pev signals the detection via fd in auger is at the detection edge .
the short and near air - shower from the telescope due to the lower energy makes the shower timing brief , sharp and at small zenith angle , and , upgoing . on the contrary
the unique inclined horizontal down - going hadron shower are far away , diluted in air density and in time scale as well as at possible bifurcate ( by geomagnetic field ) structure . no way to be confused .
three or more times duration and morphology of _ up or down _ signature will disentangle any rare neutrino lights arriving from their unique but unusual sky : the earth .
their rate , timing , energy and inclination may teach us on the real nature ( nucleon or lighters nuclei ) of uhecr . in this hope we suggest ( a ) to implement the present auger array with additional array searching for shadows of inclined hadronic from ande in auger ; ( b ) to add mini - cherenkov telescope arrays facing ande screen to reveal tau air - showers signals beyond the mountain chain also via direct cherenkov flashes.(c ) locating a new telescope array few km distance to reduce the energy thresholds ( d ) introducing novel trigger detection by fast horizontal track time signature.(e ) locating sd nearest to fd in order to enlarge their ability to discover the inclined events ( blazing cherenkov flashes ) both in fd and in cherenkov lights as well as in muon tracks in nearby sd array elements .
it must be remind that the amiga muon versus sd cherenkov signal may well signal the electromagnetic versus muonic nature of horizontal air - shower , offering an additional tool to disentangle neutrino versus hadronic air - showers .
we foresee in very near years , possibly this and next one , the tau neutrino astronomy birth , if accurate attention on trigger thresholds , and amiga - heats implementation will be concluded soon . | uhecr may be either nucleons @xcite or nuclei ; in the latter case the lightest nuclei , as @xmath0 or @xmath1 , li , be , explains at best the absence of virgo signals and the crowding of events around cen - a bent by galactic magnetic fields @xcite .
this model fit the observed nuclear mass composition discovered in auger .
however uhecr nucleons above gzk produce eev neutrinos while heavy nuclei , as fe uhecr do not produce much .
uhecr he nuclei at few @xmath2 suffer nuclear fragmentation ( producing low energetic neutrino at tens pevs ) but it suffer anyway photo - pion gzk suppression ( leading to eev neutrinos ) once above one - few @xmath3 .
both these cosmogenic uhe secondary neutrinos signals may influence usual predicted gzk @xcite,@xcite tau neutrino astronomy @xcite in significant and detectable way ; the role of resonant antineutrino electron - electron leading to tau air - shower may also rise . |
the system we analyze is an open spin - chain of @xmath0 elements , whose hamiltonian reads @xmath1+\sum^n_{i=1}b(t)\hat{z}_i.\ ] ] here , @xmath2 and @xmath3 are the coupling strengths of the pairwise interaction between adjacent spins and @xmath4 is a magnetic field . in our notation ,
@xmath5 and @xmath6 denote respectively the @xmath7 and @xmath8-pauli matrix . for the sake of simplicity , we consider @xmath0 as an odd number .
however , all the results presented here can be straightforwardly adapted to the case of an even number of spins in the chain .
physical units are chosen throughout the article so that @xmath9 .
it is important to note that the inter - spin couplings and the amplitude of the magnetic field are site independent .
this is a feature that differenciates the system at hand from a few previous proposals for perfect state transfer available in the literature @xcite .
although eq .
( [ modelloxy ] ) is spatially unmodulated ( in analogy with @xcite ) , the price to pay in order to achieve unit transfer fidelity for any length of the chain is the time dependence of the interaction strengths .
we will show later in this article that such a request can also be relaxed : perfect state transfer can be achieved even if only the magnetic field is time - dependent while any other term is constant .
( [ modelloxy ] ) does not preserve the number of spin - excitations as it does not commute with the total @xmath8-component of the spin of the system .
frequently , when one faces the case of dynamics ruled by spin - preserving models , on the assumption of proper boundary conditions , it is convenient to diagonalize the coupling hamiltonian by means of a sequence of operation comprising wigner - jordan , fourier , and bogoliubov transforms @xcite . in our case , due to the spin - non preserving nature of @xmath10 , the mutual coupling among sectors of the hilbert space labeled by different quantum spin numbers has to be considered . rather than applying techniques for the exact diagonalization of eq .
( [ modelloxy ] ) , here we tackle the evolution of the system by means of an _ information - flux approach _ , which is specifically designed for multi - spin interactions @xcite .
our method does not rely on the explicit analysis of the energy spectrum of the chain and allows us to gather an intuitive picture of the dynamics at hand .
such an approach has already proved its flexibility and efficiency in identifying hamiltonian configurations suitable for state transfer and entanglement generation @xcite .
the information - flux approach requires the time - evolved form of specific operators @xmath11 in the heisenberg picture , which is given by @xmath12 with @xmath13 the time evolution operator , @xmath14 and @xmath15 . using this method ,
one is able to understand the dependence of @xmath16 on any @xmath17 and design the proper set of interactions in such a way that it becomes possible to _ drive _ a desired evolution by means of engineered quantum interference @xcite . in our case
, we will exploit the time dependence of the couplings in order to induce precisely this effect .
our task here is the transmission of quantum information from the first spin to the last one in the chain .
this requires the study of @xmath18 s , which can be decomposed into the operator basis built out of all possible tensor products of @xmath19 with @xmath20 the @xmath21 identity operator applied to spin @xmath22 .
we have @xmath23 where @xmath24 is the initial state of the whole chain , the first spin being in @xmath25 and the rest of the chain in @xmath26 ( this definition can be easily generalized to the case of mixed states of the chain ) .
the coefficient @xmath27 is defined as the information flux at time @xmath28 from @xmath29 to @xmath30 @xcite .
if our system achieves @xmath31 when @xmath32 and for @xmath33 , we have perfect @xmath34 state transfer .
we can determine which are the terms of the operator basis that are involved in the decomposition of @xmath18 s by representing such operators in an oriented weighted graph , whose construction is straightforward .
each node of the graph corresponds to an operator involved in the decomposition .
if the interest is focused on the evolution of @xmath30 , it should occupy the first node of the graph .
any operator resulting from the commutators of @xmath30 with the hamiltonian should be included in the graph and linked to it with an oriented edge .
these steps are repeated as necessary and the construction process ends when no new operator is created on commutation .
a numerical weight , determined by the performed commutations and needed in order to calculate the information flux , is attached to each edge of the graph .
finally , an outgoing ( incoming ) edge corresponds to a + ( - ) sign .
due to the particular form of @xmath35 in eq .
( [ modelloxy ] ) , @xmath36 and @xmath37 will generate the oriented graph shown in fig .
[ doublechain ] ( where the case @xmath38 has been considered ) .
clearly , @xmath39 already appears in the graph corresponding to @xmath40 and we do not need to construct a dedicated graph for it . for the sake of clarity
, here we show how to evaluate the information flux for a time - independent hamiltonian such as eq .
( [ modelloxy ] ) for @xmath41 .
the method can be explicitly extended to any time - dependent model . however ,
in our numerical analysis we divide the total time of the evolution in a series of short steps within which the hamiltonian is taken as constant . in this way , in the limit of a large number of steps and correspondingly short time steps , the flux of information can be accurately evaluated by means of only a light computational effort . a cut - off in the number of considered temporal steps
is chosen so as to have no noticeable difference in the results if a larger number is taken . for a time - independent hamiltonian , @xmath42 and , by means of the operator expansion formula @xcite
, we have @xmath43 with @xmath44 the nested commutator of order @xmath45 between @xmath10 and @xmath46 and @xmath47 . here ,
for the sake of a clear presentation of our proposal and without loss of generality , we focus on the case of @xmath38 ( the case of an arbitrary @xmath0 being easily extrapolated from our example ) and study the dynamics of @xmath48 under the model in eq .
( [ modelloxy ] ) .
the terms in the operator expansion are deduced by @xmath49=-2i\,j_y\,\hat{y}_4\hat{z}_5 + 2i\,b\,\hat{y}_5,\\ \hat{\cal c}_2(\hat{x}_5)&=[{\hat{\cal h}},[{\hat{\cal h}},\hat{x}_5]]=2ib(2i\,j_x\,\hat{x}_4\hat{z}_5 - 2i\,b\,\hat{x}_5)\\ & -2i\,j_y(2i\,j_y\,\hat{x}_5 + 2i\,j_x\,\hat{x}_3\hat{z}_4\hat{z}_5 - 2i\,b\,\hat{x}_4\hat{z}_5),\\ & \vdots \end{split } \label{commutators}\ ] ] clearly , the only operators involved in this iterative sequence are those in the graph of fig . [ doublechain ] .
therefore , it is possible to write the evolved operators @xmath50 as @xmath51 when the parameters @xmath52 can not be analytically evaluated due to the difficulties of the evolution , it is still possible to approximate them by means of recurrence formulas .
we have @xmath53 where @xmath54 is a proper cut - off and @xmath55 with @xmath56 ( @xmath57 ) for @xmath58 ( @xmath59 ) .
it is immediate to note that these recurrence formulas come directly from the commutation rules in eq .
( [ commutators ] ) .
they can thus be easily derived from the graph in fig .
[ doublechain ] . the same formulas , but with the initial conditions @xmath56 ( @xmath57 ) for @xmath60 ( @xmath61 ) , describe the evolution of the operator @xmath62 .
the approach described so far constitutes the basis for the analysis of perfect quantum state transfer via temporally controlled information kicking , which is the focus of the next section .
let us study the action of @xmath63 in eq .
( [ modelloxy ] ) when only one of the inter - spin coupling terms differs from zero .
the graph in fig .
[ doublechain ] shows that , in these conditions , every node is linked to just one nearest neighbor .
in fact , the full oriented graph for the information flux analysis corresponding to eq .
( [ modelloxy ] ) can be seen as the juxtaposition of three mutually disconnected subgraphs , each associated with only one of the coupling terms in @xmath10 and represented by a different color in fig .
[ doublechain ] .
it is easy to track the operator dynamics associated with these configurations .
let us consider , for instance , the case @xmath64 and @xmath65 .
[ doublechain ] ( black color ) shows that @xmath48 and @xmath62 are mutually linked and , by using the formal quantitative analysis of the previous section , we have @xmath66\hat{x}_5+\sin[2\beta(t)]\hat{y}_5,\\ & \hat{\cal y}_5(t)=-\sin[2\beta(t)]\hat{x}_5+\cos[2\beta(t)]\hat{y}_5 \end{split}\ ] ] with @xmath67 . clearly , an isolated node in a graph corresponds to an operator that is constant in time [ for instance , @xmath68 when @xmath69 , @xmath70 , as in fig .
[ doublechain ] ( light green ) ] .
this result is key to our investigation .
if we let the interaction terms be alternatively non - zero for a time window @xmath71 $ ] such that @xmath72 , the action of @xmath63 in eq .
( [ modelloxy ] ) will correspond to _ kicks of information _ between pairs of connected nodes of the graphs discussed above . in fig .
[ doublechainjumps ] we show an example for @xmath73 being alternately non - zero for five time intervals such that @xmath74 for @xmath75 [ @xmath76 for @xmath77 , with @xmath78 [ @xmath79 whenever @xmath80 [ @xmath81 . the negligibility of the mutual overlaps of the coupling functions ensures that the information kicks occur according to the ordered scheme in fig . [ doublechainjumps ] .
a similar transport of information from the leftmost to the rightmost part of the graph is achieved by kicking information using a pattern given by fixing @xmath82 ( @xmath83 ) and alternately changing the amplitude of @xmath84 ( @xmath85 ) and @xmath86 in a way so as to satisfy the conditions stated above . after a sequence of @xmath0 ( for time - dependent @xmath84 and @xmath85 ) or @xmath87 kicks ( for time - dependent @xmath86 and either @xmath84 or @xmath85 ) , one obtains @xmath88 where @xmath89 is the total time of the evolution . by initially preparing the @xmath90 spin of the chain ( except the first ) in an eigenstate of @xmath91 ,
there is a complete end - to - end transport of information across the chain .
this evolution interaction can thus be exploited for a perfect state transfer exactly as it happens for the hamiltonian model of ref .
@xcite . quite interestingly ,
any other state - transfer protocol relying on the same information - flux structure , as those in refs .
@xcite , can be mapped into the very same temporal pattern of information kicking illustrated here .
this is the main result of our investigation , proving that an alternative to the standard architecture of properly arranged interaction - strength patterns for quantum state transfer is actually possible .
the key is the introduction of an alternated series of non - overlapping temporal kicks of information .
the equivalence of the two scenarios has here been proven in terms of an isomorphism of the respective information fluxes and is , as such , completely general .
depending on the details of a specific practical realization of the intra - chain couplings , temporal control of the form highlighted here can be more convenient than pre - fabrication of a specific pattern of static coupling strengths .
this is the case , for instance , for quasi - unidimensional optical lattices loaded with neutral atoms , where inter - site couplings are achievable in a time - controlled way via external optical potentials or cold atomic collisions . in perspective
, this could also be a viable option in solid - state structures ( such as arrays of josephson junctions ) , where fabrication of exactly the pattern required for ideal state transfer would be demanding ( if not prohibitive ) already at moderate chain lengths . achieving control via proper voltage / magnetic pulses inducing time - controlled information kicks would be a possibility to exploit instead .
in order to go beyond the idealization of the scheme assessed in the previous section and make our analysis closer to more realistic situations , the ideal conditions on the pattern of time - dependent couplings invoked before will be relaxed here .
let us first consider @xmath92 with @xmath2 and @xmath4 following the behavior dictated by @xmath93^m,\\ & b(t)=b_{max}[\cos(t+\pi/4)]^m , \end{split } \label{sincos}\ ] ] where @xmath94 are the maximum values of @xmath2 and @xmath4 ( properly chosen in order to satisfy the integral condition ) and @xmath95 is an even number .
pulse shapes of this form ( for moderate values of @xmath95 ) are routinely generated by commercial pulse generators .
a large value of @xmath95 obviously implies small overlap between the two coupling functions .
a period of @xmath96 will correspond to a `` horizontal '' and a `` vertical '' kick in the associated operator graph . for a numerical analysis of the information - flux behavior under the action of this time - dependent hamiltonian
, we divided the total time interval @xmath97 $ ] in short steps . within each of them , both @xmath2 and @xmath4 are taken as constant and numerically equal to the average values they assume in the respective time window . in this way , the recurrence formulas presented in sec . [ informationflux ] allow us to estimate the evolution of @xmath40 or @xmath39 for the duration of the step . clearly , the shorter is the time step , the more accurate is the estimate .
the method is then iterated for each step and the total evolution of @xmath40 or @xmath39 is finally reconstructed . as the number of operators involved in each decomposition is equal to @xmath98 , a light computational power is required . in order to evaluate the performance of our scheme , both the coefficient @xmath99 , attached to @xmath100 in the decomposition of @xmath101 , and @xmath102 , attached to @xmath103 in the decomposition of @xmath104 , should be studied .
however , one can equally well characterize the whole state transfer process simply by considering @xmath105 only .
the maximum value of @xmath105 for two values of @xmath95 is studied against @xmath0 in figs .
[ plotpowers ] * ( a ) * and * ( b)*. * ( a ) * 0.5 cm it is matter of performing an explicit calculation to see that the maximum of @xmath102 follows similar behaviors .
we then quantify the mean transmission fidelity @xmath106 by averaging the state fidelity @xmath107 between a pure input state @xmath108 to transfer and the state @xmath109 of spin @xmath0 at time @xmath28 . by assuming a uniform distribution of @xmath110 s @xcite
, we find @xmath111.\ ] ] for instance , if we take @xmath112 , we obtain for any length of the chain @xmath113 a value of @xmath114 . as @xmath115 , the average fidelity converges to @xmath57 , resulting in perfect state transfer . the ability to temporally control both @xmath84 and @xmath86 can however be experimentally demanding . in particular
, while the arrangement of external potentials with a desired time - dependence can be relatively straightforward , this might not be the case for @xmath116 s , which are in general determined by coupling mechanisms internal to the chain itself and hardly controllable with the degree of accuracy required by our scheme .
we thus make our requests lighter by taking @xmath84 as constant and allowing a time dependence only for the external magnetic field .
obviously , it is formally the same to take a constant @xmath86 with time - varying @xmath2 .
however , pragmatically our choice seems to us to be the least demanding .
we thus adapt the strategy assessed before to the case of @xmath117 .
our quantitative study shows that a slight modification to the sinusoidal trend described before is sufficient to retain the efficiency of the protocol .
we take well - spaced square - shaped pulses for @xmath4 as shown in fig .
[ plotbandjbehavior ] .
the parameter @xmath118 quantifies the width of magnetic field pulses .
it can be roughly seen as half of the overall period divided by the time in which @xmath86 is close to its maximum value .
clearly , a large value of @xmath118 corresponds to a more ideal case , as the overlap between the two functions is smaller , hence affecting the maximum value that @xmath105 can achieve and , in turn , the transfer fidelity .
the dependence of @xmath119 on @xmath118 is shown in fig . [ 5qubitsdelta5to20 ] for @xmath38 .
0.25 cm the fidelity @xmath120 is already larger than @xmath121 for @xmath122 . of course , the effect of the overlap between the two functions depends on @xmath0 : the longer the chain , the lower the transmission fidelity .
the dependence of the fidelity on the length of the chain is plotted in fig .
[ 5to15qubitsdelta5to20 ] . 0.25 cm we note that , for @xmath0 up to @xmath123 , a value of @xmath124 is large enough to obtain the maximum of @xmath125 @xmath126 ( corresponding to @xmath127 ) . to give a complete overview of the protocol
, one can finally analyze , for a fixed @xmath118 , the behavior of the maximum of @xmath125 against @xmath0 .
the case of @xmath128 has been considered , and the results are shown in fig .
[ 5to25qubitsdelta20 ] . in these conditions ,
the corresponding transmission fidelity for a chain with @xmath129 spins is @xmath130 .
we have also studied the case in which the non - zero interaction terms are @xmath84 and @xmath85 , with @xmath131 .
the transmission fidelity is larger than the one obtained with non - zero @xmath84 ( @xmath85 ) and @xmath86 , as the number of kicks required is smaller and the effect of the overlap is thus reduced .
however , we presented the results for non - zero @xmath84 ( @xmath85 ) and @xmath86 as we consider the choice of a time - dependent magnetic field the least demanding from an experimental point of view .
in this section we discuss another interesting aspect of the proposed model , namely the possibility to generate genuine multi - partite entanglement .
bi - partite as well as multi - partite entanglement creation in spin chains has been recently studied from different points of view .
for instance , one can generate entanglement by means of a proper pattern of coupling parameters @xcite , by driving the evolution through a resonant interaction @xcite , by performing a global quench @xcite , or by leaving two empty sites in a uniformly filled chain @xcite . by assuming the ability to initialize the state of the chain before the application of the pulse sequence in the protocol proposed in this article ( we necessarily have more demanding requirements for this particular task )
, greenberger - horne - zeilinger ( ghz ) states can be generated @xcite .
such a resource is well - known to be useful for multi - agent protocols for distributed qip like quantum secret sharing , remote implementation of unknown operations and quantum average estimation @xcite .
let us consider again the hamiltonian @xmath10 in eq .
( [ modelloxy ] ) under the ideal conditions mentioned above .
it is interesting to notice that , after the proper kicking of information , the evolution of _ all _ the spin operators associated with the elements of the chain have exactly the same form as those obtained by using the model proposed in refs .
@xcite .
we now want to evaluate the dynamics of a particular pure state of the chain under the action of this hamiltonian . we will need to abandon the assumption that at most a single spin excitation populates the system .
it is straightforward to check that , whenever spin @xmath22 is in an eigenstate of @xmath132 , the corresponding symmetric spin @xmath133 will end up in the same state .
whenever two or more spins are initially in eigenstates of @xmath134 or @xmath135 , the corresponding symmetric spins will end up in an entangled state .
let us consider , for definiteness , an even number of spins .
we call @xmath136 the indices corresponding to spins initially prepared in eigenstates @xmath137 of @xmath6 and @xmath138 those labeling spins initially prepared in eigenstates @xmath139 of @xmath134 ( the analysis can be straightforwardly generalized to the case of odd number of spins and eigenstates of @xmath135 )
. the initial state of the chain is thus @xmath140 it is matter of a straightforward calculation to see that , in this case , the final state corresponds to a mirror - inversion operation on the state @xmath141,\ ] ] with @xmath142 depending on the actual form of @xmath143 .
clearly , @xmath144 is the tensor product of a separable state for spins @xmath136 and an @xmath95-particle ghz - like state for spins @xmath138 .
by means of the information flux approach , we have presented a new strategy for obtaining perfect state transfer in a finite , open chain of spins .
we have shown that the evolution of spin operators under the action of a site - dependent hamiltonian that is already known to allow perfect state transfer can be mimicked by the use of a homogeneous time - dependent hamiltonian .
the engineering of the coupling strengths , necessary in the first scenario , is thus converted to a temporal arrangement of the interaction terms .
we have discussed a particularly interesting case where such a time dependence can be reduced to the requirement of just a non - constant magnetic field affecting the spins of the chain .
interestingly , this very same model can also be used in order to create a ghz - like entangled state shared by a sub - set of spins of our choice , leaving all the other spins in a separable state .
we believe that our proposal to achieve perfect state transfer and quantum correlation sharing in many - body registers of interacting spins widens the possibilities for efficient and non - demanding short / medium - haul quantum communication .
m. christandl , n. datta , a. ekert , and a. j. landahl , , 187902 ( 2004 ) ; c. albanese , m. christandl , n. datta , and a. ekert , , 230502 ( 2004 ) ; g. m. nikolopoulos , d. petrosyan , and p. lambropoulos , europhys .
lett . * 65 * , 297 ( 2004 ) ; g. m. nikolopoulos , d. petrosyan , and p. lambropoulos , j. phys . :
matter * 16 * , 4991 ( 2004 ) ; m. christandl , n. datta , t. c. dorlas , a. ekert , a. kay , and a. j. landahl , , 032312 ( 2005 ) .
m. hillery , v. buek , and a. berthiaume , , 1829 ( 1999 ) ; c. schmid , p. trojek , m. bourennane , c. kurtsiefer , m. zukowski , and h. weinfurter , , 230505 ( 2005 ) ; a. m. wang , , 062323 ( 2007 ) ; c. di franco , m. paternostro , and m. s. kim , , 052316 ( 2007 ) . | we propose a strategy for perfect state transfer in spin chains based on the use of an unmodulated coupling hamiltonian whose coefficients are explicitly time dependent .
we show that , if specific and non - demanding conditions are satisfied by the temporal behavior of the coupling strengths , our model allows perfect state transfer .
the paradigma put forward by our proposal holds the promises to set an alternative standard to the use of clever encoding and coupling - strength engineering for perfect state transfer .
multiple - spin systems , in particular spin chains , have recently been the object of extensive studies . from the quantum information processing ( qip )
viewpoint , such systems embody valuable media for quantum protocols .
in fact , it has been found that specific forms of built - in and permanent intra - register couplings , such as those typical of spin - chain models , could be used for the purposes of quantum computation @xcite and communication in quantum networks .
this second possibility is particularly interesting .
in fact , while photons are ideal candidates for the long - haul transmission of information among different local nodes @xcite , their use in hybrid architectures for quantum networks ( where the nodes are embodied by matter - like systems ) requires the use of quantum interfaces . despite the impressive experimental successes reported in this area , the realization of an interface is usually accompanied by technical problems and errors in transmission . on the other hand , for short - distance quantum communication ,
an alternative arrangement where information carriers and processors are embodied by physical systems having the same nature could be more advantageous . in the seminal work in ref .
@xcite , the idea of using spin chains as _
quantum wires _ has been proposed as a way to avoid the interfacing problems mentioned above .
the original idea has then been extended along various directions ( for more details , see ref .
@xcite ) . in particular , refs .
@xcite showed that , by properly engineering the strength of the couplings in the chain , perfect state transfer could be achieved . a proposal to bypass the initialization of the medium by means of local operations and measurements on the extremal spins of the chain
has been recently put forward @xcite .
a similar result can be obtained by properly encoding the state to transmit in two spins at one end of the chain @xcite . here , an alternative strategy to the one proposed in ref .
@xcite is presented . instead of the pre - arrangement of the interaction strengths across the chain
, we consider a uniform distribution of time - dependent couplings .
we show that perfect transfer of information is achieved in our time - dependent architecture as well and that a formal mapping of the dynamics achieved in the two models is in order , thus letting our proposal emerge as an alternative paradigma for perfect quantum state transfer .
we also study a scenario where the coupling strengths of the hamiltonian are constant , while an external time - dependent magnetic field is used in order to provide the necessary temporal modulation required for information transmission .
we believe that the introduction of a time - dependent term in the intra - chain couplings allows to bypass the intrinsic rigidity of protocols based on unmodulated and pre - engineered spin media .
in fact , it is usually the case that spin - chain models are arranged so as to implement a specific communication or computational task and can not be recycled " and used for a different one .
moreover , should an experimentally - prepared pattern of coupling strengths be found not accurate enough , the whole medium should be discarded ( unless a non - ideal performance of the quantum protocol could be tolerated ) . differently , by allowing the presence of time - dependent hamiltonian terms , a more dynamical adjustment " of the medium performances would be possible by implementing a simple feedback loop : state transmission can be tested using a given functional form of the time - dependent terms of the hamiltonian and , according to the results , this can be tuned so as to converge toward better performances . a time - dependent scheme for quantum state transfer has been discussed by lyakhov and bruder in ref .
@xcite .
building up on general result described in ref .
@xcite , they only let the first and last spin - pair of a chain to experience time - dependent couplings .
our scheme , on the other hand , is different as we either consider a full set of time - dependent interaction strengths or a fully homogeneous pattern of constant couplings with a time - dependent external magnetic field .
the remainder of this article is organized as follows . in sec . [ system ]
we describe the model used in our proposal . in sec .
[ informationflux ] we shortly review the main features of the information - flux approach , which is the main tool used in our investigation , and adapt it to our time - dependent analysis .
sec .
[ results ] illustrates the way we obtain perfect state transfer when the coupling strengths depend on the interaction time via sharply rising / lowering pulses , while sec .
[ realistic ] is dedicated to the investigation of a much more realistic setting that considers finite rising times and pulse duration . in sec .
[ multipartite ] , the proposed model is exploited in order to create a genuine multi - partite entangled state .
finally , sec .
[ remarks ] summarizes our results . |
the development of quantum - based devices is becoming increasingly necessary for future technological advances and , as a result , enormous effort is being made to find suitable experimental platforms , one of which is cold , neutral atoms .
strong coupling can be viewed as an essential requirement to devise quantum networks based on neutral atoms @xcite and , until recently , free - space cavities were the most promising method for enabling an atomic system to enter this regime @xcite .
very recently , kato and aoki @xcite demonstrated strong coupling for a single trapped atom and an all - fibred optical cavity based on an optical nanofibre ( onf ) spliced between two fibre bragg grating mirrors , thereby illustrating the potential offered by onfs for future quantum technological advances .
+ another distinct advantage of optical nanofibres is that they facilitate the study of nonlinear phenomena in an atomic medium without requiring a tightly focussed laser beam , as in free - space optics .
if light is to be tightly focussed , the interaction distance is essentially limited by the rayleigh length , @xmath0 , given by @xmath1 , where @xmath2 is the beam radius at the narrowest point and @xmath3 is the laser wavelength . if we consider a 780 nm beam , corresponding to the d@xmath4 line in @xmath5rb , the interaction length is limited to about 4 @xmath6 m if the beam is focussed to a 1 @xmath6 m radius .
onfs allow us to overcome this limitation @xcite , due to the strong evanescent field at their waist ( i.e. the thinnest region of the nanofibre ) which can contain most of the power injected into them @xcite
. this evanescent field may interact with atoms in the vicinity of the fibre @xcite and spontaneously emitted light from excited atoms can preferentially couple into the nanofibre - guided modes @xcite .
such properties have led , in recent years , to an explosion of interest in integrating onfs into cold atomic systems in order to fully exploit the advantages that they offer .
+ in this review paper , we present a brief introduction to mode propagation in onfs and provide the reader with pertinent references in relation to the fabrication of such fibres , whether for fundamental mode or higher order mode ( hom ) propagation .
next , we summarise some of the earlier onf work with atomic vapours and cold atoms , leading up to the realisation of onf - based cold atom traps and the achievement of strong coupling between a single atom and an onf - based cavity
. very recent work on nonlinear optics phenomena , such as electromagnetically induced transparency ( eit ) and slow light using onfs with cold atoms , is then discussed .
finally , we aim to provide some insight into possible future directions of the field .
note that , depending on the wavelength being used , hom propagation may require the fibre waist to be larger than a micron , in which case it should be referred to as an optical microfibre ( omf ) . for clarity , we will assume that all fibres discussed in this review are submicron in diameter , i.e. onfs , unless explicitly stated otherwise .
in the following , we derive expressions that describe the evanescent field and guided modes in a standard optical fibre .
an extensive development of the mathematics presented here can be found , for example , in @xcite .
an optical fibre can be approximated to a perfectly cylindrical , step - index fibre in which the refractive index of the core , @xmath7 , is greater than the refractive index of the cladding , @xmath8 .
mathematically , the fibre is represented by the following index profile @xmath9 where @xmath10 denotes the core radius , @xmath11 is the radial position from the fibre centre and @xmath12 is the cladding radius as defined from the centre of the fibre .
if we assume the fibre to be isotropic and source - free , maxwell s equations for light propagating within this system can be reduced to the following set of helmoltz equations @xmath13 with @xmath14 representing the transverse laplace operator in cylindrical coordinates , @xmath15 is the wave number , @xmath16 is the modal propagation constant and @xmath17 represents the @xmath18component of either the electric or magnetic field of the input light , respectively .
this equation is separable into its individual components @xcite and solutions may take the form @xmath19 with @xmath20 being the radial part and the exponential term representing the azimuthal part .
here @xmath21 , and @xmath22 , @xmath11 and @xmath23 denote the longitudinal , radial and azimuthal components of the field , respectively . inserting this into eq .
gives @xmath24r(r)=0,\ ] ] which is a bessel differential equation , the solutions of which are bessel functions of order @xmath25 .
two cases can be distinguished depending on the sign of @xmath26 and their general solutions can be calculated to yield @xcite @xmath27 where @xmath28 , @xmath29 , the @xmath30 represent constants , @xmath31 and @xmath32 are bessel functions of order @xmath25 of the first and second kind , respectively , and @xmath33 and @xmath34 are modified bessel functions of order @xmath25 of the first and second kind , respectively . + for confined propagation of light at the fibre s waist , the effective refractive index , @xmath35 , should be larger than the cladding s refractive index , which leads to the case @xmath36 . for @xmath37 ,
@xmath33 diverges and leads to nonphysical solutions .
therefore , @xmath38 and we are left with the term @xmath39 in the radial part of the field ( see eq . ) describing the exponential decay of the evanescent field at the waist of the fibre . + solving eq . for the fields in the core
yields four eigenvalue equations , corresponding to the four possible guided modes inside a cylindrical optical fibre , namely the he , eh , te and tm modes .
we list their analytical expressions here , without further development @xcite : + @xmath40 modes : @xmath41 @xmath42 modes : @xmath43 @xmath44 modes : @xmath45 @xmath46 modes : @xmath47 @xmath48^{1/2}. \end{split}\ ] ] in these expressions , @xmath49 denotes @xmath50 .
+ fabrication of an onf usually consists of heating a commercially available optical fibre , stripped of its acrylic coating , to a temperature close to the glass melting point ( for example , 1550@xmath51c for silica ) . under these conditions
, the fibre is in a plastic regime and can be elongated without breaking .
as the fibre is pulled , three regions can be distinguished , as illustrated in fig.[fig : mnf ] .
these are the pigtail , the taper and the waist , which , if small enough , leads to the generation of a significant evanescent field when light is launched into the fibre @xcite .
several fabrication methods have been reported @xcite and the one used generally depends on one s needs , aims and resources . for cold atom experiments , the so - called _ flame - brushing technique _
@xcite is usually preferred since it offers fine control over the taper shape and the waist size , two important parameters for modal selection .
routinely , fibres of the order of 350 nm can be fabricated .
similar advantages were recently reported using a microheater - based fibre - pulling rig @xcite .
+ when light is launched into a fibre , it can couple into several of the available guided modes @xcite .
light propagates differently in each of these modes , a property characterised by the modal propagation constant , @xmath16 .
this difference in propagation gives rise to modal interference and may result in unwanted beam profiles at the output of the fibre .
however , one can limit the number of propagating modes by reducing the dimensionless v - number , obtained from solving maxwell s equations for an optical fibre @xcite .
the v - number is determined from @xmath52 where @xmath53 is the fibre waist diameter , @xmath3 is the wavelength of light , and @xmath7 , @xmath54 are the refractive indices of the fibre core and cladding , respectively . as shown in fig .
[ fig : vnum](a ) , the modal limitation is easily understood when one plots the effective refractive index , @xmath55 , versus the v - number . for a given wavelength , the v - number and , consequently , the number of allowed modes , can be reduced by decreasing the fibre diameter .
this is achieved during the onf fabrication as this process consists of tapering the radius of a normal optical fibre until it reaches a few hundreds of nanometers in diameter .
+ , tm@xmath56 , he@xmath57 and he@xmath58 collectively referred to as the lp@xmath59 group .
( b ) simulated and ( c ) experimental images of the output of an onf when light is guided in the lp@xmath59 group of modes . reproduced from @xcite.,scaledwidth=49.0% ] changing the structure of an optical fibre influences its birefringence properties and may result in losses if not done carefully .
these losses can be minimised by ensuring that the adiabaticity criterion is met during fabrication @xcite .
as stated above , light injected into an onf can couple into different fibre modes , each yielding a different propagation constant .
for example , fig .
[ fig : vnum](b ) shows the simulated and fig .
[ fig : vnum](c ) the experimental output profiles obtained when one limits the fibre - guided modes to the lp@xmath59 group by reducing the fibre diameter , thereby reducing the v - number to a value close to 2.771 , i.e. the he@xmath60 cut - off .
the double - lobe pattern is the result of interference between the different submodes of this group . at the start of the taper region ,
these modes are not in phase and consequently experience beating , which results in losses . the beat length , @xmath61 , has to be kept as short as possible compared to the taper length , @xmath62 , where @xmath63 is the core radius and @xmath64 is the taper angle ( see fig.[fig : adiab ] ) . in the ideal case
@xmath65 ; this is achieved when the taper angle is as small as possible and , in this case , light at the taper couples entirely to a single mode .
+ , the core radius , @xmath63 , and taper length , @xmath66 . reproduced from @xcite.,scaledwidth=40.0% ] our group has shown that similar reasoning can be applied to derive the adiabatic criterion for efficient hom propagation @xcite .
it is possible to limit propagation of the he@xmath60 mode by careful control of the pulling length and using commercially available 80 @xmath6 m diameter fibre @xcite , which has a different cladding - to - core ratio , instead of the usual 125 @xmath6 m . as plotted in fig.[fig:125v80 ] , reducing the cladding- to - core ratio shifts the adiabatic condition on the local core taper angle .
this is a result of the fact that fewer modes can propagate within the cladding , thereby confining them in the core and reducing the beat length .
the choice of this ratio is , therefore , crucial for hom optical nanofibre experiments .
the technique was later refined by ravets _
et al . _ , but for 50 @xmath6 m diameter fibres @xcite .
they matched the hom adiabatic criterion even better and obtained hom propagation efficiency for the lp@xmath59 group up to 97.8@xmath67 .
recently , a heat - and - pull fibre rig for ultra - high transmission of both fundamental mode and hom fibres was reported by hoffman _
_ @xcite . using the same fibre - pulling rig ,
the group reported observation of the spatial evolution of homs in the lp@xmath59 group , over the whole fibre length , via rayleigh scattering imaging @xcite .
additionally , they were able to identify and control which submodes of the lp@xmath59 group were propagating within the omf .
onfs have attracted considerable interest for particle trapping and manipulation and , more recently , for the development of quantum technologies @xcite .
such fibres are particularly praised for their potential role as efficient and reliable photonic communication channels between distant quantum systems @xcite .
atoms localised close to the onf s surface can emit light that couples to its guided modes @xcite .
this , in turn , enhances the photon emission rate @xcite , and can be engineered to generate a lateral translation - invariant force on the atoms @xcite .
the coupled photons in these modes are correlated in either a bunching or an anti - bunching fashion , as demonstrated experimentally by nayak _
_ @xcite . due to the enhanced evanescent field produced at the waist ,
the guided light can interact with surrounding atoms via absorption @xcite .
the first experimental demonstration of this property was made in 2007 by sagu _
@xcite for spectroscopy purposes .
they created a cloud of cold caesium atoms in a magneto - optical trap ( mot ) which was superimposed with an onf .
a probe beam was then launched into the onf and its absorption measured on an avalanche photodiode ( apd ) . emitted light
subsequently couples back into the fibre , the propagation direction of which is dictated by the polarisation of the initially injected light @xcite or the atoms arrangement in the immediate vicinity of the fibre @xcite .
+ extensive work on in - fibre spectroscopy of hot vapours @xcite was also carried out in parallel . however , introduction of an onf in an atomic vapour often results in the atoms sticking to the onf s surface , thereby reducing its efficiency as a probing tool .
several ways of dealing with this problem have been explored by the umbc group .
for example , they demonstrated the desorption of rubidium atoms by launching low power , non - resonant light in the onf @xcite , or by placing it on a flange - mounted heater @xcite .
they also proposed to switch to a vapour of metastable xenon atoms@xcite , since it is a noble gas and consequently does not interact as much with the fibre .
they showed saturated absorption at the nw level in this system .
the same group recently reported the observation of a nonlinear optical effect , eit , in a rubidium vapour @xcite .
+ due to the ultra - low pressures needed in cold atom experiments in contrast to vapour experiments , the problem of atom adsorption on the nanofibre does not occur to a detrimental level .
this makes the onf an excellent tool to probe and characterise atom clouds created in a mot @xcite .
several experiments carried out by russell _
@xcite used an onf embedded in such a system to measure the temperature of a cloud of @xmath68rb atoms using different techniques . in @xcite , temperature variations of the order of @xmath6k created by changes in the cooling - lasers alignment , intensities or detuning , were detected by comparing the fluorescence signals coupling into the onf throughout the different stages of a release - recapture process .
a dark - mot @xcite of the same atomic species was also characterised ( via the atom density and loading time ) using similar methods .
more recently , grover _ et al . _
@xcite performed temperature measurements in a mot , but this time on @xmath5rb atoms .
the characterisation method , although relying on fluorescence collected from an onf , used the intensity - intensity correlation function to extract the velocity of atoms surrounding the fibre and , hence , their temperature .
it has also been proposed to use an onf as position sensors of trapped atoms in an optical lattice @xcite .
+ the interaction of atoms with an evanescent light field has long been known @xcite . in the 1990 s ,
it was proposed @xcite , and later demonstrated @xcite , that atoms could be guided inside a hollow - core fibre using the optical dipole force resulting from the interaction with the guided light .
in these experiments , a laser beam , red - detuned from the average transition frequency of the d@xmath4 spectroscopic lines for @xmath68rb and @xmath5rb , was launched inside the hollow core fibre .
the red - detuned light generated an attractive dipole force for atoms and they were guided inside the fibre .
later , ito _ _ et al.__@xcite designed a similar experiment using blue - detuned light to repel the atoms from the fibre s inner surface and guide them along the centre of the core .
+ loading atoms inside a fibre is nontrivial and , to overcome some of the obvious limitations , balykin _ et al . _
@xcite proposed trapping @xmath69cs atoms outside the onf by propagating a red - detuned laser beam in the fundamental guided mode ( i.e. he@xmath59 ) .
this beam creates a dipole force gradient within the evanescent field which attracts the surrounding atoms towards the fibre .
+ to understand the forces at play , consider a single atom placed in an electric field , * e*(*r * ) .
this electric field induces a dipole moment @xmath70 on the atom such that @xcite @xmath71 where @xmath72 is the fundamental electric charge , @xmath73 is the electric permittivity in vacuum and @xmath74 is the scalar polarisability .
the interaction between this dipole and the electric field yields a potential energy , u(*r * ) , given by @xmath75 where the factor of @xmath76 arises due to the dipole moment being induced rather than permanent .
assume the electric field is a plane wave , propagating along the @xmath22-direction and polarised in the @xmath77-direction with angular frequency , @xmath78 , and wave number , @xmath79 , giving @xmath80 .
then , the force acting on the atom in the @xmath22-direction is expressed as @xmath81 .
\end{aligned}\ ] ] the expression in eq . can be written with the dipole moment expressed in terms of its components in the bloch sphere and time averaged to get the final expression ( see for example p.198 of @xcite for a detailed development ) @xmath82 where @xmath83 is the natural width of the excited state , @xmath84 is the rabi frequency and @xmath85 is the detuning between the laser frequency , @xmath86 , and the resonant frequency , @xmath87 , of the considered transition .
the first term of eq . describes the dipole force , whereas the second term is the scattering force playing a central role in the initial laser cooling process .
it is important to note that the dipole force depends on the gradient of the field intensity and the detuning . for negative detuning ( i.e. a red - detuned beam ) ,
an atom placed in the vicinity of the onf and interacting with its evanescent field will experience a force pointing in the same direction as the gradient of intensity , i.e. towards the onf s centre . for fibres with a waist diameter
roughly half the light s wavelength , this force can be balanced by the centrifugal force acting on atoms revolving around the onf and , hence , can be used to trap them .
however , the number of atoms that can be trapped using this scheme is restricted to those possessing an initial angular momentum along the fibre axis within an appropriate range .
le kien _ et al . _
@xcite proposed an improved version of the onf trap , relying on a two - colour scheme ; a red - detuned beam to attract the atoms towards the onf and a blue - detuned beam to repel and prevent them from sticking to its surface . in this case , the van der waals potential created by the fibre @xcite acts together with the red - detuned beam to attract the atoms , while the blue - detuned beam prevents them from hitting its surface .
both proposed traps are shown in fig.[fig : firstonf ] .
+ the experimental achievement of the two - colour trapping scheme was made in 2010 @xcite .
laser - cooled @xmath69cs atoms were trapped around the waist of an onf using the evanescent field resulting from the combination of a blue - detuned beam and two counter - propagating red - detuned beams with respect to the @xmath69cs d@xmath4 line .
the counter - propagating beams were used to generate periodically - spaced potential wells .
the red- and blue - detuned beams were both linearly polarised , but orthogonal to each other .
this gives better azimuthally localised trapping sites , as opposed to a parallel configuration .
in addition , the trapping scheme was done in a collisional - blockade regime @xcite , which allows only one atom at most per trapping site at a time . a weak probe beam
, whose frequency was scanned around the @xmath69cs d@xmath4 line , was launched into the fibre to generate fluorescence from the trapped atoms for imaging purposes ( fig.[fig : arnotrap ] ) and to provide an estimation of their number by saturated absorption .
the resulting profile showed an asymmetry in the red - detuned region , which the authors attributed to a state - dependent light - shift created by the trapping beams .
subsequently , the group tuned their probe beam to the stark - shifted resonant frequency and measured approximately 2000 atoms trapped around the onf .
atoms trapped in this configuration showed ground state coherence times on the order of milliseconds @xcite .
this trapping scheme was extended to create a conveyor belt @xcite and led to the transportation of cold caesium atoms over a few mm along the onf .
_ later adapted the two - colour scheme and trapped @xmath5rb atoms@xcite .
in contrast to the experiment in @xcite , they arranged the trapping beams to have linear polarisations close to the parallel configuration .
they justified this choice by the fact that the cold atom cloud had a better optical depth ( od ) in this configuration , but were not able to explain the physical reason accounting for it .
an even more pronounced asymmetry in the absorption profile was observed and used to estimate the number of trapped atoms .
only 300 atoms were trapped in this experiment , that is 7 times less than reported in @xcite for @xmath69cs atoms .
this was attributed to the difference in atomic structure between @xmath69cs and @xmath5rb , the latter being subject to larger vector light shifts .
it was also proposed @xcite to use the fictitious magnetic field generated by a nanofiber - guided light field in combination with an external magnetic bias field to create trapping potentials along the waist .
similar potential depths as in @xcite were reached , though in this work the power of the trapping field was an order of magnitude lower .
in addition , this method allowed for better control over the trapped atoms positions .
+ the above - mentioned light shifts can be problematic for the transfer of coherent information over long distances @xcite .
one solution is to use the so - called magic wavelengths @xcite .
when the trapping laser is operating at these specific wavelengths , hyperfine levels of interest undergo the same ac stark shift , thus leaving the transition frequency unchanged .
several groups have been working on perfecting these state - insensitive traps , showing improved trap lifetimes and lower trapping powers in the two - colour scheme for both rubidium and caesium atoms @xcite .
although extremely efficient , state - insensitive traps are restrictive as one is required to work with a specific trapping wavelength .
+ cs atoms in the evanescent field of an onf . reproduced from @xcite.,scaledwidth=49.0% ]
homs offer a more flexible way to deal with the coherence issue .
these modes have an extended and highly directional evanescent field at the waist compared to the fundamental mode .
this is illustrated in fig.[fig : ravihom ] , where the intensity profiles of the fundamental mode , he@xmath59 , and the first set of homs , namely , te@xmath56 , tm@xmath56 and he@xmath60 , are presented , along with their respective intensity distribution across the fibre waist .
these properties could improve the coupling with laser - cooled atoms .
masalov and minogin predicted that , if atoms were placed close to a fibre supporting homs , they would couple 5 to 10 times more light in these modes than in the fundamental mode @xcite .
our group demonstrated experimentally that , if one considers the onf s lp@xmath59 family of guided modes , the coupled fluorescence from laser - cooled @xmath5rb was six times more than if one only considers lp@xmath56 @xcite .
the same experiment also showed that by injecting a laguerre - gaussian beam of index 1 ( lg@xmath56 ) , i.e. a doughnut beam , to excite the lp@xmath59 group of fibre modes , more light is absorbed by the atoms surrounding the fibre , in accordance with the theoretical predictions in @xcite .
we assumed this was due to the further extension of the evanescent field into the atom cloud .
+ red - detuned beams.(a ) both beams have circular polarisation with the same handedness yielding periodic ring - like trapping potentials .
( b ) both beams have linear polarisation yielding periodic quadrupole - like potentials .
( c ) both beams have circular polarisation but with opposite handedness yielding a double - helix trapping potential . reproduced from @xcite.,scaledwidth=49.0% ]
interference patterns resulting from the combination of two co - propagating blue - detuned lp@xmath59 sub - modes , or one of the submodes and the fundamental mode , should yield deeper trapping potentials at lower powers @xcite than that obtained relying solely on fundamental guided modes . moreover , these modes offer the possibility to design trapping geometries that allow for better control over the atom trap positions .
for example , the distance between each atom and their position relative to the fibre can be more accurately controlled by modifying the trapping beams polarisations @xcite .
this is depicted in fig.[fig : helix ] , which shows the geometries proposed in @xcite .
the two - colour scheme is used with counter - propagating red - detuned beams in a hom and a blue - detuned beam in the fundamental mode .
one can choose from a periodically spaced ring , a periodic quadrupole , or a double - helix trap geometry by choosing the red - detuned trapping beams polarisations as circular - circular with same handedness , linear - linear , or circular - circular with opposite handedness , respectively .
also proposed to use a red - detuned beam in a superposition of both the te@xmath56 and he@xmath59 modes , in order to create a 1d optical lattice around the waist of an 820 nm diameter fiber @xcite . to experimentally create these proposed traps , one has to be able to isolate each submode of the lp@xmath59 group , namely te@xmath56 , tm@xmath56 , and he@xmath60 .
this can be done either by injecting a free - space circularly polarised lg@xmath56 beam into a few - mode fibre that supports the lp@xmath59 group , and filter unwanted modes at the output via a half - wave plate @xcite , or by mixing two lg@xmath56 beams with opposite topological charge ( i.e. lg@xmath56 and lg@xmath88 ) and opposite circular polarisation @xcite .
+ among the numerous features offered by onfs , one of the most interesting is their potential as tools for exploring nonlinear optics phenomena in atomic media .
for example , it was proposed by russell _ et al . _ to use onfs to provide intense light fields as needed for two - photon absorption by laser - cooled atoms @xcite .
this can be done using extremely low powers compared to similar experiments performed in free - space .
our group demonstrated this effect @xcite by showing autler - townes splitting in laser - cooled @xmath5rb atoms via an onf - mediated two - photon process .
the autler - townes effect appeared for powers as low as 20 nw and frequency up - conversion was , in addition , observed with powers as low as 200 pw .
+ the development of an efficient quantum memory is fundamental to the creation of any quantum - based computing system . in a quantum network ,
a quantum memory acts as a node " connecting quantum channels in which information is written , stored and read out @xcite .
quantum information is easily lost over long distances due to the imperfect nature of information - carrying media such as optical fibres .
this problem can be resolved by setting quantum repeaters along a transmission channel @xcite .
these repeaters store information twice so that one half can be destroyed " by a local measurement , which in turn allows the other half to be teleported to a distant node .
the role of quantum memory in such a device is therefore obviously important .
light - atom systems are good candidates for the above mentioned purposes and extensive work on the implementation of optical quantum memories , reviewed in @xcite , has been done to date . in particular , it has been proposed to store information , carried from light , in atoms via eit @xcite .
+ the eit process , which was first demonstrated in 1990 with strontium atoms @xcite , relies , in its simplest form , on three atomic levels , namely two ground states and an excited level .
this is the commonly used @xmath89-configuration ( see fig.[fig : gamma_config ] ) , which will be used for the description hereafter .
note that other configurations such as the ladder- or v - configuration have also been presented @xcite . in the @xmath89-configuration
, the two ground states can only couple via the single excited state of the system .
the first ground state , lower in energy than the second one , is coupled to the excited state via a weak probe field .
the second ground state , however , is coupled to the excited state via a strong control field .
typical transmission profiles of the probe beam when the control field is switched off and on are shown in fig.[fig : gamma_config ] .
it appears clearly on these profiles that , for a definite range of frequencies , the optical medium is transparent to the probe field . aside from this transparency window
where absorption would normally be observed , eit also features a drastic reduction in the group velocity of the light beam passing through the material @xcite .
+ in 2000 , lukin and imamo@xmath90lu @xcite proposed to take advantage of this feature to create highly entangled photons .
the main idea was to drive an eit process in @xmath68rb and @xmath5rb so that one isotope would serve as a slowing medium for a weak field , which in turn would be used to generate eit in the second isotope .
this work led to the seminal paper by duan , lukin , cirac and zoller describing the dlcz protocol , at the heart of quantum repeaters @xcite .
as the phenomenon can be observed with a lesser contribution from doppler broadening , eit in cold atomic systems was also studied @xcite .
+ it is also possible to generate eit , and the subsequent slow light feature , in an atomic medium using the evanescent field at the waist of an onf @xcite .
our group recently demonstrated a ladder - type eit process in laser - cooled @xmath5rb atoms @xcite , using powers below 1 @xmath6w in each of the required beams .
we used a weak 780 nm wavelength laser as our probe beam to excite the transition @xmath91 , and a strong 776 nm wavelength laser as our coupling beam to excite the transition @xmath92 .
these beams were injected , counter - propagating , in a single - mode onf , with orthogonal linear polarisations .
in addition to the eit phenomenon , we showed an all - fibred optical switch in this system .
although our probe beam was kept on during the entire duration of the experiment , modulating the coupling beam yielded similar modulation in the transmission profile of the probe beam .
storage of guided light in a cold atomic system undergoing eit was also demonstrated recently @xcite . in @xcite ,
fibre - trapped @xmath69cs atoms , using a scheme similar to that presented in @xcite , were used . as explained earlier
, this scheme creates an optical lattice localised close to the waist of the onf . additionally to the trapping beams ,
co - propagating probe and control beams were also launched in the onf to create the eit process ( fig.[fig : memory ] ) .
this resulted in the probe pulse being stored and retrieved on demand .
however , the recovery efficiency was limited to 3% . in @xcite , a retrieval efficiency as high as 10%
is reported . as opposed to the experiment in @xcite , this experiment used an ensemble of cold atoms instead of trapped atoms in a fibre - based optical lattice .
another difference was that only the probe beam was guided through the fibre while the control beams were in free - space .
+ although onfs offer similar features to those of cavities , albeit in a more scalable manner , there is an interest to combine both in order to improve the light and matter interactions .
some proposals have highlighted the advantages of creating a cavity network , in which each cavity is linked to others via fibres @xcite .
le kien _ et al .
_ have been studying in detail an alternative solution , consisting of onfs with a fibre - bragg - grating ( fbg ) positioned at either end of the waist @xcite .
the system combines both aspects in one and demonstrates strong coupling regimes with relatively low cavity finesse .
such devices could lead to improvements in the generation of eit @xcite , or entangled photons @xcite , and thus pave the way to quantum information generation and storage in an all - fibred network .
these proposals are experimentally achievable as nayak _ et al . _
have reported fabrication of fbg - onfs using a focussed ion beam ( fib ) milling technique @xcite ( see fig.[fig : fbgfiber ] ) .
recently , a single caesium atom was trapped in a similar device @xcite and strong coupling was achieved .
the trap relied on a two - colour scheme analogous to that described in @xcite , but with the red - detuned beams set at a magic wavelength in order to compensate for any light - shifts .
as the atom was probed , it yielded a vacuum rabi splitting , which is a signature of strong coupling between the cavity and the atom .
polarisation dependent fbgs can also be designed @xcite by creating optical nanocavities at the surface of an onf or omf via femtosecond laser ablation .
+ 560 nm and each slit is @xmath93150 nm wide and @xmath93100 nm deep .
two neighboring slits are separated by @xmath93360 nm . reproduced from @xcite.,scaledwidth=48.0% ] our group recently proposed a design consisting of a rectangular cavity at the centre of an omf s waist@xcite .
this slotted tapered optical fibre ( stof ) , depicted in fig.[fig : slotfiber ] , combines advantages from both the onf and hollow - core fibres for atom trapping .
the use of a two - colour trapping scheme in this device should lead to improved interactions between atoms and guided light compared to the case of atoms trapped outside the waist of an onf .
short range interactions , such as van der waals and casimir - polder effects , can easily be compensated by launching a blue - detuned beam in the stof .
the fabrication methods and achieved transmissions were recently reported @xcite .
to date , evidence of 100 nm colloidal particle trapping in such a device exists@xcite . + representing the slot s length , and n@xmath94 and
n@xmath4 the refractive indices of the fibre and the surrounding medium , respectively .
( b ) cross - section of the slot in which w@xmath95 represents the width of the slot . reproduced from @xcite .
( c ) sem image of an actual stof fabricated using the fib milling technique.,scaledwidth=47.0% ]
as presented in this review article , the large panel of features offered by onfs makes them excellent tools for studies on cold atomic systems .
their potential for the creation of an atom - photon hybrid quantum system , with the advantage of it being fibred and thereby easily networked to other systems , is also clearly evident .
a great amount of work is still to be done before achieving all - fibred networks .
for example , a study on rydberg atoms trapped next to onfs @xcite could lead to applications in quantum information @xcite .
one could envision creating rydberg atoms within a series of slots along the waist of an optical nanofibre @xcite to form a 1d quantum network .
alternatively , this system would provide an ideal testbed for exploring the quantum statistics of light in a strongly nonlinear atomic medium @xcite .
+ likewise , to date very limited research has been conducted on hom propagation in optical nano- or microfibres .
fibre - traps resulting from combinations of homs would lead to several distinct geometries and an overall better control over the trapped atoms positions in the vicinity of the fibre @xcite .
moreover , the first set of homs , namely the lp@xmath59 group of modes , can be easily excited by launching an lg@xmath56 beam into the onf .
this beam carries orbital angular momentum ( oam ) @xcite and provides extremely promising applications for atom trapping @xcite and quantum information processing @xcite .
the details are beyond the scope of this review article and have been covered extensively elsewhere @xcite .
+ another promising aspect of the atom - fibre system worth investigating is the cooperativity between atoms trapped along an onf since , due to the tight spatial confinement of the optical modes , the coupling strength per atom is increased when compared to free space beams .
taking advantage of this effect qi _ et al .
_ estimated that a peak squeezing of about 5 db should be attainable for a standard configuration consisting of 2500 atoms trapped 180 nm from an onf with diameter 450 nm @xcite .
+ in conclusion , while the field of onfs , or even omfs , integrated into atomic systems has reached a certain level of maturity in the last couple of years there are many more aspects that still remain to be explored and the full versatility of these ultrathin optical fibres in quantum engineering remains to be discovered in what promises to be an exciting field .
the authors declare no conflicts of interest .
100 url # 1#1urlprefix[2][]#2 cirac i , zoller p , kimble j and mabuchi h 1997 _ phys .
* 78 * 32213224 deasy k , watkins a , morrissey m , schmidt r and nic chormaic s 2010 few atom detection and manipulation using optical nanofibres _ quantum communication and quantum networking _ ( _ lecture notes of the institute for computer sciences , social informatics and telecommunications engineering _
vol 36 ) ( springer berlin heidelberg ) pp 200209 | optical nanofibres are increasingly being used in cold atom experiments due to their versatility and the clear advantages they have when developing all - fibred systems for quantum technologies .
they provide researchers with a method of overcoming the rayleigh range for achieving high intensities in a focussed beam over a relatively long distance , and can act as a noninvasive tool for probing cold atoms . in this review article , we will briefly introduce the theory of mode propagation in an ultrathin optical fibre and highlight some of the more significant theoretical and experimental progresses to date , including the early work on atom probing , manipulation and trapping , the study of atom - dielectric surface interactions , and the more recent observation of nanofibre - mediated nonlinear optics phenomena in atomic media . the functionality of optical nanofibres in relation to the realisation of atom - photon hybrid quantum systems is also becoming more evident as some of the earlier technical challenges are surpassed and , recently , several schemes to implement optical memories have been proposed .
we also discuss some possible directions where this research field may head , in particular in relation to the use of optical nanofibres that can support higher - order modes with an associated orbital angular momentum . |
stellar winds of massive stars leave their imprints on the surrounding medium before they explode . after the explosion , the outer shocks of the supernova probe the circumstellar medium , thereby mapping the late stages of the stars evolution . because of its youth
, the remnant of sn 1987a provides a unique opportunity to investigate these late stages .
the case of sn 1987a is especially interesting , as it exhibits the enigmatic three ring structure visible in optical images @xcite .
this structure probably originates from the interaction of a wind from the slow red supergiant phase with the faster wind from the blue supergiant phase @xcite .
the morphology of this structure might provide clues to a binary - merger scenario for the progenitor @xcite . however , this three ring structure might also result from mass - loss from a fast - rotating star @xcite .
currently , the explosion is sweeping up the inner equatorial ring that was formed by the late stages of the star s evolution . as the inner equatorial ring is about a hundred times denser than its surroundings ( e.g. , * ? ? ?
* ) , the interaction of the explosion with this structure has a direct impact on the evolution and observational properties of the remnant . in 1995 ,
the outer ejecta began to light up this structure . by october 2000
, 12 spots had lit up at visible wavelengths @xcite .
infrared observations indicate that the full interaction with the ring started around mid-2002 @xcite .
a year and a half later , the expansion in x - rays slowed down dramatically @xcite while , at the same time , the soft x - ray light curve turned up @xcite .
based on past _ chandra _ observations , @xcite established that the x - rays are emitted from a flat spatial structure .
a model consisting of three components has been proposed for interpreting the high resolution x - ray spectroscopy from sn 1987a @xcite .
the first component is emitted by material in the dense ring that has been heated by a slow shock ( 500 km s@xmath2 ) .
this component has a low electron temperature ( @xmath4 kev ) .
the second component is emitted by the shock - heated circumstellar medium located above and below the equatorial ring .
these shocks have not been slowed down as much and therefore the plasma is hotter . also , the line emission of this component has to be broad ( thousands of km s@xmath2 ) .
this allowed @xcite to estimate the current contribution of this component to be of the order of @xmath5% of the total emission in the 0.5 - 2.0 kev band , based on high resolution high energy transmission grating ( hetg ) spectra and reflection grating spectrometer ( rgs ) spectra .
further investigation of the x - ray spectra of sn 1987a shows that there is need for a third component with a higher ( @xmath0 3 kev ) temperature , but without broad line emission .
@xcite proposed that this is shocked plasma , reheated by a reflected shock caused by the interaction with the equatorial ring .
alternatively , @xcite proposed that this high temperature component arises from a fast shock , traveling through the uniform equatorial ring material ( in contrast to the low temperature component , which they consider being emitted by shocked denser clumps in the equatorial ring ) .
the _ chandra _ x - ray observatory has observed sn 1987a at a regular pace of about two observations per year , resulting in a denser coverage of the evolution of the remnant than with any other x - ray observatory .
@xcite summarized the observations through 2010 . in the 0.5 - 2.0 kev light curve
, they found a break ( i.e. , the pace with which the flux increases suddenly slowed down ) around 8200 days after the explosion ( @xmath0 y2009 ) .
based on this result , they concluded that the shocks responsible for the soft x - ray emission must have recently encountered less dense material .
however , the may 2012 revision to the acis calibration ( caldb 4.4.10 ) ] introduced a significant change to the time - dependent model for the transmission of the optical blocking filter .
the new model predicts lower transmission than the previous model ( a 15% change in the optical depth in 2012 ) , reflecting an apparent increase in the rate of deposition on the filter . in this paper , we reconsider the light curve break result @xcite by reanalyzing previously published epochs using the newest calibration and by adding three recent observations ( 2011 march and september , 2012 april ) . we show that the break seen by @xcite is due primarily to the ( unmodeled ) increased rate of deposition on the optical blocking filter and not to a different phase in the evolution of the remnant .
table [ tabel ] lists the 23 monitoring observations used for this study .
we chose not to use the observations earlier in the mission ( obsids 1387 and 122 at 4608 and 4711 days after the explosion ) , as these observations were taken at a focal plane temperature of -110@xmath6 c and no charge transfer inefficiency adjustments are available for observations taken at this temperature .
however , we report the spatial distribution of the x - ray data at these epochs in section [ imaging ] .
since the remnant has thus far brightened by a factor of @xmath0 34 during the _ chandra _ mission , the observing configuration was changed several times to mitigate photon pileup effects .
first , this was done by reading out an increasingly smaller subarray of the acis - s3 chip .
all observations from day 7799 onward used the acis - hetg configuration , which produces both a zeroth - order image and a dispersed spectrum .
the dispersed spectrum provides an invaluable cross - check for our acis results . from day 8434
, the observation was moved to the bottom of the chip , to decrease the frame time to 1s .
the time of this transition coincides with the accelerated pace of the increase of the contamination on the filter . for the data reduction
, we largely follow the methods that were used before on _ chandra _ data of sn 1987a @xcite .
we filtered for background flares , and ignored the flags set by the _ chandra _
afterglow filter ( status=000000000000xxxx0000000000000000 ) as this filter tends to filter real events for bright sources ( e.g. , * ? ? ?
* ) . with an angular diameter of @xmath7 ,
sn 1987a is barely resolved in raw _
chandra _ acis images . to extract as much spatial information from the observations as possible , we adopt the same approach as in previous studies @xcite .
we remove the pixel randomization that is usually added by the _ chandra _ software . additionally , we improved the spatial resolution using split - pixel events @xcite .
we then deconvolve the image for the point spread function of the telescope using the lucy - richardson iterative deconvolution algorithm @xcite .
for the september 2011 and march 2012 observations , we analyze only the longest exposure , to avoid distortions caused by residual alignment errors between exposures .
figure [ overview ] shows the deconvolved images of all epochs of _ chandra _ observations to date . to characterize these images
, we fit the image with a model that consists of four lobes and a ring @xcite .
figure [ radii ] shows the best - fit radii of all the images considered in this study .
spectra were extracted from each observation within a circular aperture ( radius @xmath8 ) sized to encompass nearly all the light from the target in all epochs , using ciao version 4.4 @xcite and caldb version 4.4.10 .
background spectra were extracted from each observation within an annular region ( radius @xmath9 to @xmath10 ) .
even though we adjusted the observing modes to minimize pileup , pimms suggests that a modest pileup is present in many of the observations .
we first tried to take this into account during the fitting process by using the pileup model described by @xcite .
this approach adds to the overall model several pileup parameters which then have to be either assigned appropriate values or included in the fitting ; for details see the chandra abc guide to pileup " . as an example
, the parameter nregions encodes the spatial size of the source in units of @xmath11 pixel cells .
based on the acis image we can estimate that nregions should be in the range of 2 to 5 , but we can not determine the exact value a priori . including nregions in the fitting however adds a degeneracy between nregions and the other model parameters such that the fit and its flux are not well constrained
. an alternative , parameter - free strategy for addressing pileup in the analysis of acis data has been described by ( * ? ? ?
* appendix a ) .
first , a high - fidelity simulator of the acis detector @xcite , including charge transfer inefficiency @xcite , is used to infer an input photon spectrum that would cause acis to produce the observed spectrum and event rate under the specific configuration of the observation .
that photon spectrum is very flexible ; no spectral model is assumed .
the spatial distribution of the simulated photons incident on acis was produced by the marx package , using the corresponding sn 1987a morphology model described in
[ imaging ] .
marx and the ccd simulation were configured to match the parameters of the observation , such as the target location on the focal plane , frame time , and dead area / time .
second , with the input photon spectrum fixed , the simulation is repeated with only a single photon arriving in each frame , eliminating pileup .
the resulting simulated spectrum , which we refer to as `` pileup corrected '' , is then fit using standard methods and calibration files .
figure [ pileup ] shows the original and pileup - corrected spectrum of the observation that suffered worst from pileup ( obsid 4614 ) .
we applied this method to all acis observations reported here .
the flux correction attributed to pileup in each observation is reported in table [ tabel ] .
this table shows that the correction for the soft band is more smooth than the correction of the harder band .
this probably indicates that our pileup correction method is more stable for the soft band than the hard band .
we fitted the resulting pileup - corrected spectra with an absorbed two component spectral model , using the x - ray spectral fitting package xspec , version 12.6.0 @xcite .
the spectral model included a component in collisional equilibrium ionization ( ` vequil ` ) and a non - equilibrium ionization component ( ` vpshock ` ) using ` neivers ` 2.0 , updated with atomic data for inner shell processes ( c.f . * ? ? ?
* ) .we fixed the abundances to those found in the high resolution x - ray letg spectra investigated by @xcite and we fixed the absorption to 2.35@xmath12 @xmath3 @xcite .
figure [ pileup ] shows an example of one of the fits .
we measured the flux in the 0.5 - 2.0 and 3.0 - 8.0 kev energy bands from the best - fit model for all spectra and estimated flux uncertainties for a 68% confidence level .
figure [ lightcurve ] shows the evolution of the fluxes in the low and high energy bands for the observations corrected for pileup effects . for this study
, we do not need the high signal - to - noise that would be required for a detailed spectral analysis .
therefore , we decided to only correct the longest observation per epoch for pileup even if there were more observations available in the same epoch
. this explains the large error bars of the flux at day 7997 , as this epoch was divided into 5 observations . for comparison , we overplotted the fluxes of the last six epochs of _ xmm newton _ epic pn fluxes @xcite .
note that this instrument has been shown to be extremely stable ( e.g. , * ? ? ?
* ; * ? ? ?
after the previously described procedures , the 0.5 - 2.0 kev light curve increases more or less linearly from day 6400 on , but there still appears to be a break around day 8200 , which does not appear in the _ xmm - newton _ epic pn light curve @xcite . since the pileup corrections ( table [ tabel ] ) after the break
are relatively low ( 711% ) compared to some earlier observations , insufficient pileup correction is an unlikely explanation for the break .
we can confirm this directly using the data taken with the acis - hetg configuration , since the dispersed spectra in each of these observations have no pileup , they provide invaluable cross - checks for our pileup - corrected acis fluxes .
the dispersed spectra were extracted using the standard ciao tools and extraction sizes .
the medium energy grating ( meg ) plus and minus data are combined and the fluxes in the 0.79 to 2.1 kev range are determined directly from the flux - corrected data , as was done in @xcite .
in addition to the monitoring observations listed in table [ tabel ] , we analyzed the deep observations taken in 2007 ( day 7335 ) and 2011 ( day 8774 ) .
these observations have been described in @xcite .
the light curve obtained from the dispersed spectra ( figure [ hetgacis ] ) shows a break around day 8200 similar to that seen in figure [ lightcurve ] .
since this break occurs at the same time that we moved the aimpoint on the ccd , we can not disentangle any effects caused by that change from those caused by the increased contamination deposition rate , but we can be confident that this break is not caused by the increased count rate and insufficient pileup correction .
@xcite predicted a flattening of the 0.5 - 2.0 kev light curve around day 7000 and predicted the light curve to reach its maximum around day 8000 .
also , @xcite predicted the soft x - ray light curve to reach a maximum around day 9000 , but with a flux of at up to a factor of ten _ above _ the current x - ray flux .
indeed , @xcite showed that the flux in the 0.5 - 2.0 kev energy band flattened around day 8200 and argued that this flattening is a result of the transmitted shocks entering a lower - density region .
the improved calibration of the optical blocking filter now available removes a substantial part of this flattening
. however , there still appears to be a break in the _ chandra _ light curves , both in the piledup data ( bare acis and zeroth order hetg ) and in the pileup - free dispersed data .
note that all the _ chandra _ data presented here rely on the same calibration model for the optical blocking filter .
the _ xmm - newton _ epic - pn data show no evidence for for a sudden break around day 8200 . given that this break is visible in both dispersed and zeroth order , but not in the _ xmm - newton _ epic pn data , we conclude that this is not likely an effect of an insufficient pileup correction , but rather is caused by a miscalibration of the increase in contamination on the optical blocking filters , combined with a change in instrument setup .
note that this contamination has a wavelength dependent effect on the spectrum and hence has influence on the best - fit spectral parameters . also , this contamination has less of an effect on the 3.0 - 8.0 kev flux .
figure [ overview ] shows the evolution of sn 1987a s morphology . to interpret these images , one has to keep in mind that some of the structure in the images might be an artifact of the deconvolution method we used to correct for the extension of the point spread function of
chandra_. moreover , the point spread function of _ chandra _ is not entirely symmetric , as since december 2000 , there appeared a hook - shaped extension to the point spread function , which is not accounted for in marx 4.4 .
this hook contains about 6% of the total flux and will probably show up after the deconvolution as a small extending structure to the remnant .
the individual spots on the west side of the remnant appear to have merged in the last three epochs .
one question that comes to mind when looking at figure [ radii ] is `` how does this relate to the inner - ring structure as observed by @xcite ? '' .
@xcite compared radii of x - ray and radio images of sn 1987a and found comparable sizes , with radii larger than the radius of the equatorial ring in 2008 .
however , @xcite showed a strong break in the evolution of the soft x - ray light curve after the shocks have overcome the equatorial ring , and we do not yet see such evolution in the 0.5 - 2.0 kev light curve presented in figure [ lightcurve ] .
to estimate how far the shocks progressed through the equatorial ring , we need to know when the interaction started . @xcite carried out a study on multiple epochs of hubble space telescope ( hst ) images , ranging from 1994 february and 2000 february .
the first spot lit up by the blast wave is located in the northeast quadrant of the remnant and was detected in a hst observations taken 2933 days after the explosion .
remarkably , this spot is located 0.06@xmath13 inside of the main ring as visible in optical @xcite .
the second spot was only found in a hst observation taken 1350 days later and by day 4999 , 12 spots were visible in the hst observations @xcite . a lower limit to the width of the ring , as visible in the hst image , is of the order of @xmath14 ( @xmath15 km , * ? ? ?
fitting a broken linear function to the radii of figure [ radii ] results in a velocity of @xmath16 km s@xmath2 until day @xmath17 , after which the velocity is @xmath18 km s@xmath2 .
assuming that the shock wave reached the ring at day 4999 when 12 spots were lit up ( hereby ignoring the spots that were hit first , as those were protruding inwards ) , the radius has increased by @xmath19 km ( @xmath0120% of the width reported by * ? ? ?
* ) since the interaction started .
we used a monte carlo simulation to estimate the uncertainties on this value .
note that this distance relies on the assumption that the 12 spots reported by @xcite are part of the ring reported by @xcite .
@xcite argued , based on the evolution of the infrared emission , that the forward shock only fully started to interact with the equatorial ring at day @xmath05600 .
this appears consistent with the increase in soft x - ray flux around day 6000 @xcite .
the latter was considered by @xcite to be consistent with the lighting up of the optical spots @xmath0 1000 days earlier , as their models show some delay between the shock impact of dense material and the onset of x - ray emission .
nevertheless , if we take day 5600 as conservative assumption for the start of the explosion to interact with the ring , we estimate the increase of the radius of the x - ray emission since then to be @xmath20 km ( @xmath074% of the width reported by * ? ? ?
it is not straightforward to interpret the increase of the radius of the x - ray emission in terms of the 3 components as described in the introduction .
the velocity is too high for the component with low ( @xmath0 0.3 - 0.5 kev ) temperature @xcite .
@xcite reported the contribution of a high velocity component to be @xmath5 % in the hetg spectra , but its contribution to the broadband flux might be higher .
note that they measured the spectral lines to be much broader than 1700 km s@xmath2 .
therefore , the measured increase of the radius is likely a combination of the increase in radius of all three components .
the radius of the x - ray emission is identified in @xcite as following the progress of the forward shock / contact discontinuity as it moves through the uniform - with - clumps equatorial ring .
figure 9 ( top ) of @xcite shows the forward shock / contact discontinuity locations along with the @xcite radii and there is reasonable agreement . the speed of the forward shock there is of order 1700 km s@xmath2 .
current and future hst observations will reveal how far the shock has progressed , and how many of the optical spots have merged by now . a combined study with optical
, x - ray and infrared data will reveal a detailed picture of the structure of this equatorial ring , leading to better constraints on the last stages in the life of the progenitor of sn 1987a .
we report our results for the three newest epochs of _ chandra _ data on sn 1987a . together with our reanalysis of older epochs
, we paint a picture of the evolution of sn 1987a as witnessed by _
chandra _ in the past 10 years . based on this analysis
, we reach the following conclusions : * the increase of the radius of the ring visible in x - rays since the explosion first interacted with the preexisting equatorial ring is 74 - 120% of the lower limit to the thickness of this preexisting equatorial ring as measured by @xcite . *
given that the shocks have now traveled through a substantial amount of the preexisting equatorial ring , a multiwavelength study including high spatial resolution optical images is necessary to investigate how far the shock has traveled through the ring . *
the sudden break in the 0.5 - 2.0 kev light curve around day 8200 reported by @xcite was probably an instrumental effect .
the absence of a break in the light curve likely also indicates that the shocks traveling through the dense parts of the ring have not fully overcome the extent of the ring yet .
, p. s. , townsley , l. k. , feigelson , e. d. , getman , k. v. , garmire , g. p. , preibisch , t. , smith , n. , babler , b. l. , hodgkin , s. , indebetouw , r. , irwin , m. , king , r. r. , lewis , j. , majewski , s. r. , mccaughrean , m. j. , meade , m. r. , & zinnecker , h. 2011 , , 194 , 2 , a. , mcdowell , j. c. , allen , g. e. , brickhouse , n. s. , burke , d. j. , davis , j. e. , durham , n. , elvis , m. , galle , e. c. , harris , d. e. , huenemoerder , d. p. , houck , j. c. , ishibashi , b. , karovska , m. , nicastro , f. , noble , m. s. , nowak , m. a. , primini , f. a. , siemiginowska , a. , smith , r. k. , & wise , m. 2006 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 6270 , society of photo - optical instrumentation engineers ( spie ) conference series , p. , albrecht , r. , barbieri , c. , blades , j. c. , boksenberg , a. , crane , p. , deharveng , j. m. , disney , m. j. , kamperman , t. m. , king , i. r. , macchetto , f. , mackay , c. d. , paresce , f. , weigelt , g. , baxter , d. , greenfield , p. , jedrzejewski , r. , nota , a. , sparks , w. b. , kirshner , r. p. , & panagia , n. 1991 , , 369 , l63 , k. , tsunemi , h. , miyata , e. , baluta , c. j. , burrows , d. n. , garmire , g. p. , & chartas , g. 2001 , in astronomical society of the pacific conference series , vol .
251 , new century of x - ray astronomy , ed .
h. inoue & h. kunieda , 576 | we report on the present stage of sn 1987a as observed by the _
chandra _ x - ray observatory .
we reanalyze published _
chandra _ observations and add three more epochs of _ chandra _ data to get a consistent picture of the evolution of the x - ray fluxes in several energy bands .
we discuss the implications of several calibration issues for _ chandra _ data . using the most recent _ chandra _ calibration files , we find that the 0.5 - 2.0 kev band fluxes of sn 1987a have increased by @xmath0 6 @xmath1 erg s@xmath2 @xmath3 per year since 2009 .
this is in contrast with our previous result that the 0.5 - 2.0 kev light curve showed a sudden flattening in 2009 .
based on our new analysis , we conclude that the forward shock is still in full interaction with the equatorial ring . |
small free clusters of molecules with very high symmetry ( octahedral ) have been found to resemble bulk of the same substance even in the case of small number of ingredients ( 30 to 80 ) @xcite .
the number of particles ( molecules or / and atoms ) that mimics bulk properties like phase transitions is strongly dependent on the range of the potential .
the effect of decreasing the range of the potential is to destabilize strained structures @xcite . in the case of long - range potentials ( coulomb @xcite )
one detects structures , determined by the co - operative ( collective ) interaction - the system is expected to transform easier than in the case of short - range interactions . on the other side ,
the surface of free clusters plays much more important role for long - range potentials .
various parameters of the potential , for instance electric charge value , govern the cluster response to external changes .
an example is the melting temperature increase for higher charge values @xcite . in the present work we study clusters consisting of less symmetrical molecules ( tetrahedral ) in order to check how the lower symmetry of the ingredients influences cluster s behavior at different temperatures .
the interacting potential contains a long - range term ( coulomb ) , which makes it necessary to study the size influence - clusters of 50 , 137 , 229 methane molecules have been simulated with the help of the isoenergetic molecular dynamics method described in the next section .
simulations of methane clusters have been performed more than two decades ago @xcite .
they showed that the lower limit of the classic approach is @xmath210k . in those days the computers were less fast and the researchers tended to go for cheap rather than reliable potentials .
nowadays more sophisticated potentials are available and better results could be obtained @xcite .
this justifies re - visiting the problem of how the cluster size influences the structural phase changes .
another intriguing question is how a free cluster , e.g. a cluster at zero external pressure , behaves when the temperature is rather close to the ultimate low limit of the classic mechanics calculations .
+ solid ( bulk ) methane exhibits several crystalline phases based mainly on the difference in the orientation of molecules in the unit cell .
infrared investigation of solid methane performed by bini and co - workers @xcite let him propose a phase diagram , which has been additionally elucidate by nakahata _
et al . _ with a help of optical and x - ray diffraction studies @xcite . the liquid methane freezes at 90.6 k and @xmath3 mpa @xcite into a _ fcc _ orientational disordered phase i @xcite , characterized with rotational diffusion of the molecules .
a transition to the orientational ordered phase ii occurs at 20.4 k @xcite at zero pressure .
the phase ii is the james - keenan cubic structure , in which one out of four molecules ( @xmath4 ) is undergoing a weakly hindered rotation , while the other three molecules ( @xmath5 ) have an anti - ferro order @xcite .
the state of orientational disorder takes place for the systems of spherically- symmetric molecules ( @xmath1 ) with a symmetry lower than the site in the crystal structure @xcite . at cooling
such systems could undergo phase transitions to a partially - ordered or fully ordered state .
hindered rotations occur because of weakly angular - dependent intermolecular forces and a large rotational energy as it is in methane @xcite .
is such a behavior inherent for small systems as well ?
this is the question we study in the current work .
due to their spherical symmetry methane molecules resemble in many aspects the noble gases ( ar , kr , xe ) . the intra - molecular frequencies ( @xmath6 hz )
are one order of magnitude higher than the frequency of the intermolecular vibrations .
that is why , we consider the molecules as rigid bodies ( c - h distance b=1.094 _ _ ) .
suitable intermolecular potentials for regarding molecular space orientations could be the lenard - jones potential @xcite or a more sophisticated 3-body rmk potential @xcite . in both approaches
complicated procedures had been followed to obtain parameters values .
needless to say , these parameters are not entirely satisfactory . in our study , the total potential is a sum of pair atom - atom potentials @xmath7 : @xmath8 where @xmath9 , @xmath10 denote either a carbon or a hydrogen atom ; @xmath11 is the distance between the @xmath12-th and the @xmath13-th atom ; @xmath14 is either @xmath15 or @xmath16 in elementary charge units .
+ the parameters have been obtained in a way to fit experimental data .
for instance , the charge and the bond - length are determined from the octopole moment value .
@xmath17 in _ _ , @xmath18 in _ mev _ are @xcite : + @xmath19 , @xmath20 , @xmath21 ; + @xmath22 , @xmath23 , @xmath24 .
+ two starting configurations with randomly oriented molecules have been designed : 1 ) a _ bcc _ lattice , which resembles the most spherically - symmetric shape of a cluster and 2 ) _ fcc _ , which is a possible lattice for a bulk methane .
we execute the velocity verlet algorithm @xcite for numerical integration of classic equations of motion in a special micro - canonical ensemble - n , e , p=0 ( free clusters ) at temperatures above 10_k _ to avoid quantum effects .
the integration step of @xmath25 guarantees an accuracy greater than 0.003% of the energy conservation for runs up to @xmath26 0.5_ns _ ; the records are taken every 100 _
liquid- and solid - like phases can be distinguished with the help of a specially constructed formula , which is an extension of the lindemann index @xmath27 @xcite : @xmath28 where @xmath11 is the distance between molecular centers of mass . in bulk
, the lindemann index is computed on the basis of the mean deviation of the @xmath12-th atom from its ideal lattice position and @xmath29 indicates a solid state . in a free cluster
the surface plays a destabilizing role and the cluster is solid if @xmath30 @xcite .
+ + the lindemann index , plotted in the figs.([lin - b ] - [ 137lin ] ) , shows three distinguishable phases : two solid - like phases below 30_k _ and a liquid - like phase above 30_k _ for all three sizes and both starting configurations .
+ the caloric curves ( total energy per particle as a function of temperature ) are given in the fig.[cal - b ] and fig.[cal - f ] .
although the caloric curves are too smooth ( a change of the slope is hardly seen , no more than 6% ) , we observe a region of frustration in the temperature interval ( 18,25 ) _
k_. in our computations the average temperature of the system is computed from its average kinetic energy : @xmath31 , with @xmath32 - the boltzmann constant . analyzing the trajectories at different total energies
, we see that the system temperature jumps unevenly in the above interval .
this is an indication for changes much more clearly seen with the help of the lindemann index .
there is no hysteresis on heating - cooling of the system - fig.[nohyst ] .
+ the different phase structures are distinguished on the basis of their radial distribution functions - eq .
[ gr ] : the radial distribution of molecular centers of mass shows the structure of the cluster at a specific temperature , while the atom - atom radial distribution reveals orientational order ( disorder ) of the molecules .
@xmath33 + fig.[gr - mm ] shows the existence of a solid - like structure below 30 _ k _ and the absence of that structure above 30 _ k_. the volume of the cluster enlarges at heating as well .
fig.[gr - at ] shows typical distribution of the atoms for the oriented ( solid line in the figure ) and disoriented ( dashed line in the figure ) solid - like phases .
the first peak of the distribution correspond to the h - h neighbors , the second peak corresponds to the c - h neighbors , and the third one is for the c - c neighbors .
the same maximum of the h - h and c - h peaks indicates a unique mutual orientation of the molecules , e.g the cluster is orientationally ordered .
the phase transition at @xmath2 20k causes disorientation of the molecules as it is seen in the fig.[gr - at ] - the dotted curve . + + finally , we present the density of states ( dos ) function obtained by a normal mode analysis of the quenched system @xcite .
the fig.[vra - spec ] shows evidently the existence of a rovibrational coupling @xcite .
the fig.[vib - spec ] shows broadening of the vibrational spectrum at heating .
implementing molecular dynamics method to a microcanonical state of free methane clusters , we show that they exhibit transformations between an orientationally - ordered ( at @xmath2 10k ) and disordered phase ( above 20k ) .
the lack of hysteresis in the cooling - heating regime indicates a continuous transition .
however , such a statement can be confirmed either by studying infinite systems ( i.e. periodic boundary conditions ) or by applying the finite - size scaling theory @xcite .
the authors would like to thank prof .
arias group from the cornell university for the access to the computer facilities .
the work is supported by a special grant of the bulgarian ministry of education and science ( f-3 , 2003 ) .
10 a. proykova and r.s . berry ( 1997 )
_ z. phys.d _ * 40 * , 215 .
doye , d.j .
wales and r.s . berry ( 1995 ) _ j. chem .
phys . _ * 103 * , 4234 - 4249 . s. pisov and a. proykova ( 2002 ) _
comp.phys.comm._*147/1-2 * , 238 - 241 .
ruth lynden - bell , private communication m. p. allen and tildesley ( 1994 ) _ computer simulation of liquids _ , clarendon : oxford .
r. bini and g. pratesi ( 1997 ) _ phys .
b _ * 55 * , 14800 .
i. nakahata , n. matsui , y. akahama , h.kavamura ( 1999 ) _ chem .
lett . _ * 302 * , 359 .
r.l.amey and r.h.cole ( 1964 ) _ j. chem .
phys . _ * 40 * , 146 .
r. m. lynden - bell and k. h. michel ( 1996 ) _ rev .
phys . _ * 66 * , 721 .
k. clusius ( 1929 ) _ z. phys .
chem.b _ * 3 * , 41 . w. press ( 1972 ) _
* 56 * , 2597 .
r. righini , k. maki , m. l. klein ( 1981 ) chem .
* 80 * , 301 .
l. verlet ( 1967 ) _ phys .
_ * 159 * , 98 .
garzon and j. jellinek ( 1991)_z .
d _ * 20 * , 235 - 238 .
a. proykova , s. pisov and s. berry ( 2001 ) _ j. chem . phys .
_ * 115 * , 8583 .
stillinger and t.a . weber ( 1981 ) _ kinam _ * 3 * , 159 .
ferdinand and m.e .
fisher ( 1969 ) _ phys .
rev . _ * 185 * , 832 . | low - temperature ( @xmath0 60k ) phase changes of free clusters containing 50 , 137 , and 229 @xmath1 molecules have been observed in isoenergetic molecular dynamics computations .
bulk solid methane exhibits structural phase transformation at 20.4 k. however , clusters of 50 molecules already melt at @xmath2 25 k , which makes the observation of solid - solid transformations rather challenging . |
the anomalous magnetic moment of the muon defined as @xmath5 , can be accurately measured and , within the sm framework , precisely predicted@xcite .
the experimental value of @xmath2 ( @xmath6 ) measured at the brookhaven laboratory differs from the sm estimates by 3.2 - 3.6 @xmath7 @xcite .
if the deviation is confirmed with higher precision it would signal of new physics .
the main theoretical uncertainty for @xmath2 comes from hadronic contributions .
the leading order hadronic term can be derived from a combination of experimental cross section data , related to @xmath8 annihilation to hadrons . at @xmath3
the differential cross section ( as a function of @xmath9 ) for the @xmath10 initial state radiation ( isr ) process is measured .
then , the dipion cross section @xmath11 has been obtained from : @xmath12 where the radiator function _
h _ is computed from qed with complete nlo corrections and depends on the initial @xmath8 center of mass energy squared @xmath13 .
the dipion cross section @xmath14 obtained from eq .
[ dipion_cross_section ] requires the correction for final state radiation ( fsr )
[ dipion_cross_section ] is also valid for the @xmath15 and @xmath16 processes with the same radiator function _
h_. thus , we can determine @xmath14 from the ratio of the @xmath17 and @xmath18 differential cross sections for the same value of the dipion and dimuon invariant mass .
+ the pion form factor can then be determined using the following equation : @xmath19 where @xmath20 is the vacuum polarization ( vp ) correction , @xmath21 accounts for the fsr radiation assuming point - like pions .
@xmath22 is a bare cross section , i.e. corrected for the running of @xmath23 and inclusive of fsr , defined as @xcite @xmath24 @xmath25 where @xmath26 .
+ many radiative corrections drop out for this ratio method : contributions due to the radiator function ( this allows to suppress the related systematic uncertainty of 0.5% for the direct @xmath14 measurement ) , to the integrated luminosity ( since the data for the @xmath17 and @xmath18 processes are collected simultaneously ) and finally to the vacuum polarization .
in the 2008 and 2010 two analyses of the @xmath28 have been performed at @xmath3 with the kloe detector . + a cross section of the detector in the y , z plane is shown in fig.[detector08 ] .
+ the kloe08 analysis @xcite used a data sample corresponding to an integrated luminosity of @xmath29 pb@xmath30 collected at @xmath31 in 2002 and selection cuts in which the photon is emitted within a cone of @xmath32 around the beamline ( narrow cones in fig.[detector08 ] ) and the two charged pion tracks have @xmath33 ( wide cones in fig . [ detector08 ] ) . in this configuration ,
the photon is not detected and the photon momentum is reconstructed from missing momentum : @xmath34 .
these selection cuts provide high statistics data sample for the isr signal events , and significantly reduce contamination from the resonant process @xmath35 . from the bare cross section , the dipion contribution to the muon anomaly @xmath36
is measured : @xmath37 .
the kloe10 analysis @xcite was performed requiring events that are selected to have a photon at large polar angles between @xmath38 ( wide cones in fig.[detector08 ] ) , in the same angular region as the pions .
+ this selection allow to access the two pion threshold .
however , compared to the measurement with photons at small angles , this condition reduces statistics and increases the background from the process @xmath39 .
the dispersion integral for @xmath36 is computed as the sum of the values for @xmath40 times the kernel k(s ) , times @xmath41 @xmath42 the following value for the dipion contribution to the muon anomaly @xmath36 was found : @xmath43 . the last kloe measurement of the @xmath27 cross section ( kloe12 ) has been obtained from the ratio between the pion and muon isr differential cross section .
the data sample is the same as for the kloe08 analysis . +
the separation between the @xmath44 and @xmath45 events is obtained assuming the final state with two charged particles with equal mass @xmath46 and one photon .
the @xmath47 identifies the muons and @xmath48 the pions .
the selection procedure has been compared to other techniques , such as a kinematic fit or applying a quality cut on the helix fit for both tracks , all leading to consistent results .
+ trigger , particle identification and tracking efficiencies have been checked using control data samples .
+ the differential @xmath45 cross section is obtained from the observed event count @xmath49 and background estimate @xmath50 , as : @xmath51 where _ l _ is the integrated luminosity from ref .
@xcite and @xmath52 the selection efficiency .
the bare cross section @xmath40 ( inclusive of fsr , with vp effects removed ) is obtained from the bin - by - bin ratio of the kloe08 @xmath44 and the described above @xmath45 differential cross sections .
the bare cross section is used in the dispersion integral to compute @xmath36 . the pion form factor @xmath53
is extracted using eq .
( [ f_pi ] ) .
[ delta_amu ] gives @xmath54 in the interval @xmath55 .
for each bin contributing to the integral , statistical errors are combined in quadrature and systematic errors are added linearly .
+ the last three kloe estimations on the @xmath36 ( kloe08 , kloe10 , kloe12 ) have been compared and are consistent ( as you can see in the table [ tab1 ] ) .
.comparison of @xmath36 between the kloe12 and the previous kloe measurements ( kloe08 , kloe10 ) . [ cols="<,<,<",options="header " , ] * @xmath56 * + the preliminary combination of these kloe results is reported in the figure [ combination]@xcite .
it is obtained using the best linear unbiased estimate ( _ blue _ ) method @xcite . in the fig . [ combination ] ( left )
the pion form factor measurements for the three kloe analysis and the fractional difference ( right ) are shown@xcite .
the cross section ratio method used in the kloe12 measurement reduces significantly the theoretical and the systematic error .
+ the following @xmath57 values are found : + @xmath58 @xmath59 .
precision measurements of the pion vector form factor using the initial state radiation ( isr ) have been performed by the kloe / kloe-2 collaboration during the last 10 years .
the preliminary consolidation of the last analysis ( kloe12 ) with two previously published ( kloe08 , kloe10 ) ones has been presented .
the result confirms the current discrepancy ( @xmath60 ) between the standard model ( sm ) calculation and the experimental value of the muon anomaly @xmath2 measured at bnl .
+ in the near future the @xmath61 physics program of the kloe-2 experiment@xcite will further shed light in this field , with e.g. the study of the radiative width of pseudoscalar mesons and of the transition form factors @xcite , thanks to the luminosity upgrade of da@xmath4ne and the kloe upgrade with the addition of new detectors : low energy @xcite and high energy @xcite @xmath8 taggers , an inner tracker @xcite , crystal calorimeters ( ccalt ) @xcite , and tile calorimeters ( qcalt ) @xcite .
f. jegerlehner , a. nyffeler , phys . rept . * 477 * ( 2009 ) 1 .
+ m. davier , a. hoecker , b. malaescu , z. zhang , eur . phys . j. c * 71 * ( 2011 ) 1515 .
+ k. hagiwara , et al .
, j. phys .
g * 38 * ( 2011 ) 085003 .
bennett , et al .
, muon g-2 collaboration , phys .
d * 76 * ( 2006 ) 072003 .
brodsky and e. de rafael , phys .
rev . * 168 * , 1620 ( 1968 ) . for low energies collaboration , eur
. phys . j. c * 66 * ( 2010 ) 585 .
m. adinolfi , et al .
, kloe collaboration , nucl .
methods a * 488 * ( 2002 ) 51 .
m. adinolfi , et al . ,
kloe collaboration , nucl .
methods a * 482 * ( 2002 ) 364 .
d.babusci et al.,kloe and kloe-2 collaborations , physics letters b * 720 * ( 2013 ) 336 .
f. ambrosino , et al . ,
kloe collaboration , phys .
b * 670 * ( 2009 ) 285 .
f. ambrosino , et al . ,
kloe collaboration , phys .
b * 700 * ( 2011 ) 102 .
f. ambrosino , et al . ,
kloe collaboration , eur .
j. c * 47 * ( 2006 ) 589 .
s. mueller , contribution to `` mini - proceedings , 15th meeting of the working group on rad . corrections and mc generators for low energies , '' , arxiv:1406.4639 [ hep - ph ] .
editors : s. e. mueller and g. venanzoni ; + https://agenda.infn.it/getfile.py/access?contribid=25&sessionid=13 & + resid=0&materialid = slides&confid=7800 a. valassi nim a * 500 * ( 2003 ) 391 . g. dagostini nim a * 346 * ( 1994 ) 306 . g. amelino - camelia _ et al . _
( kloe2 collaboration ) , eur .
j. c * 68 * , 619 - 681 ( 2010 ) d. babusci et al .
( kloe2 collaboration ) eur .
j. c ( 2012 ) 72:1917 d. babusci et al
. nucl.instrum.meth . *
a617 * 81 ( 2010 ) f. archilli et al .
* a617 * 266 ( 2010 ) a. balla et al . ,
jinst 9 , c01014 ( 2014 ) f. happacher _
suppl . * 197 * , 215 ( 2009 ) .
m. cordelli _
et al . _ ,
instr . & meth .
* a 617 * , 105 ( 2010 ) . | the measurement of the @xmath0 cross section allows to determine the pion form factor @xmath1 and the two pion contribution to the muon anomaly @xmath2 .
such a measurement has been performed with the kloe detector at @xmath3 , the frascati @xmath4-factory .
+ the preliminary results on the combination of the last analysis ( kloe12 ) with two previous published ( kloe08 , kloe10 ) will be presented in the following . |
the casimir effect@xcite is the force and also the specific polarization of the vacuum arising in restricted quantization volumes and originating from the zero - point oscillations of quantized fields .
this force acts between two closely spaced macrobodies , between an atom or a molecule and macrobody or between two atoms or molecules . during more than fifty years ,
passed after the discovery of the casimir effect , it has attracted much theoretical attention because of numerous applications in quantum field theory , atomic physics , condensed matter physics , gravitation and cosmology , mathematical physics , and in nanotechnology ( see monographs and reviews ) . in multidimensional kaluza - klein supergravity
the casimir effect was used@xcite as a mechanism for spontaneous compactification of extra spatial dimensions and for constraining the yukawa - type corrections to newtonian gravity.@xcite in quantum chromodynamics the casimir energy plays an important role in the bag model of hadrons.@xcite in cavity quantum electrodynamics the casimir interaction between an isolated atom and a cavity wall leads to the level shifts of atomic electrons depending on the position of the atom near the wall.@xcite both the van der waals and casimir forces are used@xcite for the theoretical interpretation of recent experiments on quantum reflection and bose - einstein condensation of ultracold atoms on or near the cavity wall of different nature .
in condensed matter physics the casimir effect turned out to be important for interaction of thin films , in wetting processes , and in the theory of colloids and lattice defects.@xcite the casimir force was used to actuate nanoelectromechanical devices@xcite and to study the absorption of hydrogen atoms by carbon nanotubes.@xcite theoretical work on the calculation of the casimir energies and forces stimulated important achievements in mathematical physics and in the theory of renormalizations connected with the method of generalized zeta function and heat kernel expansion.@xcite all this made the casimir effect the subject of general interdisciplinary interest and attracted permanently much attention in the scientific literature .
the last ten years were marked by the intensive experimental investigation of the casimir force between metallic test bodies ( see refs .
) . during this time the agreement between experiment and theory on the level of 1 - 2% of the measured force
was achieved .
this has become possible due to the use of modern laboratory techniques , in particular , of atomic force microscopes and micromechanical torsional oscillators .
metallic test bodies provide advantage in comparison with dielectrics because their surfaces avoid charging . in refs .
, where the importance of the casimir effect for nanotechnology was pioneered , it was demonstrated that at separations below 100 nm the casimir force becomes larger than the typical electrostatic forces acting between the elements of microelectromechanical systems .
bearing in mind that the miniaturization is the main tendency in modern technology , it becomes clear that the creation of new generation of nanotechnological devices with further decreased elements and separations between them would become impossible without careful account and calculation of the casimir force .
successful developments of nanotechnologies based on the casimir effect calls for more sophisticated calculation methods of the casimir forces .
most of theoretical output produced during the first decades after casimir s discovery did not take into account experimental conditions and real material properties of the boundary bodies , such as surface roughness , finite conductivity and nonzero temperature .
the basic theory giving the unified description of both the van der waals and casimir forces was elaborated by lifshitz@xcite shortly after the publication of casimir s paper .
it describes the boundary bodies in terms of the frequency dependent dielectric permittivity @xmath0 at nonzero temperature @xmath1 . in the applications of the lifshitz theory to dielectrics it was supposed that the static dielectric permittivity ( i.e. , the dielectric permittivity at zero frequency ) is finite .
the case of ideal metals was obtained from the lifshitz theory by using the so - called schwinger s
prescription,@xcite i.e. , that the limit @xmath2 should be taken first and the static limit @xmath3 second . for ideal metals the same result , as follows from the lifshitz theory combined with the schwinger s prescription , was obtained independently in the framework of thermal quantum field theory in matsubara formulation.@xcite however , the cases of real dielectrics and metals ( which possess some nonzero dc conductivity at @xmath4 and finite dielectric permittivity at nonzero frequencies , respectively ) remained practically unexplored for a long time .
the case of semiconductor boundary bodies was also unexplored despite of the crucial role of semiconductor materials in nanotechnology .
starting in 2000 , several theoretical groups in different countries attempted to describe the casimir interaction between real metals at nonzero temperature in the framework of lifshitz theory .
they have used different models of the metal conductivity and arrived to controversial conclusions . in ref . , using the dielectric function of the drude model , quite different results than for ideal metals were obtained . according to ref .
, at short separations ( low temperatures ) the thermal correction to the casimir force acting between real metals is several hundred times larger than between ideal ones .
in addition , at large separations of a few micrometers ( high temperatures ) a two times smaller magnitude of the thermal casimir force was found than between ideal metals ( the latter is known as `` the classical limit''@xcite ) . in refs . , the dielectric permittivity of the plasma model was used to describe real metals and quite different results were obtained . at short separations
the thermal correction appeared to be small in qualitative agreement with the case of ideal metals . at large separations for real metals
the familiar classical limit was reproduced .
later the approach of ref . was supported in refs . , .
the plasma model approach can be used at such separations @xmath5 that the characteristic frequency @xmath6 belongs to the region of infrared optics .
later a more general framework , namely the impedance approach was suggested@xcite which is applicable at any separation larger than the plasma wave length .
it was supported in refs . , , .
in the region of the infrared optics , the impedance approach leads to practically the same results as the plasma model approach . as was shown in refs .
, , the drude model approach leads to the violation of the nernst heat theorem when applied to perfect metal crystal lattices with no impurities .
this approach was also excluded by experiment at 99% confidence in the separation region from 300 to 500 nm and at 95% confidence in the wider separation region from 170 to 700nm.@xcite on the contrary , the plasma model and impedance approaches were shown to be in agreement with thermodynamics and consistent with experiment . the polemic between different theoretical approaches to the description of the thermal casimir force in the case of real metals can be found in refs . , . these findings on the application of the lifshitz theory to real metals have inspired a renewed interest in the casimir force between dielectrics . as was mentioned above , at nonzero temperature dielectrics possess an although small but not equal to zero dc conductivity . in ref .
, the van der waals force arising from the dc conductivity of a dielectric plate was shown to lead to large effect in noncontact atomic friction,@xcite a phenomenon having so far no satisfactory theoretical explanation .
this brings up the question : is it necessary or possible to take into account the dc conductivity of dielectrics in the lifshitz theory ? recall that in the case of a positive answer the static dielectric permittivity of a dielectric material would be infinitely large .
it is amply clear that the resolution of the above issue should be in accordance with the fundamentals of thermodynamics . for this reason ,
it is desirable to investigate the low - temperature behavior of the casimir free energy and entropy for two dielectric plates both with neglected and included effects of the dc conductivity .
a major breakthrough in the investigation of this problem was achieved in the year 2005 . in ref .
a new variant of perturbation theory was developed in a small parameter proportional to the product of the separation distance between the plates and the temperature . as a result
, the behavior of the casimir free energy , entropy and pressure at low temperatures was found analytically .
if the static dielectric permittivity is finite , the thermal correction was demonstrated to be in accordance with thermodynamics .
this solves positively the fundamental problem about the agreement between the lifshitz theory and thermodynamics for the case of two dielectric plates . in ref .
it was shown that , on the contrary , the formal inclusion of a small conductivity of dielectric plates at low frequencies into the model of their dielectric response leads to a violation of the nernst heat theorem .
this result gives an important guidance on how to extrapolate the tabulated optical data for the complex refractive index to low frequencies in numerous applications of the van der waals and casimir forces .
all these problems and related ones arising for semiconductor materials are discussed in this review . in sec .
2 we derive the analytical behavior of the casimir free energy , entropy and pressure in the configuration of two parallel dielectric plates at both low and high temperature .
it is demonstrated that if the static dielectric permittivity is finite the lifshitz theory is in agreement with thermodynamics .
sec . 3 contains the derivation of the low - temperature behavior for the casimir free energy and entropy between two dielectric plates with included dc conductivity . in this case the lifshitz theory is found to be in contradiction with the nernst heat theorem .
the conclusion is made that the conductivity properties of a dielectric material at a constant current are unrelated to the van der waals and casimir forces and must not be included into the model of dielectric response . in sec .
4 we consider the thermal casimir force between metal and dielectric .
this problem was first investigated in ref . .
it was found that for dielectric plate with finite static dielectric permittivity the nernst heat theorem is satisfied but the casimir entropy may take negative values . here
we not only reproduce an analytical proof of the nernst heat theorem but also find the next perturbation orders in the expansion of the casimir free energy and entropy in powers of a small parameter .
the results obtained in high - temperature limit are also provided .
5 is devoted to the casimir interaction between metal and dielectric plates with included dc conductivity of the dielectric material .
we demonstrate that in this case the nernst heat theorem is violated .
6 contains the discussion of semiconductors which present a wide variety of electric properties varying from metallic to dielectric .
we consider the casimir interaction between metal and semiconductor test bodies and formulate the criterion when it is appropriate to include the dc conductivity of a semiconductor into the model of dielectric response .
in doing so , the results of recent experiments@xcite on the measurement of the casimir force between metal and semiconductor test bodies are taken into account . to this point
the assumption has been made that metal , dielectric and semiconductor materials of the casimir plates possess only temporal dispersion , i.e. , can be described by the dielectric permittivity depending only on frequency . in sec .
7 we discuss recent controversial results by different authors ( see , e.g. , refs .
) attempting to take into account also spatial dispersion . as is shown in this section , the way of inclusion of spatial dispersion into the lifshitz theory , used in refs .
, is unjustified .
we argue that the account of spatial dispersion can not influence theoretical results obtained with the help of usual , spatially local , lifshitz theory within presently used ranges of experimental separations .
8 contains our conclusions and discussion .
we consider two thick dielectric plates ( semispaces ) described by the frequency - dependent dielectric permittivity @xmath0 and restricted by the parallel planes @xmath7 with a separation @xmath5 between them , in thermal equilibrium at temperature @xmath1 .
the lifshitz formula for the free energy of the van der waals and casimir interaction between the plates is given by@xcite @xmath8%\right .
+ \ln\big[1-r_{\bot}^{2}(\xi_l , k_{\bot})e^{-2aq_l}\big ] \right\ } , \nonumber\end{aligned}\ ] ] where the reflection coefficients for two independent polarizations of electromagnetic field are defined as @xmath9 here @xmath10 is the magnitude of the wave vector in the plane of plates , @xmath11 are the matsubara frequencies , @xmath12 is the boltzmann constant , @xmath13 , and @xmath14 the problems in the application of the lifshitz theory to real materials discussed in the introduction are closely connected with the values of the reflection coefficients at zero matsubara frequency . for later use
we discuss it for the various cases .
* for ideal metals@xcite it holds @xmath15 * for real metals described by the dielectric function of the drude model , @xmath16 } , \label{eq5}\ ] ] where @xmath17 is the plasma frequency and @xmath18 is the relaxation parameter , it holds@xcite @xmath19 eq . ( [ eq6 ] ) results in the discontinuity between the cases of ideal and real metals and leads to the violation of the nernst heat theorem for metallic plates having perfect crystal lattices.@xcite * for real metals described by the dielectric function of the plasma model , @xmath20 from eq .
( [ eq2 ] ) it follows:@xcite @xmath21 here , in the limit of ideal metals ( @xmath22 ) the continuity is preserved because @xmath23 in eq .
( [ eq8 ] ) goes to unity .
the free energy ( [ eq1 ] ) calculated with the permittivity ( [ eq8 ] ) is also consistent with thermodynamics . * for dielectrics and semiconductors
the dielectric permittivities at the imaginary matsubara frequencies are given by the ninham - parsegian representation,@xcite @xmath24 where the parameters @xmath25 are the absorption strengths satisfying the condition @xmath26 and @xmath27 are the characteristic absorption frequencies .
here , the static dielectric permittivity @xmath28 is supposed to be finite .
although eq .
( [ eq10 ] ) is an approximate one , it gives a very accurate description for many materials.@xcite by the substitution of eq .
( [ eq9 ] ) in eq .
( [ eq2 ] ) one arrives at @xmath29 note that the vanishing of the transverse reflection coefficient for dielectrics at zero frequency in eq .
( [ eq11 ] ) has another meaning than for the drude metals in eq .
( [ eq6 ] ) . for drude metal
the parallel reflection coefficient is equal to the physical value for real photons at normal incidence , i.e. , to unity , and the transverse one vanishes instead of taking unity , its physical value .
this results in the violation of the nernst heat theorem for perfect crystal lattices . in the case of dielectrics
both reflection coefficients at zero frequency in eq .
( [ eq11 ] ) depart from the physical value for real photons which is equal to @xmath30 . in this case
, however , one of them is larger and the other one is smaller than the physical value . as we will see below , this leads to the preservation of nernst s heat theorem confirming that eq .
( [ eq9 ] ) , despite being approximate , describes the material properties of dielectric and semiconductor plates in a thermodynamic consistent way .
now we derive the analytic representation for the casimir free energy in eq .
( [ eq1 ] ) at low temperatures . for convenience in calculations ,
we introduce the dimensionless variables @xmath31 where @xmath32 is the characteristic frequency , @xmath33 , and @xmath34 was defined in eq .
( [ eq3 ] ) .
then the lifshitz formula ( [ eq1 ] ) takes the form @xmath35 where @xmath36 , \label{eq22}\end{aligned}\ ] ] and reflection coefficients ( [ eq2 ] ) , in terms of variables ( [ eq12 ] ) , being given by @xmath37 to separate the temperature independent contribution and thermal correction in eq .
( [ eq13 ] ) we apply the abel - plana formula,@xcite @xmath38 where @xmath39 is an analytic function in the right half - plane . here , taking it as @xmath40 and using eq .
( [ eq15 ] ) , we can identically rearrange eq .
( [ eq13 ] ) to the form @xmath41 where @xmath42 is the energy of the van der waals or casimir interaction at zero temperature , @xmath43 and @xmath44 is the thermal correction to this energy , @xmath45 note that , in fact , eq . ( [ eq19 ] ) describes the dependence of the free energy on the temperature arising from the dependence on temperature of the matsubara frequencies .
thus , @xmath44 in ( [ eq19 ] ) coincides with the thermal correction to the energy , defined as @xmath46 , only for plate materials with temperature independent properties .
the asymptotic expressions for the energy @xmath42 at both short and large separations are well known.@xcite below we find the asymptotic expressions for the thermal correction ( [ eq19 ] ) under the conditions @xmath47 and @xmath48 . taking into account the definition of @xmath49 in eq .
( [ eq12 ] ) , the asymptotic expressions at @xmath47 are applicable both at small and large separations if the temperature is sufficiently low .
we begin with condition @xmath47 .
let us substitute eq .
( [ eq9 ] ) in eqs .
( [ eq21 ] ) ( [ eq14 ] ) , expand the function @xmath50 in powers of @xmath51 , and than integrate the obtained expansion with respect to @xmath52 from @xmath53 to infinity in order to find @xmath54 in eq .
( [ eq20 ] ) and @xmath55 in eq .
( [ eq19 ] ) .
it is easy to check that @xmath56 does not contribute to the leading , second , order in the expansion of @xmath55 in powers of @xmath53 .
thus , we can restrict ourselves by the consideration of the expansion @xmath57 where @xmath58 was defined in eq .
( [ eq11 ] ) .
note that for simplicity we consider here only one oscillator in eq .
( [ eq9 ] ) and put @xmath59 .
the case of several oscillator modes can be considered in an analogous way . as a next step ,
we integrate eq .
( [ eq23 ] ) term by term according to eq .
( [ eq20 ] ) , expand the partial results in powers of @xmath53 and sum up the obtained series .
thereby we obtain the following expressions : @xmath60 , \label{eq28}\end{aligned}\ ] ] where li@xmath61 is the polylogarithm function and ei@xmath62 is the exponential integral function
. from these equations it follows @xmath63 and , thus , @xmath64 and @xmath65 do not contribute to the leading order in the expansion of @xmath55 .
the latter is determined by @xmath66 only . as a result
, we arrive at @xmath67 where @xmath58 was substituted from eq .
( [ eq11 ] ) and @xmath68 was introduced for the still unknown real coefficient of the next to leading order resulting from @xmath64 and @xmath66 as well as , possibly , from @xmath56 . at this stage
it is difficult to determine the value of this coefficient because all powers in the expansion of @xmath50 contribute to it .
remarkably , the two leading orders depend only on the static dielectric permittivity @xmath69 and are not influenced by the dependence of the dielectric permittivity on the frequency contained in @xmath65 .
substituting eq .
( [ eq31 ] ) in eq .
( [ eq19 ] ) and using eq .
( [ eq16 ] ) , we obtain @xmath70 , \label{eq32}\ ] ] where @xmath71 and @xmath72 is the riemann zeta function .
so far we have considered the free energy .
the thermal pressure is obtained as @xmath73 , \label{eq33}\ ] ] where @xmath74 is the casimir pressure at zero temperature . in order to determine the value of the coefficient @xmath75 of the leading term
, we express the pressure directly through the lifshitz formula @xmath76\!. \label{eq34}\end{aligned}\ ] ] again , applying the abel - plana formula ( [ eq15 ] ) , we represent the pressure as follows , @xmath77 where the thermal correction to @xmath78 , the pressure at zero temperature , is @xmath79 and the function @xmath80 is given by @xmath81 first , we determine the leading term of the expansion of @xmath82 in powers of @xmath53 . for this purpose ,
let us introduce the new variable @xmath83 and note that the reflection coefficient @xmath84 depends on @xmath53 only through the frequency dependence of @xmath85 given by eq .
( [ eq9 ] ) .
thus , we can rewrite and expand eq .
( [ eq37 ] ) as follows : @xmath86 where , according to eq .
( [ eq14 ] ) , @xmath87 integration in eq .
( [ eq39 ] ) with account of eq .
( [ eq40 ] ) results in @xmath88\,\frac{x^3}{6 } + \mbox{o}(x^4)\ , , \label{eq41}\ ] ] from which it follows : @xmath89\frac{x^3}{3 } + \mbox{o}(x^5 ) .
\label{eq42}\ ] ] the expansion of @xmath90 from eq .
( [ eq37 ] ) in powers of @xmath53 is somewhat more cumbersome .
it can be performed in the following way . as is seen from the second equality in eq .
( [ eq29 ] ) , the dependence of the dielectric permittivity on frequency contributes to @xmath55 starting from only the 5th power in @xmath53 .
bearing in mind the connection between free energy and pressure , we can conclude that the dependence on the frequency contributes to @xmath91 starting from the 4th order .
we are looking for the lowest ( third ) order expansion term of @xmath91 .
because of this , it is permissible to disregard the frequency dependence of @xmath92 and describe the dielectric by its static dielectric permittivity . to begin with , we identically rearrange @xmath90 in eq .
( [ eq37 ] ) by subtracting and adding the two first expansion terms of the function under the integral in powers of @xmath53 , @xmath93 \nonumber \\ & & \phantom{aaaaa } + \int_{x}^{\infty}y^2dy\frac{r_0 ^ 2}{e^y - r_0 ^ 2}- x^2\frac{2{\ee}_0}{{\ee}_0 + 1 } \int_{x}^{\infty}dy \frac{r_0 ^ 2e^{-y}}{\left(1-r_0 ^ 2e^{-y}\right)^2}\ , , \label{eq43}\end{aligned}\ ] ] and consider these three integrals separately .
the first integral in terms of the new variable @xmath83 reads @xmath94 , \label{eq44}\ ] ] where , in accordance with eq .
( [ eq14 ] ) , @xmath95 expanding @xmath96 in powers of @xmath53 and explicitly calculating the remaining integrals for the lowest , third , power of @xmath53 results in @xmath97+\mbox{o}(x^4 ) \label{eq47 } \\ & = & \frac{x^3}{6}\bigg[1-\frac{\sqrt{{\ee}_0}}{2}\left(1 + 3{\ee}_0 -2\,{\ee}_0 ^ 2 \right)+ % \frac{x^3}{3 } \frac{2\,r_0 ^ 2}{1-r_0 ^ 2 } - % x^3 \frac{12\,{\ee}_0}{{\ee}_0 + 1}\frac{r_0 ^ 2}{\left(1-r_0 ^ 2\right)^2}\bigg]\ !
+ \mbox{o}(x^4 ) .
\nonumber\end{aligned}\ ] ] the second and third integrals on the right - hand side of eq .
( [ eq43 ] ) are simply determined with the following result : @xmath98 + \mbox{o}(x^4)\ , .
\label{eq51}\end{aligned}\ ] ] substituting eqs .
( [ eq47 ] ) , ( [ eq49 ] ) and ( [ eq51 ] ) into @xmath99 , we arrive at @xmath100 \frac{x^3}{3 } + \mbox{o}(x^4 ) .
\label{eq52}\ ] ] then , by summing eqs .
( [ eq42 ] ) and ( [ eq52 ] ) , the result is obtained @xmath101 now we substitute eq .
( [ eq53 ] ) in eq .
( [ eq36 ] ) and perform integration . finally , from eq .
( [ eq35 ] ) the desired expression for the casimir pressure is derived @xmath102 .
\label{eq54}\ ] ] by comparison with eq .
( [ eq33 ] ) the explicit form of the coefficient @xmath75 is found as @xmath103 and , thus , both two first perturbation orders in the expansion for the free energy ( [ eq32 ] ) are determined .
equations ( [ eq32 ] ) , ( [ eq54 ] ) and ( [ eq55 ] ) solve the fundamental problem of the thermodynamic consistency of the lifshitz theory in the case of two dielectric plates . from eqs .
( [ eq32 ] ) and ( [ eq55 ] ) the entropy of the van der waals and casimir interaction between plates takes the form @xmath104 . \nonumber\end{aligned}\ ] ] as is seen from eq .
( [ eq56 ] ) , in the limit @xmath105 ( @xmath106 ) the lower order contributions to the entropy are of the second and the third powers in the small parameter @xmath49 .
thus , the entropy vanishes when the temperature goes to zero as it must be in accordance with the third law of thermodynamics ( the nernst heat theorem ) .
a similar behavior was obtained for ideal metals@xcite and for real metals described by the plasma model.@xcite for example , in the case of plates made of ideal metal the entropy at low temperatures is given by@xcite @xmath107 .
\label{eq57}\ ] ] note , however , that the expansion coefficients in eq .
( [ eq57 ] ) can not be obtained as a straightforward limit @xmath108 in eq .
( [ eq56 ] ) and the above equations for the free energy and pressure .
the mathematical reason is that it is impermissible to interchange the limiting transitions @xmath105 and @xmath108 in the power expansions of functions depending on @xmath69 as a parameter . remarkably , the low - temperature behavior of the free energy , pressure and entropy of nonpolar dielectrics in eqs .
( [ eq32 ] ) , ( [ eq54 ] ) and ( [ eq56 ] ) is universal , i.e. , is determined only by the static dielectric permittivity .
the absorption bands included in eq .
( [ eq9 ] ) do not influence the low - temperature behavior .
a more simple derivation of the results ( [ eq32 ] ) , ( [ eq54])([eq56 ] ) for dielectrics with constant @xmath92 is contained in ref . .
as was demonstrated above , all these results remain unchanged if the dependence of dielectric permittivity on frequency is taken into account . in ref .
more general results were obtained related to two dissimilar dielectric plates with dielectric permittivities @xmath109 and @xmath110 . for brevity here we present only the final expressions for the low - temperature behavior of the casimir free energy , pressure and entropy between dissimilar plates .
they are as follows:@xcite @xmath111 , \nonumber \\ & & p(a , t)=p_0(a)-\frac{\hbar c}{32\pi^2a^4}\left[c_4\tau^4 + \mbox{o}(\tau^5)\right ] , \phantom{\frac{(\ezf -1)(\ezs
-1)}{8\pi^2(\ezf + \ezs ) } } \label{eq58 } \\ & & s(a , t)=\frac{k_b}{2\pi a^2}\,\tau^2 % \,\times \nonumber \\ & & ~~ \times \left[\frac{3\zeta(3)}{32\pi^2}\ ; \frac{\ezf+\ezs+2\ezf\ezs}{(\ezf+1)(\ezs+1)}\ ; \frac{(\ezf-1)(\ezs-1)}{(\ezf+\ezs)}- c_4\,\tau+\mbox{o}(\tau^2)\right ] .
\nonumber\end{aligned}\ ] ] here @xmath112 and the coefficient @xmath75 is given by@xcite @xmath113\right\}\ , .
\nonumber\end{aligned}\ ] ] it is easily seen that in the limit @xmath114 equations ( [ eq58 ] ) , ( [ eq59 ] ) coincide with equations ( [ eq32 ] ) , ( [ eq54])([eq56 ] ) having obtained above .
note that in the application region of low - temperature asymptotic expressions the entropy of the casimir interaction between dielectric plates is nonnegative .
the obtained analytic behavior of the free energy , pressure and entropy at low temperatures can be compared with the results of numerical computations using the lifshitz formula .
dielectric properties of the plates can be described by the static dielectric permittivity or more precisely using the optical tabulated data for the complex index of refraction . as an example , in fig .
1 we present the thermal corrections to the casimir energy ( a ) and pressure ( b ) at a separation @xmath115 nm as functions of temperature in the configuration of two dissimilar plates made of high - resistivity si and sio@xmath116 .
the dielectric permittivities of both materials along the imaginary frequency axis were computed in ref . using the optical data of ref . .
the precise thermal corrections computed by taking into account these permittivities are shown by the solid lines and corrections computed by our analytical asymptotic expressions are shown by the long - dashed lines .
short - dashed lines indicate the results computed by the lifshitz formula with constant dielectric permittivities of si and sio@xmath116 equal to @xmath117 and @xmath118 , respectively . as is seen in fig .
1a , b , at @xmath119k the results obtained using the analytical asymptotic expressions practically coincide with the solid lines computed using the tabulated optical data for the materials of the plates .
now we return to the case of two similar dielectric plates and consider the asymptotic expressions under the condition @xmath48 , i.e. , at high temperatures ( large separations ) .
it is well known@xcite that in this case the approximation of static dielectric permittivity works good and the main contribution is given by the zero - frequency term of the lifshitz formula ( [ eq13 ] ) @xmath120 ( the other terms being exponentially small ) . performing the integration in eq .
( [ eq59a ] ) we obtain @xmath121 in a similar manner for the casimir pressure and entropy at @xmath48 it follows @xmath122 equations ( [ eq59b ] ) and ( [ eq59c ] ) are simply generalized@xcite for the case of two dissimilar dielectric plates by performing the replacement @xmath123 , where @xmath124 are defined by eq .
( [ eq11 ] ) with the static dielectric permittivities of dissimilar plates @xmath125 .
as was discussed in the previous section , the zero - frequency term in formula ( [ eq13 ] ) , i.e. , the contribution with @xmath126 , is of prime importance and determines many of the basic properties of the casimir interaction . in the above consideration we have described dielectric materials by eq .
( [ eq9 ] ) with finite static dielectric permittivity @xmath127 .
this resulted in eq .
( [ eq11 ] ) where one reflection coefficient at zero frequency is larger and the other one is smaller than the physical value for real photons at normal incidence .
however , in the lifshitz theory , the departure of both coefficients from their physical values is coordinated in such a way that the nernst heat theorem remains valid .
it is common knowledge that at nonzero temperatures dielectric materials possess a negligibly small but not equal to zero dc conductivity . from physical intuition
it is reasonable to expect that the influence of this conductivity on the van der waals and casimir forces should be also negligible . in ref .
it was shown that , on the contrary , the inclusion of small dielectric dc conductivity in the model of dielectric response leads to a large effect in dispersion forces .
this raises the question if dc conductivity is related to dispersion forces or if , on the contrary , the zero - frequency contribution should be understood not literally but as an analytic continuation from the region of high frequencies determining the physical phenomenon of dispersion forces . to illustrate this problem in more details
, we consider the asymptotic behavior of the free energy and entropy at low temperature with included dc conductivity .
what this means is that , instead of the dielectric permittivity @xmath128 given in eq .
( [ eq9 ] ) , one uses@xcite @xmath129 where @xmath130 is the dc conductivity of the plate material and @xmath131 .
the conductivity of dielectrics depends on temperature as @xmath132 where @xmath133 is determined by the energy gap @xmath134 which differs for different materials.@xcite the smallness of the dc conductivity of dielectrics can be illustrated@xcite by the example of sio@xmath116 where at @xmath135k it holds @xmath136 . thus , the role of the dc conductivity is really negligible for all @xmath137 .
in addition , @xmath138 quickly decreases with decrease of @xmath1 and , as a consequence , remains negligible at any @xmath1 . in spite of this , the substitution of eq .
( [ eq60 ] ) into the reflection coefficients ( [ eq14 ] ) leads to different result than in eq .
( [ eq11 ] ) : @xmath139 equation ( [ eq61 ] ) is in some analogy to eq .
( [ eq6 ] ) , obtained for metals described by the drude model , which leads to the violation of the nernst heat theorem in the case of perfect crystal lattices .
now we substitute the dielectric permittivity @xmath140 in eqs .
( [ eq13 ] ) ( [ eq14 ] ) instead of @xmath141 and find the casimir free energy @xmath142 with included dc conductivity . for convenience , we separate the zero - frequency term , subtract and add the usual zero - frequency contribution for dielectric without dc conductivity .
the result is @xmath143 \nonumber \\ & + & \frac{k_bt}{16\pi a^2 } \int_{0}^{\infty}ydy \ln\left(1-r_0 ^ 2e^{-y}\right ) \label{eq62}\\ & + & \frac{k_bt}{8\pi a^2}\sum\limits_{l=1}^{\infty } \int_{\zeta_l}^{\infty}ydy\left\ { \ln\big[1-\tilde{r}_{\|}^{2}(\zeta_l , y ) e^{-y}\big]+ \ln\big[1-\tilde{r}_{\bot}^{2}(\zeta_l , y)e^{-y}\big ] \right\ } , \nonumber\end{aligned}\ ] ] where @xmath58 was defined in eq .
( [ eq11 ] ) .
let us expand the last , third , integral on the right - hand side of eq .
( [ eq62 ] ) in powers of the small parameters @xmath144 . then , we combine the zero - order contribution in this expansion with the second integral on the right - hand side of eq .
( [ eq62 ] ) and obtain the casimir free energy @xmath145 calculated with the dielectric permittivities @xmath146 .
the first integral on the right - hand side of eq .
( [ eq62 ] ) is calculated explicitly .
then , eq . ( [ eq62 ] ) can be rewritten as @xmath147+r(a , t ) .
\label{eq63}\ ] ] here , @xmath148 is of order @xmath149 .
it represents the first and higher - order contributions in the expansion of the third integral on the right - hand side of eq .
( [ eq62 ] ) in powers of @xmath144 . restricting its explicit form to the first order contribution we get @xmath150\ , , \label{eq64 } \\%\end{equation } % \begin{align } & & r_1(a , t ) = \frac{k_bt}{4\pi a^2 } \sum\limits_{l=1}^{\infty } \frac{\beta}{l}\int_{\zeta_l}^{\infty}\!\ !
\frac{dy\,y^2e^{-y}}{\sqrt{y^2+\zeta_l^2({\ee}_l-1 ) } } \left\{\frac{(2-{\ee}_l)\,\zeta_l^2- 2y^2 } { \big[{\sqrt{y^2+\zeta_l^2({\ee}_l-1)}+\ee}_ly\,\big]^2}\,\ , \right . \nonumber \\ & & \phantom{aaaa}\left .
\times \frac{r_{\|}(\zeta_l , y)}{1- r_{\|}^2(\zeta_l , y)e^{-y } } -\frac{\zeta_l^2}{\big[\sqrt{y^2+\zeta_l^2({\ee}_l-1)}+y\,\big]^2}\,\ , \frac{r_{\bot}(\zeta_l , y)}{1-r_{\bot}^2(\zeta_l , y)e^{-y } } \right\}. \label{eq65}\end{aligned}\ ] ] calculating the entropy by the first equality in eq .
( [ eq56 ] ) , we arrive at @xmath151- \frac{\partial r(a , t)}{\partial t } , \label{eq66}\ ] ] where @xmath152 is the entropy for the plates with the dielectric permittivity @xmath141 given by eq .
( [ eq56 ] ) .
let us now show that the quantity @xmath148 exponentially goes to zero with the decrease of @xmath1 .
first we consider only the integral in eq .
( [ eq65 ] ) , expand the integrated function in powers of @xmath49 ( we recall that @xmath153 ) , restrict ourselves to the main contribution , resulting for @xmath154 , and rearrange it appropriately : @xmath155 substituting this in eq .
( [ eq65 ] ) , we find @xmath156 % \label{eq69 } \nonumber \\ & & \phantom{aaaaaaa } = \frac{k_b\,\mbox{li}_2\left(r_0 ^ 2\right)}{2\pi a^2\left({\ee}_0 ^ 2 -1\right)}\ ; t\beta\,\ln\tau+t\beta\,\mbox{o}(\tau^0)\ , .
\label{eq71}\end{aligned}\ ] ] here , the last line is obtained by using the equality @xmath157 substituting only its leading term and observing the definition of the integral logarithm , @xmath158 .
taking into account that @xmath159 , we get the conclusion that the temperature dependence of @xmath160 is given by @xmath161 thus , both @xmath160 and its derivative with respect to @xmath1 in eqs .
( [ eq63 ] ) and ( [ eq66 ] ) go to zero .
the terms of the second and higher powers in @xmath162 in eq .
( [ eq64 ] ) go to zero even faster than @xmath163 when @xmath106 .
finally , in the limit @xmath106 from eq .
( [ eq66 ] ) it follows @xmath164>0 . \label{eq73}\ ] ] the right - hand side of this equation depends on the parameter of the system under consideration ( the separation distance @xmath5 ) and implies a violation of the nernst heat theorem .
an analogous result was obtained@xcite in the case of two dissimilar dielectrics .
the violation of the nernst heat theorem in the casimir interaction for dielectrics originates from the inclusion of the dc conductivity in the model of dielectric response .
this violation is , however , of different nature than the one discussed above in the case of drude metals . in the case of dielectrics
the entropy at zero temperature is positive but in the case of drude metals it is negative . in the case of metals
the violation is caused by the vanishing contribution from the transverse electric mode at zero frequency whereas the other reflection coefficient takes the physical value 1 [ see eq .
( [ eq6 ] ) ] . for dielectrics
the situation is quite opposite . in this case
the transverse reflection coefficient at zero frequency is always equal to zero [ compare eqs .
( [ eq11 ] ) and ( [ eq61 ] ) in the absence and in the presence of the contribution from dc conductivity ] . here
the violation occurs due to the unity value of the parallel reflection coefficient at zero frequency in eqs .
( [ eq61 ] ) which departs from the value @xmath165 coordinated with the zero value of the transverse coefficient in eq .
( [ eq11 ] ) .
one can conclude that the dc conductivity of a dielectric is not related to the nature of the van der waals and casimir forces and must not be included in the model of dielectric response . ignoring this rule results in a violation of thermodynamics .
physically it is amply clear that there is no fluctuating field of zero frequency and that for such high - frequency phenomena as the van der waals and casimir forces the low - frequency behavior should be obtained by analytic continuation from the region of high frequencies .
this permits to conclude that the correct procedure consists in the substitution of the finite static dielectric permittivities into the zero - frequency term of the lifshitz formula , as lifshitz and his collaborators really did.@xcite
the casimir interaction between metal and dielectric plates suggests the interesting opportunity to verify the thermodynamic consistency of the lifshitz theory with different models of the dielectric response .
this configuration was first investigated in ref . where it was proved that the casimir entropy is in accordance with the demands of the nernst heat theorem if the static permittivity of the dielectric plate is finite . in ref .
, however , only the first leading terms in the low - temperature asymptotic expressions for the free energy and entropy were obtained and the casimir pressure was derived only in the dilute approximation . here
we derive the more precise low - temperature behavior for the casimir free energy , pressure and entropy in the configuration of one plate made of ideal metal and another plate made of dielectric with any finite static dielectric permittivity . for the configuration of metal and dielectric plates the lifshitz formula takes the form@xcite analogical to eq .
( [ eq13 ] ) @xmath166 + \ln\big[1-r_{\bot}^{m}(\zeta_l , y)\,r_{\bot}^{d}(\zeta_l , y)\,e^{-y}\big ] \right\}. \nonumber\end{aligned}\ ] ] here the reflection coefficients @xmath167 for metal and dielectric , respectively , are given by eqs .
( [ eq14 ] ) where @xmath141 should be changed for @xmath168 . for an ideal metal @xmath169 and eq .
( [ eq74 ] ) takes the more simple form @xmath170 + \ \ln\left[1-r_{\bot}(\zeta_l , y ) e^{-y}\right]\right\ } \nonumber\end{aligned}\ ] ] ( here and below we omit the index @xmath171 near the reflection coefficient and permittivity of a dielectric plate ) .
we admit that the dielectric permittivity calculated at matsubara frequencies @xmath172 , i.e. , is equal to its static value and find the asymptotic behavior of eq .
( [ eq75 ] ) at small @xmath49 .
[ in analogy with sec . 2 it is possible to prove that the deviations of @xmath173 from @xmath127 at high frequencies do not influence the low - temperature behavior of the casimir free energy , pressure and entropy .
it can be shown also that the results of this section are valid not only for ideal metal plate but for plate made of real metal as well . ]
the free energy ( [ eq75 ] ) can be represented by eqs .
( [ eq20 ] ) ( [ eq19 ] ) , with the function @xmath174 replaced by @xmath175 + y\ln\left[1-r_{\bot}(\zeta , y)e^{-y}\right ] \nonumber \\ & & \phantom{aaaaa } \equiv { \hat f}_{\|}(\zeta , y)+{\hat f}_{\bot}(\zeta , y ) . \label{eq76}\end{aligned}\ ] ] in the case of one dielectric and one metal plate both @xmath176 and @xmath177 contribute to @xmath55 .
the expansion of @xmath178 in powers of @xmath53 takes the form @xmath179 now we integrate eq .
( [ eq77 ] ) in accordance with eq .
( [ eq20 ] ) to find the function @xmath54 .
the integral of the first term on the right - hand side of eq .
( [ eq77 ] ) is evaluated using the new variable @xmath180 : @xmath181 where the coefficient near the first - order contribution in @xmath53 vanishes . as a result
, this term could contribute to @xmath55 only starting from the third expansion order .
the integrals of the second - order terms on the right - hand side of eq .
( [ eq77 ] ) are simply calculated using the formulas @xmath182 finally , we obtain @xmath183 where the unknown third order expansion coefficient is designated as @xmath184 .
substituting eq .
( [ eq80 ] ) in eq .
( [ eq19 ] ) and using eq .
( [ eq16 ] ) , we find the free energy in the system metal - dielectric in the form @xmath185 , \label{eq81}\ ] ] where @xmath186 . the casimir pressure in the configuration of metal and dielectric plates obtained from eq .
( [ eq81 ] ) is equal to @xmath187 .
\label{eq82}\ ] ] the direct application of the lifshitz formula gives the expression for the pressure analogical to eq .
( [ eq34 ] ) , @xmath188\!. \label{eq83 } \end{aligned}\ ] ] using the abel - plana formula ( [ eq15 ] ) , eq .
( [ eq83 ] ) can be represented in the form of eqs .
( [ eq35 ] ) , ( [ eq36 ] ) where @xmath189 again , we deal first with @xmath82 . by adding and subtracting the asymptotic behavior of the integrated function at small @xmath53 , @xmath190 and introducing the new variable @xmath83 , the function @xmath82 can be identically rearranged and expanded in powers of @xmath53 as follows : @xmath191 \nonumber\\ % \label{eq86 } % \end{equation } % \noindent % expanding in eq.~(\ref{eq86 } )
we obtain % \begin{equation } % \phi_{\bot}(x ) & & \phantom{aaaaa}=\frac{1}{4}(\ez-1)x^2(1-x)+x^3 \int_{1}^{\infty}dv\left [ \frac{v^2r_{\bot}(v)}{1-r_{\bot}(v)}- \frac{\ez-1}{4}\right]+\mbox{o}(x^4 ) .
\label{eq87}\end{aligned}\ ] ] the integral on the right - hand side of eq .
( [ eq87 ] ) is converging and can be simply calculated with the result @xmath192 to deal with @xmath90 we add and subtract in eq .
( [ eq84 ] ) the two first expansion terms of the integrated function in powers of @xmath53 , @xmath193 \label{eq89 } \\ & & \phantom{aaaa } + \int_{x}^{\infty}y^2dy \bigg[\frac{r_{\|}(x , y)}{e^y - r_{\|}(x , y)}- \frac{r_0}{e^y - r_0}+ \frac{{\ee}_0r_0e^{-y}x^2}{y^2(\ez+1)\left(1-r_0e^{-y}\right)^2 } \bigg ] . \nonumber\end{aligned}\ ] ] the asymptotic expansion of the first integral on the right - hand side of eq .
( [ eq89 ] ) is given by @xmath194 and of the second one by @xmath195x^3+\mbox{o}(x^4 ) .
\label{eq91}\ ] ] by summing eqs .
( [ eq90 ] ) and ( [ eq91 ] ) we find @xmath196\,x^3+\mbox{o}(x^4 ) .
\nonumber\end{aligned}\ ] ] finally we add eq .
( [ eq92 ] ) to eq .
( [ eq88 ] ) and arrive at @xmath197 substituting this in eq .
( [ eq36 ] ) and using eq .
( [ eq35 ] ) , the asymptotic expression for the casimir pressure is obtained @xmath198 .
\label{eq94}\ ] ] comparing eqs .
( [ eq94 ] ) and ( [ eq82 ] ) , the explicit expression for the coefficient @xmath199 reads @xmath200 now we are in a position to find the asymptotic behavior of the entropy in the limit of low temperatures in the configuration of two parallel plates one of which is metallic and the other one dielectric . by calculating the negative derivative of eq .
( [ eq81 ] ) with respect to temperature , one arrives at @xmath201 .
\nonumber\end{aligned}\ ] ] this equation is in analogy to eq .
( [ eq56 ] ) obtained for the case of two dielectrics .
as is seen from eq .
( [ eq96 ] ) , the entropy of the casimir and van der waals interactions between metal and dielectric plates vanishes when the temperature goes to zero , i.e. , the nernst heat theorem is satisfied .
note that the first term of order @xmath202 on the right - hand side of eq .
( [ eq96 ] ) was already obtained in ref . .
it is notable also that at low temperatures the entropy goes to zero remaining positive . at the same time , as was shown in ref .
, at larger temperatures entropy is nonmonotonous and may take negative values .
this interesting property distinguishes the configuration of metal and dielectric plates from two dielectric plates . in the latter configuration
the negative entropy appears only for nonphysical dielectrics with anomalously large and frequency independent dielectric permittivities.@xcite now let us consider the case @xmath48 , i.e. , high temperatures ( large separations ) . in the same way as for two dielectric plates , here the main contribution to the free energy is given by the term with @xmath126 in eq .
( [ eq75 ] ) , @xmath203 performing the integration in eq .
( [ eq96a ] ) , we obtain @xmath204 from this equation , for the casimir pressure and entropy at @xmath48 it follows @xmath205 the results ( [ eq96b ] ) and ( [ eq96c ] ) are analogous to ( [ eq59b ] ) and ( [ eq59c ] ) for two dielectric plates .
in the previous section it was supposed that the static dielectric permittivity @xmath127 of the dielectric plate is finite .
now we will deal with the configuration of metal and dielectric plates with included dc conductivity of the dielectric material . in doing so
the permittivity of the dielectric plate is given by @xmath206 where all notations were introduced in and below eq .
( [ eq60 ] ) .
thus , the reflection coefficients at zero frequency satisfy eq .
( [ eq4 ] ) for a plate made of ideal metal and eq .
( [ eq61 ] ) for a plate made of dielectric with included dc conductivity .
let us find the low - temperature behavior of the casimir entropy and verify the consistency of the lifshitz theory with thermodynamics in this nonstandard situation .
for this purpose we substitute the dielectric permittivity ( [ eq97 ] ) in eq .
( [ eq75 ] ) instead of @xmath127 and find the casimir energy @xmath142 with included dc conductivity of a dielectric plate . in the same way as in sec .
3 , it is convenient to separate the zero - frequency term of @xmath142 and subtract and add the usual zero - frequency contribution for metal - dielectric plates computed with the dielectric permittivity @xmath127 , @xmath207 \nonumber\\ & + & \frac{k_bt}{16\pi a^2 } \int_{0}^{\infty}ydy \ln\left(1-\rz e^{-y}\right ) \label{eq98}\\ & + & \frac{k_bt}{8\pi a^2}\sum\limits_{l=1}^{\infty } \int_{\zeta_l}^{\infty}ydy\left\ { \ln\left[1-\tilde{r}_{\|}(\zeta_l , y ) e^{-y}\right]+ \ln\left[1-\tilde{r}_{\bot}(\zeta_l , y ) e^{-y}\right ] \right\}. \nonumber\end{aligned}\ ] ] here the reflection coefficients @xmath208 are calculated with the permittivity ( [ eq97 ] ) . we expand the third integral on the right - hand side of eq .
( [ eq98 ] ) in powers of the small parameter @xmath144 . the zero - order contribution in this expansion together with the second integral of eq .
( [ eq98 ] ) form the casimir free energy @xmath145 calculated with dielectric permittivity @xmath127 .
the first integral on the right - hand side of eq .
( [ eq98 ] ) is calculated explicitly . as a result ,
( [ eq98 ] ) is rearranged to @xmath209+q(a , t ) , \label{eq99}\ ] ] where @xmath210 contains the first and higher - order contributions in the expansion of the third integral on the right - hand side of eq .
( [ eq98 ] ) in powers of @xmath144 .
the explicit form of the main first - order term in @xmath210 is the following : @xmath211 ^ 2}\,\ , \right .
\nonumber \\ & & \phantom{aaa}\left .
\times\frac{1}{1-r_{\|}(\zeta_l , y)e^{-y } } -\frac{\zeta_l^2}{\big[\sqrt{y^2+\zeta_l^2({\ee}_0 - 1)}+y\,\big]^2}\,\ , \frac{1}{1-r_{\bot}(\zeta_l , y)e^{-y } } \right\}. \label{eq100}\end{aligned}\ ] ] in the same way as in sec . 3 , we expand the integrated function in eq .
( [ eq100 ] ) in powers of @xmath49 ( bearing in mind that @xmath153 ) and preserve only the main contribution at @xmath154 : @xmath212 . \nonumber \end{aligned}\ ] ] dealing with this expression in the same way as with eq .
( [ eq71 ] ) , we arrive at @xmath213 the casimir entropy in the configuration of metal and dielectric plates with included dc conductivity of the dielectric plate is obtained as minus derivative of eq .
( [ eq99 ] ) with respect to temperature , @xmath214- \frac{\partial q(a , t)}{\partial t}. \label{eq103}\ ] ] using eq .
( [ eq102 ] ) , the calculation of the limiting value at @xmath106 is straightforward : @xmath215>0 .
\label{eq104}\ ] ] from this equation it follows that the inclusion of the dc conductivity of dielectric plate in the configuration metal - dielectric results in a violation of the nernst heat theorem . in the above this result was obtained for a metallic plate made of ideal metal .
it can be shown that it remains valid for a metallic plate made of real metal with finite conductivity . thus , both configurations ( two dielectric plates or one metal plate and one dielectric ) lead to the same conclusion that when the dc conductivity is included in the model of dielectric response of the dielectric plate , the lifshitz theory loses its consistency with thermodynamics .
this confirms the conclusion made in sec . 3 that the actual properties of dielectric materials at very low frequencies are in fact not related to the van der waals and casimir forces .
as was mentioned in the introduction , semiconductors possess a wide variety of electric and optical properties ranging from metallic to dielectric .
this opens the possibility to modulate the van der waals and casimir forces by changing the charge carrier density .
bearing in mind the discussed above problems on the consistency of the lifshitz theory with thermodynamics , semiconductors can provide us with a test for the validity of different approaches .
thus , if for good dielectric the dc conductivity does not play any real role in the van der waals and casimir forces , the question arises on how much it should be increased in order to become a relevant factor in the description of dispersion forces . in ref .
the casimir force acting between two si plates was calculated using the simple analytic expression for si dielectric permittivity as a function of frequency .
the complete tabulated optical data of si were used in ref . to calculate the van der waals interaction of different atoms with a si wall . the first attempt to measure the van der waals force between a glass lens and a si plate and to modify it by light due to the change of carrier density
was undertaken in ref . .
however , glass is a dielectric and therefore the electric forces due to localized point charges could not be controlled .
this might explain that no force change occured on illumination at small separations where the effect should be most pronounced .
the first measurements of the casimir force between a gold coated sphere and a single crystal si plate were performed in refs . by means of the atomic force microscope .
the experiments used a @xmath216-type b doped si plate of resistivity @xmath217 cm .
the chosen resistivity of the plate is in some sense intermediate between the resistivity of metals ( which is usually two or three orders of magnitude lower ) and the resistivity of dielectrics ( it can be by about a factor of @xmath218 larger ; for instance , high - resistivity `` dielectric '' si has the resistivity @xmath219 cm ) .
thus , the used si plate had a relatively large absorption typical for semiconductors but it was also enough conductive to avoid the accumulation of charges.@xcite in fig . 2 , taken from ref . ,
the differences of the theoretical and mean experimental casimir forces acting between au sphere and si plate are presented as functions of separation . in fig .
2a the theoretical force @xmath220 is computed using the lifshitz formula and the dielectric permittivity of a si plate with the relatively low resistivity @xmath221 used in experiment .
this dielectric permittivity goes to infinity as @xmath222 with decreasing frequency ( see the solid line in fig . 3 ) . in fig .
2b the theoretical casimir force @xmath223 is computed using the dielectric permittivity of the si plate made of `` dielectric '' si with high resistivity @xmath224 .
the dielectric permittivity of high - resistivity si is shown by the dashed line in fig .
it is characterized by the finite static value @xmath225 .
the solid and dashed lines in fig .
2a , b indicate the 95% and 70% confidence intervals , respectively . as is seen from fig .
2 , the theoretical approach using the dielectric permittivity of high - resistivity `` dielectric '' si is excluded by experiment within the separation range from 60 to 110 nm at 70% confidence . at the same time , the theory using the dielectric permittivity of si with a low resistivity @xmath221 is consistent with experiment .
the above results suggest an approach on how to correctly determine the possible role of the low - frequency conductivity properties in dispersion forces . as is seen from fig . 3 , the dielectric permittivity of low - resistivity si ( solid line ) significantly departs from the dielectric permittivity of `` dielectric '' si in the region around the important dimensional parameter of the problem , the characteristic frequency @xmath226rad / s . that is why the low - resistivity sample can not be described at low frequencies by the static dielectric permittivity of si equal to @xmath225 . to describe it , the term @xmath227 , like in eq .
( [ eq97 ] ) , should be added to @xmath228 .
note that in this case @xmath229 at the first matsubara frequencies with @xmath230 and , thus , this quantity can not be considered as a small parameter . at the same time , for a high - resistivity sample the inclusion of the dc conductivity would lead to deviations from the dashed line in fig .
3 only at frequencies @xmath231rad / s , which are much less than the characteristic frequency .
this comparison permits to make a conclusion in what experimental situations the conductivity properties of semiconductors at low frequencies should be taken into account and when they should be omitted as being not related to dispersion forces .
the future experiments on the modification of semiconductor charge carrier density by laser light@xcite will bring a more clear understanding of this problem on the connection between the low - frequency material properties and dispersion forces .
the presented above new analytic results on the low - temperature behavior of the casimir free energy , pressure and entropy between dielectrics or between dielectric and metal are based on the conventional lifshitz theory which describes dielectric materials by means of the frequency dependent dielectric permittivity .
in fact , the assumption that the material of the plates possesses only the frequency dispersion means that the components of electric displacement are connected with the components of electric field by the relation @xmath232 this equation is central in all different derivations of the lifshitz formula ( see , e.g. , refs . , , ) .
the effects of spatial nonlocality ( spatial dispersion ) are in fact essential only at shortest separations between the plates comparable with atomic dimensions and also for metals at sufficiently large separations ( typically of about 23@xmath233 m ) in the frequency region of the anomalous skin effect .
the casimir force in the latter region was described by the lifshitz theory reformulated in terms of the leontovich impedance.@xcite recently the spatial dispersion came to the attention in connection with the thermal casimir force.@xcite in particular , it was claimed@xcite that for real metals at any separation the account of spatial dispersion leads to practically the same result ( [ eq6 ] ) for the reflection coefficients at zero frequency as was obtained earlier using the drude model dielectric function ( [ eq5 ] ) .
this conclusion , if it is correct , not only returns us to the contradiction with experiment@xcite but also casts doubts on all results obtained by means of the conventional lifshitz theory accounting for only the frequency dispersion .
it is natural when the spatial dispersion contributes a small fraction of a percent as it is generally believed in numerous applications of the lifshitz theory .
it is , however , quite another matter when the account of the spatial dispersion results in some `` dramatic effects'',@xcite i.e. , in several hundred times larger thermal correction than is obtained in the local case .
below we demonstrate that the conclusions of refs .
are in fact not reliable because they use the lifshitz theory of dispersion forces outside of its application range.@xcite to find the electromagnetic modes associated with an empty gap between the plates , refs .
use the standard continuity boundary conditions , @xmath234 which are commonly applied in the derivation of the lifshitz formula for spatially local nonmagnetic materials . here
@xmath235 is the magnetic induction , @xmath236 is the normal to the boundary directed inside the medium , the subscripts @xmath237 refer to the normal and tangential components , respectively , the subscript 1 refers to the vacuum and subscript 2 to the plate material . in refs .
the spatial dispersion is described by the longitudinal and transverse dielectric permittivities depending on the wave vector and frequency : @xmath238 . however , as is shown below , in the theory of the casimir effect both the boundary conditions ( [ eq106 ] ) and permittivities @xmath238 are inapplicable .
we start from the boundary conditions and recall the set of maxwell equations in a metal describing the casimir effect , @xmath239 equations ( [ eq107 ] ) do not contain any external , i.e. , independent on @xmath240 , @xmath241 and @xmath242 , current or charge densities .
the definition of the electric displacement is @xmath243 where the volume current @xmath244 is induced by @xmath240 and @xmath242 and takes into account the conduction electrons . in electrodynamics with spatial dispersion
the electric field and magnetic induction are finite at the boundary surfaces , whereas the electric displacement can tend to infinity.@xcite then , integrating eqs .
( [ eq107 ] ) over the thickness of the boundary layer as is done in ref . , we reproduce the first two conditions in eq .
( [ eq106 ] ) and arrive at the modified third and fourth conditions,@xcite @xmath245=\frac{4\pi}{c}\mbox{\boldmath{$j$ } } , \label{eq109}\ ] ] where the induced surface charge and current densities are given by @xmath246dl , \quad \mbox{\boldmath{$j$}}=\frac{1}{4\pi}\int_{1}^{2 } \frac{\partial \mbox{\boldmath{$d$}}}{\partial t}dl . \label{eq110}\ ] ] note that the boundary conditions ( [ eq109 ] ) , ( [ eq110 ] ) are obtained from the macroscopic maxwell equations for physical fields .
they should not be mixed with the boundary conditions arising in perturbative theories and for the fictitious fields ( see below ) . in linear electrodynamics for a medium with time - independent properties without spatial dispersion the material equation connecting the electric displacement and electric field takes the form @xmath247 according to this equation , the electric displacement at a point @xmath248 and moment @xmath249 is determined by the electric field at the same point @xmath248 at different moments @xmath250 ( the spatial dispersion is absent but the temporal may be present ) .
it is easily seen that the substitution of eq .
( [ eq111 ] ) in eq .
( [ eq110 ] ) leads to @xmath251 , @xmath252@xmath253 and , as a result , the boundary conditions ( [ eq109 ] ) coincide with the standard continuity conditions ( [ eq106 ] ) .
it is unjustified , however , to use conditions ( [ eq106 ] ) in the presence of spatial dispersion . in refs .
a few examples are presented illustrating that in this case neither @xmath254 nor @xmath252 is equal to zero .
we now turn to a discussion of the use of dielectric permittivity @xmath255 in the theory of the casimir effect with account of spatial dispersion . in the presence of only frequency dispersion , it is possible to perform the fourier transformation of the fields @xmath256 in eq .
( [ eq111 ] ) and arrive at eq .
( [ eq105 ] ) where @xmath257 is the frequency - dependent dielectric permittivity and @xmath258 .
in fact eqs .
( [ eq105 ] ) , ( [ eq112 ] ) and ( [ eq113 ] ) are used in parallel with the boundary conditions ( [ eq106 ] ) in all derivations of the lifshitz formula.@xcite if the material of the plates is characterized not only by temporal but also spatial dispersion , eq .
( [ eq111 ] ) is generalized to @xmath259 if the material medium were uniform in space ( i.e. , all points were equivalent ) , the kernel @xmath260 would not depend on @xmath248 and @xmath248@xmath261 separately , as in eq .
( [ eq114 ] ) , but on the difference @xmath262 . in this case , by performing the fourier transformation , @xmath263 and substituting it in eq .
( [ eq114 ] ) , one could introduce the dielectric permittivities @xmath264 as refs .
do , and rearrange eq .
( [ eq114 ] ) to the form @xmath265 in the casimir effect , however , the material medium is not uniform due to the presence of a macroscopic gap between the two plates ( half spaces ) . because of this , the assumption that the kernel @xmath260 depends only on @xmath266 and @xmath49 is wrong . as a result ,
it is not possible to introduce the dielectric permittivity @xmath267 depending on the wave vector and frequency .
in fact , for systems with spatial dispersion in the presence of boundaries the kernel @xmath260 depends not only on @xmath266 and @xmath49 but also on the distance from the boundary.@xcite in this complicated situation the following approximate phenomenological approach is sometimes applicable.@xcite for electromagnetic waves with a wavelength @xmath268 the kernel @xmath269 in eq .
( [ eq114 ] ) differs essentially from zero only in a certain vicinity of the point @xmath248 with characteristic dimensions @xmath270 ( for nonmetallic condensed media @xmath271 is of the order of the lattice constant ) .
then it is reasonable to assume that @xmath260 is a function of @xmath266@xmath272@xmath248@xmath248@xmath261 , except for a layer of thickness @xmath271 adjacent to the boundary surface .
if one is mostly interested in bulk phenomena and neglects the role and influence of a subsurface layer , the quantity @xmath267 may be employed as a reasonable approximation .
this approximate phenomenological approach is widely applied in the theoretical investigation of the anomalous skin effect ( see , e.g. , ref . ) .
note that in ref .
some kind of fictitious infinite system was introduced and electromagnetic fields in this system are discontinuous on the boundary surface .
this discontinuity should not be confused with the discontinuity of physical fields of a real system in the presence of spatial dispersion described by eqs .
( [ eq109 ] ) and ( [ eq110 ] ) .
( there is also another approach to the description of the anomalous skin effect in polycrystals using the generalizations of the local leontovich impedance which takes into account the shape of fermi surface@xcite ) .
the frequency and wave vector dependent dielectric permittivity in the presence of boundaries is also approximately applied in the theory of radiative heat transfer@xcite or in the study of electromagnetic interaction of molecules with metal surfaces.@xcite in all these applications the boundary effects are usually taken into account by the boundary conditions ( [ eq109 ] ) supplemented by so called `` additional boundary conditions '' .
it is unlikely , however , that the approximate phenomenological approach using such quantity as @xmath267 in the presence of boundaries would be applicable in the theory of the casimir force where the boundary effects on the zero - point electromagnetic oscillations are of prime importance .
it is notable also that this approach faces serious theoretical difficulties including the violation of the law of conservation of energy.@xcite it is then not surprising that the application of this approach in refs .
results in the contribution to the casimir free energy from the transverse electric mode which is in contradiction with experiment.@xcite one more shortcoming of refs .
is that they substitute the dielectric permittivity @xmath267 , depending on both wave vector and frequency , into the conventional lifshitz formula derived in the presence of only temporal dispersion . in the famous review paper@xcite
it has been noticed , however , that with the inclusion of spatial dispersion the free energy of a fluctuating field takes the form @xmath273 , where @xmath274 is given by the conventional lifshitz expression derived in a spatially local case and written in terms of the fresnel reflection coefficients , and @xmath275 is an additional term which can be expressed in terms of the thermal green s function of the electromagnetic field and polarization operator .
review@xcite calls as not reliable the results of , e.g. , ref .
obtained by the substitution of dielectric permittivity @xmath267 , taking account of spatial dispersion , into the conventional lifshitz formula .
it can be true that the lifshitz formula written in terms of general reflection coefficients is applicable in both spatially local and nonlocal situations .
however , as far as the exact reflection coefficients in a spatially nonlocal case are unknown , the use of some approximate phenomenological models , elaborated in literature for applications different than the casimir effect , may lead to incorrect results for @xmath275 and create inconsistencies with experiment . to conclude this section , the results of refs .
on the influence of spatial nonlocality on the casimir interaction are shown to be not reliable .
although at present there is no fundamental theory of the thermal casimir force incorporating spatial dispersion , there is no reason to expect that it can play any significant role in the frequency region of infrared optics ( experimental separations ) or normal skin effect ( i.e. , at separations between plates greater than 45@xmath233 m ) .
in the foregoing we have presented the derivation of analytic asymptotic expressions for the free energy , pressure and entropy of the casimir interaction between two dielectric plates and between metal and dielectric plates at low and high temperatures .
it was shown that the low - temperature behavior of the casimir interaction between dielectrics and between dielectric and metal is determined by the static dielectric permittivities of nonpolar dielectrics .
the obtained results were shown to be in agreement with thermodynamics when the static dielectric permittivities of dielectrics are finite .
in particular , the entropy of the casimir interaction goes to zero when the temperature vanishes , i.e. , the nernst heat theorem is satisfied .
this demonstrates the consistency of the original lifshitz s approach to the van der waals forces between dielectrics which disregards the small conductivity of dielectrics at constant current .
the second important result shown above is that the inclusion of the dc conductivity of dielectrics into the model of dielectric response leads not to some small corrections to the characteristics of the casimir interaction , as one could expect , but makes the lifshitz theory inconsistent with thermodynamics leading to the violation of the nernst heat theorem .
this reveals that real material properties at very low , quasistatic frequencies are in fact not related to the phenomenon of the van der waals and casimir forces which is actually determined by sufficiently high frequencies . in this case the zero - frequency contribution to the casimir force should be understood not literally but as analytic continuation to zero of the material physical behavior in the region around the characteristic frequency .
the presented results provide a basis for the calculation of the van der waals and casimir forces between real materials .
such calculations are much needed for numerous applications of the casimir force discussed in the introduction , in particular for the applications in nanotechnology and for constraining predictions of fundamental physical theories beyond the standard model .
bearing in mind that semiconductors are the main constituent materials in nanotechnological devices , it is a subject of high priority to understand the casimir effect with semiconductor boundaries . in this connection
we have discussed new experimental results and theoretical ideas on the casimir interaction between a metal sphere and semiconductor plate .
it was stressed that by changing the charge carrier density in the semiconductor it is possible to bring it in different intermediate states between metallic and dielectric . in this case
the problem arises when the conductivity properties of semiconductor are not related to dispersion forces and when they are becoming relevant .
a criterion for the resolution of this problem was formulated based on the relation between the typical frequency at which the dc conductivity properties come into play and the characteristic frequency of the casimir effect .
in fact the thermal casimir interaction between semiconductors remains an open question and much work should be done both in experiment and theory to gain a better insight into this subject .
the last major problem discussed in this review is whether or not the spatial dispersion influences essentially the thermal casimir force between real materials . in the present literature
there is no agreement on this subject .
we adduced arguments in favor of the statement that in the region of experimental separations the influence of the spatial dispersion on the casimir force is negligible small .
the statements on the opposite , contained in the literature , were shown to be not reliable because they are obtained by the application of the lifshitz theory outside of its application range . at the same time
it was ascertained that at the moment there is no consistent fundamental theory of the van der waals and casimir forces taking the spatial dispersion into account .
this is the problem to solve in the foreseeable future .
g.l.k . and v.m.m .
are grateful to the center of theoretical studies and the institute for theoretical physics , leipzig university for their kind hospitality .
this work was supported by deutsche forschungsgemeinschaft grant 436rus113/789/0 - 2 .
99 h. b. g. casimir , _ proc .
_ * 51 * , 793 ( 1948 ) .
w. milonni , _ the quantum vacuum _ ( academic press , san diego , 1994 ) .
v. m. mostepanenko and n. n. trunov , _ the casimir effect and its applications _ ( clarendon press , oxford , 1997 ) .
k. a. milton , _ the casimir effect _ ( world scientific , singapore , 2001 )
. m. kardar and r. golestanian , _ rev .
phys . _ * 71 * , 1233 ( 1999 ) .
m. bordag , u. mohideen and v. m. mostepanenko , _ phys . rep . _
* 353 * , 1 ( 2001 ) .
s. k. lamoreaux , _ rep .
phys . _ * 68 * , 201 ( 2005 ) .
p. candelas and s. weinberg , _ nucl .
phys . _ * b237 * , 397 ( 1984 ) .
m. bordag , b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
_ * d58 * , 075003 ( 1998 ) ; * d60 * , 055004 ( 1999 ) ; * d62 * , 011701(r ) ( 2000 ) .
j. c. long , h. w. chan and j. c. price , _ nucl .
phys . _ * b539 * , 23 ( 1999 ) .
e. fischbach , d. e. krause , v. m. mostepanenko and m. novello , _ phys .
_ * d64 * , 075010 ( 2001 ) .
v. m. mostepanenko , _ int . j. mod
. phys . _ * a17 * , 722 ( 2002 ) ; * a17 * , 4143 ( 2002 ) . g. l. klimchitskaya and u. mohideen , _ int
phys . _ * a17 * , 4143 ( 2002 ) .
j. f. babb , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * a70 * , 042901 ( 2004 ) .
m. antezza , l. p. pitaevskii and s. stringari , _ phys .
_ * a70 * , 053619 ( 2004 ) .
m. krech , _ the casimir effect in critical systems _ ( world scientific , singapore , 1994 )
. h. b. chan , v. a. aksyuk , r. n. kleiman , d.j .
bishop and f. capasso , _ science _ * 291 * , 1941 ( 2001 ) ; _ phys .
lett . _ * 87 * , 211801 ( 2001 ) .
e. v. blagov , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
* b71 * , 235401 ( 2005 ) .
e. elizalde , _ ten physical applications of spectral zeta functions _
( springer , berlin , 1995 ) . s. k. lamoreaux , _ phys .
lett . _ * 78 * , 5 ( 1997 ) .
u. mohideen and a. roy , _ phys .
_ * 81 * , 4549 ( 1998 ) .
g. l. klimchitskaya , a. roy , u. mohideen and v. m. mostepanenko , _ phys .
rev . _ * a60 * , 3487 ( 1999 ) .
a. roy and u. mohideen , _ phys .
* 82 * , 4380 ( 1999 ) .
a. roy , c .- y .
lin and u. mohideen , _ phys .
_ * d60 * , 111101(r ) ( 1999 ) .
b. w. harris , f. chen and u. mohideen , _ phys .
_ * a62 * , 052109 ( 2000 ) .
t. ederth , _ phys .
rev . _ * a62 * , 062104 ( 2000 ) .
g. bressi , g. carugno , r. onofrio and g. ruoso , _ phys .
_ * 88 * , 041804 ( 2002 )
. f. chen , u. mohideen , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
_ * 88 * , 101801 ( 2002 ) .
f. chen , u. mohideen , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * a66 * , 032113 ( 2002 ) .
r. s. decca , e. fischbach , g. l. klimchitskaya , d. e. krause , d. lpez and v. m. mostepanenko , _ phys .
rev . _ * d68 * , 116003 ( 2003 ) .
f. chen , g. l. klimchitskaya , u. mohideen and v. m. mostepanenko , _ phys .
rev . _ * a69 * , 022117 ( 2004 ) .
r. s. decca , d. lpez , e. fischbach , g. l. klimchitskaya , d. e. krause and v. m. mostepanenko , _ ann . phys . _
( n.y . ) * 318 * , 37 ( 2005 ) .
g. l. klimchitskaya , r. s. decca , d. lpez , e. fischbach , d. e. krause and v. m. mostepanenko , _ int .
phys . _ * a20 * , 2205 ( 2005 ) .
y. srivastava , a. widom and m. h. friedman , _ phys .
_ * 55 * , 2246 ( 1985 ) .
y. srivastava and a. widom , _ phys .
rep . _ * 148 * , 1 ( 1987 ) .
e. m. lifshitz , _ sov .
* 2 * , 73 ( 1956 ) .
i. e. dzyaloshinskii , e. m. lifshitz and l. p. pitaevskii , _ sov .
usp . _ * 4 * , 153 ( 1961 ) .
e. m. lifshitz and l. p. pitaevskii , _ statistical physics _ ,
part ii ( pergamon , oxford , 1980 ) .
j. schwinger , l. l. deraad and k. a. milton , _ ann .
( n.y . ) * 115 * , 1 ( 1978 ) .
l. s. brown and g. j. maclay , _ phys .
rev . _ * 184 * , 1272 ( 1969 ) .
m. bostrm and b. e. sernelius , _ phys .
_ * 84 * , 4757 ( 2000 ) .
j. feinberg , a. mann and m. revzen , _ ann .
( n.y . ) * 288 * , 103 ( 2001 ) .
a. scardiccio and r. l. jaffe , _ nucl .
phys . _ * b743 * , 249 ( 2006 ) . c. genet , a. lambrecht and s. reynaud , _ phys .
rev . _ * a62 * , 012110 ( 2000 ) .
m. bordag , b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ phys . rev
. lett . _ * 85 * , 503 ( 2000 ) .
j. s. hye , i. brevik , j. b. aarseth and k. a. milton , _ phys .
rev . _ * e67 * , 056116 ( 2003 ) .
i. brevik , j. b. aarseth , j. s. hye and k. a. milton , _ phys .
_ * e71 * , 056101 ( 2005 ) .
v. b. bezerra , g. l. klimchitskaya and c. romero , _ phys .
rev . _ * a65 * , 012111 ( 2002 ) .
b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
_ * a67 * , 062102 ( 2003 ) .
j. r. torgerson and s. k. lamoreaux , _ phys .
* e70 * , 047102 ( 2004 ) .
s. k. lamoreaux and w. t. buttler , _ phys .
rev . _ * e71 * , 036109 ( 2005 ) .
v. b. bezerra , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * a66 * , 022112 ( 2002 ) . v. b. bezerra , g. l. klimchitskaya , v. m. mostepanenko and c. romero , _ phys .
* a69 * , 022119 ( 2004 ) .
v. b. bezerra , r. s. decca , e. fischbach , b. geyer , g. l. klimchitskaya , d. e. krause , d. lpez , v. m. mostepanenko and c. romero , _ phys .
rev . _ * e73 * , 028101 ( 2006 ) .
i. brevik , s. a. ellingsen and k. a. milton , _ new j. phys . _
* 8 * , 236 ( 2006 ) . v. m. mostepanenko , v. b. bezerra , r. s. decca , e. fischbach , b. geyer , g. l. klimchitskaya , d. e. krause , d. lpez and c. romero , _ j. phys .
_ * a39 * , 6589 ( 2006 ) .
j. r. zurita - snchez , j .- j .
greffet and l. novotny , _ phys .
rev . _ * a69 * , 022902 ( 2004 ) . k. joulain , j .-
mulet , f. marquier , r. carminati and j .- j .
greffet , _ surf .
rep . _ * 57 * , 59 ( 2005 ) .
b. s. stipe , h. j. mamin , t. d. stowe , t. w. kenny and d. rugar , _ phys .
lett . _ * 87 * , 096801 ( 2001 ) .
b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * d72 * , 085009 ( 2005 ) .
b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * a72 * , 022111 ( 2005 ) .
f. chen , u. mohideen , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
_ * a72 * , 020101(r ) ( 2005 ) ; * a73 * , 019905(e ) ( 2006 ) .
f. chen , u. mohideen , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
_ * a74 * , 022103 ( 2006 ) . b. e. sernelius , _ phys . rev . _ * b71 * , 235114 ( 2005 ) .
v. b. svetovoy and r. esquivel , _ phys .
rev . _ * e72 * , 036113 ( 2005 ) .
r. esquivel and v. b. svetovoy , _ phys .
* a69 * , 062102 ( 2004 ) .
r. esquivel , c. villarreal and w. l. mochn , _ phys .
_ * a68 * , 052103 ( 2003 ) .
a. m. contreras - reyes and w. l. mochn , _ phys .
rev . _ * a72 * , 034102 ( 2005 ) .
j. mahanty and b. w. ninham , _ dispersion forces _
( academic , new york , 1976 ) .
v. a. parsegian and b. w. ninham , _ nature _ * 224 * , 1197 ( 1969 ) .
l. bergstrm , _ adv .
interface sci .
_ * 70 * , 125 ( 1997 ) .
h. mitter and d. robaschik , _ eur .
phys . j. _
* b13 * , 335 ( 2000 ) .
b. w. ninham and j. daicic , _ phys .
rev . _ * a57 * , 1870 ( 1998 ) .
g. l. klimchitskaya , b. geyer and v. m. mostepanenko , _ j. phys . _ * a39 * , 6495 ( 2006 ) .
a. o. caride , g. l. klimchitskaya , v. m. mostepanenko and s. i. zanette , _ phys .
rev . _ * a71 * , 042901 ( 2005 ) . , ed .
e. d. palik ( academic , new york , 1985 ) .
b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ int .
phys . _ * a16 * , 3291 ( 2001 ) . j. c. slater , _ insulators , semiconductors and metals . quantum theory of molecules and solids _ , vol . 3 , ( mcgraw - hill , new york , 1967 ) . ,
j. f. shackelford and w. alexander ( crc , boca raton , 2001 ) .
b. geyer , g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * a65 * , 062109 ( 2002 ) .
n. inui , _ j. phys
jpn . _ * 72 * , 2198 ( 2003 ) ; * 73 * , 2198 ( 2004 ) ; * 75 * , 024004 ( 2006 ) . w. arnold , s. hunklinger and k. dransfeld , _ phys .
rev . _ * b19 * , 6049 ( 1979 ) .
f. chen and u. mohideen , _ j. phys .
_ * a39 * , 6233 ( 2006 ) . c. genet , a. lambrecht and s. reynaud , _ phys .
rev . _ * a67 * , 043811 ( 2003 ) .
e. i. kats , _ sov .
. jetp _ * 46 * , 109 ( 1977 ) .
g. l. klimchitskaya and v. m. mostepanenko , _ phys .
rev . _ * b75 * , 036101 ( 2007 ) . v.
m. agranovich and v. l. ginzburg , _ crystal optics with spatial dispersion , and excitons _
( springer , berlin , 1984 ) .
j. a. stratton , _ electromagnetic theory _
( mcgraw - hill , new york , 1941 ) . v. l. ginzburg , _ physics and astrophysics _ ( pergamon press , oxford , 1985 ) . v. m. agranovich and v. l. ginzburg , _ sov .
. jetp _ * 36 * , 440 ( 1973 ) .
k. l. kliewer and r. fuchs , _ phys .
rev . _ * 172 * , 607 ( 1968 ) .
i. m. kaganova and m. i. kaganov , _ phys .
rev . _ * b63 * , 054202 ( 2001 ) .
s. m. rytov , yu . a. kravtsov and v. i. tatarskii , _ principles of statistical radiophysics _ , vol . 3 ( springer , berlin , 1989 ) .
g. w. ford and w. h. weber , _ phys .
rep . _ * 113 * , 195 ( 1984 ) .
j. t. foley and a. j. devaney , _ phys .
rev . _ * b12 * , 3104 ( 1975 ) . yu .
s. barash and v. l. ginzburg , _ sov .
usp . _ * 18 * , 305 ( 1975 ) .
g. g. kleinman and u. landman , _ phys .
lett . _ * 33 * , 524 ( 1974 ) . | we review recent results obtained in the physics of the thermal casimir force acting between two dielectrics , dielectric and metal , and between metal and semiconductor .
the detailed derivation for the low - temperature behavior of the casimir free energy , pressure and entropy in the configuration of two real dielectric plates is presented . for dielectrics with finite static dielectric permittivity
it is shown that the nernst heat theorem is satisfied .
hence , the lifshitz theory of the van der waals and casimir forces is demonstrated to be consistent with thermodynamics .
the nonzero dc conductivity of dielectric plates is proved to lead to a violation of the nernst heat theorem and , thus , is not related to the phenomenon of dispersion forces .
the low - temperature asymptotics of the casimir free energy , pressure and entropy are derived also in the configuration of one metal and one dielectric plate .
the results are shown to be consistent with thermodynamics if the dielectric plate possesses a finite static dielectric permittivity .
if the dc conductivity of a dielectric plate is taken into account this results in the violation of the nernst heat theorem .
we discuss both the experimental and theoretical results related to the casimir interaction between metal and semiconductor with different charge carrier density .
discussions in the literature on the possible influence of spatial dispersion on the thermal casimir force are analyzed . in conclusion ,
the conventional lifshitz theory taking into account only the frequency dispersion remains the reliable foundation for the interpretation of all present experiments . |
we recall some notions and definitions .
a function @xmath9 is said to be completely monotonic on an interval @xmath10 if @xmath9 has derivatives of all orders on @xmath11 and @xmath12 for all @xmath13 and @xmath14 .
if @xmath15 for some nonnegative integer @xmath16 is completely monotonic on an interval @xmath10 , but @xmath17 is not completely monotonic on @xmath11 , then @xmath18 is called a completely monotonic function of @xmath16-th order on an interval @xmath11 .
a function @xmath9 is said to be logarithmically completely monotonic on an interval @xmath10 if its logarithm @xmath19 satisfies @xmath20^{(k)}\ge0\ ] ] for all @xmath13 and @xmath21 .
a function @xmath22 is called a bernstein function on @xmath11 if @xmath18 has derivatives of all orders and @xmath23 is completely monotonic on @xmath11 .
[ stieltjes - dfn ] a stieltjes function is a function @xmath24 which can be written in the form @xmath25 where @xmath26 are nonnegative constants and @xmath27 is a nonnegative measure on @xmath28 such that @xmath29 .
[ x - degree - dfn ] let @xmath30 be a nonnegative function and have derivatives of all orders on @xmath28 .
a number @xmath31 is said to be the completely monotonic degree of @xmath30 with respect to @xmath32 if @xmath33 is a completely monotonic function on @xmath28 but @xmath34 is not for any positive number @xmath35 . in what follows , for convenience
, we denote the sets of completely monotonic functions on @xmath10 , logarithmically completely monotonic functions on @xmath10 , stieltjes functions , and bernstein functions on @xmath10 by @xmath36 $ ] , @xmath37 $ ] , @xmath38 , and @xmath39 $ ] respectively .
now we briefly describe some basic relationships between the above defined classes of functions and list a characterization of bernstein functions on @xmath28 . in @xcite , any logarithmically completely monotonic function on an interval @xmath11
was once again proved to be completely monotonic on @xmath11 . in @xcite ,
the set of all stieltjes functions was proved to be a subset of all logarithmically completely monotonic functions on @xmath28 .
see also ( * ? ? ?
* remark 4.8 ) .
conclusively , @xmath40\subset\mathcal{c}[(0,\infty)].\ ] ] it is obvious that any nonnegative completely monotonic function of first order is a bernstein function .
the relation between bernstein functions and logarithmically completely monotonic functions was discovered in ( * ? ? ?
* , theorem 3 ) and ( * ? ? ?
* , proposition 5.17 ) , which reads that the reciprocal of any positive bernstein function is logarithmically completely monotonic . in other words , @xmath41\longrightarrow \frac1f\in\mathcal{l}[i].\ ] ] a relation between @xmath38 and @xmath42 $ ] was given by ( * ? ? ?
* theorem 5.4 ) which may be recited as @xmath43.\ ] ] it is easy to see that the degree of any completely monotonic function on @xmath28 is at least zero .
conversely , if a nonnegative function @xmath30 on @xmath28 has a nonnegative degree @xmath44 , then it must be a completely monotonic function on @xmath28 .
see @xcite .
bernstein functions can be characterized by ( * ? ? ?
* , theorem 3.2 ) which states that a function @xmath45 is a bernstein function if and only if it admits the representation @xmath46 where @xmath47 and @xmath27 is a measure on @xmath28 satisfying @xmath48 for information on characterizations of the classes @xmath49 $ ] and @xmath50 $ ] , please refer to related texts in @xcite and references cited therein .
we recall from @xcite that the extended mean value @xmath51 may be defined by @xmath52^{{1/(s - r ) } } , & rs(r - s)(x - y)&\ne 0 ; \\
e(r,0;x , y)&=\biggl[\frac{y^r - x^r}{r(\ln y-\ln x)}\biggr]^{{1/r } } , & r(x - y)&\ne 0 ; \\
e(r , r;x , y)&=\frac1{e^{1/r}}\biggl(\frac{x^{x^r}}{y^{y^r}}\biggr)^ { { 1/(x^r - y^r ) } } , & r(x - y)&\ne 0 ; \\
e(0,0;x , y)&=\sqrt{xy}\ , , & x&\ne y ; \\
e(r , s;x , x)&=x , & x&=y;\notag\end{aligned}\ ] ] where @xmath53 are positive numbers and @xmath54 . because this mean was first defined in @xcite , so it is also called stolarsky s mean by a number of mathematicians .
many special means with two positive variables are special cases of @xmath55 , for example , @xmath56 for more information on @xmath55 , please refer to the monograph @xcite , the papers @xcite , and a lot of closely - related references therein .
it is easy to see that the arithmetic mean @xmath57 is a trivial bernstein function of @xmath6 for @xmath58 . in (
* , remark 6 ) , it was pointed out that , 1 . by standard arguments ,
it is easy to verify that the reciprocal of the exponential mean @xmath59^ { { 1/(x - y)}}\ ] ] for @xmath58 with @xmath60 is a logarithmically completely monotonic function of @xmath6 ; 2 . from the newly - discovered integral representation
@xmath61 it is easy to obtain that the exponential mean @xmath62 for @xmath63 with @xmath60 is also a completely monotonic function of first order ( that is , a bernstein function ) . in (
* , remark 3.7 ) , the logarithmic mean @xmath64 was proved to be increasing and concave in @xmath63 for @xmath58 with @xmath60 . more strongly ,
the logarithmic mean @xmath65 was proved in ( * ? ? ? * theorem 1 ) to be a completely monotonic function of first order on @xmath66 for @xmath58 with @xmath60 . therefore , the logarithmic mean @xmath65 is a bernstein function of @xmath6 . by
* , theorem 3 ) or ( * ? ? ?
* , proposition 5.17 ) , the logarithmically complete monotonicity of the exponential mean @xmath62 and the logarithmic mean @xmath65 can be deduced respectively from their common property that they are bernstein functions .
the goals of this paper are to prove that the harmonic mean @xmath67 and the geometric mean @xmath68 are all bernstein functions of @xmath69 on @xmath66 for @xmath58 with @xmath60 , and to establish integral representations of @xmath7 and @xmath8 .
in order to prove our main results , the following lemmas are needed .
[ geom - mean - cm - lem - h(t ) ] for @xmath70 , the @xmath71-th derivatives of the functions @xmath72 the reciprocal @xmath73 , and @xmath74 on @xmath28 may be computed by @xmath75^{(i)}&=\frac{(-1)^{i+1}}{2^it^i(1+t)^ih(t)}\sum_{k=0}^{i-1}b_{i , k}t^k , \label{h(t)-recip - i - d}\\ h^{(i)}(t)&=\frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \sum_{k=0}^{i-1}c_{i , k}t^k , \label{h(t)-expression}\end{aligned}\ ] ] where @xmath76 consequently , the functions @xmath77 and @xmath78 are completely monotonic on @xmath28 , and the reciprocal @xmath73 is a bernstein function on @xmath28 . a direct calculation yields @xmath79 , which means that @xmath80 so , the formulas and are valid for @xmath81 and @xmath82 .
differentiating on both sides of gives @xmath83 ' & = \biggl[\frac{(-1)^i}{2^it^{i+1}(1+t)^{i-1}h(t)}\sum_{k=0}^{i-1}a_{i , k}t^k\biggr]'\\ & = \frac{(-1)^{i+1}}{2^{i+1}t^{i+2}(1+t)^{i}h(t)}\sum_{k=0}^{i-1}[1 + 2(i - k)+2(2i - k)t]a_{i , k}t^k\\ & = \frac{(-1)^{i+1}}{2^{i+1}t^{i+2}(1+t)^{i}h(t)}\sum_{k=0}^{i}a_{i+1,k}t^k.\end{aligned}\ ] ] because @xmath84a_{i , k}t^k = \sum_{k=0}^{i-1}[1 + 2(i - k)]a_{i , k}t^k + \sum_{k=0}^{i-1}2(2i - k)a_{i , k}t^{k+1}\\ = \sum_{k=0}^{i-1}[1 + 2(i - k)]a_{i , k}t^k + \sum_{k=1}^{i}2(2i - k+1)a_{i , k-1}t^{k}\\ = ( 1 + 2i)a_{i,0}+\sum_{k=1}^{i-1}\{[1 + 2(i - k)]a_{i , k}+2(2i - k+1)a_{i , k-1}\}t^k + 2(i+1)a_{i , i-1}t^{i},\end{gathered}\ ] ] we obtain @xmath85 and , for @xmath86 , @xmath87a_{i , k}+2(2i - k+1)a_{i , k-1}.\ ] ] combining with and results in @xmath88 and @xmath89 taking @xmath90 in and using give @xmath91 from , it is easily deduced that @xmath92 . substituting this into and recurring repeatedly lead to @xmath93 taking @xmath94 in and using show @xmath95 from , it is readily deduced that @xmath96 . substituting this into and recurring repeatedly reveal @xmath97 taking @xmath98 in and using show @xmath99 from , it is immediately obtained that @xmath100 . substituting this into and recurring repeatedly
yield @xmath101 by the same arguments as above , we may obtain @xmath102 and @xmath103 inductively , we can derive that @xmath104 for @xmath86 .
specially , we have @xmath105 replacing @xmath16 by @xmath106 in yields @xmath107 for @xmath108 .
substituting into leads to @xmath109\lambda_{i,\ell}+\ell(2i-\ell+1)\lambda_{i,\ell-1}=i(i+1)\lambda_{i+1,\ell}\ ] ] for @xmath108 .
the equality is equivalent to @xmath110 for @xmath86 .
the quantities in implies that @xmath111 , that is , @xmath112 is independent of @xmath71 .
then the equality may be written as @xmath113\mu_{k+1}=k(1+k)\mu_{k+1}\ ] ] for @xmath86 .
recurring by @xmath114 reveals @xmath115 for @xmath86 . as a result , by
, we conclude that @xmath116 for @xmath86 .
replacing @xmath16 by @xmath106 in shows @xmath117 for @xmath108 .
it is easy to verify that the sequence for @xmath118 meets the recursion formulas , , and .
the formulas and for general terms are thus proved .
it is obvious that @xmath119 which is equivalent to @xmath120 therefore , using the formulas and just verified , we have @xmath121^{(i)}&=-2\bigl[t^2h'(t)\bigr]^{(i)}\\ & = -2\sum_{\ell=0}^i\binom{i}{\ell}\bigl(t^2\bigr)^{(\ell)}h^{(i-\ell+1)}(t)\\ & = -2\biggl[\binom{i}{0}t^2h^{(i+1)}(t)+2\binom{i}{1}th^{(i)}(t)+2\binom{i}{2}h^{(i-1)}(t)\biggr]\\ & = -2\biggl[\frac{(-1)^{i+1}}{2^{i+1}t^{i}(1+t)^{i}h(t)}\sum_{k=0}^{i}a_{i+1,k}t^k + \frac{(-1)^ii}{2^{i-1}t^{i}(1+t)^{i-1}h(t)}\sum_{k=0}^{i-1}a_{i , k}t^k \\ & \quad+\frac{(-1)^{i-1}(i-1)i}{2^{i-1}t^{i}(1+t)^{i-2}h(t)}\sum_{k=0}^{i-2}a_{i-1,k}t^k\biggr]\\ & = \frac{(-1)^{i+1}}{2^it^i(1+t)^ih(t)}\biggl[4i(1+t ) \sum_{k=0}^{i-1}a_{i , k}t^k-\sum_{k=0}^{i}a_{i+1,k}t^k\\ & \quad-4(i-1)i(1+t)^2\sum_{k=0}^{i-2}a_{i-1,k}t^k\biggr]\\ & = \frac{(-1)^{i+1}}{2^it^i(1+t)^ih(t)}\biggl\{4ia_{i,0}-4i(i-1)a_{i-1,0}-a_{i+1,0}\\ & \quad+[4i(a_{i,1}+a_{i,0})-4i(i-1)(a_{i-1,1}+2a_{i-1,0})-a_{i+1,1}]t \\ & \quad+[4i(a_{i , i-1}+a_{i , i-2})-4i(i-1)(a_{i-1,i-3}+2a_{i-1,i-2})-a_{i+1,i-1}]t^{i-1}\\ & \quad+[4ia_{i , i-1}-4i(i-1)a_{i-1,i-2}-a_{i+1,i}]t^i + \sum_{k=2}^{i-2}[4i(a_{i , k}+a_{i , k-1 } ) \\ & \quad-4i(i-1)(a_{i-1,k}+2a_{i-1,k-1}+a_{i-1,k-2})-a_{i+1,k}]t^k\biggr\}\\ & = \frac{(-1)^{i+1}}{2^it^i(1+t)^ih(t)}\sum_{k=0}^{i-1}\frac{(i-1)!i!(2i-2k-3)!!}{(i - k-1)!(i - k)!k!}2^kt^k.\end{aligned}\ ] ] hence , the general formulas and are obtained . adding the two formulas and yields @xmath122^{(i)}=\frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \biggl[(1+t)\sum_{k=0}^{i-1}a_{i , k}t^k -t\sum_{k=0}^{i-1}b_{i , k}t^k\biggr]\\ \begin{aligned } & = \frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \biggl[\sum_{k=0}^{i-1}(a_{i , k}-b_{i , k})t^{k+1}+\sum_{k=0}^{i-1}a_{i , k}t^{k}\biggr]\\ & = \frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \biggl[\sum_{k=1}^{i}(a_{i , k-1}-b_{i , k-1})t^k+\sum_{k=0}^{i-1}a_{i , k}t^{k}\biggr ] \end{aligned}\\ \begin{aligned } & = \frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \biggl[a_{i,0}+\sum_{k=1}^{i-1}(a_{i , k-1}-b_{i , k-1}+a_{i , k})t^k+(a_{i , i-1}-b_{i , i-1})t^i\biggr]\\ & = \frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \biggl\{(2i-1 ) ! !
+ \sum_{k=1}^{i-1}\frac{(i-1)!(i+1)!(2i-2k-1)!!}{(i - k-1)!(i - k+1)!k!}2^{k}t^k\biggr\}\\ & = \frac{(-1)^i}{2^it^{i+1}(1+t)^{i}h(t ) } \sum_{k=0}^{i-1}\frac{(i-1)!(i+1)!(2i-2k-1)!!}{(i - k-1)!(i - k+1)!k!}2^{k}t^k . \end{aligned}\end{gathered}\ ] ] this implies that the function @xmath78 is completely monotonic on @xmath28 .
the proof of lemma [ geom - mean - cm - lem - h(t ) ] is completed . in (
* , remark 2.4 ) , it was collected as an example that the function @xmath123 is a stieltjes function for @xmath124 .
the property ( iv ) in section 3 of @xcite ( see also the property ( vii ) in ( * ? ? ? * theorem 1.3 ) ) reads that if @xmath125 then @xmath126 for @xmath127 .
specially for @xmath128 and @xmath129 , we have @xmath130 .
the property ( i ) in section 3 of @xcite ( see also the property ( i ) in ( * ? ? ? * theorem 1.3 ) ) states that if @xmath131 then @xmath132 .
applying this property to @xmath133 brings out @xmath134 which means , by the relation from the very ends of the inclusions , that @xmath135 $ ] and , by the relation , that @xmath136 $ ] . in ( * ? ? ?
* , remark 3.11 ) , it was listed as examples that @xmath137 $ ] for @xmath138 and @xmath139 $ ] .
the item ( iii ) of corollary 3.7 in @xcite write that if @xmath140 $ ] then @xmath141 $ ] . applying @xmath142 and @xmath143 respectively to @xmath144 and @xmath145
reveals once again that @xmath146 $ ] .
taking @xmath147 and @xmath148 .
it is easy to see that @xmath149 $ ] and @xmath150 .
a part of theorem 3.6 in @xcite asserts that if @xmath151 $ ] then @xmath152 $ ] for every @xmath153 $ ] . since @xmath154 $ ]
, applying @xmath9 and @xmath155 in this assertion respectively to @xmath156 and @xmath145 leads to @xmath157 $ ] .
the proof of lemma [ geom - mean - cm - lem - h(t ) ] is completed .
[ lem2-zhang - li - qi - eq ] for @xmath158 $ ] , the complex functions @xmath159 and @xmath160 have integral representations @xmath161 and @xmath162 consequently , the functions @xmath77 and @xmath163 are stieltjes functions and the complex function @xmath164 has the integral integral representation @xmath165 for @xmath158 $ ] , where @xmath166e^{-us}\operatorname{d\mspace{-1mu}}u = \int_0^{1/2 } q\biggl(\frac12-u\biggr)\bigl(e^{us}-e^{-us}\bigr)e^{-s/2}\operatorname{d\mspace{-1mu}}u\ ] ] is nonnegative on @xmath28 and @xmath167 on @xmath168 . by standard arguments , we immediately obtain that @xmath169=\lim_{z\to0}\sqrt{z^2+z}\,=\sqrt{\lim_{z\to0}(z^2+z)}\,=0,\\ \lim_{z\to0}\frac{z}{h(z)}=\lim_{z\to0}\sqrt{\frac{z^3}{1+z}}\ , = \sqrt{\lim_{z\to0}\frac{z^3}{1+z}}\,=0 , \label{frac - z = h(z)=0}\\ \lim_{z\to\infty}\sqrt{1+\frac1z}\,=\sqrt{1+\lim_{z\to\infty}\frac1z}\,=1 , \label{sqrt-1=frac - z}\\ \lim_{z\to\infty}\frac1{\sqrt{1+\frac1z}}\,=\frac1{\sqrt{1+\lim_{z\to\infty}\frac1z}}\,=1 , \label{sqrt-1=frac - z - recip}\\ h(\overline{z})=\overline{h(z ) } , \label{h(z)-h(z)-conjugate}\\ \frac1{h(\overline{z})}=\overline{\biggl[\frac1{h(z)}\biggr]}. \label{h(z)-frac - conjugate}\end{gathered}\ ] ] for @xmath170 and @xmath35 , we have @xmath171\\ = \exp\biggl\{\frac12\biggl[\ln\biggl|\frac{t^2+\varepsilon^2-t}{t^2+\varepsilon^2 } -i\frac{\varepsilon}{t^2+\varepsilon^2}\biggr| + i\arg\biggl(\frac{t^2+\varepsilon^2-t}{t^2+\varepsilon^2}-i\frac{\varepsilon } { t^2+\varepsilon^2}\biggr)\biggr]\biggr\}\\ = \exp\biggl\{\frac12\biggl[\ln p(t,\varepsilon ) + i\arg\biggl(\frac{t^2+\varepsilon^2-t}{t^2+\varepsilon^2 } -i\frac{\varepsilon}{t^2+\varepsilon^2}\biggr)\biggr]\biggr\}\\ = \begin{cases } \exp\biggl\{\dfrac12\biggl[\ln p(t,\varepsilon)\ , + i\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}\biggr]\biggr\ } , & t^2+\varepsilon^2-t>0,\\ \exp\biggl\{\dfrac12\biggl[\ln p(t,\varepsilon)\ , + i\biggl(\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}-\pi\biggr)\biggr]\biggr\ } , & t^2+\varepsilon^2-t<0,\\ \exp\biggl\{\dfrac12\biggl(\ln\dfrac{\varepsilon}{t^2+\varepsilon^2}\ , -i\dfrac\pi2\biggr)\biggr\ } , & t^2+\varepsilon^2-t=0 , \end{cases}\end{gathered}\ ] ] where @xmath172 hence , @xmath173 \sin\biggl(\dfrac12\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}\biggr ) , & t^2+\varepsilon^2-t>0;\\ \exp\biggl[\dfrac12\ln p(t,\varepsilon)\,\biggr ] \sin\biggl(\dfrac12\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}-\dfrac\pi2\biggr ) , & t^2+\varepsilon^2-t<0;\\ -\exp\biggl(\dfrac12\ln\dfrac{\varepsilon}{t^2+\varepsilon^2}\,\biggr)\sin\dfrac\pi4 , & t^2+\varepsilon^2-t=0 .
\end{cases}\ ] ] accordingly , @xmath174 similarly , for @xmath170 and @xmath35 , we have @xmath175\\ & = \begin{cases } \exp\biggl\{-\dfrac12\biggl[\ln p(t,\varepsilon)\ , + i\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}\biggr]\biggr\ } , & t^2+\varepsilon^2-t>0;\\ \exp\biggl\{-\dfrac12\biggl[\ln p(t,\varepsilon)\ , + i\biggl(\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}-\pi\biggr)\biggr]\biggr\ } , & t^2+\varepsilon^2-t<0;\\ \exp\biggl\{-\dfrac12\biggl(\ln\dfrac{\varepsilon}{t^2+\varepsilon^2}\ , -i\dfrac\pi2\biggr)\biggr\ } , & t^2+\varepsilon^2-t=0 . \end{cases}\end{aligned}\ ] ] therefore , @xmath176=\\ \begin{cases } -\exp\biggl[-\dfrac12\ln p(t,\varepsilon)\,\biggr ] \sin\biggl(\dfrac12\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}\biggr ) , & t^2+\varepsilon^2-t>0;\\ -\exp\biggl[-\dfrac12\ln p(t,\varepsilon)\,\biggr ] \sin\biggl(\dfrac12\arctan\dfrac{\varepsilon}{t^2+\varepsilon^2}-\dfrac\pi2\biggr ) , & t^2+\varepsilon^2-t<0;\\ \exp\biggl(-\dfrac12\ln\dfrac{\varepsilon}{t^2+\varepsilon^2}\,\biggr)\sin\dfrac\pi4 , & t^2+\varepsilon^2-t=0 .
\end{cases}\end{gathered}\ ] ] consequently , @xmath177= \begin{cases } \sqrt{\dfrac{t}{1-t } } , & 0<t<1;\\ \infty , & t=1;\\ 0 , & t>1 .
\end{cases}\ ] ] let @xmath178 be a bounded domain with piecewise smooth boundary .
the famous cauchy integral formula ( see @xcite ) reads that if @xmath179 is analytic on @xmath178 , and @xmath179 extends smoothly to the boundary of @xmath178 , then @xmath180 for any fixed point @xmath158 $ ] , choose @xmath181 and @xmath182 such that @xmath183 , and consider the positively oriented contour @xmath184 in @xmath185 $ ] consisting of the half circle @xmath186 for @xmath187 $ ] and the half lines @xmath188 for @xmath189 until they cut the circle @xmath190 , which close the contour at the points @xmath191 , where @xmath192 as @xmath193 .
see figure [ note-on-li-chen-conj.eps ] . ,
title="fig:",scaledwidth=75.0% ] + by the above mentioned cauchy integral formula , we have @xmath194}^{\arg[-r(\varepsilon)+i\varepsilon]}\frac{ir e^{i\theta}h\bigl(re^{i\theta}\bigr)}{re^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta\biggr].\notag\end{aligned}\ ] ] by the limit , it follows that @xmath195 in virtue of the limit , it can be derived that @xmath196}^{\arg[-r(\varepsilon)+i\varepsilon]}\frac{ir e^{i\theta}h\bigl(re^{i\theta}\bigr)}{re^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta = \lim_{r\to\infty}\int_{-\pi}^{\pi}\frac{ir e^{i\theta}h\bigl(re^{i\theta}\bigr)}{re^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta=2\pi i.\ ] ] making use of the limits and yields that @xmath197\operatorname{d\mspace{-1mu}}x\\ \begin{aligned } & = \int_{-r(\varepsilon)}^0\frac{(x - i\varepsilon - z)h(x+i\varepsilon ) -(x+i\varepsilon - z)h(x - i\varepsilon ) } { ( x+i\varepsilon - z)(x - i\varepsilon - z)}\operatorname{d\mspace{-1mu}}x\\ & = \int_{-r(\varepsilon)}^0\frac{(x - z)[h(x+i\varepsilon)-h(x - i\varepsilon ) ] -i\varepsilon[h(x - i\varepsilon)+h(x+i\varepsilon ) ] } { ( x+i\varepsilon - z)(x - i\varepsilon - z)}\operatorname{d\mspace{-1mu}}x\\ & = 2i\int_{-r(\varepsilon)}^0\frac{(x - z)\im h(x+i\varepsilon ) -\varepsilon\re h(x+i\varepsilon ) } { ( x+i\varepsilon - z)(x - i\varepsilon - z)}\operatorname{d\mspace{-1mu}}x\\ & \to2i\int_{-r}^0\frac{\lim_{\varepsilon\to0^+}\im h(x+i\varepsilon)}{x - z}\operatorname{d\mspace{-1mu}}x\\ & = -2i\int^r_0\frac{\lim_{\varepsilon\to0^+}\im h(-t+i\varepsilon)}{t+z}\operatorname{d\mspace{-1mu}}t\\ & \to-2i\int^\infty_0\frac{\lim_{\varepsilon\to0^+}\im h(-t+i\varepsilon)}{t+z}\operatorname{d\mspace{-1mu}}t\\ & = 2i\int_0 ^ 1\sqrt{\frac1t-1}\,\frac{\operatorname{d\mspace{-1mu}}t}{t+z } \label{level0lines } \end{aligned}\end{gathered}\ ] ] as @xmath198 and @xmath199 . substituting equations , , and into and simplifying produce the integral representation . similarly , by the above mentioned cauchy integral formula , we have @xmath200}{\varepsilon e^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta + \int_{-r(\varepsilon)}^0 \frac{1/h(x+i\varepsilon)}{x+i\varepsilon
- z}\operatorname{d\mspace{-1mu}}x \label{h(z)-frac - cauchy - apply}\\ & \quad+\int_0^{-r(\varepsilon)}\frac{1/h(x - i\varepsilon)}{x - i\varepsilon - z}\operatorname{d\mspace{-1mu}}x + \int_{\arg[-r(\varepsilon)-i\varepsilon]}^{\arg[-r(\varepsilon)+i\varepsilon]}\frac{ir e^{i\theta}\bigl[1/h\bigl(re^{i\theta}\bigr)\bigr]}{re^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta\biggr].\notag\end{aligned}\ ] ] from the limit , it follows that @xmath201}{\varepsilon e^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta=0.\ ] ] by virtue of the limit , it may be deduced that @xmath202}^{\arg[-r(\varepsilon)+i\varepsilon]}\frac{ir e^{i\theta}\bigl[1/h\bigl(re^{i\theta}\bigr)\bigr]}{re^{i\theta}-z}\operatorname{d\mspace{-1mu}}\theta=2\pi i.\ ] ] employing the limits and yields that @xmath203 -\varepsilon\re [ 1/h(x+i\varepsilon ) ] } { ( x+i\varepsilon - z)(x - i\varepsilon - z)}\operatorname{d\mspace{-1mu}}x\notag\\ & \to2i\int_{-r}^0\frac{\lim_{\varepsilon\to0^+}\im [ 1/h(x+i\varepsilon)]}{x - z}\operatorname{d\mspace{-1mu}}x \quad \text{as $ \varepsilon\to0^+$}\notag\\ & \to-2i\int^\infty_0\frac{\lim_{\varepsilon\to0^+}\im h(-t+i\varepsilon)}{t+z}\operatorname{d\mspace{-1mu}}t \quad \text{as $ r\to\infty$}\notag\\ & = -2i\int_0 ^ 1\sqrt{\frac{t}{1-t}}\,\frac{\operatorname{d\mspace{-1mu}}t}{t+z}. \label{level0lines = frac}\end{aligned}\ ] ] substituting equations , , and into and simplifying produce the integral representation . adding and leads to @xmath204e^{-zs}\operatorname{d\mspace{-1mu}}s.\end{aligned}\ ] ] utilizing @xmath205 for @xmath206 or @xmath207 for @xmath208 results in @xmath209\operatorname{d\mspace{-1mu}}u = \int_0^{1/2 } q(u)\bigl[1-e^{-(1 - 2u)s}\bigr]e^{-us}\operatorname{d\mspace{-1mu}}u \ge0\end{gathered}\ ] ] or @xmath210\operatorname{d\mspace{-1mu}}u\\ & = \int_0^{1/2 } q\biggl(\frac12-u\biggr)\bigl(e^{us}-e^{-us}\bigr)e^{-s/2}\operatorname{d\mspace{-1mu}}u\\ & \ge0.\end{aligned}\ ] ] the proof of lemma [ lem2-zhang - li - qi - eq ] is thus completed .
the property ( x ) in ( * ? ? ? * theorem 1.3 ) formulates that if @xmath125 then @xmath211 for @xmath212 .
since @xmath213 , see , and , by the property ( i ) in ( * ? ? ?
* theorem 1.3 ) , @xmath214 , replacing @xmath9 by @xmath215 , making use of the easy fact that @xmath216 , and letting @xmath217 yield @xmath218 . for a stieltjes function @xmath9 given by , by the stieltjes - perron inversion formula in @xcite
, we can determine the scalars @xmath219 $ ] and @xmath220 and the measure @xmath221 as done in @xcite .
specially , for the function @xmath222 , since @xmath223=0 $ ] and @xmath224 , we have @xmath225 for @xmath226 , where @xmath227 when @xmath228 and @xmath229 when @xmath230 because taking @xmath231 we obtain @xmath232-si}{\tau^2+s^2}}\ , \to-\sqrt{\frac{\tau(1-\tau)}{\tau^2}}\,\ ] ] when @xmath228 and @xmath233 when @xmath230 .
thus we find @xmath234 when @xmath235 and @xmath236 when @xmath237 .
substituting @xmath238 in the representation results in the formula .
the formula for @xmath160 or for @xmath239 can be derived in a similar way as above .
the rest is the same as in the first proof .
lemma [ lem2-zhang - li - qi - eq ] is proved once again .
our results on the harmonic mean @xmath7 may be stated as the theorem below .
[ harmonic - bernstein - thm ] the harmonic mean @xmath7 defined by is a bernstein function of @xmath69 on @xmath66 for @xmath58 with @xmath60 and has the integral representation @xmath240 consequently , @xmath241 the harmonic mean @xmath7 meets @xmath242}{(x+y+2t)^2 } = 1+\frac{(x - y)^2}{(x+y+2t)^2}>1.\ ] ] it is obvious that the derivative @xmath243 is completely monotonic with respect to @xmath69 . as a result ,
the harmonic mean @xmath7 is a bernstein function of @xmath69 on @xmath66 for @xmath58 with @xmath60 . in @xcite , it was listed that , for @xmath244 and @xmath245 , the classical euler gamma function @xmath246 this formula can be rearranged as @xmath247 for @xmath244 and @xmath248 . combining with yields @xmath249 and so , by integrating with respect to @xmath250 on both sides of
, the formula follows .
letting @xmath251 on both sides of and using the limit @xmath252=a(x , y ) $ ] generate the formula .
taking @xmath253 in produces .
theorem [ harmonic - bernstein - thm ] is thus proved . by (
* , theorem 3 ) or ( * ? ? ?
* , proposition 5.17 ) , it can be derived that the reciprocal of the harmonic mean @xmath7 , that is , the function @xmath254 , is logarithmically completely monotonic .
this logarithmically complete monotonicity can also be proved by considering @xmath255'=\frac{x^2+y^2 + 2(x+y)t+2t^2}{(x+t)(y+t)(x+y+2 t ) } = \frac12\biggl(\frac1{x+t}+\frac1{y+t}\biggr ) \biggl[1+\frac{(x - y)^2}{(x+y+2t)^2}\biggr]\ ] ] and that the product and sum of finitely many completely monotonic functions are also completely monotonic functions .
moreover , from , it follows readily that @xmath256 is an increasing function in @xmath6 for @xmath58 with @xmath60 .
our results on the geometric mean @xmath8 can be summarized as two theorems . by a direct differentiation ,
we have @xmath257 taking the logarithm on both sides of the above equality creates @xmath258 in @xcite , it was collected that for @xmath124 and @xmath259 , @xmath260 using this formula in leads to @xmath261/2}}{2v}\operatorname{d\mspace{-1mu}}v.\ ] ] since the function @xmath262 is convex on @xmath263 , we have @xmath264/2}\ge0.\ ] ] therefore , we have @xmath265^{(k)}=\frac{(-1)^k}2 \int_0^\infty\bigr\{e^{-(x+t)v}+e^{-(y+t)v}-2e^{-v[(x+t)+(y+t)]/2}\bigr\ } v^{k-1}\operatorname{d\mspace{-1mu}}v.\ ] ] this means that the derivative @xmath266 is logarithmically completely monotonic , and so it is also completely monotonic . as a result ,
the geometric mean @xmath8 is a bernstein function .
it is clear that the geometric mean @xmath8 satisfies @xmath267 and @xmath268'=\frac12\biggl(\frac1{x+t}+\frac1{y+t}\biggr),\ ] ] where @xmath269 if @xmath270 , then @xmath271 for @xmath272 and @xmath273 is completely monotonic in @xmath272 . on the other hand ,
the function @xmath274 is positive and @xmath275\\ & = \frac{(-1)^{i}(2i-3)!!}{2^{i+1 } } \frac1{u^{(2i-1)/2 } } \biggl(\frac{2i-1}{u}-1\biggr)\end{aligned}\ ] ] for @xmath70 , which implies that the function @xmath274 is completely monotonic on @xmath168 ; a ready modification of a conclusion in @xcite yields the following conclusion : if @xmath155 and @xmath276 are completely monotonic functions such that @xmath277 is defined on an interval @xmath11 , then @xmath278 is also completely monotonic on @xmath11 ; so , when @xmath279 , the derivative @xmath266 is completely monotonic and the geometric mean @xmath8 is a bernstein function .
consequently , considering the symmetric property @xmath280 , it is easily obtained that the geometric mean @xmath8 for @xmath6 with @xmath60 is a bernstein function . by and
, it follows that @xmath282 = \frac12h\biggl(\frac{y+t}{x - y}\biggr)\ ] ] and @xmath283^{(i)}=\frac1{2(x - y)^i}h^{(i)}\biggl(\frac{y+t}{x - y}\biggr)\ ] ] for @xmath284 . by the formula in lemma [ geom - mean - cm - lem - h(t ) ] , we have @xmath285^{(i)}=\frac{(-1)^i}{2^{i+1}\bigl(\frac{y+t}{x - y}\bigr)^{i+1 } ( x+t)^{i}h\bigl(\frac{y+t}{x - y}\bigr)}\\ \times\sum_{k=0}^{i-1}\frac{(i-1)!(i+1)!(2i-2k-1 ) ! ! } { ( i - k-1)!(i - k+1)!k!}2^{k}\biggl(\frac{y+t}{x - y}\biggr)^k,\end{gathered}\ ] ] which means that , when @xmath286 , the derivative @xmath266 is completely monotonic .
since @xmath280 , when @xmath287 , the derivative @xmath288 is also completely monotonic .
this implies that the geometric mean @xmath8 is a bernstein function of @xmath6 .
[ geometric - integral - thm ] for @xmath289 and @xmath290 $ ] , the geometric mean @xmath291 has the integral representation @xmath292 where the function @xmath293 is defined by .
consequently , the geometric mean @xmath8 is a bernstein function of @xmath69 on @xmath66 . for @xmath289 and @xmath290 $ ]
, making use of @xmath294 = \frac12h\biggl(\frac{y+z}{x - y}\biggr)\ ] ] and gives @xmath295 integrating with respect to @xmath296 from @xmath297 to @xmath298 on both sides of the above equation and interchanging the order of integrals yield @xmath299\operatorname{d\mspace{-1mu}}s\\ & = w+\frac{x - y}{2\pi}\int_0^\infty\frac{\rho((x - y)s)}s e^{-ys}\bigl(1-e^{-ws}\bigr)\operatorname{d\mspace{-1mu}}s.\end{aligned}\ ] ] since @xmath300 , the integral representation is readily deduced .
m. abramowitz and i. a. stegun ( eds ) , _ handbook of mathematical functions with formulas , graphs , and mathematical tables _ , national bureau of standards , applied mathematics series * 55 * , 9th printing , washington , 1970 . c. berg , _ stieltjes - pick - bernstein - schoenberg and their connection to complete monotonicity _
, in : _ positive definite functions . from schoenberg to space - time challenges _ , s. mateu and e. porcu ( eds . ) , dept .
jaume i , castelln de la plana , spain , 2008 , 24 pages .
chen , f. qi , and h. m. srivastava , _ some properties of functions related to the gamma and psi functions _ , integral transforms spec . funct . * 21 * ( 2010 ) , no . 2 , 153164 ; available online at http://dx.doi.org/10.1080/10652460903064216 .
b .- n . guo and f. qi , _ a completely monotonic function involving the tri - gamma function and with degree one _ , appl .
* 218 * ( 2012 ) , no . 19 , 98909897 ; available online at http://dx.doi.org/10.1016/j.amc.2012.03.075 .
b .- n . guo and f. qi , _ a property of logarithmically absolutely monotonic functions and the logarithmically complete monotonicity of a power - exponential function _
, politehn .
bucharest sci . bull .
phys . * 72 * ( 2010 ) , no . 2 , 2130 .
b .- n . guo and f. qi , _ the function @xmath306 : logarithmic convexity and applications to extended mean values _ , filomat * 25 * ( 2011 ) , no . 4 , 6373 ; available online at http://dx.doi.org/10.2298/fil1104063g
. g. a. kalugin , d. j. jeffrey , and r. m. corless , _ bernstein ,
pick , poisson and related integral expressions for lambert w _
, integral transforms spec .
* 23 * ( 2012 ) , no . 11 , 817829 ; available online at http://dx.doi.org/10.1080/10652469.2011.640327 .
g. a. kalugin , d. j. jeffrey , r. m. corless , and p. b. borwein , _ stieltjes and other integral representations for functions of lambert w _ ,
integral transforms spec .
* 23 * ( 2012 ) , no .
8 , 581593 ; available online at http://dx.doi.org/10.1080/10652469.2011.613830 . f. qi and b .-
guo , _ complete monotonicities of functions involving the gamma and digamma functions _ , rgmia res . rep
* 7 * ( 2004 ) , no . 1 , art . 8 , 6372 ; available online at http://rgmia.org/v7n1.php .
f. qi , s. guo , and s .- x .
chen , _ a new upper bound in the second kershaw s double inequality and its generalizations _ , j. comput .
appl . math .
* 220 * ( 2008 ) , no . 1 - 2 , 111118 ; available online at http://dx.doi.org/10.1016/j.cam.2007.07.037 . | it is general knowledge that the harmonic mean @xmath0 and that the geometric mean @xmath1 , where @xmath2 and @xmath3 are two positive numbers . in the paper ,
the authors show by several approaches that the harmonic mean @xmath4 and the geometric mean @xmath5 are all bernstein functions of @xmath6 and establish integral representations of the means @xmath7 and @xmath8 . |
magnetic fields are being recognized as one of the key components in star formation theories ( e.g. , * ? ? ?
nevertheless , their exact role in the formation and evolution of molecular clouds is still a matter of debate in the literature . on the observational side ,
one of the difficulties is to accurately measure the magnetic field strength .
the zeeman effect still provides the only known method of directly measuring magnetic field strengths along the line of sight in a molecular cloud ( e.g. , * ? ? ?
* ; * ? ? ?
a statistical method , aiming at separating a three - dimensional uniform field and a non - uniform field component is given in @xcite and was later extended and applied to optical polarization data in @xcite .
the goldreich - kylafis effect , the linear polarization in molecular line spectral emission @xcite , is a mechanism to trace the magnetic field morphology . up to now , the field structures of only a few sources have been detected with this effect ( e.g. , * ? ? ?
* ; * ? ? ?
polarization of dust thermal emission at infrared and submillimeter wavelengths provides another method to study magnetic field properties ( e.g. , * ? ? ?
the dust grains are thought to be aligned with their shorter axis parallel to the magnetic field lines .
therefore , the emitted light appears to be polarized perpendicular to the field lines @xcite .
radiative torques are likely to be responsible for the dust alignment @xcite .
dust polarization observations have been obtained both for cores together with their surrounding diffuser material ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and for higher - resolution collapsing cores ( e.g. , * ? ? ?
* ; * ? ? ?
complementary to zeeman splitting , dust polarization measurements are probing the plane of sky projected magnetic field structure .
they , however , do not directly yield a magnetic field strength , but only some field morphology . additional modeling
is needed in order to derive the actual field strength perpendicular to the line of sight .
a three - dimensional model of an hourglass magnetic field configuration is fitted to dust observations in , e.g. @xcite to yield a mean field strength . with pinched field configurations
detected , comparing gravitating force and field tension gives an upper limit for the field strength ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
another commonly used technique is the chandrasekhar - fermi ( cf ) method @xcite , where a field strength is derived from the field dispersion ( around a mean field direction ) combined with a turbulent velocity information .
various modifications and corrections to this method are discussed in the literature ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and tested through magneto - hydrodynamics ( mhd ) simulations @xcite .
the work by @xcite further refines this method noting that a globally derived dispersion might not reflect the true contribution from magneto - hydrodynamic waves and/or turbulence by introducing the concept of a dispersion function ( structure function ) around local mean magnetic field orientations
. however , the cf method and all its variations are statistical in nature , relying on ( large ) ensembles of polarization segments .
thus , they typically provide one _ ensemble - averaged _
value for the magnetic field strength for an entire map .
consequently , zones of weaker and stronger field strengths can not be separated .
it is , therefore , non - trivial to evaluate the relative , generally position - dependent role of the magnetic field ( compared to , e.g. , gravity , turbulence ) .
the goal of this paper is to introduce a new method which leads to a _
local _ magnetic field strength and , therefore , can map the position - dependent field strength over an entire molecular cloud the paper is organized as follows .
section [ source ] summarizes the submillimeter array ( sma ) polarization observation of the source w51 e2 which serves as an illustration case .
section [ method ] introduces our new method , starting with an observed correlation , the dust - mhd connection and a qualitative analysis of the predicted features of the method . a possible extension of the method to an arbitrary cloud shape
is discussed in section [ general ] .
the results are presented in section [ results ] .
we follow up with a discussion in section [ discussion ] which addresses application regime and limitations of the method , and which also proposes a measure to evaluate the dynamical role of the magnetic field in different zones of a molecular cloud .
summary and conclusion are given in section [ summary ] .
the w51 e2 map ( figure [ w51_e2 ] ) analyzed in the following sections was obtained with the sma interferometer @xcite .
the observation was carried out at a wavelength of 0.87 mm where the polarization of dust thermal emission is traced . in the extended array configuration ,
the synthesized beam resolution was about @xmath5 with a primary beam ( field of view ) of about @xmath6 at 345 ghz .
a detailed description of the data analysis with a discussion and interpretation of the obtained images is given in @xcite .
the re - constructed features depend on the range of @xmath7-sampling and weighting .
nevertheless , these data currently provide the highest angular resolution ( @xmath8 ) information on the morphology of the magnetic field in the plane of sky obtained with the emitted polarized light in the w51 star formation site .
this study is part of the program on the sma @xcite to study the structure of the magnetic field at high spatial resolutions .
w51 e2 is one of the strongest mm / submm continuum sources in the w51 region . located at a distance of 7 kpc @xcite , 1@xmath9 is equivalent to @xmath10 0.03 pc .
@xcite measured the polarization at 850@xmath11 m with scuba on jcmt across the molecular cloud at a scale of 5 pc with a binned resolution @xmath123 . at this scale ,
the polarization shows a morphology that changes from the dense cores to the surrounding diffuser gas .
@xcite reported a higher angular resolution ( 3@xmath9 ) polarization map at 1.3 mm obtained with bima , which resolved out large scale structures .
in contrast to the larger scale polarization map , the polarization appears to be uniform across the envelope at a scale of 0.5 pc , enclosing the sources e2 and e8 . in the highest angular resolution map obtained with the sma at 870 @xmath11 m with @xmath137 ,
the polarization patterns appear to be pinched in e2 and also possibly in e8 @xcite .
in particular , the structure detected in the collapsing core e2 reveals hourglass - like features . a statistical analysis based on a polarization structure function ( of second order ) shows a turbulent to mean magnetic field ratio which decreases from the larger ( bima ) to the smaller ( sma ) scales from @xmath14 to @xmath15 @xcite , possibly demonstrating that the role of magnetic field and turbulence evolves with scale .
the apparent axis of the hourglass along the north - west south - east direction coincides with the axes of the bipolar outflows in co(2 - 1 ) @xcite , co(3 - 2 ) and hcn(4 - 3 ) @xcite .
collapse and/or accretion flow signatures , possibly orthogonal to that , have been reported for various molecules in @xcite .
this might correspond roughly to the depolarization regions along the north - east south - west direction in figure [ w51_e2 ] .
the highest resolution sma map of w51 e2 serves as a testbed for the method developed in the following section .
figure [ w51_e2 ] reproduces the dust continuum stokes @xmath0 map with the magnetic field segments ( red ) from @xcite .
detected polarization segments are rotated by 90@xmath16 in order to derive the field directions .
typical measurement uncertainties of individual position angles are in the range of @xmath17 to @xmath18 . additionally shown in the map
are features relevant for the method in the following section .
in this section we introduce a new method which allows us to measure the ( projected ) magnetic field strength in a molecular cloud as a function of the position in a map . starting from an observed correlation in section [ motivation ] ,
the details of the method and its link to observed dust continuum stokes @xmath0 and polarization maps are given in section [ method_detail ] .
we follow up with a qualitative analysis of the features of the method in section [ analysis ] .
the stokes @xmath0 dust continuum contours and the magnetic field segments ( red ) from the dust polarization emission @xcite are shown in figure [ w51_e2 ] for w51 e2 .
interestingly , field segments tend to be perpendicular to stokes @xmath0 emission contours , or equivalently , polarization is preferentially tangential to the emission contour lines . instead of further working with intensity contours , we alternatively choose to investigate the intensity gradients ( blue thick segments in figure [ w51_e2 ] ) .
the top panel in figure [ correlation_b_i_grad ] displays the magnetic field position angles ( @xmath19s ) versus the intensity gradient @xmath19s .
a close correlation becomes manifest .
both @xmath19s are defined counter - clockwise starting from north in a range of 0 to 180@xmath16 .
the few data points ( black empty squares ) which originally do not seem to follow the trend , belong to pairs where both segments are close to @xmath20 ( vertical orientation ) , but with one of the @xmath19s slightly rotated to the left and the other one slightly rotated to the right hand side of the vertical . despite being closely correlated in orientation ( small absolute differences ) , these pairs would mimic a bad correlation simply due to the definition of the @xmath19s . for these cases , the original range of the magnetic field @xmath19 ( 0 to 180@xmath16 ) is extended by adding their relative difference in @xmath19s to their intensity gradient @xmath19s .
e.g. , a pair with an intensity gradient and magnetic field @xmath19 of 175@xmath16 and 5@xmath16 has a relative difference of 10@xmath16 , and is therefore displayed as ( 175@xmath16,185@xmath16 ) .
this re - definition does not alter the correlation in any way , but is merely a consequence of the ambiguity in the conventional definition of the @xmath21 @xmath19 range .
we note that the correlation is not limited to those pairs directly in the core collapsing region , but seemingly still holds for the most northern segments where the emission contours start to deviate from circular shapes .
the correlation in the definition of pearson s linear correlation coefficient is 0.95 .
it is important to remark that , despite this tight correlation , the distribution of the differences ( between magnetic field @xmath19s and intensity gradient @xmath19s ) is non - gaussian with an absolute mean of @xmath22 .
a ks - test rejects the null hypothesis ( differences following a gaussian distribution ) with a @xmath23value of @xmath24 . as a matter of fact
, the relative differences show a bimodal distribution ( bottom panel in figure [ correlation_b_i_grad ] ) , which results from a systematic clockwise and counter - clockwise rotation of the magnetic field with respect to the intensity gradient . this further points toward a systematic origin of this correlation .
all this leads to the conclusion that magnetic field and intensity gradient can not simply be radially aligned with the apparent misalignment just being the result of some measurement uncertainty .
it is the focus of an on - going study where this correlation is further analyzed on a larger data sample in the context of the evolution of molecular clouds ( koch et al .
2012 ; in preparation ) . in the following section , this observed correlation and the measured non - gaussian differences in position angles of the magnetic field and the intensity gradient
are further used to investigate the field strength .
molecular clouds are the sites of interactions of various forces .
gravity , magnetic fields , expanding shells such as from expanding hii regions or supernovae , accretion and outflows all add to the dynamics , and their relative importance can vary with the evolutionary stage of the star formation site . generally , magneto - hydrodynamics ( mhd ) equations - possibly for various molecules and dust with individual coupling terms - are needed to provide an accurate and complete description of the system . in the following , aiming at making use of dust polarization measurements , a single mhd equation for the dust component is adopted . under the assumptions of negligible viscosity and infinite conductivity ( ideal mhd case ) we focus on the force equation : @xmath25 where @xmath26 and @xmath27 are the dust density and velocity , respectively .
@xmath28 is the magnetic field .
@xmath29 is the hydrostatic dust pressure .
@xmath30 is the gravitational potential resulting from the total mass contained inside the star forming region .
@xmath31 denotes the gradient . as usual
, the left hand side in equation ( [ mhd_momentum ] ) describes the resulting action based on the force terms on the right hand side .
these include the gradients of the hydrostatic pressure terms of the gas , the magnetic field and the gravitational potential together with the magnetic field tension term ( last term on the right hand side ) .
in the following , an interpretation of equation ( [ mhd_momentum ] ) is derived which can be matched to a polarization observation as in figure [ w51_e2 ] .
we start by noting that the magnetic field tension term , @xmath32 , can be rewritten as : @xmath33 where a magnetic field line , @xmath34 , is directed along the unity vector @xmath35 with the generalized coordinate @xmath36 along the field force line . a change in direction , @xmath37 , measures the magnetic field curvature @xmath38 and is directed normal to @xmath35 along the unity vector @xmath39 . in an analogous way
, @xmath40 can be reformulated with the velocity vector @xmath41 directed along the unity vector @xmath42 .
generally , the gradient of the isotropic field pressure , @xmath43 , in equation ( [ mhd_momentum ] ) is not necessarily directed along the field direction @xmath35 , but can also have a component orthogonal to it .
therefore , when combining the equations ( [ b_curvature ] ) and ( [ mhd_momentum ] ) , a term @xmath44 remains which needs to be compared to the curvature term @xmath45 .
both @xmath39 and @xmath46 are orthogonal to @xmath35 , but they are not necessarily collinear . assuming @xmath47 and @xmath48 to be given approximately by the resolution of an observation , the significance of this additional term can be estimated as @xmath49 .
thus , the term can be omitted if the local change in the normal field component , @xmath50 , is small compared to the total field strength .
this holds for any smooth and slowly varying function of field strength .
equation ( [ mhd_momentum ] ) can then be transformed into : @xmath51 where the resulting directions of the gradients of pressure and gravitational potential are given with @xmath52 and @xmath53 , respectively . in the above formulation ,
generalized coordinates @xmath54 along the directions of the unity vectors @xmath55 are used .
generally , these vectors are not collinear , but have different three - dimensional directions with the condition that the vector sum of the right hand side adds up to the resulting direction on the left hand side .
each term in equation ( [ mhd_2 ] ) is intentionally written with its corresponding unity vector , defining its specific direction .
this will later be used when interpreting equation ( [ mhd_2 ] ) together with an observed map .
equivalently , all unity vectors can be expressed in any regular coordinate system ( e.g. spherical coordinates @xmath56 , @xmath8 , @xmath30 ) should this be of advantage .
the unity vectors @xmath57 can be interpreted as volume averaged directions , where the size of a volume element will depend on the resolution of an observation . in the above derivation
we have neglected the partial time derivative , @xmath58 , assuming stationarity .
equation ( [ mhd_2 ] ) describes the basic interaction of forces and the resulting motion .
linking some of these terms ( in direction and strength ) with observations will then allow us to solve for others .
our goal is to isolate the magnetic field strength @xmath59 in the term @xmath60 .
therefore , the remaining terms in equation ( [ mhd_2 ] ) need to be identified and characterized .
dust in star formation sites is reacting to all the force terms in equation ( [ mhd_2 ] ) .
maps of observed dust emission are reflecting the overall result of gravity , pressure and magnetic field .
consequently , they represent a measure of the resulting motion , the left hand side in equation ( [ mhd_2 ] ) .
we make the fundamental assumption that a change in emission intensity is the result of the transport of matter driven by the combination of the above mentioned forces . adopting this
, it then follows that the gradient in emission intensity defines the resulting direction of motion on the left hand side in equation ( [ mhd_2 ] ) . for the directions of the gradients of the pressure and the gravitational potential we assume here , for simplicity , a spherically symmetrical molecular cloud , where the center can be identified from the peak emission .
this assumption is relaxed in section [ general ] , where the method is generalized to an arbitrary cloud shape .
it is important to remark that the derivation so far is generally valid for 3 dimensions .
projection effects in both the integrated stokes @xmath0 and the polarized emission inevitably are present in an observation . in the following we first proceed by identifying equation ( [ mhd_2 ] ) with an observed map in 2 dimensions .
projection effects are later addressed in section [ projection ] .
figure [ schematic_b ] illustrates the further steps .
the derivation of @xmath59 is shown for a measured polarization ( red ) and intensity gradient direction ( blue ) in the first quadrant ( with respect to the gravity center which here is supposed to coincide with the emission peak ) .
for each measured pair of polarization and intensity gradient , its distance and direction from the gravity center defines the gravitational pull , @xmath61 .
generally , gravity , magnetic field and the intensity gradient directions are different .
the schematic in figure [ schematic_b ] illustrates how the vector sum of the various terms in equation ( [ mhd_2 ] ) is constructed .
solving for the magnetic field term then relies on measurable angles in the orientations between polarization and the intensity gradient ( @xmath62 ) and the difference between the gravity and the intensity gradient directions ( @xmath63 ) .
applying a sinus theorem to the closed triangle , @xmath64 , leads to the expression for the magnetic field strength : @xmath65 where , in the most general case , all variables are functions of positions in a map .
we note that although illustrated for a simple , close to spherically symmetric case in figure [ schematic_b ] with @xmath66 and @xmath67 aligned and pointing toward the same center equation ( [ b ] ) is generally valid for any directions of @xmath66 and @xmath67 .
thus , the above derivation is not restricted to a spherical collapsing core , but can be applied to any configuration where the various local force directions can be identified ( see also sections [ general ] and [ application ] ) . as an important further outcome of the method
, the angle factor @xmath68 in equation ( [ b ] ) has a direct physical meaning . with the magnetic field tension force term @xmath69 and
the gravitational and pressure forces @xmath70 , equation ( [ b ] ) can be reformulated as : @xmath71 where we have introduced @xmath72 to define the field significance .
this directly quantifies the local impact of the magnetic field force compared to the other forces involved ( gravity and pressure gradient ) . remarkably , the ratio is free of any model assumptions and simply relies on the two measured angles .
it is purely based on the geometrical imprint by the various forces left in the field and intensity gradient morphologies .
the angle factor , thus , also provides a model - independent qualitative criterion whether the magnetic field can prevent an area in a molecular cloud from gravitational collapse @xmath73 or not @xmath74 .
the field significance @xmath72 and its implications for the mass - to - flux ratio and the star formation efficiency are investigated in @xcite .
we briefly illustrate the expected features of our method when applied to a representative field and intensity configuration in a molecular cloud ( figure [ qualitative ] ) .
a summary is given in table [ table_analysis ] .
field lines like ( i ) and ( ii ) are expected in a core collapse when gravity is dragging inward the magnetic field which is flux - frozen to the dust particles . in the case of ( i ) , the field line is close to the pole , little bent , and the gas is mostly moving along the field .
intensity gradients are mostly radial , pointing toward the center of gravity .
therefore , in the outer region ( a ) , @xmath75 const , @xmath76 small ( @xmath77 ) , @xmath63 small ( @xmath78 const ) and thus @xmath79 const .
the field strength @xmath59 will scale as @xmath80 , where @xmath81 describes a power - law profile resulting from density and gravitational potential gradient ( section [ result_field_strength ] ) .
small ( random ) changes in angles are possible due to local turbulence and can lead to some scatter . in the inner part ( b ) , @xmath48 might slightly decrease with the other parameters remaining roughly constant .
this will lead to similar scalings for @xmath72 and @xmath59 .
a field line of type ( ii ) is expected around the accretion direction .
the outer field ( a ) is still little bent ( @xmath75 const ) , @xmath82 const and small , @xmath63 small , and therefore @xmath79 const and @xmath83 as in ( ia ) . in ( b )
the field lines start to bend more significantly ( @xmath48 decreases ) and @xmath76 increases ( @xmath84 decreases ) .
@xmath63 is small ( @xmath85 const ) or possibly increasing if a a non - spherical central core is being formed .
thus , @xmath72 can increase as @xmath86 .
the field strength is additionally modulated by @xmath87 which can lead to a steeper or shallower profile than @xmath88 depending on the exact values of @xmath48 and @xmath62 . in the most inner section ( c ) ,
the field radius is minimized , field and intensity gradient directions are close to orthogonal to each other ( @xmath89 modulo some turbulent dispersion , @xmath90 ) , and @xmath63 is small .
the force ratio @xmath72 can then possibly further increase locally .
the field strength is thus mostly controlled by @xmath91 and the power - law exponent @xmath92 , which will lead to highest values here .
overall , for ( i ) and ( ii ) we thus expect the field strength to generally increase with smaller radius toward the center .
additional features are seen in figure [ w51_e2 ] , which are possibly not directly linked to the above cases .
region ( iii ) corresponding to the ne patch around the offset ( -0.3,1.8 ) shows a group of parallel b field segments with parallel intensity gradients . in this outer region
the collapse is likely initiated , but does not yet reveal clear signatures .
@xmath63 is changing , @xmath82 const ( @xmath93 const ) , @xmath75 const , and @xmath72 can therefore increase or decrease .
the field strength will generally be small here because of the low density .
the relative field significance can still be important here , as gravity has not yet fully taken control to shape and align intensity gradient and field direction .
area ( iv ) corresponds to the most northern part in figure [ w51_e2 ] .
this region is possibly disconnected from the main core , likely belonging to another smaller core being formed here .
the @xmath72 values are dominated by large changes in @xmath63 when linked to the gravitational center of the main core . due to the lack of a clear identification of a local gravity center ( in which case @xmath63
would be much smaller ) , @xmath72 is probably overestimated for the two most northern segments .
possible solutions to this issue are addressed in section [ local_gravity ] .
the field strength is again small here due to the low density .
this section focuses on the 2 remaining parameters in equation ( [ b ] ) , which are not yet explicitly locally evaluated : the magnetic field radius @xmath48 ( curvature @xmath94 ) and the gravitational potential @xmath30 .
in particular , the initial assumption of a spherically symmetrical molecular cloud from section [ method ] is given up .
it has to be stressed that the method in section [ method ] derived from the equation ( [ mhd_momentum ] ) remains unchanged , and is valid for both average or locally precise values in @xmath48 and @xmath30 . from the derivation in section [ method ] , both angles @xmath62 and @xmath63 , the gravitational potential and the pressure @xmath29 are generally functions of positions @xmath95 in a projected plane of sky map .
the magnetic field line radius @xmath48 ( curvature @xmath96 ) will generally also change with positions .
it is , however , not possible to define a curvature based on a single isolated polarization segment . for polarization observations with a sufficiently large number of segments ( as e.g. in figure [ w51_e2 ] ) , a curvature resulting from two adjacent ( or more ) polarization segments can be calculated .
two neighboring magnetic field segments , separated by distance @xmath97 , are assumed to be tangential to a field line of radius @xmath48 .
the difference in their position angles , @xmath98 , is a measure of how much the direction of the field line has changed over @xmath97 .
the curvature @xmath99 can then be written as ( figure [ schematic_curvature ] ) : @xmath100\right ) \label{c}.\ ] ] with this definition of a local curvature , the calculation of the field strength @xmath59 in equation ( [ b ] ) is no longer limited to an average field curvature , but can be precise to any locally varying field structure .
we remark that the above definition for @xmath99 is not necessarily unique . in particular , determining which segments connect to a field line
can be non - trivial .
additionally , for some neighboring segments , as it can also be seen in figure [ w51_e2 ] , it is obvious that they can not be both tangential to the same field line .
the curvature as defined in equation ( [ c ] ) is therefore rather a local averaged curvature . in this view
, a local curvature can not only be defined with two segments which are roughly lining up , but also with two segments side by side .
the local field radius map in figure [ local_factors ] , middle left panel , is calculated by averaging the local curvatures between two adjacent segments over all available closest neighbors .
a closest neighbor is defined through the grid spacing of the map .
depending on the resolution and the detection of a polarization structure , the definition of @xmath99 can possibly be adjusted or refined . for w51 e2 ,
the local field radius derived from equation ( [ c ] ) is in the range between @xmath101 and @xmath102 ( figure [ local_factors ] , middle panels ) .
whereas most areas in the map show rather uniform patches with @xmath103 , regions in the north - east and south - west around the accretion plane show larger radii .
this is possibly a consequence of the accretion flow which tends to align field lines until some limiting inner radius where the field will bend .
yet more likely , this is a boundary effect because only two almost parallel segments are available here for averaging .
this leads to unphysically large radii .
the local radius averaged over the map is @xmath104 .
discarding the extreme values gives @xmath105 .
unless the gas distribution in a molecular cloud is ( close to ) perfectly symmetrical in azimuth around an emission peak , @xmath106 will differ in direction and strength from a simple radius - only dependence in the gravitational potential .
thus , for an arbitrary mass distribution , the direction and strength of the local gravitational pull , @xmath61 in equation ( [ mhd_2 ] ) , can not any longer simply be constructed with the direction and distance to the reference center ( emission peak in figure [ w51_e2 ] ) . instead , at any position in a map , the distribution of all the surrounding mass has to be taken into account to calculate the resulting local gravitational force . since the method in section [ method ]
is based on measuring angles between various force directions ( including gravity ) at all positions in a map , any substantial local deviations from a simple global gravity model can be relevant . for a discrete mass distribution ( @xmath107 ) ,
the resulting total potential @xmath30 at a location @xmath108 is : @xmath109 , where @xmath110 is the gravitational constant .
we assume that a detected dust emission ( like in figure [ w51_e2 ] ) is proportional to the gravitating mass , and that it further traces reasonably well the overall gravitating potential . additionally , it is assumed that the dust temperature is roughly constant over the emission map . in this case
, there is no weighting for @xmath107 when linking it to the integrated dust emission @xmath111 . the local gravity force ( per unit mass ) , @xmath112 is then : @xmath113 where @xmath114 is a conversion factor linking dust emission and total mass , and @xmath115 is the direction between each detected position with dust emission ( @xmath116 ) and the local position ( @xmath108 ) where the gravity force is evaluated .
we remark that , if @xmath117 over the entire map , the local gravity direction can be derived without prior knowledge of @xmath118 . for a typical observation , the mass ( intensity ) distribution is obtained by pixelizing a map with the beam resolution . figure [ local_factors ] , bottom left panel , shows the deviations from an azimuthally symmetrical potential for w51 e2 when calculating the local gravity directions following equation ( [ g ] ) for @xmath117 .
this analysis shows that , although w51 e2 ( figure [ w51_e2 ] ) appears to be a rather spherically symmetrical collapsing core , the deviations are systematically positive or negative on one or the other side of the hourglass axis .
minimum deviations are roughly along the south - east north - west direction which is consistent with the hourglass axis .
the average absolute deviation is @xmath119 .
as further discussed in the sections [ dynamic_b ] and [ b_gravity ] , already this simple symmetry analysis points toward a systematic influence of the magnetic field .
these deviations provide a correction for the angle @xmath63 ( angle between the gravity direction and the intensity gradient , figure [ schematic_b ] ) , which originally is calculated assuming an azimuthally symmetrical potential ( figure [ angles ] , top left panel ) .
consequently , also the magnetic field strength @xmath59 in equation ( [ b ] ) needs to be recalculated .
the angle @xmath62 , as defined relative to the intensity gradient , remains unchanged with this local gravity correction .
we finally note that further refinement of the local gravity correction introduced here is possible .
local over - densities which follow the flow of the global dynamics
can eventually locally collapse once they pass the jeans mass .
the north - west extension in figure [ w51_e2 ] is possibly such a case .
using equation ( [ g ] ) , however , the resulting local gravity direction will always be dominated by the main mass ( emission ) .
consequently , the angle @xmath63 might be over- or underestimated .
this defect can be corrected by identifying local collapse centers with a local maxima search and a suitable threshold criterion .
this section applies the method developed in section [ method ] to the case of w51 e2 .
starting from the expression for the magnetic field strength @xmath59 in equation ( [ b ] ) , results are first presented for separated factors and then for the final field strength .
whereas in the original observed map in figure [ w51_e2 ] the magnetic field is gridded to half of the synthesized beam resolution , the maps in the following are interpolated for an enhanced visual impression .
this can emphasize features along the field lines where polarization is detected at equally separated spacings . on the other hand
, it might lead to non - physical values in depolarization zones along the south - west north - east direction or in the center , where no data are available .
the angle @xmath63 measures the difference between the orientations of the intensity gradient and gravity .
small values in @xmath63 ( or @xmath121 ) therefore indicate that a change in the local intensity structure is closely following global gravity .
the top panels in figure [ angles ] show a map of @xmath121 together with its trends as a function of radius and azimuth .
the peak emission ( figure [ w51_e2 ] ) is again assumed as the reference center , and the gravity direction is simply taken as the radial direction from the center to any measured intensity location .
small values of @xmath121 - in the range between 0 and 0.2 - are found for most of the magnetic field segments pointing toward the center .
an x - pattern is revealed made up from the majority of the segments .
there is a trend of larger @xmath121 values in the north - east and south - west where likely accretion is ongoing .
the magnetic field segments in the far north show @xmath121 values close to one . here , the intensity gradient directions are very different ( by up to 90@xmath16 ) from the global gravity directions .
this northern patch , therefore , seems to be decoupled from the main collapsing core ( section [ analysis ] , case ( iv ) ) .
@xmath121 averaged over the entire map gives @xmath122 ( @xmath123 ) .
the average decreases to about 0.19 ( @xmath124 ) if the six polarization segments in the far north are ignored .
a hint of an increase in @xmath121 with radius is apparent in the top right panel in figure [ angles ] .
again , the most northern patch dominates values beyond a distance of @xmath125 .
no clear structure in azimuth is revealed .
since @xmath63 measures a deviation from gravity , mass distributions differing from an azimuthal symmetry can introduce corrections to the local gravity direction ( section [ local_gravity ] ) .
the top left panel in figure [ angles_local_gravity ] displays the @xmath120 map calculated from the generalized gravitational potential in equation ( [ g ] ) .
the overall patterns are still similar to the uncorrected @xmath120 map , but with smaller values in the accretion areas . as a result
, the average is reduced to @xmath126 ( @xmath124 ) and to about 0.09 ( @xmath127 ) when discarding the most northern patch .
the difference in orientations between the intensity gradient and the magnetic field is denoted with @xmath76 ( figure [ schematic_b ] ) .
the angle @xmath62 is its complement to 90@xmath16 . or , in other words , @xmath62 is the difference in orientations between a measured ( projected ) polarization and the intensity gradient .
values of @xmath62 close to its maximum of 90@xmath16 ( @xmath128 ) , indicate that the projected magnetic field orientation and the intensity gradient are closely aligned .
this again reflects the initially presented correlation in section [ motivation ] .
the middle panels in figure [ angles ] show a map of @xmath84 with its radial and azimuthal dependence . around the collapsing core ,
the values of @xmath84 are between 0.85 and 1 .
they decrease to about 0.7 in the most northern area .
the intensity and magnetic field orientations are therefore very similar over most of the map , with an average value @xmath129 ( @xmath130 ) .
leaving out the most northern part gives @xmath131 ( @xmath132 ) .
possibly there is a slight trend of decrease in @xmath84 with radius .
no clear structures appear in azimuth .
@xmath62 is not affected by local gravity corrections as outlined in section [ local_gravity ] .
the angle factor @xmath133 measures in a model - independent way the local impact of the magnetic field force ( section [ local_field_importance ] ) .
the top panels in figure [ local_factors ] illustrate the results .
values for the relative field significance , @xmath72 , range from @xmath134 to @xmath135 , with an average over the entire map @xmath136 .
thus , the magnetic field can balance only about one third of the gravitational force .
consequently , it can not prevent the core from any further collapse .
we stress that this result is obtained without any assumption about the mass involved .
the top panels in figure [ local_factors ] further show that the impact of the magnetic field changes as a function of position .
generally , the field influence seems to become weaker toward the center where gravity likely becomes more and more dominant .
this result is further explored in @xcite .
maps for the magnetic field strength @xmath59 are presented in this section for two scenarios : firstly , only the newly introduced force ratio @xmath72 is taken into account , leaving the other parameters in equation ( [ b ] ) constant .
this will illustrate the influence of @xmath72 .
secondly , the field strength is calculated by combining @xmath72 together with an analysis of the density profile and the resulting gravitational potential .
a first order estimate of the field strength , comparing the gravitational force @xmath137 and the magnetic field tension @xmath138 is given in @xcite . since w51 e2 is a collapsing core , @xmath139 ,
an upper limit of the field strength , @xmath140 mg , can be derived .
@xmath141 refers to the gas mass enclosed within a radius @xmath142 .
the radius of a magnetic flux tube is @xmath143 .
the mass density @xmath26 is estimated with the gas volume number density @xmath144 @xmath145 .
@xmath110 is the gravitational constant .
the above value , @xmath140 mg , will serve as a reference for the magnetic field strength maps presented in the following .
we remark that equation ( [ b ] ) is consistent with the above estimate , @xmath146 , when setting @xmath147 and neglecting the pressure gradient @xmath66 .
in the first scenario , the @xmath59 map for w51 e2 is calculated by taking into account only the angles @xmath63 and @xmath62 .
field radius , potential and density are set constant as above .
assuming local changes in temperature and density to be negligible compared to gravity , the pressure gradient @xmath66 is omitted from equation ( [ b ] ) in the following . from the measured ranges in
@xmath63 and @xmath62 ( figure [ angles ] , @xmath148 , @xmath149 ) , it is then obvious that mostly @xmath63 will dominate the range and structure in a magnetic field strength map ( figure [ map_b ] ) .
values of the factor @xmath150 range from @xmath134 to @xmath135 , with an average over the entire map @xmath136 ( figure [ local_factors ] , top panels ) .
the resulting patterns are very similar to the ones in the @xmath121-map ( figure [ angles ] , top left panel ) .
resulting variations in @xmath59 are then from a few mg to @xmath151 mg .
the map average value is @xmath152 mg . compared to the reference value above , the correction factor @xmath68 generally leads to a lower field strength . discarding the most northern patch
where @xmath59 is likely overestimated due to the lack of a clear local gravity center ( section [ analysis ] ) the value becomes @xmath153 mg .
obviously , when leaving curvature , gravitational potential and density constant , the radial profile of the field strength is determined by @xmath154 . as the field - to - gravity force ratio decreases toward the center , consequently the field strength also decreases and its largest values tend to be in the outer zones ( figure [ map_b ] ) .
this result needs to be compared with the following complete analysis . in the second scenario ,
the field strength is calculated by additionally taking into account the profiles of density and gravitational potential . as discussed later in section [ projection ] , the force ratio @xmath72 is little or not at all affected by projection effects .
therefore , we proceed to reconstruct 3-dimensional radial profiles assuming spherical symmetry . for that purpose ,
the azimuthally averaged dust emission profile from figure [ w51_e2 ] is deprojected with an abel integral . the resulting 3-dimensional radial profile ( @xmath56 )
is well described with a functional form @xmath155 , whereas the initial projected map ( @xmath48 ) is well fit with a profile @xmath156 .
the fitting is performed on profiles with values binned to half of the synthesized beam resolution . assuming the total mass of the e2 core to be distributed similar to its dust component
, the cumulative mass @xmath157 and the resulting gradient of the potential @xmath158 can be derived . combining all together ( equation([b ] ) ) , the magnetic field strength profile scales as @xmath159 . whereas the dust emission map allows us to derive this functional form
, its absolute normalization remains to be determined .
adopting @xmath136 ( section [ result_mf_force ] ) i.e. the observed averaged magnetic field strength can balance only about one third of the gravitating mass @xmath141 we normalize @xmath59 so that @xmath160 , where the latter field strength is equivalent to @xmath161 @xcite .
this gives @xmath162 mg .
the final field strength map and its radial and azimuthal trends are shown in the left panels in the figures [ b_map_local ] and [ b_map_local_profile ] .
an increase in field strength with smaller radius becomes apparent .
a profile @xmath3 is overlaid for illustration .
a hint of an increase in field strength around @xmath163 possibly points toward the forming core in the northwest .
this is found by still assuming a dominating gravity center at the emission peak .
it is obvious that the first map ( figure [ map_b ] ) can not represent the correct field strength , but merely reflects the relative field significance . only by adding profiles for the density and the gravitational potential can the magnetic field profile correctly be derived .
density and gravitational potential only ( setting @xmath164 ) would lead to a field profile @xmath165 .
taking into account the decreasing field - to - gravity force ratio with radius , leads to the shallower scaling @xmath166 . with the density profile , @xmath167
, this connects field and density as @xmath168 . whereas the dust emission shows little variations in azimuth ( close to spherical symmetry )
, the local position - dependent measurement of @xmath72 shows some variations .
consequently , features in the magnetic field strength map start to be revealed , with possibly regions of increased strength in the accretion zones in the north - east and the south - west directions . in order to trace the magnetic field to the very center
, higher resolution observations will be needed .
we remark that the projected dust emission can also reasonably well , although with larger residuals , be fit with a @xmath169 profile .
the deprojected profile @xmath170 then gives a magnetic field profile @xmath171 and @xmath172 ( not shown ) .
this leads to about 15% lower values in the field strength in the center with a similar mean field strength @xmath173 mg .
all the other features are left unchanged . in the above final field strength map we have neglected the variations in the field radius and assumed it to be constant ( figure [ local_factors ] , top panels ) .
this ignores two zones where large field radii are likely a consequence of boundary effects ( section [ local_curvature ] ) .
finally , the influence of the local gravity correction is illustrated in the right panels in the figures [ b_map_local ] and [ b_map_local_profile ] .
generally , features tend to be smoother . the field strengths in the accretion zones are less pronounced and a single center becomes apparent .
the average strength is about 7.7 mg . the radial profile of the binned values remains close to @xmath166 with maximum values around 19 mg . structures in azimuth remain similar although with smaller variations .
we stress that , although we are making use of the total gravitating mass @xmath174 which sets an upper limit for the field strength and defines its order of magnitude only the newly derived force ratio @xmath72 leads to a proper normalization and opens the possibility to derive a radial profile .
this has become possible by solving the mhd force equation by the method introduced in section [ method_detail ] .
errors @xmath175 in the final @xmath59 field maps are estimated by propagating typical measurement uncertainties through equation ( [ b ] ) .
the individual polarization @xmath19 uncertainty of @xmath17 to @xmath18 is assumed for @xmath176 .
a slightly smaller uncertainty , @xmath177 to @xmath17 , can be expected for @xmath178 because of the averaging process in the interpolation when calculating the intensity gradient .
this leads to @xmath179 , where the numerical factor follows from the above uncertainties combined with map averaged values for @xmath180 and @xmath181 . for any of the above scenarios and @xmath182 mg , we derive an average systematic error @xmath183 mg .
errors in the profiles for density and gravitational potential are small after binning and azimuthal averaging .
they are , therefore , neglected here .
besides this systematic error , there is an additional statistical uncertainty in the field orientation resulting from the turbulent field dispersion .
adopting @xmath17 to @xmath18 for this @xcite , this contributes to the final error budget with @xmath184 mg .
we note that all uncertainties in angles are smaller than the mean value ( @xmath185 ) of the differences between magnetic field and intensity gradient @xmath19s ( section [ motivation ] ) .
this further confirms that the correlation in figure [ correlation_b_i_grad ] can not only result from systematic and statistical uncertainties . in conclusion , density and gravitational potential profiles
determine the order of magnitude of the field strength .
the local features result from the angle factor @xmath68 .
these features remain robust with typical total errors ( systematic and statistical ) of @xmath186 mg .
the derived field strengths are also consistent with the level of magnetic fields typically detected by zeeman effects on oh masers in compact hii regions ( a few mg up to @xmath187 mg , @xcite ) .
the method presented in section [ method ] leads to a local magnetic field strength .
it , nevertheless , uses strictly speaking a not exactly local but larger scale property of the field , namely its curvature ( figure [ schematic_curvature ] ) .
the shape of a field line is the result of various forces acting on the dust particles coupled to the magnetic field .
these can be both long - ranging forces ( like gravity ) as well as locally acting perturbations ( like turbulence ) . in order to solve equation ( [ mhd_momentum ] ) ,
we need to determine the larger scale curvature .
the locally turbulent field can here act as a contaminant .
this is in contrast to , e.g. the cf method which precisely uses the field dispersion around an average field direction , in combination with some turbulent velocity information , to derive a magnetic field strength . in principle , a curvature can be calculated from a minimum of two segments ( section [ local_curvature ] ) , but the local turbulent field can introduce some uncertainty here . assuming an isotropic turbulent field component , a local curvature can be derived by averaging over several neighboring curvatures .
as long as reasonably smooth and connected polarization patterns are detected , this will reveal the larger scale field curvature .
this is the case for the field morphology in figure [ w51_e2 ] where a clear signature of dragged - in field lines is apparent .
higher resolution data are then needed to trace the field in the inner most part of the collapsing cloud where the curvature is expected to change faster within a small area .
the method further relies on identifying a noticeable effect from the various forces in equation ( [ mhd_momentum ] ) which shape the field on large scales .
when assuming flux freezing , this means that @xmath188 . in this weak field case of flux - freezing ,
the field lines are forced to move along with the gas ( at least up to some scale where ambipolar diffusion sets in ) . as a consequence , gravity
, intensity gradient and field orientations can be individually determined which is needed to calculate the angles in equation ( [ b ] ) . in the strong field case , @xmath189 , where matter can mostly move only along field lines
, the method will yield @xmath190 , @xmath191 , @xmath192 , and therefore @xmath193 .
( see also section [ analysis ] . ) unless the pressure gradient @xmath66 can be identified , which then allows to measure @xmath194 , the method can not further constrain @xmath59 ( see also table [ table_analysis ] ) .
geometrically , the method is based on closing a force triangle ( figure [ schematic_b ] ) . thus , if @xmath195 or @xmath90 or if both angles are zero at the same time , the triangle is no longer defined and the method fails . in summary , whereas the cf method is based on the field dispersion but has to isolate a mean large - scale field direction , the method discussed here uses the large - scale field morphology shaped by various forces . therefore , the method is more directly applicable to the weak field case .
a successful extension to the strong field case is possible if pressure and/or gravitational potential gradients can be localized .
generally , coherent patches with smooth changes in field and intensity gradient directions are of advantage as this leads to a clear identification of the field curvature and the acting forces .
the cf method needs both dust polarization continuum and molecular line velocity observations .
the method here uses the dust continuum stokes @xmath0 and its polarized emission .
whereas the cf method provides one single statistically averaged value , the approach here leads to a map of field strength values .
finally , we conclude this section by noting that the method here is not limited to dust continuum polarization .
as long as a force triangle can be identified , the method is applicable to any measurement which reveals the large - scale field morphology in combination with gravity and/or pressure forces .
notably , molecular cloud polarization maps obtained from the goldreich - kylafis ( gk ) effect , i.e. from spectral line linear polarization @xcite , can be analyzed in the same way .
this might open a way to study outflows which are often not observed in dust continuum due to their relatively small masses .
the only complication possibly rises from the difficulty that it can be non - trivial to conclude whether the magnetic field is parallel or perpendicular to the polarization detected from the gk effect . in an even broader context
, the method introduced here can also be applied to field morphologies detected from polarization by synchrotron emission or from faraday rotation measures .
the method is , therefore , not limited to molecular clouds , but can potentially also be adopted for , e.g. galaxy rotation measures . the results in section [ results ]
were derived assuming the plane of sky projected geometry in figure [ schematic_b ] . here
, we address the question of how an original 3-dimensional configuration affects the parameters in equation ( [ b ] ) to calculate the magnetic field strength .
the schematic in figure [ schematic_3d ] illustrates the connection .
we note , that like in the 2-dimensional ( 2d ) case , gravitational pull ( and pressure gradients ) , magnetic field tension and the resulting intensity gradient have to form a closed triangle
. the projected 2d triangle can generally be deprojected with 2 different inclination angles : @xmath196 and @xmath197 , corresponding to the inclinations of the deprojected orientations of the magnetic field tension and the gravitational pull , respectively . strictly speaking , projection effects are likely to be present to some extent in all variables in equation ( [ b ] ) . having deprojected the intensity emission profile and adopted an average field radius ( section [ result_field_strength ] ) , we limit the discussion to the remaining factor @xmath68 . in the following , lower indices _ 3 _ or _
2 _ refer to 3-dimensional or 2-dimensional projected quantities . the deprojection of the side @xmath198 of the triangle ( figure [ schematic_3d ] ) can be written as @xmath199 and @xmath200 , where @xmath201 is the length of the projection along the line of sight ( @xmath202-direction ) .
analogous relations hold for the side @xmath203 . linking a sinus theorem in the 3d triangle , @xmath204 , to the previously used 2d expression @xmath205 ,
then yields : @xmath206 where @xmath207 $ ] .
it becomes obvious that for any inclination angles , the projection effects cancel out as long as gravitational pull ( @xmath197 ) and field tension ( @xmath196 ) are inclined by the same angles , i.e. @xmath208 .
furthermore , if @xmath196 and @xmath197 are different , projection effects are negligible as long as @xmath196 and @xmath197 are small . for @xmath196 and @xmath197 within about @xmath209 ,
the projection effect is within about @xmath210% .
the left panel in figure [ projection_effect ] displays the relevant function @xmath211 , which shows a large relatively flat saddle around one in the center . for extreme values ,
i.e. inclination angles @xmath212 , the projection effect becomes severe ( @xmath213 or @xmath214 ) . besides having two totally different inclination angles @xmath196 and @xmath197
, we also investigate the projection effect as a function of the difference between the two inclination directions , @xmath215 ( figure [ projection_effect ] , right panel ) . since the gravitational pull and the field tension directions are dynamically coupled , their inclination angles are likely to be not completely independent .
the right panel in figure [ projection_effect ] shows that projection effects are within about @xmath216% for a range in the differences @xmath217 up to about 100@xmath16 or more , up to inclinations of about @xmath218 . for larger inclinations
@xmath118 grows again .
these situations , however , might not be of a concern , because at such large inclinations of the field lines , polarization emission ( perpendicular to the field lines ) might not be detectable .
the analysis might , therefore , not be applicable . in summary , for identical inclination angles projection effects cancel out .
the effect is small ( up to @xmath219% ) or even negligible for one inclination up to @xmath220 and the other one varying within about @xmath221 and @xmath222 ( range in @xmath217 up to about @xmath223 ) . only for extreme inclinations larger than about @xmath224 projection effects
can severely affect the method .
such situations might , however , not be detectable .
it , thus , has to be emphasized that it is unlikely that severe projection effects are present if regular polarization patterns are observed .
consequently , the method introduced here leads to a total magnetic field strength . without any further modeling ,
the method does not include components such as e.g. , rotation or temperature variations .
effects due to rotation are not explicitly considered in equation ( [ mhd_momentum ] ) .
nevertheless , an additional term due to rotation on the right hand side of equation ( [ mhd_momentum ] ) would primarily lead to a change in the resulting direction of motion , @xmath225 in equation ( [ mhd_2 ] ) , which was identified with the intensity gradient direction .
technically , this is equivalent to an additional inclination in the intensity gradient and can thus be approximated as a projection effect .
similarly , the directions of the magnetic field lines dragged along with the rotating gas would be subject to an additional projection effect . as discussed in section
[ projection ] , the resulting uncertainty in the field strength is relatively small for a large parameter space of inclination angles . alternatively ,
if the rotation velocity and axis are known , equation ( [ mhd_momentum ] ) could be written in a rotating frame with an additional contribution from a centrifugal force .
density and temperature combined make up for the detected dust emission .
significant variations in temperature could mimic mass concentrations which might then lead to wrong gravity directions and , therefore , to wrong magnetic field strength values .
if concentrated locally such as e.g. , in the case of radiation pressure from a star , this is unlikely to alter the global features in the field strength over an entire map .
similarly , small shifts in the position of a dominating emission peak ( like in figure [ w51_e2 ] ) introduce only minor variations in the field strength .
only if there are very irregular and significant changes in temperature over a map , the method is likely to fail .
additional information ( such as from spectra ) might then help to isolate the temperature effect .
following up on section [ application ] , we expect the strong and weak field case of flux - freezing to be present with different characteristics . in the strong field case gas moving along the field lines the intensity gradient direction is presumably closely aligned with the magnetic field orientation . on the other hand , in the weak field case
field lines dragged along and shaped by the forces acting on the gas intensity gradient and field orientations can be at any angle . in the extreme case they can be at a 90@xmath16 angle , when e.g. a gas flow is bending a field line ( case ( iic ) in figure [ qualitative ] ) .
the angle @xmath226 , as previously introduced in figure [ schematic_b ] , measures the absolute difference in orientations between the intensity gradient and the magnetic field . since the cases of weak and strong field in a weak field region is not necessarily weaker than the field in a strong field zone . ]
flux - freezing categorize the relative importance of the field , we propose here the angle @xmath76 as a diagnostic measure for the dynamical role of the @xmath59 field .
small angles ( @xmath227 ) will point toward a dominating field , whereas larger and larger angles mean that the field is being overwhelmed by other forces .
the map in figure [ flux_freezing ] displays @xmath76 , revealing some areas with @xmath228 .
these areas roughly coincide with the expected hourglass at larger radii . here
, the gas is possibly collapsing along the field lines .
additionally , it is likely that gravity has initially led to this field and intensity gradient configuration , with @xmath228 being a mere consequence of that . both at smaller radii toward the center and away from the hourglass toward the accretion plane , values for @xmath76 increase .
this might be an indication that the field is dynamically less important here and that gravity has largely taken control .
it needs to be stressed that this does not necessarily mean that the field strength is smaller here , because @xmath59 further also depends on @xmath48 , @xmath106 and @xmath66 ( see also previous footnote ) .
moreover , the field can still provide pressure support which can slow down the gravitational collapse .
this picture with generally smaller @xmath76 values around the core and larger values toward the north - west corner is consistent with the angle factor @xmath68 ( section [ local_field_importance ] and top panels in figure [ local_factors ] ) , which identifies the field force to be less significant in the core .
figure [ flux_freezing ] is necessarily complementary to the middle panels in figure [ angles ] , because @xmath229 .
we have to note that the angle @xmath76 ( like @xmath62 ) is subject to a projection correction . in the notation of section [ projection ] : @xmath230 with @xmath231 . in the particular case of a close to spherical collapse , @xmath232 , because both the 3-dimensional and the projected intensity show a clear azimuthal symmetry ( figure [ w51_e2 ] ) .
assuming @xmath233 to be approximately constant over a map ( section [ projection ] ) , then gives @xmath234 .
thus , despite the projection effect , the dynamical importance of the field can still be understood from the relative differences in @xmath76 .
the significance of the angle @xmath76 as a diagnostic measure is further analyzed on a larger data set together with the correlation from section [ motivation ] in koch et al .
, 2012 ( in preparation ) .
the method developed in section [ method ] is based on identifying the three orientations of gravity , the magnetic field and the emission intensity gradient . in particular ,
two angles are relevant in order to isolate and calculate the magnetic field strength @xmath59 : the angle @xmath63 between the gravity direction and the intensity gradient , and the angle @xmath62 between the polarization orientation and the intensity gradient . given the three orientations , an additional angle , @xmath236 , can be read off from a map .
@xmath236 measures the difference in orientations between gravity and the magnetic field ( figure [ schematic_b ] ) .
thus , it is a measure for how effectively gravity is able to shape the field lines and aligning them with the direction of the gravitational pull .
small angles ( small values of @xmath235 ) indicate close alignment .
the bottom panel in figure [ angles ] shows @xmath235 for the case of w51 e2 , assuming the center of gravity to be at the emission peak .
the bottom panels in figure [ angles_local_gravity ] display again @xmath235 when a local gravity direction ( section [ local_gravity ] ) is adopted .
the two figures are qualitatively similar .
applying a local gravity direction seems to sharpen the contrasts . the closest alignment ( @xmath237 ) is found at the outer radii close to the accretion plane in the west and in the north - west direction around the axis of the hourglass - like structure .
an obvious poor alignment is in the northern area .
averaged over the entire map and excluding the northern patch we derive @xmath238 ( @xmath239 ) and @xmath240 ( @xmath241 ) , respectively .
interestingly , after the local gravity correction , there is a very close resemblance between the @xmath235-map and the @xmath242-map ( figure [ flux_freezing ] ) .
since @xmath236 measures the magnetic field - gravity alignment and @xmath76 measures the magnetic field - intensity gradient alignment , the very similar maps , of course , indicate that the intensity structure is mostly resulting from gravity .
this is also shown with the small values in the @xmath121-map ( figure [ angles_local_gravity ] , top panels ) .
this is , of course , a consequence of the geometry in figure [ schematic_b ] because all the 3 angles are dependent : @xmath243 with @xmath244 .
nevertheless , the clear systematic differences in structures between the @xmath235 and the @xmath121-map ( figure [ angles_local_gravity ] ) lead to the conclusion that the @xmath59 field at least in some regions has kept its own dynamics and is not yet aligned and overwhelmed by gravity .
this result is then also consistent with the finding of zones of weak and strong flux - freezing ( section [ dynamic_b ] ) which characterizes the dynamical role of the magnetic field .
the systematic influence of the field is further manifest in the systematic deviations from an azimuthally symmetric emission structure , seen in the bottom panels in figure [ local_factors ] .
projection effects ( section [ projection ] ) are again likely to change only the absolute numbers , leaving relevant features and comparisons between maps still intact .
dust polarization observations in molecular clouds often show position angles of magnetic field segments which are close to perpendicular to the dust continuum intensity contours .
this correlation does not seem to be arbitrary , but can be interpreted in the context of ideal magneto - hydrodynamics ( mhd ) . based on this
, we put forward a new method to derive a local magnetic field strength .
the key points are summarized in the following . 1 . _
new method for magnetic field strength mapping : _ a new method is proposed which leads to a position - dependent estimate of the magnetic field strength . the approach is based on measuring the angle between the polarization and the stokes @xmath0 intensity gradient orientations .
assuming that the intensity gradient is a measure for the resulting direction of motion in the mhd force equation , this angle quantifies how tightly the direction of motion is coupled to the field orientation . in combination with the angle in between the orientations of gravity ( and/or the pressure gradient ) and the intensity gradient , the two angles can be linked to the field strength by identifying a force triangle .
application regime : _ the method relies on separating noticeable effects from various components in the mhd force equation , in particular from gravitational pull and field tension .
consequently , the method is more directly applicable to the case of weak field flux freezing , where the field morphology is clearly shaped by these forces . collapsing ( spherical ) cores in star formation sites are of most immediate interest because the directions of gravity and magnetic field are relatively easily determined
nevertheless , the technique is generally valid and applicable to any measurement ( also strong field flux freezing ) , where the directions of magnetic field , pressure gradients and/or gravitational pull can be localized . with the concepts of local curvature and local gravity direction , the method can be further generalized .
moreover , the method is not limited to dust polarization measurements , but can potentially also be applied to , e.g. molecular line linear polarization maps from the gk effect or even to faraday rotation measure maps for galaxies .
+ here , the method is applied to sma dust polarization data of the collapsing core w51 e2 .
averaging over the entire core , a field strength of @xmath2 mg is found .
variations in the field strength are detected with an azimuthally averaged radial profile @xmath166 .
maximum values toward the center are about @xmath4 mg .
the current data lack resolution in order to probe the innermost part of the core , where the largest field strength is expected .
comparison with chandrasekhar - fermi ( cf ) method : _ the method introduced here uses the large - scale field curvature . the local turbulent field dispersion is acting as a contaminant .
this is in contrast to the cf method which precisely uses the field dispersion after separating a mean field orientation in combination with a turbulent velocity information in order to derive a field strength .
whereas the cf method yields one statistically averaged field strength for an ensemble of polarization segments , the method here provides a field strength at each location of a detected polarization segment .
limitations and shortcomings of the method : _ technically , the method fails where the force triangle can not be closed .
this leads to the unphysical value @xmath193 .
a triangle degenerating to a single line corresponds to the situation where the field is very weak and , therefore , the intensity gradient does not deviate from the gravity direction .
the method then still consistently yields the limit @xmath245 . + the method is not entirely free of projection effects . in the most general case , the force triangle has to be deprojected with two different inclination angles .
nevertheless , the projection correction enters as the ratio of the two inclination angles .
thus , in the most favorable case with identical inclination angles for gravity and field orientations , the projection effect cancels out and the method leads to a total magnetic field strength .
even if the inclination angles differ from each other , the possible error due to unknown inclinations is relatively modest , within @xmath216% to @xmath246% , for a large parameter space . in any case , if the inclination angles do not vary significantly within the same source , the method always provides a way to detect relative differences in the field strength .
the accuracy and reliability of the method are limited by how well projection effects and forces defining the field morphology can be identified and quantified .
additional effects like , e.g. rotation and local variations in temperature are likely of minor importance . 5 . _ weak and strong flux freezing : _ the dynamical role of the magnetic field in the sense of weak or strong flux freezing can be assessed through the angle between the magnetic field and the intensity gradient orientations .
angles close to zero point toward a dominating role of the field ( or an overall controlling gravitational pull ) .
larger angles indicate that the field lines are more significantly shaped by other forces .
although this criterion alone does not constrain the field strength , it provides a measure to characterize the role of the magnetic field as a function of position in a map .
_ dynamical significance of magnetic field : _ the method additionally provides a model - independent criterion to quantify the magnetic field force compared to the sum of any other forces involved . as a result of the force triangle , the local significance of the magnetic field can be evaluated simply with the ratio of the angle between intensity gradient and gravity direction , and the angle between intensity gradient and polarization orientation .
these angles reflect the imprint of the various forces onto the dust and polarization morphologies .
this measure can then constrain the local impact of the magnetic field and quantify its support against gravitational collapse . the collapsing core w51 e2 @xcite .
contours show the stokes @xmath0 dust continuum at 0.87 mm with a synthesized beam resolution of about @xmath5 ( gray ellipse in the lower left corner ) with natural weighting .
colors correspond to the color wedge on the right hand side with units in jy / beam .
contours are linearly spaced in steps of 130 mjy / beam . overlaid
are the magnetic field segments ( thick red segments ) at the locations where polarized emission was detected .
magnetic field segments are plotted by rotating the polarization segments by 90@xmath16 .
only polarization data with a flux above @xmath247 , the rms noise of the polarized intensity , are included .
the blue vector field displays the gradient directions of the dust continuum emission , with most vectors pointing toward the emission peak .
highlighted are those emission intensity gradient directions ( thick blue segments ) at the locations of the magnetic field segments .
the length of the segments is arbitrary , and for visual guidance only .
the axes offset positions ( in arcsec ) are with respect to the original phase center of the observation at right ascension ( j2000)= @xmath248 , declination ( j2000)=@xmath249 ] top panel : the correlation ( blue filled squares ) between the intensity gradient @xmath19s ( blue segments in figure [ w51_e2 ] ) and the magnetic field @xmath19s ( red segments in figure [ w51_e2 ] ) .
the black empty squares belong to pairs with @xmath19s close to @xmath20 , one @xmath19 being slightly rotated to the left and the other one to the right hand side of the vertical . in order to properly display their correlations , the magnetic field @xmath19 is re - defined beyond the 0 to 180@xmath16 range for these cases ( blue filled squares above 180@xmath16 and below 0@xmath16 ) . for visual guidance added is the straight blue line , representing a perfect correlation . also shown are the absolute differences ( @xmath250 ) between the @xmath19s for each pair ( red crosses ) .
both @xmath19s are defined counter - clockwise starting from north .
bottom panel : the bimodal distribution of the relative differences ( in the range between @xmath251 to @xmath252 ) between the intensity gradient and the magnetic field @xmath19s , reflecting a systematic clockwise and counter - clockwise rotation of the field with respect to the intensity gradient .
, title="fig : " ] top panel : the correlation ( blue filled squares ) between the intensity gradient @xmath19s ( blue segments in figure [ w51_e2 ] ) and the magnetic field @xmath19s ( red segments in figure [ w51_e2 ] ) .
the black empty squares belong to pairs with @xmath19s close to @xmath20 , one @xmath19 being slightly rotated to the left and the other one to the right hand side of the vertical . in order to properly display their correlations , the magnetic field @xmath19 is re - defined beyond the 0 to 180@xmath16 range for these cases ( blue filled squares above 180@xmath16 and below 0@xmath16 ) . for visual guidance added
is the straight blue line , representing a perfect correlation . also shown
are the absolute differences ( @xmath250 ) between the @xmath19s for each pair ( red crosses ) .
both @xmath19s are defined counter - clockwise starting from north .
bottom panel : the bimodal distribution of the relative differences ( in the range between @xmath251 to @xmath252 ) between the intensity gradient and the magnetic field @xmath19s , reflecting a systematic clockwise and counter - clockwise rotation of the field with respect to the intensity gradient .
, title="fig : " ] illustration of the equation ( [ b ] ) with the relevant angles in order to solve for the magnetic field strength @xmath59 .
a pair of magnetic field ( red solid line ) and emission intensity gradient segments ( blue solid line ) are shown at a location @xmath253 with respect to a common reference center . the originally measured polarization orientation ( rotated by 90@xmath16 with respect to the field orientation )
is indicated as a red dashed line .
position angles are measured counter - clockwise with respect to north ( n ) .
@xmath254 measures the angle in between the intensity gradient and the field orientation .
@xmath62 is its complement to @xmath255 .
@xmath256 are assumed to be known in direction and strength . for simplicity ,
as suggested from figure [ w51_e2 ] for the case of w51 e2 , the emission peak is assumed as the gravity center , and thus defines direction and strength . here
, it also serves as the reference center .
@xmath63 denotes the deviation of the intensity gradient from the direction of @xmath256 .
the corresponding deviation for the magnetic field is measured with the angle @xmath236 .
the force triangle @xmath257 pointing toward the gravitating center , the resulting intensity gradient orientation ( left hand side in equation ( [ mhd_2 ] ) ) and the direction of the restoring magnetic field tension ( normal to the field segment ) can then be closed by intercepting the red dotted and black dotted lines , where the former one is shifted parallel to the red dashed line . ]
qualitative analysis ( section [ analysis ] ) : illustration of the expected features of the method .
the black solid lines are intensity emission contours .
blue and red segments are intensity gradient and magnetic field orientations , respectively , as in the figures [ w51_e2 ] and [ schematic_b ] .
the labeling refers to the cases discussed in section [ analysis ] and listed in table [ table_analysis ] .
all cases have similar equivalents in the w51 e2 map in figure [ w51_e2 ] . ]
illustration of the local curvature derivation .
position angles ( @xmath19s ) , separated by an observed distance @xmath97 , are measured counter - clockwise with respect to north ( n ) .
the 2 magnetic field segments ( solid red lines ) are assumed to be tangential to a connecting field line ( thick solid black line ) with radius @xmath48 .
the angle @xmath258 measures the difference in between 2 @xmath19s , @xmath259 .
@xmath48 is read from the isosceles triangle with @xmath260 , where @xmath261 .
this then leads to equation ( [ c ] ) .
note that the derivation uses the measured relative quantities @xmath97 and @xmath262 .
technically , this allows to detach the magnetic field segments from their absolute positions , by keeping their relative separation and @xmath262 , and ensures that a connecting field line can be constructed . strictly speaking ,
even this definition of curvature is not exactly local because it involves at least 2 segments separated by a distance @xmath97 . ]
position dependence of parameters in equation ( [ b ] ) .
magnetic field segments are shown in white .
top left panel : the combined factor @xmath150 , which quantifies the significance of the magnetic field force compared to the other forces ( section [ local_field_importance ] ) .
@xmath63 is measured with respect to a spherically symmetrical gravitational potential .
middle left panel : the local magnetic field radius calculated from the method in section [ local_curvature ] .
units are in arcsec .
bottom left panel : the difference in the direction of the gravitational pull between a spherically symmetrical potential and a local gravity direction as defined in section [ local_gravity ] .
units are in degree .
the difference is shown relative to the symmetrical assumption , revealing a symmetry axis along the southeast- northwest direction .
the right panels illustrate the tendencies for @xmath72 , @xmath59 field radius and difference in gravity direction ( from top to bottom ) as a function of radius ( distance from emission peak ) and azimuth angle ( measured counter - clockwise from west ) .
mean values averaged over the entire maps are shown with the red solid lines .
for the local gravity difference the mean is calculated from the absolute differences . ]
relevant angles for the derivation of the magnetic field strength , calculated for w51 e2 ( figure [ w51_e2 ] ) .
magnetic field segments are shown in white .
top left panel : @xmath121-map , where @xmath63 @xmath263 ( figure [ schematic_b ] ) is the difference between the orientations of the intensity gradient and gravity .
middle left panel : @xmath84-map .
@xmath264 @xmath263 measures the deviation between the magnetic field and the intensity gradient orientations .
bottom left panel : @xmath235-map , where @xmath236 is the difference between the magnetic field and gravity orientations .
the right panels show the corresponding tendencies for each angle as a function of radius ( distance from emission peak ) and azimuth angle ( measured counter - clockwise from west ) .
mean values averaged over the entire maps are shown with the red solid lines . ]
angles corrected for local gravity directions .
magnetic field segments are shown in white .
top panels : @xmath121 measuring the deviation between the local gravity and the intensity gradient orientations .
bottom panels : @xmath235 measuring the deviation between the local gravity and the magnetic field orientations . ] the magnetic field strength ( b ) derived by taking into account only variations in the field - to - gravity force factor @xmath265 .
units are in mg .
all the other parameters in equation ( [ b ] ) are constant over the map .
magnetic field segments are shown in white .
the right panels show the field strength tendencies as a function of radius ( distance from emission peak ) and azimuth angle ( measured counter - clockwise from west ) .
mean values averaged over the entire map are shown with the red solid lines . ]
final magnetic field strength maps calculated from equation ( [ b ] ) .
units are in mg . taken into account are the deprojected profile for the density , the resulting gravitational potential gradient profile and variations in the force ratio @xmath72 ( left panel ) . in the right panel , corrections due to the local gravity direction are additionally considered .
mean field strengths are about 9.4 mg and 7.7 mg for the left and right map , respectively .
the field radius is set constant .
magnetic field segments are shown in white .
typical measurement uncertainties lead to an average systematic error in the field strength of @xmath266 mg . ]
radial and azimuthal trends for the magnetic field strength maps presented in figure [ b_map_local ] .
radial profiles @xmath3 are overlaid for illustration in the top panels .
mean values averaged over the entire map are shown with the red solid lines . ]
the angle @xmath76 ( absolute difference in between the intensity gradient and the magnetic field orientations ( figure [ schematic_b ] ) ) as a diagnostic tool to distinguish between weak and strong case of flux - freezing .
magnetic field segments are shown in white .
the values for @xmath242 are displayed in the left panel in order to facilitate a comparison with the @xmath84 ( figure [ angles ] , middle panels ) and the @xmath235 maps ( figure [ angles ] , bottom panel ; figure [ angles_local_gravity ] , bottom panel ) . ] projected and deprojected triangles for the method presented in section [ method ] .
lower indices _ 3 _ or _
2 _ refer to 3-dimensional or 2-dimensional projected quantities .
the projected triangle ( dashed lines ) in the @xmath267-plane is deprojected along @xmath201 and @xmath268 ( parallel to the line of sight ) to form a 3d triangle ( solid lines ) .
the angles @xmath269 and @xmath270 are identical to @xmath63 and @xmath62 in figure [ schematic_b ] .
@xmath197 and @xmath196 are the inclination angles of the gravitational pull and the magnetic field tension , respectively . ]
left panel : the projection function @xmath271 for the two inclination angles @xmath196 and @xmath197 ( figure [ schematic_3d ] ) .
the relatively flat saddle in the center shows that projection effects are small ( @xmath272 ) for @xmath196 and @xmath197 within about @xmath209 .
highlighted in black and red are the parameter spaces where the projection effects are within @xmath216% ( @xmath118 within @xmath273 ) and within @xmath246% ( @xmath118 within @xmath274 ) , respectively . for much larger inclination angles polarization is not easily detectable .
it is , thus , unlikely that severe projection effects are present if regular polarization patterns are observed .
right panel : the projection effect as a function of the difference @xmath217 between the two inclination directions , @xmath215 , for various fixed inclinations @xmath196 .
the dashed line marks @xmath275 .
a @xmath216% bound is shown with the dotted lines . ]
ccccccc general + + & @xmath276 & @xmath277 & @xmath48 & @xmath276 & @xmath276 , @xmath278 & @xmath72 and @xmath279 with larger angle @xmath63 + & @xmath277 & @xmath276 & @xmath48 & @xmath280 & @xmath280 , @xmath281 & @xmath72 and @xmath279 with smaller angle @xmath62 + & & & & & & + specific cases + + ( i ) ( a ) & small , @xmath85 const & @xmath282 & @xmath85 const & @xmath85 const & @xmath85 const@xmath283 & field line close to pole , @xmath79 const , @xmath279 with smaller @xmath56 + ( b ) & small , @xmath85 const & @xmath282 & @xmath85 const & @xmath85 const & @xmath85 const@xmath283 & + ( ii ) ( a ) & small , @xmath85 const & @xmath282 & @xmath85 const & @xmath85 const & @xmath85 const@xmath283 & field line around accretion direction , @xmath79 const , @xmath279 with smaller @xmath56 + ( b ) & small , @xmath85 const & @xmath284 & @xmath284 & @xmath285 & @xmath286 & @xmath287 , @xmath279 with smaller @xmath56 steeper or shallower than @xmath88 ( depending on @xmath48 and @xmath62 ) + ( c ) & small , @xmath85 const & @xmath284 , @xmath227 & @xmath284 , @xmath288 & @xmath285 & @xmath286 & @xmath287 , maximum field strength @xmath59 + ( iii ) & @xmath289 & @xmath85 const & @xmath85 const & @xmath289 & @xmath289 , @xmath290 & outer area , no clear signatures yet , @xmath59 generally small due to @xmath88 + ( iv ) & large & @xmath85 const & @xmath85 const & large & @xmath290 & disconnected area , no clear gravity center localized , @xmath59 small due to @xmath88 + & & & & & & + limits + + & @xmath291 & @xmath277 & ( i ) @xmath292 , ( ii ) @xmath293 & @xmath291 & ( i ) @xmath294 , ( ii ) @xmath293 & triangle degenerates to single line + & @xmath277 & @xmath291 & ( i ) @xmath48 , ( ii ) @xmath295 & @xmath295 & ( i ) @xmath295 , ( ii ) @xmath295 & method fails , triangle not closed + & @xmath291 & @xmath291 & ( i ) @xmath48 , ( ii ) @xmath295 & @xmath2961 & ( i ) @xmath297 , ( ii ) @xmath295 & method fails , triangle not closed and degenerated to single line + | dust polarization orientations in molecular clouds often tend to be close to tangential to the stokes @xmath0 dust continuum emission contours .
the magnetic field and the emission gradient orientations , therefore , show some correlation .
a method is proposed , which in the framework of ideal magneto - hydrodynamics ( mhd ) connects the measured angle between magnetic field and emission gradient orientations to the total field strength .
the approach is based on the assumption that a change in emission intensity ( gradient ) is a measure for the resulting direction of motion in the mhd force equation . in particular
, this new method leads to maps of position - dependent magnetic field strength estimates .
when evaluating the field curvature and the gravity direction locally on a map , the method can be generalized to arbitrary cloud shapes .
the technique is applied to high - resolution ( @xmath1 ) submillimeter array polarization data of the collapsing core w51 e2 . a tentative @xmath2 mg field strength is found when averaging over the entire core .
the analysis further reveals some structures and an azimuthally averaged radial profile @xmath3 for the field strength .
maximum values close to the center are around @xmath4 mg . the currently available observations lack higher resolution data to probe the innermost part of the core where the largest field strength is expected from the method .
application regime and limitations of the method are discussed . as a further important outcome of this technique
, the local significance of the magnetic field force compared to the other forces can be quantified in a model - independent way , from measured angles only .
finally , the method can potentially also be expanded and applied to other objects ( besides molecular clouds ) with measurements that reveal the field morphology , as e.g. faraday rotation measurements in galaxies . |
prostate cancer ( cap ) is the second leading cause of cancer death among american men , killing 29,000 men each year .
the us preventive services task force ( uspstf ) cited the substantial rate of prostate cancer overdiagnosis ( 1750 % ) and overtreatment as justification for discouraging the use of prostate specific antigen ( psa ) for early detection of prostate cancer @xcite.the problem of overdiagnosis is caused by the inability to distinguish aggressive cancer from non - aggressive cancer .
preferential detection of aggressive cancers would improve outcomes in men who need treatment while sparing men with non - aggressive cancers the harms of treatment .
this is critically important because non - aggressive cancers are slow growing and rarely harmful , while aggressive cancers pose a substantial risk of morbidity and mortally @xcite . yet , current methods of prostate cancer detection , including the use of conventional ` blind ' biopsy , detect more non - aggressive cancers than aggressive cancers .
multiparametric mr is a powerful imaging technique that is useful for detection of aggressive cap .
however , evaluation of the images is plagued by inter - radiologist variability .
state - of - the - art computerized learning models for multiparametric imaging are often affected either by the inaccuracy introduced by improper registration of multiparametric mr images , or by overfitting caused by the incorporation of a large set of potentially relevant imaging features @xcite . moreover ,
current computerized diagnosis systems are hindered by the computational complexity of processing a large set of features ; in fact , it is believed that only a fraction of them are important for the characterization of prostate cancer aggressiveness . to tackle these issues ,
a common trend in the literature is to compute only a few representative features from the multi parametric mr images for cap diagnosis ( e.g. statistical features @xcite , histogram based features @xcite , edge - based @xcite , texture features @xcite ) .
however , the manual selection of features is biased since the selected small feature subset captures only a restricted characteristic of the prostate cancer lesion . to the best of our knowledge ,
no systematic study exists that evaluates the performance of a wide range of distinct quantitative imaging features for efficient characterization of aggressive cancer tissue in the prostate gland .
the purpose of this study is to develop an automated multiparametric image analysis model that can recognize aggressive cap ( gleason score @xmath0 7 ) based on a set of discriminative imaging features computed from three multiparameteric mr image sequences : t2 weighted imaging , diffusion weighted imaging ( dwi ) , and apparent diffusion coefficient mapping ( adc ) .
the aims are : ( i ) perform precise selection of the most discriminative group of quantitative imaging features from a large feature space that captures a comprehensive description of the lesion and its surrounding tissue extracted from multiparametric mr images ; ( ii ) incorporate the discriminating features into a prediction model that can provide fast , efficient and reproducible diagnosis of cap .
we acquired multiparametric mri exams in 79 patients : 40 patients with aggressive cancer ( gleason score @xmath0 7 ) and 39 patients with non - aggressive cancer ( gleason score < 7 ) .
for each patient , the multiparametric mr image set includes t2-weighted ( 3.0 t , 3@xmath1 slice thickness ) images , diffusion - weighted images ( b values @xmath2 ) and apparent diffusion coefficient maps .
all the images were acquired in the axial plane , and an expert radiologist identified and circumscribed the suspicious lesion in a single slice through the largest section of the lesion on each sequence .
the histopathology result of the image - targeted biopsy using mri - ultrasound fusion was recorded for each patient , including the gleason score .
the gleason score varies from 0 - 10 , with 10 representing the most aggressive cancer and 0 representing benign lesion .
our proposed approach is shown in fig . [ fig : workflow1 ] where the key components are the following : * creation of feature matrix * - we compute a comprehensive set of quantitative imaging features from the multiparametric mr images that are potentially relevant in delineating an exhaustive characterization of a lesion in the prostate gland and the normal tissue surrounding the lesion .
table [ table : features ] shows the quantitative features that we extract , which is a broad and overlapping range of features that captures a comprehensive set of characteristics of the lesion and the surrounding tissue . in total , we extract 488 dimensional feature vector from each mr image sequence and concatenate them into a large feature matrix with 1464 features : @xmath3=\{f_{adc}[488],f_{dwi}[488],f_{t2}[488]\}$ ] , where @xmath4 $ ] , @xmath5 $ ] , @xmath6 $ ] are the features extracted from adc , dwi and t2 images . after the computation , we normalized the feature matrix @xmath3 $ ] so that each feature vector ( a column in @xmath3 $ ] ) are centered at 0 with a standard deviation of 1 .
* selection of discriminative features and train the computerized model * - we have created a prediction model based on elasticnet regularized regression @xcite that combines @xmath7 penalties of lasso and @xmath8 penalties of ridge to overcome the limitations of incorporating high dimensional features for relatively small number of samples .
based on the gleason score derived from the biopsy results , the patient cohort is divided into two groups : ( i ) aggressive - gleason score @xmath0 7 , and ( 2 ) non - aggressive - gleason score @xmath9 7 .
the biopsy based classification ground truth and the normalized feature matrix ( @xmath3 $ ] ) are the only inputs to our predictive model .
we performed shrinkage of the large quantitative feature matrix ( @xmath3 $ ] ) while maintaining the pairwise correlation between features , and selected a set of discriminative features for distinguishing aggressive from non - aggressive tumor type .
finally , training and testing of the model was conducted using 10-fold cross - validation , and we evaluated the accuracy of cap aggressiveness diagnosis by constructing a receiver operating characteristic ( roc ) curve .
the ranking of the imaging features is derived by analyzing the coefficient of the features computed by the 10 fold cross validated elastic net .
we achieved 58% sensitivity and 87% specificity at an roc operating point where the model gave the best trade off between sensitivity and specificity . without registering the multiparametric mr images
, the roc curve for the cap aggressiveness diagnosis yielded an area under the curve ( auc ) of 0.73 ( see figure [ fig:2].a . ) which is comparable with the auc value achieved by a published method @xcite that performs elastic registration between multiparametric mr images .
moreover , our model not only derives a diagnosis , but it allows the clinicians to reason on the computerized diagnosis outcome by presenting the multiparametric image feature ranking .
figure [ fig:2].b shows a trace plot of a grid of lambda values ( 100 values of lambda ) used in the model and the corresponding feature coefficient values generated by the model . by exploiting the trace plot , in figure
[ fig:2].c we present digest version of the feature selection results of the model which produced maximum auc value .
the feature acronyms are the same as in table [ table : features ] , concatenated with image sequence identifier at the front . on total
, only 44 discriminative predictors were selected from the feature matrix @xmath3 $ ] . the lesion boundary roughness feature computed from the diffusion weighted image
is ranked as the most dominant feature ( figure [ fig:2].c .
row highlighted in blue ) . lesion shape and texture extracted from
t2 mr images appeared to be the second and third most discriminative features for detection .
texture and contrast profile of the lesion computed from adc maps are ranked within top ten features , whereas the lesion contrast profile is not discriminative for t2 . in our model with multiparametric mr imaging , diffusion weighted imaging is ranked as the most informative sequence for cap aggressiveness detection , while when the diffusion weighted imaging anlyzed alone , the auc value was 0.60 .
in the paper , we propose a method for classifying the cap aggressiveness from quantitative analysis of features extracted from multiparametric mri images .
our method uses three different multiparametric mr image sequences , extracts quantitative image features , and incorporates learning from the aggressiveness classification based on the gleason score by utilizing the most discriminative features .
interestingly , the top five discriminative features are extracted from either dwi or t2 mri , and the feature with largest coefficient value , dwi lesion boundary , represents a novel feature for predicting prostate cancer aggressiveness that is not currently examined by radiologists to assess prostate lesions .
the model allows real time feature extraction as well as automatic classification for unknown cases .
our method is also advantageous in that we need not register the multiparametric mr images .
our results appear promising on our limited sized dataset .
studies in larger independent datasets would be helpful to confirm our results .
our study also was limited to analysis of 2d images , and volumetric analysis of features could improve the results . in the future
, we will extend our model by including a more comprehensive volumetric characterization of the lesion .
k. chamie , g. a. sonn , d. s. finley , n. tan , d. j. margolis , s. s. raman , s. natarajan , j. huang , and r. e. reiter , `` the role of magnetic resonance imaging in delineating clinically significant prostate cancer , '' _ urology _ , vol .
83 , no . 2 ,
pp . 369375 , 2014 .
s. wang , k. burtt , b. turkbey , p. choyke , and r. m. summers , `` computer aided - diagnosis of prostate cancer on multiparametric mri : a technical review of current research , '' _ biomed research international _ , vol .
2014 , 2014 .
p. c. vos , t. hambrock , c. a. hulsbergen - van de kaa , j. j. ftterer , j. o. barentsz , and h. j. huisman , `` computerized analysis of prostate lesions in the peripheral zone using dynamic contrast enhanced mri , '' _ medical physics _
35 , no . 3 , pp .
888899 , 2008 .
p. liu , s. wang , b. turkbey , k. grant , p. pinto , p. choyke , b. j. wood , and r. m. summers , `` a prostate cancer computer - aided diagnosis system using multimodal magnetic resonance imaging and targeted biopsy labels , '' in _ spie medical imaging_.1em plus 0.5em minus 0.4em international society for optics and photonics , 2013 , pp .
86701g86701 g .
p. tiwari , j. kurhanewicz , m. rosen , and a. madabhushi , `` semi supervised multi kernel ( sesmik ) graph embedding : identifying aggressive prostate cancer via magnetic resonance imaging and spectroscopy , '' in _ international conference on medical image computing and computer - assisted intervention_.1em plus 0.5em minus 0.4emspringer , 2010 , pp .
666673 .
e. niaf , o. rouvire , f. mge - lechevallier , f. bratan , and c. lartizien , `` computer - aided diagnosis of prostate cancer in the peripheral zone using multiparametric mri , '' _ physics in medicine and biology _ , vol .
57 , no . 12 , p. 3833
, 2012 .
s. viswanath , b. n. bloch , j. chappelow , p. patel , n. rofsky , r. lenkinski , e. genega , and a. madabhushi , `` enhanced multi - protocol analysis via intelligent supervised embedding ( empravise ) : detecting prostate cancer on multi - parametric mri , '' in _ spie medical imaging_.1em plus 0.5em minus 0.4eminternational society for optics and photonics , 2011 , pp . | we propose an automated method for detecting aggressive prostate cancer(cap ) ( gleason score @xmath07 @xcite ) based on a comprehensive analysis of the lesion and the surrounding normal prostate tissue which has been simultaneously captured in t2-weighted mr images , diffusion - weighted images ( dwi ) and apparent diffusion coefficient maps ( adc ) .
the proposed methodology was tested on a dataset of 79 patients ( 40 aggressive , 39 non - aggressive ) .
we evaluated the performance of a wide range of popular quantitative imaging features on the characterization of aggressive versus non - aggressive cap .
we found that a group of 44 discriminative predictors among 1464 quantitative imaging features can be used to produce an area under the roc curve of 0.73 . |
initially formulated by john wheeler@xcite , the program of deriving quantum formalism from information - theoretic principles has been receiving lately much attention .
thus , jozsa@xcite promotes a viewpoint which `` attempts to place a notion of information at a primary fundamental level in the formulation of quantum physics '' .
fuchs @xcite presents his program as follows : `` the task is not to make sense of the quantum axioms by heaping more structure , more definitions ... on top of them , but to throw them away wholesale and start afresh
. we should be relentless in asking ourselves : from what deep _ physical _ principles might we _ derive _ this exquisite mathematical structure ? ..
i myself see no alternative but to contemplate deep and hard the tasks , the techniques , and the implications of quantum information theory . '' in a similar fashion , rovelli@xcite distinguishes a philosophical problem of interpretation from a mathematical problem of derivation of quantum mechanical formalism from the first principles .
he writes , `` ... quantum mechanics will cease to look puzzling only when we will be able to _ derive _ the formalism of the theory from a set of simple physical assertions ( postulates , principles ) about the world .
therefore , we should not try to append a reasonable interpretation to the quantum mechanics _ formalism _ , but rather to _ derive _ the formalism from a set of experimentally motivated postulates '' .
rovelli refers to his own work as a point of view and not as interpretation : `` from the point of view discussed here , bohr s interpretation , consistent histories interpretations , as well as many worlds interpretation , are all correct '' .
rovelli s _ point of view _
, i.e. informational treatment of quantum mechanics , thus serves a formal criterion or a filter that permeates certain interpretations and not others .
in other words , treatment of quantum mechanics on information - theoretic grounds entails that some interpretations of quantum theory will be with certainty inapplicable but a number of other interpretations will all remain possible .
such a result can be naturally expected from any novel formal development of quantum theory that remains in the area of science as opposed to philosophy .
any formal derivation of quantum mechanics , in particular those using bayesian methods@xcite and quite promising for someone who believes in information - theoretic foundations of physics , requires a definite conceptual background on which such a derivation will further operate . as it is often the case , to give a rigorous axiomatic system that could provide the necessary background ,
is a difficult task .
below we analyze some three proposed solutions , by rovelli@xcite , by fuchs @xcite and by brukner and zeilinger @xcite .
elementary act of measurement is understood by rovelli as yes - no question .
brukner and zeilinger use the term `` proposition '' which generalizes the notion of binary question .
still , if one looks into where from the term `` proposition '' appears , one finds in @xcite two formulations of zeilinger s fundamental principle for quantum mechanics : fp1 : : an elementary system represents the truth value of one proposition .
fp2 : : an elementary system carries one bit of information .
it seems that zeilinger s choice of these two principles strongly suggests that the following phrase in bz reflects the view of the authors on the fundamental issue and thus puts them very close to rovelli s position : `` yes - no alternatives are representatives of basic fundamental units of all systems . ''
fuchs starts directly with the hilbert space and the full structure of quantum mechanics .
he desribes measurements not by projectors but by positive operator - valued measures .
this allows one to think that he will not agree with a definition of primitive measurements as consisting of exclusive yes - no alternatives , where the word `` exclusive '' leads to mathematically representing yes - no questions as orthogonal projectors .
still , fuchs mentions some of the basic assumptions that he makes in his derivation .
rovelli and bz each then pose two axioms .
axiom 1 : : : + * rovelli : `` there is a maximum amount of relevant information that can be extracted from a system . ''
* fuchs : does nt follow the axiomatic approach ; states that `` there is maximal information about a system . ''
* brukner and zeilinger : `` the information content of a quantum system is finite . ''
axiom 2 : : : + * rovelli : `` it is always possible to acquire new information about a system . '' * fuchs : does nt follow the axiomatic approach ; states that `` there will always be questions that we can ask of a system for which we can not predict the outcomes . ''
* brukner and zeilinger : introduce the notion of total information content of the system ; state that there exist mutually complementary propositions ; state that total information content of the system is invariant under a change of the set of mutually complementary propositions . in spite of a quite striking analogy between the axioms chosen by different authors , as for the following derivation of quantum mechanics , they do not proceed in the same manner .
we shall now have a closer look at the axioms and derivation techniques .
axiom 1 marks a crucial point of departure from classical physics .
newtonian physics employs mathematics of continuum to represent the world and , therefore , any calculation of complete information about , say , a particle position would require an infinitely long computation .
this fact has profoundly disturbed many physicists , with most prominently feynman saying , `` it always bothers me that , according to the laws as we understand then today , it takes a computing machine an infinite number of logical operations to figure out what goes on in no matter how tiny a region of space and no matter how tiny a region of time , ... why should it take an infinite amount of logic to figure out what one tiny piece of space - time is going to do ? ''
axiom 1 also goes in line with wheeler s `` no continuum '' principle@xcite .
while it seems intuitively plausible to accept axiom 1 , its interpretation is not straightforward .
each author imposes his own interpretation by choosing a suitable translation into his language ; for rovelli it means that there exist complete sets of yes - no questions that could provide one , abstractly , with complete information about a system ; for brukner and zeilinger it means that everything that is there to a system is represented by a complete set of mutually complementary propositions .
however , it appears that axiom 1 can raise yet a different issue .
our intuition is that essential finiteness applies , not to the system to which we address yes - no questions but to the system plus the observer who asks these questions .
formal development of this idea will appear in section [ sect3 ] .
philosophical argument goes as follows : it is not true , that in order to know a newtonian coordinate we need the knowledge of infinitely many decimal digits .
the latter should not make us worry , for we are endowed with an ability to create a special code ( a new concept ) , which will substitute in the thinking the undesired infinity .
the same works for computation : feynman s argument from the infinity of logical operations must include the possibility of `` hiding '' newtonian infinities under the `` and so forth '' concept , for which one may specify operational rules . consequently , the requirement of finiteness applies to the observer - observed system . in a similar manner , in bz terms , essential finiteness applies to system plus the one who chooses the propositions to be tested on the system , i.e. the observer .
we do not have the intuition that `` one can not know the infinity '' but , rather , that one can not have infinite knowledge .
how is this understanding of the finiteness axiom related to the one adopted by rovelli and other authors ? with the assumption of universality of quantum theory , one can deduce the old point of view .
indeed , universality allows us to treat the border line between theory and meta - theory in quantum mechanics as flexible @xcite . any given observer ( meta - theoretical entity ) can be included in the theory proper by taking the point of view of an observer external to the one in question .
if the amount of information remains finite in spite of the arbitrary choice of the frontier between the system and the observer , we can eliminate from consideration any previously given observer at the price of redefining the question - asking party .
imagine , for example , a computer solving in maple software some field theory renormalization problem .
renormalization is about removing infinities , so if a computer were let to solve this task without conceptualizing infinities by means of a previously learned renormalization technique , it would have never arrived at any result .
abstract amount of information in the system is infinite in this example , but the amount of relevant information is finite .
what is relevant , decides the observer who translates relevancy into concepts that he employs for the computation on a system , or equivalently into a specific manner to ask some yes - no questions and not others ( see section [ sect44 ] ) .
going to the extreme of definitions , what is information and what is not decides the observer , and it is because of this that the amount of information is finite .
had we had the liberty to call information anything we want , there would be no intuitively clear argument showing why this `` anything '' must be finite .
finiteness thus has to include the observer , and thanks to the universality of quantum mechanics can be in the limit reduced to finiteness in the sense of rovelli .
axiom 2 beautifully corresponds to wheeler s dictum ( adopted after philip anderson ) `` more is different '' .
if one wants to get more information , this will be different information ; or one can always get more information , and it will be different information .
though in the original `` more is different '' was used in the context of complexity theory , it can as well , as a basic principle , apply to information - based quantum theory . at it
will be seen below , for rovelli axiom 2 allows to introduce probabilities and to deduce an analogue of the born rule ; for bz axiom 2 leads to imposing a certain structure on the information space .
this axiom is responsible for the departure from classicality , which is not yet fully accommodated by axiom 1 alone .
fuchs uses as _ a priori _ given the structure of hilbert space ; his task is to deduce some of the operational structure of quantum mechanics , namely , density matrices .
rovelli , on the contrary , is interested in deducing quantum mechanics from the axioms and does not show a way to deduce most of further structure , to start with the superposition principle ( apart from introducing it as an axiom ) .
fuchs uses a decision - theoretic ( bayesian ) approach to derive the superposition principle .
he refers to rovelli s paper in his own , and one is left free to suggest that many of his axiomatic assumptions , on which he does nt clearly comment , might be similar to rovelli s ones , apart from the key issue of how to define measurement .
indeed , fuchs insists on the fundamental character of positive operator - valued measures .
this may not seem intuitively evident .
but because fuchs leaves the axiomatic foundations of the bayesian approach open , even if we dismiss the necessity to define measurement as povm , there still remains an opportunity to introduce the latter in the theory .
povm have a natural description as conventional von neumann measurements on an ancilla system@xcite , and thus to rovelli s axiomatic derivation of the hilbert space structure one may try to add an account of inevitability of ancilla systems and naturally obtain from this the needed povm description .
this will be attempted in section [ sect3 ] .
brukner and zeilinger proceed differently . if information is primary , they argue , then any formalism should be a formalism dealing with information and not with some other notions .
therefore bz construct an _ information space _ where they apply the axioms and use the formalism to deduce testable predictions .
bz do not refer to physical space or to hilbert space in their construction .
thus they do not have access to algebra allowing a reconstruction of the state space out of the operator hilbert space .
therefore , because of this change of scenery , they are bound to postulate more properties of mutually complementary propositions than rovelli or fuchs .
namely , they postulate the homogeneity of parameter space .
bz s self - imposed terminological limitation to abstract information space does not seem viable for philosophical , i.e. extra - scientific , reasons : it renders the formalism less transparent in use , while introduction of supplementary axioms does not make it conceptually clearer than traditional formalisms . to continue
, the question is how to extract a useful approach from the juxtaposition of rovelli s and fuchs s proposals .
rovelli , as said before , shows a way to construct the hilbert space structure from two axioms . unlike brukner and zeilinger
, we do not call this hilbert space information space but simply physical space , for there is no other space in the whole construction that would be the physical space .
bz s information space is what the physical space _ _
is__. next , with hilbert space in hand , we use fuchs s derivation based on gleason s theorem to deduce density matrices and their properties .
this requires essentially one more step : we need to introduce povm as measurements , as discussed above .
once we ve introduced ancilla systems , we can operationally redefine measurement as described by povm . to move further in combining rovelli s and fuchs s proposals , after the axiomatic stage , we either need a sort of decision - theoretic approach to derive the formal consequences of the necessary intersubjective accord of measurement results ( for fuchs , for example , via a version of the de finetti theorem ) , or we need to use an algebraic approach so that the constructed hilbert space be treated as space of operators corresponding to observables .
this latter option will be investigated below .
we are guided by the computer metaphor .
indeed , the strategic task is to give a reformulation of quantum theory in information - theoretic terms . a theory that operates with the notion of information
can be compared to software as opposed to a theory that operates with the notion of energy which can be compared to hardware .
ideally one would wish to see all `` hardware '' or energetic language disappear from the formulation of the theory , so that only `` software '' or informational language remain .
we are usually interested in information about ( knowledge of ) the chosen system and we disregard particular ways in which we have obtained this information .
all that counts is knowledge that can be useful in future or , in other words , relevant knowledge or relevant information .
this is why one usually does not pay attention to the very process of interaction between the system being measured and the measuring system , and one treats measuring system as a meta - theoretic , i.e. non - physical , apparatus . to give an example , for some experiment a physicist may need to know the proton mass but he will not at all be interested in how this quantity was measured , unless he is a narrow specialist whose interest is in measuring particle masses .
particular ways to gain knowledge are irrelevant , while knowledge itself is highly relevant and useful .
some of the experiments where one is interested in the measurement as a physical process are discussed in@xcite . from now on we assume that measurement details are irrelevant , perhaps at the price of redefining what is measurement . in a practical setting , though , information is _ always _
this is to say that there always is some physical support of information , some hardware .
the necessity of the physical support requires that we proceed in the following manner : first , treat the measurement interaction as physical ; then , disregard the fact that it was physical and reformulate the theory in terms of measurement results only .
to start , make a distinction between two parts of the world : quantum system @xmath0 , which is the system of interest , and the observer .
the observer , in the spirit of the software - hardware metaphor , consists of an informational agent ( `` i - observer '' ) and of the physical realization of the observer ( `` p - observer '' ) .
there is no i - observer without p - observer .
reciprocally , there is no sense in calling p - observer an observer unless there is i - observer ( otherwise p - observer is just a physical system as any ) .
hence , the two components of the `` larger observer '' are not in any way separate or orthogonal to each other ; on the contrary , these are merely two viewpoints , and the difference is but descriptive .
p - observer interacts with the quantum system and thus provides for the physical basis of measurement .
i - observer is only interested in the measurement result , i.e. information per se , and he gets information by reading it from p - observer .
the act of reading or getting information is here a common linguistic expression and not a physical process since i - observer and p - observer are not physically distinct .
in fact , the concept of `` being physical '' only applies to p - observer , and by definition the physical content of the `` larger observer '' is all contained in p - observer .
i - observer as informational agent is meta - theoretic , and hence the fact that its interaction with p - observer , or the act of `` reading information '' , is unphysical . to give a mathematical meaning to this act ,
we assume that getting information is described as yes - no questions asked by i - observer to p - observer . to follow rovelli s construction@xcite
, the set of questions will be denoted @xmath1 . according to axiom 1
, there is a finite number @xmath2 that characterizes p - observer s informational capacity .
the number of questions in @xmath3 , though , can be much larger than @xmath2 , as some of these questions are not independent .
in particular , they may be related by implication ( @xmath4 ) , union ( @xmath5 ) , and intersection ( @xmath6 ) .
one can define an always false ( @xmath7 ) and an always true question ( @xmath8 ) , negation of a question ( @xmath9 ) , and a notion of orthogonality as follows : if @xmath10 , then @xmath11 and @xmath12 are orthogonal ( @xmath13 ) .
equipped with these structures , and under the non - trivial assumption that union and intersection are defined for every pair of questions , @xmath14 is an orthomodular lattice .
one needs to make a few more steps to obtain the hilbert space structure .
as follows from axiom 1 , one can select in @xmath14 a set @xmath15 of @xmath2 questions that are independent from each other .
in the general case , there exist many such sets @xmath15 , @xmath16 , etc .
if i - observer asks the @xmath2 questions in the family @xmath15 then the obtained answers form a string @xmath17_c.\ ] ] this string represents the `` raw '' information that i - observer got from p - observer as a result of asking the questions in @xmath15 .
note that this is not yet information about the quantum system @xmath0 that the i - observer ultimately wants to have , but only a process due to functional separation within the `` larger observer '' .
the string @xmath18 can take @xmath19 values and , since these outcomes are by construction mutually exclusive , we can define new questions @xmath20 such that the yes answer to @xmath21 corresponds to the string of answers @xmath22 : @xmath23 to these questions we refer as to `` complete questions '' . by taking all possible unions of sets of complete questions
@xmath21 of the same family @xmath15 one constructs a boolean algebra that has @xmath21 as atoms .
alternatively , one can consider a different family @xmath16 of n independent yes - no questions and obtain another boolean algebra with different complete questions as atoms .
it follows , then , from axiom 1 that the set of questions @xmath14 that can be asked to p - observer is algebraically an orthomodular lattice containing subsets that form boolean algebras .
this is precisely the algebraic structure formed by the family of linear subsets of hilbert space .
it is interesting to note that in approaches that start with an abstract @xmath24-algebra of operators one needs to use the gelfand - naimark - segal construction to obtain a representation of this algebra as algebra of operators on a hilbert space . in the present approach ,
information - theoretic axioms are evoked to obtain a similar result , namely , to show that operators form a hilbert space . from the second rovelli s axiom
it follows immediately that there are questions such as answers to these questions are not determined by @xmath18 .
define , in general , as @xmath25 the probability that a yes answer to @xmath26 will follow from the string @xmath22 .
given two complete strings of answers @xmath18 and @xmath27 , we can then consider the probabilities @xmath28 from the way it is defined , the @xmath29 matrix @xmath30 can not be completely arbitrary .
first , we must have @xmath31 then , if information @xmath32 is available about the system , one and only one of the outcomes @xmath33 may result .
therefore @xmath34 if we assume that @xmath35 then we also get @xmath36 if pursued further in an attempt to deduce probability amplitudes , this derivation , however , encounters some difficulties .
to get the result , rovelli postulates explicitly the superposition principle .
we , too , introduce a new assumption to obtain more of the structure of quantum theory .
namely , we postulate non - contextuality and use gleason s theorem to deduce density matrices .
it remains an open question if non - contextuality as an intuitively made assumption is welcome or must be rejected as too strong@xcite . in mathematical terms , it states that probabilities can be defined for a projector independently of the family of projectors of which it is a member , or that in @xmath37 with fixed @xmath38 probability will be the same had the fixed question belonged not to the family @xmath39 but to some other family @xmath16 .
one can then prove a theorem due to gleason@xcite : * theorem ( gleason ) * _ let @xmath40 be any function from 1-dimensional projections on a hilbert space of dimension @xmath41 to the unit interval , such that for each resolution of the identity @xmath42 then there exists a unique density matrix @xmath43 such that @xmath44 _
one last step before we move to quantum theory of the system @xmath0 is to obtain unitary dynamics .
following rovelli , any question can be labelled by the time variable @xmath45 indicating the time at which it is asked .
denote as @xmath46 the one - parameter family of questions defined by the same procedure performed at different times .
assume that time evolution is a symmetry in the theory .
in the context of our approach the latter word `` theory '' includes theory of p - observer _ and _ of the quantum system @xmath0 . then recall that the set @xmath14 has the structure of a set of linear subspaces in the hilbert space , and the set of all questions at time @xmath47 to the p - observer part of the physically interacting conjunction of two systems , must be isomorphic to the set of all questions at time @xmath48 .
therefore , the corresponding family of linear subspaces must have the same structure ; it follows that there must be a unitary transformation @xmath49 such that @xmath50 it is straightforward to see that these unitary matrices form an abelian group and @xmath51}$ ] , where @xmath52 is a self - adjoint operator in the hilbert space , the hamiltonian . in a practical
setting , it is from the past or the future of a given experiment , in particular from the intentions of the experimenter , that one can learn which information about the experiment is relevant and which is not .
what is relevant can either be encoded in the preparation of the experiment or selected by the experimenter _ a posteriori_. in all cases ,
the notion of relevance does not enter into the formalism which solely describes the measurement within the context of the experiment .
all that is `` allowed to be known '' inside the formal framework is that there ( a ) _ is _ ( b ) _ some _ relevant information .
what is relevant is reflected in the choice of questions that are asked by i - observer . interaction between p - observer and the quantum system
should be viewed as physical interaction between just any two physical systems . still ,
because i - observer then reads information from p - observer and because we are nt interested in the posteriority of relations between p - observer and the quantum system , we can treat p - observer as an ancillary system in course of its interaction with @xmath0 .
such an ancillary system would have interacted with @xmath0 and then would be subject to a standard measurement described mathematically on its hilbert space via a set of orthogonal yes - no projection operators .
so far , for p - observer we have the hilbert space and the standard born rule . the fact that p - observer is treated as ancillary system allows to transfer some of this structure on the quantum system @xmath0 .
a new non - trivial assumption has to be made , that the time dynamics that has previously arisen in the context of p - observer alone , also applies to the i - observer and to @xmath0 . in other words , there is only one time in the system .
time of i - observer is the one in which one can grasp the meaning of the words `` past '' and `` future '' as used above in relation with the experimental setting and the notion of relevant information : it is in this time that there is a `` before the experiment '' and an `` after the experiment '' .
times of physical systems , such as @xmath0 or p - observer , are times in which their dynamics takes place .
now , both the physical interaction of p - observer with @xmath0 and the process of asking questions by i - observer to p - observer take place in one and the same time . since ( a ) until i - observer asks the question that he chooses to ask , sets of questions at different times are isomorphic and evolution is unitary , and ( b ) time at which i - observer asks the question only depends on i - observer and considerations of relevance that must not enter into the formalism
, then one concludes that the interaction between the quantum system and p - observer must respect the unitary character all until the decoupling of the ancilla .
now write , @xmath53 after asking a question corresponding to a projector @xmath54 , probability of the yes answer will be given by @xmath55 because the systems decouple , trace can be decomposed into @xmath56 where all presence of the ancilla is hidden in the operator @xmath57 which acts on the quantum system @xmath0 alone .
this operator is positive - semidefinite , and a family of such operators form resolution of identity .
they are not , however , mutually orthogonal .
such operators form positive operator - valued measures ( povm)@xcite .
we have shown how to obtain a description of quantum measurement via povm at the condition of disregarding completely the physical interaction during measurement and the existence of p - observer .
if one is only interested in a formal description of how i - observer acquires information about the quantum system @xmath0 , this is done via povm and the born rule following from gleason s theorem . to be mentioned here , gleason s theorem also admits a generalization from von neumann s orthogonal projector measures to povm @xcite .
one gets therefore a description of measurement as used in quantum information theory , and one can now continue the development of the theory in the conventional way@xcite . in agreement with
the intuition expressed in the key metaphor , all `` hardware '' language is eliminated and the theory can be formulated in the `` software '' language alone . formal deduction of the results concerning the hilbert space , however , was not completely rigorous .
rovelli@xcite acknowledges it in his disclaimer , `` i do not claim any mathematical nor philosophical rigor '' .
indeed , the fact that yes - no questions form an orthomodular lattice containing subsets that form boolean algebras only commits one to the structure of union of hilbert spaces and not of a single hilbert space . thus , this can happen to be the union of primitive hilbert spaces , which allow for a classical and not a quantum interpretation .
generally speaking , the structure will be the one of the hilbert space with superselection rules .
one needs then to use axiom 2 to show that the possibility to ask in every situation some new informative question excludes classicality .
completion of this program remains an open problem .
introduction of space and time `` by hand '' in any algebraic approach to quantum mechanics is certainly quite unsatisfactory . one would wish to see how space and time arise naturally from the formalism
this may require a fully rigorous algebraic approach involving von neumann algebras and the gns construction for the hilbert space .
however , the author is only aware of one way to introduce time in this framework@xcite .
this leaves open the question of link between time and space , and of the possibility to use the two notions together to obtain the evolution equation .
the author would like to thank carlo rovelli and christopher fuchs for discussion and suggestions .
10 j.a .
wheeler , _ information , physics , quantum : the search for links _
, in anthony j.g .
hey , editor , _ feyman and computation : exploring the limits of computers .
_ , perseus books , reading , massachusets , 1998 . | information - theoretic derivations of the formalism of quantum theory have recently attracted much attention .
we analyze the axioms underlying a few such derivations and propose a conceptual framework in which , by combining several approaches , one can retrieve more of the conventional quantum formalism . |
a major obstacle to a symplectic theory of finite dimensional integrable hamiltonian systems is that _ differential topological _ and _ symplectic _ problems appear side by side , but smooth and symplectic methods do not always mesh well .
morse - bott theory represents a success in bringing together in a cohesive way continuous and differential tools , and it has been used effectively to study properties of dynamical systems . but incorporating symplectic information into the context of dynamical systems is far from automatic . however , many concrete examples are known for which computations , and numerical simulations , exhibit a close relationship between the symplectic dynamics of a system , and the differential topology of its bifurcation set . in the 1980s and 1990s , the fomenko school developed a morse theory for regular energy surfaces of integrable systems . moreover ,
theoretical successes ( in any dimension ) for compact periodic systems in the 1970s and 1980s by atiyah , guillemin , kostant , sternberg and others , gave hope that one can find a mathematical theory for bifurcations of integrable systems in the symplectic setting .
this paper develops a symplectic bifurcation theory for integrable systems in dimension four compact or not .
because it recently has become clear that the bifurcation diagram of an integrable system is the natural setting to study symplectic invariants ( see for instance @xcite ) , this paper provides a setting for the study of quantum integrable systems .
semiclassical quantization is a strong motivation for developing a systematic bifurcation theory of integrable systems ; the study of bifurcation diagrams is fundamental for the understanding of quantum spectra @xcite .
moreover , the results of this paper may have applications to mirror symmetry and symplectic topology because an integrable system without hyperbolic singularities gives rise to a toric fibration with singularities .
the base space is endowed with a singular integral affine structure .
these singular affine structures are studied in symplectic topology , mirror symmetry , and algebraic geometry ; for instance , they play a central role in the work of kontsevich and soibelman @xcite .
we refer to section [ sec : remarks ] for further analysis of these applications , as well as a natural connection to the study of solution sets in real algebraic geometry .
the development of the theory requires the introduction of methods to construct morse - bott functions which , from the point of view of symplectic geometry , behave well near the singularities of integrable systems .
these methods use eliasson s theorems on linearization of non - degenerate singularities of integrable systems , and the symplectic topology of integrable systems , to which many have contributed .
the first part of this paper is concerned with the connectivity of joint level sets of vector - valued maps on manifolds , when these are defined by the components of an integrable system .
the most striking previous result in this direction is atiyah s 1982 theorem which guarantees the connectivity of the fibers of the momentum map when the integrable system comes from a hamiltonian torus action .
the second part of the paper explains how the pioneering results proven in the seventies and eighties by atiyah , guillemin , kostant , kirwan , and sternberg describing the image of the momentum of a _ hamiltonian compact group action _ also hold in the context of integrable systems on four - dimensional manifolds , when there are no hyperbolic singularities .
the conclusions of the theorems in this paper are essentially optimal .
moreover , there is only one transversality assumption on the integrable system : that there should be no vertical tangencies on the bifurcation set , up to diffeomorphism .
if this condition is violated then there are examples which show that one can not hope for any fiber connectivity . the work of atiyah , guillemin , kostant , kirwan , and sternberg exhibited connections between symplectic geometry , combinatorics , representation theory , and algebraic geometry .
their work guarantees the convexity of the image of the momentum map ( intersected with the positive weyl chamber if the compact group is non - commutative ) .
although this property no longer holds for general integrable systems , an explicit description of the image of the singular lagrangian fibration given by an integrable system with two degrees of freedom can be given .
the understanding of this image , which corresponds to the bifurcation diagram of the dynamics in the physics literature , is essential for the description of the system , as it has proven to be the best framework to define new symplectic invariants of integrable systems , and hence to quantize ; see the recent work on semitoric integrable systems @xcite .
the most striking known result for fiber connectivity of vector - valued functions is atiyah s famous connectivity theorem @xcite proved in the early eighties .
suppose that @xmath0 is a compact , connected , symplectic , @xmath1-dimensional manifold . for smooth functions @xmath2 ,
let @xmath3 be the flow of the hamiltonian vector field @xmath4 , where @xmath4 is defined by the equation @xmath5 .
we denote by @xmath6 the @xmath7-dimensional torus @xmath8 .
suppose that @xmath9 , where @xmath10 , defines a @xmath6-action on @xmath11 .
then the fibers of the map @xmath12 are connected .
this theorem has been generalized by a number of authors to general compact lie groups actions and more general symplectic manifolds .
indeed , a hamiltonian @xmath13-torus action on a @xmath1-manifold may be viewed as a very particular integrable system .
a long standing question in the integrable systems community has been to what extent atiyah s result holds for integrable systems , where checking fiber connectivity by hand is extremely difficult ( even for easily describable examples ) .
this paper answers this question in the positive in dimension four : if an integrable system has no hyperbolic singularities and its bifurcation diagram has no vertical tangencies , then the fibers of the integrable system are connected . to state this result precisely , recall that a map @xmath14 is a _ integrable hamiltonian system _ if @xmath15 are point - wise almost everywhere linearly independent and for all indices @xmath16 , the function @xmath17 is invariant along the flow of the hamiltonian vector field @xmath18 .
( recall : @xmath4 is the vector field defined by @xmath5 ) . in general , fiber connectivity is no longer true for integrable systems , due to the existence of singularities .
for instance , consider the manifold is @xmath19 . choose the following coordinates on @xmath20 : @xmath21 $ ] and @xmath22 .
choose coordinates @xmath23 on @xmath24 and let @xmath25 be the symplectic form on @xmath11 , where @xmath7 is a positive integer .
the map @xmath26 , defines an integrable system .
the fiber over any regular value of @xmath27 is @xmath7 copies of @xmath28 .
the image @xmath29 is an annulus ( see figure [ fig : annulus - only ] ) . of integrable system with disconnected fibers , on the compact manifold @xmath30 .
this integrable system has only singularities of elliptic and transversally elliptic type , so it is remarkable how while being quite close to be a toric system , fiber connectivity is lost.,width=170,height=170 ] an annulus has the property that it has vertical tangencies , and it can not be deformed into a domain without such tangencies .
remarkably , if such tangencies do not exist in some deformation of @xmath29 , fiber connectivity still holds .
this is for instance the case for hamiltonian torus actions .
next we state this precisely . in this paper , manifolds are assumed to be @xmath31 and second countable .
let us recall here some standard definitions .
a map @xmath32 between topological spaces is _ proper _ if the preimage of every compact set is compact .
let @xmath33 , @xmath34 be smooth manifolds , and let @xmath35 .
a map @xmath36 is said to be _ smooth _ if at any point in @xmath37 there is an open neighborhood on which @xmath38 can be smoothly extended .
the map @xmath38 is called a _
diffeomorphism onto its image _ when @xmath38 is injective , smooth , and its inverse @xmath39 is smooth . if @xmath33 and @xmath34 are smooth manifolds , the _ bifurcation set _
@xmath40 of a smooth map
@xmath41 consists of the points of @xmath33 where @xmath38 is not locally trivial ( see definition [ def : bifurcation ] ) . it is known that the set of critical values of @xmath38 is included in the bifurcation set and that if @xmath38 is proper this inclusion is an equality ( see ( * ? ? ?
* proposition 4.5.1 ) and the comments following it ) .
second , recall that an integrable system @xmath42 is called _ non - degenerate _ if its singularities are non - degenerate ( see definition [ def : nondegpoi ] ) .
if @xmath27 is proper and non - degenerate , then @xmath43 is the image of a piecewise smooth immersion of a 1-dimensional manifold ( proposition [ prop : strata ] ) .
we say that a vector in @xmath44 is _ tangent to @xmath43 _ whenever it is directed along a left limit or a right limit of the differential of the immersion .
we say that the curve @xmath45 has a _ vertical tangency _ at a point @xmath46 if there is a vertical tangent vector at @xmath46 .
our first main result is the following .
[ theo : main - connectivity0 ] suppose that @xmath47 is a compact connected symplectic four - manifold .
let @xmath48 be a non - degenerate integrable system without hyperbolic singularities .
denote by @xmath43 the bifurcation set of @xmath27 .
assume that there exists a diffeomorphism @xmath49 onto its image such that @xmath50 does not have vertical tangencies ( see figure [ fig : diffeo ] ) .
then @xmath27 has connected fibers . of @xmath27 consists precisely of the boundary points in the left figure ( which depicts @xmath29 ) .
the diffeomorphism @xmath51 transforms @xmath29 to the region on the right hand side of the figure , in order to remove the original vertical tangencies on @xmath43.,scaledwidth=50.0% ] if @xmath52 in theorem [ theo : main - connectivity ] is the momentum map of a hamiltonian @xmath53-torus action then @xmath54 .
this is no longer true for general integrable systems ; the simplest example of this is the spherical pendulum , which has a point in the bifurcation set in the interior of @xmath29 .
we denote by @xmath55 the cone in figure [ fig : cone ] , i.e. , the intersection of the half - planes defined by @xmath56 and @xmath57 on the plane @xmath58 .
this cone will be called _ proper _ , if @xmath59 , @xmath60 .
theorem [ theo : main - connectivity ] can be extended to non - compact manifolds as follows .
[ theo : main - connectivity ] suppose that @xmath47 is a connected symplectic four - manifold .
let @xmath61 be a non - degenerate integrable system without hyperbolic singularities such that @xmath27 is a proper map .
denote by @xmath43 the bifurcation set of @xmath27 .
assume that there exists a diffeomorphism @xmath62 onto its image such that : 1 .
the image @xmath63 is included in a proper convex cone @xmath64 ( see figure [ fig : cone ] ) ; 2 .
the image @xmath50 does not have vertical tangencies ( see figure [ fig : diffeo ] ) .
then @xmath27 has connected fibers .
note that theorem [ theo : main - connectivity ] clearly implies theorem [ theo : main - connectivity0 ] .
lies in the convex cone @xmath55 and has no vertical tangencies .
see condition 1 in theorem [ theo : main - connectivity].,scaledwidth=30.0% ] the assumption in theorem [ theo : main - connectivity ] is optimal in the following sense : if there exist vertical tangencies then the system can have disconnected fibers ( examples [ ex0 ] , [ ex2 ] ) ; see theorem [ theo : striking ] .
other than these exceptions we do not know of an integrable system with disconnected fibers and which does not violate our assumptions .
further in this article , we introduce a weaker transversality condition that allows us to deal with some cases of vertical tangencies . using this condition and theorem [ theo : main - connectivity ]
, we will prove the following .
[ theo : striking ] suppose that @xmath47 is a compact connected symplectic four - manifold .
let @xmath61 be a non- degenerate integrable system without hyperbolic singularities .
assume that 1 .
the interior of @xmath29 contains a finite number of critical values ; 2 .
there exists a diffeomorphism @xmath51 such that @xmath63 is either a disk , a disk with a conic point , a disk with two conic points , or a compact convex polygon ( see figure [ fig:4images ] )
. then the fibers of @xmath27 are connected . here
a neighborhood of a _ conic point _ is by definition locally diffeomorphic to some proper cone @xmath55 .
atiyah proved his connectivity theorem @xcite simultaneously with the so called convexity theorem of atiyah , guillemin , and sternberg @xcite ; it is one of the main results in symplectic geometry .
their convexity theorem describes the image of the momentum map of a hamiltonian torus action . altogether , this result generated much subsequent research , in particular it led kirwan to prove a remarkable non - commutative version @xcite .
suppose that @xmath0 is a compact , connected , symplectic , @xmath1-dimensional manifold .
for smooth functions @xmath2 , let @xmath3 be the flow of the hamiltonian vector field @xmath4 , where @xmath4 is defined by the equation @xmath5 .
we denote by @xmath6 the @xmath7-dimensional torus @xmath8 .
suppose that @xmath9 , where @xmath10 , defines a @xmath6-action on @xmath11 .
then the image of @xmath65 is a convex polytope .
an important paper prior to the work of atiyah , guillemin , kirwan , and sternberg dealing with convexity properties in particular instances is kostant s @xcite , who also refers to preliminary questions of schur , horn and weyl .
these convexity results were used by delzant @xcite in his classification of _ symplectic toric manifolds_. all together , these papers revolutionized symplectic geometry and its connections to representation theory , combinatorics , and complex algebraic geometry . for a detailed analysis of symplectic toric manifolds in the context of complex algebraic geometry ,
see duistermaat - pelayo @xcite . in this paper
we prove the natural version of the atiyah - guillemin - sternberg convexity theorem in the context of integrable systems . before stating it we recall that the _ epigraph @xmath66 of a map @xmath67 _ consists of the points lying on or above its graph , i.e. , the set @xmath68 .
similarly , the _ hypograph @xmath69 of a map @xmath70 _ consists of the points lying on or below its graph , i.e. , the set @xmath71 .
[ theo : main - image0 ] suppose that @xmath47 is a compact connected symplectic four - manifold .
let @xmath52 be a non - degenerate integrable system without hyperbolic singularities .
denote by @xmath43 the bifurcation set of @xmath27 .
assume that there exists a diffeomorphism @xmath62 onto its image such that @xmath50 does not have vertical tangencies ( see figure [ fig : diffeo ] ) .
then : * the image @xmath29 is contractible and the bifurcation set can be described as @xmath72 where @xmath73 is a finite set of rank @xmath74 singularities which is contained in the interior of @xmath29 ; * let @xmath75 and let @xmath76 $ ] .
then the functions @xmath77 \to { { \mathbb r}}$ ] defined by @xmath78 and @xmath79 are continuous and @xmath29 can be described as @xmath80 .
figure [ fig : fig1 ] shows a possible image @xmath29 , as described in theorem [ theo : main - image0 ] . in the case on non - compact manifolds we have the following result .
[ theo : main - image ] suppose that @xmath47 is a connected symplectic four - manifold .
let @xmath52 be a non - degenerate integrable system without hyperbolic singularities such that @xmath27 is a proper map .
denote by @xmath43 the bifurcation set of @xmath27 .
assume that there exists a diffeomorphism @xmath81 onto its image such that : 1 .
the image @xmath63 is included in a proper convex cone @xmath82 ( see figure [ fig : cone ] ) ; 2 .
the image @xmath50 does not have vertical tangencies ( see figure [ fig : diffeo ] ) .
equip @xmath83 with the standard topology . then : * the image @xmath29 is contractible and the bifurcation set can be described as @xmath84 where @xmath73 is a countable set of rank zero singularities which is contained in the interior of @xmath29 ; * let @xmath75 .
then the functions @xmath85 defined on the interval @xmath86 by @xmath87 and @xmath88 are continuous and @xmath29 can be described as @xmath89 .
note that theorem [ theo : main - image ] clearly implies theorem [ theo : main - image0 ] .
the rest of this paper is devoted to proving theorem [ theo : main - connectivity ] , theorem [ theo : striking ] and theorem [ theo : main - image ] .
+ + * acknowledgements . *
we thank denis auroux , thomas baird , and helmut hofer for enlightening discussions .
the authors are grateful to helmut hofer for his essential support that made it possible for tsr and vns to visit ap at the institute for advanced study during the winter and summer of 2011 , where a part of this paper was written .
additional financial support for these visits was provided by washington university in st louis and by nsf .
the authors also thank msri , the mathematisches forschungsinstitut oberwolfach , washington university in st louis , and the universit de rennes i for their hospitality at different stages during the preparation of this paper in 2009 , 2010 , and 2011 .
ap was partly supported by an nsf postdoctoral fellowship , nsf grants dms-0965738 and dms-0635607 , an nsf career award , a leibniz fellowship , spanish ministry of science grant mtm 2010 - 21186-c02 - 01 , and by the csic ( spanish national research council ) .
was partly supported by a msri membership and swiss nsf grant 200020 - 132410 .
vns was partly supported by the nonaa grant from the french anr and the institut universitaire de france .
+ + the paper combines a number of results to arrive at the proofs of these theorems .
the following diagram describes the structure of the paper .
@xmath90 \ar @/_1.5cm/ [ ddd ] \ar[d ] \ar[rddd ] & & & \\
\text{\bf theorem~\ref{theo : interior } } \ar @/^1.7cm/ [ ddddd ] & \text{proposition~\ref{prop : strata } } \ar[d ] & & \\ & \text{\bf theorem~\ref{prop : bott } } \ar[d ] & & & \\
\text{lemma~\ref{lemm : semibounded } } \ar[d ] & \text{proposition~\ref{prop : bott - index } } \ar[d ] \ar[dddrr ] & & \\
\text{proposition~\ref{prop : levelset } } \ar @/_1.2cm/ [ dd ] & \text{proposition~\ref{prop : j - connected } } \ar[dd ] \ar @/^1cm/ [ dddd ] & & \\
\text{lemma~\ref{tricky } } \ar[d ] & & & \\
\text{\bf theorem~\ref{theo : fibers } } \ar[r ] & \text{\bf theorem~\ref{theo : main - connectivity } } \ar[r ] & \text{\bf theorem~\ref{theo : main - connectivity0 } } \ar[r ] & \text{\bf theorem~\ref{theo : striking}}\\ \text{lemma~\ref{lemma : critical - set } } \ar[d ] & & & \\
\text{\bf theorem~\ref{imagemomentum : theorem } } \ar[r ] & \text{\bf theorem~\ref{theo : main - image } } \ar[r ] & \text{\bf theorem~\ref{theo : main - image0 } } } \ ] ]
in this section we prove some basic results that we need in of section [ fibers : sec ] and section [ sec3 ] .
let @xmath47 be a connected symplectic 4-manifold .
a smooth map @xmath52 is _ toric _ if there exists an effective , integrable hamiltonian @xmath91-action on @xmath11 whose momentum map is @xmath27 .
it was proven in @xcite that if @xmath27 is a proper momentum map for a hamiltonian @xmath91-action , then the fibers of @xmath27 are connected and the image of @xmath27 is a rational convex polygon
. we shall be interested in maps @xmath92 that are not toric yet retain enough useful topological properties . in the analysis carried out in the paper
we shall need the concept of non - degeneracy in the sense of williamson of a smooth map from a 4-dimensional phase space to the plane .
[ def : nondegpoi ] suppose that @xmath47 is a connected symplectic four - manifold .
let @xmath93 be an integrable system on @xmath47 , and @xmath94 a critical point of @xmath27 .
if @xmath95 , then @xmath13 is called _ non - degenerate _ if the hessians @xmath96 span a cartan subalgebra of the symplectic lie algebra of quadratic forms on the tangent space @xmath97 .
if @xmath98 one can assume that @xmath99 .
let @xmath100 be an embedded local @xmath53-dimensional symplectic submanifold through @xmath13 such that @xmath101 and @xmath102 is transversal to @xmath103 .
the critical point @xmath13 of @xmath27 is called _ transversally non - degenerate _ if @xmath104 is a non - degenerate symmetric bilinear form on @xmath105 .
one can check that definition [ def : nondegpoi ] does not depend on the choice of @xmath106 .
the existence of @xmath106 is guaranteed by the classical hamiltonian flow box theorem ( see e.g. , ( * ? ? ?
* theorem 5.2.19 ) ; this result is also called the darboux - caratheodory theorem ( * ? ? ? * theorem 4.1 ) ) .
it guarantees that the condition @xmath107 ensures the existence of a symplectic chart @xmath108 on @xmath11 centered at @xmath13 , i.e. , @xmath109 , such that @xmath110 and @xmath111 .
therefore , since @xmath112 , @xmath106 can be taken to be the local embedded symplectic submanifold defined by the coordinates @xmath113 .
definition [ def : nondegpoi ] concerns symplectic four - manifolds , which is the case relevant to the present paper . for the notion of non - degeneracy of a critical point in arbitrary dimension see @xcite , ( * ? ? ?
* section 3 ) .
non - degenerate critical points can be characterized ( see @xcite ) using the williamson normal form @xcite .
the analytic version of the following theorem by eliasson is due to vey @xcite .
[ h ] is a regular point ( rank 2 ) ; on the second figure , @xmath13 is a focus - focus point ( rank @xmath74 ) ; on the third one , @xmath13 is a transversally elliptic singularity ( rank @xmath114 ) ; on the right most figure , @xmath13 is an elliptic - elliptic point ( rank 0).,title="fig:",width=566 ] [ singularities_theorem ] the non - degenerate critical points of a completely integrable system @xmath115 are linearizable , i.e. if @xmath116 is a non - degenerate critical point of the completely integrable system @xmath117 then there exist local symplectic coordinates @xmath118 about @xmath13 , in which @xmath13 is represented as @xmath119 , such that @xmath120 , for all indices @xmath16 , where we have the following possibilities for the components @xmath121 , each of which is defined on a small neighborhood of @xmath122 in @xmath123 : * elliptic component : @xmath124 , where @xmath125 may take any value @xmath126 . *
hyperbolic component : @xmath127 , where @xmath125 may take any value @xmath126 .
* focus - focus component : @xmath128 and @xmath129 where @xmath125 may take any value @xmath130 ( note that this component appears as `` pairs '' ) . *
non - singular component : @xmath131 , where @xmath125 may take any value @xmath126 .
moreover if @xmath13 does not have any hyperbolic component , then the system of commuting equations @xmath120 , for all indices @xmath16 , may be replaced by the single equation @xmath132 where @xmath133 and @xmath51 is a diffeomorphism from a small neighborhood of the origin in @xmath123 into another such neighborhood , such that @xmath134 .
if the dimension of @xmath11 is @xmath135 and @xmath27 has no hyperbolic singularities which is the case we treat in this paper we have the following possibilities for the map @xmath136 , depending on the rank of the critical point : * if @xmath13 is a critical point of @xmath27 of rank zero , then @xmath137 is one of * * @xmath138 and @xmath139 .
* * @xmath140 and @xmath141 ; on the other hand , * if @xmath13 is a critical point of @xmath27 of rank one , then * * @xmath138 and @xmath142 . in this case , a non - degenerate critical point is respectively called _ elliptic - elliptic , focus - focus _ , or _
transversally - elliptic _ if both components @xmath143 are of elliptic type , @xmath144 together correspond to a focus - focus component , or one component is of elliptic type and the other component is @xmath145 or @xmath146 , respectively .
similar definitions hold for _ transversally - hyperbolic , hyperbolic - elliptic _ and _ hyperbolic - hyperbolic _ non - degenerate critical points .
suppose that @xmath47 is a connected symplectic four - manifold .
an integrable system @xmath52 is called _ almost - toric _ if all the singularities are non - degenerate without hyperbolic components .
[ prop : elliptic ] suppose that @xmath47 is a connected symplectic four - manifold .
let @xmath61 be an integrable system .
if @xmath27 is a toric integrable system , then @xmath27 is almost - toric , with only elliptic singularities .
this follows from the fact that a torus action is linearizable near a fixed point ; see , for instance @xcite .
a version of the following result is proven in @xcite for almost - toric systems for which the map @xmath147 is proper . here
we replace the condition of @xmath27 being proper by the condition that @xmath29 is a closed subset of @xmath58 ; this introduces additional subtleties .
our proof here is independent of the argument in @xcite .
[ theo : interior ] suppose that @xmath47 is a connected symplectic four - manifold .
assume that @xmath148 is an almost - toric integrable system with @xmath149 closed .
then the set of focus - focus critical values is countable , i.e. we may write it as @xmath150 , where @xmath151 . consider the following statements : * the fibers of @xmath27 are connected ; * the set @xmath152 or regular values of @xmath27 is connected ; * for any value @xmath46 of @xmath27 , for any sufficiently small disc @xmath153 centered at @xmath46 , @xmath154 is connected ; * the set of regular values is @xmath155 .
moreover , the topological boundary @xmath156 of @xmath157 consists precisely of the values @xmath158 , where @xmath13 is a critical point of elliptic - elliptic or transversally elliptic type
. then statement ( i ) implies statement ( ii ) , statement ( iii ) implies statement ( iv ) , and statement ( iv ) implies statement ( ii ) .
if in addition @xmath27 is proper , then statement implies statement .
it is interesting to note that the statement is optimal in that no other implication is true ( except @xmath159 which is a consequence of the stated implications ) .
this gives an idea of the various pathologies that can occur for an almost - toric system . from the local normal form [ singularities_theorem ] ,
focus - focus critical points are isolated , and hence the set of focus - focus critical points is countable ( remember that all our manifolds are second countable ) .
moreover , the image of a focus - focus point is necessarily in the interior of @xmath157 .
let us show that @xmath160 this is equivalent to showing that any value in @xmath161 is a critical value of @xmath27 . since @xmath157 is closed , @xmath162 , so for every @xmath163 we have that @xmath164 is nonempty . by the darboux - carathodory theorem ,
the image of a regular point must be in the interior of @xmath157 , therefore @xmath164 can not contain any regular point : the boundary can contain only singular values . since a point in @xmath161 can not be the image of a focus - focus singularity , it has to be the image of a transversally elliptic or an elliptic - elliptic singularity .
we now prove the implications stated in the theorem .
@xmath165 : since @xmath27 is almost - toric , the singular fibers are either points ( elliptic - elliptic ) , one - dimensional submanifolds ( codimension 1 elliptic ) or a stratified manifold of maximal dimension 2 ( focus - focus and elliptic ) . this is because none of the critical fibers can contain regular tori since the fibers are assumed to be connected by hypothesis ( i ) .
the only critical values that can appear in one - dimensional families are elliptic and elliptic - elliptic critical values ( see figure [ singularities_in_image ] ) .
the focus- focus singularities are isolated .
therefore the union of all critical fibers is a locally finite union of stratified manifolds of codimension at least 2 ; therefore this union has codimension at least 2 .
hence the complement is connected and therefore its image by @xmath27 is also connected .
@xmath166 : there is no embedded line segment of critical values in the interior of @xmath157 ( which would come from codimension 1 elliptic singularities ) because this is in contradiction with the hypothesis of local connectedness .
therefore @xmath167 . hence by , @xmath168 as desired , and all the elliptic critical values must lie in @xmath161 .
@xmath169 : as we saw above , @xmath170 contains only critical points , of elliptic type .
because of the local normal form , the set of rank 1 elliptic critical points in @xmath11 form a 2-dimensional symplectic submanifold with boundary , and this boundary is equal to the discrete set of rank 0 elliptic points .
therefore @xmath171 is connected .
this set is equal to @xmath172 , which in turn implies that @xmath173 is connected . by hypothesis
, this ensures that @xmath152 is connected .
assume for the rest of the proof that @xmath27 is proper .
@xmath174 : assume does not hold . in view of
, there exists an elliptic singularity ( of rank 0 or 1 ) @xmath46 in the interior of @xmath157 .
let @xmath175 be the corresponding fiber .
since it is connected , it must entirely consist of elliptic points ( this comes from the normal form theorem [ singularities_theorem ] ) .
the normal form also implies that @xmath46 must be contained in an embedded line segment of elliptic singularities , and the points in a open neighborhood @xmath176 of @xmath175 are sent by @xmath27 in only one side of this segment .
since @xmath46 is in the interior of @xmath157 , there is a sequence @xmath177 on the other side of the line segment that converges to @xmath46 as @xmath178 .
hence there is a sequence @xmath179 such that @xmath180 .
since @xmath27 is proper , one can assume that @xmath181 converges to a point @xmath13 ( necessarily in @xmath182 ) .
by continuity of @xmath27 , @xmath13 belongs to the fiber over @xmath46 , and thus to @xmath176 , which is a contradiction . in an open disk.,width=472 ]
in this section we study the structure of the fibers of an almost - toric system . we shall need below the definition and basic properties of the bifurcation set of a smooth map .
[ def : bifurcation ] let @xmath11 and @xmath183 be smooth manifolds . a smooth map @xmath184 is said to be _ locally trivial _ at @xmath185
if there is an open neighborhood @xmath186 of @xmath187 such that @xmath188 is a smooth submanifold of @xmath11 for each @xmath189 and there is a smooth map @xmath190 such that @xmath191 is a diffeomorphism .
the _ bifurcation set _ @xmath40 consists of all the points of @xmath183 where @xmath38 is not locally trivial . note , in particular , that @xmath192 is a diffeomorphism for every @xmath193 . also , the set of points where @xmath194 is locally trivial is open in @xmath183 .
recall that @xmath195 is a closed subset of @xmath196 .
it is well known that the set of critical values of @xmath194 is included in the bifurcation set ( see ( * ? ? ? * proposition 4.5.1 ) ) . in general , the bifurcation set strictly includes the set of critical values .
this is the case for the momentum - energy map for the two - body problem @xcite .
however ( see ( * ? ? ?
* page 340 ) ) , if @xmath197 is a smooth _ proper _ map , then the bifurcation set of @xmath38 is equal to the set of critical values of @xmath38 .
recall that a smooth map @xmath198 is _ morse _ if all its critical points are non - degenerate . the smooth map @xmath38 is _ morse - bott _ if the critical set of @xmath38 is a disjoint union of connected submanifolds @xmath199 of @xmath11 , on which the hessian of @xmath38 is non - degenerate in the transverse direction , i.e. , @xmath200 the _ index of _
@xmath13 is the number of negative eigenvalues of @xmath201 .
the goal of this section is to find a useful morse theoretic result , valid in great generality and interesting on its own , that will ultimately imply the connectedness of the fibers of an integrable system ( see theorem [ theo : fibers ] ) .
here we do not rely on fomenko s morse theory @xcite , because we do not want to select a nonsingular energy surface .
instead , the model is ( * ? ? ? * lemma 5.51 ) ; however , the proof given there does not extend to the non - compact case , as far as we can tell .
we thank helmut hofer and thomas baird for sharing their insights on morse theory with us that helped us in the proof of the following result .
[ lemm : semibounded ] let @xmath202 be a morse - bott function on a connected manifold @xmath11 .
assume @xmath38 is proper and bounded from below and has no critical manifold of index 1 .
then the set of critical points of index 0 is connected .
we endow @xmath11 with a riemannian metric .
the negative gradient flow of @xmath38 is complete . indeed , along the the flow the function @xmath38 can not increase and , by hypothesis , @xmath38 is bounded from below .
therefore , the values of @xmath38 remain bounded along the flow . by properness of @xmath38 ,
the flow remains in a compact subset of @xmath11 and hence it is complete .
let us show , using standard morse - bott theory , that the integral curve of @xmath203 starting at any point @xmath204 tends to a critical manifold of @xmath38 . in the compact set in @xmath205 , there must be a finite number of critical manifolds contained .
if the integral curve avoids a neighborhood of these critical manifolds , by compactness it has a limit point , and by continuity the vector field at the limit point must vanish ; we get a contradiction , thus proving the claim .
thus , we have the disjoint union @xmath206 , where @xmath207 is the set of critical points of index @xmath208 , and @xmath209 is its stable manifold : @xmath210 where @xmath211 is any distance compatible with the topology of @xmath11 ( for example , the one induced by the given riemannian metric on @xmath11 ) and @xmath212 is the flow of the vector field @xmath213 .
since @xmath38 has no critical point of index 1 , we have @xmath214 the local structure of morse - bott singularities given by the morse - bott lemma @xcite ) implies that @xmath209 is a submanifold of codimension @xmath208 in @xmath11 . hence @xmath215 can not disconnect @xmath11 .
since all local minima of @xmath38 are in @xmath216 , we see that @xmath216 must be the set of global minima of @xmath38 ; thus @xmath216 must be equal to the level set @xmath217 .
[ prop : levelset ] let @xmath11 be a connected smooth manifold and @xmath202 be a proper morse - bott function whose indices and co - indices are always different from 1
. then the level sets of @xmath38 are connected .
let @xmath46 be a regular value of @xmath38 ( such a value exists by sard s theorem ) .
then @xmath218 is a morse - bott function . on the set @xmath219 ,
the critical points of @xmath51 coincide with the critical points of @xmath38 and they have the same index . on the set @xmath220 , the critical points of @xmath51 also coincide with the critical points of @xmath38 and they have the same coindex .
the level set @xmath221 is clearly a set of critical points of index 0 of @xmath51 .
of course , @xmath51 is bounded from below .
thus , by lemma [ lemm : semibounded ] , the set of critical points of index 0 of @xmath51 is connected ( it may be empty ) and hence equal to @xmath222 . therefore @xmath223 is connected .
this shows that all regular level sets of @xmath38 are connected .
( as usual , a regular level set or regular fiber is a level set that contains only regular points , i.e. the preimage of a regular value . ) finally let @xmath224 be a critical value of @xmath38 ( if any ) .
since @xmath38 is proper and has isolated critical manifolds , the set of critical values is discrete .
let @xmath225 such that the interval @xmath226 $ ] does not contain any other critical value .
consider the manifold @xmath227 , and , for any @xmath228 , let @xmath229 where @xmath230 is the vertical component on the sphere @xmath231 .
notice that @xmath232 is a morse function with indices 0 and 2 .
thus @xmath233 is a morse - bott function on @xmath183 with indices and coindices of the same parity as those of @xmath38 .
thus no index nor coindex of @xmath234 can be equal to 1 . by the first part of the proof
, the regular level sets of @xmath233 must be connected .
the definition of @xmath235 implies that @xmath74 is a regular value of @xmath233 .
thus @xmath236 is connected .
since @xmath38 is proper , @xmath237 is also compact . because a non - increasing intersection of compact connected sets is connected
, we see that @xmath238 is connected .
there is no a priori reason why the fibers @xmath239 of @xmath27 should be connected even if @xmath240 and @xmath241 have connected fibers ( let alone if just one of @xmath240 or @xmath241 has connected fibers ) .
however , the following result shows that this conclusion is holds . to prove it
, we need a preparatory lemma which is interesting on its own .
[ tricky ] let @xmath242 be a map from a smooth connected manifold @xmath33 to @xmath123 .
ler @xmath152 be the set of regular values of @xmath38 .
suppose that @xmath38 has the following properties .
* @xmath38 is a proper map . * for every sufficiently small neighborhood @xmath153 of any critical value of @xmath38
, @xmath243 is connected . *
the regular fibers of @xmath38 are connected .
* the set @xmath244 of critical points of @xmath38 has empty interior
. then the fibers of @xmath38 are connected .
we use the following `` fiber continuity '' fact : that if @xmath176 is a neighborhood of a fiber @xmath223 of a continuous proper map @xmath38 , then the fibers @xmath245 with @xmath246 close to @xmath46 also lie inside @xmath176 . indeed ,
if this statement were not true , then there would exist a sequence @xmath247 and a sequence of points @xmath248 , @xmath249 , such that there is a subsequence @xmath250 .
however , by continuity @xmath251 which is a contradiction .
assume a fiber @xmath252 of @xmath38 is not connected .
then there are disjoint open sets @xmath253 and @xmath254 in @xmath33 such that @xmath73 lies in @xmath255 but is not contained in either @xmath253 or @xmath254 . by fiber continuity
, there exists a small open disk @xmath153 about @xmath256 such that @xmath257 .
since the regular fibers are connected , we can define a map @xmath258 which for @xmath259 is equal to @xmath114 if @xmath260 , and is equal to @xmath74 if @xmath261 . the fiber continuity says that the sets @xmath262 and @xmath263 are open , thus proving that @xmath264 is continuous .
by ( 2 ) , the image of @xmath264 must be connected , and therefore @xmath264 is constant .
we can hence assume without loss of generality that all regular fibers above @xmath153 are contained in @xmath253 .
now consider the restriction @xmath265 of @xmath38 on the open set @xmath266 .
because of the above argument , it can not take any value in @xmath154 .
thus this map takes values in the set of critical values of @xmath38 , which has measure zero by sard s theorem .
this requires that on @xmath266 , the rank of @xmath267 be strictly less that @xmath7 , which contradicts ( 4 ) , and hence proves the lemma : @xmath268 has to be connected .
now we are ready to prove one of our main results .
[ theo : fibers ] suppose that @xmath47 is a connected symplectic four - manifold .
let @xmath269 be an almost - toric integrable system such that @xmath27 is a proper map .
suppose that @xmath240 has connected fibers , or that @xmath241 has connected fibers .
then the fibers of @xmath27 are connected . without loss of generality
, we may assume that @xmath240 has connected fibers .
[ [ step-1 . ] ] step 1 .
+ + + + + + + we shall prove first that for every _ regular value _
@xmath270 of @xmath27 , the fiber @xmath271 is connected . to do this
, we divide the proof into two cases .
+ + _ case 1a . _
assume @xmath272 is a regular value of @xmath240 .
then the fiber @xmath273 is a smooth manifold .
let us show first that the non - degeneracy of the critical points of @xmath147 and the definition of almost - toric systems implies that the function @xmath274 is morse - bott .
let @xmath152 be the set of regular values of @xmath27 .
let @xmath275 be a critical point of @xmath276 .
then there exists @xmath277 such that @xmath278 .
thus @xmath275 is a critical point of @xmath27 ; it must be of rank @xmath114 since @xmath279 never vanishes on @xmath273 . since @xmath27 is an almost - toric system , the only possible rank @xmath114 singularities are transversally elliptic singularities , i.e. , singularities with one elliptic component and one non singular component in theorem [ singularities_theorem ]
; see figure [ singularities_in_image ] .
thus , by theorem [ singularities_theorem ] , there exist local canonical coordinates @xmath280 such that @xmath281 for some local diffeomorphism @xmath51 of @xmath58 about the origin and fixing the origin ; thus the derivative @xmath282 note @xmath283 implies that @xmath284 .
therefore , by the implicit function theorem , the submanifold @xmath273 is locally parametrized by the variables @xmath285 and , within it , the critical set of @xmath276 is given by the equation @xmath286 ; this is a submanifold of dimension 1 .
the taylor expansion of @xmath276 is easily computed to be @xmath287 thus , the coefficient of @xmath288 is @xmath289 which is non - zero and hence the hessian of @xmath276 is transversally non - degenerate .
this proves that @xmath276 is morse - bott , as claimed .
second , we prove , in this case , that the fibers of @xmath27 are connected . at @xmath275 , the transversal hessian of @xmath276
has either no or two negative eigenvalues , depending on the sign of @xmath289 .
this implies that each critical manifold has index @xmath74 or index @xmath53 .
if this coefficient is negative , the sum of the two corresponding eigenspaces is the full @xmath53-dimensional @xmath290-space .
note that @xmath291 is a proper map : indeed , if @xmath292 is compact , then @xmath293 is compact because @xmath27 is proper .
thus @xmath291 is a smooth morse - bott function on the connected manifold @xmath273 and @xmath276 and @xmath294 have only critical points of index @xmath74 or @xmath53 .
we are in the hypothesis of proposition [ prop : levelset ] and so we can conclude that the fibers of @xmath291 are connected .
now , since @xmath295 , it follows that @xmath271 is connected for all @xmath296 whenever @xmath272 is a regular value of @xmath240 .
+ + _ case 1b . _
assume that @xmath297 is not a regular value of @xmath240 .
note that there exists a point @xmath298 in every connected component @xmath299 of @xmath152 such that @xmath300 is a regular value for @xmath240 ; otherwise @xmath279 would vanish on @xmath301 , which violates the definition of @xmath152 .
the restriction @xmath302 is a locally trivial fibration since , by assumption , @xmath27 is proper and thus the bifurcation set is equal to the critical set .
thus all fibers of @xmath303 are diffeomorphic .
it follows that @xmath271 is connected for all @xmath304 .
this shows that all inverse images of regular values of @xmath27 are connected .
[ [ step-2 . ] ] step 2 .
+ + + + + + + we need to show that @xmath305 is connected if @xmath306 is not a regular value of @xmath27 .
we claim that there is _ no _ critical value of @xmath27 in the interior of the image @xmath29 , except for the critical values that are images of focus - focus critical points of @xmath27 .
indeed , if there was such a critical value @xmath307 , then there must exist a small segment line @xmath308 of critical values ( by the local normal form described in theorem [ singularities_theorem ] and figure [ singularities_in_image ] ) .
now we distinguish two cases .
+ + _ case 2a . _
first assume that @xmath308 is not a vertical segment ( i.e. , contained in a line of the form @xmath309 ) and let @xmath310 .
then @xmath311 contains a small interval around @xmath272 , so by sard s theorem , it must contain a regular value @xmath312 for the map @xmath240 .
then @xmath313 is a smooth manifold which is connected , by hypothesis . by the argument earlier in the proof
( see step 1 , case a ) , both @xmath241 and @xmath314 restricted to @xmath313 are proper morse - bott functions with indices @xmath74 and @xmath53 .
so , if there is a local maximum or local minimum , it must be unique .
however , the existence of this line of rank 1 elliptic singularities implies that there is a local maximum / minimum of @xmath315 ( see formula ) .
since the corresponding critical value lies in the interior of the image of @xmath315 , it can not be a global extremum ; we arrived at a contradiction .
thus the small line segment @xmath308 must be vertical .
+ + _ case 2b . _
second , suppose that @xmath308 is a vertical segment ( i.e. , contained in a line of the form @xmath309 ) and let @xmath310 .
we can assume , without loss of generality , that the connected component of @xmath308 in the bifurcation set is vertical in the interior of @xmath29 ; indeed , if not , apply case 2a a above . from figure [ singularities_in_image ]
we see that @xmath316 must contain at least one critical point @xmath37 of transversally elliptic type together with another point @xmath157 either regular or of transversally elliptic type . by the normal form of non - degenerate singularities
, @xmath313 must be locally path connected . since it is connected by assumption , it must be path connected .
so we have a path @xmath317\mapsto j^{-1}(x_0)$ ] such that @xmath318 and @xmath319 .
near @xmath37 we have canonical coordinates @xmath320 and a local diffeomorphism @xmath51 defined in a neighborhood of the origin of @xmath58 and preserving it , such that @xmath321 and @xmath322 .
write @xmath323 .
the critical set is defined by the equations @xmath286 and , by assumption , is mapped by @xmath27 to a vertical line .
hence @xmath324 is constant , so @xmath325 since @xmath51 is a local diffeomorphism , @xmath326 , so we must have @xmath327 .
thus , by the implicit function theorem , any path starting at @xmath37 and satisfying @xmath328 must also satisfy @xmath329 .
therefore , @xmath45 has to stay in the critical set @xmath286 .
assume first that @xmath330)$ ] does not touch the boundary of @xmath29 .
then this argument shows that the set of @xmath331 $ ] such that @xmath332 belongs to the critical set of @xmath27 is open .
it is also closed by continuity of @xmath333 .
hence it is equal to the whole interval @xmath334 $ ] .
thus @xmath157 must be in the critical set ; this rules out the possibility for @xmath157 to be regular .
thus @xmath157 must be a rank-1 elliptic singularity .
notice that the sign of @xmath335 indicates on which side of @xmath308 ( left or right ) lie the values of @xmath27 near @xmath37 .
thus , even if @xmath336 itself is not globally defined along the path @xmath45 , this sign is locally constant and thus globally defined along @xmath45 .
therefore , all points near @xmath316 are mapped by @xmath27 to the same side of @xmath308 , which says that @xmath308 belongs to the boundary of @xmath29 ; this is a contradiction .
finally , assume that @xmath330)$ ] touches the boundary of @xmath29 .
from the normal form theorem , this can only happen when the fiber over the contact point contains an elliptic - elliptic point @xmath337 .
thus there are local canonical coordinates @xmath338 and a local diffeomorphism @xmath51 defined in a neighborhood of the origin of @xmath58 and preserving it , such that @xmath339 near @xmath340 .
we note that the same argument as above applies : simply replace the @xmath146 component by @xmath341 .
thus we get another contradiction .
therefore there are no _ critical _ values @xmath46 in the interior of the image @xmath29 other than focus - focus values ( i.e. , images of focus - focus points ) .
[ [ step-3 . ] ] step 3 .
+ + + + + + + we claim here that for any critical value @xmath46 of @xmath27 and for any sufficiently small disk @xmath153 centered at @xmath46 , @xmath154 is connected .
first we remark that step 2 implies that item in theorem [ theo : interior ] holds , and hence item must hold : the set of regular values of @xmath27 is connected .
if @xmath46 is a focus - focus value , it must be contained in the interior of @xmath29 , therefore it follows from step 2 that it is isolated : there exists a neighborhood of @xmath46 in which @xmath46 is the only critical value , which proves the claim in this case .
we assume in the rest of the proof that @xmath342 is an elliptic ( of rank 0 or 1 ) critical value of @xmath27 . since we have just proved in step 2 that there are no critical values in the interior of @xmath29 other than focus - focus values
, we conclude that @xmath343 .
moreover , the fiber can not contain a regular liouville torus .
then , again by theorem [ singularities_theorem ] , the only possibilities for a neighborhood of @xmath46 in @xmath29 are superpositions of elliptic local normal forms of rank 0 or 1 ( given by theorem [ singularities_theorem ] ) in such a way that @xmath343 .
if only one local model appears , then the claim is immediate .
let us show that a neighborhood @xmath253 of @xmath46 can not contain several different images of local models .
indeed , consider the possible configurations for two different local images @xmath344 and @xmath345 : either both @xmath344 and @xmath345 are elliptic - elliptic images , or both are transversally elliptic images , or @xmath344 is an elliptic - elliptic image and @xmath345 is a transversally elliptic image .
step 2 implies that the critical values of @xmath27 in @xmath344 and @xmath345 can only intersect at a point , provided the neighborhood @xmath253 is taken to be small enough .
let us consider a vertical line @xmath308 through @xmath344 which corresponds to a regular value of @xmath240 .
any crossing of @xmath308 with a non - vertical boundary of @xmath29 must correspond to a local extremum of @xmath346 , and by step 2 this local extremum has to be a global one . since only one global maximum and one global minimum are possible , the only allowed configurations for @xmath344 and @xmath345 are such that the vertical line @xmath347 through @xmath46 separates the regular values of @xmath344 from the regular values of @xmath345 ( see figure [ fig : c1c2 ] ) .
are connected.,scaledwidth=45.0% ] since @xmath152 is path connected , there exists a path in @xmath152 connecting a point in @xmath348 to a point in @xmath349 . by continuity
this path needs to cross @xmath347 , and the intersection point @xmath350 must lie outside @xmath253 .
therefore there exists an open ball @xmath351 centered at @xmath352 .
suppose for instance that @xmath353 .
then for @xmath354 , @xmath355 can not be a maximal value of @xmath356 , which means that for each of @xmath344 and @xmath345 , only local minima for @xmath356 are allowed .
this can not be achieved by any of the local models , thus finishing the proof of our claim .
the statement of the theorem now follows from lemma [ tricky ] since all fibers are codimension at least one ( this follows directly from the independence assumption in the definition of an integrable system .
but , in fact , we know from the topology of non - degenerate integrable systems that fibers have codimension at least two @xcite ) .
it is not true that an almost - toric integrable system with connected regular fibers has also connected singular fibers .
see example [ ex0 ] below .
the following are examples of almost - toric systems in which the fibers of @xmath27 are not connected . in the next section
we will combine theorem [ theo : fibers ] with an upcoming result on contact theory for singularities ( which we will prove too ) in order to obtain theorem [ theo : main - connectivity ] of section [ sec : intro ] .
[ ex0 ] this example appeared in ( * ? ? ?
* chapter 5 , figure 29 ) .
it is an example of a toric system @xmath357 on a compact manifold for which @xmath240 and @xmath241 have some disconnected fibers ( the number of connected components of the fibers also changes ) . because this example is constructed from the standard toric system @xmath358 by precomposing with a local diffeomorphism , the singularities are non - degenerate . in this case
the fundamental group @xmath359 has one generator , so @xmath29 is not simply connected , and hence not contractible .
see figure [ anneau ] .
an extreme case of this example can be obtained by letting only two corners overlap ( figure [ fig : anneau - extreme ] ) .
we get then an almost - toric system where all regular fibers are connected , but one singular fiber is not connected . by precomposing with a local diffeomorphism.,width=434 ] by a one - point identification , and which has one disconnected fiber .
all the other fibers are connected.,width=94 ] [ ex2 ] the manifold is @xmath360 .
choose coordinates @xmath21 $ ] , @xmath361 on @xmath20 ; these are action - angle coordinates for the hamiltonian system given by the rotation action .
choose coordinates @xmath362 on @xmath24 .
for each @xmath363 , @xmath364 , the 2-form @xmath365 is symplectic .
let @xmath366 and note that @xmath240 is the momentum map of an @xmath367-action rotating the sphere about the vertical axis and the first component of @xmath368 .
note that @xmath27 maps @xmath11 onto the annulus in figure [ fig : annulus ] .
topologically @xmath27 winds the first copy of @xmath367 exactly @xmath7 times around the annulus and maps @xmath20 to radial intervals .
the fiber of any regular value is thus @xmath7 copies of @xmath369 , while the preimage of any singular value is @xmath367 .
one can easily check that all singular fibers are transversely elliptic .
the hamiltonian vector fields are : @xmath370 which are easily checked to commute ( the coefficient functions are invariants of the flow ) .
observe that neither integrates to a global circle action .
as for the components , they all look alike ( the system has rotational symmetry in the @xmath300 coordinate ) .
consider for instance the component @xmath371 which has critical values @xmath372 and is morse - bott with critical sets equal to @xmath7 copies of @xmath367 and with morse indices @xmath373 , and @xmath374 , respectively . the sets @xmath375 and @xmath376 are empty .
for any @xmath377 or @xmath378 , the fiber of @xmath379 is equal to @xmath7 copies of @xmath380 and for @xmath381 the fiber @xmath379 is equal to @xmath382 copies of @xmath383 .
we thank thomas baird for this example .
much of our interest on this topic came from questions asked by physicists and chemists in the context of molecular spectroscopy ( * ? ? ? * section 1 ) , @xcite .
many research teams have been working on this topic , to name a few : mark child s group in oxford ( uk ) , jonathan tennyson s at the university college london ( uk ) , frank de lucia s at ohio state university ( usa ) , boris zhilinskii s at dunkerque ( france ) , and marc joyeux s at grenoble ( france ) . for applications to concrete physical models the theorem is far reaching . [
fig : example ] the reason is that is essentially impossible to prove connectedness of the fibers of concrete physical systems . for the purpose of applications , a computer may approximate the bifurcation set of a system and find its image to a high degree of accuracy .
see for example the case of the coupled spin - oscillator in figure [ fig : example ] , which is one of the most fundamental examples in classical physics , and which recently has attracted much attention , see @xcite .
in this section we study the structure of the image of an almost - toric system .
[ h ] of an almost - toric momentum map @xmath384 .
the set of regular values of @xmath27 is denoted by @xmath152 .
the marked dots @xmath385 inside of @xmath157 represent singular values , corresponding to the focus - focus fibers .
the set @xmath152 is equal to @xmath157 minus @xmath386.,title="fig:",width=207 ] we star with the following observation .
[ lemma : critical - set ] let @xmath11 be a connected smooth manifold and let @xmath387 be a morse - bott function with connected fibers .
then the set @xmath216 of index zero critical points of @xmath38 is connected . moreover , if @xmath388 , the following hold : * if @xmath389 then @xmath390 . * if @xmath391 then @xmath392 .
the fiber over a point is locally path connected .
if the point is critical this follows from the morse - bott lemma and if the point is regular , this follows from the submersion theorem .
let @xmath13 be a critical point of index @xmath74 of @xmath38 , i.e. , @xmath393 , and let @xmath394 , @xmath395 .
let @xmath396 \to \lambda$ ] . since @xmath175 is connected and locally path connected , it is path connected .
let @xmath397 \to \lambda$ ] be a continuous path starting at @xmath13 . by the morse - bott lemma , @xmath398 .
therefore , since @xmath175 is path connected , @xmath399 .
each connected component of @xmath216 is contained in some fiber of @xmath38 and hence @xmath175 is the connected component of @xmath216 that contains @xmath13 .
we shall prove that @xmath216 has one one connected component .
assume that there is a point @xmath400 such that @xmath401 and let @xmath402 \to m$ ] be a continuous path from @xmath13 to @xmath400 .
let @xmath403 then @xmath404 and , by definition , for every @xmath405 there exists @xmath406 $ ] such that @xmath407 .
let @xmath408 .
let @xmath253 be a neighborhood of @xmath409 in which we have the morse - bott coordinates given by the morse - bott lemma centered at @xmath409 .
for @xmath410 small enough , @xmath411 ) \subset u$ ] and therefore for @xmath412 $ ] we have that @xmath413 which is a contradiction .
note that the following result is strictly morse - theoretic ; it does not involve integrable systems .
a version of this result was proven in ( * ? ? ?
* theorem 3.4 ) in the case of integrable systems @xmath414 for which @xmath415 is both a proper map ( hence @xmath416 is proper ) and a momentum map for a hamiltonian @xmath367-action .
the version we prove here applies to smooth maps , which are not necessarily integrable systems .
[ imagemomentum : theorem ] let @xmath11 be a connected smooth four - manifold .
let @xmath417 be a smooth map .
equip @xmath418 with the standard topology .
suppose that the component @xmath240 is a non - constant morse - bott function with connected fibers .
let @xmath419 be the functions defined by @xmath420 and @xmath421 .
the functions @xmath422 , @xmath423 are lower semicontinuous . moreover ,
if @xmath29 is closed in @xmath58 then @xmath424 , @xmath423 are upper semicontinuous ( and hence continuous ) , and @xmath29 may be described as @xmath425 in particular , @xmath29 is contractible .
first we consider the case where @xmath29 is not necessarily closed ( part 1 ) . in part 2
we prove the stronger result when @xmath29 is closed .
[ [ part-1 . ] ] part 1 .
+ + + + + + + we do not assume that @xmath29 is closed and prove that @xmath424 is lower semicontinuous ; the proof that @xmath426 is lower semicontinuous is analogous . since , by assumption , @xmath240 is non - constant , the interior set @xmath427 of @xmath86 is non - empty .
the set @xmath427 is an open interval @xmath298 since @xmath11 is connected and @xmath240 is continuous .
lower semicontinuity of @xmath424 is proved first in the interior of @xmath86 ( case a ) and then at the possible boundary ( case b ) .
+ + _ case a. _ let @xmath428 , and let @xmath429 .
let @xmath430 . by the definition of supremum ,
there exists @xmath431 with @xmath432 such that if @xmath433 then @xmath434 ( see figure [ fig : lowersemicontinuity ] ) .
here we have assumed that @xmath435 ; if @xmath436 , we just need to replace @xmath437 by an arbitrary large constant .
let @xmath438 . then @xmath439 .
endow @xmath11 with a riemannian metric and , with respect to this metric , consider the gradient vector field @xmath440 of @xmath240 .
let @xmath441 such that the flow @xmath442 of @xmath440 starting at @xmath443 exists for all @xmath444 .
now we distinguish two cases .
.,scaledwidth=60.0% ] * assume @xmath445 .
since @xmath446 , the set @xmath447 is a neighborhood of @xmath312 .
+ let @xmath157 the ball of radius @xmath448 centered at @xmath449 .
let @xmath450 , which contains @xmath13 .
let @xmath451 be small enough such that @xmath452 for all @xmath453 .
the set @xmath454 is a neighborhood of @xmath312 , so there is @xmath405 such that @xmath455 .
+ let @xmath456 ; there exists @xmath453 such that @xmath457 by definition of @xmath454 . since @xmath458
we conclude that @xmath459 , so @xmath460 for all @xmath272 with @xmath461 .
+ thus we get @xmath462 for all @xmath272 with @xmath461 , which proves the lower semicontinuity . *
assume @xmath463 .
by lemma [ lemma : critical - set ] we conclude that @xmath13 is not of index @xmath74 , for otherwise @xmath439 would be a global minimum in @xmath86 which contradicts the fact that , by assumption , @xmath428 . +
thus the hessian of @xmath240 has at least one negative eigenvalue and therefore there exists @xmath441 such that @xmath464 is an open neighborhood of @xmath312 and we may then proceed as in case a.1 .
hence @xmath424 is lower semicontinuous .
+ + _ case b. _ we prove here lower semicontinuity at a point @xmath312 in the topological boundary of @xmath86 .
we may assume that @xmath465 and that @xmath466 .
by lemma [ lemma : critical - set ] , @xmath467 , where @xmath216 denotes the set of critical points of @xmath240 of index @xmath74 .
if @xmath468 and @xmath253 is a small neighborhood of @xmath13 , it follows from the morse - bott lemma that @xmath469 is a neighborhood of @xmath470 in @xmath86 .
hence we may proceed as in case a to conclude that @xmath424 is lower semicontinuous .
[ [ part-2 . ] ] part 2 .
+ + + + + + + assuming that @xmath29 is closed we shall prove now that @xmath424 is upper semicontinuous .
suppose that @xmath424 is not upper semicontinuous at some point @xmath471 ( so we must have @xmath472 )
. then there exists @xmath225 and a sequence @xmath473 converging to @xmath312 such that @xmath474 .
first assume that @xmath475 is finite .
since @xmath476 is upper semicontinuous , for @xmath272 close to @xmath312 we have @xmath477 thus we have @xmath478 let @xmath479 . because of , and since @xmath480 is an interval whose closure is @xmath481 $ ] ( remember that @xmath240 has connected fibers and @xmath241 is continuous ) , we must have @xmath482 .
therefore @xmath483 .
since @xmath29 is closed , the limit of this sequence belongs to @xmath29 ; thus @xmath484 .
hence @xmath485 , a contradiction .
if @xmath486 , we just replace in the proof @xmath487 by some constant @xmath37 such that @xmath488 .
hence @xmath422 is upper semicontinuous .
therefore @xmath422 is continuous .
the same argument applies to @xmath476 . in order to prove ,
notice that for any @xmath489 , we have the equality @xmath490 therefore , if @xmath29 is closed , @xmath491 must be closed and hence equal to @xmath492 $ ] .
the equality follows by taking the union of the identity over all @xmath493 .
finally , we show that @xmath29 is contractible .
since @xmath494 is homeomorphic to a compact interval , @xmath29 is homeomorphic to a closed subset of the strip @xmath495 $ ] by means of a homeomorphism @xmath51 that fixes the first coordinate @xmath272 .
thus , composing by this homeomorphism we get @xmath496 for some continuous functions @xmath497 and @xmath498 . then the map @xmath499 \ni((x , y),t ) \mapsto ( x , t(y - h^-(x ) ) ) \in \mathbb{r}^2\ ] ] is a homotopy equivalence with the horizontal axis . since @xmath51 is a homeomorphism ,
we conclude that @xmath29 is contractible . we can give a stronger formulation of theorem [ imagemomentum : theorem ] .
first , recall that if @xmath500 is a smoothly immersed @xmath114-dimensional manifold in @xmath58 , we say that @xmath500 _ has no horizontal tangencies _ if there exists a smooth curve @xmath501 such that @xmath502 and @xmath503 for every @xmath504 .
note that @xmath500 has no horizontal tangencies if and only if for every @xmath505 the @xmath114-manifold @xmath500 is transverse to the horizontal line @xmath506 .
we start with the following result , which is of independent interest and its applicability goes far beyond its use in this paper .
we begin with a description of the structure of the set @xmath507 of critical values of an integrable systems @xmath508 ; as usual @xmath509 denotes the set of critical points of @xmath27 .
let @xmath510 and @xmath511 a small closed ball centered at @xmath512 .
for each point @xmath513 we choose a chart about @xmath13 in which @xmath27 has normal form ( see theorem [ singularities_theorem ] ) .
there are seven types of normal forms , as depicted in figure [ singularities_in_image ] .
since @xmath514 is compact , we can select a finite number of such chart domains that still cover @xmath514 . for each such chart domain @xmath176 ,
the set of critical values of @xmath515 is diffeomorphic to the set of critical values of one of the models described in figure [ singularities_in_image ] , which is either empty , an isolated point , an open curve , or up to four open curves starting from a common point .
since @xmath516 it follows that @xmath517 is a finite union of such models .
this discussion leads to the following proposition .
[ prop : strata ] let @xmath0 be a connected symplectic four - manifold .
let @xmath518 be a non - degenerate integrable system .
suppose that @xmath27 is a proper map .
then @xmath519 is the union of a finite number of stratified manifolds with 0 and 1 dimensional strata .
more precisely , @xmath43 is a union of isolated points and of smooth images of immersions of closed intervals ( since @xmath27 is proper , a 1-dimensional stratum must either go to infinity or end at a rank - zero critical value of @xmath27 ) .
[ defi : tangent ] let @xmath520 .
a vector @xmath521 is called _ tangent to @xmath43 _ if there is a smooth immersion @xmath522\to\sigma_f$ ] with @xmath523 and @xmath524 .
here @xmath525 is smooth on @xmath334 $ ] , when @xmath334 $ ] is viewed as a subset of @xmath526 .
notice that a point @xmath46 can have several linearly independent tangent vectors .
[ defi : contact ] let @xmath45 be a smooth curve in @xmath44 .
* if @xmath45 intersects @xmath43 at a point @xmath46 , we say that the intersection is _ transversal _ if no tangent vector of @xmath43 at @xmath46 is tangent to @xmath45 .
otherwise we say that @xmath46 is a _
tangency point_. * assume that @xmath45 is tangent to a 1-stratum @xmath527 of @xmath43 at a point @xmath528 .
near @xmath46 , we may assume that @xmath45 is given by some equation @xmath529 , where @xmath530 .
+ we say that @xmath45 has a _ non - degenerate contact _ with @xmath527 at @xmath46 if , whenever @xmath531 is a smooth local parametrization of @xmath527 near @xmath46 with @xmath532 , then the map @xmath533 has a non - degenerate critical point at @xmath534 . *
every tangency point that is not a non - degenerate contact is called _ degenerate _
( this includes the case where @xmath45 is tangent to @xmath43 at a point @xmath46 which is the end point of a 1-dimensional stratum @xmath527 ) . with this terminology , we can now see how a morse function on @xmath44 can give rise to a morse - bott function on @xmath11 . [
prop : bott ] let @xmath0 be a connected symplectic four - manifold .
let @xmath535 be an integrable system with non - degenerate singularities ( of any type , so this statement applies to hyperbolic singularities too ) such that @xmath27 is proper .
let @xmath536 be the set of critical values of @xmath27 , i.e. , @xmath519 .
let @xmath537 be open .
suppose that @xmath538 is a morse function whose critical set is disjoint from @xmath43 and the regular level sets of @xmath38 intersect @xmath539 transversally or with non - degenerate contact .
then @xmath540 is a morse - bott function on @xmath541 . here by regular level set of @xmath38 we mean a level set corresponding to a regular value of @xmath38 .
let @xmath542 .
writing @xmath543 we see that if @xmath13 is a critical point of @xmath544 then either @xmath545 is a critical point of @xmath38 ( so @xmath546 ) , or @xmath547 and @xmath548 are linearly dependent ( which means @xmath549 ) . by assumption
, these two cases are disjoint : if @xmath545 is a critical point of @xmath38 , then @xmath550 which means that @xmath13 is a regular point of @xmath27 .
thus @xmath551 is a disjoint union of two closed sets . since hausdorff manifolds are normal , these two closed sets have disjoint open neighborhoods .
thus @xmath552 is a submanifold if and only if both sets are submanifolds , which we prove next .
[ [ study - of - f-1textupcritf . ] ] study of @xmath553 .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath554 and @xmath555 .
we assume that @xmath512 is a critical point of @xmath38 , i.e. , @xmath556 . by hypothesis ,
since the rank is lower semicontinuous , there exists a neighborhood @xmath176 of @xmath275 in which @xmath558 for all @xmath559 .
thus , on @xmath176 , @xmath542 is critical at a point @xmath13 if and only if @xmath158 is critical for @xmath38 .
since @xmath38 is a morse function , its critical points are isolated ; therefore we can assume that the critical set of @xmath544 in @xmath176 is precisely @xmath560 .
since @xmath561 is a compact regular fiber ( because @xmath27 is proper ) , it is a finite union of liouville tori ( this is the statement of the action - angle theorem ; the finiteness comes from the fact that each connected component is isolated ) . in particular , @xmath562 is a submanifold and we can analyze the non - degeneracy component - wise . given any @xmath563 , the submersion theorem ensures that @xmath240 and @xmath241 can be seen as a set of local coordinates of a transversal section to the fiber @xmath561 .
thus , using the taylor expansion of @xmath38 of order 2 , we get the 2-jet of @xmath564 : @xmath565 where @xmath566 again , since @xmath567 are taken as local coordinates , we see that the transversal hessian of @xmath544 in the @xmath567-variables is non - degenerate , since @xmath568 is non- degenerate by assumption ( @xmath38 is morse ) .
thus we have shown that @xmath553 is a smooth submanifold ( a finite union of liouville tori ) , transversally to which the hessian of @xmath544 is non - degenerate .
[ [ study - of - textupcritl - cap - textupcritf . ] ] study of @xmath569 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let @xmath554 be a critical point for @xmath27 , and let @xmath555 . by assumption
, @xmath512 is a regular value of @xmath38 .
thus , there exists an open neighborhood @xmath254 of @xmath512 in @xmath44 that contains only regular values of @xmath38 .
therefore , the critical set of @xmath544 in @xmath570 is included in @xmath571 . in what follows
, we choose @xmath254 with compact closure in @xmath44 and admitting a neighborhood in the set of regular values of @xmath38 .
[ [ case-1-rank-1-critical - points . ] ] _ case 1 : rank 1 critical points . _
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + there are 2 types of rank 1 critical points of @xmath27 : elliptic and hyperbolic . by the normal form theorem [ singularities_theorem ]
, there are canonical coordinates @xmath572 at in a chart about @xmath275 in which @xmath27 takes the form @xmath573 where @xmath246 is either @xmath341 ( elliptic case ) or @xmath574 ( hyperbolic case ) , and @xmath575 is a local diffeomorphism of a neighborhood of the origin to a neighborhood of @xmath576 , @xmath577 .
we see from this that @xmath578 is the 1-dimensional submanifold @xmath579 .
now consider the case when the level sets of @xmath38 in @xmath254 ( which are also 1-dimensional submanifolds of @xmath44 ) are transversal to this submanifold .
we see that the range of @xmath580 is directed along the first basis vector @xmath581 in @xmath44 , which is precisely tangent to @xmath582 .
hence @xmath267 can not vanish on this vector and hence @xmath583 this shows that @xmath544 has no critical points in @xmath570 .
now assume that there is a level set of @xmath38 in @xmath254 that is tangent to @xmath582 with non - degenerate contact at the point @xmath584 .
the tangency gives the equation @xmath585 .
since @xmath586 , the equation of @xmath587 is @xmath588 since @xmath589 on @xmath254 and @xmath51 is a local diffeomorphism , we have @xmath590 in a neighborhood of the origin .
but the contact equation gives @xmath591 so , taking @xmath254 small enough , we may assume that @xmath592 does not vanish .
hence the second condition in is equivalent to @xmath593 , which means @xmath594 ( and hence @xmath595 ) . by definition ,
the contact is non - degenerate if and only if the function @xmath596 has a non - degenerate critical point at @xmath534 .
therefore , by the implicit function theorem , the first equation @xmath597 has a unique solution @xmath598 .
thus , the critical set of @xmath544 is of the form @xmath599 , where @xmath600 is arbitrary in a small neighborhood of the origin ; this shows that the critical set of @xmath544 is a smooth 1-dimensional submanifold .
it remains to check that the hessian of @xmath544 is transversally non - degenerate .
of course , we take @xmath601 as transversal variables and we write the taylor expansion of @xmath544 , for any @xmath602 : @xmath603 we know that @xmath604 and , by the non - degeneracy of the contact , @xmath605 recalling that @xmath606 or @xmath607 , we see that the @xmath601-hessian of @xmath544 is indeed non - degenerate .
[ [ case-2-rank-0-critical - points . ] ] _ case 2 : rank 0 critical points . _ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + there are 4 types of rank 0 critical point of @xmath27 : elliptic - elliptic , focus - focus , hyperbolic - hyperbolic , and elliptic - hyperbolic , giving rise to four subcases . from the normal form of these singularities ( see theorem [ singularities_theorem ] , we see that all of them are isolated from each other .
thus , since @xmath27 is proper , the set of rank 0 critical points of @xmath27 is finite in @xmath570 .
again , let @xmath275 be a rank 0 critical point of @xmath27 and @xmath608 . a. _ elliptic - elliptic subcase . _
+ in the elliptic - elliptic case , the normal form is @xmath609 where @xmath610 .
the critical set of @xmath27 is the union of the planes @xmath611 and @xmath612 ( we use the notation @xmath613 ) .
the corresponding critical values in @xmath254 is the set @xmath614 ( the @xmath51-image of the closed positive quadrant ) .
+ the transversality assumption on @xmath38 amounts here to saying that the level sets of @xmath38 in a neighborhood of @xmath512 intersect @xmath582 transversally ; in other words , the level sets of @xmath615 intersect the boundary of the positive quadrant transversally . up to further shrinking of @xmath254 , this amounts to requiring @xmath616 , @xmath617 for all @xmath618 , where @xmath619 is the canonical @xmath44-basis . +
any critical point @xmath13 of @xmath27 different from @xmath275 is a rank 1 elliptic critical point .
since the level sets of @xmath38 do nt have any tangency with @xmath582 , we know from the rank 1 case above that @xmath13 can not be a critical point of @xmath544 .
hence @xmath275 is an isolated critical point for @xmath544 .
+ the hessian of @xmath544 at @xmath275 is calculated via the normal form : it has the form @xmath620 , with @xmath621 and @xmath622 .
the hessian determinant is @xmath623 .
the transversality assumption implies that both @xmath470 and @xmath300 are non - zero which means that the hessian is non - degenerate .
b. _ focus - focus subcase . _
+ the focus - focus critical point is isolated , so we just need to prove that the hessian of @xmath544 is non - degenerate . but
the 2-jet of @xmath544 is @xmath624 thus , in normal form coordinates ( see theorem [ singularities_theorem ] ) , as in the previous case , it has the form @xmath625 , where this time @xmath626 and @xmath627 are the focus - focus quadratic forms given in theorem [ singularities_theorem](iii ) .
the hessian determinant is now @xmath628 , which does not vanish . c. _ hyperbolic - hyperbolic subcase . _ + here the local model for the foliation is @xmath629 , @xmath630 .
however , the formulation @xmath631 may not hold ; this is a well - known problem for hyperbolic fibers .
nevertheless , on each of the 4 connected components of of @xmath632 , we have a diffeomorphism @xmath633 , @xmath634 such that @xmath635 .
these four diffeomorphisms agree up to a flat map at the origin ( which means that their taylor series at @xmath636 are all the same ) .
+ thus , the critical set of @xmath27 in these local coordinates is the union of the sets @xmath637 and @xmath638 : this is the union of the four coordinate hyperplanes in @xmath639 .
the corresponding set of critical values in @xmath254 is the image of the coordinate axes : @xmath640 where @xmath272 and @xmath355 both vary in a small neighborhood of the origin in @xmath526 .
+ for each @xmath641 we let @xmath642 .
as before , the transversality assumption says that the values @xmath643 and @xmath644 ( which do nt depend on @xmath645 at the origin of @xmath44 ) do nt vanish in @xmath254 .
thus , as in the elliptic - elliptic case , the level sets of @xmath38 do nt have any tangency with @xmath582 . hence no rank 1 critical point of @xmath27
can be a critical point of @xmath544 , which shows that @xmath275 is thus an isolated critical point of @xmath544 .
+ the hessian determinant of @xmath620 is again @xmath623 with @xmath646 and @xmath647 ; thus the hessian of @xmath544 at @xmath275 is non - degenerate .
d. _ hyperbolic - elliptic subcase . _ + we still argue as above .
however , the hessian determinant in this case is @xmath648 . summarizing
, we have proved that rank 0 critical points of @xmath27 correspond to isolated critical points of @xmath544 , all of them non - degenerate .
putting together the discussion in the rank 1 and 0 cases , we have shown that the critical set of @xmath544 consist of isolated non - degenerate critical points and isolated 1-dimensional submanifolds on which the hessian of @xmath544 is transversally non - degenerate .
this means that @xmath544 is a morse - bott function .
since we have calculated all the possible hessians , it is easy to compute the various indices that can occur .
we shall need a particular case , for which we introduce another condition on @xmath38 .
[ defi : outward ] let @xmath0 be a connected symplectic four - manifold .
let @xmath518 be an almost - toric system with critical value set @xmath43 .
a smooth curve @xmath45 in @xmath44 is said to have an _ outward contact _ with @xmath29 at a point @xmath649 when there is a small neighborhood of @xmath46 in which the point @xmath650 is the only intersection of @xmath45 with @xmath29 . in the proof below
we give a characterization in local coordinates .
[ prop : bott - index ] let @xmath0 be a connected symplectic four - manifold .
let @xmath518 be an almost - toric system with critical value set @xmath43 .
let @xmath38 be a morse function defined on an open neighborhood of @xmath651 such that 1 .
[ bott - index - hyp1 ] the critical set of @xmath38 is disjoint from @xmath43 ; 2 . [ bott - index - hyp2 ] @xmath38 has no saddle points in @xmath29 ; 3 .
[ bott - index - hyp3 ] the regular level sets of @xmath38 intersect @xmath43 transversally or have a non - degenerate outward contact with @xmath29 .
( see definitions [ defi : contact ] and [ defi : outward ] . )
then @xmath652 is a morse - bott function with all indices and co - indices equal to 0 , 2 , or 3 . because of theorem [ prop : bott ]
, we just need to prove the statement about the indices of @xmath38 . at points of @xmath553
, we saw in that the transversal hessian of @xmath540 is just the hessian of @xmath38 . by assumption
, @xmath38 has no saddle point , so its ( co)index is either @xmath74 or @xmath53 .
we analyze the various possibilities at points of @xmath653 .
there are two possible rank 0 cases for an almost - toric system : elliptic - elliptic and focus - focus . at such points ,
the hessian determinant is positive ( see theorem [ singularities_theorem ] ) , so the index and co - index are even . in the rank 1 case , for an almost - toric system ,
only transversally elliptic singularities are possible .
we are interested in the case of a tangency ( otherwise @xmath540 has no critical point ) .
the hessian is computed in and we use below the same notations .
the level set of @xmath38 through the tangency point is given by @xmath654 .
we switch to the coordinates @xmath655 , where the local image of @xmath27 is the half - space @xmath656 .
let @xmath657 the level set of @xmath38 is @xmath658 , and @xmath234 satisfies @xmath659 , @xmath660 , @xmath661 ( this is the non - degeneracy condition in definition [ defi : contact ] ) . by the implicit function theorem ,
the level set @xmath662 near the origin is the graph @xmath663 , where @xmath664 this level set has an outward contact if and only if @xmath665 or , equivalently , @xmath666 and @xmath667 have the same sign .
from we see that the index and coindex can only be 0 or 3 .
we conclude by proving the two theorems in the introduction . both will rely on the following result . in the statement below
, we use the stratified structure of the bifurcation set of a non - degenerate integrable system , as given by proposition [ prop : strata ] .
[ prop : j - connected ] let @xmath0 be a connected symplectic four - manifold .
let @xmath668 be an almost - toric system such that @xmath27 is proper .
denote by @xmath43 the bifurcation set of @xmath27 .
assume that there exists a diffeomorphism @xmath62 onto its image such that : 1 .
[ hyp : cone ] @xmath63 is included in a proper convex cone @xmath82 ( see figure [ fig : cone ] ) .
[ hyp : tangencies ] @xmath50 does not have vertical tangencies ( see figure [ fig : diffeo ] ) .
write @xmath669 .
then @xmath240 is a morse - bott function with connected level sets .
let @xmath670 .
the set of critical values of @xmath671 is @xmath672 .
we wish to apply proposition [ prop : bott - index ] to this new map @xmath671 .
let @xmath673 be the projection on the first coordinate : @xmath674 , so that @xmath675 . since @xmath38 has no critical points , it satisfies the hypotheses and of proposition [ prop : bott - index ] .
the regular levels sets of @xmath38 are the vertical lines , and the fact that @xmath676 has no vertical tangencies means that the regular level sets of @xmath38 intersect @xmath676 transversally . thus the last hypothesis of proposition [ prop : bott - index ] is fulfilled and we conclude that @xmath240 is a morse - bott function whose indices and co - indices are always different from 1 .
now , since @xmath671 is proper , the fact that @xmath677 is included in a cone @xmath55 easily implies that @xmath678 is proper .
thus , using proposition [ prop : levelset ] , we conclude that @xmath240 has connected level sets . by proposition [ prop :
j - connected ] , we get that @xmath240 has connected level sets .
it is enough to apply theorem [ theo : fibers ] to conclude that @xmath671 ( and thus @xmath27 ) has connected fibers . using again proposition [ prop : j - connected ] ,
we conclude that @xmath240 is a morse - bott function with connected level sets . by the definition of an integrable system
, @xmath240 can not be constant ( its differential would vanish everywhere )
. thus we can apply theorem [ imagemomentum : theorem ] , which yields the desired conclusion .
it turns out that even in the compact case , theorem [ theo : main - connectivity ] has quite a striking corollary , which we stated as theorem [ theo : striking ] in the introduction .
the last two cases ( a disk with two conic points and a polygon ) can be transformed by a diffeomorphism as in theorem [ theo : main - connectivity0 ] to remove vertical tangencies , and hence the theorem implies that the fibers of @xmath27 are connected .
for the first two cases , we follow the line of the proof of theorem [ theo : main - connectivity0 ] .
the use of proposition [ prop : bott - index ] is still valid for the same function @xmath674 even if now the level sets of @xmath38 can be tangent to @xmath43 . indeed one can check that here only non - degenerate outward contacts occur . then one can bypass proposition [ prop : j - connected ] and directly apply proposition [ prop : levelset ] .
therefore the conclusion of the theorem still holds .
the goal of this section is to prove that the spherical pendulum is a non - degenerate integrable system to which our theorems apply .
the configuration space is @xmath20 .
we identify the phase space @xmath679 with the tangent bundle @xmath680 using the standard riemannian metric on @xmath20 naturally induced by the inner product on @xmath681 .
denote the points in @xmath681 by @xmath682 .
the conjugate momenta are denoted by @xmath683 and hence the canonical one- and two - forms are @xmath684 and @xmath685 , respectively .
the manifold @xmath686 has its own natural exact symplectic structure .
it is easy to see that @xmath687 is a symplectic embedding since @xmath688 coincides with the canonical one - form on @xmath680 .
the action given by rotations about the @xmath689-axis is given by @xmath690 the equations of motion determined by the infinitesimal generator of the lie algebra element @xmath691 for the lifted action to @xmath692 are @xmath693 the @xmath367-invariant momentum map of this action is given by ( see , e.g. , ( * ? ? ? * theorem 12.1.4 ) ) @xmath694 .
note that @xmath680 is invariant under the flow of and thus @xmath695 given for all @xmath696 is the momentum map of the @xmath367-action on @xmath680 .
in particular , the equations of motion of the hamiltonian vector field @xmath697 on @xmath680 are .
we also have @xmath698 ; indeed , if @xmath699 and @xmath700 , then @xmath701 is an arbitrary element of @xmath702 .
the momentum map @xmath240 is _ not _ proper : the sequence @xmath703 does not contain any convergent subsequence and the sequence of images @xmath704 is constant , hence convergent .
let us describe the equations of motion .
the hamiltonian of the spherical pendulum is @xmath705 where @xmath706 is the standard orthonormal basis of @xmath681 with @xmath689 aligned with and pointing opposite the direction of gravity and we set all parametric constants equal to one .
the equations of motion for @xmath707 are given by @xmath708 [ sphericalpendulum : theorem ] let @xmath709 be the singular fibration associated with the spherical pendulum . * @xmath27 is an integrable system .
* the singularities of @xmath27 are non - degenerate . *
the singularities of @xmath27 are of focus - focus , elliptic - elliptic , or transversally elliptic - type so , in particular , @xmath27 has no hyperbolic singularities .
there is precisely one elliptic - elliptic singularity at @xmath710 , one focus - focus singularity at @xmath711 , and uncountably many singularities of transversally - elliptic type .
* @xmath241 is proper and hence @xmath27 is also proper , even though @xmath240 is _ not _ proper . *
the critical set of @xmath27 and the bifurcation set of @xmath27 are equal and given in figure [ critical_set_spherical_pendulum.figure ] : it consists of the boundary of the planar region therein depicted and the interior point which corresponds to the image of the only focus - focus point of the system ; see point ( 3 ) above . *
the fibers of @xmath27 are connected . *
the range of @xmath27 is equal to planar region in figure [ critical_set_spherical_pendulum.figure ] .
the image under @xmath27 of the focus - singularity is the point @xmath712 .
the image under @xmath27 of the elliptic - elliptic singularity is the point @xmath713 .
we prove each item separately .
[ [ integrability . ] ] _ ( 1 ) integrability . _ + + + + + + + + + + + + + + + + + + + + since the hamiltonian @xmath241 is @xmath367-invariant , the associated momentum map @xmath714 is conserved , i.e. , @xmath715 .
to prove integrability , we need to show that @xmath716 and @xmath717 are linearly independent almost everywhere on @xmath680 .
note that if @xmath718 , the one - forms @xmath719 and @xmath720 are linearly dependent precisely when the vector fields and are linearly dependent .
thus we need to determine all @xmath721 such that there exist @xmath722 , not both zero , satisfying @xmath723 , @xmath724 , @xmath725 , @xmath726 .
a computation gives that the set of points @xmath696 for which @xmath716 and @xmath717 are linearly dependent is the measure zero set : @xmath727 [ [ non - degeneracy . ] ] _ ( 2 ) non - degeneracy . _ + + + + + + + + + + + + + + + + + + + + + all critical points of @xmath728 are non - degenerate .
we prove it for rank 1 critical points .
the non - degeneracy at the rank 0 critical points @xmath711 and @xmath710 are similar exercices ( the computation for the point @xmath729 may also be found in @xcite ) .
so we consider the singularities @xmath730 given by @xmath731 , @xmath732 , where @xmath733 ( and hence @xmath734 ) .
we have @xmath735 and @xmath736 .
as explained at the beginning of the section , the component @xmath240 is a momentum map for a hamiltonian @xmath367-action .
the flow of @xmath240 rotates about the @xmath689 axis , so any vertical plane is transversal to the flow in the @xmath682-coordinates .
the surface @xmath500 obtained by intersecting the level set of @xmath240 with the plane @xmath737 is chosen as the symplectic transversal surface to the flow of @xmath738 .
the restriction of @xmath739 to @xmath500 is @xmath740 , so that @xmath741 is constant on @xmath500 .
we work near @xmath742 , so @xmath743 , and hence on @xmath500 we have @xmath744 so @xmath500 may be smoothly parametrized by the coordinates @xmath745 using .
let us compute the hessian of @xmath241 at @xmath746 in these coordinates .
the taylor expansion of @xmath241 at @xmath746 is @xmath747 by we get @xmath748 and hence the critical point is @xmath749 to simplify notation , let us denote @xmath750 $ ] ; thus @xmath751 are the smooth local coordinates near @xmath752 and the expression on the hamiltonian is @xmath753 the first term in is equal to @xmath754 while the taylor expansion of the second term @xmath755 with respect to @xmath756 is @xmath757 the coefficient @xmath758 vanishes since @xmath759 , so the hessian @xmath760 of @xmath761 is of the form @xmath762 , where @xmath763 and @xmath764 , which is non - degenerate , as we wanted to show . [ [ the - nature - of - the - singularities - of - f . ] ] _ ( 3 ) the nature of the singularities of @xmath27 . _ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the statement in the theorem follows from the computations in ( 2 ) .
the equilibrium @xmath711 is of focus - focus type and the equilibrium @xmath710 is of elliptic - elliptic type .
the other critical points are of transversally - elliptic type .
[ [ h - and - f - are - proper - maps . ] ] _ ( 4 ) @xmath241 and @xmath27 are proper maps . _
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the properness of @xmath241 follows directly from the defining formula ; indeed , on @xmath680 , the map @xmath765 takes values in a compact set , hence if @xmath766 is compact , there is a compact set @xmath767 such that @xmath768 is closed in @xmath769 and hence is compact .
thus @xmath241 is proper .
then for any compact set @xmath770 , @xmath771 is compact , thus @xmath27 is proper as well .
( @xmath772 is the projection on the second factor ) . on the other hand
it is clear that the level sets of @xmath240 are unbounded , so @xmath240 can not be proper . [ [ critical - and - bifurcation - sets . ] ] _ ( 5 ) critical and bifurcation sets . _ +
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the critical set is the image of by the map @xmath773 .
we have @xmath774 , @xmath775 .
in addition , @xmath776 for @xmath777 which is a parametric curve with two branches .
we can give an cartesian equation for this curve .
an analysis of @xmath778 for @xmath779 shows that @xmath780 . eliminating @xmath781 yields the two branches @xmath782 for @xmath783 .
the critical set is given in figure [ critical_set_spherical_pendulum.figure ] .
the graph intersects the horizontal momentum axis at @xmath784{3}/3 $ ] and at the lower tip @xmath785 the graph is not smooth .
[ h ] is its boundary plus the focus - focus value @xmath786.,title="fig:",width=151 ] since @xmath27 is proper , the bifurcation set equals the set of critical values of the system , see figure [ critical_set_spherical_pendulum.figure ] .
[ [ the - fibers - of - f - are - connected . ] ] _ ( 6 ) the fibers of @xmath27 are connected . _
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + this follows from item ( 5 ) and theorem [ theo : fibers ] .
compare this with the result ( * ? ? ?
* , table 3.2 ) whose proof is quite difficult ( however , it also gives the description of the fibers of @xmath27 ) .
[ [ range - of - the - momentum - energy - set . ] ] _ ( 7 ) range of the momentum - energy set . _
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the range of the momentum - energy set is the epigraph of the critical set shown in figure [ critical_set_spherical_pendulum.figure ] : this follows from item in theorem [ theo : interior ] .
to motivate further our connectivity results ( theorem [ theo : main - connectivity0 ] and theorem [ theo : main - connectivity ] ) , consider on the real plane with coordinates @xmath270 the following question : _ is the solution set of the polynomial equation @xmath787 connected ?
_ surely the answer is yes , since the solution set consists of two parallel vertical lines and an intersecting horizontal line : @xmath788 .
one can modify this equation slightly to consider the equation @xmath789 , @xmath790 . in this case
the solution set @xmath791 is disconnected , so a small perturbation of the original equation leads to a disconnected solution set ( see figure [ fig : solution ] ) . as is well known in real algebraic geometry ,
the connectivity question is not stable under small perturbations , so any technique to detect connectivity must be sensitive to this issue .
this question may be put in a general framework in two alternative , equivalent , ways .
consider functions @xmath794 and constants @xmath795 , where @xmath11 is a connected manifold .
is the solution set @xmath796 of @xmath797 a connected subset of @xmath11 ?
equivalently , are the fibers @xmath798 of the map @xmath799 defined by @xmath800 connected ? if @xmath27 is a scalar valued function ( i.e. , if there is only one equation in the system ) a well - known general method exists to answer it when @xmath27 is smooth , namely , _
morse - bott theory_. we saw in proposition [ prop : levelset ] that if @xmath11 is a connected smooth manifold and @xmath202 is a proper morse - bott function whose indices and co - indices are always different from 1 , then the level sets of @xmath38 are connected . in order to motivate the idea of this result further consider the following example : the height function defined on a @xmath53-sphere , and the same height function defined on a @xmath53-sphere in which the north pole is pushed down creating two additional maximum points and a saddle point , as in figure [ fig : morse ] . when the height function is considered on the @xmath53-sphere , it has connected fibers . on the other hand , when it is considered on the right figure , many of its fibers are disconnected .
the `` essential '' difference between these two examples is that in the second case the function has a saddle point , while in the first case it does nt .
a saddle point has index 1 .
next , let us look at an example of a vector valued function .
consider @xmath801 with coordinates @xmath802 .
is the solution set @xmath803 of @xmath804 connected ?
in other words , is the set @xmath805 connected , where @xmath806 defined by @xmath807 one is tempted to again use morse - bott theory to check this connectivity , but morse - bott theory for which function ? _ the first goal of the present paper was to give a method to answer connectivity questions of this type in the case @xmath27 is an integrable system .
_ in the paper we introduced a method to construct morse - bott functions which , from the point of view of symplectic geometry , behave well near singularities .
we saw how the behavior of an integrable system near the singularities has a strong effect on the global properties of the system .
the fibers of @xmath808 are , by the theorem [ theo : main - connectivity ] , connected , because its image can be easily seen to be given by example [ fig : example ] .
the question answered by theorem [ theo : main - connectivity ] is probably the most basic _ topological _ question to ask about a solution set ( except for its non - emptyness ) .
semiclassical quantization is a strong motivation for undergoing the systematic study of integrable systems in this paper .
consider the quantum version of the connectivity problem .
the functions @xmath809 are replaced by `` quantum observables '' , i.e. , self - adjoint operators @xmath810 on a hilbert space @xmath811 , and the system of fiber equations is replaced by the system @xmath812 so @xmath813 should be an eigenvalue for @xmath814 and the eigenvector @xmath264 should be the same for all @xmath815 . there is a chance to solve this when the operators @xmath814 pairwise commute .
this is the quantum analogue of the poisson commutation property for the integrable system given by @xmath809 . in the last thirty years
, semiclassical analysis has pushed this idea quite far .
it is known that to any regular liouville torus of the classical integrable system , one can associate a _
quasimode _ for the quantum system , leading to approximate eigenvalues @xcite .
thus , the study of bifurcation sets ( described in theorem [ theo : main - image0 ] and theorem [ theo : main - image ] ) is fundamental for a good understanding of the quantum spectrum .
more recently , quasimodes associated to singular fibers have been constructed ; see @xcite and the references therein . in some case , it can be shown that the quantum spectrum completely determines the classical system , see zelditch @xcite .
when the system has symmetries , the quantum spectrum typically exhibits degenerate ( or almost degenerate ) eigenvalues . in the setting of integrable systems ,
this is revealed by the non - connectedness of some fibers .
thus , connectivity results are of primary importance from the quantum perspective .
they predict almost degenerate ( or _ clusters of _ ) eigenvalues .
for instance , in the inverse spectral result of @xcite , detecting the number of connected components was the key point ( and the most subtle ) of the analysis .
theorem [ theo : main - connectivity0 ] and theorem [ theo : main - connectivity ] may have applications to mirror symmetry and symplectic topology .
an integrable system without hyperbolic singularities gives rise to a toric fibration with singularities .
the base space is endowed with a singular integral affine structure .
remarkably , these singular affine structures are of key importance in various parts of symplectic topology , mirror symmetry , and algebraic geometry ; for example they play a central role in the work of kontsevich and soibelman @xcite .
these singular affine structures have been studied in the context of integrable systems ( in particular by nguyn tin zung @xcite ) , but also became a central concept in the works by symington @xcite and symington - leung @xcite in the context of symplectic geometry and topology , and by gross - siebert , castao - bernard , castao - bernard - matessi @xcite , among others , in the context of mirror symmetry and algebraic geometry .
fiber connectivity is ( usually ) assumed in theorems about lagrangian fibrations , so the theorem above gives a method to test whether a given result in that context applies to a lagrangian fibration arising from an integrable system .
see also eliashberg - polterovich @xcite for relevant examples in the theory of symplectic quasi - states .
non - degenerate integrable systems without hyperbolic singularities are prominent in the literature , both in mathematics and in physics ; see for instance @xcite .
they are usually called _ almost - toric systems _ , although the choice of name is somewhat misleading as they may not have any periodicity from a circle or torus action ; on the other hand , these systems retain some properties of toric systems , at least in the compact case .
the most well studied case of almost - toric system is that of toric systems coming from a hamiltonian torus actions .
a proof of theorem [ theo : main - connectivity ] for the so called `` semitoric '' systems was given by the third author in @xcite using the theory of hamiltonian circle actions .
semitoric systems are integrable systems on four - manifolds which have no hyperbolic singularities and for which one component of the system generates a proper periodic flow .
the proof in @xcite uses in an essential way the periodicity assumption and it can not be extended , as far as we know , to deal with the general systems treated in theorem [ theo : main - connectivity ] . a classification of semitoric systems in terms of five symplectic invariants
was given by the first and third authors in @xcite .
abraham , r. and marsden , j.e . : _ foundation of mechanics _ , second edition , revised and enlarged .
with the assistance of tudor ratiu and richard cushman .
benjamin / cummings publishing co. , inc .
, advanced book program , reading , mass .
, 1978 .
child , m. s. , weston , t. , and tennyson , j. : quantum monodromy in the spectrum of @xmath817 and other systems : new insight into the level structure of quasi - linear molecules , _ mol .
_ , * 96*(3 ) ( 1999 ) , 371379 .
fitch , n. j. , weidner , c. a. , parazzoli , l. p. , dullin , h. r. , and lewandowski , h. j. : experimental demonstration of classical hamiltonian monodromy in the 1:1:2 resonant elastic pendulum , _ phys . rev .
_ , * 103 * ( 2009 ) , 034301 .
fomenko , a. t. the topology of surfaces of constant energy of integrable hamiltonian systems and obstructions to integrability .
( russian ) _ izv .
nauk sssr ser
* 50 * ( 1986 ) , no . 6 , 12761307 , 1344 .
lazutkin , v. f. : _ kam theory and semiclassical approximations to eigenfunctions_. with an addendum by a. i. shnirelman .
ergebnisse der mathematik und ihrer grenzgebiete ( 3 ) , * 24*. springer - verlag , berlin , 1993 .
symington , m. : four dimensions from two in symplectic topology , _ topology and geometry of manifolds ( athens , ga , 2001 ) _ , proc .
pure math .
, * 71 * , 153208 , amer . math . soc . ,
providence , ri , 2003 .
+ lvaro pelayo + school of mathematics + institute for advanced study + einstein drive + princeton , nj 08540 usa .
+ + and + + washington university , mathematics department + one brookings drive , campus box 1146 + st louis , mo 63130 - 4899 , usa .
+ _ e - mail _ : ` [email protected] ` v ngoc san + institut universitaire de france + + and + + institut de recherches mathmatiques de rennes + universit de rennes 1 + campus de beaulieu + f-35042 rennes cedex , france + _ e - mail : _ ` [email protected] ` | this paper develops a symplectic bifurcation theory for integrable systems in dimension four . we prove that if an integrable system has no hyperbolic singularities and its bifurcation diagram has no vertical tangencies , then the fibers of the induced singular lagrangian fibration are connected . the image of this singular lagrangian fibration is , up to smooth deformations , a planar region bounded by the graphs of two continuous functions .
the bifurcation diagram consists of the boundary points in this image plus a countable collection of rank zero singularities , which are contained in the interior of the image .
because it recently has become clear to the mathematics and mathematical physics communities that the bifurcation diagram of an integrable system provides the best framework to study symplectic invariants , this paper provides a setting for studying quantization questions , and spectral theory of quantum integrable systems . |
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.